Richard C.

Benson
Lynn B. Yuhr

Site Characterization
in Karst and
Pseudokarst Terraines
Practical Strategies and Technology
for Practicing Engineers,
Hydrologists and Geologists
Site Characterization in Karst and Pseudokarst
Terraines
Richard C. Benson • Lynn B. Yuhr

Site Characterization in Karst

and Pseudokarst Terraines
Practical Strategies and Technology
for Practicing Engineers, Hydrologists
and Geologists
Richard C. Benson Lynn B. Yuhr
Technos, Inc. Technos, Inc.
Miami, FL, USA Miami, FL, USA

ISBN 978-94-017-9923-2 ISBN 978-94-017-9924-9 (eBook)

DOI 10.1007/978-94-017-9924-9
Library of Congress Control Number: 2015948377

Springer Dordrecht Heidelberg New York London

© Springer Science+Business Media Dordrecht 2016
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction
on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation,
computer software, or by similar or dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and
regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed
to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty,
express or implied, with respect to the material contained herein or for any errors or omissions that may have been
made.

Printed on acid-free paper

Springer Science+Business Media B.V. Dordrecht is part of Springer Science+Business Media (www.springer.com)
Preface

The purpose of this book is to present both an insight into the complexities of karst and pseu-
dokarst conditions and a broad strategy and practical approach needed to carry out an effective
site characterization within these or any other complex geologic conditions. Our intent is to
provide a basic understanding of these topics and introduce the vast array of tools available to
solve the geologic puzzle. We believe that the geology is the key to a site characterization,
since the geology is what controls the engineering and hydrologic conditions at a site. If the
geology is properly understood, then all subsequent work can be done with more confidence,
including design, construction, contaminant assessment, remediation, modeling, and long-
term operation.
This book is divided into three parts:

• Part I provides a brief introduction to some of the many karst and pseudokarst conditions
that may be encountered. It provides a brief overview of karst development, various condi-
tions that may be present, and their wide range of scale. While there are many benefits to a
karst landscape (groundwater resources, minerals, and recreation), our focus is on the dam-
aging impacts to both engineered structures and groundwater resources. By developing an
understanding of the nature of karst and appreciating its potential impact, we have a starting
point for an effective site characterization effort.
• Part II is the core of the book and presents a strategy to carry out an effective site character-
ization. The strategy emphasizes the use of a wide range of measurements and technology,
over a range of scales. We introduce many of the off-the-shelf technologies that are pres-
ently available to help characterize karst conditions. This section includes numerous exam-
ples and mini-case histories to illustrate the strategies and methods presented.
• Part III presents three site characterization case histories that include:
– A landfill that was being developed over an abandoned limestone mine in the Kansas
City area, which had experienced a major mine-roof collapse due to a deep-seated paleo-
collapse feature
– A geotechnical assessment for a new bridge into the Florida Keys in an area of suspected
karst
– An EPA Superfund site in west central Florida that required assessment of geologic,
hydrologic, and karst conditions prior to initiating remediation efforts

The Authors

This book is, for the most part, based upon the authors’ combined experience of more than
80 years as consulting geologists. The authors are a father and daughter team, who has worked
together since 1978 providing consulting services through their company Technos, Inc. Richard
Benson pioneered many of the early applications of various geophysical methods in the 1960’s
and 1970’s. He founded Technos Inc. in 1971 specializing in applied earth sciences. Their

v
vi Preface

hands-on experience in site characterization, with both karst and pseudokarst, is the basis for
the strategies and array of methods presented in this book. This personal experience is repre-
sented by the many technical examples included in the book. These examples are from sites in
the USA, the Bahamas, much of the Caribbean, and Guam.
The authors have developed all tables, diagrams, and photos, unless specified. Most of the
figures shown are simplified. We recognize that the real-world, especially in karst and psuedo-
karst, is much more complex. In many cases, the examples do not contain the specific location
or name of the owner in order to avoid concerns of liability.

Miami, FL, USA Richard C. Benson

Lynn B. Yuhr
Acknowledgments

We dedicate this book to our families and thank them for their unconditional patience and sup-
port. With a special dedication to Bonnie Benson (wife of Richard), Eric Yuhr (husband of
Lynn) and Marisol Sanchez (adopted family member and office manager for over 20 years). To
our friends and colleagues, we thank you for the continuous positive nods, pats on the back,
and review of chapters that were often incomplete and incoherent. To our publisher, thank you
Petra and Hermine for your patience, support, and allowing us to check one off of our bucket list.
A special thank you is extended to our reviewers and contributors who include:

David Carrier, Argila Enterprises

Charles Finkl, Professor Emeritus of Geosciences, Florida Atlantic University, and editor in
chief of the Journal of Coastal Research
Thomas Aley, Ozark Laboratories
Tony Waltham, Geophotos, www.geophotos.co.uk
Chris Anderson, www.darklightimagery.net
Jim Lococo, Mount Sopris
Peter Annan and Steve Cosway, Sensors & Software, Ltd.
Dwain Butler, (retired) US Army Waterway Experiment Station
Peter Fenning, (retired) consulting geophysicist in UK
Devraj Sharma, founder Principia Mathematica
Gary Olhoeft, (retired) Colorado School of Mines
Paul Beam, Dept. of Energy
Richard Woods, University of Michigan
Lewis Land, NCKRI
George Veni, NCKRI

Authors can be contacted at:

Richard C. Benson, CPG, PG

[email protected]
Lynn B. Yuhr, PG
[email protected]

vii
Contents

Part I A Brief Overview Karst and Pseudokarst

1 Some Common Terms .............................................................................................. 3
1.1 Karst .............................................................................................................. 3
1.2 Paleokarst ...................................................................................................... 3
1.3 Pseudokarst ................................................................................................... 4
References .................................................................................................................. 4
2 The Development of Karst Conditions ................................................................... 7
2.1 Carbonates and Other Soluble Rock ............................................................. 7
2.2 Post Deposition Processes ............................................................................. 7
2.2.1 Dissolution of Limestone ............................................................. 7
2.2.2 Dissolution of Other Soluble Rocks ............................................ 10
2.2.3 Dissolution of “Non-soluble Rocks” ........................................... 10
2.2.4 Mechanical Erosion and Transport .............................................. 10
2.2.5 Geomorphology ........................................................................... 10
2.3 Some Properties of Karst Rock ..................................................................... 13
References .................................................................................................................. 14
3 Types of Karst Features ........................................................................................... 17
3.1 Sinkholes ....................................................................................................... 17
3.1.1 A Wide Range of Sizes ................................................................ 18
3.1.2 Sinkhole Densities and Linear Trends ......................................... 22
3.1.3 Sinkhole Susceptibility Maps and Databases .............................. 23
3.2 Sinking Streams and Springs ........................................................................ 23
3.2.1 Springs in Florida......................................................................... 25
3.3 The Epikarst Zone ......................................................................................... 27
3.4 Caves ............................................................................................................. 29
3.4.1 Cave Geometry and Densities ...................................................... 30
3.4.2 Large Cave Systems Develop in Thick Massive Limestone ........ 31
3.4.3 Other Types of Caves ................................................................... 31
3.4.4 Secrecy and Discretion as a Cave Management Tool .................. 32
References .................................................................................................................. 32
4 Karst Maturity and Development........................................................................... 35
4.1 Karst Maturity ............................................................................................... 35
4.2 Karst Development Time Scale ..................................................................... 35
References .................................................................................................................. 40
5 Areas Affected by Karst and Pseudokarst ............................................................. 41
5.1 United States ................................................................................................. 41
5.2 Worldwide ..................................................................................................... 41
References .................................................................................................................. 43

ix
x Contents

6 Karst and Its Many Benefits ................................................................................... 45

6.1 Springs........................................................................................................... 45
6.2 Caves ............................................................................................................. 46
6.3 Sinkholes ....................................................................................................... 46
6.4 Karst Aquifers and Groundwater Resources ................................................. 46
6.4.1 The Edwards Aquifer ................................................................... 46
6.4.2 The Floridan Aquifer.................................................................... 46
6.5 Mineral Resources......................................................................................... 47
References .................................................................................................................. 47
7 Karst and Its Damaging Impact ............................................................................. 49
7.1 Structural Impacts ......................................................................................... 49
7.1.1 Private Residences ....................................................................... 49
7.1.2 Surface Water Management System ............................................ 50
7.1.3 Sinkholes on an Airport Runway ................................................. 50
7.1.4 Multiple Collapse at a Housing Development ............................. 50
7.1.5 Seepage and Collapse at Dams .................................................... 50
7.1.6 Elevated Expressway Failure ....................................................... 52
7.1.7 Deep Unknown Paleokarst ........................................................... 54
7.1.8 Excessive Grout Quantities .......................................................... 54
7.1.9 Reactivation of Sinkholes and the Drainage
of Sinkhole Lakes ........................................................................ 54
7.1.10 Problems with Man-Made Lakes ................................................. 55
7.2 Groundwater Contamination ......................................................................... 55
7.2.1 Mining Wastes.............................................................................. 55
7.2.2 Regional Contamination: The Woodville Karst Plain (WKP) ..... 57
7.3 Pseudokarst Impacts (Natural and Man-Made)............................................. 58
7.3.1 Naturally Occurring Pseudokarst ................................................. 58
7.3.2 Smaller Man-Induced Pseudokarst .............................................. 59
7.3.3 Larger Man-Induced Pseudokarst ................................................ 60
References .................................................................................................................. 64
8 Triggering Mechanisms for Sinkholes .................................................................... 67
8.1 Statistics ........................................................................................................ 67
8.2 Water-Related Triggering Mechanisms ......................................................... 67
8.2.1 Changes in Water Levels .............................................................. 68
8.2.2 Pumping ....................................................................................... 68
8.2.3 Dewatering ................................................................................... 68
8.2.4 Surface Water Run-Off................................................................. 68
8.2.5 Leaky Water Pipes and Sewers .................................................... 69
8.3 A Guideline to Minimize Sinkholes Triggered by Water .............................. 70
8.4 Other Triggering Mechanisms....................................................................... 70
8.4.1 Drilling Operations ...................................................................... 70
8.4.2 Impact of Vibrations, Blasting and Earthquakes .......................... 71
8.5 Size and Rate of Sinkhole Collapse .............................................................. 71
8.5.1 The Size of a Sinkhole ................................................................. 71
8.5.2 The Speed of a Sinkhole Collapse ............................................... 71
References .................................................................................................................. 73
9 Cave and Cavern Collapse....................................................................................... 75
9.1 Breakdown Domes ........................................................................................ 75
9.1.1 A Conceptual Model of a Large Cavern ...................................... 75
9.2 Mechanics of Cavern Breakdown ................................................................. 78
9.2.1 Two Modes of Breakdown: Fixed Beam and Cantilever Beam ... 78
9.3 Thickness of Rock Needed to Prevent Surface Subsidence or Collapse....... 78
Contents xi

9.4 Experience from Mine Failures ..................................................................... 79

9.4.1 Bulking of Fallen Roof Rock ....................................................... 81
9.5 Propagation of Subsidence and Collapse from Great Depths ....................... 82
References .................................................................................................................. 84
10 Insight into the Nature of Cover Collapse Sinkholes ............................................ 87
10.1 Introduction ................................................................................................... 87
10.2 Insight from Scale Model Sinkhole Tests ..................................................... 87
10.3 Insight from Mine Backfill Stabilization....................................................... 88
10.3.1 Comparison Between Sinkholes and Mine Backfilling ............... 88
10.4 Conceptual Models of Cover Collapse Sinkholes ......................................... 90
10.4.1 Small Cover Collapse Sinkholes .................................................. 90
10.4.2 Intermediate Cover Collapse Sinkholes ....................................... 91
10.4.3 Large Cover Collapse Sinkholes .................................................. 91
10.4.4 Very Large Cover Collapse Sinkholes ......................................... 91
10.4.5 Extremely Large Cover Collapse Sinkholes ................................ 92
10.4.6 Conceptual Models and Their Limitations................................... 92
References .................................................................................................................. 95

Part II The Strategy and Methods for Site Characterization

11 What Is Site Characterization................................................................................. 99
11.1 Introduction ................................................................................................... 99
11.2 Uncertainties in Site Characterization........................................................... 100
11.3 The Technical Literature ............................................................................... 100
11.4 Concepts and Strategies for Site Characterization by Others ....................... 101
11.5 The Site Characterization Team .................................................................... 103
11.6 Some Pitfalls of Site Characterization .......................................................... 104
11.6.1 Lack of Interdisciplinary Approach ............................................. 105
11.6.2 Impact of Computers.................................................................... 105
References .................................................................................................................. 106
12 The Strategy.............................................................................................................. 107
12.1 The Detection Dilemma ................................................................................ 107
12.1.1 Direct Detection ........................................................................... 108
12.1.2 Indirect Detection......................................................................... 109
12.1.3 Statistical Approach ..................................................................... 109
12.2 Appropriate, Adequate and Accurate Data .................................................... 109
12.2.1 Appropriate Data .......................................................................... 110
12.2.2 Adequate Data Density ................................................................ 113
12.2.3 Accuracy of Data ......................................................................... 115
12.3 Key Steps in the Site Characterization Process ............................................. 118
12.3.1 Project Preparation ....................................................................... 118
12.3.2 The Conceptual Model................................................................. 120
12.3.3 In the Office ................................................................................. 121
12.3.4 The Field Effort............................................................................ 121
12.3.5 Conversion of Data to Useful Information................................... 122
12.3.6 Additional Studies........................................................................ 122
12.4 Summary ....................................................................................................... 122
References .................................................................................................................. 123
13 The Desk Study......................................................................................................... 125
13.1 What We Know and Don’t Know.................................................................. 125
13.2 Sources of Existing Information and Data .................................................... 125
13.3 Type of Data Available .................................................................................. 127
xii Contents

13.3.1 Topographic and Geologic Maps ................................................. 127

13.3.2 Aerial Photos................................................................................ 127
13.3.3 Geologic and Hydrologic Reports ............................................... 128
13.3.4 Sinkhole Databases ...................................................................... 129
13.3.5 Cave Maps and Databases............................................................ 129
13.3.6 Databases of Borings, Water Wells, and Monitor Wells .............. 129
13.3.7 Cultural Features and Changes .................................................... 129
13.3.8 Anecdotal Information ................................................................. 131
13.4 Data Mining and Review............................................................................... 131
13.5 The Preliminary Conceptual Model .............................................................. 132
References .................................................................................................................. 132
14 Aerial Photography and Remote Sensing Data ..................................................... 133
14.1 Availability .................................................................................................... 133
14.2 Scale .............................................................................................................. 134
14.3 Coverage........................................................................................................ 134
14.4 Aerial Photos ................................................................................................. 134
14.4.1 Fracture Trace and Lineament Analysis....................................... 135
14.4.2 Oblique Aerial Photos .................................................................. 137
14.4.3 Aerial Photos and Video from Small Unmanned Aircraft ........... 137
14.5 Beyond Black and White Aerial Photos (Other Formats and Methods) ....... 140
14.5.1 Aerial Thermography ................................................................... 140
14.5.2 InSAR .......................................................................................... 140
14.5.3 LiDAR .......................................................................................... 141
References .................................................................................................................. 142
15 Site Walkover ............................................................................................................ 145
15.1 The Initial Site Walkover............................................................................... 145
15.2 Importance of Observations .......................................................................... 145
15.3 Some Tools for the Field ............................................................................... 146
15.4 On-Site Walkovers and Off-Site Drives ........................................................ 148
15.5 Site Coverage ................................................................................................ 148
15.6 Observations and Mapping............................................................................ 148
15.6.1 Geologic Observations ................................................................. 148
15.6.2 Hydrologic Observations ............................................................. 150
15.6.3 Inventory of Karst Features .......................................................... 153
15.6.4 Indications of Subsidence and Sinkhole Activity ........................ 154
15.6.5 Cultural Factors, Utilities and Other Site Limitations ................. 156
15.6.6 Eye Witness and Anecdotal Information...................................... 157
15.7 Fly Over ........................................................................................................ 158
15.8 Updating the Conceptual Model ................................................................... 158
15.9 Updating the Work Plan ................................................................................ 158
References .................................................................................................................. 159
16 Surface Geophysical Methods ................................................................................. 161
16.1 Introduction ................................................................................................... 161
16.2 A Brief History of the Surface Geophysical Methods .................................. 162
16.3 An Overview of Surface Geophysics ............................................................ 163
16.3.1 Parameter Measured..................................................................... 167
16.3.2 Anomalies .................................................................................... 167
16.3.3 Direct and Indirect Detection of Anomalous Conditions ............. 168
16.3.4 Penetration of Measurements ....................................................... 168
16.3.5 Resolution .................................................................................... 168
16.3.6 Processing and Presentation of Data ............................................ 170
Contents xiii

16.3.7 Use of a Survey Grid .................................................................... 171

16.3.8 Location of Utilities, Buried Drums, Tanks and Trash ................ 172
16.3.9 Ease of Use .................................................................................. 173
16.3.10 Surface Geophysical Data Can Be Acquired Over Water ............ 173
16.3.11 Limitations ................................................................................... 173
16.4 Guidelines for the Selection of the Surface Geophysical Methods............... 173
16.5 Application of Surface Geophysical Methods .............................................. 175
16.5.1 Soil Piping and Collapse Within the Sediments .......................... 176
16.5.2 Conditions Within the Epikarst, Top of Rock or Rockhead ......... 181
16.5.3 Fractures and Cavities Within the Rock ....................................... 184
16.5.4 Buried Sinkholes and Paleokarst ................................................. 190
16.5.5 Pseudokarst Conditions................................................................ 195
References .................................................................................................................. 198
17 Invasive Methods ...................................................................................................... 201
17.1 Introduction ................................................................................................... 201
17.2 Direct Push Methods ..................................................................................... 201
17.2.1 Percussion Driven Direct Push Methods ..................................... 202
17.2.2 Cone Penetrometer Testing (CPT) ............................................... 204
17.3 Borings .......................................................................................................... 207
17.3.1 A Drilling Plan ............................................................................. 209
17.3.2 Drilling Methods .......................................................................... 209
17.3.3 Indications of Karst When Drilling .............................................. 213
17.3.4 Special Considerations When Drilling or Using Drilling Data.... 214
17.3.5 An Optimum Approach for Drilling and Sampling ..................... 221
17.4 Excavations and Trenches ............................................................................. 222
References .................................................................................................................. 225
18 Geophysical Logging ................................................................................................ 227
18.1 Introduction ................................................................................................... 227
18.2 Geophysical Logging Measurements ............................................................ 228
18.2.1 Key Aspects of Geophysical Logging .......................................... 231
18.3 Various Applications for Geophysical Logs.................................................. 234
18.3.1 Mapping Stratigraphy .................................................................. 234
18.3.2 Low Density Zones, Fractures and Cavities ................................ 234
18.3.3 An Alternate to Core Samples or Oriented Core ......................... 236
18.3.4 Groundwater Flow and Contaminants ......................................... 237
18.3.5 Pseudokarst Due to Acid Leaks ................................................... 239
18.3.6 Corrections Due to Borehole Deviation ....................................... 240
18.3.7 Reconstructing Geohydrologic and Well Construction Data ....... 242
18.4 Downhole, Crosshole and Tomographic Measurements ............................... 245
18.4.1 Downhole and Uphole Measurements ......................................... 245
18.4.2 Crosshole (Hole to Hole) Measurements ..................................... 246
18.4.3 Tomographic (Imaging) Measurements ....................................... 248
References .................................................................................................................. 249
19 Assessment of Larger Open Voids and Structures ................................................ 251
19.1 A Variety of Methods .................................................................................... 251
19.2 Visual Inspection ........................................................................................... 251
19.2.1 Concerns for Deep Foundation Piles ........................................... 253
19.2.2 Power Plant Ocean Water Intake System ..................................... 253
19.2.3 Mapping by Cavers and Cave Divers ........................................... 253
19.2.4 Tarpon Springs Bridge Failure ..................................................... 253
19.2.5 Road Widening Adjacent Sinkhole .............................................. 254
xiv Contents

19.3 Photographic and Video Documentation....................................................... 256

19.3.1 Its Not Always Easy ..................................................................... 258
19.4 Cave Mapping Systems ................................................................................. 259
19.4.1 Inertial Navigation Systems (INS) ............................................... 259
19.4.2 Cave Radio ................................................................................... 259
19.5 Laser and Sonar Systems .............................................................................. 260
19.5.1 Lasers ........................................................................................... 260
19.5.2 Sonar ............................................................................................ 261
19.6 Remotely Operated and Autonomous Vehicles for Inspection...................... 261
19.6.1 Remotely Operated Vehicles (ROVs) ........................................... 261
19.6.2 Autonomous Vehicles................................................................... 263
References .................................................................................................................. 263
20 Engineering Measurements and Monitoring ......................................................... 265
20.1 In-Situ Geotechnical Measurements and Monitoring ................................... 265
20.1.1 Drilling and the Installation of Instrumentation........................... 267
20.2 Monitoring Subsidence ................................................................................. 267
20.2.1 Regional Subsidence .................................................................... 268
20.2.2 Site-Specific Subsidence .............................................................. 268
20.2.3 Localized Settlement or Subsidence ............................................ 272
References .................................................................................................................. 272
21 Hydrologic Characterization and Measurements ................................................. 275
21.1 A Complex System ....................................................................................... 275
21.2 Karst Is a Multiple Porosity System.............................................................. 276
21.3 Lets Revisit the Issue of Scale....................................................................... 277
21.4 Temporal Aspects .......................................................................................... 278
21.5 Hydrologic Measurements ............................................................................ 279
21.6 Surface Water ................................................................................................ 280
21.6.1 Submerged Spring Flow Within Rivers, Lakes and Off-Shore .... 280
21.7 The Unsaturated Zone ................................................................................... 280
21.7.1 Groundwater Monitoring for a Landfill in Karst ......................... 281
21.8 The Saturated Zone ....................................................................................... 282
21.8.1 Example of Fracture Flow in Southeastern Minnesota ................ 283
21.8.2 An Example of Flow from an Artesian Well ................................ 285
21.8.3 Equivalent Porous Media ............................................................. 287
21.9 Groundwater Contaminants .......................................................................... 287
21.9.1 Inorganic Contaminants ............................................................... 287
21.9.2 Organic Contaminants.................................................................. 290
21.10 Aquitards and Barriers .................................................................................. 290
References .................................................................................................................. 293
22 Dye Tracing ............................................................................................................... 295
22.1 Introduction ................................................................................................... 295
22.2 Considerations for Dye Tracing .................................................................... 295
22.2.1 Water Sampling Prior to Introducing Dye ................................... 297
22.2.2 Estimating Quantity of Dye ......................................................... 297
22.2.3 Handling of Dye and Avoiding Cross Contamination ................. 298
22.2.4 Sampling and Analysis Methods.................................................. 299
22.3 Results and Analysis of Dye Trace Studies ................................................... 300
22.3.1 Dye Traces Confirming Connections at a Dam Site .................... 300
22.3.2 Assessing Vulnerability of a Cave Habitat ................................... 301
22.3.3 Dye Traces in the Woodville Karst Plain ..................................... 301
Contents xv

22.3.4 Evaluation of 2, 10 and 100 Year Capture Zone

for Silver Springs, Florida............................................................ 303
22.4 Limitations of Dye Traces ............................................................................. 306
References .................................................................................................................. 306
23 The Conversion of Data to Useful Information ..................................................... 307
23.1 An Assessment of All Data ........................................................................... 307
23.2 Managing the Data ........................................................................................ 308
23.3 Assembly of Data .......................................................................................... 308
23.3.1 The Use of Graphics .................................................................... 308
23.3.2 Selecting Scales ........................................................................... 308
23.3.3 Developing Graphics.................................................................... 308
23.4 Processing of Data......................................................................................... 310
23.5 The Final Interpretation and Conceptual Model ........................................... 310
23.5.1 Integration of Independent Data Sets ........................................... 310
23.5.2 Final Interpretation....................................................................... 312
23.5.3 Final Conceptual Model............................................................... 312
23.6 Visualization and Presentation of Data ......................................................... 315
23.7 Documentation: A Final Report .................................................................... 317
References .................................................................................................................. 317
24 Risk Assessment........................................................................................................ 319
24.1 Definition of Risk .......................................................................................... 319
24.2 Objective and Subjective Methods for Risk Assessment .............................. 319
24.3 Regional Risk Assessments ........................................................................... 320
24.3.1 Sinkhole Databases ...................................................................... 320
24.3.2 Regional Sinkhole Risk Maps ...................................................... 320
24.3.3 Groundwater Vulnerability in Karst ............................................. 321
24.3.4 Advantages and Limitations of Regional Methods ...................... 322
24.4 Site-Specific Risk Assessment ...................................................................... 322
24.4.1 Examples of Site-Specific Risk Assessment ................................ 324
References .................................................................................................................. 329

Part III Case Histories

25 The Development of a Landfill over an Abandoned Limestone Mine ................. 333

25.1 Background ................................................................................................... 333
25.2 An Assessment of the CCA and the Surface Fissures ................................... 336
25.2.1 Review of Regional and Local Geology ...................................... 336
25.2.2 Regional Geomorphology ............................................................ 337
25.2.3 Mapping and Trenching of the Surface Fissures.......................... 337
25.2.4 Developing a Preliminary Conceptual Model .............................. 338
25.2.5 Aerial Photo Analysis .................................................................. 339
25.2.6 Site-Specific Geology .................................................................. 341
25.2.7 Confirmation of the Preliminary Conceptual Model ................... 343
25.2.8 Hydraulic Connection of the Paleocollapse Fractures ................. 344
25.2.9 Subsidence Measurements ........................................................... 346
25.3 An Assessment of the Mine Conditions ........................................................ 349
25.3.1 Development of a Detailed Mine Map ......................................... 349
25.3.2 Monitoring of Changes in Mine Conditions over 3 Years............ 350
25.3.3 Determining the Types of Mine Collapse .................................... 350
25.3.4 Determining the Mechanism of Mine Collapse ........................... 352
xvi Contents

25.3.5 Bulking Measurements ................................................................ 353

25.3.6 Further Support of the CCA Conceptual Model .......................... 354
25.3.7 Determining Sources of Water Filling the Mine .......................... 355
25.3.8 Water Quality Measurements ....................................................... 356
25.4 A Groundwater Monitoring Plan .................................................................. 357
25.5 Subsidence Risk Assessment......................................................................... 357
25.5.1 Mine Stability Assessment ........................................................... 357
25.5.2 Two Examples of Subsidence ...................................................... 358
25.6 The Mine Backfilling Program...................................................................... 359
25.6.1 Initial Efforts with Fly-Ash .......................................................... 359
25.6.2 Crushed Rock Slurry Backfill ...................................................... 360
25.6.3 A QA/QC Program for Mine Backfilling..................................... 360
25.6.4 Summary of Mine Backfilling ..................................................... 361
25.7 Conclusions ................................................................................................... 361
References .................................................................................................................. 362
26 Site Characterization Along Bridge Alignment ..................................................... 365
26.1 Background ................................................................................................... 365
26.2 An Initial Site Assessment ............................................................................ 367
26.2.1 Findings by USGS ....................................................................... 367
26.2.2 Initial Site Visit ............................................................................ 367
26.3 The Approach ................................................................................................ 367
26.4 Phase I Reconnaissance Investigation ........................................................... 368
26.4.1 Review of Regional and Local Geology ...................................... 368
26.4.2 The Regional Geomorphology ..................................................... 370
26.4.3 A Site Fly Over ............................................................................ 370
26.4.4 Reconnaissance Marine Seismic Reflection Survey .................... 370
26.4.5 Microgravity Data ........................................................................ 371
26.4.6 Aerial Photo Analysis .................................................................. 371
26.4.7 A Detailed Review of Existing Boring Data ................................ 372
26.4.8 The Correlation of Anomalous Conditions .................................. 373
26.5 Phase II Confirmation Phase ......................................................................... 374
26.5.1 Additional Marine Seismic Reflection Data ................................ 374
26.5.2 Additional Microgravity Data ...................................................... 375
26.5.3 A Conceptual Model .................................................................... 376
26.6 Phase III Detailed Investigation .................................................................... 376
26.6.1 Deeper Seismic Reflection Data .................................................. 376
26.6.2 Drilling of Four Additional Boreholes ......................................... 377
26.6.3 Geophysical Logs from the Four Boreholes ................................ 378
26.6.4 Support for the Final Conceptual Model ..................................... 379
26.7 Risk Assessment ............................................................................................ 380
26.7.1 Limitations ................................................................................... 382
26.8 Conclusions ................................................................................................... 382
References .................................................................................................................. 383
27 EPA Superfund Site ................................................................................................. 385
27.1 Background ................................................................................................... 385
27.1.1 Record of Decision (ROD) .......................................................... 386
27.2 Objectives of the Overall Investigation ......................................................... 387
27.2.1 Objectives of the Geologic and Karst Investigation..................... 387
27.2.2 Objectives of the Hydrologic Investigation.................................. 388
27.2.3 The Owner’s Goals ...................................................................... 388
27.2.4 Review and Oversight Committee ............................................... 388
Contents xvii

27.3 Technical Approach ....................................................................................... 388

27.3.1 Core Team .................................................................................... 388
27.3.2 Methodologies.............................................................................. 389
27.4 Site Preparation ............................................................................................. 389
27.5 The Desk Study ............................................................................................. 392
27.5.1 The General Site Setting .............................................................. 392
27.5.2 Regional Geology ........................................................................ 392
27.5.3 Regional Hydrology ..................................................................... 393
27.5.4 Regional Geomorphology ............................................................ 393
27.5.5 Reported Sinkholes and Sinkhole Trends in the Region .............. 395
27.5.6 Review of Corporate Files and Interviews
with Previous Workers ................................................................. 397
27.6 The Preliminary Conceptual Models ............................................................ 398
27.6.1 Geologic Conceptual Model ........................................................ 399
27.6.2 Hydrologic Conceptual Model..................................................... 399
27.6.3 Karst Conceptual Model .............................................................. 400
27.7 Shallow Geohydrologic Conditions .............................................................. 400
27.7.1 The Sands and Surficial Aquifer .................................................. 401
27.7.2 The Semi-confining Layer ........................................................... 401
27.7.3 Pond Boundaries .......................................................................... 403
27.7.4 Buried Drums ............................................................................... 404
27.7.5 Contaminants ............................................................................... 405
27.8 Deeper Geohydrologic Conditions................................................................ 405
27.8.1 Anomalous Areas ......................................................................... 406
27.8.2 The Sinkhole Pond ....................................................................... 409
27.8.3 Meyers Cove ................................................................................ 410
27.8.4 Within the Anclote River Adjacent to the Site ............................. 410
27.8.5 Summary of Karst Conditions ..................................................... 411
27.9 The Conceptual Model for the Site ............................................................... 411
27.10 Sinkhole Risk Assessment............................................................................. 412
27.10.1 Summary of Risk Assessment...................................................... 413
27.11 The Ability of Geology to Support the Proposed Remedy ........................... 413
27.12 About the Site Characterization Strategy ...................................................... 414
References .................................................................................................................. 415

Index .................................................................................................................................. 417

 Part I
A Brief Overview Karst and Pseudokarst

Part I of this book provides the reader with a brief overview of some of the wide range of karst
and pseudokarst conditions that can exist along with their benefits and damaging effects. The
scale of these features that can impact engineering works or environmental issues can range
from small voids or permeable zones of centimeters to caverns with dimensions of hundreds of
meters or more. Most karst conditions will have evolved over thousands of years or more while
those involving mans activities (pseudokarst) may be formed over much shorter periods of
time from a few years to a few hundred years. Having an understanding of the type of karst or
pseudokarst features that can develop or are already present in a particular setting will provide
a starting point for an effective site characterization effort.
 Some Common Terms
1

Abstract
There is a wide range of terminology encountered in the field of karst. The following terms
will set the stage for the topics to be covered. Additional terms will be presented throughout
the book. The two key divisions are karst and pseudokarst. Karst conditions are formed due
to a natural process of dissolution of soluble rock. Pseudokarst may look or act like karst,
but it may be formed naturally in non-soluble rock or may be caused by man’s activities. In
addition, the term paleokarst is used to indicate those karst features that are very old, in
geologic terms, and are typically inactive. For a complete discussion of karst terminology
see Monroe (A glossary of karst terminology. US Geologic Survey Water Supply Paper
1899K, 1970), Lowe and Waltham (Dictionary of karst and caves. In: Judson D (ed) British
Cave Research Association cave studies series 10. British Cave Research Association,
Malvern, 2002), and Fields (A lexicon of cave and karst terminology with special reference
to environmental karst hydrogeology. US EPA/600/R-02/003, 2002).

1.1 Karst forming process has been occurring over much longer
periods of many thousands of years or more. However,
The word karst stems from the meaning stone, and its deriva- certain conditions in evaporate rocks, such as gypsum or
tives are found in many languages of Europe and the Middle salt and some pseudokarst conditions can result in subsid-
East (Ford and Williams 2007). The term karst is commonly ence, sinkholes, and collapse that can develop over a human
used to describe a wide range of surface landforms that have time scale.
developed from the dissolution of soluble rock. These surface Traditionally karst refers to the landforms that result from
features may include bare rock, closed depressions or collapsed the dissolution of soluble rock, these landforms are result
sinkholes. These features result in unique hydrologic conditions from subsurface dissolution features. For this book, the term
such as sinking streams, springs, or conduit flow providing high karst will be used to include both surface and subsurface fea-
porosity. White (1988) and LaMoreaux (1995) provide a brief tures, which are an interconnected system. The presence of
review of historical development of karst literature. the surface landforms requires the presence of the subsurface
The most common karst images are those of sinkholes in features. Therefore, we are using the term in a broader sense
the news or from our visits into caves. In a broader sense, to include both surface and subsurface features when refer-
karst conditions include both surface and subsurface features ring to karst terrains.
including sinking streams, springs, highly irregular top of
rock profile, the epikarst, dissolutionally enlarged joints or
bedding planes, conduits and caves (Fig. 1.1). Sediments or 1.2 Paleokarst
younger rock strata may cover karst features with no visible
evidence of their presence at the surface. Paleokarst refers to old or mature karst (sinkholes or caves)
While subsidence, sinkholes, and collapse can occur on a that has developed over geologic periods of time and has
human time scale, the dissolution of carbonate rock and void become inactive. That is, paleokarst has been uncoupled

© Springer Science+Business Media Dordrecht 2016 3

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_1
4 1 Some Common Terms

Spring
Epikarst Zone (discharge)
River
(recharge) Sinkhole
Lake

Numerous small Cave System Isolated Large Chamber

isolated void sand fractures

Fig. 1.1 A conceptual sketch of a karst system showing some of the many geologic and hydrologic complexities

from the dissolution process and as a result there is little, if from industrial sites. Many of the sinkholes reported in the
any, further dissolutioning taking place. However, subsid- evening news are commonly pseudokarst “sinkholes” caused
ence or collapse may still occur if there are large open cav- by leaky water pipes or sewer lines.
erns present. While karst alone can present large problems,
Paleokarst may have undergone tectonic changes with pseudokarst covers a broader range of conditions and is
uplifts or subsidence and/or sea-level changes. While probably as pervasive as, if not more than, karst, particu-
typically buried, it may be exposed at the surface or at road larly in developed areas. The number of pseudokarst events
cuts or high in the mountains. Paleokarst can sometimes be is probably equal to those of karst and may be of greater
reactivated by natural causes or man’s actions. Paleokarst is impact and cost than that of karst alone. Palmer (2007) pro-
of importance because of its association with deposits of vides a brief discussion of pseudokarst related to non-solu-
oil and gas. In addition, minerals such as silver, lead, zinc ble rocks. Borchers (1998), Holzer (1984), and Johnson
and bauxite are found in zones of paleokarst and breccias and Neal (2003) all provide papers on pseudokarst related
pipes (circular zones of collapse filled with broken rock to fluid withdrawal, hydro-compaction, earth fissures,
fragments). ground cracks and mining.
Palmer and Palmer (1989) discuss the extent of paleokarst
within the United States. Bosak et al. (1989) and James and
Choquette (1988) provide a general discussion of paleokarst References
terminology, its presence worldwide, along with associated
mineral deposits, hydrogeology and engineering hazards. Borchers JW (ed) (1998) Land subsidence: case studies and current
research. In: Proceedings of the Dr. Joseph F. Poland symposium on
land subsidence, Association of Engineering Geologists, special
publication no 8. Star Publishing Co., Belmont
1.3 Pseudokarst Bosak P et al (eds) (1989) Paleokarst a systematic and regional review.
Developments in earth surface processes 1. Elsevier, New York
Fields MS (2002) A lexicon of cave and karst terminology with special
Pseudokarst conditions are those that resemble or act like
reference to environmental karst hydrogeology. US
karst but were not formed by the natural dissolution of rock. EPA/600/R-02/003
They are also known as “false-karst” or are sometimes Ford D, Williams P (2007) Karst hydrogeology and geomorphology.
referred to as “analogous karst”. Pseudokarst may be associ- Wiley, New York
Holzer TL (1984) Ground failure induced by ground water withdrawal
ated with natural conditions such as caves in lava tubes, sea
from unconsolidated sediment. In: Holzer T (ed) Man-induced land
caves, and ice caves. Pseudokarst may be associated with fis- subsidence, The Geologic Society of America, Reviews in engineer-
sures and caves in sandstone, or even granite caused by ing geology VI. , Geologic Society of America, Boulder, pp 67–105
structural changes or erosion. In some cases, it may be asso- James NP, Choquette PW (eds) (1988) Paleokarst. Springer,
New York
ciated with man’s activities such as subsidence or collapse
Johnson KS, Neal JT (eds) (2003) Evaporate karst and engineering/
due to underground mines, fissures and subsidence due to environmental problems in the United States, Oklahoma Geological
withdrawal of groundwater, oil or gas or even acid leaks Survey, Circular 109
References 5

LaMoreaux PE (1995) Historical references to karst studies In: Beck Palmer AN (2007) Cave geology. Cave Books, an Affiliate of the Cave
BF (ed.) Proceedings of the 5th multidisciplinary conference on Research Foundation, Dayton
sinkholes and the engineering and environmental impacts of karst: Palmer MV, Palmer AN (1989) Paleokarst of the United States. In:
karst geohazards, Gatlinburg, Tennessee. Balkema, Rotterdam, Bosak P, Ford D, Glazek J, Horacek I (eds) Paleokarst, a systematic
pp 11–16 and regional review. Developments in earth surface processes 1.
Lowe D, Waltham T (2002) Dictionary of karst and caves. In: Judson D Elsevier, New York, pp 337–363
(ed) British Cave Research Association cave studies series 10. White WB (1988) Geomorphology and hydrology of karst terrains.
British Cave Research Association, Lancaster Oxford University Press, New York
Monroe WH (1970) A glossary of karst terminology. US Geologic
Survey Water Supply Paper 1899K
 The Development of Karst Conditions
2

Abstract
The following is a brief introduction to the development of karst conditions from deposition
to the variety of post-depositional processes that have created what we see today. This is a
very broad topic and is specific to the area in which it occurs as well as the maturity of the
process. This includes the types of soluble rock, rates of dissolution and geomorphic pro-
cesses that may affect the dissolution process. The benefit of this information lies in the
ability to understand what karst conditions may exist or be expected at a particular site.
Often the various geologic settings provide clues as the type, depth, shape or relationship of
expected karst features that may be present.

2.1 Carbonates and Other Soluble Rock 1963). The islands of the Bahamas contain a wide range of
karst conditions in terms of size and depth. The well known
Carbonate rocks include limestone, dolomites, and chalks. and spectacular Blue Holes found in the Bahamas are deep
Other readily soluble rocks include evaporates such as gyp- solution conduits as much as 100 m in diameter and more
sum, anhydrite and salt. All of which are found in a wide than 100 m deep. Dean’s Blue Hole on Long Island is the
range of environments worldwide. Nearly all carbonate sedi- deepest known at about 200 m (Wilson 1994).
ment is derived from calcareous algae and calcium carbonate Chalk is a marine limestone consisting of fine grain car-
(CaCO3) producing animals living at or near the site of depo- bonates. It is often not well cemented, friable, and is highly
sition. Deposition of carbonates typically occurs in shallow, porous. Because of its high porosity and low strength large
marine environments in tropical and subtropical climates. caves do not generally develop in chalk. As a result, chalk is
On land, both carbonate and evaporate deposits can also not strong enough to give rise to typical karst landforms.
occur in lakes (Braithwaite 2005). Waltham et al. (2005) provide further details on chalk.
The term limestone is used for rocks in which the carbon- Evaporate rocks include gypsum (CaSO4 2H2O), or anhy-
ate fraction is primarily calcite (CaCO3). The term dolomite drite (CaSO4), and salt (NaCl). Both salt and gypsum are
is used for rocks in which the carbonate fraction is primarily much more soluble than limestone. These evaporates are
magnesium (MgCa(CO3)2). The original deposition of lime- found in 32 of the 48 contiguous states (Johnson 2003) and
stone is as an unconsolidated mud. It is then altered to a con- over 25 % of the continental surface of the world (Ford and
solidated rock by compaction, cementation, dissolution and Williams 2007).
replacement.
Today active deposition is taking place over the Bahamas
platform, which extends more than 966 km from Florida to 2.2 Post Deposition Processes
the island of Hispaniola. This area has developed along the
subsiding continental margin of North America and consists 2.2.1 Dissolution of Limestone
of a series of shallow-water carbonate banks. It is the largest
carbonate platform in the world and is also one of the thick- While limestone is nearly insoluble in distilled water,
est ever developed with a possible thickness of 10 km (Purdy groundwater will contain dissolved carbon dioxide that

© Springer Science+Business Media Dordrecht 2016 7

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_2
8 2 The Development of Karst Conditions

forms a weak carbonic acid (H2CO3) in which limestone is or enlargement of caves or subsidence cannot occur within
slightly soluble. While some carbon dioxide is derived from the 100 year lifetime of most engineered structures (Sowers
the atmosphere during rainfall resulting in a weak carbonic 1996; Waltham et al. 2005). However, dissolution and sub-
acid, most of the carbon dioxide is derived from organics sidence may be accelerated by man-made changes such as:
within the soil and by decaying organic material. A detailed
discussion of the chemical reactions and the complex disso- • Turbulent flow (due to pumping or reservoir leakage at
lution process can be found in White (1988), Ford and dams).
Williams (2007), and Palmer (2007). • Changes or concentration of surface water runoff from
The rock mass is initially formed with some degree of roads, roofs and paved parking lots (primarily add to the
primary porosity as it is deposited and modified by compac- erosional effects).
tion and cementation. It is then modified with cracks, joints, • Acids that have leaked from industrial facilities can result
fractures, and bedding planes (secondary porosity) over time in rapid dissolution creating cavities over tens of years.
by a variety of stresses including tectonic stress, loading and
rebound from glaciations and the constant small flexure of Based upon the very slow process of natural dissolution
daily earth tides (about 2 cm) over geologic time (Fett 1992, for limestone and dolomites, the subsurface voids, conduits,
personal communication). The resulting openings (second- cavities or caves that may impact a site are already present
ary porosity) in the rock serve as preferential means for and have taken more than 10,000–100,000 years to develop
water to flow through the rock system. (White 1988). It is these well-developed features that become
When the flow of slightly acidic water is concentrated the focus of many of our site characterization projects in
through the zones of secondary porosity, it leads to an karst settings.
increase in local porosity (tertiary porosity) through dissolu-
tion. Dissolution begins with laminar flow of water through 2.2.1.2 Two Models for Dissolution
small cracks. Over time the crack widens and flow becomes The classic model for dissolution of carbonate rocks is that
more turbulent and the process of dissolution accelerates. of surface water with elevated levels of carbonic acid that
Selective zones of secondary porosity then become dominant infiltrate into the rock resulting in dissolution and karst. This
flow paths becoming a self-perpetuating means of dissolu- is referred to as epigenic karst (Fig. 2.1a). However, in the
tion resulting in a network of connected conduits. These ter- last few decades, an alternate model of karst development
tiary pathways can be significantly enlarged by a combination has become recognized. The alternate model has dissolution
of dissolution, mechanical erosion and transport of rock. dominated by upward flow and gradients from deeper waters
Their size can range from a few centimeters to tens of meters which are referred to as hypogenic karst (Fig. 2.1b).
in width. Large volumes of rapidly moving fresh water are Hypogenic karst develops from the upward flow of deeper
required to create these tertiary porosity features. The effec- groundwater systems that are independent of recharge from
tiveness of the acidic water is greatest when it first comes the surface.
into contact with rock and is reduced as the water flows Additional models have suggested a combination of pro-
deeper and becomes saturated with dissolved calcite. cesses involving upward migration of fluids of higher tem-
peratures, or carrying sulfuric acid in addition to carbonic
2.2.1.1 Rate of Dissolution in Limestone acid (Palmer 2007). The concepts of hypogene karst devel-
The rate of dissolution (limestone removal) is a function of opment are summarized by Klimchouk (2009) and discussed
the chemical dissolution (or corrosion) and the mechanical at two symposiums:
erosion (corrasion) process. Both the dissolution and the
mechanical erosion processes are a function of a number of • National Cave and Karst Symposium 1: Advances in
factors that make the exact rates difficult to determine. Many Hypogene Karst held in 2009 and published by National
environmental factors play an important role in the dissolu- Cave and Karst Research Institute (Stafford et al. 2009).
tion of carbonate rocks. • Hypogene Speleogenesis and Karst Hydrogeology of
Dissolution rates of limestone have been measured at a Artesian Basins held in 2009 and published by Ukrainian
number of locations throughout the world based upon the Institute of Speleology and Karstology: Special Paper 1
amount of surface denudation or by the chemistry of springs. (Klimchouk and Ford 2009).
The results are commonly reported in units of centimeters
per thousand years (cm/ka−1). While the natural dissolution The depth of active dissolution in the epigenic model is
of carbonate rock is highly variable, 1–3 cm/1,000 years is generally thought of as being shallow. Water will accomplish
an average range of values (Ford and Williams 2007). 50–80 % of its solutional work within about 10 m of the sur-
In general, the natural dissolution of limestone is a very face (Ford and Williams 2007). Milanovic (2004) suggests
slow geologic process, so slow that significant new cavities that the most significant dissolution from a hydrogeologic
2.2 Post Deposition Processes 9

a Infiltration of slightly b
acidic rain water along with
acids from organic debris
results in dissolution

Ground Surface

Soluble Rock
Clay Layer

Cave Formation in Soluble Rock

Cave Formation in Soluble Rock

Deep acidic waters

flow into soluble rock controlled by
secondary porosity (joints, fractures, faults)

Epigenic Model Hypogenic Model

Fig. 2.1 Two models of karst development, the classic model of epigenic karst development results from the infiltration of surface waters (a) and
a more recent model of hypogenic karst development results from the upward flow of deeper waters (b)

standpoint will occur in the shallow environment, generally Biologic microbes in certain caves have been found to
less than 100 m deep. The depth of active dissolution in the produce acids that also result in dissolution of limestone.
hypogene model is much deeper. Deep-water circulation sys- Such microbial activity may occur at much greater depths in
tems have been recently recognized with karst water circula- carbonate aquifers. Palmer (2007) provides further discus-
tion detected to depths greater than 2 km (Glazek 1989). sion of microbial effects on karst chemistry.
The Carlsbad Caverns and nearby Lechuguilla Cave in
New Mexico are two of the largest cave systems which have 2.2.1.3 The Fresh Water Lens and Mixing
been formed, in part, by circulating deep waters and the pres- Dissolution
ence of hydrogen sulfide. The origin of the acids is from The fresh water lens is an important source of drinking water
much deeper oil deposits developed by bacteria activity. The within coastal areas and on islands (Fig. 2.2). The thickness
resulting hydrogen sulfide gas migrates upward into the thick of the fresh water lens is approximated by the Gheyben-
overlaying limestone where it comes in contact with ground- Herzberg equation, that is, for every meter of fresh water
water and produces sulfuric acid. Dissolution of the lime- head, (above sea level) the fresh water lens is approximately
stone by sulfuric acid is about eight times faster than that due 40 m thick (a 1:40 ratio due to density differences between
to carbonic acid, which can account for the large cave sys- fresh and salt water) (Palmer 2007). This can account for
tems encountered in these two cave systems. large amounts of fresh water on islands and coastal areas.
With the development of hypogenic karst model, it is Dissolution is associated with fresh waters flowing from
clear that the karst system is much more complex than origi- the surface (Fig. 2.2a), but when infiltrating fresh water
nally thought. In fact, some of the existing karst systems are becomes saturated with CO2 further dissolution ceases.
beginning to be reinterpreted to include both epigenic and However, where fresh water and saline waters come in con-
hypogenic karst conditions (Barr et al. 2008). tact at the fresh water/salt water interface, further dissolution
10 2 The Development of Karst Conditions

d = Water table height above

Mean Sea Level
Dissolution occurs Ground Surface
at and below
the water table (a)

Ocean Fresh Ground Fresh Water

Water Aquifer
Water Flow
Increased porosity zones Saline Ground D = d x 40
which result in reef (Ghyben-Herzberg
Water Flow
development (c) at relationship)
lower sea level stands

Freshwater/Saltwater Interface

Mixing dissolution occurs at the

freshwater/saltwater interface
resulting in dissolution (b)

Fig. 2.2 A fresh water lens is found along the coast and on islands and its thickness is determined by the Ghyben-Herzberg relationship. Dissolution
of limestone occurs near the surface of fresh water (a) and at the freshwater saltwater interface (b)

can occur even when the fresh water has become saturated 2.2.3 Dissolution of “Non-soluble Rocks”
(Fig. 2.2b). This is referred to as mixing dissolution. White
(1988) and Palmer (2007) provide a discussion on the pro- Development of karst conditions in non-soluble rock such as
cess by which mixing dissolution can take place. sandstone and granite are unusual, but due occur (Ford and
The depth of the saltwater interface is a function of fresh Williams 2007). Work by Alexander et al. (2005) have iden-
water level above sea level at a 1:40 ratio. As the fresh water tified karst conditions including sinkholes and rapid ground-
lens becomes thicker, the interface will become deeper, water flow by dye traces in the Hinckley sandstone in east
resulting in a deeper mixing dissolution. The freshwater/salt- central Minnesota. A number of caves in the St. Peter
water interface can vary from a very sharp boundary to a Sandstone occur under the streets and buildings of down-
broad diffuse boundary depending upon site-specific condi- town Minneapolis, Minnesota. Some of these caves are
tions and can change overtime (Yuhr and Benson 1995). thought to be pseudokarst, created by the piping of the poorly
cemented St. Peter Sandstone into a main sanitary sewer line
for the downtown area (Brick 2000).
2.2.2 Dissolution of Other Soluble Rocks

Evaporites, such as halite (salt) and gypsum, are much more 2.2.4 Mechanical Erosion and Transport
soluble in contrast to carbonates such as limestone, dolo-
mites and chalks. Halite is the most soluble of karst rock, Mechanical erosion (corrasion) and transport of rock occurs
about three orders of magnitude more soluble than lime- under rapid flow conditions which is common in karst.
stone. Large cavities within salt deposits have been known to Rapidly flowing water can mechanically remove bedrock by
develop over a few years to tens of years associated with abrasion and could exceed rates of dissolution (Palmer
extraction of salt water for oil and gas drilling (Johnson 2007). Under flood conditions the increased buoyancy pro-
1998). vided by water reduces the weight of the rock and enables
The dissolution of gypsum is approximately two orders of rapidly flowing flood waters to move large blocks of rock. As
magnitude greater than limestone. Under high groundwater a result, the mechanical erosion and transport of rock may
gradients such as beneath dams, dissolution rates for gypsum add to the removal of limestone by dissolution alone by an
may be as high as 10 mm/year (Dreybrodt et al. 2002). order of magnitude or more.
Because gypsum dissolves faster, sinkholes and caves have
been known to form within a few years where gypsum is
located beneath dams (Brune 1965). 2.2.5 Geomorphology
The many aspects of dissolution and cave development
in other soluble rocks are covered by Klimchouk et al. Karst development occurs over a geologic time scale and can
(2000). be modified over that time by the dynamics of the earth
2.2 Post Deposition Processes 11

(Fookes et al. 2005). This includes changes in sea level, by of the lower sea levels (Fig. 2.3). Stalactites found in some
uplift of a land mass and by climate change. Recognizing underwater caves are evidence for a water table (sea level)
these past processes will aid in understanding the karst con- that was once much lower than today.
ditions that may be present. Throughout the last episode of sea level rise there were
intermediate intervals of stability. During these times, as
2.2.5.1 Changes in Sea Level the fresh water flowed seaward dissolution took place cre-
Sea levels have changed considerably over geologic time. Due ating local horizontal zones of tertiary porosity (Fig. 2.2c).
to multiple glacial and interglacial stages, sea level stands have As the fresh water exited the sea floor, it carried nutrients
ranged from about 60 m above current sea level to at least that created ideal conditions for reefs to grow. Evidence for
130 m below. Recent analysis of sea-level data from the north- this is seen on Florida’s southeast coast. As sea level rose,
ern Gulf of Mexico by Balsillie and Donoghue (2004) used three main reefs are found at approximately 6, 12 and 24 m
carbon-14 (C-14) dating along with other historic data. These below MSL. On-shore there are horizontal zones of
changes in sea levels are tied to the last period of glaciation. increased tertiary porosity found at approximately 6, 12
Their findings suggest that sea level has been as much as 130 m and 24 m below MSL, (Finkl and Esteves 1997; Esteves
below present day sea level. These results from the northern and Finkl 1999; Finkl and Krupa 2003). Today these hori-
Gulf of Mexico have significance beyond the local region since zontal zones of tertiary porosity are often encountered
the data compare favorably with data from the Red Sea. when pouring cement piles for foundations resulting in
Both Florida and the Bahamas, which have not been excessive quantities of cement (Berkowitz 1994 personal
impacted by uplift, have deeper karst development as a result communication).

Land Surface

d Present Sea Level

e
D = d x 40 ac
Freshwater nte rf Saltwater
I

Change in Sea Level

as much as 130 meters

d Ancient Sea Level

Freshwater Saltwater
D = d x 40 e
fac
er
Int

The combination of lower sea levels and a

freshwater head of 3 m can account for deep
cave development of more than 250 m

Fig. 2.3 During periods of lower sea level stands, with a thick freshwater lens, there can be extensive dissolution deep within the limestone
12 2 The Development of Karst Conditions

The combination of lower sea levels and a thick fresh water Similar conditions of large paleosinkholes and underwater
lenses (Fig. 2.3) can account for the deeper zones of dissolu- cave systems can be expected at other locations with soluble
tion in Florida and the Bahamas. A few examples include: rock throughout the world.

• Key Largo, Florida – A sinkhole found by United States 2.2.5.2 Tectonic Uplift of Land Masses
Geological Survey (USGS) off of Key Largo in the reefs Some land masses and islands have been uplifted by tectonic
was more than 55 m deep with a C-14 date of 15 ka (Shinn forces essentially resulting in an apparent lowering of sea
et al. 1996); level. The effects of uplift will provide greater potential
• Tarpon springs, Florida – A paleosinkhole more than 49 m energy and increased hydraulic gradients within an aquifer
deep with a C-14 date of more than 40 ka (Yuhr et al. 2003); resulting in a lowering of the base level of karst develop-
• Bahamas – the deep Blue holes up to 200 m deep (Wilson ment. For example, limestone cliffs up to 180 m high exist
1994); around the perimeter of northern Guam (Taborosi 2004).
• Fort Lauderdale, Florida – Two sinkholes with diameters of These cliffs have been uplifted and contain flank margin
365 and 670 m were observed in the Gulf Stream approxi- caves that were developed when the land was at a lower level
mately 9 and 13 km offshore from Fort Lauderdale at depths (Fig. 2.4). Similar results can be seen on the Isla de Mona off
of approximately 250–300 m (US Coast Guard 2007). of Puerto Rico (Mylroie 2005; Palmer 2007).

Evidence of
Past Sea Level
Tectonic Uplift of the land mass has occurred

Evidence of
Past Sea Level

Current Sea Level

Fig. 2.4 Sea caves originally developed at sea level can been seen at higher elevations due to tectonic uplift
2.3 Some Properties of Karst Rock 13

2.2.5.3 Climate increased vegetation, along with larger islands to the north
Regional climate has a strong influence on the development result in a larger fresh water lens provides conditions for
of karst landforms. The most mature karst occurs in wet trop- deeper bedrock dissolution at the fresh/saltwater interface
ical environments. The principal climatic factor that impacts (Whitaker and Smart 1997).
the development of karst appears to be the mean annual run-
off rate (Whitaker and Smart 1997). As the earth’s climate
has evolved, some areas that were once optimum for devel- 2.3 Some Properties of Karst Rock
opment of karst conditions are now dry deserts and moun-
tains that contain the presence of paleokarst. The physical properties vary widely for limestones and dolo-
Even local changes in climate and associated vegetation mites. “Although there are general correlations between age,
are cited for the differences in hydraulic conductivity in the bulk density, compressive strength and permeability, these
Bahamian limestone. The average hydraulic conductivity is may vary within the scale of samples” (Braithwaite 2005).
two orders of magnitude greater in the northern Bahamas The range of limestone strength and its properties can vary
(with rainfall of 1,550 mm) than in the relatively more arid significantly: As an example:
southern islands (with a rainfall of 810 mm). The greater
rainfall to the north results in greater vegetation with an • The Bethany Falls limestone from the Kansas City area is
increase in soil CO2 providing conditions for shallow disso- an older (Pennsylvania Period), massive, low porosity and
lution. In addition, the combination of greater rainfall, hard limestone (1,229 psi, 139.3 MPa) (Fig. 2.5a),

Fig. 2.5 Variations in limestone range from an older, massive strong limestone Kansas City, Kansas (a), a younger, highly weathered weak, almost
friable limestone, in Puerto Rico (b) and a very young weaker, porous, Miami limestone (c) in south Florida
14 2 The Development of Karst Conditions

• The Aymamon Limestone from Puerto Rico is a younger Bell FG (2004) Engineering geology and construction. Spoon Press,
(early Miocene), and at some surface exposures it is London
Braithwaite CJR (2005) Carbonate sediments and rocks, a manual for
highly weathered and almost friable limestone earth scientists and engineers. Whittles Publishing, Dunbeath, Scotland
(Fig. 2.5b), and Brick G (2000) Subterranean twin cities. University of Minnesota
• The Miami Limestone from Southeast Florida is a very Press, Minnesota
young (Pleistocene age), extremely porous and relatively Brune G (1965) Anhydrite and gypsum problems in engineering geol-
ogy. Eng Geol 2(1):26–38
weaker limestone (Fig. 2.5c). Deere DU, Miller RP (1966) Engineering classification and index prop-
erties for intact rock. Air Force Weapons Laboratory Technical
Limestone strength in unconfined compression can range Report AFWL-TR-65-116
between 1.5 MPa, the lower limit for rock by its engineering Dreybrodt W et al (2002) Karstification below dam sites: a model of
increasing leakage from reservoirs. Environ Geol 42(5):518–524
definition, to more than 150 MPa, greater than concrete Esteves LS, Finkl CW (1999) Late Cenozoic depositional paleoenvi-
(Sowers 1996). Sowers’ describes the young, porous Miami ronments of Southeast Florida interpreted from core logs. Rev Bras
Limestone as having very different porosities above and Geocienc 29(2):129–134
below the water table. Above the water table porosity is Finkl CW, Esteves LS (1997) A preliminary analysis of neogene depos-
ital environments in Broward County, Florida. Fla Sci
about 15 % versus below the water table where circulation is 60(2):94–103
the greatest may be as high as 75 % and appears sponge-like Finkl CW, Krupa SL (2003) Environmental impacts and coastal plain
(Fig. 2.5c). Mean values for compressive strength of lime- activities on sandy beach systems: hazards, perception and mitiga-
stones are on the order of 100 MPa, with a few ranging up to tion. J Coast Res Spec Issue 35:132–150
Fookes PG et al (eds) (2005) Geomorphology for engineers. Whittles
200 MPa. General values for limestone and dolomite strength Publishing, Dunbeath, Scotland
characteristics and bearing capacity of cave roofs can be Ford D, Williams P (2007) Karst hydrogeology and geomorphology.
found in Deere and Miller (1966), Bell (2004), Sowers Wiley, New York
(1996), and Waltham et al. (2005). Glazek J (1989) Tectonic conditions for karst origin and preservation.
In: Bosak P, Ford DC, Glazek J, Horacek I (eds) Paleokarst, a sys-
When the primary porosity of the rock is high, such as in tematic and regional review. Elsevier Academia, Amsterdam,
chalk or the Miami Limestone, the flow of groundwater will pp 569–574
tend to be widely diffuse and tertiary porosity development Johnson KS (1998) Land subsidence above man-made salt dissolution
will be minimized (Sowers 1996). Cave systems do not com- cavities. In: Borchers JW (ed) Proceedings of the Poland sympo-
sium on land subsidence, Land Subsidence Case Studies and
monly form in weaker carbonate rocks and those that do Current Research. AEG Special Pub. No. 8, Star Pub. Belmont,
form will not persist. These weaker rocks lack the strength to California, pp 385–392
span large cavities and result in limited caves and surface Johnson KS (2003) Evaporite-karst problems in the United States. In:
karst features. Johnson KS, Neal JT (eds) Evaporate karst and engineering/envi-
ronmental problems in the United States, Oklahoma Geological
In strong rocks that are well cemented with low porosity, Survey, Circular 109:1–20
flow is restricted to fractures that then become enlarged by Klimchouk AB (ed) (2000) Speleogenesis evolution of karst aquifers.
dissolution and erosion resulting in a cave system. Most National Speleological Society, Huntsville, Alabama
caves form in the stronger rocks with low primary porosities. Klimchouk AB (2009) Principle features of hypogene speleogenesis.
In: Klimchouk AB, Ford D (eds) Hypogene speleogenesis and karst
The greatest karst development is generally seen in massive hydrogeology of artesian basins. Ukrainian Institute of Speleology
thick-bedded limestone (Waltham et al. 2005). and Karstology: Special Paper 1, Simferopol, Ukraine, pp 7–16
Klimchouk AB, Ford D (eds) (2009) Hypogene speleogenesis and karst
hydrogeology of artesian basins. Ukrainian Institute of Speleology
and Karstology: Special Paper 1, Simferopol, Ukraine
References Milanovic T (2004) Water resources engineering in karst. CRC Press,
Boca Raton
Alexander EC Jr et al (2005) Dye tracing sewage lagoon discharge in a Mylroie JE (2005) Coastal caves. In: David CC, White WB (eds)
sandstone karst, Askov, Minnesota. In: Beck B (ed) Proceedings of Encyclopedia of caves. Elsevier, New York, pp 122–125
the 10th multidisciplinary conference on sinkholes and the engi- Palmer AN (2007) Cave geology. Cave Books, an Affiliate of the Cave
neering and environmental impacts of karst, San Antonio, Texas Research Foundation, Dayton
Sept 2005, vol 144. ASCE Geo-Institute Geotechnical Special Purdy EG (1963) Recent calcium carbonate facies of the Great Bahamas
Publication, Reston, Virginia, pp 449–458 Bank. Sediment Facies J Geol 71:472–497
Balsillie JH, Donoghue JF (2004) High resolution sea–level history for Shinn EA et al (1996) A giant sediment trap in the Florida Keys. J Coast
the Gulf of Mexico since the last glacial maximum, vol 103, Report Res 12:953–959
of investigation. Florida Geologic Survey, Tallahassee Sowers GF (1996) Building on sinkholes, design and construction of
Barr K et al (2008) Hypogenic karst and its implications for Minnesota foundations in karst terrain. ASCE Press, Reston (out of print)
hydrogeology. In: Yuhr L, Alexander Jr CE, and Beck BF (eds) Stafford KW et al (eds) (2009) National cave and karst symposium 1:
Proceedings of the 11th multidisciplinary conference on sinkholes advances in hypogene karst. National Cave and Karst Research
and the engineering and environmental impacts of karst, ASCE Institute, Carlsbad, New Mexico
Geo-Institute, Geotechnical Special Publication No. 183:43–53
References 15

Taborosi D (2004) Field guide to caves and karst of Guam. Bess Press, Wilson WL (1994) Morphology and hydrology of the deepest know
Honolulu cave in the Bahamas: Dean’s Blue Hole, Long Island. Bahamas J
US Coast Guard (2007) Draft environmental impact statement calypso Sci 2(1):10–14
LNG deepwater port license application, Volume I impact analysis Yuhr L, Benson RC (1995) Saltwater intrusion: concepts for
US Coast Guard, Washington, DC measurements and regional characterization for Broward County,
Waltham AC et al (2005) Sinkholes and subsidence, karst and cavern- Florida. In: Proceedings of the symposium for the application of
ous rocks in engineering and construction. Springer-Praxis, geophysics to environmental and engineering projects, EEGS
Chichester, UK Annual Meeting, Orlando, FL
Whitaker FF, Smart PL (1997) Climatic control of hydraulic conductiv- Yuhr L et al (2003) A case history of a large karst investigation.
ity of Bahamian limestones. Groundwater 35(5):859–868 Presented at the 3rd international conference on applied geophysics,
White WB (1988) Geomorphology and hydrology of karst terrains. Orlando, FL
Oxford University Press, New York
 Types of Karst Features
3

Abstract
There are a wide variety of surface and subsurface features associated with various karst
terrains. They occur at a wide range of scale due to processes both past and present.
However, not all features will be found at all locations. In general, the types of karst features
can be divided into surface landforms and subsurface features. The surface landforms
include sinkholes (dolines), areas of subsidence, sinking streams, springs and cave entrances.
The subsurface features are the areas of dissolution within the rock. These subsurface land-
forms include tertiary porosity (enlarged fracture systems), open conduits some with flow-
ing water and large caverns as well as buried sinkholes. Some of these subsurface landforms
may have already caused subsidence or collapse features at the surface. The following chap-
ter discusses the typical karst features that may be encountered. Understanding the type of
karst features that can exist in a particular geologic setting will provide insights that can be
used as a basis to guide the site characterization efforts.

3.1 Sinkholes c­ onditions and the nature or mechanism of their collapse.

There has also been a wide range of terminology used to
Sinkholes are one of the most common features we think of describe the variety of sinkholes (Table 3.1). See Waltham
when working in an area of karst geology. The terms doline et al. (2005) for a more complete discussion of sinkhole ter-
and sinkhole are generally used interchangeably with sink- minology used by others. In the United States, the term sink-
hole dominantly used in the United States. They are the vis- hole is widely used to describe any event of subsidence or
ible evidence of the impact of karst and are often the topic of ground collapse whether truly associated with dissolution of
the evening news and newspapers. Big catastrophic collapse rock or not. The term doline is used by much of the karst
and loss of life are rare, however, even a small depression of community. For all practical purposes the terms sinkhole and
a few centimeters on a high speed expressway can result in doline are the same. We will use the term sinkhole to mean
devastating vehicle accidents. Most sinkholes develop slow an area of obvious physical surface collapse (Fig. 3.2a). The
enough (typically over a few hours or even days) so that peo- term subsidence will be used to describe an area of gentle
ple have time to avoid or escape them, but on occasion they depression (Fig. 3.2b).
have occurred almost instantaneously. Property damage Sinkholes can evolve over time. For example, a sinkhole
ranges from inconvenience to catastrophic. As late as the may remain open and dry providing access into the cave sys-
1970s sinkholes were commonly considered “Acts of God” tem for exploration by cavers (Fig. 3.3a) or it may be filled
and their prediction and assessment were thought to be with water limiting access to cave divers (Fig. 3.3b). The
beyond the geologic and engineering capabilities of the day. throat of the sinkhole may be closed off with sediments
Waltham et al. (2005) have characterized six types of allowing the sinkhole to be filled with water resulting in a
sinkholes (Fig. 3.1). This classification of sinkholes provides sinkhole lake (Fig. 3.3c). In some cases, a sinkhole lake will
a common basis of terminology and a means of characteriz- periodically drain and refill. Depending upon the age and
ing the collapse features depending upon their geologic location of the collapse, it could also be naturally filled by

© Springer Science+Business Media Dordrecht 2016 17

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_3
18 3  Types of Karst Features

Fig. 3.1  Six types of sinkholes have been classified by Waltham and Fookes (2003) (Courtesy of T Waltham Geophoto)

sediments and buried so that there is little, if any, surface remediated using a small amount of backfill material.
evidence remaining (Fig. 3.3d). Figure  3.4 shows a quarry in northern Florida where the
rock surface has been exposed. These numerous small
diameter pits were created by dissolution due to vegeta-
3.1.1 A Wide Range of Sizes tion. Similar pits are seen along canal banks in south
Florida and in the easternmost Everglades. When these
Sinkholes can range in size from almost insignificant to small local voids are present and soil cover is thin, mate-
gigantic in both width and depth. The smaller subsidence rial easily settles into the voids and small sinkholes
or sinkholes are often simply a nuisance and are typically develop. These conditions often result in small nuisance-
3.1 Sinkholes 19

Table 3.1  Various terms used to describe sinkholes (After Waltham et al. 2005)
Sinkhole type (Fig. 3.1) Other terms in use
Solution sinkhole Dissolution sinkhole, doline, cockpit, or subsidence
Dropout sinkhole Subsidence sinkhole, cover collapse sinkhole, alluvial sinkhole
Collapse sinkhole Cave collapse sinkhole, cenote
Suffosion sinkhole Subsidence sinkhole, cover subsidence sinkhole, alluvial sinkhole
Caprock sinkhole Subjacent collapse sinkhole, interstratal collapse, breccia pipe, caprock collapse
Buried sinkhole Filled sinkhole, compaction sinkhole, paleosinkhole

Fig. 3.2  A sinkhole is commonly thought of as an obvious collapse (a) while a gentile depression can be thought of as subsidence (b) (Photo (a)
courtesy of T Scott, Florida Geological Survey)

type sinkholes. Figure 3.5 shows two ­examples of small of a large drop-out or cover collapse sinkhole and can be
nuisance sinkholes that are generally backfilled for seen in Fig. 3.7b.
remediation. There are also examples of much larger sinkholes. While
When the sediment cover becomes thicker, (6–15 m or encountering such conditions would be a very rare event,
more) the sinkholes become larger and can reach tens of they are mentioned here only to illustrate that such extreme
meters across with losses up to 100,000 m3 or more. conditions exist.
Figure 3.6 shows an old large sinkhole in England. Note the
person in the photo for scale reference. New sinkholes in • Investigators described “Crveno Jezero” the Red Lake
thick soil mantle constitute the most wide spread karst haz- located in the Croatian Coastal area as the world’s largest
ard (Waltham et al. 2005) and become a major hazard when sinkhole. It is 518 m deep from the upper rim to water
they occur in populated areas. levels with a diameter of 300 m at lake level. Besides its
The Winter Park Sinkhole in central Florida (Fig. 3.7a) is huge dimensions, the investigation of this great sinkhole
a very large sinkhole with dimensions of 100 × 106 m in is a fascinating story of perseverance, logistics, rigging,
diameter and 30 m deep. It consumed more than 228,000 m3 diving technology and adventure (Aspacher et al. 2000).
of sediments along with a house, large trees, a few vehicles, • Li (2004) described extremely large areas of karst in
and half of an Olympic swimming pool. This is an example China and indicated that there are about 50 known large
20 3  Types of Karst Features

a b

c d

Fig. 3.3  Sinkholes can remain open or evolve over time by filling with water or sediment. (a) Open accessible cave (b) Water-filled cave accessible
by cave divers (c) Sinkhole lake (d) Sediment-filled “buried sinkhole”

Fig. 3.4  The top of rock in a North Florida quarry was cleared of sediments exposing extensive small pits (Photo courtesy of B. Wisner 1972
Florida Department of Transportation)
Fig. 3.5  Examples of small sinkholes are due to small pits or isolated cavities at or near the top of shallow rock (a) Sinkholes along roadway in
Tennessee (b) Sinkhole along railroad in south Florida

Fig. 3.6  A very large sinkhole in England estimated to be 100 m in diameter (note person for scale)
22 3  Types of Karst Features

Fig. 3.7  The Winter Park Sinkhole in central Florida is typical of a very large sinkhole (a) Senior author at edge of collapse (b) Aerial view of
collapse (Photo courtesy of T. Scott, Florida Geological Survey)

areas of karst referred to as tiankeng in China. These are Mitchell Plain showing the pervasiveness of sinkholes
giant sinkholes, three of which are more than 500 m deep (closed depressions) throughout the area. However, sink-
and 500 m in diameter. holes may be rare to absent in many karst areas.
The number of sinkholes present and their densities are
tied to site-specific geologic conditions such as thickness
of soil cover, type of underlying rock, depth to water
3.1.2 Sinkhole Densities and Linear Trends table, the degree of dissolution or karst maturity. In many
cases, where sinkholes are present they tend to line up as
The presence of sinkholes is a clear indicator of a karst ter- a linear trend or other geometric patterns. These linear
rain. The Mitchell Plain in southern Indiana, where local trends or other g­ eometric patterns often follow main joints
sinkhole densities are over 380 per km2 is a classic example that control groundwater flow resulting in preferential
of a sinkhole plain. Figure 3.8 is an aerial photo over the dissolution.
3.2 Sinking Streams and Springs 23

Fig. 3.8  Sinkhole densities can be quite high such as the Mitchell Plain in southern Indiana (Courtesy of T Waltham Geophoto)

For example a topographic map from central Florida The sinkhole databases provide information such as the
(Fig. 3.9) shows extensive sinkholes and sinkhole lakes. The location, date of occurrence, size, shape and depth, along
sinkholes form very distinct linear trends. The linear trends with the circumstances of collapse. They provide an excellent
are associated with the subsurface geologic conditions such overview of location and concentration of sinkhole occur-
as preferential dissolution. There are also major linear trends rence and can be used for risk assessment for collapse or for
associated with topographic changes that range up to 24 m. vulnerability of groundwater contamination. These databases
A known north-south fault runs parallel less than 1 km to the are also used to calculate new sinkholes per km2 per year
west of the main road. (NSH). This term was developed by Wilson (1995) as a means
These sinkhole patterns and densities may provide initial of characterizing the sinkhole activity within an area.
clues regarding areas of dissolution and therefore areas of These maps and databases do have limitations such as
higher risk. If a site is to be located within an area of higher under-reporting, non-technical observations and errors within
sinkhole density or on a linear pattern of sinkholes it is not a the database itself. Estimates of under-reporting sinkhole
guarantee of a major cave system at depth being present. events range from 2.5 to 22 in two different areas of Florida
However, the likelihood of some dissolution feature being and from 5 to 8.5 in eastern Tennessee (Wilson 1995). The
present is much higher. databases are also biased toward more developed areas where
sinkholes are typically of greater concern and reported more
frequently. As a result of these limitations such data must be
3.1.3 S
 inkhole Susceptibility Maps used with caution. However, they are a good place to start
and Databases when assessing the potential for sinkhole activity.

Sinkhole susceptibility or risk maps have been developed for

some karst areas. An early sinkhole risk map of Florida 3.2 Sinking Streams and Springs
(Fig.  3.10) was developed by Sinclair and Stewart (1985)
and is based upon sediment cover thickness and type of Sinking streams and springs represent the recharge and dis-
­sinkholes. Many county, states and regions now have GIS- charge points for groundwater flow. A spring is where
based sinkhole information and maps which are readily groundwater discharges to the surface, while a sinking
available for use. Sinkhole susceptibility or risk maps are stream is where surface water enters the subsurface
generally regional in nature, identifying areas of sinkhole ­recharging the groundwater system. Sinking streams are also
concentration or frequency of occurrence. referred to as swallowholes or swallets.
24 3  Types of Karst Features

Fig. 3.9  A topographic map from central Florida reveals possible alignment of sinkholes (USGS Lake Louisa Quadrangle, 7.5 min Topographic,
revised 1980)

Regional recharge occurs over a sinkhole plain with makes the groundwater within a karst basin vulnerable to
d­iffuse infiltration. Local recharge is commonly concen- contamination. White (1988) discusses the various water
trated into topographic lows or individual depressions, col- resource and contaminant problems in karst.
lapse features, and along fractures. Figure 3.11a shows a In the US, springs are common in Alabama, Kentucky,
sinking stream in southeastern Minnesota where stream flow Missouri, Tennessee, Texas, Virginia, West Virginia and
enters a small sinkhole recharging the groundwater. Florida. Those states with a large number of springs also
The majority of springs are located along the perimeter of have significantly large areas of karst. However, it should be
the erosion base at the outer boundary of the karst basin or at noted that springs are not unique to karst areas but are also
the seacoast (Milanovic 2004). Figure 3.11b shows a spring found in other geologic settings.
discharging from a hillside in England. Springs are also At sea level, fresh water flows from caves and enlarged
known to occur underwater and off-shore at the outer bound- fractures into the sea. Considerable loss of fresh water occurs
ary of a karst basin. by flow from fractures and springs in coastal settings. Taborosi
In a karst environment, the presence of sinks or springs (2004) estimates that some fracture springs in Northern Guam
is controlled by site-specific hydrologic conditions that are discharge over 8 million l per day. Submarine springs have
simply taking advantage of a flow path along a fracture, been discovered at many locations throughout the world. In
joint, conduit, or cave. As hydrologic conditions change ancient times, the Phoenicians drew potable water from the
(naturally or due to man’s influence) the flow into sinks and springs in the sea bottom (Milanovic 2004). Historically sail-
out of springs will vary, can become intermittent and some- ing ships would stop in south Florida to fill flasks with fresh
times dry up all together or move to another location. water from springs within Biscayne Bay (Parks 1977).
Identifying and understanding the location and size of sinks Springs are proportional to the size of the karst water basin
and springs provides initial insight to a site’s hydrology. and are classified by their flow volume (Scott et al. 2004).
These hydrologic connections between surface and Their point of discharge is further classified as vents or seeps.
groundwater allow rapid transmission of waters within a Vents are defined as a larger cave like opening while seeps are
karst basin. This allows karst aquifers to rapidly recharge small openings with more diffuse discharge. They are also
and provide an excellent source of drinking water. This also classified as to their location on-shore or off-­shore (Scott
3.2 Sinking Streams and Springs 25

Fig. 3.10  An early sinkhole risk map was developed for the State of Florida (Sinclair and Stewart 1985)

et al. 2004). A very small flow rate would be considered a state (Scott et al. 2004). In the early years of Florida’s t­ ourism
seep and classified as a magnitude 8 spring having a flow rate development, the springs became popular locations for rec-
of less than 8 ml/s. A first magnitude spring has a very large reation. Today many of Florida’s springs continue to be
flow rate of 2,800 l/s. First magnitude springs have been major recreational areas and tourist attractions, including
mapped in Florida, Arkansas, Missouri and even Idaho. Silver Springs with its glass bottom boats, and Weeki Wachee
Springs with its mermaids. Divers have explored many of the
springs and sinkholes in the state making major contribu-
3.2.1 Springs in Florida tions to understanding our karst system (Stamm 1994).
Numerous submarine springs exist off Florida’s Atlantic
More than 700 springs have been recognized in the state of and Gulf of Mexico coastlines (Scott et al. 2004). While
Florida, with 33 first magnitude springs, more than any other most offshore springs today are concentrated in the northern
26 3  Types of Karst Features

Fig. 3.11  A small sinking stream in southeastern Minnesota (a) and a moderate spring flowing from a hillside in England (b)

portion of the state, there is also a history of springs in from the offshore spring) at an elevation of approximately
­southeast Florida, where sailing vessels stopped in the Miami 33 m above sea level (Swarzenski and Holmes 2000).
area to replenish their fresh water supply. Historic photos Unfortunately some of the major spring systems in
have shown sailors filling barrels with fresh water from a Florida are being affected by surface water runoff leading
spring in the Biscayne Bay in South Florida (Parks 1977). As to pollution of the spring system. An increase of tenfold in
a result of development in the early twentieth century, water nitrate concentration has occurred in 13 of the first magni-
levels were lowered which were the driving force for these tude springs of Florida (Scott et al. 2004). Extensive study
springs. All of the springs in the southeastern portion of the of the Wakulla spring system has clearly demonstrated
state now have little to no flow. man’s impact on the springs of Florida (Kincaid et al.
In contrast, one large offshore spring still flows in the 2012).
Gulf Stream 4 km off of Crescent Beach in northeastern An understanding of the presence and location of springs
Florida. The spring occurs at a water depth of 18 m with the begins to form the basis of characterizing the karst basin,
vent extending 38 m below the bottom. Its discharge is con- groundwater flow regime and level of risk from potential
fined to a localized area and flow greatly exceeds that of a contamination. While many older springs are no longer flow-
first order spring. Recharge occurs from lakes and sinkholes ing, their permeable pathways remain and may impact both
in an area east of Gainesville Florida, (some 120 km inland geotechnical projects as well as groundwater flow.
3.3 The Epikarst Zone 27

3.3 The Epikarst Zone of highly-fractured and dissolved bedrock (Fig. 3.14).

However, the depth of weathering may exceed 100 m espe-
The term epikarst is used to describe the “skin of the karst”. cially in humid tropical climates (Fookes et al. 2005). The
Engineers commonly referred this zone as the “top of rock” amount of rock removed by dissolution within the epikarst
or “rockhead”. This is the dissolutionally weathered upper varies from less than 1 % to more than 50 %. The percent of
portion of the bedrock where the waters are more aggressive the bedrock void volume that is filled with sediment within
and maximum dissolution has taken place. the epikarst can range from less than 5 % to more than 95 %,
Exposed rock without sediment cover can be highly dis- with the higher percent values being the most common
solutioned and is referred to as karren karst. Figure 3.12 (Aley 1997).
shows two examples of extreme karren karst, one is from The epikarst zone typically has lower overall permeabil-
Manitoulin Island, Canada and the other is from England. ity than underlying portions of the bedrock and can function
However, many of the areas of dissolutioned rock are com- as a perched aquifer providing appreciable water storage. It
monly covered by sediment and referred to as the epikarst. may be separated from the saturated zone. In contrast, the
The epikarst can range from thin, almost non-existent to epikarst zone may also provide a means to convey water lat-
tens of meters thick or more. The top of rock conditions can erally over large distances. Flow into the epikarst zone is
range from hard (older, well consolidated and cemented more rapid than flow out of it. Discharge from the epikarst is
(Fig. 3.13a) to highly weathered rock (Fig. 3.13b). Typically by limited highly permeable vertical pathways transmitting
the epikarst is commonly about 10–15 m thick, and consists water downward (Aley 1997).

Fig. 3.12  Exposed limestone with extensive dissolution along secondary fractures is referred to as limestone pavement or karren fields (a) Karren
karst in Manatoulin Island, Canada (b) Extreme karren karst in England (Photo courtesy of T Waltham Geophoto)
28 3  Types of Karst Features

Fig. 3.13  Epikarst, top of rock or rockhead can range from massive rock (a) to highly weathered rock (b)

Industrial sites or roadside spills underlain by complex aquifers is a layer of clay over the top of highly weathered
epikarst zones can have significant difficulties with environ- rock forming an aquitard. At this site, the epikarst has been
mental remediation. The epikarst can provide storage, lateral referred to as the semi-confining layer (SCL). Its presence is
flow and even multidirectional flow. Therefore, a single critical for limiting the downward migration of contaminants
point-source of contamination can often result in a pattern of as well as minimizing sinkhole development.
contaminant distribution, that suggests multiple sources of The highly variable geologic and hydrologic conditions
contamination. In contrast, contaminants within the epikarst found within the epikarst make it an important part of any
may also move slowly or remain trapped in pockets. site characterization effort. Highly variable rock conditions
In some cases, the epikarst zone may function as an aqui- beneath a soil cover inevitably provide greater geohazards
tard or confining layer (see case history in Chap. 27). At this because of conditions are obscured and difficult to define by
site, there is a surficial aquifer consisting of sands over a drilling alone (Waltham and Fookes 2003). In addition, the
deeper limestone aquifer. The epikarst separating these two highly variable hydrologic conditions present a complex
3.4 Caves 29

a b

Soil

Epikarst Zone Usually High

Water Storage

Local discharge
Unweathered of water via
Rock tertiary fractures

Fig. 3.14  Epikarst conditions are complex both in terms of geologic and hydrologic conditions (a) Geologic model of epikarst (b) Hydrologic
model of epikarst

zone for environmental cleanup of contaminants. In some

cases, the true conditions will only be known when the rock
is exposed. For a further discussion of the epikarst see Jones
et al. (2004).

3.4 Caves

Caves are only explored and mapped where an opening can

be found to provide human access. By definition, caves are
openings into rock that can be explored by humans for an
appreciable distance. This implies that the cave has a mini-
mum dimension of 0.5 m, and often much larger. It should be
recognized that this size limitation for direct exploration
implies that there are probably many smaller conduits and
fissures that are part of the cave network that have not been
identified. It is also possible that some larger portions of a
cave system cannot be explored because of restricted access
or fallen rock blocking a passage (Fig. 3.15). As a result, a
cave map may not necessarily represent the complete cave
Fig. 3.15  Piles of fallen rock can nearly block off access in a cave and
system. in some cases will completely block off access (Courtesy of T Waltham
Culver et al. (1999) have developed a map of caves in the Geophoto)
US that shows the location of nearly 45,000 caves. The loca-
tion, mapping and characterization of caves provide us with
considerable insight to regional karst conditions. These maps who have formed local groups called grottos. These
can be invaluable in the assessment of local and site-specific groups are actively involved in locating, mapping, and
karst conditions, if located nearby. managing caves. The Grotto locations, names and contact
In the US, a great deal of information about caves can information can be found by contacting the NSS on-line at
be obtained from state agencies who maintain such www.caves.org. In addition, there are cave organizations,
records and by consultation with the National Speleological speleological societies, throughout the world with similar
Society (NSS) and their caving members and cave divers missions.
30 3  Types of Karst Features

If available, cave maps can provide considerable levels that are controlled by nearly horizontal bedrock
i­ nformation, well beyond its location. Cave development can ­conditions. Figure  3.17 is a partial cave cross section from an
occur along bedding planes, fractures, or structure and can early cave map that shows the periodic development of larger
be a combination of them. The level of information that has chambers in the underwater Alachua Cave System north of
come from cave explorations and mapping provides evi- Gainesville, Florida. These periodic caverns within the cave
dence of dissolution patterns, preferred depth of dissolution have most likely developed at intersections of fractures with
and preferred formation for dissolution. The cross section of zones of weakened rock. One of these enlarged chambers has
a cave will allow us to make an assessment of minimum and broken through to the surface providing the divers access. By
maximum depths of a cave and its relation to the geologic noting the spacing of the larger cavern development, one
­stratigraphy. It will also reveal the location of any large might be able to predict where the next large cavern with a
chambers, if they are present. higher risk of subsidence might be located. One might also
expect to find near surface indications of the presence of
these caverns associated with lineaments caused by the frac-
3.4.1 Cave Geometry and Densities ture zones and vegetation associated with their recharge.
By understanding local geologic fabric or geomorphol-
The development of most caves is influenced by local geol- ogy of a cave system one can begin to make informed deci-
ogy (Palmer 2007). This includes the geologic strata (bed- sions as to the trends of a cave system beyond its mapped
ding planes), fractures, faults, folds and structure along with extent and identify what possible unique features may be
dip and strike which all play a role in cave development and present for a particular site. Invaluable as they are, cave maps
its geometry. Conduit development begins along the second- only represent those areas of a cave system that are accessi-
ary porosity zones driven by recharge and groundwater flow. ble and only those caves that have been explored and mapped.
Those conduits with higher flow eventually become domi- The density of cave conduits has been estimated to be a
nant and take on various forms. Cave maps can often reveal relatively small portion of the total surface area, which is
many characteristics such as whether the cave is fracture or known to contain caves. Quinlan (1991, personal
bedding plain controlled, which geologic formation it is in ­communication) suggests that the Mammoth Cave System
and its minimum depth of rock cover. Cave systems are dis- has approximately 585 km of conduit within an area of
cussed in detail by Palmer (2007). approximately 90 km2. Assuming an average conduit dimen-
For example, Fig. 3.16 is the cross section of Sorcerers sion of 7.5 m this implies that about 5 % of the surface has a
Cave in southwest Texas that was mapped by Veni (1980). conduit under it. Worthington et al. (2000) suggest that the
This cave system has horizontal development at two main cave passages of Mammoth Cave underlie between 0.36 %

Entrance
0

-20
Elevation in Meters (MSL)

-40
Projected Profile

-60
Two depths of lateral cave
-80 development dominate

-100

-120

-140

0 30 60
Scale in Meters

Fig. 3.16  The cave map of Sorcerer’s Cave in southwest Texas indicates linear cave development at two levels. Such data provides insight as to
other possible areas of development in the area (Veni 1980)
3.4 Caves 31

+15

0 Water Level
Depth in Meters
Limestone
-15

Water
Filled
-30

-45
Approx. 30m continues on

-60

Fig. 3.17  An initial cross section map of Alachua cave developed by cave divers shows enlarged caverns occurring at intervals along the cave
system (Mount 1973)

and 3.3 % of the surface area. The extremely enlarged listing of longest underwater caves, mostly in Florida and
­caverns or dome areas within a cave system are limited to an Mexico, some lengths of more than 106 km (combined
even smaller percentage of area, probably something on the total passage length) with depths ranging from 8 to 120 m.
order of 0.1–1 % or less. Cave systems in the Yucatan peninsula exceed 200 km.
• The Deepest Cave: By 2007 there were several caves over
1,500 m deep, with early reports of one 2,000 m deep.
3.4.2 L
 arge Cave Systems Develop in Thick The current record for the deepest cave system is Krubera
Massive Limestone (Voronja) Cave in the Western Caucasus, Republic of
Georgia at about 2,191 m (Williams 2012).
Large cave systems develop best in massive competent, • The Largest Chamber: The largest cave chamber is the
limestone (Waltham and Fookes 2003). An extensive system Sarawak Chamber in Malaysia, which is approximately
of solution cavities and large caverns is known to exist in the 700 m long, 400 m wide and at least 70 m high, (Dixon
thick (over 120 m) Mississippian limestone and dolomite 2011).
beds throughout the mid-continent of the US (Gentile 1984).
Finding the deepest, longest and largest caves has been
the “holy grail” of those adventurers who are pushing the
limits of cave knowledge. The following are some examples 3.4.3 Other Types of Caves
of extreme dimensions and depths of larger caves. While
these large features are rare they are included to illustrate the While the majority of caves are formed as a result of dissolu-
possible worse case conditions for geotechnical concerns. tion (karst), caves can be formed in a number of other ways
and in a wide variety of materials. These caves include:
• The Longest Cave: The benchmark for the longest cave
system has been 100 km. There are currently 14 cave sys- • Lava tube caves which form in volcanic rock as magma is
tems longer than 100 km with Mammoth Cave system in deposited at the surface during an eruption and then is
Kentucky being the longest, in fact, longer than the next cooled.
two combined. The Mammoth Cave System includes • Coastal caves which include both sea caves and flank
approximately 643 km of complex maze passageways margin caves. Sea caves are formed in all types of rock
(Gulden 2012). and are due to mechanical erosion of a weak zone within
• Longest Underwater Caves: As a result of the introduction the rock. Flank margin caves are formed at the outer edges
of SCUBA equipment in mid-1950s, divers began to ven- of the fresh-saltwater interface due to dissolution (Mylroie
ture into springs and sinkholes. Exley (2004) provides a 2005).
32 3  Types of Karst Features

• Structural caves which form due to tectonic movement of encourages its members to cooperate with the federal
rock forming a void or cave system. ­inventory (Cave Conservationist Newsletter 1994).
• Ice caves which form in glacial areas.
• Caves which form in sandstones, quartzite and granite.
References
White and Culver (2005) as well as Palmer (2007)
describe the various types of caves in more detail. Those Aley T (1997) Groundwater tracing in the epikarst. In: Beck B, Stephenson
caves formed by a process other than dissolution would be JB (eds) Proceedings of the 6th multidisciplinary conference on sink-
considered natural pseudokarst. While all caves are not nec- holes and the engineering and environmental impacts of karst, the
engineering geology and hydrogeology of karst terranes, Springfield,
essarily formed the same way, the void space they create can Missouri. Balkema, Rotterdam, Netherlands, pp 207–211
potentially impact a geotechnical or environmental project in Aspacher B et al (2000) Beyond the blue. NSS News 58(5):13–20
the same way. Cave Conservationist (1994) Newsletter of cave conservation and man-
agement. NSS 13(1), February, 24 p.
Culver DC et al (1999) Distribution map of caves and cave animals in
the United States. J Cave Karst Stud 61(3):139–140
3.4.4 S
 ecrecy and Discretion as a Cave Dixon K (2011) Measuring Sarawak Chamber. The Mulu caves project:
Management Tool exploring the world’s largest caves. www.mulucaves.org. Accessed
21 June 2013
Exley S (2004) The taming of the slough, a comprehensive history of
Obtaining cave locations and maps from both state and pri- peacock springs. In: Poucher S (ed) National Speleological Society,
vate groups is becoming more difficult because of concerns Huntsville, Alabama, p 170
regarding damage to the cave systems as their locations Fookes PG et al (eds) (2005) Geomorphology for engineers. Whittles
become more commonly known. The issues of concern range Publishing, Dunbeath, Scotland
Gentile RJ (1984) Paleocollapse structures: longview region, Kansas
from destruction of delicate geologic features, plundering City, Missouri. Bull Assoc Eng Geol 21(2):229–247
archaeologically significant caves, (some of which contain Gookin J (1997) Secrecy and discretion as a cave management tool.
pictographs and burial remains), destruction of unique habi- NSS News 55(4):114–136
tats and contamination. Gookin (1997) provides a summary Gulden B (2012) Worlds longest caves. www.caverbob.com Accessed 3
June 2006
of some of the threats to caves. Jones WK et al (eds) (2004) Proceedings of the 2003 epikarst sympo-
Even as a member of the NSS, the senior author has sium. Karst Waters Institute Special Publication 9
encountered some problems with secrecy in attempting to Kincaid TR et al (2012) Demonstrating interconnection between a
determine the presence of caves near a low level radioactive wastewater application facility and a first magnitude spring in a
karstic watershead: tracer study of the southeast farm wastewater
waste site for the Department of Energy in Missouri. While reuse facility, Tallahassee, Florida. Florida Geologic Survey, Report
the state agency provided the number of caves within a spec- of Investigation 111
ified radius of the site they would not provide the locations Li F (2004) Introduction to karst Tiankeng in China. NSS News
and even shielded the computer printout from view. On 62(8):220–221
Milanovic T (2004) Water resources engineering in karst. CRC Press,
another site in West Virginia the local cavers provide the Boca Raton
number of caves and general location within a specified Mount T (1973) Preliminary Alachua cave map. Consultant report
radius of the site but avoided providing any further details. Mylroie JE (2005) Coastal caves. In: David CC, White WB (eds)
In general, the caving community is reasonably helpful in Encyclopedia of caves. Elsevier, New York, pp 122–125
Palmer AN (2007) Cave geology. Cave Books, an affiliate of the cave
providing limited data if one clearly identifies themselves research foundation, Dayton, Ohio
and explains the need for the data and does not become too Parks AM (1977) The forgotten frontier, Florida through the lens of
intrusive with the inquiry. Work closely with the caving com- Ralph Middleton Monroe. Banyan Books, Miami
munity and do not publish cave locations. Scott TM et al (2004) Springs of Florida. Florida Geologic Bulletin
No. 66
Under the Federal Cave Resources Protection Act (1988), Sinclair WC, Stewart JW (1985) Sinkhole type, development and distri-
federal land managers are required to inventory all known bution in Florida, Florida Geologic Survey, Map Series 110
caves so they can be mapped and safeguarded against van- Stamm D (1994) The springs of Florida. Pineapple Press, Sarasota
dalism and exploitation. Many cavers insist that the best way Swarzenski PW, Holmes CW (2000) Re-examining the submarine
spring at Crescent Beach, Florida. USGS Open File Report 00-158
to protect caves is to keep their locations secret. After years Taborosi D (2004) Field guide to caves and karst of Guam. Bess Press,
of struggling with the issue, the National Speleological Honolulu
Society established a federal cave management policy that Veni G (1980) Sorcerer’s cave map. In: The Texas caver, February 25(1)
References 33

Waltham AC, Fookes PG (2003) Engineering classification of karst Wilson WL (1995) Sinkhole and buried sinkhole densities and new
ground conditions. Q J Eng Geol Hydrol 36:101–118 ­sinkhole frequencies in karst of northwest Peninsular Florida. In:
Waltham AC et al (2005) Sinkholes and subsidence, karst and cavernous Beck BF (ed) Proceedings of the 5th multidisciplinary conference
rocks in engineering and construction. Springer-Praxis, Chichester, UK on sinkholes and the engineering and environmental impacts of
White WB (1988) Geomorphology and hydrology of karst terrains. karst: karst geohazards, Gatlinburg, Tennessee. Balkema,
Oxford University Press, New York Rotterdam, pp 79–91
White WB, Culver DC (2005) Cave definition of. In: Culver DC, White Worthington SRH et al (2000) Matrix, fracture and channel compo-
WB (eds) Encyclopedia of caves. Elsevier, New York, pp 81–84 nents of storage and flow in a Paleozoic limestone aquifer. In:
Williams RA (2012) Braving caves. National Geographic Magazine, Sasowsky ID, Wicks CM (eds) Groundwater flow and contaminate
National Geographic Society, Washington, DC, 224(2):122–123, transport in carbonate aquifers. Balkema, Rotterdam, Netherlands,
August pp 113–128
 Karst Maturity and Development
4

Abstract
This chapter introduces the concepts of maturity and karst development, which can span
millions of years. The type of karst features present at a site can provide clues as to where
in the process of development it is. Understanding the maturity of a site will aid in under-
standing the conditions that exist and aid in assessing the degree of their stability or instabil-
ity. These concepts help us understand this complex geologic system and put it into the
context of development from initiation to decay.

4.1 Karst Maturity in southern Indiana (Fig. 3.8) is most likely Juvenile or
Youthful Karst. The karst class in Puerto Rico (Fig. 4.2) is
Waltham and Fookes (2003) both engineering geologists, most likely Mature Karst. The karst class in Guilin, China
have proposed an engineering classification of karst that (Fig. 4.3) is Extreme Karst. These evolutionary stages repre-
describes the karst in terms of degree of maturity, complexity sent increasing geologic complexities and concern for engi-
and geomorphology along with specific features that may neering projects.
impact engineering projects. Waltham et al. (2005) have con- It should also be noted that very mature karst conditions
tinued these ideas in their recent book on sinkholes and sub- could have a low number of new sinkholes per km2 per year
sidence. They have classified karst maturity into five levels, (NSH) (Table 4.1). Such examples occur in Puerto Rico and
(Table 4.1 and Fig. 4.1). The five levels of karst classification the cockpit karst of Jamaica. This does not mean that mature
generally increase in number, size and variety of karst fea- karst is completely stable. Old, existing mature karst, sink-
tures from Juvenile to Extreme. This classification presents hole lakes or paleokarst can contain open voids that may
the karst as a system illustrating the range features and their impact engineering projects and can sometimes be
development. It provides an excellent focus for one’s early reactivated.
thinking about site conditions and also provides a means for
effective communication with team members and discus-
sions with the client/owner. 4.2 Karst Development Time Scale
In many cases, without intimate site knowledge and expe-
rience, it may be difficult to decide upon one of the five clas- It is important to place the stages of karst development and
sifications. It may be better to start by using just three levels, maturity in context with time. Geologists have divided
Juvenile, Mature, and Extreme. For example, the karst class Earth’s history into several eras, which are further partitioned

© Springer Science+Business Media Dordrecht 2016 35

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_4
36 4 Karst Maturity and Development

Table 4.1 Karst classification based upon maturity (Waltham and Fookes 2003)
NSH (new sinkholes per
Classification Sinkholes Caves Rockhead relief km2 per year)
Juvenile Karst Few, barely developed Few, 1 m wide, minimum Minimal <0.001
collapse
Youthful Karst Common but <10 m dia Larger cave development Some fissures and relief 0.001–0.05
beginning
Mature Karst Frequent, 100 m diameter Large caves present Fissures and relief is 0.05–1.0
common
Complex Karst Large sinkhole dominant, Large caves common Fissures and extreme 0.5–2.0
1 km diameter relief
Extreme Karst Largest size and variety Large caves common with Extensive fissures and >>1.0
of features, only found in cavernous zones and collapse extreme relief
wet tropics

into periods. The most recent periods have been further The process proceeds from initiation to enlargement over
divided into epochs (Fig. 4.4). Most caves and karst date a period of 3,000–5,000 years. Eventually degradation begins
from the late Cenozoic Era (Neogene Period) and represent after about 10,000–100,000 years and finally decay after
less than 0.1 % of the Earth’s history. However, the rocks that about 1–10 million years (White 1988; Esteban and Wilson
contain caves can be of almost any age. The oldest rock that 1993) (Fig. 4.5). This corresponds to the evolution of karst
contain large solution caves are Precambrian, with ages of up stages described by Waltham and Fookes (2003) from
to about 2.3 billion years, although the caves themselves are Juvenile, Youthful, Mature, Complex and Extreme Karst.
much younger (Palmer 2007). These evolutionary stages also represent increasing concern
A karst system is initiated and evolves by a variety of con- for engineering and environmental projects.
ditions over time. Karst voids and caverns are created by a The eventual decay and collapse of a karst system may
combination of dissolution (corrosion), erosion (corrosion) evolve in various ways:
and incasion (collapse). Figure 4.5 illustrates the major
events in the life of a karst system from its initiation to decay. • After collapse, it may remain partially open as small to
Initially laminar flow occurs through secondary fractures. As large isolated voids.
dissolution occurs the fractures are widened. At some critical • Bulking of rock may occur filling the cavern with porous
threshold turbulent flow begins to occur leading to accelerat- rubble or breccia which may become recemented.
ing rates of enlargement resulting in tertiary fractures. As • The system may be eroded from above resulting in a sta-
dissolution continues the fractures become enlarged and ble valley.
selective ones become dominant resulting in large conduits • Or some combination of these may occur.
and in some cases larger caverns may develop.
4.2 Karst Development Time Scale 37

Fig. 4.1 Five conceptual models of karst maturity developed by Waltham and Fookes (2003). These engineering classifications of karst range from
simple young juvenile system to a well developed extreme system (Courtesy of T Waltham Geophoto)
38 4 Karst Maturity and Development

Fig. 4.2 The karst of Puerto Rico is likely a Mature Karst. Note two of the magotes (the term used in Puerto Rico to describe these features) have
been cut away for road development

Fig. 4.3 The karst of Guilin, China is likely Extreme Karst

4.2 Karst Development Time Scale 39

ERA PERIOD EPOCH

0 Holocene
Quaternary Pleistocene

Paleogene gene
Neo-
Cenozoic
Miocene Pliocene

Tertiary Oligocene
50 Eocene

Time - millions of years ago

Paleocene

Mesozoic
Cretaceous
100

Jurassic
200
Triassic
Permian
300 Pennsylvanian
Paleozoic Mississippian
400 Devonian
Silurian
Ordovician
500
Cambrian
600
cambrian
Pre-

Fig. 4.4 Abbreviate geologic time table

Initial Karst Enlargement Degredation & Decay

100

Collapse

Large Caverns
t
en
ul
rb
10 Tu
Conduit Diameter (m)

Large Conduits

ty
1 ci
lo
Tertiary Porosity/ Ve
Open Fractures
ow
Fl

Dry Abandoned
Cave
0.1
Secondary Porosity/
Fractures
ar
in
m
La

0.01
103 104 105 106 107
Time (years)
Karst maturity Juvenile
terms used by Youthful
Waltham and Mature
Complex
Fookes (2003) Extreme Senile

Fig. 4.5 Karst development and its evolution, compiled from information presented by White (1988), Esteban and Wilson (1993), and Waltham
et al. (2005)
40 4 Karst Maturity and Development

References Waltham AC, Fookes PG (2003) Engineering classification of karst

ground conditions. Q J Eng Geol Hydrol 36:101–118
Waltham T et al (2005) Sinkholes and subsidence, karst and cavernous
Esteban M, Wilson JL (1993) Introduction to karst systems and paleo-
rocks in engineering and construction. Springer-Praxis, Chichester
karst reservoirs In: Fritz RD, Wilson JL, Yurewicz DA (eds)
White WB (1988) Geomorphology and hydrology of karst terrains.
Paleokarst related hydrocarbon reservoirs, SEPM Core Workshop
Oxford University Press, New York
No. 18:1–9
Palmer AN (2007) Cave geology. Cave Books, an affiliate of the cave
research foundation, Dayton
 Areas Affected by Karst
and Pseudokarst 5

Abstract
This chapter briefly discusses the presence of karst and pseudokarst in the US and
worldwide. Soluble rocks exist at or near the surface over approximately 25 % of the
US. If pseudokarst and deep conditions are included up to 75 % of the US may be affected.
Similar conditions may exist worldwide. Knowing where these conditions exist will help
in the planning and management for land use, groundwater resources, mineral resources
and unique habitats.

5.1 United States Institute (AGI) presented this map in a document Living with
Karst. Weary and Doctor (2014), with the USGS developed
Davies et al. (1984) developed a map of the US showing new digital maps of the United States, Puerto Rico and the
areas affected by karst and pioneered the topic of karst and US Virgin Islands. They identified areas having karst or the
its engineering aspects during the 1960s and 1970s. Based potential for karst and pseudokarst using GIS maps devel-
upon his map, limestone or other soluble rock is present at or oped for various states. Their digital maps are available as
near the surface over 25 % of the United States. The most downloadable files at various scales from USGS. Figure 5.1
extensive karst in the United States occurs in the limestone’s is the 2014 digital map of karst for the contiguous 48 states.
of Mississippian age (about 325–345 million years old) These karst maps are an invaluable resource of regional
(Gentile 1984) which covers the mid-US continent. trends for use by engineers, geologists, land planners and
Davies (1977) in a keynote address at an early Symposium others as a starting point for an investigation.
on Detection of Subsurface Cavities at the US Army The mapping of karst, paleokarst and pseudokarst condi-
Waterways Experiment Station in Vicksburg, Mississippi tions is continually being updated. Improvements are being
estimated that: made as terminology is unified, map projections are stan-
dardized and more accurate, complete and detailed informa-
• Limestone and other soluble rock were present at or near tion is gathered.
the surface over approximately 25 % of the United States
• Adding natural pseudokarst (such as lava caves) increases
the area to about 32 %; 5.2 Worldwide
• When withdrawal of fluids (water and oil) and subsidence
due to mining are included the area increases to about Carbonate rock is widespread and karst occurs to some
54 %; and extent in almost every part of the world. Estimates of the
• If deeper conditions to 1,000 ft and more (paleokarst) are amount of carbonate rocks found at or near the earth’s sur-
included, about 75 % of the U.S. is affected. face range from 15 %, (Sweeting 1973) to 20 % (Ford and
Williams 2007). An estimated 25 % of the world’s popula-
The 1984 map produced by Davies et al. was updated by tion is supplied water from karst aquifers (Ford and Williams
Veni et al. (2001) and included both karst (carbonate and 2007). Ford and Williams (2007) as well as Palmer (2007)
evaporate) and pseudokarst. The American Geological discuss the distribution of karst and caves worldwide.

© Springer Science+Business Media Dordrecht 2016 41

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_5
42 5 Areas Affected by Karst and Pseudokarst

Humid Climate Karst

Carbonate rocks
Unconsolidated calcareous or
carbonate rocks
Evaporite rocks
Quartz sandstone
Dry Climate Karst
Carbonate rocks
0 500
Unconsolidated calcareous or carbonate rocks
Kilometers
Evaporite rocks

Fig. 5.1 A map of the contiguous United States shows a general representation of karst and pseudokarst areas. Although most karst is exposed at
the land surface, some is buried under layers of sediment and rock (Weary and Doctor 2014)

Karst conditions are commonly thought of as occurring in world. The impact of mans activities results in a wide range
a tropical or semi-tropical environment such as Mexico, of subsidence and collapse problems.
Florida, Guam, the Caribbean and Central America, South Efforts are being made to improve the worldwide maps of
Pacific, and Asia. However, we also find karst in temperate karst. One such effort was a thesis supported by The Nature
climates and even in the deserts of the middle east and Conservancy and the University of Arkansas which began
Afghanistan. In desert environments the sinkholes and caves efforts to develop the Karst Regions of the World (KROW)
were formed under historically different climatic conditions (Hollingsworth et al. 2008). This was an effort to develop and
when the area had a wetter climate. Paleokarst, occurs over populate an interactive database as well as output maps on a
much of the world between 30 and 60° latitude (Bosak et al. continental and worldwide scale showing areas of near-surface
1989). Gypsum, anhydrite and salt underlie more than karst (within 100 m of the surface). This work however does
20–25 % of the worlds land surface. More than 90 % of evap- not include buried or deep-seated karst and focuses on surface
orates, including all of the salt, is buried under other rocks areas of karst that contain fragile ecosystem. This effort is
(Kozary et al. 1968). being continued by a team of researchers that include the
Pseudokarst caused by mans activities such as pumping USGS, the National Cave and Karst Research Institute and the
of groundwater or oil and gas as well mining of coal, salt, Nature Conservancy as part of the University of South Florida
gypsum, and other resources also extends across much of the libraries’ Karst Information Portal (www.karstportal.org).
References 43

References Hollingsworth E et al (2008) Karst regions of the world (KROW): global

karst datasets and maps to advance the protection of karst species and
habitats worldwide. In: Kuniansky E (ed) US Geological Survey Karst
Bosak P et al (eds) (1989) Paleokarst a systematic and regional review,
Interest Group Proceedings, Bowling Green, Kentucky May 27–29.
vol 1, Developments in earth surface processes. Elsevier, New York
USGS Scientific Investigations Report 2008–5023. Reston, Virginia
Davies WE (1977) Keynote presentation to the symposium on detection
Kozary MT et al (1968) Incidence of saline deposits in geologic time.
of subsurface cavities. US Army Engineer Waterways Experiment
Geol Soc Am Spec Pap 88:43–57
Station, Vicksburg (Lecture recorded by Benson RC)
Palmer AN (2007) Cave geology. Cave Books (An affiliate of the Cave
Davies WE et al (1984) Engineering aspects of karst, a map of the karst
Research Foundation), Dayton
and pseudokarst of the United States. US Geologic Survey US
Sweeting MM (1973) Karst landforms. Columbia University Press,
Geological Survey, National Atlas of the United States of America,
New York
scale 1:7,500,00
Veni G et al (2001) Living with karst, a fragile foundation, Awareness
Ford D, Williams P (2007) Karst hydrogeology and geomorphology.
Series No. 4. American Geologic Institute, Alexandria
Wiley, New York
Weary DJ, Doctor DH (2014) Karst in the United States: a digital map
Gentile RJ (1984) Paleocollapse structures: longview region, Kansas
compilation and database. USGS Open File Report 2014–1156
City, Missouri. Bull Assoc Eng Geol 21(2):229–247
 Karst and Its Many Benefits
6

Abstract
There are a number of benefits to karst, from its abundant source of water to its mineral
resources and recreational aspects. Karst landforms provide a wide array of picturesque
scenery and conditions providing many opportunities for recreational activities such as
swimming, canoeing, fishing and boating. Most of us have experienced these benefits either
directly or indirectly by spending recreational time at a spring; visiting a cave as a tourist
attraction; utilizing groundwater from a karst spring or aquifer; benefiting from the mineral
resources found in some karst terrains.
In addition, some caves and pseudokarst features such as man-made caves, mines and
tunnels have all provided benefit to man for shelter or commercial use. This chapter briefly
summarizes some of the benefits that have been realized in karst terraines.

streams, and in sinkholes of Florida. Sites have been found

6.1 Springs along ledges within sinkholes and at springs and streams
which are now well below sea level. These sites were occu-
Springs discharging from karst aquifers and their associated pied during lower sea levels. Little Salt Spring is located near
streams and rivers have attracted both animals and humans to the west coast of Florida some 96 km south of Tampa. The
their cool refreshing source of fresh water. The vital role of sinkhole has an hour glass shape which is about 73 m diam-
karst groundwater dates back to pre-historic times. Desert eter at the surface and is as much as 67 m deep. There are
people all came to know the critical locations of fresh water. ledges at approximately 16 and 26 m deep where preserved
Indians and settlers located their villages at springs and along ancient artifacts dating back 8,000 to more than 12,000 years
rivers that provided both food and water. have been found. More than 10,000 years ago the water table
Today much of the bottled water comes from springs. One was much lower and the ledges within the sinkhole were
bottled-water company in Florida lists five springs as the accessible to animals and humans (Gifford 2008 personal
sources of their water. Kentucky spring water is perfect for communication).
bourbon distilling because it is free of minerals that affect The springs and their rivers of Florida are used exten-
taste. Bourbon whiskey has been called “one of America’s sively for recreation including swimming, fishing, various
unique cultural contributions to the world” (Fryar 2009). forms of boating, canoeing and floating. In addition, many of
Springs have provided resources to pre-historic people the springs and water-filled sinkholes provide access for cave
and wildlife during a time when sea level was much lower. divers to explore and map these underwater cave systems
Archeology finds have been made at many springs, spring (Stamm 1994).

© Springer Science+Business Media Dordrecht 2016 45

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_6
46 6 Karst and Its Many Benefits

6.2 Caves stone. The depression provided a natural geometry for the
305 m reflector. Originally built in the early 1960s, with a
Tourist caves are open to the public and are found throughout primary goal to study the ionosphere. Today, with many
many countries. These caves provide safe access for tourists upgrades, it is the site of the world’s largest single unit radio
to view the wonders of the underground network of caverns telescope, which employs its reflector to listen for radio sig-
and their spectacular formations of stalactites and stalag- nals from possible life in space (NAIC 2014).
mites, large chambers, and hidden rivers.
Caves have also been used as shelters since the dawn of
mankind. Many of the world’s greatest archeological sites 6.4 Karst Aquifers and Groundwater
have been found in caves that have preserved evidence of Resources
prehistoric cultures including early art and records of mans
early activity. The Dead Sea Scrolls were placed in a cave for Karst aquifers supply about 25 % of the worlds drinking
safe storage and found by a Bedouin boy in 1947 (Tushingham and irrigation water and are among the most prolific and
1958). To the Maya Indians, the cave was a realm between important aquifers. Many regions including arid coastal
the supernatural and the source of life. Archeologists have areas and island nations of the Caribbean depend almost
discovered brightly painted clay pottery and stone alters dat- entirely upon water from karst aquifers (Ford and Williams
ing back 2,000 years, evidence of cave rituals carried out by 2007). The extremely high porosity and often cavernous
the Maya (Roberts 2004). nature of a karst aquifer allows large volumes of water to be
Natural caves have been used as places of refuge from stored underground. In some parts of the world, cave
prosecution in biblical times through World War II. During streams are large enough to economically allow damming
World War II a group of Jews hid and survived in a cave in to store water or to utilize the flow for hydroelectric power
the Ukraine. On May 5, 1943 38 Ukrainian Jews started a (Milanovic 2004).
344 day underground stay to avoid capture. The cave was
called Priest’s Grotto and is one of the fourth longest caves
(124 km) in the world today. Cris Nicola tells this fascinating 6.4.1 The Edwards Aquifer
story in the NSS News (Steele 2005), where he discusses his
experience with the cave and it’s few living survivors. The Edwards aquifer of central Texas is one of the most pro-
Other uses include caves on islands in the South Pacific lific aquifers in North America. It occupies 1,035 km2, ranges
which were very effectively used by the Japanese to defend from 137 to 183 m thick and is the primary source of water
against attach by US forces in World War II. Today some for approximately 1.7 million who live in the region and pro-
caves continue to be used as emergency shelters during hur- vides much of agricultural and industrial water needs. A well
ricanes. In 2004, residence of the Cayman Islands took shel- drilled in 1991 is reportedly the worlds greatest flowing well
ter in the island caves during a hurricane, as they had in with a discharge of about 95,000 l/min.
previous storms. The Edwards Aquifer has an extremely high porosity and
In an era prior to modern refrigeration, the brewing indus- permeability, characteristic of many karst aquifers. The aqui-
try used the caves of St Louis, Missouri. In the early 1800s fer has been subjected to several uplifts, major faulting and
the beer making industry utilized these caves for cold stor- karstification. Faulting and subsequent dissolution along
age. By 1860, there were some 40 breweries operating and fractures create a very heterogeneous and anisotropic perme-
using caves for storage (Rother and Rother 2004). ability distribution. This karst system includes sinkholes,
sinking streams, caves, springs, and an extensive system of
subsurface drainage. The stratigraphic and structural fea-
6.3 Sinkholes tures serve to control the distribution of recharge features
along with water chemistry. There are extensive cave sys-
Sinkholes have found many uses throughout history. The tems that support diverse ecosystems (Hovorka et al. 1996).
sinkholes (cenotes) of Yucatan were used as ritual places and
it is believed that they were used for human sacrifices. The
ancient Maya considered the sinkholes sacred entrances to 6.4.2 The Floridan Aquifer
the underworld (Vesilind 2003).
The Arecibo Observatory in northwest Puerto Rico is part The Floridan aquifer is the primary aquifer in the southeast-
of the National Astronomy and Ionosphere Center (NAIC). It ern United States and extends under all of Florida, and por-
is located in a large circular karst depression in the lime- tions of Alabama, Georgia, and South Carolina (Miller
References 47

1990). The Floridan aquifer in Florida yields over 950 mil- conditions. Hydrocarbon reserves are sometimes found
lion l/day to wells. An estimated 8 quadrillion liters of fresh within deep-seated paleokarst collapse zones, which consist
water are contained within the Floridan aquifer system. of brecciation. This brecciation sometimes has an increased
Porosity due to dissolution ranges from isolated vugs to porosity providing a trap for hydrocarbons, (Fritz et al.
caverns several meters across. The aquifer consists of a num- 1993). Large areas of the United States are affected by paleo-
ber of very high permeability zones, which generally con- karst (Palmer and Palmer 1989). James and Choquette (1988)
form to bedding planes, which are commonly either solution and Bosak et al. (1989) provide detailed discussions of
riddled or fractured. These karst conditions extend deep into paleokarst.
the Floridan aquifer system and were formed at lower sea
level stands (Miller 1990). In South Florida, the lower
Floridan aquifer at depths of 760–900 m is not considered References
part of the freshwater system. This zone contains saline
water and is used extensively for disposal of treated munici- Bosak P et al (eds) (1989) Paleokarst a systematic and regional review,
developments in earth surface processes 1. Elsevier, New York
pal wastes along Florida’s southeast coast via injection wells.
Ford D, Williams P (2007) Karst hydrogeology and geomorphology.
Wiley, New York
Fritz RD et al (1993) Paleokarst related hydrocarbon reservoirs. Soc
6.5 Mineral Resources Econ Paleontol Miner Core Work 18:1–275
Fryar AE (2009) Springs and the origin of bourbon. Ground Water
47(4):605–610 (July–August)
Resources have historically been mined from caves such as Hovorka SD et al (1996) Edwards aquifer ground-water resources:
flint to make tools, minerals for medicine, and paint pig- geologic controls on porosity development in platform carbonates,
ments. Bat guano from caves was also used as fertilizer in the south Texas. The University of Texas at Austin, Bureau of Economic
Geology, Report of Investigation 238
eighteenth and nineteenth centuries. During the Revolutionary
James NP, Choquette PW (eds) (1988) Paleokarst. Springer, New York
and Civil wars extensive mining of saltpeter (nitrates) from Milanovic T (2004) Water resources engineering in karst. CRC Press,
caves was used in the production of gun powder. Boca Raton
Bauxite (an aluminum ore) occurs as a clay-like material Miller JA (1990) USGS ground water atlas of the United Stated:
Floridan aquifer system, Alabama, Florida, Georgia and South
produced by intense weathering in warm humid climates and
Carolina. USGS HA 730-G
can be found in carbonates in numerous locations throughout National Astronomy and Ionosphere Center (NAIC) (2014) Arecibo
the world. The largest bauxite region is on the island of observatory, www.NAIC.edu. Accessed 12 May 2014
Jamaica where vast bauxite deposits are found within paleo- Palmer MV, Palmer AN (1989) Paleokarst of the United States. In:
Bosak P, Ford D, Glazek J, Horacek I (eds) Paleokarst, a systematic
sinkhole pits 10–30 m deep.
and regional review, developments in earth surface processes 1.
Mineral deposits are often found in zones of paleokarst Elsevier, New York, pp 337–363
and breccia. Breccia zones include sinkholes and collapse Roberts D (2004) Descent into the Maya underworld. Natl Geogr Mag
where angular fragments of carbonate rocks containing lead 206(5):36–53
Rother H, Rother C (2004) Lost caves of St Louis: a history of the cities
and zinc deposits are found in a few isolated areas of the
forgotten caves. Virginia Publishing Co, St. Louis, 119p
United States, Canada and Europe. The Tri-state district of Sangster DF (1988) Breccia-hosted lead zinc deposits in carbonate
Missouri, Oklahoma, and Arkansas, are one of the best rocks. In: James NP, Choquette PV (eds) Paleokarst. Springer,
known ore producers. Within the Tri-State District, the brec- New York, pp 102–116
Stamm D (1994) The springs of Florida. Pineapple Press, Sarasota
cias zones can be up to 30 m deep, 150 m wide and several
Steele B (2005) Spelean spotlight interview with Chris Nicola (discuss-
kilometers long (Sangster 1988). Some uranium deposits are ing the priest’s grotto). NSS News, Jan, pp 6–12
also found in breccia pipes in the United States. Tushingham, AD (1958) The men who hid the dead sea scrolls. Natl
Many of the oil and gas fields over the world are devel- Geogr Mag 114(6):785–808
Vesilind PJ (2003) Maya water world. Natl Geogr Mag 204(4):82–100
oped in the high permeability’s associated with paleokarst
 Karst and Its Damaging Impact
7

Abstract
Karst and pseudokarst can impact sites in a number of ways. The problems can range from
small nuisance type of subsidence or collapse that simply require routine maintenance to
catastrophic collapse which have major economic impact and in some cases loss of life. The
sudden collapse impacting structures (buildings, roads, etc.) are those that get everyone’s
attention including the attention of the media. Other impacts due to karst which are just as
important include flooding and contamination issues. The following chapter highlights a
variety of real-life impacts due to karst and psuedokarst related activity. These examples are
not meant to be inclusive, but are meant to show a variety of impacts in terms of type of
karst feature, size and type of impact.

7.1 Structural Impacts 7.1.1 Private Residences

Nuisance sinkholes are typically small (Fig. 3.5) don’t The impact of sinkholes on individual residential structures
necessarily have a large impact on structures or possess a is quite common in Florida. Insurers, in the state of Florida,
high risk for injury. They can often be “managed” by sim- have been required to offer coverage for damages resulting
ple back-filling measures. This type of sinkhole is often from sinkholes since 1981 (Florida Statutes 2014). Sinkhole
common in surface water retention ponds, parking lots, insurance claims have increased substantially in number and
roads, etc. cost over the past decades. The majority of the claims have
However, the size of a sinkhole does not always correlate occurred in central Florida in counties with obvious sinkhole
to its impact, but is typically related to the proximity to struc- activity. However, even areas not subject to dramatic sink-
tures, infrastructure, and culturally developed areas. For hole activity, like southeastern Florida, the number of claims
example, a large sinkhole in the middle of a grassy field may has also increased. Insurers have reported claims increased
have little to no impact. A similar size sinkhole or even a from 2,360 in 2006 to 6,694 in 2010 (Florida Senate 2011).
much smaller one within in a developed area may have cata- When there is an obvious sinkhole collapse present, such
strophic effects. Even a small sinkhole in the middle of a as a house about to disappear (Fig. 3.2a) or the Winter Park
high-speed highway or industrial waste pond could have a sinkhole (Fig. 3.7), there is little dispute regarding insurance
significant impact. coverage. However, if the evidence of a sinkhole is simply
The following section illustrates a number of different cracks in a walkway or the house walls (Fig. 7.1) without the
structural impacts that include houses, landfills, runways, obvious presence of a collapsed sinkhole there can be con-
bridges and dams. These examples are not meant to be inclu- siderable differences of opinion as to the cause. Many factors
sive but to show a range of serious impacts due to karst may cause small differential settlement of a structure such
activity. as: shrink/swelling clays, post construction settlement, ero-

© Springer Science+Business Media Dordrecht 2016 49

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_7
50 7 Karst and Its Damaging Impact

Fig. 7.1 In some cases the structure may be impacted by settlement resulting in cracks

sion, decomposition of organic materials, faulty construc- began to decay, local collapse features occurred along the
tion, and thermal expansion/contraction of construction runway. Their frequency of occurrence increased after a
materials. Even the removal of trees or large shrubs can lead major storm had gone through the islands. Upon landing, an
to swelling of clay which can go on for decades after their aircraft encountered one of these small sinkholes. Figure 7.3
removal (Clayton et al. 1995). There has been an expanding shows the damage to the propeller of the aircraft after
industry of consultants, remediation firms, and lawyers to encountering a small sinkhole.
deal with these issues. Many other states have also developed
regulations for sinkholes or mine related subsidence.
7.1.4 Multiple Collapse at a Housing
Development
7.1.2 Surface Water Management System
A new housing development in west central Florida
The management of surface water runoff at a landfill included (Hernando County) experienced the formation of 760
a concrete liner to guide waters off of and away from the sinkholes in less than 24 h over an area of approximately
landfill into a retention pond. Small leaks in the liner resulted 8 ha (Fig. 7.4). This may be the greatest number of sink-
in sinkholes developing beneath the concrete lining impact- holes to occur in a limited area within this short period of
ing and eventually breaching the concrete liner. Once time. The sinkholes ranged from 1 m (Fig. 7.5a) to more
breached, the surface water runoff was concentrated into than 30 m in diameter (Fig. 7.5b), with the majority of
these locations rapidly enlarging the sinkholes and extending them smaller in size. There was no loss of life or equip-
the damage (Fig. 7.2). ment in this case and the incident occurred prior to hous-
ing construction.

7.1.3 Sinkholes on an Airport Runway

7.1.5 Seepage and Collapse at Dams
Small sinkholes at a local airport runway on Andros Island in
the Bahamas became more than a nuisance. Small pits in the Many of the early well known problems with karst are asso-
surface of the rock are commonly referred to as “banana ciated with the failure of dams. There are records of historic
holes”. These small pits, typically on the order of 1 m in dam failures with some dating before 4000 B.C. Legget and
diameter or so, had been filled with organic debris and were Hatheway (1988) provide a summary of dam failures.
underlying the asphalt runway. Once the organic debris
7.1 Structural Impacts 51

Fig. 7.2 Seepage of water through a concrete lining used to transfer surface water away from a landfill resulted in washing away of sediment and
ultimately the collapse of the lining

Fig. 7.3 Damage to an airplane propeller that encountered a small sinkhole (banana hole) on the runway at Congo Town airport, Andros, Island
Bahamas

• The first masonry dam in the world in Egypt was constructed • The Austin dam constructed in 1893 started leaking
before 4000 B.C. It failed because it was built on limestone before it was finished and a flood carried away portions of
and much of the water flowed underneath the dam. it in 1900. Legget (1962) provides further details.
• The first big problem that occurred in the United States in • More than 54 sinkholes were documented in the reservoir
the early 1890s was a reservoir in New Mexico. The dam area behind Anchor Dam in Wyoming, one as big as 90 m
failed to hold water because it was built adjacent to a in diameter. While built, it holds only a fraction of the
rather large gypsum outcrop (Davies 1977). water that was intended for it (Davies 1977).
52 7 Karst and Its Damaging Impact

0 30
Approx. Scale in Meters

Fig. 7.4 A new housing development in west central Florida 760 sinkholes occurred within 24 h as a result of drilling and completing a well for
irrigation purposes

In addition, the Hales Bar Dam was constructed on 7.1.6 Elevated Expressway Failure
limestone on the Tennessee River. Thousands of cavities
(>8,000) were identified in the 2,000 borings and several A pile failure occurred during the construction of the new,
hundred thousand barrels of cement were injected into elevated Tampa Bay Crosstown Expressway on the west coast
these cavities and fissures. Leakage at the dam had become of Florida in April of 2004. A single massive column support-
serious a little more than 10 years after completion in ing a three-lane road over a 30-m span, sank into the ground
1926. After many attempts to stop the leakage, hot liquid about 5 m over a 10 s period along with the newly constructed
asphalt was used with success. Unfortunately, leakage roadway resting on top of the column (Fig. 7.6). The other 60
began again in 1941. TVA then became the new owner and piers that were in-place showed no signs of any problems.
decided on a final remediation for the dam. A continuous The columns extended down into a caisson that was
curtain of concrete was install across the face of the dam to founded on strong limestone about 18 m below the surface.
stop the leaks. When this failed in 1963 the TVA decided Prior to construction, the geotechnical investigation included
to give up and build the Nickajack Dam downstream. The drilling 22 m deep to test subsurface conditions (Stein et al.
Hales Bar Dam was demolished in 1968 (Legget and 2004). The fatal flaw in the limestone was obviously deeper
Hatheway 1988). than the test borings.
7.1 Structural Impacts 53

Fig. 7.5 The majority of the 760 sinkholes in Fig. 7.4 were 1–3 m in diameter (a) along with a few larger ones of 15–30 m in diameter (b)

Fig. 7.6 Damage occurred during construction of the Crosstown Expressway in Tampa, Florida when one pile slowly sunk about 6 m
54 7 Karst and Its Damaging Impact

Expressway officials and road builders said the collapse However, on occasion the amount of grout required will
has as much chance of happening again “as someone getting greatly exceed estimates. In these instances, the understand-
hit by a meteor”. Other comments in the press included; “A ing of subsurface conditions may not be clear. Over the years
sinkhole as deep as this is undetectable”, “It was just a working in South Florida, stories of excessive grout were
bizarre event”, “A small problem with the soil, something continually cropping up. A tall building at a local marina in
1–6 m across, is easy to miss. If you try to find everyone, you Ft Lauderdale, Florida encountered foundation problems
could never afford the project”, and “It was an act of God”. when borings encountered voids. Unusually large quantities
These comments showed a complete lack of understanding of grout were needed to stabilize conditions.
of karst conditions, (Waltham et al. 2005). The cost of this During expansion of the Macarthur Causeway bridge in
single catastrophic event was estimated at about $100 mil- Miami, an excessive amount of grout was needed to fill voids
lion, but the cost of additional investigation (deeper initial encountered (Berkowitz 1994, personal communication).
borings or employing other methods) prior to construction is The need for excessive grout was most likely associated with
negligible in comparison when dealing with possible karst high permeable zones developed at lower sea level stands.
conditions at such critical structures. These permeable zones are laterally discontinuous and are
found at depths of roughly 6, 12 and 24 m below mean sea
level associated with older sea level stands (Finkl and Esteves
7.1.7 Deep Unknown Paleokarst 1997; Esteves and Finkl 1999).

Buried paleokarst are usually not noticed and can be a

surprise in many cases. The Straights of Mackinac Bridge in 7.1.9 Reactivation of Sinkholes
Michigan was built on an area that had collapsed into the vast and the Drainage of Sinkhole Lakes
solution chambers of evaporates about 450 million years
ago. Subsequently new rock was deposited on top of it. Piers Most sinkhole lakes are stable because the opening or throat of
for the bridge should have been on solid rock, but the ancient the sinkhole has been filled with sediments. The stability of a
buried karst was encountered leading to problems in con- sinkhole lake is associated with the thickness and clay content
struction (Davies 1977). Black (2012) provides further of the sediments filling the throat of the sinkhole as well as the
details on this regional problem and its history. hydraulic head differences between the surficial and deeper
aquifers. On occasion however, an old sinkhole may re-activate
(or possibly a new sinkhole occurs within an existing sinkhole
7.1.8 Excessive Grout Quantities lake) draining much or all of the water from the lake.
An example from 1999 is that of Lake Jackson, north of
It is quite common to encounter a void, cavity or open frac- Tallahassee, Florida. Lake Jackson is a closed karst lake of
ture in test borings and then remediate with grouting. approximately 16 km2 that has no surface outflows (Fig. 7.7a).

Fig. 7.7 Lake Jackson, a sinkhole lake in north Florida (a), has periodically drained allowing geologists to investigate some of the sinkholes (b)
7.2 Groundwater Contamination 55

The lake had been a valuable recreational resource for aquifer commonly measured in meters per year. Because of
northern Florida and southern Georgia, well known for its the rapid flow, and minimal filtration, karst aquifers are par-
bass fishing and water sports. The triggering mechanism that ticularly vulnerable to contamination. Contaminants such as
reactivated the sinkholes is thought to be drought, resulting in nutrients, pesticides, heavy metals, acids, organic solvents,
increasing differences between the water table and the poten- petroleum products and fecal bacteria have been documented
tiometric surface (Scott 2013, personal communication). in karst aquifers. In addition, there is the potential for acciden-
In 1999, the lake completely drained. Once the outflow tal spills to occur along roads and railways that pass through
ceased, geologists and biologists found a rare opportunity to karst areas. Once contaminants have reached the groundwater
investigate the dry lake basin. State geologists entered the the conduits provide a pathway for contaminants to move
sinkhole to explore the system (Fig. 7.7b). State geologists quickly over long distances and contaminate large areas.
have estimated that there are hundreds of buried sinks in the In many engineering designs, surface runoff from roofs,
Lake Jackson Basin. Drilling on the dry lake bottom revealed paved parking lots and roadways are disposed of in nearby
a highly irregular, karstic surface. Historically, the lake sinkholes or within retention ponds. While clearly not the
drains into a sinkhole approximately every 25 years (Scott best of environmental practice, it has been a very common
2013 personal communication). practice in order to manage surface water. This storm water
runoff is a source of non-point pollution transporting solids,
heavy metals, nutrients, bacteria, road salt, herbicides, and
7.1.10 Problems with Man-Made Lakes hydrocarbons from highways. Stephenson and Beck (1995)
provide a detailed discussion of the problem along with
Man-made lakes are sometimes created to improve a new extensive references.
housing development and provide recreation. Cedar Lake The continued rapid development of real estate leads to
was a man-made lake of about 30 ha and up to 17 m deep increased runoff from built up areas. As the runoff from
located 72 km west of Oklahoma City. It was designed as a roofs, roads, parking lots and buildings exceeds 15 % of the
recreational lake for a privately owned residential commu- watershed area, water quality degradation begins to take
nity but one night it drained. Initial investigations indicated place (Veni et al. 2001).
that the lake had been saturating the underlying gypsum for Both agricultural and industrial sites underlain by epi-
about 20 years creating a crevice where the water leaked out karst zones can have significant environmental impacts.
traveling 0.4 km underground before flowing out of a canyon Agricultural application of pesticides, fertilizers, farm wastes
wall (Etter 1986). This was a major economic impact to the and their associated pathogens can easily impact the ground-
local community. water system. The waste from massive cattle, pig and chicken
At a site in North Carolina, numerous small diameter farms and their point recharge via closed depressions and
sinkholes (0.3–0.6 m in diameter) were developing at the thin patchy soil cover can have an even greater impact. Leaks
surface. These sinks were associated with voids and cavities and on-site waste disposal from industrial sites present simi-
within the upper portion of the limestone that was at a depth lar long-term contamination issues (Coxon 2011).
of about 12 m below grade. The primary triggering mecha- Because of the multidirectional nature of water move-
nism for these small sinkholes was the installation of a dam ment in the epikarst, a single point-source of contamination
that created a lake to enhance property values. The lake arti- can result in a pattern of contaminant distribution, which
ficially increased water levels changing hydraulic gradients may initially suggest that there are multiple sources of con-
and thereby initiating sinkhole activity. In addition to higher tamination (Aley 1997). Because of the extreme variability
water levels, which was the primary cause of sinkhole activa- within the epikarst zone and deeper conduit flow, contami-
tion, a secondary factor contributing to the rate of soil piping nant flow can be rapid and complex making the characteriza-
may have been the low frequency vibrations of the heavily tion, modeling and remediation much more difficult.
loaded freight trains passing directly over the area of sink-
hole development.
7.2.1 Mining Wastes

7.2 Groundwater Contamination Waste impoundments are commonly utilized to manage

waste streams from mining and industrial facilities. Many of
Karst aquifers are inherently susceptible to contamination. these impoundments are simple excavations into the existing
Rapid recharge via fractures or sinkholes provides a direct soils and are unlined. While these waste streams are com-
pathway for contaminants to enter the groundwater system. posed of a variety of materials, those that are acidic in nature
The flow of karst groundwater through conduits and fractures can be of great concern in karst areas. Release of acids is
is often measured in kilometers per day as compared to controlled by neutralization, however, this is often delayed
groundwater movement through a granular (diffuse-flow) or neglected. While small discharges of acids are quickly
56 7 Karst and Its Damaging Impact

neutralized by limestone bedrock, long-term leakage can

result in both geotechnical and environmental problems.
Sasowsky et al. (1995) provide a discussion of acid mine
drainage and its impact within carbonate geology. They
found that “acid waters can persist far into the basin, particu-
larly if contact with the limestone is minimized by clastic or
precipitated coatings on the rock”.

7.2.1.1 Jamaica Bauxite Wastes

Bauxite is found in abundance within the karst depressions of
the western part of Jamaica. The refining of bauxite to alumina
results in a waste known locally as “red mud”, which is dis-
posed of in unsealed mined out karst depressions in the lime-
stone. The “red mud”, which is more than 70 % water is highly
caustic and infiltrates into the groundwater. Groundwater con-
taminated by “red mud”, shows an increased sodium concen-
tration and increased pH. Two areas have been affected by
such contamination resulting in over 20 km2 of aquifer being
unsuited for groundwater development, (Fernandez 1991).

7.2.1.2 Phosphate Mining in Florida

Phosphate is a key ingredient in fertilizer and many other
products and has been mined in central Florida since the
Fig. 7.8 A large sinkhole developed in a gypstack, a waste material
1880s. Phosphate deposits are found about 7.5 m below the from the phosphate industry in Florida (Tihansky 1999)
surface and giant draglines are used to uncover and dig up
the matrix, which consists of phosphate, sand and clay.
Gypsum waste slurry (silt to sand size particles) results the surface to about 32 m diameter at a depth of 18 m and then
from the manufacturing process. This waste slurry is depos- extended vertically to a depth of 55 m where blocks of gypsum
ited in huge diked ponds for dewatering that are referred to were encountered. The sinkhole extended more than 122 m
as gypstacks. The suspended materials settle out and form a below the top of the stack. An estimated 113,000 m3 of phos-
weak porous rock-like mass. The remaining water at the sur- phogypsum and an undetermined amount of contaminated
face is acidic with a pH of 1.5–2.0 (Fuleihan et al. 1997). water disappeared down the sinkhole (Fuleihan et al. 1997).
The gypstacks range in size from less than 40 to more Hayward Baker carried out remediation by installing angle
than 280 ha and range in height up to 60 m above grade. borings and grouting with the focus upon sealing of the con-
More than 20 gypstacks are located within 40,400 ha in cen- fining layer (Fig. 7.9). Over 2,890 m3 of concrete grout were
tral Florida. All new gypstacks are lined at their base and injected through 50 injection points to seal the base of the
have water-circulation systems to prevent the escape of waste sinkhole within the confining layer. The remaining sinkhole
slurry. With exception of six new stacks, all of the others are was then filled with gypsum. Fuleihan et al. (1997) provides
unlined (Florida Institute of Phosphate Research Library a complete discussion of the collapse event, the groundwater
1999 personal communication). contamination and remediation.
Gypstacks have had sinkhole failures (Sowers 1996), the Sinkholes are not uncommon in this area of Florida. The
most recent formed in June of 1994. A major collapse occurred original conceptual model showed a narrow throat going into
at a 168 ha, 60 m high gypstack at the New Wales plant of IMC- limestone. However, there was almost certainly a pre-existing
Agrico Company. During a routine morning visual inspection, subsurface karst feature (i.e., an enlarged well intercon-
an elongated erosion feature was seen in the surface of the gyp- nected fracture or cave system) that enabled the 113,000 m3
stack. That afternoon a sinkhole developed which was approxi- of materials to rapidly disappear into the underlying lime-
mately 48 m in diameter and tapered to 32 m wide at the base of stone (Fig. 7.9). The combination of the additional increased
the gypstack (Fig. 7.8), (Fuleihan et al. 1997). weight of the 60 m high gypsum stack, along with possible
Six borings including four angle borings along with water level fluctuations and some of the dilute acid reaching
piezometers were used to determine the dimensions of the col- the underlying limestone could all have contributed to the
lapse and assess hydrologic conditions. Figure 7.9 shows a failure below the gypstack. Sowers (1996) described a previ-
conceptual cross section of the conditions at the failed gyp- ous gypstack failure resulting in a collapse feature within the
stack. The nearly vertical shaft tapered from 48 m diameter at limestone of 30 m in diameter.
7.2 Groundwater Contamination 57

48 m

32 m

Gypsum 65 m
Focus of
Blocks of Remediation
Gypsum by grouting

Sand 12 m

Confining Layer 40 m
Original
Conceptual
Model
Intermediate Aquifer 21m
Clay Unit 4.5 m
213m
Deeper Aquifer
Throat connected
Extension to to deeper
Conceptual Model cavernous zones
needed to
acommodate
large amounts
of lost material

Fig. 7.9 A conceptual cross section of the gypsum stack collapse

7.2.2 Regional Contamination: The Woodville the groundwater to contamination. Three possible sources of
Karst Plain (WKP) contamination to Wakulla Springs have been recognized: a
waste water disposal spray field located south of Tallahassee,
The Woodville Karst Plain (WKP), in Northern Florida, cov- Florida; storm water runoff that rapidly recharges the
ers approximately 1,200 km2. Wakulla Spring is the largest Floridan aquifer through sinkholes; and discharge from pri-
inland spring within the WKP and is the third largest spring vate septic tanks.
in Florida. Unfortunately, the water quality has been The Wakulla Springs Woodville Karst Plain Symposium
declining within Wakulla Spring. Nitrates that come from was held in October of 1998 (Schmidt et al. 2000). The
human waste and fertilizer have been blamed for the degra- Hydro Geo Consortium Group at the Florida Geologic
dation of the spring waters that includes an increase in Survey and the Florida State University initiated a regional
hydrilla and algae resulting in a 30 % decrease in water clar- assessment of the sinkholes, springs and conduit flow.
ity. In the past 30 years the spring has seen a fivefold increase Extensive studies have been completed to characterize the
in nitrogen levels (Chelette et al. 2002). Woodville Karst Plain. Cave divers have made tremendous
Recharge to the Floridan aquifer in the Woodville Karst contributions to the descriptions and connections of the
Plain occurs by sinking streams, direct infiltration of precipi- underwater network of caves within the WKP. The results are
tation through sinkholes, infiltration through the variable described by Wisenbaker and McKinlay (2006) and
thick sands and soils overlaying the aquifer, and groundwa- Wisenbaker et al. (2007). Microgravity surveys have been
ter flow into the WKP from the north. An estimated 1,000 carried out to detect the presence of possible karst conduits
sinkholes are located within the WKP. Several of these sink- not mapped by cave divers between the spray field near
holes are known to receive water from disappearing surface Tallahassee and Wakulla Springs a distance of about 16 km
streams that drain upland regions (Gerami 1984; Davis (Yuhr et al. 2008). Quantitative dye tracing has also been
1996). This pervasive recharge increases the vulnerability of performed in the WKP in order to gain defensible estimates
58 7 Karst and Its Damaging Impact

of the contributions of specific pathways to the discharge at In areas with poorly consolidated sediments and little or
Wakulla Spring and to measure hydraulic parameters for use no vegetative cover, extremely rapid weathering and mass
in numeric modeling (Kincaid et al. 2012). transport may occur under flash flood conditions. Erosional
processes can lead to the development of pipe-like openings
within the unconsolidated material such as silt, clay and
7.3 Pseudokarst Impacts (Natural loess along with certain shales, claystones and volcanic ash
and Man-Made) and tuff. Soil piping is a major factor in the erosional pro-
cess. Although the forms resemble karst landforms they are
Besides subsidence and collapse associated with the dissolu- generally not large or expansive in area. These pseudokarst
tion of soluble rock, there are a number of conditions which features may develop and change rapidly as a result of a sin-
resemble or behave like karst but are not caused by natural gle heavy rainstorm.
dissolution of the rock. These karst-like features (pseudo- Loess usually contains a carbonate cement and frequently
karst) may occur due to natural processes or man-induced gives rise to small sinkholes and caves (Sweeting 1973). In
processes and may impact a site as dramatically as natural loess, the erosion process may develop to the level where
karst conditions. The pseudokarst that is man-induced tends sinks, pipes, and caverns develop looking like classic karst
to be more pervasive in developed areas. and are known in China as “loess karst” (Fookes et al. 2005).
The natural processes creating pseudokarst can occur in soil These erosion features present problems to engineering
or rock. Naturally occurring processes in soils include fissures works when they lie undetected just below the surface.
and erosional processes. Karst-like conditions can also occur
within insoluble rocks such as lava, sandstone, or even granite. 7.3.1.2 Swelling (or Expansive)
Man-induced processes creating pseudokarst can be and Shrinking Soils
divided into small and large sizes. The small pseudokarst can Some clay soils undergo slow volume changes with a change
range from very small, localized problems with dimensions in moisture content. A decrease in moisture causes shrinkage
of cm to tens of meters. The smaller scale pseudokarst and fissuring while an increase in moisture causes swelling
includes dissolution due to acid spills or leaks, broken water and heave. While swelling soils do not present the ultimate
or sewer lines, boring or tunneling and voids behind concrete hazard like a large sinkhole, they can cause considerable dif-
such as tunnel liners. The large pseudokarst can be quite dra- ferential settlement, cracking and damage to structures simi-
matic and range from tens of meters to regional in nature lar to the impact from early phase of natural karst subsidence
over several square km. This type of pseudokarst includes or collapse.
subsidence or collapse associated with petroleum activities Factors causing changes in moisture content include:
and underground mining as well as regional subsidence due
to withdrawal of groundwater, oil and gas. • Changes in precipitation
• Wet and dry seasons
• Poor drainage
7.3.1 Naturally Occurring Pseudokarst • Leaking underground pipes and
• Tree and vegetation cover and its removal
7.3.1.1 Earth Fissures and Erosional Features
The term “earth fissure” is defined as a crack at or near the Transpiration from vegetation cover is a major cause of
ground surface, generally occurring in unconsolidated earth water loss from soils in semi-arid regions. The removal of
material. These fissures develop below the ground and trees or large shrubs from a site removes their desiccating
migrate towards the surface. Earth fissures have been widely effect, therefore a subsequent increase in moisture results in
observed in a number of the western United States. They go swelling (Bell and Culshaw 1998). The effects from removal
undetected before they break through to the surface and can of vegetation can go on for decades after their removal
cause problems with foundations (Shlemon 2004). (Clayton et al. 1995).
Fissures are relatively long, curvilinear pattern of discon- The depth of the active zone in which swelling and shrink-
tinuous, relatively narrow, open steep-sided cracks in earth age occurs varies by season and may extend over 6 m deep in
materials. Holzer (1984) suggests that fissures due to subsid- some arid regions. In temperate regions such as Britain with
ence are generally long and may be on the order of 1,000 m, a damp climate, changes are typically restricted to the upper
while desiccation fissures or giant polygons are on the order 1–1.5 m in clay soils (Bell 2004).
of 10 m. They are typically 1– 2 m wide and 2–3 m deep. Swelling soils are present in much of the United States as
Major mechanisms to explain earth fissuring are groundwater well as the rest of the world. They impact structures resulting
depletion and erosion associated with the rapid flow of sur- in cracking much like slow subsidence and early sinkhole
face water. activity. When cracks to structures occur, they can often be
7.3 Pseudokarst Impacts (Natural and Man-Made) 59

interpreted as a sinkhole problem. This is common in the ging data. Extensive remediation grouting was required to
west central area of Florida where expansive soils are pres- stabilize the foundation.
ent. Often an insurance claim of a sinkhole without the pres-
ence of obvious collapse can lead to disputes as to the actual
cause. Is the problem due to a sinkhole, swelling soils or 7.3.2.2 Sinkholes Caused by Broken Water
poor foundation conditions due to fill and debris? Bell and and Sewer Lines
Culshaw (1998), Noe and Dodson (1999), Bell (2004), and Broken water and sewer lines are a common cause of much
Hunt (2005) provide further discussion on the subject of pseudokarst collapse in developed areas and are commonly
swelling soils. reported in the news as “sinkholes”. Voids are created by soils
washing away due to the broken water or sewer lines. These
voids ultimately collapse and look just like traditional sink-
7.3.2 Smaller Man-Induced Pseudokarst holes. While most problems are relatively small and localized
there are occasional examples of catastrophic collapse.
7.3.2.1 Dissolution by Release of Acids In 2005, residents in Guatemala City began to complain
Significantly accelerated dissolution of limestone can occur about rumbling and shaking in their homes. Because the city
when acids are introduced whether natural or man induced. is located in a seismic zone and has volcanic activity
Here the focus is on the release of acids, which can occur complaints were ignored until February 2007 when a large
under a variety of conditions such as during transportation, sinkhole collapse occurred. The sinkhole was 30 m diameter,
from waste impoundments or through pipelines at industrial and 60 m deep, five homes collapsed and three people died.
facilities. Whether produced by the oxidation of sulfides from Then in May of 2010 a second sinkhole collapsed (18 m in
mine wastes or rocks containing sulfides, the presence of sul- diameter and 36 m deep) including a three-story building and
furous acids will dissolve limestone rapidly (Sowers 1996). the disappearance of at least three people. These were both
The release of acids is controlled by neutralization. If large, straight-sided collapse features due to a deep sewer
addressed in a timely manner, the impact can be minimized. system (approximately 40–50 m below ground) with a diam-
In some cases, leaks can go undetected leading to the devel- eter of 2.25 m (Hermosilla 2012).
opment of large cavities, subsidence or collapse as well as In Atlanta, a parking lot at a hotel developed a large sink-
contamination problems. It is this long-term leakage that can hole due to a leaky sewer line. Problems with the sewer line
have the greatest impact, creating both geotechnical and were known and a reinforced fabric liner was placed under
environmental problems. the parking lot designed to support a 6 m diameter sinkhole
It is not unusual for manufacturing facilities to develop allowing time for evacuation. A number of smaller sinkholes
leaks in their processing systems. In some cases, acids leak- developed which the liner supported until they all coalesced.
ing from pipelines at commercial facilities have dramatically The large sinkhole, 45 × 60 m, caused two fatalities. A hotel
increased dissolution of limestone under the facility leading employee was found crushed in a car at the bottom of the
to the formation of large cavities and sinkholes. The follow- sinkhole and another employee was found roughly 3 km
ing examples illustrate some of the problems encountered: away in a combined sanitary and storm water line (Melvin
1993).
• Cavities had developed within the limestone underlying a
power plant chemical laboratory due to an acid leak. 7.3.2.3 Settlement and Collapse
Some cavities were up to 2 m in diameter with channels Due to Horizontal Borings and Tunneling
extending latterly as much as 3–6 m. A remedial grouting Horizontal borings are commonly used for the installation of
program was carried out utilizing acid resistant cement to utilities and pipelines. Some of these borings have resulted in
stabilize the structure. significant subsidence and sizeable sinkholes, which can
• Electroplating facilities utilize large vats of acids. A plant occur during installation or after completion.
had been in operation for a number of years with acid A subway tunnel was being dug below existing city streets
leakage occurring. This created significant voids under in Los Angeles. The construction induced a large collapse,
the concrete floor that resulted in foundation problems for about 24 m wide, in the overlaying soils and breaking both
the large tanks and equipment within the building. water and gas lines. The collapse was stabilized with a low
• One large manufacturing plant had unknowingly lost a strength cement and construction proceeded. During subse-
large amount of acids (unknown volume) over the years. quent modifications to the liner a second collapse occurred at
This resulted in dissolution of the limestone and the the same location (Gordon and Kennedy 1995).
development of a large cavity under the building. The cav- Unsuspected vertical settlement of up to 12 cm was
ity was estimated to be more than 600 m3 based upon observed in fractured granitic gneiss several hundred meters
microgravity measurements, borings and geophysical log- over the Gotthard highway tunnel in central Switzerland.
60 7 Karst and Its Damaging Impact

Settlement was probably related to consolidation of fractures table has declined more than 30 m since the early 1950s
resulting from fluid drainage into the tunnel and pore pres- (Londquist 1995).
sure changes after tunnel construction (Zangerl et al. 2003). A large portion of Rogers Lakebed is utilized as earthen
runways at Edwards Air Force Base (EAFB). One of the run-
7.3.2.4 Voids Behind Concrete ways is 11 km long. By 1990 groundwater levels at EAFB
Small voids of a few centimeters to a meter or so are com- were more than 21 m below the lakebed surface. Near the
monly encountered behind many concrete structures. These southern edge of Rogers Lakebed, the land had subsided
voids behind concrete are typically small but can have a sig- more than 0.6 m between 1961 and 1989 due to groundwater
nificant impact on structures such as highways, runways, sea withdrawal. This contributed to the formation of surface fis-
walls and concrete tunnel linings. For example, hydraulic sures on the runways (Kratochvil 1989).
pumping of concrete slabs on highways and airport runways The fissures would develop below the surface hidden
can lead to a thin void space under the concrete slab leaving from view. They would eventually break through the surface
one side of the slab lower than the other. This causes bumps with polygonal patterns (Fig. 7.10a). They would then grad-
in our highways and backbreaking shocks to pilots on land- ually become larger and more obvious (Fig. 7.10b) eventu-
ing when the nose wheel crosses the variations of up to ally reaching dimensions of about a meter or so deep and a
2.5 cm in the top of the concrete slabs. The 727 stretch air- meter wide (Fig. 7.10c). Once obvious they would be
craft are known as concrete breakers of runways. excavated and repaired. Undetected, they could be a signifi-
Delaminating of concrete in bridge decks is also a problem. cant hazard to fighter aircraft, bombers and shuttle landings.
A small industry has developed to carry out non-destructive On three occasions the shuttle’s wheels broke through the
testing to evaluate the presence of such voids. desert floor, fortunately without damage (Kratochvil et al.
1992).

7.3.3 Larger Man-Induced Pseudokarst 7.3.3.2 Pseudokarst Associated with Petroleum

Industry Activity
7.3.3.1 Regional Subsidence Due to Withdrawal Evaporites are present in 32 of the 48 contiguous states and
of Groundwater underlie approximately 35–40 % of the land area. Natural
Pumping of groundwater has lead to significant regional sub- evaporite karst is present in almost all areas in which evapo-
sidence in more than 14 areas of the United States (Holzer rates are found. The most wide spread and pronounced
1984). Land subsidence of one type or another has been examples of problems in both gypsum and salt are in the
reported in at least 37 of the 50 states. This type of subsid- Permian basin of the southwestern United States (Johnson
ence is expected to increase as the demands for water supply 2003).
and natural resources increase all over the world (Johnson Subsidence and collapse features resulting from petro-
1998a). leum activity have been well documented and are considered
Subsidence was first recognized in the Houston-Galveston man-made pseudokarst. Examples include the Wink Sinks in
area in 1918 as a result of oil field withdrawals and then sub- Texas, brine well sites in or near Carlsbad New Mexico and
sequently pumping of groundwater covering 9,420 km2 subsidence over the Ekofisk Oil Field in Norway.
(Holzschuh 1991). A classic case of subsidence up to 9 m
due to groundwater pumping has occurred in the San Joaquin Wink Sinks
valley of southern California encompassing 25,800 km2 The Wink Sinks are two very large sinkholes in west Texas
(Johnson 1998a). This type of subsidence is regional in that occurred in 1980 and 2002 in the giant Hendrick oil field
nature but can have local impacts such as fissures and ero- near the town of Wink, Texas. Wink Sink 1 occurred in 1980
sion or impact to infrastructure. and was 110 m in diameter and 34 m deep with an estimated
For example, Rogers Dry Lakebed in the Mojave Desert volume of 159,000 m3. Wink Sink 2 formed in 2002 around
north of Los Angeles consists, in part, of pluvial clay depos- a water supply well. It was approximately 238 × 186 m wide
ited during a period when the area was a permanent lake. A and 61 m deep with an estimated volume of 1,330,000 m3. In
thin layer of silt covers the clay. Below the pluvial clay lies a addition, broad subsidence features occurred with sags of
layer of gravel containing an extensive aquifer. Ward et al. 7–8.5 m.
(1995) describe the geologic setting. The origin is believed to be from naturally occurring cavi-
The groundwater pumping from the regional valley has ties within a salt/anhydrite (gypsum) bed about 400–655 m
exceeded the estimated annual recharge almost every year below the surface. Unconsolidated clastic materials overlie
since the early 1920s. After World War II, pumping expanded the soluble deposits. Natural dissolution is known to exist
driven by the growth in California. Originally groundwater from deep fresh water upwelling through fractures and dis-
was under artesian flow conditions. In some areas, the water solving of the overlaying salts.
7.3 Pseudokarst Impacts (Natural and Man-Made) 61

Fig. 7.10 Fissures at the Edwards Air Force Base in California stage of development displaying open fissures (b), and a fully developed
extended over many of the earthen runways. A typical fissure early in fissure with subsidence (c)
its development with polygonal patterns at the surface (a), a second

It is thought that oil production over the area subsequently

aided in further dissolution and ultimately to an unstable
state. The area has a large oil field including one well within
the original sink. Freshwater from surficial aquifers moved
down through fractures created by the borehole. Large evap-
oration ponds added to the downward flow of fresher water
that also greatly enhanced dissolution (Johnson 1987;
Johnson et al. 2003).

Brine Wells in New Mexico

Numerous brine wells are located in New Mexico. These
wells are drilled into massive salt beds. Fresh water is
pumped into the well to dissolve the salt. The salt water is
recovered from a nearby well and is used as drilling fluid.
The process can develop large cavernous zones within the
salt bed. Without proper engineering control the cavern can
become large enough so that it collapses resulting in large
sinkholes. Fig. 7.11 A large sinkhole north of Carlsbad New Mexico was a result
A site 35 km northeast of Carlsbad, New Mexico was of dissolution mining of salt brine to be used as drilling fluids in drilling
for oil (Land 2013)
pumping water from a brine well when a worker on-site
noted a rumbling noise and quickly left the site. Minutes
later a large sinkhole abruptly formed, engulfing the brine (Fig. 7.11). An estimated 432,400 m3 of sediment was lost. A
well and structures. The initial sinkhole was several tens of few months later another sinkhole formed about 17 km to the
meters in diameter and filled with water to within 12 m below northeast of the initial site (Land 2013).
the surface. There were large concentric fractures around its As a precaution, another brine well within the city of
perimeter. The sinkhole grew to about 111 m in diameter and Carlsbad has been shut down and is being monitored. This
45 m deep and became dry as the water drained from it site is at the intersection of three state highways with
62 7 Karst and Its Damaging Impact

adjacent mobile homes, a drainage canal and a church. lar coal mine can result in classic looking sinkholes when the
Signs along the roads warn of the potential danger roof span between pillars fails and when the overlaying rock
(Fig. 7.12). Surface geophysical surveys were made to is not thick or strong enough to prevent the collapse from
determine the lateral extent of the cavern and the site is reaching the surface. Subsidence has occurred as early as a
being monitored by inclinometers installed within the decade after mining and as late as a century. More than half
boreholes with their outputs sent to critical emergency the subsidence incidents took place 50 or more years after
staff phones (Land 2013). mining (Gray and Bruhn 1984; Dunrud 1984).
Long-wall mining inherently leads to subsidence as the
Subsidence Over the Ekofisk Oil Field entire coal seam is removed as mining advances. Immediately
The Ekofisk Oil Field located in the North Sea off of Norway after the coal is removed the mine-roof begins to collapse
withdraws oil from a chalk reservoir. Subsidence due to and the collapse begins to propagate slowly to the surface.
withdrawal of oil was first recognized in 1984 and was 2.6 m About 90 % of the subsidence occurs in the first 3 months
maximum. The subsidence depression was 6 km long and after extraction, while residual subsidence continues for up
4 km wide. Subsidence at sea impacts critical drilling rig to 3 years at the deeper sites (Gray and Bruhn 1984; Dunrud
platforms, bottom mounted storage tanks and piping, 1984).
(Whittaker and Reddish 1989). In one area of southern Italy, limestone was mined under-
ground to maximize land space for farming. Urban expan-
7.3.3.3 Pseudokarst Associated sion over time resulted in buildings over the mines. Over the
with Underground Mines decades memory of the existence and lateral extent of the
Pseudokarst conditions have developed as a result of subsid- mines was lost. In addition weathering processes and illegal
ence and collapse over underground mining operations and discharge of waste caused a decrease in strength of the
mine shafts. There has been subsidence and collapse over limestone resulting in failures. Vattano et al. (2013) describe
abandoned limestone, gypsum and coal mines in the United the history and investigation of these abandoned mines.
States. Underground coal mining has occurred beneath 3.2
million hectares of land in the US, almost 1 million hectares Underground Coal Mine Fires
of which have been affected by subsidence. Most of the coal Fires in coal mines are reported from nearly all parts of the
mining has taken place in the eastern half of the US where world were coal is being mined. Coal mine fires have lead to
thousands of hectares in urban areas are threatened by sub- subsidence as a result of decreasing the volume of coal
sidence, (Gray and Bruhn 1984). through burning. Local communities have had to be moved
Room and pillar mining in the US averages 40–60 % because of environmental issues. Stracher (2007) provides a
extraction but can be as great as 80–90 %. The room and pil- review of coal fires from around the world.

Fig. 7.12 Another site within the city of Carlsbad, New Mexico where salt was dissolved for drilling purposes has created a large cavity and a
road sign alerts drivers to potential of sinkhole occurring in this area
7.3 Pseudokarst Impacts (Natural and Man-Made) 63

Gold Mining in South Africa inflow of fresh water. The mine filled in approximately
The richest gold mines in the world are located in the “Far 21 months with an estimated flow of 113,000,000 l/day
West Rand” of South Africa about 64 km west of (Nieto and Young 1998).
Johannesburg. Gold is mined from the conglomerates, which Very little geotechnical drilling data were available prior
underlie the thick (1,000 m or more) dolomites. The mining to 1995. However, the literature contained numerous refer-
takes place between thick dykes that divide the dolomite into ences describing stress conditions and up-warps beneath
isolated compartments. By 1960, the mines were deep sedimentary valleys. Several deep, water wells located
enough that it was necessary to start extensive pumping and 1–5 km from the mine had water level declines of 15 m or
the water table was lowered through the dolomites to provide more. Initial indicators of impacts on private wells were dis-
access to the gold deposits. The impacts were first seen when regarded until a widespread pattern of surface subsidence
springs began to dry up and then sinkholes began to occur as and groundwater changes became obvious. Published reports
the water level was lowered (Brink 1984; De Bruyn and Bell from the nineteenth century reveal that natural brine and gas
2001). pools existed in the region long before mining began (Gowan
The most catastrophic sinkhole event occurred in and Trader 2000). The unfavorable geologic and hydrologic
December of 1962, when a three-story crusher plant along conditions within the overburden along with the existing
with 29 occupants was swallowed by a sinkhole. More than stress regime near the collapse could have been anticipated
200 other sinkholes have destroyed buildings and taken (Nieto and Young 1998).
additional lives. These were all very rapid collapses, each
occurring in a period of minutes and not allowing time for Weeks Island
people to escape. Other sinkholes have occurred in the area Sinkholes have occurred over the U. S. Strategic Petroleum
over the past 25 years without the loss of life (Brink 1984; Reserves storage facility at Weeks Island, Louisiana. The
De Bruyn and Bell 2001). Weeks Island salt dome (a room and pillar salt mine) was one
of six original storage sites selected for the US Strategic
7.3.3.4 Pseudokarst Associated Petroleum Reserve, which was established after the middle-
with Underground Salt Mines east oil embargo of 1973–1974. The salt mine was originally
In the US, salt deposits underlie some part of 25 of the 48 opened in 1902 and salt was extracted until 1977 when the
contiguous states. Some of the salt deposits are extensive and mine was converted to oil storage. Petroleum reserves are
rank among the greatest salt deposits of the world (Johnson stored in salt mines more than 152 m below grade (Neal
2003). Salt was first obtained by surface mining and evapora- 1997).
tion ponds, then by underground mining. There are two meth- In 1992 a sinkhole formed over the outer edge of the for-
ods for underground mining of salt, the room and pillar mining mer salt mine. A second smaller sinkhole formed early in
where the salt is physically extracted and lifted to the surface 1995. These sinkholes were associated with a combination
and by solution where the salt is dissolved and pumped to the of factors including anomalous geologic conditions, salt
surface as a brine. Collapse of room and pillar mines are rare creep (the gradual movement of salt), and stress cracks prop-
unless they become flooded. Examples of dry salt mines that agated upward leading to collapse. The sinkholes were stabi-
have developed sinkholes are the Retsof Mine (New York), lized and oil reserves were finally moved to another location
Weeks Island (Louisiana) and Jefferson Island Diamond Salt and the mine was filled with 85 % saturated brine (Neal
Mine collapse (Louisiana). These mines are now flooded and 1997).
are closed (Johnson 2003; Neal 1997).
Jefferson Island Diamond Salt Mine Collapse
Retsof Mine of New York A room and pillar salt mine was in operation 400 m under the
The Retsof Mine is owned by Akzo Nobel Salt Inc. and is the freshwater Lake Peigneur in Louisiana. Oil exploration was
largest salt mine in the western hemisphere. It is located taking place within the shallow lake in 1980. A 120 m error
south of Rochester in New York. The mine was in operation in the location of an oil exploration borehole resulted in the
from 1882 to 1995. The large (26 km2) room and pillar mine borehole hitting the salt mine at the 400 m level. The result
with a mined thickness of 3.5 m is located about 335 m was a huge whirlpool that drained the entire lake along with
below grade. A complex system of geology and aquifers additional water flowing through a canal from the Gulf of
more than 305 m thick overlies the mine which includes Mexico. This water formed a 45 m high waterfall. The five
approximately 182 m of shale and limestone with a cover of million dollar drill rig vanished into the giant “sinkhole”
glacial sediments. In 1994, the mine-roof failed in the east- along with barges and a tugboat. Adjacent property of 26 ha
ern portion of the mine that was located beneath a buried with 150 year old trees were washed away in a landslide. A
glacial valley with a fresh water aquifer. The failure occurred local fisherman saw the calamity and headed to shore. He
with the inrush of brine and gas, followed by a sustained had just tied his boat to a tree and walked away when he
64 7 Karst and Its Damaging Impact

turned around his boat and the tree were gone. Compressed Coxon C (2011) Agriculture and karst. In: van Beynen PE (ed) Karst
air from the mine resulted in a huge geyser of water from the management. Springer, Dordrecht, pp 103–138
Davies WE (1977) Keynote presentation to the symposium on detection
mine shaft. Although there were more than 50 workers in the of subsurface cavities. US Army Engineer Waterways Experiment
mine along with the drilling crew on the lake, there was no Station Vicksburg, Mississippi. (lecture recorded by Benson RC)
loss of life (Johnson 2003). A commentary and video can be Davis H (1996) Hydrogeologic investigation and simulation of ground-
seen on the Modern Marvels Engineering Disasters 5 (A&E water flow in the upper Floridan aquifer of north-central Florida and
southwestern Georgia and delineation of contributing areas for
Television Networks 2000/2001). selected city of Tallahassee, Florida. Water Supply Wells: USGS
In this case the rapid flow of large quantities of water and Water-Resources Investigation Report 95-4296. Denver, Colorado
sediment quickly created a large diameter hole from a small DeBruyn IA, Bell FG (2001) The occurrence of sinkholes and subsid-
35 cm borehole, large enough to allow the drill rig, two ence depressions in the far west Rand and Gauteng Province, South
Africa, and their engineering implications. Environ Eng Geosci
barges and a tugboat to be lost down the hole. 7(3):281–295
Dunrud CR (1984) Coal mine subsidence – western United States. Geol
Solution Mining of Salt Soc Am Rev Eng Geol 6:151–194
Cavities developed during solution mining of salt are typi- Ege JR (1984) Mechanisms of surface society of subsidence resulting
from solution extraction of salt. In: Holzer TL (ed) Man induced
cally 10–100 m in diameter and are 10–600 m high. land subsidence. Reviews in engineering geology, vol 6. The
Unfortunately the cavity sometimes becomes too large and Geologic America, Boulder, pp 203-221
the roof collapses. Most solution mining collapses have been Esteves LS, Finkl CW (1999) Late Cenozoic depositional paleoenvi-
a result from cavities formed 50–100 years ago before ronments of southeast Florida interpreted from core logs. Rev Bras
Geosci 29(2):129–134
modern-day engineering safe guards were developed. Proper Etter J (1986) Geological mishap drains lake. NewsOK (Oklahoma), 16
design has virtually eliminated this problem in new facilities. Nov 1986
Examples of subsidence resulting from solution mining are Fernandez BF (1991) Caustic waste contamination of karstic limestone
provided by Johnson (1998b) and include the Cargill Sink in aquifers in two areas of Jamaica. In: Proceedings of the 3rd confer-
ence on hydrogeology, ecology, monitoring, and management of
Kansas, along with the Grand Saline Sinks in Texas. Ege ground water in karst terranes. US EPA and Ground Water Scientists
(1984) provides examples of the Meade Salt Well and the and Engineers a division of NGWA, pp 389–402
Hutchinson sink both in Kansas along with the two sinkholes Finkl CW, Esteves LS (1997) A preliminary analysis of neogene depos-
on Grosse Ile, an island in the Detroit River, Michigan. ital environments in Broward County, Florida. Fla Sci
60(2):94–103
Florida Senate (2011) Issues relating to sinkhole insurance. Interim
Report 2011-104 Committee on Banking and Insurance Dec 2010
References Florida Statutes (2014) Title XXXVII, Insurance, Chapter 627, insur-
ance rates and contracts, 627.706 Sinkhole insurance; catastrophic
Aley T (1997) Groundwater tracing in the epikarst. In: Beck B, ground cover collapse. www.leg.state.fl.us/statutes. Accessed 12
Stephenson JB (eds) Proceedings of the 6th multidisciplinary con- Oct 2014
ference on sinkholes and the engineering and environmental impacts Fookes PG et al (eds) (2005) Geomorphology for engineers. Whittles
of karst, The Engineering Geology and Hydrogeology of Karst Publishing, Dunbeath, Scotland
Terranes. Balkema, Springfield/Rotterdam, pp 207–211 Fuleihan NF et al (1997) The hole story: how a sinkhole in a phospho-
Arts & Entertainment (2000/2001) Modern marvels engineering disas- gypsum pile was explored and remediated. In: Beck BF, Stephenson
ters 5, The History Channel A&E Television Networks DVD, Cat. JB (eds) Proceedings of the 6th multidisciplinary conference on
No. AAE-73136 sinkholes and the engineering and environmental impacts of karst,
Bell FG (2004) Engineering geology and construction. Spon Press, the engineering geology and hydrology of karst terranes. Balkema,
London Rotterdam, pp 363–369
Bell FG, Culshaw MG (1998) Some geohazards caused by soil mineral- Gerami A (1984) Hydrogeology of the St. Marks river basin, northwest
ogy, chemistry and microfabric: a review. In: Maud JG, Eddleston Florida. M.S. thesis, Florida State University, Tallahassee, FL
M (eds) Geohazards in engineering geology. Engineering geology Gordon L, Kennedy JM (1995) Street vanishes in subway sinkhole. Los
special publication, vol 15. Geological Society, London, Angeles Times, 23 Jun 1995
pp 427–441 Gowan SW, Trader SM (2000) Mine failure associated with a pressur-
Black TJ (2012) Deep karst system research, Michigan, USA. Carbonate ized brine horizon: Retsof salt mine, western New York. Environ
Evaporite 27(2):119–122 Eng Geosci 6(1):57–70
Brink ABA (1984) A brief review of the South African sinkhole prob- Gray RE, Bruhn RW (1984) Coal mine subsidence–eastern United
lem. In: Beck BF (ed) Proceedings of the 1st multidisciplinary con- States. Geol Soc Am Rev Eng Geol 6:123–149
ference on sinkholes: their geology, engineering and environmental Hermosilla RG (2012) The Guatemala city sinkhole collapses.
impact, Orlando, Florida. Balkema, Rotterdam, pp 123–127 Carbonate Evaporite 27(2):103–107 (Special issue on integrating
Chelette A et al (2002) Nitrate loading as an indicator of nonpoint science and engineering to solve karst problems)
source pollution in the lower Saint Marks-Wakulla rivers watershed. Holzer TL (1984) Ground failure induced by ground water withdrawal
Northwest Florida Water Management District, Water Resources from unconsolidated sediment. In: Holzer T (ed) Man-induced land
Special Report 02-1. Havanna, Florida subsidence. Reviews in engineering geology, vol 6. The Geologic
Clayton CRI et al (1995) Site investigations, a handbook for engineers, Society of America, Boulder, pp 67–105
2nd edn. Blackwell Science, Cambridge Holzschuh JC (1991) Field trip guidebook for the 4th international
symposium on land subsidence, 12–17 May, Houston
References 65

Hunt RE (2005) Geotechnical engineering investigation handbook, 2nd Colorado. Colorado Geological Survey, Denver, Colorado (Special
edn. Taylor & Francis, Boca Raton Publication 42)
Johnson KS (1987) Development of the wink sink in west Texas due to Sasowsky ID, White WB, Webb JA (1995) Acid mine drainage in karst
salt dissolution and collapse. In: Beck BF, Wilson WL (eds) terranes: geochemical considerations and field observations. In:
Proceedings of the 2nd multidisciplinary conference on sinkholes Beck BF (ed) Proceedings of the 5th multidisciplinary conference
and the environmental impacts of karst, karst hydrogeology: engi- on sinkholes and the engineering and environmental impacts of
neering and environmental applications. Orlando, Florida. Balkema, karst: karst geohazards, Gatlinburg, Tennessee. Balkema,
Rotterdam, pp 127–136 Rotterdam, pp 241–247
Johnson AI (1998a) Land subsidence due to fluid withdrawal in the Schmidt et al (2000) The Wakulla springs Woodville karst plain sympo-
United States, an overview. In: Borchers JW (ed) Proceedings of the sium transactions, 9 Oct 1998. Florida Geological Survey Special
Dr. Joseph F. Poland symposium on land subsidence, Land subsid- Publication 46
ence case studies and current research. Star Pub, Belmont, Shlemon RJ (2004) Subsidence and ground fissures in southern
California, pp 51–57 California: an engineering and geologic perspective. In: The
Johnson KS (1998b) Land subsidence above man-made salt dissolution Shlemon conference earth fissures, sponsored by the Association of
cavities. In: Borchers JW (ed) Proceedings of the Poland sympo- Engineering Geologists and Engineering Geology Foundation, AEG
sium on land subsidence, Land subsidence case studies and current special publication
research. Star Pub, Belmont, California, pp 385–392 (AEG Special Sowers GF (1996) Building on sinkholes, design and construction of
Publication No. 8) foundations in karst terrain. ASCE Press, Reston (out of print)
Johnson KS (2003) Evaporite-karst problems in the United States. In: Stein L et al (2004) Unseen sinkhole downs span. St. Petersburg Times,
Johnson KS, Neal JT (eds) Evaporate karst and engineering/envi- 14 Apr 2004
ronmental problems in the United States, Oklahoma Geological Stephenson JB, Beck BF (1995) Management of the discharge quality
Survey, Circular 109:1–20 of highway runoff in karst areas to control impacts to ground water,
Johnson KS et al (2003) Sinkholes and land subsidence owing to salt a review of relevant literature. In: Beck BF (ed) Proceedings of the
dissolution near wink, west Texas, and other sites in western Texas 5th multidisciplinary conference on sinkholes and the engineering
and New Mexico. In: Johnson KS, Neal JT (eds) Evaporate karst and environmental impacts of karst: karst geohazards, Gatlinburg,
and engineering/environmental problems in the United States, Tennessee. Balkema, Rotterdam, pp 297–321
Oklahoma Geological Survey Circular 109:183–195 Stracher GB (ed) (2007) Geology of coal fires: case studies from around
Kincaid TR et al (2012) Demonstrating interconnection between a the world, Reviews in Engineering Geology XVIII. Geological
wastewater application facility and a first magnitude spring in a Society of America, Boulder, Colorado
karstic watershed: tracer study of the southeast farm wastewater Sweeting MM (1973) Karst landforms. Columbia University Press,
reuse facility. Florida Geologic Survey, Tallahassee, Report of New York
Investigation 111 Tihansky AB (1999) Sinkholes, west-central Florida. In: Galloway D
Kratochvil GL (1989) Geological evaluation of potential hazards to (ed) Land subsidence in the United States, USGS circular 1182.
shuttle landing operations on Rogers Dry Lakebed, Edwards Air U.S. Geological Survey, Reston
Force Base, California. Astronaut Office NASA National Aeronautics Vattano M et al (2013) Examples of anthropogenic sinkholes in Sicily
and Space Administration, Johnson Space Center, Houston and comparison with similar phenomena in southern Italy. In: Land
Kratochvil GL et al (1992) Fishing for fissures. Mil Eng 84:66–68 L, Doctor DH, Stephenson BJ (eds) Proceedings of the 13th multi-
Land L (2013) Geophysical records of anthropogenic sinkhole forma- disciplinary conference on sinkholes and the engineering and envi-
tion in the Delaware Basin region, southeast New Mexico and West ronmental impacts of karst, National Cave and Karst Research
Texas, USA. Carbonate Evaporite 28(1):183–190 (Special Issue on Institute, Symposium 2:263–271
Integrating Science and Engineering to Solve Karst Problems) Veni G et al (2001) Living with karst, a fragile foundation, vol 4,
Legget RF (1962) Geology and engineering. McGraw-Hill, New York Awareness Series. American Geologic Institute, Alexandria
Legget RF, Hatheway AW (1988) Geology and engineering, 3rd edn. Waltham AC et al (2005) Sinkholes and subsidence, karst and cavern-
McGraw-Hill, New York ous rocks in engineering and construction. Springer-Praxis,
Londquist CJ (1995) Hydrology and land subsidence, Antelope Valley Chichester, UK
California, In: Prince KR, Galloway DL, Leake SA (eds) US geo- Ward AW et al (1995) Geologic setting of east Antelope Basin, with
logical survey subsidence interest group conference, Edwards Air emphasis on fissuring on Rogers Lake, Edwards Air Force Base,
Force Base, Antelope Valley, California, 18–19 Nov 1992. US Mojave Desert, California. In: Prince KR et al (eds) US Geological
Geologic Survey Open-File Report 94–532:38–39 Survey Subsidence Interest Group Conference, Edwards Air Force
Melvin D (1993) Storm sewer collapse, killing 1, hole swallows 3 cars Base, Antelope Valley, California, 18–19 Nov 1992. US Geologic
outside Atlanta hotel. Sun-Sentinel 15 June 1993. Available online Survey Open-File Report 94–532, p 35–37
at www.articles.sun-sentinel.com. Accessed April 12, 2012 Whittaker BN, Reddish DJ (1989) Subsidence – occurrence, prediction
Neal JT (1997) Mine induced sinkholes over the US strategic petroleum and control, vol 56, Developments in Geotechnical Engineering.
reserve (SPR) storage facility at Weeks Island, Louisiana: geologic Elsevier, New York
mitigation and environmental monitoring. In: Beck BF, Stephenson Wisenbaker M, McKinlay C (2006) A history of exploration in
JB (eds) Proceedings of the 6th multidisciplinary conference on America’s longest underwater cave, Part 3. NSS News 64(8):13
sinkholes and the engineering and environmental impacts of karst, Wisenbaker M et al (2007) Explorers connect America’s longest under-
the engineering geology and hydrology of karst terranes. Balkema, water cave to Wakulla Springs. NSS News 65(12):4–11
Rotterdam, pp 357–361 Yuhr L et al (2008) Mapping and drilling of conduits within the
Nieto AS, Young RA (1998) Retsof salt mine collapse and aquifer Woodville Karst Plain of north Florida. In: Yuhr LB, Alexander EC
dewatering, Genesee Valley, Livingston County, New York. In: Jr, Beck BF (eds) Proceedings of the 11th multidisciplinary confer-
Borchers JW (ed) Proceedings of the Dr. Joseph F. Poland sympo- ence sinkholes and the engineering and environmental impacts of
sium on land subsidence, case studies and current research, karst. ASCE Geotechnical Special Publication No. 183:340–350
pp 309–325 Zangerl C et al (2003) Ground settlements above tunnels in fractured
Noe DC, Dodson MD (1999) Heaving-bedrock hazards associated crystalline rock: numerical analysis of coupled hydromechanical
with expansive, steeply dipping bedrock in Douglas County, mechanisms. Hydrogeol J 11:162–173
 Triggering Mechanisms for Sinkholes
8

Abstract
Dealing with the engineering issues associated with karst collapse features requires an
understanding of the sinkhole development process including an understanding of the
mechanisms that trigger sinkhole collapse. Sinkhole collapse is often thought of as a natural
process. However, collapse can be triggered by either natural conditions or man’s activities
and is often a complex combination of both. If we can gain an understanding of the likely
triggering mechanisms, we can possibly avoid areas of potential collapse, design around it
or at least minimize its impact.

8.1 Statistics collapse. The causes of these sinkholes were classified into
six categories (Table 8.1). The data show that at least 66 % of
Statistics have been collected by a number of different groups the collapses were directly caused by water related factors
regarding the triggering mechanisms for collapse. (low water table and heavy rainfall). While these statistics in
Overwhelmingly heavy rain, drought or other changes in Florida generally represent roadway-related occurrences,
water conditions play a major role. Davies (1977) cites they indicate that natural and man-made changes in surface
causes of collapse recorded by the Missouri Geological water and groundwater are the dominant factors in precipi-
Survey, which had kept excellent records of collapse. Their tating collapse. They also tend to agree with independent
findings indicate that about 56 % of the collapses were findings by Newton (1976).
induced by some activity by man. The remaining 44 % were
due to natural causes.
Newton (1976, 1986), a geologist with the USGS had 8.2 Water-Related Triggering
reviewed sinkhole phenomena for the Alabama Highway Mechanisms
Department. He found that water was the dominant trigger-
ing factor and indicated that collapse was accelerated by a Water is a powerful and key component that triggers subsid-
decline in the water table that resulted in: ence and collapse of sinkholes. Changes in water levels can
be caused by many factors that include heavy rainfall,
• A loss of buoyant support to unconsolidated deposits, drought, pumping and dewatering, along with the concentra-
• An increase in the velocity of movement of water, tion of surface water run-off, to name a few. A number of
• Water-level fluctuations at the base of unconsolidated man-made structures can also lead to changes in water lev-
deposits adding to erosion and soil piping, and els. Industrial impoundments, cooling ponds, man-made
• Induced recharge. lakes and dams where the hydraulic head has been artificially
raised can all trigger sinkhole activity. While a singular event
Wisner and Denahan (1985 personal communication) may ultimately trigger a sinkhole it is usually a combination
geologists with the Florida Department of Transportation of events that has lead up to developing unstable conditions
compiled statistics over a 10-year period on sinkhole prior to the final triggering of collapse.

© Springer Science+Business Media Dordrecht 2016 67

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_8
68 8 Triggering Mechanisms for Sinkholes

Table 8.1 Causes of collapse resulting stress can contribute to breaching the overlying
Cause Percentage (%) sediments within the voids of the epikarst, creating a path-
Blasting 5 way for sediments to move downward. These rapid fluctua-
Drilling 5 tions in the water table may be natural or man-induced (such
Low water table 8 as purging of a well).
Construction 11
Other 11
Heavy rain fall 58 8.2.2 Pumping

During Florida winters, farmers pump large quantities of

8.2.1 Changes in Water Levels groundwater to spray on crops to prevent freezing. This low-
ers the water table rapidly and commonly results in sinkhole
Sinkhole development often tends to be seasonal in nature activity. An increased pumping of 11 million liters/day in a
and follow the trends in rainfall. For example, during drought well field near Tampa resulted in 64 new sinkholes within
periods, the groundwater levels are lowered by lack of 1 month (Beck and Sinclair 1986).
recharge and pumping for drinking water and agriculture
purposes. This results in a loss of buoyancy support to both
unconsolidated materials and rock. The buoyant force is 8.2.3 Dewatering
approximately 1 ton/m3. Therefore, if the water table were
reduced by 30 m the increased load would be approximately Dewatering associated with quarries, mining operations or
30 tons/m3 (Hunt 2005). In periods of heavy rainfall, ground- construction have been known to cause sinkholes nearby.
water level is raised. Such cycling of the water level cause Again, the triggering mechanism is the change in water level.
changes in hydraulic gradients between surficial and deeper A limestone quarry leased land approximately 760 m
aquifers including artesian head. These conditions create from a local college in northwest Georgia. The water table
stresses due to drying and wetting as well as changes in head was within 1.5–3 m from the surface and rock lay 3–6 m
pressure within the unconsolidated materials. The combina- below grade. As the quarry excavated deeper to find better
tion of these stresses in an area where dissolution has already quality rock it became necessary to increase pumping to
created cavities results in a high risk of materials raveling dewater the quarry and keep it dry. A few sinkholes begin to
into void spaces and ultimately resulting in sinkhole occur near the quarry and were filled with quarry spoil. With
collapse. time the sinkhole activity began to approach the college.
In many areas of Florida, especially along the elevated Eventually, a man-made lake on campus went dry and sink-
sandy ridges, sinkhole development peaks when the potenti- hole activity increased on campus. At one point, quarry
ometric surface of the Floridan is the lowest and groundwa- blasting resulted in columns of water spurting from the
ter demand is highest usually in April and May. In the ground on campus (Sowers 1996).
Orlando area, 42 % of all new sinkholes occur during the The December Giant sinkhole (or “Golly Hole”) occurred
months of April and May during the peak of the rainy season. in the isolated woods of Shelby County about 48 km south of
In the coastal lowlands, a double peak occurs in the seasonal Birmingham, Alabama in 1972. A farmer was startled by his
sinkhole frequency. One peak occurs in the dry season, but a house shaking one evening and he heard the sound of break-
second occurs, usually in August, which is the height of the ing trees. The next morning he walked in the direction of the
summer thunder-storm season (Wilson 1991). sound and found a giant sinkhole that was more than 91 m in
The heavy rainfall associated with hurricane Frances in diameter and 36 m deep (about 86,700 m3). Several obvious
September 2004 caused extensive sinkholes to occur along fissures were observed around the circumference of the sink-
its path through central and the west coast of Florida. Thirty hole. Sowers (1996) reported the presence of a large lime-
centimeters of heavy rain fell within 8-h near Ocala, Florida stone quarry about 2.5 km from the sinkhole which was more
which triggered about 200 sinkholes (Wilson 1991). Again in than 30 m below the water table and had been pumping large
2012, Tropical Storm Debbie dumped tremendous amounts amounts of water to maintain the quarry dry and operational.
of rain, at a time that Florida was experiencing a drought This pumping probably triggered the Golly Hole.
(Wilson 1991). The combination of a lowered water table
due to the drought and the heavy rainfall triggered numerous
sinkholes. 8.2.4 Surface Water Run-Off
Rapid changes in water levels, sometimes referred to as
water hammer, can create both compressional and tensional Changes in drainage patterns that result from construction
stress on trapped water and air (Chen et al. 1995). The such as major roads and their associated drainage systems
8.2 Water-Related Triggering Mechanisms 69

can result in sinkhole activity. Most highway drainage and perimeter of the sump. Localized subsidence became
repairs are (or should be) designed to utilize the natural karst numerous and continued to cause problems (Fig. 8.1). The
systems recharge. Yet inadequate drainage lines along high- initial subsidence and localized collapse were repaired only
ways commonly trigger failures (Moore 1987, 1988, 2006). to have subsequently worse problems. Major remediation
Concentrated surface water drainage from roofs of large was undertaken consisting of excavation to the top of rock
buildings and their associated paved parking lots are also a and backfilling with gravel. Water was then redirected into
factor causing sinkholes. Where buildings, parking areas and the central portion of the drainage area.
roadways are present 50–80 % of the surface area may be
covered from which surface water is being diverted and con-
centrated. Retention ponds are often used to collect surface 8.2.5 Leaky Water Pipes and Sewers
waters and divert this concentration of water away from
important structures. However, the retention ponds them- Broken water and sewer lines are the cause of much pseudo-
selves are frequently plagued with sinkholes. karst collapse. A break in a storm sewer or sanitary sewer
A commercial office complex near Frederick, Maryland line provides an opening, allowing unconsolidated materials
collected rainwater from approximately a few hectares of to flow into the sewer. An alternative is when a water line
building roofs and paved parking area that was diverted into breaks, the flow of water washes material away resulting in a
a central sump of about 0.5 ha. The top of rock was shallow sinkhole. Many of the sinkholes reported in the news are, in
(about 3 m below grade) and highly irregular with intermedi- fact, caused by such leaks. While most problems are rela-
ate size voids and small conduits. Within 1 year the concen- tively localized there are a few examples of catastrophic col-
tration of surface water began to take its toll around the lapse (see Sect. 7.3.2).

Fig. 8.1 A local storm water disposal area near Fredrick, Maryland resulted in settlement and numerous small sinkholes
70 8 Triggering Mechanisms for Sinkholes

8.3 A Guideline to Minimize Sinkholes

Triggered by Water

In South Africa, surface water is also a major triggering

mechanism for sinkhole collapse. So much so that detailed
guidelines have been developed to control the impact of
water, including leaky water lines and management of
surface water runoff. The dominant factor in the guideline
focuses upon design installation and monitoring of water ser-
vices and runoff, (South Africa Institute of Engineering and
Environmental Geologists 2003).
The following are some examples from their guidelines.

• Water bearing services must be placed above ground or in

sleeves.
• Drainage from rooftops must discharge onto sealed sur-
faces or lined furrows and then into storm water systems.
• No trees shall be planted within 1.5 times their eventual
height from the line of storm-water services.
• All ponds, water courses and road surfaces shall be
impervious.
• Criteria are provided for the design of swimming pools.
All water pipes, pumps and connections should be
installed in the open, on the surface or in service canals
where they may be inspected.

Note that the guideline focuses upon controlling a variety

of sources of water as well as the potential breaching of pipe-
lines due to the root systems of trees.

Fig. 8.2 The Guest Sinkhole in west central Florida triggered by drill-
ing a monitor well (Tihansky 1999)
8.4 Other Triggering Mechanisms

While water is the major factor in developing of karst and was down about 62 m on the third day of drilling when the
triggering of sinkholes, many other factors also may trigger ground began to give way. Within minutes, the sinkhole
a sinkhole including: swallowed the well-drilling rig, a water truck, a trailer loaded
with drill pipe and a 12 m pine tree. Figure 8.2 shows the
• Construction activities such as drilling and grouting, and beginning of the collapse around the drill rig. The sinkhole
• Vibrations, blasting or earthquakes was about 91 m in diameter and 7.5 m deep (taking about
200,160 m3 of sediment). As the sinkhole formed, it was
originally dry, however, within 24 h it had filled with water
8.4.1 Drilling Operations seeping into the sinkhole from the surficial aquifer. Ten small
satellite sinks as much as 6 m in diameter formed along a
Drilling operations can trigger collapse by creating a pathway NNW-SSE trend through the main sinkhole (Scott 2009, per-
through a confining layer, between a surficial and deeper sonal communication, Tihansky 1999).
groundwater zones. Breaching a confining layer and allowing Although there are obvious sinkholes in the region, there
sudden downward drainage from a surficial aquifer can result was no surface indication of a problem at this site. No one
in a sinkhole. Alternatively, when an artesian aquifer is breached could have anticipated the event without a more detailed inves-
the upward flow of water can also lead to a collapse. tigation of the area. An example of a large sinkhole located
The Guest Sinkhole in west central Florida occurred when about 16 km to the south is Eagles Nest, which is one of the
a saltwater monitoring well was being drilled. The drill bit premiere cave diving spots in Florida. The entrance to Eagles
8.5 Size and Rate of Sinkhole Collapse 71

Nest is typical of a large dome breakdown, a shallow lake at the Cave divers in north-central Florida reported seeing
surface with a narrow opening into a very large chamber. blocks of rock <1 m in diameter spalling off of the cave roof
Lateral tunnel development occurs at depths of 60–88 m as a train passed overhead (Mount 1973 personal
(Florea 2008). It is most likely that drilling at Guest Sink had communication). In contrast, the New Madrid earthquakes of
encountered one of the large chambers similar to those found at 1811 and 1812 had an estimated magnitude of 7.5–8.0 and
Eagles Nest, which triggered the collapse and allowed the rapid occurred within 240 km of Mammoth Cave in Kentucky.
loss of all of the equipment, trees and sediments. Workers present within Mammoth Cave at the time reported
An example of multiple sinkhole collapse at a new hous- no rock failure. Observations in Appalachian caves show no
ing development in west central Florida (Hernando County) rock falls although local earthquakes have taken place during
was presented in Sect. 7.1.4. A total of 760 sinkholes occurred the period of observations (Davies 1951).
in less than 24 h (Fig. 7.4). A deep well was being drilled for Further details are provided by Bolt (1993) who discusses
irrigation purposes. Air rotary drilling had proceeded below earthquakes, Bell (2004) who discusses blasting and earth-
the sand sediments and a clayey epikarst layer into the mas- quakes, and Hunt (2005) who discusses vibration and
sive limestone. During the subsequent development of the earthquakes.
well (a procedure that involved flushing of the well to obtain
maximum production) sinkholes began to occur.
The process of air rotary drilling had probably loosened 8.5 Size and Rate of Sinkhole Collapse
the clayey sediments within the epikarst zone due to air pres-
sure from below. The repeated flushing pressures during well 8.5.1 The Size of a Sinkhole
development continued to aggravate these conditions. The
combination of these two activities was sufficient to initiate The size of a dropout sinkhole or cover collapse sinkhole is
collapse. As the collapse process continued the head differ- related to the thickness of unconsolidated sediments and the
ence between the surficial aquifer and the deeper Floridan size or volume of the void space in the underlying rock and
aquifer aided in sinkhole development. its interconnections. When the sediments are thin and the
The many smaller sinkholes (Fig. 7.5a) were likely asso- localized voids in the top of rock are limited in size, the sink-
ciated with small isolated voids near the top of weathered holes are small.
limestone at a depth of about 12 m. The larger sinkholes Some dropout sinkhole or cover collapse sinkholes may
(Fig. 7.5b) were likely related to cavernous zones that are initially be quite small but commonly enlarge with time. At a
known to exist deeper within the lower third of the Suwannee site in North Carolina where sediments were about 12 m
Limestone, at depths of approximately 30–60 m below grade thick, small (0.3–0.6 m diameter) soil piping sinkholes
(Wetterhall 1964). would slowly migrate to the surface (Fig. 8.3a). Then a slow
process of raveling of sediments would take place resulting
in a funnel shape sinkhole of a 3–6 m diameter within weeks
8.4.2 Impact of Vibrations, Blasting (Fig. 8.3b). The sinkholes in Figs. 3.5, 7.5a, and 8.3 are all
and Earthquakes related to small isolated voids within the epikarst.
Larger dropout sinkhole or cover collapse sinkholes are
Deeper underground mines, tunnels and cave systems are gen- generally associated with thicker overburden, and occur due
erally known for the excellent stability during earthquakes. In to the presence of well-developed and interconnected open
general, vibrations, blasting and earthquakes will have little, if fracture system or an open cave system within massive rock.
any, impact to natural cave systems deeper within the rock The thicker sediments begin to slowly ravel into the fractures
mass. This is because the energy is dominated by low fre- or voids creating a void in the sediments that enlarges and
quency waves, which travel as surface waves, therefore the migrates toward the surface. At some critical point the soil
deeper caves and mines are much less affected by this energy. arch collapses forming a sinkhole. Examples of these sink-
Wisner and Denahan (1985 personal communication) have holes are seen in Figs. 3.7 and 7.5b.
noted that only 5 % of the sinkholes were caused by blasting
associated with nearby quarries. In addition, caves often evolve
into stable shapes with inherently strong and stable arched 8.5.2 The Speed of a Sinkhole Collapse
roofs. Air raid shelters have commonly been established in tun-
nels, subways, mines and caves to avoid the effects of bombs. The dissolution of rock, the formation of tertiary porosity
However vibrations, blasting and earthquakes may trigger roof- and the development of large conduits and cave systems
collapse in caves and mines which are shallow, where the rock occurs over a geologic time scale (Fig. 4.5), while sinkhole
is highly fractured and where conditions are already unstable. collapse itself occurs on a human time scale. After a
72 8 Triggering Mechanisms for Sinkholes

Fig. 8.3 Initial sinkholes at a site in North Carolina were quite small (<1 m in diameter) as a result of piping in sandy soils (a). Within a few weeks,
these sinkholes would slowly enlarge to 3–6 m in diameter as the sandy soils continued to ravel (b)

triggering event, most sinkhole collapses develop over a few • In the South Africa gold mining district, in December of
hours or a few days, most often providing ample warning for 1962 an entire Crusher plant and 29 people disappeared
human safety. However, there are a number of examples almost instantaneously. Dewatering of mine operations
where collapse has been almost instantaneous. Those unique 1 km of more below grade were likely responsible for the
situations are where both speed of development and size collapse (Brink 1984).
come together resulting in catastrophic collapse. The • In another case in South Africa, Mr. Nortjie sat in a pavil-
following illustrate some examples of rapid and large sink- ion watching four friends play a game of tennis. The
hole development: friends heard a loud noise like a pistol shot, when the dust
cleared there was nothing to be seen of the pavilion or Mr.
• On the west coast of central Florida, at a new housing Nortjie (Brink 1984).
development 760 sinkholes developed in less than 24 h • March of 2013 in a small town east of Tampa, Florida,
over an 8-ha area (Fig. 7.4). There was no loss of life or without any warning the concrete floor slab of a single
equipment in this case. story home collapses taking with it the bed and its occu-
• On the west coast of central Florida, the Guest Sink con- pant, all within a few minutes. The house was later
sumed a drill rig and two support trucks, within a matter of demolished and efforts to recover the body were aban-
10 min (Fig. 8.2) (Scott 2009, personal communication). doned. In this case the sediments had already flowed
References 73

into deeper voids resulting in a large void under the By 1 pm A fourth Porsche slipped off the edge and settled
concrete slab, which grew until eventually the slab failed 6 m below the surface, where it stayed the rest of the after-
(Altman 2013). noon. Mrs. Owens’ house was in the pit, about 7.5 m down
the slope. A fifth Porsche rolled into the pit.
In some cases, there will be precursor indications of At 2:40 pm – A portion of a large swimming pool col-
impending collapse such as cracks in the soil, pavement, or lapsed and a crane had lifted some of the vehicles out of
the structure itself. In others there may be no obvious prior the pit.
indications of impending collapse. The above examples are At 2:50 pm – There was a reverberating crash, then a
illustrations of unsuspected collapse, some of which occurred swooshing sound as Mrs. Owens’ house literally disappeared
almost instantaneously. into the abyss about 38 m below. It was about 19 h after the
In these examples, the speed at which a collapse occurs beginning of the sinkhole before Mrs. Owens’ house disap-
depends on many factors such as the character of the sedi- peared below the water. At days end, six vehicles had been
ments, clay content, partial cementation, hydraulic head dif- swallowed by the sinkhole.
ferences and the size of the opening into the rock along with Sunday Morning – The Olympic sized swimming pool,
the size and interconnection of the void space within the valued at $150,000 was ripped in half.
rock. Therefore, the speed at which a collapse takes place is In this example, there was time to prevent loss of life.
unique to the geologic setting and the temporal conditions While this was not instantaneous, there was nothing to be
(natural and man-made), which are impacting the collapse. If done to stop the collapse progress.
the sediments are loose sands, the process may proceed rap-
idly. Where the sediments contain a significant amount of
clay or are weakly cemented the mature slope may develop References
quite slowly.
Altman H (2013) Search for man under Seffner sinkhole abandoned.
The Tampa Tribune 18 Mar 2013, Tampa, Florida
8.5.2.1 The Winter Park Sinkhole Beck BF, Sinclair WC (1986) Sinkholes in Florida, an introduction.
The Winter Park Sinkhole in the Orlando, Florida area is University of Central Florida, College of Engineering, The Florida
another excellent example of a large sinkhole collapse with Sinkhole Research Institute, Orlando
thicker overburden (Fig. 3.7). It has been studied and well Bell FG (2004) Engineering geology and construction. Spon Press,
London
documented including its time sequence of events. The fol-
Bolt BA (1993) Earthquakes. W. H. Freeman and Company, New York
lowing chronological summary of the Winter Park Sinkhole Brink ABA (1984) A brief review of the South African sinkhole prob-
has been edited from the Winter Park Sinkhole Report lem. In: Beck BF (ed) Proceedings of the 1st multidisciplinary con-
(Jammal 1982). ference on sinkholes: their geology, engineering and environmental
impact, Orlando, Florida. Balkema, Rotterdam, pp 123–127
Friday at 8 pm, May 8th 1981 Mrs. Owens, a homeowner,
Chen et al (1995) Mechanical impacts of air and water compression in
heard a queer swishing noise and saw a sycamore tree disap- karst conduits. In: Beck BF (ed.) Proceedings of the 5th multidisci-
pear into the ground. She called the police department. An plinary conference on sinkholes and the engineering and environ-
initial cone shaped sinkhole 12 m in diameter and 6 m deep mental impacts of karst: karst geohazards, Gatlinburg, Tennessee.
Balkema, Rotterdam, pp 17–22
had occurred.
Davies WE (1951) Mechanics of cave breakdown. Natl Speleol Soc
Saturday at 4:00 am – Overnight the sinkhole had Bull 13:36–43
slowly enlarged to a diameter of 24 m. Mrs. Owens saw Davies WE (1977) Keynote presentation to the symposium on detection
the ground cracking on her property. Latter in the day the of subsurface cavities. US Army Engineer Waterways Experiment
Station Vicksburg, Mississippi. (lecture recorded by Benson RC)
sinkhole rapidly expanded and half of Mrs. Owens yard
Florea LJ (ed) (2008) Caves and karst of Florida. In: A guidebook for
was gone. the 2008 NSS National Convention, NSS, Huntsville
Two Porsches parked at the nearby German Car Hunt RE (2005) Geotechnical engineering investigation handbook, 2nd
Services were engulfed by the sinkhole. A Winter Park edn. Taylor & Francis, Boca Raton
Jammal & Associates, Inc. (1982) The winter park sinkhole, a report of
firefighter saw two 18 m trees sucked into the hole just
the investigation, findings, evaluation and recommendations.
like they were going down the toilet. By noon a pickup Consultant report for the City Commission of Winter Park Florida,
truck and another Porsche fell into the sinkhole. The div- March
ing board of a large swimming pool was halfway down the Moore HL (1987) Sinkhole development along ‘untreated’ highway
ditches in east Tennessee. In: Beck BF, Wilson WL (eds),
30 m slope. One corner of Mrs. Owens’ frame house was
Proceedings of the 2nd multidisciplinary conference on sinkholes
starting its descent into the sinkhole and another corner of and the environmental impacts of karst, karst hydrogeology: engi-
the house rose off the ground before sliding down the neering and environmental applications. Orlando, Florida. Balkema,
slope. Rotterdam, pp 115–120
74 8 Triggering Mechanisms for Sinkholes

Moore HL (1988) Treatment of karst along Tennessee highways. In: South Africa Institute of Engineering and Environmental Geologists
Geotechnical aspects of karst terrains: exploration, foundation (2003) Guideline for engineering-geological characterization and
design and performance, and remedial measures. ASCE Geotech development of dolomitic land. Council for Geoscience, Pretoria
Special Publ No 14, pp 133–148, ASCE, Reston, Virginia Sowers GF (1996) Building on sinkholes, design and construction of
Moore HL (2006) A proactive approach to planning and designing high- foundations in karst terrain. ASCE Press, Reston (out of print)
ways in east Tennessee karst. Environ Eng Geosci 12(2):147–160 Tihansky AB (1999) Sinkholes, west-central Florida. In: Galloway D
Mount T (1973) Preliminary Alachua cave map. Consultant report et al (eds) Land subsidence in the United States. USGS Circular
Newton JG (1976) Early detection and correlation of sinkhole problems 1182, Reston, Virginia
in Alabama with a preliminary evaluation of remote sensing appli- Wetterhall WS (1964) Geohydrologic reconnaissance of Pasco and
cations. Alabama Highway Research Report 76 southern Hernando counties, Florida. Florida Geological Survey
Newton JG (1986) Development of sinkholes resulting from man’s activi- Report of Investigation 34
ties in the eastern United States. US Geologic Survey Circular 968, Wilson WL (1991) The new picture of sinkhole development in Florida.
Reston, Virginia The Florida Specifier, Feb, Goldenrod, Florida
 Cave and Cavern Collapse
9

Abstract
This chapter presents the concept of large caverns and domes found in many caves. It
discusses the process of their development and modes of breakdown. This includes discus-
sion of factors necessary to prevent or minimize the risk of subsidence or collapse at the
surface including bulking of rock, which can sometimes restrict further collapse and the
thickness of rock needed over a cave or mine to be considered safe. Regardless of the many
rules of thumb and safety factors there are cases where collapse has propagated to the sur-
face from great depths.

9.1 Breakdown Domes may develop without the benefit of significant further
dissolution (Waltham et al. 2005).
Breakdown domes are huge caverns which form as a result These breakdown domes tend to become inherently stable
of a localized area of collapse within the rock mass. They when they develop compression arches in the roof rock. If
typically have arched roofs and have a large debris pile on there is a sufficient rock cover, they can remain stable for
the floor (Fig. 9.1a). High-arched cathedral roofs with large long periods of time. The risk of surface subsidence or col-
spans are seen in many caves and are part of their spectacular lapse is relatively low and they may constitute little, if any,
beauty. The largest dome in the eastern United States and the hazard to normal civil engineering works at the ground
second largest in the US is the Rumble Room in Rumbling surface.
Falls cave system in eastern Tennessee (Fig. 9.1b). It covers This process of developing a cavern within a cave occurs
an area of more than 1.6 ha and is approximately 120 by on a geologic time scale. A time period of 10,000–
210 m and about 48 m high (Davis 2002). 100,000 years (or more) is needed under most conditions to
These large caverns with breakdown domes tend to develop large cavernous conditions (White 1988). As the
develop at intervals along a cave system that are impacted by cavern slowly continues to expand upwards it eventually
a major fracture, the intersection of fractures, a syncline or reaches the top of the rock (Fig. 9.2c). The sediments begin
incised valley where the rock has been stressed. Such areas to fall into the cavern (Fig. 9.2c). If the sediment is sandy, a
are often more permeable allowing recharge from the surface narrow piping feature may slowly work its way toward the
over long periods of time, which further weakens the rock. surface (Fig. 9.2c). If there is clay in the sediment or it is
These zones provide an area of rock failure that slowly slightly cemented, a larger spherical void may develop and
enlarges upward resulting in a large dome-shaped cavern. remain stable for some time (Fig. 9.2e). Eventually the void
in the sediment will reach a critical state and a single event or
combination of events will act as the final triggering mecha-
9.1.1 A Conceptual Model of a Large Cavern nism resulting in a large sinkhole (Fig. 9.2f).
Once the event is triggered it may proceed rapidly in min-
In Fig. 9.2a, a fracture is shown where water slowly begins to utes or may occur over a few hours or even days. The speed
seep down into the cave weakening the ceiling rock and rock and volume of material transfer will depend upon the volume
failure begins to occur falls. Over time the localized rock fall of the void space within the rock, the degree to which the
develops into a sizeable cavern (Fig. 9.2b). This large cavern void space is open and interconnected, the size of the opening

© Springer Science+Business Media Dordrecht 2016 75

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_9
76 9 Cave and Cavern Collapse

A domed cavern with adequate overlaying

rock thickness is inherently stable, a>>b
b

Photo of large cavern Rumble Room in Rumbling Falls Cave,

Tennessee (courtesy of Chris Anderson)

Fig. 9.1 Huge caverns can develop arched domes that can be ( b ). Note person for scale standing on top of the rock debris and
inherently strong and stable for long periods of time ( a ). An exam- another person on rope (Photo courtesy of Chris Anderson, www.
ple is seen in the Rumble Room of Rumbling Falls Cave, Tennessee darklightimagery.net)
9.1 Breakdown Domes 77

a b

Rock contains
fractures

c d

e f

aven

Fig. 9.2 A conceptual model of the development of a very large cover materials. (d) Enlarged cavern continues to take soils and sinkhole
collapse sinkhole over a cavern. (a) A cave system develops along bed- expands. (e) Enlarged cavern breaks through and causes void in clayey,
ding plane. (b) Enlarged cavern develops along fracture. (c) Enlarged cohesive materials. (f) Void in soil reaches critical state and collapse
cavern breaks through and causes soil piping in sandy unconsolidated into cavern occurs
78 9 Cave and Cavern Collapse

into the rock, the properties of the sediment and the presence White (2005) provides a further summary of breakdown and
of a hydraulic head. The properties of the sediment (its range their mechanics including simple equations for analysis of
of particle size, density, viscosity, cohesiveness and degree fixed and cantilever beams.
of cementation, if any) will affect the rate and amount of the White (2005) summarizes the factors that contribute to
material that will freely flow by gravity into the void space. cave roof failure. The processes that initiate breakdown
The presence of a surficial water table will tend to fluidize include:
the sediments and hydraulically add to the downward move-
ment of sediment. This often results in a flushing noise dur- • Passage enlargement below the water table as water flows
ing the collapse. In many cases, the surficial water table through the cave
becomes depressed and the sinkhole may be dry immediately • Removal of water removes the buoyant support (about
after the collapse. The sinkhole eventually fills as the water 42 %) of the roof rock
level slowly recovers and conditions begin to stabilize. • Effects of occasional flooding can destabilize the roof
• Action of vadose water flowing into the cave from above
enlarges fractures and changes fixed beams into cantilever
9.2 Mechanics of Cavern Breakdown beams.
• Ice wedging within joints further fracture the bedrock
In 1951, Davies presented the mechanics of cavern breakdown • Crystal wedging, replacement by minerals such gypsum
by brittle fracture. He describes two types of breakdown, can fracture the bedrock
block and slab. Block breakdown occurs with blocks that are
typically short laterally and have significant thickness. Slab Under these various conditions, collapse will have slowly
breakdown occurs with blocks that are typically limited in progressed over time. The process can be observed by the
thickness to a single or a few beds and are longer laterally. slow enlarging of the cavern roof and the pile of rock debris
Ellis (1994) reported an example of slab breakdown in on the cave floor. In many caverns the debris pile on the floor
Mammoth Cave. An estimated 100-ton slab of roof rock is much smaller than the void from which the rock came.
about 6 by 21 m and 0.3 m thick fell 12 m from the ceiling of This is due to rapid flowing water through the conduit system
the rotunda near the entrance to Mammoth Cave. It is thought resulting in dissolution of the fallen rock and the transport of
that the extremely cold air caused the slab to pry loose rubble and sediment through the conduit system thereby
because of wedging effect of freezing moisture. Temperatures maintaining large open conduits.
outside had dipped as low as −16 °F and were accompanied
by heavy winds.
An example of massive block breakdown occurred instan- 9.3 Thickness of Rock Needed to Prevent
taneously over a 6.8-ha area referred to as the “central col- Surface Subsidence or Collapse
lapse area” (CCA) in the Tobin limestone mine in the Kansas
City area. The 6.8-ha CCA occurred shortly after the area Various rules of thumb have been used to predict safe levels
had been mined in 1973. It’s air blast shattered windows of of rock thickness necessary in order to prevent subsidence or
the trucks within the mine and reportedly was recorded on collapse. Waltham and Fookes (2003), Waltham et al. (2005),
seismographs some 48 km away. Further details are dis- and Waltham and Lu (2007) all provide a number of exam-
cussed in Chap. 25. ples and guidelines showing cave roof stability in limestone
with varying loads and rock quality.
Waltham and Fookes (2003) suggest a rule of thumb in
9.2.1 Two Modes of Breakdown: Fixed Beam which the ground is considered stable if the overlaying thick-
and Cantilever Beam ness of the rock is equal or greater than the roof span exclud-
ing any soil cover. This concept is conservative and would
Fixed beam breakdown occurs when the rock strata of the apply to a limestone with a normal density of fractures and
cave roof extend from one side of the cavern to the other, bedding planes and excludes any heavily fissured limestone.
supported at both ends with no fractures in between Waltham et al. (2005) provide a discussion on safe cover
(Fig. 9.3a). Cantilever beams are those supported on one side thickness over caves. Their rule of thumb suggests that the
of the cavern and extend into the cavern some distance roof thickness should exceed 70 % of the cave width.
unsupported (Fig. 9.3b). Long fixed beam strata will eventu- Guidelines for safe cover must be increased to accommodate
ally sag and fracture resulting in a cantilever beam (Davies gypsum, chalk, and weak limestone.
1951). Failure of roof rock tends to occur gradually a slab at Full scale loading tests carried out by Waltham and
a time as local stress ultimately leads to failure. The simplest Swift (2004) within man-made caves in the sandstones of
approach for roof breakdown is the brittle fracturing model Nottingham provide further insight to the thickness of rock
based upon work from the mining literature (Davies 1951). necessary for stability over such caves or mines. They
9.4 Experience from Mine Failures 79

a fixed beam scenario in which the roof beam is

continuous and supported at both ends
b

cantilever beams are only supported at one end

Fig. 9.3 A fixed beam (a) and a cantilever beam (b). The length (L) and thickness (T) of the beam as well as its density and flexural strength are
factors in roof stability

concluded that thickness of overlaying rock should exceed All such rules of thumb are only guidelines and must be
50 % of cave width for most projects on typical karst lime- tempered by site-specific data and observations. Site-specific
stone. Where higher loads are imposed, cover thickness data should include establishing bed thickness, rock quality,
would need to exceed 70 % of cave width to be considered fracture spacing and apertures along with identifying other
stable. Numeric modeling has also been done by Waltham conditions such as lineaments, major fractures, incised val-
and Lu (2007) to determine the thickness of rock cover leys and structure which may indicate weak rock conditions.
needed to avoid subsidence over a cave when additional This data must then be combined with experienced profes-
loading is imposed by engineered structures. An intact sional judgment.
rock-cover thickness that exceeds half the cave’s width
appears to be safe in most karst terrains in strong
limestone. 9.4 Experience from Mine Failures
Waltham et al. (2005) discuss the range of roof thickness
required for safe conditions with various rock mass ratings We can also learn about collapse and subsidence from expe-
(RMR). If the RMR for a typical cavernous limestone is rience with mines because there is much published literature
taken conservatively as between 30 and 40, a cover ratio of from the mining industry. Miners, engineers and geologists
roof thickness being half the cave width (t = w/2) appears to have often observed roof rock failure in mines shortly after it
be adequate for most engineering practices. occurs and have been able to quantify it. However, it is rare
80 9 Cave and Cavern Collapse

that observations and analysis of roof failure in cave systems lithologic and structural discontinuities. Lithologic condi-
have been made, although cavers have photographed the tions may include paleochannels, scours, rider seams, clay
results of some failures. veins, or crevasse splay deposits. Fractures (joints and faults)
Cavern breakdown and roof failure in a mine are similar also contribute to roof falls. Stress fields both tectonic and
with the following exceptions, Davies (1951): mining-induced can determine whether failure will occur,
and the mode of such failure. They suggest that paleoenvi-
• A lack of blasting in caves ronmental studies are necessary in determining hazardous
• The mine has a short-term development of 10’s of years roof conditions and should be incorporated into the mine
or so while the cavern has evolved over 100,000’s of years design.
or more. In summary, where mines (and natural cavities) are near
• The stress are fully developed and are in equilibrium the surface with little rock cover, roof collapse and subsid-
within the cavern ence are common. However, in those situations where there
• There is an absence of any artificial supports within the is abundant rock cover, with strong rock beds, the risk of
cavern subsidence or collapse is lower, yet it may still occur. Both
cavern and mine breakdown are a function of the site-specific
As a result, stress fields around mines tend to be unstable geologic conditions, the width of span along with fracture
compared to natural cave systems, which have developed spacing, hydrologic conditions, and the thickness of the
over long periods of time where stress tends to be in overlaying rock.
equilibrium. The authors have spent hundreds of hours in the Tobin
Much of the experience that has evolved within the min- Limestone Mine and many hours in other abandoned and
ing industry can be used to help understand the collapse of modified limestone mines in the Kansas City area. In some
mines as well as collapse of cave systems. There have been cases, stable roof spans of up to 30 m have been identified
numerous efforts to predict the magnitude and extent of mine while shorter spans of 15 m have commonly failed. However,
subsidence (mostly coal mines). While good mine design our observations in the Kansas City limestone mines show
and practice minimizes failures, the problems of subsidence the typical localized breakdown begins as localized spalling
remain pervasive. as described by Hasan et al. (1988).
Gray and Salver (1971) provide an early summary of This localized spalling is commonly initiated where
state-of-the-predictive capabilities for subsidence over coal paleochannels of rubble rock have breached the mine roof
mines. The National Coal Boards Subsidence Engineering exposing the “rubble zone”. The rubble zone is a calcite
Handbook (1975) became the definitive early reference for mudstone and silty claystone (shale). Some of the rubble
coal mine subsidence in England and elsewhere. More recent zone contains swelling clay minerals (smectite-illite mixed
work includes; an excellent review by Whittaker and Reddish layers). When exposed to the moist air within the mine,
(1989) along with a recent summary of Empirical Rock localized breakdown occurs, typically less than a cubic yard.
Failure Criteria (Sheorey 1997), which provides an excellent The zone of breakdown slowly enlarges and may ultimately
overview of approaches to mine roof failure and the resulting include up to ½ ha or more of collapse extending upward 3 m
surface subsidence. Kane et al. (1993) provide an excellent and in some cases more. The presence of the rubble zone and
review of the geologic factors affecting coal mine roof stabil- its impact on mine collapse is discussed in more detail in
ity in the eastern United States. These approaches utilize a Chap. 25. By paying attention to lithologic and structural
variety of numeric and empirical methods to predict the lat- discontinuities such as paleochannels, clay veins, joints and
eral extent and depth of surface subsidence. faults remedial action can then be taken (Kane et al. 1993).
While most of these authors mention the importance and There are three conditions that are cause for concern of
impact of geologic conditions, they do not discuss them in roof failure in caves or mines which can result in surface
any detail. For example, Whittaker and Reddish (1989) sug- subsidence or collapse (Fig. 9.4):
gests that local geology and natural strength of the immedi-
ate roof (and overlying strata) are important factors. However, • a wide roof span at shallow depth (a).
the mining dimensions and geometry of workings are of • a roof span at depth with weakened rock cover, such as an
equal importance and should be considered in making assess- incised valley and fractured rock which has allowed infil-
ments of subsidence risks above such mines. Kane et al. tration of surface waters resulting in weathering and
(1993) found that geologic factors are a major factor in fail- weakened rock (b).
ure of coal mine roofs. Geologic conditions which contrib- • the roof of a localized enlarged cavern which is approach-
uted to roof failure in underground coal mines include both ing a critical state of stability (c).
9.4 Experience from Mine Failures 81

Zones of Possible Collapse

a b c

Fig. 9.4 Examples of where roof collapse can result in surface subsidence or collapse. (a) Wide span with thin rock cover. (b) A thick overlying
rock with a localized zone of weakened rock cover such as an incised valley or major fracture. (c) Enlarged cavern in a critical state

9.4.1 Bulking of Fallen Roof Rock Piles of breakdown rock on the floor are random in size
and orientation and highly permeable. Massive breakdown
As rock falls from the roof of a cave or a mine it commonly (Fig. 3.15), that completely blocks a cave passage, is referred
falls as small to medium size blocks that rotate while falling to as terminal breakdown (White 2005). In many cases, the
and break, ending up in a random pile of rock. This broken quantity of debris on the cave floor is much less than the
rock will occupy a greater volume than they did when the volume of rock that has fallen from the roof. This indicates
rock was in place (Fig. 9.5a) this is known as bulking factor. that large quantities of rock have been removed when water
As the upward collapse continues, the greater volume of the was flowing through the cave system. Examples of such con-
fallen rock begins to fill the void space (Fig. 9.5b). At some ditions are seen in many large caverns and in sinkholes
point, the volume of bulked rock will nearly or totally fill up accessed by cave divers where the volume of fallen rock
the void space thereby, preventing further roof failure debris on the cavern floor is substantially less than the vol-
(Fig. 9.5c). ume of rock removed from the cavern roof.
Typical bulking factors from the coal mining industry are There are factors that can reduce bulking of fallen rock
0.3–0.5. That means that the expanded volume of fallen rock that include:
is 30–50 % more than its in-situ volume due to its breaking
into pieces (Piggott and Eynon 1978). This factor may be as • The bulking factor may decrease if there is flowing water
low as 0.2 if claystone and shale are present (Dunrud and through the rubble leading to further degradation and
Osterwald 1980). Blasted rock will typically have the highest transport of the fallen rock.
bulking factors of 50–60 % (Colaizzi, 1990, personal • The bulking factor will decrease as the void space
communication). becomes filled reducing the distance the rock falls and its
The authors have made extensive direct measurements tendency for the rock slabs to rotate and fracture.
and photo documentation of bulking in a limestone mine in • Some compaction and possible degradation can be
the Kansas City area. Measurements were made of the vol- expected after the fallen rock has been in place for a
ume of the void created by the fallen rock and the volume of while. Rock strata with a high shale content will have less
the rock rubble pile on the mine floor (Fig. 9.6a). These bulk- bulking factor because of weaker shale crumbling on
ing measurements yielded an average bulking factor of 0.4 impact and being compacted into voids within the rubble
(Benson et al. 2000). However, in some cases rock may fall pile without adding much to the bulking factor.
as a massive slab resulting in bulking factor close to zero
(Fig. 9.6b). More details from this case history is presented Consideration of the bulking is another factor which
in Chap. 25. Piggott and Eynon (1978) and Bell (2004) pro- allows us to estimate the risk of surface collapse for a given
vide equations for various geometries and plots of maximum rock thickness over a cavern or mine. However, without a
height of collapse for various bulking factors. detailed understanding of the local rock conditions (rock
82 9 Cave and Cavern Collapse

Fig. 9.5 Bulking of fallen rock begins to build up (a), slowly filling the void space (b), and eventually choking off further collapse (c)

strength, fracture patterns, structure and local hydrology) Some of the most extensive karst in the United States
these bulking factors should only be used as guidelines. occurs in the limestone and dolomite beds of Mississippian
age (about 325–345 million years old), which have a com-
bined thickness of over 120 m. Extensive cavern develop-
9.5 Propagation of Subsidence ment has occurred in the Mississippian rocks and is
and Collapse from Great Depths recognizable throughout the mid-continent (Fig. 9.7a)
(Gentile 1984). The Pennsylvanian System, a section of
It is relatively easy to understand the collapse reaching the shale and limestone, about 150–180 m thick lie over the
surface in the presence of thinly bedded and fractured Mississippian limestone and dolomite beds in the Kansas
strata over a shallow cave or mine. It is difficult to concep- City area. A number of large cavern collapse have occurred
tualize subsidence at the surface due to collapse of very within the thick Mississippi Limestone have propagated
deep natural cavities or mines. Especially when consider- upward through the Mississippian and Pennsylvanian System
ing the issues of bulking, thick sections of overlaying rock (more than 200 m) and have been identified in the Kansas
(100’s of m or more) with massive strata. Nevertheless, City area road cuts (Fig. 9.7b). There are places where large
collapse from great depths does occasionally propagate to blocks of bedrock have moved downward a few meters along
the surface. high angle, normal faults (Gentile 1984).
9.5 Propagation of Subsidence and Collapse from Great Depths 83

Fig. 9.6 An example of rock bulking that will eventually choke off further collapse (a) and an example of a single large slab failure of roof rock
approximately 3 by 6 by 2.25 m thick resulting in low bulking (b)

Deep mines have also resulted in surface subsidence. A weakening a zone of rock. Then surface operations require
few examples are: an impoundment be constructed or surface water drainage be
diverted concentrating it over the fracture. Or a building and
• Dewatering a deep gold mine in South Africa (>1,000 m parking lot is built concentrating large amounts of runoff into
below grade) resulted in a few rapid catastrophic col- the fracture contributing to its further weakening and ulti-
lapses (Brink 1984). mately resulting in a deep cave or mine collapse reaching the
• A room and pillar coal mine collapse 365 m underground surface. Here the presence of either the natural fracture sys-
occurred near Birmingham, Alabama, and reached the tem or the surface modification independently may not have
surface (Colaizzi, 2004, personal communication) triggered collapse, but when combined, the two, further
• The Retsof room and pillar salt mine more than 300 m weaken the rock allowing propagation of the deep collapse to
below grade in New York State resulted in surface subsid- reach the surface.
ence (Nieto and Young 1998). The structural integrity of natural cave systems and mines
are impacted by both geologic conditions and man’s activi-
Roof collapse in deep-seated mines or caves can only ties. The natural geologic and hydrologic conditions above
reach the surface when the overlaying rock has been weak- caves or mines combined with certain impacts of mans activ-
ened by some means such as the presence of a major fracture ities are, in fact, the dominant factors in a collapse propagat-
zone, fault or structural weakness. For example an unknown ing to the surface. In many cases when the details are
major fracture system may intersect a mine. Rainfall may established we find that there is a geologic flaw present,
seep downward in the fracture over long periods of time which has lead to the subsidence or collapse.
84 9 Cave and Cavern Collapse

a 0

Approx. Depth in Meters

Pennsylvanian
100
System
limestones/shales

200

Mississipian
Limestones 300

conceptual model of deep-seated

karst (based upon Gentile, 1984)
b

road cut showing evidence of deep-seated karst

Fig. 9.7 Deep-seated collapse has occurred in the Mississippian Limestone that reached the surface in and around Kansas City area (a) (Based on
Gentile 1984). Edges of such a paleocollapse along Interstate-470 near Kansas City, Missouri with arrows showing the locations of fault scarps (b)

References Gray RE, Salver HA (1971) Discussion of state of predictive art in

subsidence engineering. J Soil Mech Found Div, ASCE
97(SM1)
Bell FG (2004) Engineering geology and construction. Spon Press,
Hasan S et al (1988) Geology of greater Kansas City, Missouri and
London
Kansas, United States of America. Bull Assoc Eng Geol
Benson RC et al (2000) Stabilization against geologic instability-
XXV(3):281–341
underground documentation of a mine backfilling case history.
Kane WF et al (1993) Geologic factors affecting coal mine roof stabil-
ASCE Geo-Denver Conference Denver, Colorado, 5–8 August 2000
ity in the eastern United States. Bull Assoc Eng Geol
Brink ABA (1984) A brief review of the South African sinkhole prob-
30(2):165–179
lem. In: Beck BF (ed) Proceedings of the 1st multidisciplinary con-
National Coal Board (NCB) (1975) Subsidence engineer’s handbook.
ference on sinkholes: their geology, engineering and environmental
NCB Mining Department, London, 111p
impact, Orlando, Florida. Balkema, Rotterdam, pp 123–127
Nieto AS, Young RA (1998) Retsof salt mine collapse and aquifer
Davies WE (1951) Mechanics of cave breakdown. Natl Speleol Soc
dewatering, Genesee Valley, Livingston County, New York. In:
Bull Number 13:36–43
Borchers JW (ed) Proceedings of the Dr. Joseph F. Poland sympo-
Davis C (2002) Exposing Tennessee’s titanic cave chamber. National
sium on land subsidence, case studies and current research, pp 309–
Geographic Adventure Magazine, March 4(2): on-line extra at
325, Star Pub., Belmont, California
www.nationalgeographic.com
Piggott RJ, Eynon P (1978) Ground movements arising from the pres-
Dunrud CR, Osterwald FW (1980) Effects of coal mine subsidence in
ence of shallow abandoned mine workings. In: Geddes JD (ed)
the Sheridan, Wyoming area, USGS professional paper 1164,
Proceedings 1st international conference on large ground move-
Reston, Virginia
ments and structures. Pentech Press, London, pp 749–780
Ellis I (1994) Rockfall in Mammoth Cave. NSS News April 52(4):129
Sheorey PR (1997) Empirical rock failure criteria. Balkema,
Gentile RJ (1984) Paleocollapse structures: longview region, Kansas
Rotterdam
City, Missouri. Bull Assoc Eng Geol XXI(2):229–247
References 85

Waltham AC, Fookes PG (2003) Engineering classification of karst Waltham AC et al (2005) Sinkholes and subsidence, karst and cavernous
ground conditions. Geol Soc Lond Q J Eng Geol Hydrol 36:101–118 rocks in engineering and construction. Springer-Praxis, Chichester
Waltham AC, Lu Z (2007) Natural and anthropogenic rock collapse White WB (1988) Geomorphology and hydrology of karst terrains.
over open caves. In: Parise M, Gunn J (eds) Natural and Oxford University Press, New York
anthropogenic hazards in karst areas: recognition, analysis and miti- White EL (2005) Breakdown. In: Culver DC, White WB (eds)
gation. Geologic Society special publications 279. Geologic Society, Encyclopedia of caves. Elsevier, Burlington, pp 56–60
London, pp 13–21 Whittaker BN, Reddish DJ (1989) Subsidence – occurrence, prediction
Waltham AC, Swift GM (2004) Bearing capacity of rock over mined and control, vol 56, Developments in geotechnical engineering.
cavities in Nottingham. Eng Geol 75:15–31 Elsevier, New York
 Insight into the Nature of Cover
Collapse Sinkholes 10

Abstract
The focus of this chapter is upon providing insight to the depth, size and nature of the void
space in the rock based upon the size and speed of cover collapse sinkholes. Insight is pro-
vided by scale model tests of cover collapse sinkholes, mine backfilling as well as data on
many large and very large collapse features. Five levels of conceptual models are presented
relating the size of the sinkhole to the possible size and nature of void space within the rock.
These models can provide a first approximation to the nature of conditions at a site.

10.1 Introduction These data have provided us with the ability to develop
simplistic conceptual models relating the size of a cover col-
Cover collapse sinkholes are also known as dropout sink- lapse sinkhole to the approximate size and nature of the void
holes, subsidence sinkholes, or alluvial sinkholes, (Fig. 3.1 space within the rock.
and Table 3.1). These are sinkholes in which sediments have
collapsed into the void space within the rock. These sink-
holes have a wide range of sizes and occur under a wide 10.2 Insight from Scale Model
range of conditions. Sinkhole dimensions (diameter and Sinkhole Tests
depth) along with the volume of sediment lost are commonly
measured or estimated. In some cases, we may also know the Chen and Beck (1989) discuss the use of a 2D sand model to
size of the opening into the rock (aven) along with the time illustrate the development of a cover collapse sinkhole. They
required for the collapse to occur. What we normally do not used a plexiglass model 132 cm long by 91 cm high with
know is the size, depth, geometry and openness of the void wooden sides and bottom. Its width was adjustable from 6 to
space within the rock, which is often one of the many goals 18 cm. Six holes from 0.6 to 5 cm were drilled in the bottom
for a site characterization in karst terrain. to simulate the opening into the limestone. Twenty-nine dif-
Over the years we have had the opportunity to observe the ferent trials were run with a variety of conditions ranging
nature and behavior of a variety of cover collapse sinkholes from dry sand to saturated conditions, their results
and to investigate the conditions under which these sinkholes indicated that:
have developed. In addition, we have gained some insight
into the size and nature of void space in the rock by the • Gravitational movement of the sediment is a major factor
amount of sediment lost based upon: in sinkhole development.
• Very small initial openings (avens) in the limestone can
• A review of scale model tests showing the general behav- result in large sinkholes if there is sufficient void space
ior of cover collapse sinkholes deeper within the limestone.
• Extensive experience in backfilling of mines by others • A more cohesive sediment strata may slow the upward
• Direct observations and measurements of mine backfill development of the collapse resulting in horizontal
materials injected into an abandoned limestone mine enlargement below the more cohesive strata.

© Springer Science+Business Media Dordrecht 2016 87

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_10
88 10 Insight into the Nature of Cover Collapse Sinkholes

• Infiltrating recharge aided the collapse process. The • From mine backfilling literature which include scale
hydraulic head drops discontinuously with the surround- model tests and full scale projects by the Bureau of Mines
ing water table, creating a hydraulic jump. The water flow (Colaizzi et al. 1981)
over the opening is vertically downward with a great • Experience backfilling abandoned coal and limestone
scouring effect mines from Goodson Associates (Colaizzi, 1990, per-
sonal communication) and
Additional sand model results are shown in Whittaker and • Direct observations and documentation from mine back-
Reddish (1989) who show the time sequence of collapse filling of the Tobin limestone mine in Kansas City Kansas
with similar results. These scale model test results are in (Benson et al. 2000).
good agreement with many observed sinkholes over known
geologic conditions.
10.3.1.1 Scale Model Tests and Experience
with Full Scale Projects
10.3 Insight from Mine Backfill Colaizzi et al. (1981) presented the results of two scale model
Stabilization tests of backfilling of mines. The conclusions of these
tests were:
Insight into the amount of sediment or material that can flow
into a large open void space can be obtained from experi- • The slurry was generally distributed over an area of
ence filling abandoned room and pillar coal and limestone approximately 0.4 ha.
mines with fly-ash, sand, and crushed rock. At some point • Deposits within the model were not dependent upon the
the mine becomes sufficiently full of material so that the concentration of the sediment or injection velocity, pro-
system chokes off before the entire mine space is full. This viding the velocity was high enough to transport the slurry
data provides us with insight and practical limits to the through the injection pipe.
amount of material that can be placed into a very large or • Backfill material extended further from the injection hole
infinite open space within a mine from a single opening (a in a flooded mine as opposed to a dry mine.
borehole). • The extent to which backfill material was transported was
more or less independent of the slope of the model.
• Backfill material was transported over areas of roof falls
10.3.1 Comparison Between Sinkholes that blocked corridors.
and Mine Backfilling • Backfill material was transported into blind areas as long
as a circulation of sediment laden water occurs in and out
The mine backfilling process uses a small diameter borehole of the blind areas.
(of 20 cm or so). Larger cover collapse sinkholes usually • As resistance to flow of the slurry develops, a new chan-
pass through a larger opening in the rock (1–15 m or more). nel is formed in another direction along a line of less
The mine backfilling process also takes place over a few resistance.
days, longer than most sinkhole collapse that typically occur • At the end of each test a backpressure is developed within
over a shorter period of time of a few minutes to a day or the injection system and a final breakout of flow occurs
more. The void space in a mine is laterally uniform and along an unobstructed corridor. The entire flow from the
extensive compared to dissolution features like a cave, which injection pipe then travel down this channel transporting
are often tubular in extent and highly variable. material long distances from the injection pipe to the
The particle sizes of backfill materials (fly ash mixtures, perimeter of fill material.
sand and crushed rock) are reasonably similar to the sedi-
ment lost into a sinkhole. The backfill material is com- In addition, three full-scale backfilling projects were car-
monly fluidized similar to the sediment being fluidized by ried out by the Bureau of Mines (Colaizzi et al. 1981). The
the presence of a surficial water table. In addition, the back- largest amount of fill placed into a single fillhole at each of
fill material is usually under a high head of 15–30 m or the three sites respectively, was 29,800, 53,300 and 63,500 m3
more, similar to the sediment in a large cover collapse sink- (Table 10.1).
hole. While there are differences between the two processes
there are sufficient similarities to make useable compari- 10.3.1.2 Experience in Backfilling of Mines
sons between natural sinkholes and the mine backfilling Goodson and Associates of Denver, Colorado have had
process. extensive experience stabilizing both coal and limestone
Data from mine backfilling has been obtained from three mines with fly ash, sand and crushed rock. They have placed
sources: up to 54,500 m3 in a 2.5 m high mine through a single cased
10.3 Insight from Mine Backfill Stabilization 89

Table 10.1 Summary of mine backfill data

Maximum amount of
Name Dimensions backfill in one borehole (m3)
Bureau of Mines Report (Colaizzi et al. 1981) Two scale model test and three 29,800
full scale backfilling projects 53,400
63,500
Backfilling of coal mine, Goodson Associates, 2.5 m high coal mine with an 54,500
(Colaizzi, 1990, personal communication) estimated 60 % extraction
Backfilling with crushed limestone in the 4.25 m high with 85 % extraction Up to 7,600–15,300a
Tobin mine (Benson et al. 2000)
a
These numbers are on the low side because in all cases fill was being placed against mine walls or adjacent to previous fill on at least one side,
thereby limiting the total amount of fill material that could be put into a single fillhole

borehole (20–33 cm) before the system became choked off. point, the system would become choked-off and the fill pro-
Fly-ash mixtures and sand will flow for 30 m or more from cess would cease. The slope (angle of repose) around the
the fillhole into an air or water-filled mine. In some cases, perimeter of fill ranged from 4 to 9° and averaged 6°.
flow up to 300 m from the fillhole has been observed As part of a QA/QC process to document conditions of
(Colaizzi, 1990, personal communication). backfill for the State of Kansas, both photographic as well as
observations and measurements were made of the extent of
10.3.1.3 Direct Observations of Mine Backfilling fill after each fillhole had choked off. Additional details are
in a Limestone Mine included in Benson et al. (2000) and in Chap. 25, the first
A landfill was being constructed over an abandoned lime- case history.
stone mine in the Kansas City area. The Tobin mine was a A spacing of approximately 45–60 m was selected for
room and pillar mine with a height of 4.25 m, pillars of the fillhole borings to achieve 90–100 % filling of the mine.
approximately 7.5 by 7.5 m and spaced on about 18 m cen- The amount of fill placed in each hole would vary from
ters. Even though there was 52 m of limestone and shale over 7,600 to more than 15,300 m3. This volume is on the low
the mine, the state required that the mine be filled to 90 % or side since fill was being placed against mine walls, and or
3.8 m to prevent possible collapse at the surface. existing mine fill on one side or more. As a result the
Goodson Associates was employed to fill the mine using amount of fill in each borehole would be limited depending
a crushed rock slurry. The uppermost rock at the site upon the location of nearby mine walls and previous fill
(Argentine Limestone) was removed by ripper and crushed material. Portions of the mine were dry while other areas of
to 1.25 cm. The rock was then mixed with water and pumped the mine were filled with water. There was little difference
through a 20 cm pipe to an open borehole into the mine at between the amount of fill that could be placed in a dry or
5,680 l/min. Water was pumped from the mine to the surface wet portion of the mine.
and injected back into the mine as part of the fill process.
Direct observations, photographs and video of the mine 10.3.1.4 Conclusions from Mine Backfilling
backfilling process were made from within the mine as the The experience from backfilling mines (Table 10.1) shows
fill was being injected. Initially we were within a few meters that there are limits to the amount of fill material which can
from the injection hole. The fill material would eventually be placed into a large open space through a single opening.
reach the roof of the mine and would then cut a channel for The amounts of fill material ranged from 7,600 to 63,500 m3.
the crushed rock to flow to the perimeter (Fig. 10.1a). At These observations and measurements show that:
approximately 20 min intervals the channel would choke-off
and a new channel would be cut in a different direction • There is a maximum amount of backfill material that can
allowing crushed rock to flow to the perimeter. This sequence be placed in a single fillhole into an open mine space, and
of breakthroughs would continue to occur in a somewhat • The void space will generally be greater than the maxi-
random pattern, 360° from the injection point until an area of mum volume of backfill material.
about 23–30 m radius around the fillhole was backfilled to
the mine-roof (100 % full) and filled to the 90 % level over a This information can provide insight into understanding
distance up to 68 m from the fillhole (Fig. 10.1b). At this of cover collapse sinkholes.
90 10 Insight into the Nature of Cover Collapse Sinkholes

a Injection Borehole
Mine Back (Roof)
Pillar
Mine Floor

High pressure and flow rates

allow slurry to cut channel to
edge of fill

Typical backfill to roof

or 95% closure

Sketch of mine backfilling process

b

Crushed rock
backfill
to mine roof

Slope of
backfill

Photo of crushed rock backfill in place

Fig. 10.1 A sketch of the crushed rock fill process showing progressive amounts of fill (a). A photo of shows the crushed rock fill to the roof of
the mine and the slope of the edge of fill (b) at this time, water had filled the 4.25 m mine, more than half full

10.4 Conceptual Models of Cover Collapse • Larger cover collapse sinkholes associated with a cave
Sinkholes system
• Very large cover collapse sinkholes associated with the
Five generic conceptual models have been developed which presence and collapse of a large cavern
relate the size of a sinkhole collapse (volume of material) to • Extremely large cover collapse sinkholes associated with
possible void space in the subsurface. We have simplified evaporate karst and some types of pseudokarst
issues and are assuming a one-time collapse event where the
sediment is not washed away by subsurface flowing water.
These conceptual models are based upon our sinkhole inves-
tigation experience, insight from scale model sinkhole tests 10.4.1 Small Cover Collapse Sinkholes
and experience from backfilling of mines. The models are
simplified and assume one dominant type of dissolution fea- Smaller cover collapse sinkholes are due to movement of
ture. The many other variables that may be present have been sediments into isolated voids, fissures, and cavities in the
ignored for the sake of simplicity. The five models include: epikarst zone or upper surface of the rock. These sinkholes
are generally small (on the order of 1–5 m in diameter and
• Small cover collapse sinkholes associated with voids in volumes of 1–15 m3) and result in limited amount of sedi-
the epikarst zone or top of rock. ment lost. While the percentage of dissolved bedrock within
• Intermediate size cover collapse sinkholes associated the epikarst can range up to 50 % or more, a high percentage
with fractured rock of this void space will be filled with sediment (Aley 1997).
10.4 Conceptual Models of Cover Collapse Sinkholes 91

This results in small voids, which are commonly isolated or be in a fluidized state with a large hydraulic head to
poorly interconnected. accommodate larger volumes of sediment.
The conceptual model for this size of cover collapse
includes a shallow top of rock (about 15 m or so, our guess-
timate) and overlying sediment that is dominantly sands with 10.4.3 Large Cover Collapse Sinkholes
limited clay content or cemented layers. This allows piping
or raveling of soils to occur vertically over small isolated Large cover collapse sinkholes include those that are
void spaces and results in small individual sinkholes. 30–50 m in diameter with a material loss of greater than a
Two such examples include: few 1,000–50,000 m3. A conceptual model for this size of
collapse would be a typical open cave system.
• A site in North Carolina where the limestone was 12 m Based upon data from Mammoth Cave its mean cavern-
below grade. A small diameter sinkhole (about 0.3–0.6 m ous porosity has been estimated by Quinlan (1989, personal
in diameter) with vertical sides would slowly work its communication) at about 5 % and by Worthington et al.
way to the surface through mostly sandy soil (Fig. 8.3a). (2000) between 0.36 and 3.3 % of the subsurface area. This
Within a few weeks or so, the sinkhole would enlarge to is very similar to the fractured rock model by Waltham et al.
about 3 m in diameter with more sloped sides (Fig. 8.3b). (2005) of 4.6 %. While the fracture systems are generally
• A site in west central Florida had 760 sinkholes develop more spatially prevalent they tend to be smaller and tend to
overnight (Fig. 7.4). The vast majority of them were be in-filled thereby limiting the amount of sediment they can
small, a meter or so in diameter (Fig. 7.5a). These smaller accommodate. Caves are less spatially prevalent, but tend to
sinkholes were likely associated with small isolated voids have a larger open cross section that can accept greater
within the upper surface of the limestone about 12 m amounts of sediment than fractured rock. Therefore, it is rea-
below grade. sonable that sinkholes over a cave system would be larger
than a fractured rock system.
While sinkholes originating in the epikarst generally The open void space of a cave system would be much less
manifest themselves as small individual sinkholes, the indi- than that of an open mine due to the typical linear nature and
vidual sinkholes may some times merge and appear as larger meandering of a cave system. Since there is a limited amount
ones. In addition, when the sediment thickness is small the of fill material that can be placed into an open mine
dimensions of the cover collapse sinkholes are quite limited. (Table 10.1) this suggests that there is an inherent limit as to
As the sediment thickness increases and the void space in the the amount of sediments that can be transferred into a cave
epikarst increase the dimensions of the sinkholes increase system with less open space.
but they are almost always small and isolated.

10.4.4 Very Large Cover Collapse Sinkholes

10.4.2 Intermediate Cover Collapse Sinkholes
Very large cover collapse sinkholes, 100 m in diameter or
Intermediate cover collapse sinkholes are 10–15 m in diam- more, are those that have consumed unusually large quanti-
eter with a material loss of greater than a few 100 up to ties of sediment 100,000–200,000 m3 or more. In many cases
1,000 m3 and occur when the rock is highly fractured. A these large sinkholes have not only consumed large quanti-
tertiary fracture system within the massive rock can accom- ties of sediment, but numerous large objects such as trees,
modate a significant amount of unconsolidated material structures, vehicles and even drill rigs.
only if the fracture apertures are wide, well interconnected Many karst texts, professional papers, and reports discuss
and are not clogged with sediment. In addition, the sediment these very large collapse features. However, their illustra-
moving into the fracture system needs to be in a fluidized tions often show the geologic cross section under a large
state. cover collapse sinkhole with relatively small fissures or voids
Waltham et al. (2005) presents a conceptual model of a limited to the uppermost rock. For example, the initial con-
tertiary fractured rock system. It consists of a massive lime- ceptual model of Winter Park Sinkhole (Fig. 10.2a) showed
stone, with fissures 0.1 m wide spaced every 10 m on each of the original geologic cross section with relatively small fis-
three rectilinear orthogonal fracture systems, and with con- sures tapering to zero with no cavernous system. This sug-
duits 0.6 m in diameter along each fracture intersection. This gested a limited void space into which large amounts of
results in a mean cavernous porosity of only 4.6 %. sediment and other large debris could flow. This sinkhole
Such a fracture system could result in a cover collapse had quite an appetite and consumed two 18 m trees, a house,
sinkhole that could consume a volume of sediment on the part of an Olympic swimming pool, part of a city street, a
order of 100–1,000 m3 or more. The sediment would have to number of vehicles and more than 229,000 m3 of sediments.
92 10 Insight into the Nature of Cover Collapse Sinkholes

Based upon experience with mine backfilling (Table 10.1) Table 10.3 also includes a few examples of pseudokarst
it would be impossible to lose that much sediment and other that are noteworthy and have resulted in extremely large
debris into small narrow fissures. Figure 10.2b shows a sinkholes and losses. For example:
revised conceptual model with a relatively large open cavern
within the limestone that would be necessary to have accom- • Two sinkholes in Guatemala were the result of a sewer
modated all of this unconsolidated material as well as the line of only 2.25 m diameter about 40–50 m below the
large amount of debris. It is the presence of this large cavern, surface. Yet the sediment lost was more than 42,000 m3
that eventually failed and lead to this very large sinkhole. and included five homes, a three-story building and six
In addition, the opening into the rock must be sufficiently people (Hermosilla 2012).
large to enable such large items to pass into the deeper cavity • The three-story crusher plant at a South African Goldmine
without clogging the opening in the rock. In the case of the disappeared down a huge hole (Brink 1984).
Winter Park Sinkhole (Fig. 3.7), the initial opening within • The Jefferson Island salt mine in Louisiana, where the ini-
the uppermost sands and the deeper clay aquitard may have tial opening was a borehole of about 35 cm in diameter. It
been only a meter or so. After the surface collapse occurred then became rapidly enlarged as a very large volume of
and sediment began to move downward, the opening in the water and the abrasion of sediment scour out and enlarge
sands and aquitard are rapidly enlarged by the large volume the opening, which allowed 56,536,000 m3 of sediment, a
of water and debris. Even the size of the opening in the lime- large oil drilling rig, barges and a tug boat to disappear
stone is probably enlarged by erosion. The final size of the (Johnson 2003).
opening at the Winter Park sinkhole was measured at 16 m
(Scott, 2005, personal communication).
Table 10.2 includes examples of very large cover collapse
sinkholes that have already been discussed throughout this 10.4.6 Conceptual Models and Their
book. Included are examples where large amounts of material Limitations
have disappeared into a sinkhole and in some cases large trees,
structures, and vehicles have also disappeared. When very The general features of the five conceptual models are sum-
large amounts of sediment are lost (100,000–200,000 m3 or marized in Table 10.4. These conceptual models provide an
more) along with other large debris, a large cave system along estimate of the relative depth, type and size of open void
with a cavern must exist within the massive rock. There also space within the rock based upon the size and depth of the
must be, or must quickly develop, a sufficiently large opening cover collapse sinkhole along with the volume of
(aven) in the top of rock to allow large quantities of other sediment lost.
debris to flow into the cavernous system. In each of these cases The range of dimensions and volume of sediments in
there were large caverns that had developed over geologic Table 10.4 and Fig. 10.3 are estimates and not intended as
time and collapse was finally triggered by some event. absolute values, there can be a wide range of overlap in the
values. These models are subjective first approximations of
conditions to guide further work and there will be excep-
10.4.5 Extremely Large Cover Collapse tions due to the many variables involved. Only after we have
Sinkholes obtained additional site-specific data from the site charac-
terization effort will we be able to better characterize the
Table 10.3 includes examples of extremely large sinkholes, final nature and magnitude of the open space within the
more than 200 m in diameter that have occurred in both rock.
evaporate karst and certain pseudokarst conditions. However, A plot of the volume of material lost versus estimated
this type of collapse must be treated independently because void space within the rock is shown in Fig. 10.3. In most
of their wide range of variations from typical karst in carbon- cases, the void space will often be much greater than the
ate rock. The extremely large cover collapse sinkholes in material lost. This is supported by the scale models of mine
evaporate karst are often a result of dissolution mining of salt backfilling and actual mine backfill data that shows that a
and brine extraction which resulted in a large cavity within limited amount of material can be placed in an open mine
the salt (Fig. 7.11). Unless very carefully engineered and from a single fillhole. The system then becomes choked off
managed these large cavities have been known to collapse preventing additional material from entering the mine.
resulting in large sinkholes. Examples from west Texas and Sinkholes likely behave similarly where there is a limit on
southeastern New Mexico have resulted in sinkholes with the material that can be moved into the open void space
dimensions of as much as 240 by 186 m and estimated losses before the system is choked off leaving the void space less
of sediment up to 1,300,000 m3. than 100 % filled.
10.4 Conceptual Models of Cover Collapse Sinkholes 93

a
25
Sands

0
Elevation in Meters (MSL)

Clayey Sands

-25

Limestone
-50

-75

original conceptual model

b
25
Sands

0
Clayey Sands
Elevation in Meters (MSL)

-25

Limestone
-50

-75

-100

Note: vertical exaggeration 2:1

revised conceptual model

Fig. 10.2 The original conceptual model of the geologic conditions at more realistic geologic condition with a large cavern at depth that has
the Winter Park sinkhole shows limited void space within the rock (a). sufficient space to accommodate the large volume of sediment and
A revised conceptual cross section of the Winter Park sinkhole shows a other debris (b)
94 10 Insight into the Nature of Cover Collapse Sinkholes

Table 10.2 Examples of very large cover collapse sinkholes

Estimated amount of
Site name Size of sinkhole (m) material lost (m3) Type of material lost Rate of collapse
Winter Park Sinkhole, 100 × 106 m and >229,000 Sediments, a house, large trees, portion Over 36 h
Orlando Florida 30 m deep of a large pool and multiple vehicles
Guest sinkhole west About 90 m dia. >191,000 Sediment, a drill rig, water truck, pipe Rapid (within
central Florida and 90 m deep truck and large trees minutes)
Gypstacks in west 48 m dia 113,200 Phosphogypsum waste Approx. 6 h
central Florida

Table 10.3 Examples of extremely large cover collapse sinkholes

Estimated amount
Site name Size of sinkhole (m) of material lost (m3) Type of material lost Rate of collapse
A sinkhole north of 111 m dia. and 45 m deep 432,400 Sediment Initial collapse in minutes then
Carlsbad, New Mexico expanded for more than a week
Wink Sink 1 in Texas 110 m dia. and 34 m deep 159,000 Sediment
Wink Sink 2 in Texas 240 × 186 m 1,330,000 Sediment
Gold Mining in South Africa 55 m dia. and 30 m deep >500,000 est. A three story crusher plant Very rapid
along with 29 people
Diamond Crystal Salt Mine, Up to 800 m wide 56,536,000 Sediment and water draining Hours
Jefferson Island, Louisiana Lake Peigneur, an oil drilling
rig, barges and a tug boat
Coal Waste Impoundment, Impoundment failure 945,450 Waste coal slurry Rapid
Kentucky 30 ha
Sinkholes in Guatemala City Up to 30 m dia. and 60 m Up to 42,400 Sediment along with Rapid
deep buildings and people

10,000,000
Extremely Large Cover Collapse
Evaporite and Some Pseudokarst
Approximate Amount of Sediment/Material Lost (m3)

1,000,000
Very Large Cover Collapse
Cave System with Large Cavern
100,000

Limit of backfill into an open

mine space as a control point
10,000
Large Cover Collapse (15,000 to 60,000 m3)
Cave System

1,000
Open Void Space in Rock
Intermediate Cover Collapse
Typically > Amount of Sediment Lost
Fracture System
100

10 Small Cover Collapse

Voids within Epikarst

1
1 100 10,000 1,000,000

Estimated Size of Void Space within the Rock (m3)

Fig. 10.3 The relationship between sinkhole size (based upon the volume of material lost) and expected void space in subsurface rock for a cover
collapse sinkhole
References 95

Table 10.4 Summary of cover collapse sinkhole conceptual models

Relative size of the cover
collapse sinkhole Diameter of sinkhole (m) Volume of sediments lost (m3) Nature of the void space within the rock
Small 1–5 m 1–15 Typically associated with isolated voids within
the epikarst or top of tock
Intermediate 10–15 m 100–1,000 Open tertiary fracture system
Large 30–50 m 1,000–50,000 Cave system
Very large 100 m or more 100,000–200,000 or more Large cavern associated with a cave system
Extremely large More than 200 m 100,000 to more than 50,000,000 Evaporite and pseudokarst

Benson RC et al (2000) Stabilization against geologic instability-

General conditions and assumptions include: underground documentation of a mine backfilling case history. In:
ASCE Geo-Denver conference Denver, 5–8 August 2000
• These conceptual models are based upon a single collapse Brink ABA (1984) A brief review of the South African sinkhole prob-
event ignoring changes with time such as hydrologic lem. In: Beck BF (ed) Proceedings of the 1st multidisciplinary con-
ference on sinkholes: their geology, engineering and environmental
transport of sediments. impact, Orlando, Florida. Balkema, Rotterdam, pp 123–127
• Sinkhole size is directly related to sediment thickness Chen J, Beck BF (1989) Quantitative modeling of the cover-collapse
along with the size of the void space within the rock. process. In: Beck BF (ed) Proceedings of the 3rd multidisciplinary
• If there is a surficial aquifer or hydraulic head to fluidize conference on sinkholes and the engineering and environmental
impacts of karst, St Petersburg Beach, Florida. Balkema, Rotterdam,
the sediments they can flow more rapidly and greater pp 89–95
quantities can flow into a given void space. Colaizzi GJ et al (1981) Pumped-slurry backfilling of abandoned coal
• Evaporite karst and pseudokarst sinkholes are unique mine workings for subsidence control at Rock Springs, Wyoming.
cases and must be considered separately. US Bureau of Mines Information Circular 8846
Hermosilla RG (2012) The Guatemala city sinkhole collapses.
Carbonates Evaporates Spec Issue Integr Sci Eng Solve Karst Probl
The data in Fig. 10.3 can be utilized in conjunction with 27(2):103–107
the sinkhole classification (Fig. 3.1), the evolution of karst Johnson KS (2003) Evaporite-karst problems in the United States. In:
(Fig. 3.3) and the models of karst maturity (Fig. 4.1) to Johnson KS, Neal JT (eds) Evaporate karst and engineering/envi-
ronmental problems in the United States, Oklahoma geological sur-
develop a more complete understanding of the potential karst vey, circular 109. Norman, Oklahoma, pp 1–20
or pseudokarst conditions that may exist at a particular site. Waltham AC et al (2005) Sinkholes and subsidence, karst and cavern-
ous rocks in engineering and construction. Springer-Praxis,
Chichester, UK
Whittaker BN, Reddish DJ (1989) Subsidence – occurrence, prediction
References and control, vol 56, Developments in geotechnical engineering.
Elsevier, New York
Aley T (1997) Groundwater tracing in the epikarst. In: Beck B, Worthington SRH et al (2000) Matrix, fracture and channel compo-
Stephenson JB (eds) Proceedings of the 6th multidisciplinary con- nents of storage and flow in a Paleozoic limestone aquifer. In:
ference on sinkholes and the engineering and environmental Sasowsky ID, Wicks CM (eds) Groundwater flow and contami-
impacts of karst, the engineering geology and hydrogeology of nate transport in carbonate aquifers. Balkema, Rotterdam,
karst terranes, Springfield, Missouri. Balkema, Rotterdam, pp 113–128
pp 207–211
 Part II
The Strategy and Methods
for Site Characterization

Part II of this book outlines the site characterization strategies and methods that have been
found to be effective by the authors and their associates for the past 40 years in the fields of
geotechnical and environmental investigation, much of which was in karst and pseudokarst
conditions. The discussion on strategy introduces the concepts, strategies and pitfalls, which
need to be considered to achieve a reasonably complete and accurate site characterization.
They include obtaining appropriate and adequate data followed by a through assessment of all
data. The focus is upon understanding the sites geologic conditions, since the geology ulti-
mately controls all geotechnical aspects as well as groundwater and contaminant flow. An
understanding of the geologic conditions will enable us to make effective, informed decisions
for construction, remediation and the long-term management of the site. To accomplish this
requires a dedicated hands-on team of senior professionals with a range of expertise and famil-
iarization with the many methods available for site characterization.
The discussion on methods includes the many existing and diverse tools that are available
for our toolbox, which can be used for effective site characterization. Numerous examples and
mini-case histories are included to illustrate their applications and successful results. Emphasis
is placed upon simple approaches such as old fashion observations (almost a lost art) as well
as current technology such as a variety of remote sensing, surface and borehole geophysical
methods. However, it is this integrated approach using both traditional methods as well as cur-
rent technologies that enables us to obtain a more complete and accurate site characterization.
While the focus of the book is on karst and pseudokarst, the strategies and methodology used
are equally applicable in all geologic environments for both geotechnical and environmental
site characterization problems.
No single book, including this one, can possibly provide all of the details related to the
methods and their use, necessary to carry out a proper site characterization. The focus of the
methods section is to present a brief introduction to the many off the shelf tools we have avail-
able for site characterization. The details of how to use them are left for many other texts, short
courses and a life long learning experience.
 What Is Site Characterization
11

Abstract
Site characterization is the process of developing an understanding of the geologic, hydro-
logic and engineering properties at the site including the soil, rock, along with groundwater
and in many cases, man-modified conditions in the subsurface (e.g. utilities, structures,
mines and tunnels) that can impact site conditions. It also includes the spatial and temporal
assessment of contaminants when they are present. Various terms such as site investigation,
site assessment and site characterization have been used to describe this process and are
often used interchangeably. Many case histories have shown that most failures are caused
by not properly understanding the site conditions that can impact the project. These failures
could have been avoided by focusing attention on the geologic and hydrologic conditions.

11.1 Introduction ence of existing structures both above and below ground
limiting access, interfering with measurements as well as
Each site characterization is a unique combination of setting, increasing cost.
objectives, logistics, technical issues and non-technical
issues (budget, politics, etc.). Davies (1977) and Fookes In either case, a site characterization is required. If done
(1997) believe that every site characterization in karst must before construction or development you will save money and
be treated as unique and the unexpected should be antici- time and have the opportunity to incorporate findings into a
pated until proven otherwise. quality design. Therefore, you may as well do it right the first
There are generally, two different times when a site char- time (Davies 1977).
acterization is required for a project: Site characterization is done in a number of ways by
different professions including: geotechnical engineers,
• Before a problem has occurred, prior to construction of a geologists and hydrologists. The category of engineering
building, bridge, dam or nuclear power plant, etc. to geologists is a special group who are focused upon the
assess the potential for settlement, collapse, or leakage. process of site characterization. Culshaw et al. (2008)
This is the easiest and most cost effective time to com- describe the history of engineering geology along with the
plete a site characterization since there is usually better leading pioneers in the field. Medley (2009) provides further
site access and problems can be corrected or avoided details on the role of the engineering geologist. Knill (2003)
before construction. discusses the role of the engineering geologist with respect
• After a problem has occurred investigating the cause of to larger projects such as dams where a large team of special-
settlement, collapse or leakage at an existing facility in ists may be involved over a period of several years. Fookes
order to address problems that have already occurred and (1997) suggests that the job of the engineering geologist is to
plan remediation actions. This is usually a more difficult get the geology right. It is this focus on geology that lays the
time to complete a site characterization due to the pres- foundation for the site characterization efforts.

© Springer Science+Business Media Dordrecht 2016 99

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_11
100 11 What Is Site Characterization

Large and
What we may
Infinite
never know

Knowledge
What we can know
Small
Finite
What we know

Fig. 11.1 The nature and limitations of knowledge involved in solving a geotechnical problem. This figure is probably representative of many site
characterization projects

11.2 Uncertainties in Site Characterization geotechnical, environmental and groundwater site character-
ization work is significantly less than 50 % complete and
There are many issues, both technical and non-technical, accurate (Fig. 11.2). Frequently it is as low as 10–15 % and
which can impact the completeness and accuracy of the site rarely ever is the completeness and accuracy at the 75–95 %
characterization process. A fundamental aspect is the inherent level. These findings tend to support the concepts presented
limitations of our knowledge when attempting to solve com- in Fig. 11.1. However, the knowledge and technology is cur-
plex geotechnical problems. Figure 11.1 has been adapted rently available to achieve completeness and accuracy at the
from a presentation by George Sowers, who provided a lec- 75–95 % confidence levels. These levels are only achieved
ture to the Southeast Department of Transportation Engineers with a focused level of effort that includes a sound strategy,
on the topic of site characterization. A key point made, was a team of experienced senior professionals in the field and
that one of the most difficult tasks is to identify what we don’t most important, the owner’s commitment to the project.
know. Recognizing what we have missed in the site character-
ization and what its potential impact to the process is, may be
critical. Sowers went on to point out that unfortunately only a 11.3 The Technical Literature
small percentage of existing knowledge is commonly used
and that most knowledge is acquired by trial and error. A sig- Site characterization is not new. Modern site characterization
nificant part of the site characterization process lies in reduc- began with Carl Terzaghi, the father of soil mechanics, in the
ing uncertainties by understanding the variable conditions early 1930s (Goodman 1999). Hvorslev (1949) compiled an
and by evaluating their consequences to various project deci- early approach for geotechnical investigations. More recent
sions, be they geotechnical or environmental in character. In geotechnical insights are provided by:
order to minimize the impact of karst, the complex interaction
of the soil, rock and groundwater must be understood. Many • Clayton et al. (2005) Site Investigations, 2nd edition
failures in site characterization are related to misunderstand- • Bell (2004) Engineering Geology and Construction
ing of the geology (Sowers 1982). • Hunt (2005) Geotechnical Engineering Investigation
Others have also cited similar philosophies throughout Handbook, 2nd edition
the years. Glossop (1968) suggests that in order to minimize • Hencher (2012) Practical Engineering Geology
uncertainties, what you should be looking for will be dictated
by the geologic setting and project needs. McMahon (1985) Recent approaches for environmental site investigation
points out in his Davis Memorial Lecture, that uncertainty are presented in:
in geotechnical engineering is often large with enormous
economic consequences because the natural subsurface is • Sara (2003) Site Assessment and Remediation Handbook,
often highly variable and its sampling is typically limited. The 2nd edition and
need to remain vigilant and expect surprises throughout a site • Neilsen (2006) Practical Handbook of Environmental Site
characterization is critical. Oreskes and Belitz (2001) also Characterization and Ground Water Monitoring, 2nd edition
stress the need to be aware that there may be many unknowns.
No site investigation will ever be 100 % complete and These six books provide detailed information on a wide
accurate and there are some facts that may never be known range of geologic and hydrologic issues along with technology
about a site. Based upon over 100 years of cumulative expe- for geotechnical and environmental site characterization.
rience, Benson and Sharma (1994, 1995) suggest that most However, their discussion of karst is quite limited.
11.4 Concepts and Strategies for Site Characterization by Others 101

100
Possible Achievable Levels

Increasing completeness, Accuracy

95%

and Confidence Levels

75%

50%

Typical Levels Achieved

(Benson and Sharma 1994 and 1995)

10%
0
Increasing Level of Effort

Fig. 11.2 A hypothetical plot showing how confidence levels will increase with the increasing level of effort as the site characterization
proceeds

There are many excellent books on karst, paleokarst and also evolved. Terzaghi’s Observational Method was intro-
pseudokarst. However, the authors have only found three duced in the 1930s and 1940s when there were few, if any,
books that deal with karst from a practical engineering remote sensing, geophysical or minimally invasive means of
perspective. The authors of each of these books have exten- measurements available. Peck (1969) summarizes the steps
sive experience in karst. of the Observational Method. Terzaghi made observations
before and throughout the project as excavations and boring
• Sowers (1996) Building on Sinkholes, Design and data became available. He predicted the most likely geologic
Construction of Foundations in Karst Terrain, unfortunately scenario and the worse case scenario. As work progressed he
this publication is out of print. was gradually able to converge upon an accurate model of
• Milanovic (2004) Water Resources Engineering in Karst. geologic conditions. All this was accomplished without any
This excellent book deals with the problems of managing of the modern tools we have today. Peck (1969) attributed
water resources within the extreme karst environment of Terzaghi’s great success to his:
Yugoslavia and other areas.
• Waltham et al. (2005) Sinkholes and Subsidence, Karst • unique observational skills;
and Cavernous Rocks in Engineering and Construction. • understanding of geology and its implications to the
specific construction project; and
In addition, the semi-annual Sinkhole Conferences, • insistence upon his presence in the field coupled with his
initiated by Berry Beck beginning in 1984, have provided full, personal responsibility and authority on critical jobs.
extensive information on applied karst issues. These confer-
ences continue today under the direction of the National The observational method still fits today’s site character-
Karst and Cave Research Institute and PE LaMoreaux ization needs and remains one of the most powerful tools in
Associates. The focus of this conference is upon applied our toolbox. However, it is commonly misused with the term
solutions to karst problems. The conference has brought “observation method” cited as little more than a marketing
academics, cavers, engineers, geologists and hydrologists tool (Peck 1969).
together for effective technology transfer. For more informa- Site characterization began as a largely observational
tion on these conferences see www.sinkholeconference.com. science. In the past, observations demanded that many hours
be spent walking the site, taking notes and making sketches.
As technology has advanced, many have relied upon data
11.4 Concepts and Strategies for Site downloaded from the internet, limited site-specific data and
Characterization by Others sometimes unfortunately may not have set foot on the site.
Davies (1977) a karst geologist with USGS provided a
The strategies presented in this book are not new, they have keynote address at the US Army Corps of Engineers Facility
evolved over time as technology has advanced from the at Vicksburg, Mississippi at a Symposium on Detection of
1930s on, and our understanding of karst conditions have Subsurface Cavities. This was probably one of the earliest
102 11 What Is Site Characterization

geotechnical meetings on karst in the United States. Davies investigations. Only then, were the percussion drilling
summarized his experience and his recommended approach and geophysical logging performed along with cross-hole
for dealing with karst. The following is a summary of the key seismic measurements.
points from his lecture. • TVA used a zone approach to karst investigations. Most
karst problems are caused by a system of relatively small
• At the beginning of any project in karst, you are either going joint-controlled and or bedding plane features, rather than
to avoid the potential problems and find a better site, accept a single large cavity. As such the field investigation would
the problems, understand them and correct them before con- focus upon locating and characterizing these zones.
struction, or fix problems later after construction. • TVA had an understanding of the importance of geology
• Most consultants are concerned with the corrective action, and the necessity of a tight working relation between the
more than the initial site characterization problem. The geologist, geophysicist, and the engineer. TVA would
site characterization approach is an entirely different move the team of key personnel to the site and they would
route and it’s a rare route. It costs money but it certainly be dedicated to that site until the site characterization had
does not cost anywhere near as much as unexpected been completed.
corrective action that will ultimately have to be taken
along with political and legal delays. The Corps of Engineers published a document,
• If you are working in karst you must do one of two things “Foundation Considerations in Siting of Nuclear Facilities in
either accept the risks and go forward without a site Karst Terrains and Other Areas Susceptible to Ground
characterization and plan on spending a lot of money, or Collapse”, (Franklin et al. 1981). This document provided an
complete a thorough site investigation and then remedia- overview of the nature of karst and some pseudokarst prob-
tion, which will also cost a lot of money. However, in the lems along with a strategy for site characterization. They
latter case you will have had a chance to understand the provided extensive details of the probability and difficulty of
site conditions in greater detail prior to construction and detecting a solution cavity or channel by borings and con-
then carry out remediation before the fact when the site is cluded that a grid search was more effective than a random
completely open and accessible. search. They were also early proponents of the use of surface
and borehole geophysical methods in the site characteriza-
Other engineers and geologist who have worked in karst tion process.
echo similar ideas. Peter Fookes, a British engineering geol- The following comments were taken from Hatheway
ogist with world-wide experience, much of it in karst, pres- (1996), Perspectives No.21 “Karstic may not be Karst; When
ents an excellent overview of the site characterization is it Safe for a Landfill” a commentary section published in
problems from a practicing engineering geologists viewpoint the Association of Engineering Geologists News.
(Fookes 1997). He covers all geologic settings, emphasizes Karst, by its very definition, is a condition that represents the
the use of geomorphology and discusses some karst issues. potential for the presence of geologic conditions that could
His use of detailed sketches is excellent. This is a must read compromise the integrity and function of engineered works or
for anyone practicing site characterization. In his lecture he environmental facilities. Karst conditions can range from nil to
extreme.
states that the difficult ground conditions found in karst are It is not enough to spot one or two indications of karst in the
often poorly understood by most engineers. Karst conditions regional setting and come to the conclusion that the site itself is
are immensely variable, and require a thorough investigation unsafe. Just to say that an area is “karst”, does not declare it to
by a team that appreciates the complex characteristics of be unsuitable for sitting and operation of engineered works.
Karst features in the area, by themselves, do not indicate that the
karst and at many sites, the details of the karst conditions ground beneath a particular site has been or will be affected by
will only be known when the site has been excavated existing karst conditions or activity at any time in the future.
(Waltham and Fookes 2003). Even if karst features are found on the site itself they should not
The Tennessee Valley Authority (TVA) began its site be cited as a project-stopper rationale. A complete site characteriza-
tion may characterize the existing karst conditions sufficiently
characterization of dams in the 1930s and then nuclear power well so that they may be avoided or properly remediated avoid-
plants in the 1980s with most, if not all, developed in a karst ing impact to engineered works. It is the engineering geologist
setting. TVA’s site characterization strategy (Hopkins 1977) who must carry the message to the owner, but it is the owner who
is a phased approach and included: must make the decision to commit the funds and time sufficient
to characterize the site adequately.

• Phase I of a site investigation consists of a comprehensive Just because a site is in a karst area, or has karst around it
study of regional geologic literature. does not mean that the site itself has karst activity or cannot
• Phase II was a localized site investigation which consisted be used for an engineered purpose. Many dams, water works,
of a site visit followed by detailed geologic mapping, landfills and other critical structures have been built and
remote sensing, tectonic studies, and geophysical successfully operated within karst settings.
11.5 The Site Characterization Team 103

The Expedited Site Characterization (ESC) process ideas of ESC have been used as part of the authors overall
developed by the Department of Energy (DOE) in the 1990s approach and strategy to site characterization for decades.
incorporated many of the early concepts on site characterization The authors have had the opportunity to participate in the
by others. DOE had looked at the progress of Environmental demonstration of the ESC concept at three DOE sites in karst
Protection Agency (EPA) field investigations and had decided and pseudokarst. On these projects, Ames Laboratory teamed
that DOE’s areas of contamination were significantly greater with private consultants to carry out this work.
than that of EPA. As a result they were searching for a better
and more focused strategy for site characterization to more
rapidly converge upon the actual conditions at a site. 11.5 The Site Characterization Team
The ESC strategy emphasizes the use of a variety of meth-
odologies, which will commonly include remote sensing and An effective site characterization team of two to six senior
surface geophysics, followed by minimally invasive push tech- hands-on professionals is usually sufficient for most site
nology. Subsequently, borings and/or trenches and sampling characterization efforts. They must be sensitive to the issues
along with geophysical logging and hole-to-hole measure- of geologic uncertainty and possess the skills, experienced
ments are made with their locations based upon data, rather and persistence to pursue the details. There is no substitute
than guesses. In all cases, the use of on-site experienced hands- for good judgment based upon experience and on-site obser-
on professionals in the field managing the project and making vations. The senior professionals should directly participate
observations and measurements, (as in the Observational in data acquisition and remain involved for the duration of
Method) is absolutely essential. Daily field team meetings to the project in order to provide continuity. This critical asset
review data and change direction and approach, as needed to is often a key to success.
rapidly converge on site conditions. The ESC process is We recognize that a small team of experts, no matter how
flexible and is neither site nor technology specific. experienced will not be able to cover all technical issues.
Some of the unique features of the ESC method include: That is why it is important to know what you and your small
team know and what they don’t know and to fill the gaps
• A small core team of two to six professionals who are where necessary. The selection of members of the site char-
highly experienced, hands-on multidisciplinary geologists acterization team and its supporting members and subcon-
and engineers. tractors is a critical component for a complete and accurate
• The core team members are on-site and are directly site characterization effort.
involved in the daily field observations, data acquisition, In addition, the site characterization team must be flexible
processing and interpretation. and be able to change direction and activities at any time
• Emphasis is placed upon understanding of the site geology. as new data is obtained, if necessary. All field operations
• The work plan is dynamic and can be modified by the (including all decisions on changes), data assessment, analysis
field team on-site, as needed. and reporting must be under the control of the core team of
• A broad variety of measurements and technologies are used. senior professionals who are in the field.
• A preliminary conceptual model of site conditions is If the site characterization effort is compromised due to
developed early and is tested and improved as new data time, costs or capability; the result are almost certain to have
becomes available. a much higher risk of failure and negative impact to future
• The use of on-site laboratories or a minimum turn around efforts. Failure to effectively complete an early site characteriza-
time from an off-site lab for geotechnical or chemical tion will almost always results in greater risks, remediation
analysis is used; costs, schedule delays, along with lawsuits and other needless
• Data from the various activities are analyzed daily. Each day expenses. You will pay for a site characterization eventually
the team meets to discuss findings and modifies the site char- whether you want to or not. The most economical approach
acterization program as needed to optimize data acquisition. by far is to carry out an effective site characterization early
• A unified database/graphics system is used to manage before design and construction have begun, as part of the site
data and present results to a common scale with all data selection process (Davies 1977). By doing so, both time and
points shown; costs of the long-term project can often be easily reduced by
• The same team is involved throughout the site character- factors of 2–5 or more.
ization process from beginning to end. Figure 11.3 shows the geologic ESC team who were work-
ing at the Faultless site (an underground nuclear test site) in the
A more detailed discussion of the ESC process can be Central Nevada Test Area. The team is standing at the surface
found in (Beam et al. 1997; Benson 1997; Benson et al. casing for the 975 m deep boring used to deploy the test bomb.
1998; Yuhr 1998) and ASTM’s Standard Practice for After detonation the ground subsided about 2.5 m. The focus of
Expedited Site Characterization (ASTM 1998). The key the ESC effort was to investigate possible contamination
104 11 What Is Site Characterization

Fig. 11.3 The geologic ESC team at Faultless Site (an underground nuclear test site) in the Central Nevada Test Area. The 2.4 m diameter casing
was used to lower the bomb

associated with drilling mud pits and other waste materials. • A Lack of Understanding or Appreciating the Impact of
The work was carried out for DOE under the direction of Ames Geology.
Laboratory with Team members consisting of: • Politics Outweighing Science.
• Distrust Between Parties
• Technos, who was responsible for location and character-
ization of mud pits and other waste as well as geologic The problem of inadequate site characterization and lack
characterization of the site of attention to geologic conditions is also common in geo-
• McLaren-Hart, who was responsible for chemical analy- technical work as highlighted by Shuirman and Slosson
sis utilizing on-site mobile chemical laboratories. (1992) in their book Forensic Engineering. They cite five
areas of concern in the state of the practice in geotechnical
site characterization:

11.6 Some Pitfalls of Site Characterization • Frivolous Lawsuits:

• Incompetent Professional Work:
There are a wide variety of pitfalls that can impact the site • Outdated Codes and Regulations:
characterization process. They can be divided into technical • Educational Shortcoming:
and non-technical issues. The following is a brief summary • Most geotechnical related accidents, disasters, and failures
of some of these significant pitfalls and reasons that prevent are not “acts of God”. In almost every instance, the impact
adequate site characterization from taking place (Benson and event was recognizable, predictable and preventable if
Sharma 1994, 1995). The examples cited cover a wide range only good professional judgment (and good site charac-
of issues related to geotechnical and environmental projects. terization) had been used.
While many of the comments do not apply directly to karst,
they do apply to the general process of site characterization. More recent comments regarding pitfalls were cited in
Throughout the 1990s there has been a notable increase in three articles in Geo-Strata by:
the number of papers, articles, and editorials on ethics and
the degradation of technical work within environmental and • Mitchell and Kavazanjian (2007) “Geoengineering
engineering practices. The editorial “What has Gone Wrong” Engineering for the 21st Century”
by Freeze and Cherry (1989) presented an assessment of the • DiMaggio (2007) “Commentary: Advancements and
environmental industry and identified problems with the Disappointments in GeoEngineering and Geo-
process after a decade of experience including: Construction – A 30 Year Reflection”, and
11.6 Some Pitfalls of Site Characterization 105

• Matheson (2009) “Engineering Geology Past, Present and • The vast amount of data that is available often provides a
Future-This time it’s Different”. false sense of confidence in a project.
• Many professionals are becoming too reliant upon tech-
They suggest that two major technical pitfalls that have nology rather than judgment gained through experience.
impacted site characterization are the lack of an interdisciplin- • Engineering geology remains in part an observational
ary approach and the over reliance on the use of computers. science. There is no substitute for first-hand observations
and on-site experience. Analysis by computer alone cannot
replace on-site observations and experience.
11.6.1 Lack of Interdisciplinary Approach
There are also many non-technical pitfalls that regularly
Applying an interdisciplinary approach to site characteriza- impact site characterization efforts. Some of the many non-
tion seems like such a logical approach. However, the push technical pitfalls include:
for developing narrow specialties within a given field is
stronger than ever within our educational systems as well as • Cost, quality control and project management aspects
in practice. As such, there is a lack of being able to visualize • Fragmenting of a project into small segments, often over
the broader picture of site characterization including the gaps of time, and
integrated aspects of geology, hydrology and engineering. • Lack of senior experienced professional involvement.
This lack of understanding and communication occurs
between disciplines as well as within each profession view- Cost, quality control, and project management aspects
ing the problem with limited insight from narrow specialties. have occasionally become more important than the real tech-
Clearly, greater emphasis is needed on interdisciplinary stud- nical mission. All too often projects are being administered
ies. Most geotechnical engineers have not studied enough by personnel with little or no understanding of the technical
applied geology and geologists have not studied enough issues, with little emphasis on qualifications and experience
applied geotechnical engineering. of the consultant being retained. In addition low-bid and fast-
track contracts have been promoted in the past few decades
which tend to focus upon cost and time as opposed to getting
11.6.2 Impact of Computers things done right.
Fragmenting of a project is not unusual, with a project
Advancements in technology, computers and software is broken into small segments spread over months and years by
moving at a rapid pace. One of the impacts of the digital age the owners. Multiple consultants and contractors are com-
is that we have the ability to gather and manage vast amounts monly involved and there are often personnel changes with
of data very quickly. While advancements in technology, both the owner and the consultants. When this happens there
computers and software have obvious benefits, we also have is a lack of continuity in the project and understanding site
to recognize the disadvantages. conditions is never complete. Critical information is easily
Since the launch of “Sputnik” in 1957, there has been a shift lost, and quality suffers.
from the more practical hands-on aspects of engineering, Lack of direct involvement of senior experienced profes-
hydrology, and geology towards a more advanced theory and sionals. Firms commonly will list the most senior staff on the
use of computer programs to solve problems. Many geologists proposals and they show up at the presentations and meet-
now graduate without having the advantage of a field course and ings but are rarely ever involved in the project once awarded.
geotechnical engineers without having seen a drill rig in opera- The junior staff is then sent to the field, reserving dollars for
tion. The lack of these hands-on experiences ultimately impacts billable time of the senior staff in the office. Senior experienced
their understanding of the site characterization process. professionals in the field are almost unheard of in today’s
cost conscious world.
• There is a need for better geotechnical databases, models There are a wide variety of technical and non-technical
and information systems to handle the large data streams. pitfalls, which can significantly impact site characterization
• Another impact of this rapid change in technology is the process. Most of these limitations and pitfalls have been rec-
lack of critical thinking and practical problem solving ognized over the decades, yet they continue to persist. It
abilities by professionals. requires an awareness and commitment by both the site
• Computer software often lacks reality checks resulting in over- owner and the consultant to overcome these many limitations
confidence that the computer always yields the right answer. and pitfalls.
106 11 What Is Site Characterization

References Hatheway A (1996) “Karstic” may not be karst; when is it “safe” for a
landfill? Assoc Eng Geol News, Perspectives No.21, 39(2):103–110
Hencher S (2012) Practical engineering geology. Spoon Press, London
American Society for Testing and Materials (1998) Standard practice
Hopkins R (1977) TVA philosophy and methods for site investigation.
for expedited site characterization of vadose and ground water
In: Butler D (ed) Symposium on cavity detection of subsurface cavi-
contamination at hazardous waste contaminated sites, ASTM D
ties. US Army Waterways Experiment Station, Vicksburg, pp 25–26
6235-98a. ASTM International, West Conshohocken
Hunt RE (2005) Geotechnical engineering investigation handbook, 2nd
Beam P et al (1997) Lessons learned, a brief history of site characteriza-
edn. CRC Press, Boca Raton
tion. HazWaste World/Superfund XVIII, Washington, DC, 2–4
Hvorslev MJ (1949) Subsurface exploration and sampling of soils for
December 1997
civil engineering purposes. US Army Waterways Experiment
Bell FG (2004) Engineering geology and construction. Spoon Press,
Station, Vicksburg
London
Knill J (2003) Core values: the first Hans-Cloos lecture. Bull Eng Geol
Benson RC (1997) ESC: just another new acronym or maybe some-
Environ 62:1–34
thing meaningful. Environ Eng Geosci 3(3, Fall):453–456
Benson RC, Sharma D (1994) The pitfalls in site characterization and Matheson GM (2009) Commentary: engineering geology past, present
how to avoid them. Short course presented by Environmental and future-this time it’s different! Geo-Strata March/April, pp 14–16
Education Enterprises, Salt Lake City Utah McMahon BK (1985) Geotechnical design in the face of uncertainty. J
Benson RC, Sharma D (1995) Pitfalls of site characterization and how Aust Geomech Soc 10:7–19
to avoid them. Short course presented by Environmental Education Medley EM (2009) Commentary: engineering geology, a vital phase of
Enterprises, Charlotte, NC, November geoengineering. ASCE Geo-Inst Geo-Strata 10(2):10–12
Benson RC et al (1998) ESC: how it differs from current state of the Milanovic PT (2004) Water resources engineering in karst. CRC Press,
practice. In: Symposium on the application of geophysics to Boca Raton
environmental and engineering problems, EEGS, Denver, CO, Mitchell JK, Kavazanjian E Jr (2007) Geoengineering for the 21st cen-
pp 531–540 tury, vol 8(4). ASCE Geo-Institute, Geo-Strata, Reston, pp 14–18
Clayton CRI et al (2005) Site investigation. Available on line at www. Neilsen DM (ed) (2006) Practical handbook of environmental site char-
geotechnique.info acterization and groundwater monitoring, 2nd edn. CRC Press,
Culshaw MG et al (2008) Two hundred years of engineering geology. Boca Raton
Geol Soc Lond Q J Eng Geol Hydrol 41(2):137–142 Oreskes N, Belitz K (2001) Philosophical issues in model assessments.
Davies WE (1977) Keynote presentation to the symposium on detection In: Anderson MG, Bates PD (eds) Model validation: perspectives in
of subsurface cavities. In: Butler DK (ed) US Army Engineer hydrological science. Wiley, London, pp 23–41
Waterways Experiment Station Vicksburg, Mississippi. (lecture Peck RB (1969) Advantages and limitations of the observational
recorded by Benson RC) method in applied soil mechanics. Geotechnique 19(2):171–187
DiMaggio JA (2007) Commentary: advancements and disappointments Sara MN (2003) Site assessment and remediation handbook, 2nd edn.
in geo-engineering and geo-construction, a 30–year reflection. Geo- Lewis Publishers, New York
Strata July/August, pp 12, 13, 26 Shuirman G, Slosson JE (1992) Forensic engineering, environmental case
Fookes PG (1997) Geology for engineers: the geologic model, prediction histories for civil engineers and geologists. Academic, San Diego
and performance. Geol Soc Lond Q J Eng Geol 30(4):293–431 Sowers G (1982) Lecture presentation to the southeast department of
Franklin AG et al (1981) Foundation considerations in sitting of nuclear transportation engineers. Recorded by RC Benson
facilities in karst terrains and other areas susceptible to ground Sowers GF (1996) Building on sinkholes, design and construction of
collapse. Geotechnical Laboratory, U.S. Army Engineers Waterways foundations in karst terrain. ASCE Press, Reston (out of print)
Experiment Station NUREG/CR- 2062, Vicksburg Waltham AC, Fookes PG (2003) Engineering classification of karst
Freeze A, Cherry J (1989) What has gone wrong, a guest editorial. ground conditions. Geol Soc Lond Q J Eng Geol Hydrol
Groundwater 27(4):458–464 36:101–118
Glossop R (1968) The rise of geotechnology and its influence on Waltham AC et al (2005) Sinkholes and subsidence, karst and cavernous
engineering practice. The Institution of Civil Engineers, London. rocks in engineering and construction. Springer-Praxis, Chichester
Geotechnique 18(2):107–150 Yuhr L (1998) Managing data and its integrity in the ESC process, symposium
Goodman RE (1999) Karl Terzaghi, the engineer as artist. ASCE Press, on the application of geophysics to environmental and engineering
Reston problems. EEGS, Chicago
 The Strategy
12

Abstract
The objective of the site characterization process is to develop an understanding of the
geologic conditions (both regional and site-specific), which is sufficiently complete and
accurate so that a conceptual geologic model of the subsurface can be created with reason-
able confidence level for the intended project needs. The difficulty is one of solving the
three-dimensional geologic/hydrologic puzzle with many variables and unknowns. The
success of a site characterization effort lies in the ability to locate, describe and quantify the
natural geological, and hydrological heterogeneity at a site. Heterogeneity is a natural result
of the processes that create and modify the geologic setting, and are sometimes further
modified by man’s activities. The presence of karst typically complicates the process by
introducing a higher level of heterogeneity in the three-dimensional geologic puzzle. It can
often be the fatal flaw for a project.

Site characterization must focus upon obtaining a thorough conditions due to their limited spatial extent, which represent
and complete understanding of geologic conditions that will less than 1–5 % of the subsurface area. As a result traditional
impact the site. While uncertainties, assumptions and opin- approaches to investigation, using borings, have a very low
ions are a part of any site characterization (Fig. 12.1a), they probability of intersecting these features. Beginning in the
should be minimized. Interpretations must be supported in a 1950s the development of geophysical instrumentation and
logical and obvious way by sufficient data, which have been its application to characterization of shallow subsurface con-
tested and proven to be correct. A solid base of data enables ditions began. Since then digital technology has resulted in a
us to carry out subsequent efforts such as modeling, risk greatly expended group of geophysical tools for our site
assessment, construction, or remediation with greater confi- characterization tool box (Olhoeft 1994). This has provided
dence and accuracy while minimizing uncertainties, assump- us with three basic approaches that can be used to improve
tions and opinions (Fig. 12.1b) (Benson et al. 1996; Yuhr our ability to detect and evaluate subsurface conditions and
et al. 1996). geologic variability. They are:

• Direct Detection
12.1 The Detection Dilemma • Indirect Detection
• Statistical Assessments
One of the biggest problems with site characterization efforts
has been our ability to effectively locate and map subsurface In most site characterization efforts all three approaches
geologic conditions especially localized anomalous condi- will be employed to effectively converge upon a complete
tions. This is particularly true in fractured rock and karst and accurate conceptual model of site conditions.

© Springer Science+Business Media Dordrecht 2016 107

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_12
108 12 The Strategy

Uncertainties,
Assumptions
& Opinions
Uncertainties,
Assumptions
& Opinions
Data
Data

Appropriate, Adequate and

Accurate Data along with a detail
assessment of all available data

Fig. 12.1 Minimizing uncertainties, assumptions and opinions in site characterization by increasing the amount of use full data

Area of Site = As

Area of Target = At

As/At = 10

Probability of Detection As/At = 10 As/At = 100 As/At = 1000

100 16 160 1600
98 13 130 1300
90 10 100 1000
75 8 80 800
50 5 50 500
40 4 4
40 400

Fig. 12.2 The probability of detecting an isolated target using borings based upon the site to target area ratio. For a given probability of detection,
the number of borings increases if the site to target ratio increases

12.1.1 Direct Detection certainly should not be applied when assessing complex
geologic conditions.
Direct detection of geologic conditions can be accomplished Benson et al. ( 1984) has shown that boring densities
in a number of ways such as by observations at outcrops, are commonly inadequate to detect geologic anomalies.
conventional drilling and sampling, or trenching. This pro- For example, Fig. 12.2 shows a target that could be a
vides a very high level of confidence in the detailed data cavity or buried sinkhole, buried debris, or a contami-
from a local area or the area immediately within the borehole nant plume. The table provides the number of borings
or the trench walls. This type of detection is an integral part necessary to simply detect the target for different site to
of the site characterization process, but needs to be optimized target ratios. If the site to target ratio is 10:1 then ten
in terms of the location. regularly spaced borings will be required to provide a
Most subsurface geologic assessments are based upon detection probability of 90 %. Note that these statistics
vertical boreholes, commonly made along a survey line or only apply to the detection of the target by one boring
over a grid. In feasibility and preliminary studies, spacing and are not sufficient to define the shape or depth of the
may range from 30 to 90 m or more. Final study programs, target, which would require many additional borings.
depending upon the project type, a borehole spacing of Smaller targets, such as widely spaced joints or frac-
15–30 m is commonly recommended (Hunt 2005). While the tures, would create a much larger site to target ratio and
use of boreholes with this type of spacing is a general guide- require a 100–1,000 borings to achieve a 90 % probabil-
line, in practice, it all too often is treated as a standard and ity of detection.
12.2 Appropriate, Adequate and Accurate Data 109

Quinlan (personal communication, 1989), has indicated our attention on anomalous conditions and locating the direct
that the area of caves within Mammoth Cave system underlie sampling methods.
less than 5 % of the surface area encompassing the overall
boundaries of the cave system. Worthington (1999) and
Worthington et al. (2000) have compiled data on the extent 12.1.3 Statistical Approach
of some mapped caves. His data suggests that cave passages
occupy 3 % or less of the bedrock in which they are located. Statistical data and trends can be used to predict the risk of
Such low percentages of occurrence makes a subsurface subsidence and collapse of sinkholes. For example, a sink-
investigation using a limited number of boreholes like “look- hole database or a cave database and cave maps can be uti-
ing for a needle in a haystack” and almost assures failure. lized to provide a statistical trends for an area in which data
Not only would it be cost prohibitive to use a large number is available. This type of data may be used to calculate fre-
of borings to accurately resolve conditions, but it would quency of occurrence of sinkholes within an area. Wilson
reduce the site to swiss-cheese. If vertical fractures are tar- (1995) has utilized this approach in Florida to define the
geted by vertical borings the probabilities of detection are number of new sinkholes per year per unit area (NSH). This
even worse. See Franklin et al. (1981) for further details on approach can provide reasonable confidence levels as to
the probability of detection. what conditions might be expected over a large area. But
A more effective way to utilize the direct detection meth- such methods are not intended to provide reliable site-
ods is needed so they can be used in a focused manner on the specific data.
critical anomalous areas. This can be accomplished by the All three approaches (direct and indirect detection along
use of surface geophysical methods (indirect detection) and with statistical methods) will commonly be used to evaluate
statistical methods to aid in the detection of anomalous areas subsurface conditions. Each has its advantages and limita-
so that the direct methods can be located based upon solid tions. Adequate spatial density and coverage over an area is
data rather than rules of thumb or guesses. the key to any geologic site characterization to adequately
define the background and anomalous subsurface conditions.
Once the background and anomalous subsurface conditions
12.1.2 Indirect Detection have been accurately located, our direct sampling methods
(direct push methods, borings and trenches) can then be
Indirect detection is accomplished in a variety of ways located with confidence (rather than guesses) to further
using simple observation, remote sensing and surface geo- improve our understanding of subsurface conditions.
physical methods to obtain indirect evidence of subsurface
conditions. Indirect detection may be simple observations
of a variety of indicators present at the surface in the form 12.2 Appropriate, Adequate
of surface cracks, water puddles and flow, vegetation and Accurate Data
stress, or regional photo-lineaments (linear trends seen in
aerial photos). For example, when we use aerial photos for At this point, it should be clear that a reasonably accurate
geologic interpretation, the photo only reveals the condi- and complete site characterization is all about data. In
tions at the surface. However, we often interpret aerial order to have a sound understanding of the subsurface
photos to provide subsurface information. A photo-linea- conditions, we have to have data. In order to minimize
ment may be an indication of possible fracture zone, which uncertainties and assumptions we have to have data. But
may be highly weathered with water seeping downward. A it’s not just any data. One must have appropriate data,
linear trend of sinkholes seen in aerial photos may indicate adequate data and accurate data followed by a thorough
the presence of a cave system between the sinkholes at assessment of all data. Determining what data are appro-
depth. priate, how much data are adequate along with assuring
We often observe evidence of deeper-seated karst activity accurate data, are some of the most important steps in car-
within the shallow soil and rock due to the presence of Near rying out site characterization efforts (Benson et al. 1996;
Surface Indicators (NSI). These NSI may be present at the Yuhr et al. 1996). All data associated with any site charac-
surface such as cracks, or just below the surface (within the terization should meet these three criteria. The less one
upper couple of meters) and show up in some of the shallow relies on assumption and opinions (Fig. 12.1a), the more
geophysical data. For example, ground penetrating radar confident we can become in our site characterization
data may show dipping strata that would be a NSI of possible because of the solid base of appropriate, adequate and
deeper activity. These NSI can provide an initial insight to accurate data which we are basing our interpretations on
the subsurface and become an effective means of focusing (Fig. 12.1b).
110 12 The Strategy

12.2.1 Appropriate Data for our conceptual model. This concept is stressed by Fookes
et al. (2000) who point out that “The present ground condi-
There is an extremely wide range of technology that is avail- tions, at any site, including the soil and bedrock, are a result
able in our tool box for site characterization. All of these of the total geological and geomorphological history of the
techniques have their advantages and limitations. It is critical site, including the stratigraphy, the structure, the former and
to understand the limitations of each of the many technolo- current geomorphological processes, the past and present cli-
gies so that the most effective tools can be selected. The matic conditions over the geologic time scale. It is this total
selection will be site-specific and based upon the project geological history that is responsible for the mass and mate-
needs so that the resulting data is appropriate to the problem rial characteristics of the ground today”.
to be resolved. The following are some concepts for select- While this strategy of working from the regional to local
ing appropriate methods. investigation is not new, it is seldom followed in practice. All
too often the process is reversed and the detailed work of
12.2.1.1 Keep It Simple, Focus Upon Basics drilling, sampling and laboratory analysis occurs early in the
So often enthusiasm for some new technology by someone site characterization process, often omitting the regional set-
who has seen a demo or heard a paper drives the project. We ting and its geomorphic impact to the sites history. Hoek and
are not adverse to technological advancements, but when a Bray (1977) state that the review rock cores are often the
simple measurement or an approach can provide the needed starting point in a geotechnical investigation. While this is
information cost-effectively then it may be the most appro- critical information, they suggest that the regional frame-
priate. We recommend following the KISS principle (keep it work of the geological environment should be developed
simple stupid) and always work from the simple to the first so that subsequent information can be placed into
complex. perspective.
The scale of observations and measurements can range
12.2.1.2 The Sequence of Work from the regional setting of many square kilometers to the
In practice, we should approach a site characterization by most detailed site-specific sampling and measurements of
first understanding the regional setting and its geomorphol- soil samples or rock cores (with dimensions of <10 cm2)
ogy along with its implications to the site itself and then (Fig. 12.4). Methods such as satellite data are best applied to
focus upon the site-specific details (Fig. 12.3). The data regional coverage while aerial photography can be applied to
acquired to assess each of these settings (regional, local and regional, local and site-specific coverage. The surface
site-specific) needs to be appropriately scaled. In this way, geophysical measurements are usually limited to local or
our detailed, site-specific data and interpretations can be site-specific coverage. The invasive measurements such as
based upon and placed into the context of the regional geo- borings are usually limited to site-specific use. In some
logic setting and its geomorphology, providing a solid basis cases, limited spatial data from geologic mapping of outcrops

Regional Setting (many km2)

Local to Site Setting (1 to a few km2)

Boring

Boring

Detailed Site Data

Core
Sample

Fig. 12.3 The sequential approach to site characterization should start with the regional setting and work toward the local setting and then the
site-specific details
12.2 Appropriate, Adequate and Accurate Data 111

Local The Localized Sediment

Regional Setting Site Within the or Core
site Samples
Method Many km2 A few About 1 <10 m2 <10 cm3
km2 km2
Airborne/Satellite Measurements
Limited
Aerial Photos

Geologic Mapping/Observations
Surface Geophysics

Invasive Measurements
Geophysical Logging
Engineering Measurements
Hydrologic Measurements

Dye Tracing

Fig. 12.4 Typical methods used for site characterization and their range of scale

Fig. 12.5 The evaluation of fractures can encompass a wide range of surface geophysical data. (d) Site-specific – roadcut. (e) Site-specific
methods and scale of measurements from regional to site-specific data. detailed – cores
(a) Regional data – Landsat. (b) Local aerial photos. (c) Site-specific
112 12 The Strategy

or borings can be expanded over larger areas by interpolation Core Samples

and extrapolation. Each method has its own advantages and

Parameter or Property
limitations, which must be understood in order to select the

Being Measured
appropriate method for the scale of data needed.
Large Block or
Figure 12.5 illustrates various observations and measure- Outcrop
ments that can be used when characterizing lineaments,
joints and fractures proceeding from the regional to the site REV
itself and finally to the more site-specific details.

• A project might start with a regional lineament assess-

Sample Size
ment based upon Landsat imagery. Figure 12.5a is an
example covering more than 82,000 km2. Fig. 12.6 As the volume of the sample increases within the same rock
• Aerial photos can also be used for lineament analysis mass the value for the parameter being measured changes. A volume
Fig. 12.5b. This example covers about 10 km2. The can be reached, beyond which the test results become independent, that
volume is the REV. In this example, we have shown the REV to be a
obvious N10oW joint pattern was observed at both a large block of rock or rock outcrops
regional and local setting.
• A lineament assessment was made over 4 ha using surface
geophysical measurements (electromagnetics) to map
joints in shale (Fig. 12.5c). These data provide detailed
coverage over the site itself. Dye Tracing

Pumping Tests
• Geologic mapping along road cuts on or surrounding a Hydraulic Conductivity Tests

Slug Tests
site, expose joints and fractures, over 10’s–100’s of meters

Packer Tests
(Fig. 12.5d). Sampling is now at the localized site-specific
level.
• A core sample (Fig. 12.5e) provides data on bedrock frac-
Core Tests

tures at a very detailed level (7–10 cm diameter sample).

However, we cannot assume that they are laterally exten-
sive or are hydraulically interconnected over the scale of
the site itself unless verified by other data.

12.2.1.3 The Impact of Scale Scale of Measurement

It is important to understand the concept of scale effect
because it impacts the results of our measurements and Fig. 12.7 Hydrologic properties also vary with the size or volume
affects the appropriateness of data to the site characterization being sampled
effort. Results from various mechanical and hydrologic tests
will vary based upon the area selected for the test and the that property (Fig. 12.6). Nelson (1984) has suggested a rule
volume tested. The mechanical properties of the rock mass of thumb that the volume of the sample should be at least ten
includes fractures, their orientation, spacing, aperture which times that of the fracture spacing in order to have it be repre-
vary with scale along with deformability, strength, and inter- sentative. Such rules of thumb are reasonable when no other
nal stress. Hydrologic properties of the rock mass will also data is available. The hydrologic properties of a rock mass
vary with the size of the sample. For example, a slug test also vary with scale as illustrated by Fig. 12.7. The many
measures the hydraulic conductivity of the volume near the issues of scale effects are discussed in Da Cunha (1990b).
well screen, a packer tests measures a larger volume of a When determining appropriate data, it is recommended that:
meter or so around the well and a pump test will measure a
volume between wells which may be several meters or more. • The sequence of work is from the regional toward
The relationship between the test results and volume mea- site-specific,
sured is called the “scale effect”. • Simpler measurements or methods should be used when
DaCunha (1990a) introduced the concept of Representative and where appropriate,
Elementary Volume (REV). He shows that for the same rock • The REV for various properties and their scale effects
mass, a sampling volume can be reached where the test needs to be considered and,
results become independent of the sampling size. The REV is • The advantages and limitations of each measurement
the smallest volume that can be considered representative for should be understood.
12.2 Appropriate, Adequate and Accurate Data 113

a b

uniform grid- dense station measurements

station measurements along selected profile lines
c d

continuous measure- station or continuous

ments along uniformly measurements along profile
spaced profiles lines where access allows

Fig. 12.8 The survey grid used for data acquisition may be a uniform ous” measurements over parallel survey lines (c), or the survey may
grid of station measurements (point data) over the site (a), a series of need to accommodate existing structures and incorporate both station
parallel survey lines using dense station measurements (b), “continu- and continuous measurements, where accessible (d)

12.2.2 Adequate Data Density background conditions can then be effectively located based
upon data rather than guesses.
While a very high data density providing 100 % coverage Adequate data is typically thought of as having good lat-
would be optimum, it is not usually possible based upon both eral coverage over a site. However a site characterization
technical and economic reasons. Therefore, once appropriate includes all three dimensions as well as changes over time.
technologies are selected, the next critical decision is how Therefore, adequate data density must also address sampling
can they be used to provide the best spatial sampling. with depth and over time. These factors are particularly criti-
Adequate data simply means that there should be suffi- cal where hydrologic conditions play a role.
cient spatial sampling to reduce the uncertainty of variations
in geologic and hydrologic conditions. Insufficient spatial 12.2.2.1 Lateral Data Density
sampling can easily miss anomalous conditions leading to an Lateral data density can be accomplished by using tightly
incomplete and erroneous understanding of site conditions. spaced measurements on a grid or “continuous” (or nearly
This is where the use of some remote sensing or surface continuous) measurements on closed spaced parallel lines.
geophysical method will greatly benefit a project because of Todays digital instruments have fast sample rates, such that
their ability to dramatically increase spatial data density. the data is almost continuous (see Chap. 16 for further dis-
Satellite or aircraft measurements and imaging provide cov- cussion). This approach is used when there is no specific tar-
erage over large areas (regional conditions) and in some get of interest and the objective is to determine the presence
applications local data. Surface geophysical measurements of variable or anomalous geologic conditions. Figure 12.8
can be used to increase data density along survey lines or illustrates various approaches to optimizing lateral data den-
over a grid on a site-specific basis. If adequate data density is sity whether using station or continuous measurements. The
obtained then geologic variability as well as anomalous and reality is that there may be site access restrictions (steep
114 12 The Strategy

Fig. 12.9 The size and shape of the target will influence the approach to data acquisition and the amount of data required to be adequate. (a) Planar
features. (b) Linear features. (c) Discrete features

Fig. 12.10 Surface geophysical electromagnetic conductivity data (EM31) from a school yard near Love Canal (1980) providing essentially
100 % site coverage. Boring locations are shown by dots

topography, buildings or fencing) that require a more • If the target is linear, then the survey lines should be
creative, mixed approach. However, data density can also be oriented to cross as close to perpendicular as possible to
optimized if something is known about the site or a particular the target and should be spaced close enough to provide
target of interest. For example, the target may be: adequate definition of the target.
• If the target is a discrete feature, then the expected size of
• a generally planar surface such as a clay layer or top of the target may impact the spacing of the measurements to
rock, improve the probability of detection (Fig. 12.2).
• a linear target such a buried channel in the bedrock, fault
or fracture zone; or Once an anomalous feature is detected, it is often neces-
• a discrete feature such as an enlarged cavern or buried sary to add additional data for a more detailed characteriza-
paleosinkhole. tion of the shape, boundaries and depth.
While not an example of karst, Fig. 12.10 is a classic
Given the nature of the target one can optimize a given example of a site, which clearly shows the impact of inade-
number of measurements or survey lines (Fig. 12.9). For quate spatial sampling by borings (direct sampling) alone
example: and the benefits of indirect measurements to locate anoma-
lous conditions. There had been some reports of contamina-
• If the target is known to be planar, fewer grid points or tion in a schoolyard near Love Canal (Niagara Falls,
points along a survey line may be needed. New York). Five borings had been located, one in each corner
12.2 Appropriate, Adequate and Accurate Data 115

and one in the center of the site. A sixth boring was place the appropriate depth of interest for the project needs.
over an area of dead grass. These boring locations were a Figure 12.12 is an example from a site characterization for a
reasonable approach given no other information, however proposed hazardous waste facility in southwest Texas. A
they were simply a best guess and none of the six borings wide range of measurements with a wide range of depths
encountered any contaminated material. was used on that project.
Surface geophysical measurements were continuously
acquired using electromagnetic (EM) measurements. The 12.2.2.3 Data Density Over Time (Temporal
EM instrument used was an EM31 measuring to an approxi- Data)
mate depth of 6 m. Continuous data was acquired along sur- Temporal adequacy is concerned with sufficient data points
vey lines spaced 3 m apart. This provided virtually 100 % to avoid gaps in the data or changes in conditions over time.
coverage of the site. The resulting data in Fig. 12.10 is a If natural periodic events, such as seasonal water levels,
three dimensional perspective of the electrical conductivity spring or stream flow, are to be represented, the data must
values which clearly show the spatial extent and magnitude have an adequate sampling over the time to document these
of a high conductivity area. One new boring was located at changes. Where the water level changes slowly with time
the center of the anomalous area and indicated the presence (e.g. a decline over periods of months due to seasonal
of buried fly ash, which is an electrically conductive con- changes) fewer water level measurements may be needed
taminant. This example clearly illustrates the benefit of such than where water levels change more rapidly (e.g. in response
indirect measurements to locate direct measurements. to periodic storm events). In either case, it is important that
these temporal changes are adequately represented by the
12.2.2.2 Data Density Versus Depth measurements.
Data density versus depth is often overlooked, yet, applies to An example of temporal changes at a site, include a
all of our measurements. For example, we often encounter mountain stream that was but a trickle. One could easily
boring locations on a map. At first glance it might appear that cross it without getting your feet wet and we had done so
there is adequate spatial density. However, if we plot the daily over a 2 week period (Fig. 12.13a). Then one day, a
depth of borings we often find few, if any, borings extend to heavy rain in the mountains turned the stream into a raging
sufficient depths to meet all project objectives. torrent, which could easily sweep you off of your feet
For example, a low-level radioactive waste site in St. Louis (Fig. 12.13b). Such rapid events in stream or spring flow are
in a karst setting had 80 borings over 67 ha (Fig. 12.11a). critical in evaluating hydrology and contaminant transport.
Initially, this appears to be adequate spatial coverage across Such conditions can only be monitored by electronic water
the site. However, the borings and wells were installed by level gauge since such events occur infrequently and often
different contractors at various phases of investigation. When tend to be of short duration, which are easily missed by per-
the distribution of borings with depth was plotted it revealed sonnel manually monitoring a staff gauge.
that most of the borings were terminated well above the top
of rock and only two borings (Fig. 12.11b) had actually pen-
etrated bedrock. Furthermore, they only penetrate the upper 12.2.3 Accuracy of Data
3 m of weathered rock. Clearly, any critical assessment of
top of rock, karst conditions, groundwater or possible deeper Accuracy can be defined as “closeness to the truth”. Errors to
radioactive contaminants at the site is limited based upon some degree occur in all measurements and data. Errors will
information from only two such borings. also occur in consultant reports and even our own work. To
Often the resolution of the data with respect to its intended make an error is easy, to catch it and correct it is much more
purpose may determine if it is adequate. For example, drill- difficult and requires attention to detail. Some errors may be
ing data that samples every 1.5 m has a vertical resolution of minor with little, if any, effect on our results. Some errors are
+/−0.75 m. If the intended purpose is to determine the thick- bigger than others and can have a major effect upon the proj-
ness of an aquifer 30 m thick, a resolution of +/−0.75 m or ect. Once we recognize that all measurements and data will
5 % of the thickness may be sufficiently adequate. However have some degree of error, we need a means to assess if a
if the thickness of the aquifer is only 3 m, then the same reso- given piece of data is sufficiently accurate to be used for a
lution would be 50 % of the thickness which is clearly not given purpose.
adequate. As data is being acquired, a preliminary assessment of its
Much as with the scale issue, the question of depth of quality and interpretation should be made in the field. After
measurements needs to be addressed. Since there are a wide all data has been acquired, a thorough assessment must be
range of measurements whose depths vary from the surface made so that subsequent interpretations, conceptual models,
(for example aerial photos) to much greater depths such as and risk assessments, are not impacted by errors, gaps or
borings. We need to select measurements that will provide incomplete data. This assessment includes a critical review
116 12 The Strategy

720
meters

0 150

Scale in Meters

900 meters

Location of 80 on-site monitoring wells, test

borings and piezometers
b

720
meters

0 150

Scale in Meters

900 meters

Only two locations penetrated into rock at least 3 meters

Fig. 12.11 At first glance the distribution of 80 borings over a 67 ha karst site appears to be reasonable (a). However, only 2 of the 80 borings
penetrated bedrock by about 3 m (b)

to identify deficiencies errors and inadequacy of any of the impact conceptual models, interpretations, calculations, risk
data used in the site characterization process including all assessment and modeling resulting in inaccurate site charac-
data from others along with data obtained from all databases terization, and engineering analysis. Therefore, it is critical
(Yuhr 1998). that all data must be assessed and verified before being inte-
If errors, data gaps or inappropriate data are allowed to be grated into the project database.
entered into the project database and are not detected they Errors come from a variety of sources and may include
can impact all future analysis. The results can significantly but not be limited to:
12.2 Appropriate, Adequate and Accurate Data 117

gs
is
er

ing
din
lys
ov

gg
un
na
alk

e o
Lo
So
oA

Vid
eW

ty

al
ty
t

g
i
ho

sic
v

ole
Sit

hin
tivi

gra

gs
lP

hy

reh
34

nc
sis
0%

31

c ro

rin
ria

op

Tre
EM

Re

Bo
EM
10

Bo
Ae

Mi

Ge
0

12

Limited to borehole depth

Limited to borehole depth

Segovia Limestone - unsaturated

Depth in Meters

up to 240 m
24

36

48

60

Fig. 12.12 An example of the range of methods and their depth of measurement used on a site characterization project in southwest Texas

Fig. 12.13 A creek bed running adjacent to the site appeared almost dry most of the time (a). The conditions are very different as a result of a rare
storm event (b)
118 12 The Strategy

• spatial locations such as improper location (x, y, z) of well was found that only about 100 of the wells could be used for
or measurement; the purpose of modeling. The others were omitted due to
• reporting errors where work done in the field is not docu- incomplete or erroneous data, which could not be reconciled.
mented or incorrectly documented in a report (such as the Many of the errors identified were simply impossible data
construction of a well); such as a water level below the bottom depth of the well.
• data gaps from inadequate spatial sampling (laterally or
vertically) or inadequate temporal sampling missing some
significant features or change in conditions. 12.3 Key Steps in the Site Characterization
• impossible data can include data outside of the spatial Process
boundaries (for example water level measurements below
the total depth of the well or coordinates of a well outside Site investigations should be carried out in a logical sequence
the study area) or outside temporal boundaries (for exam- and should focus upon developing a preliminary conceptual
ple water quality measurements made before the well was model of site conditions and verifying it or updating as the
installed). project proceeds. The proposed approach utilizes older exist-
ing approaches such as the “Observational Method” by
Two types of data must be considered for evaluation, fac- Terzzagi (Goodman 1999), and strategies used by TVA
tual data (e.g. the elevation at a well head, or water level in a (Hopkins 1977) as well as the later strategies of DOE’s ESC
well) and interpreted data (e.g. a contour map of water lev- philosophy discussed in ASTM D 6235-98a (ASTM 1998).
els). Errors in the factual data need to be identified and cor- These key steps are based upon the strategies developed over
rected. Errors in interpretation are more difficult to identify. many years by others and have been modified by decades of
If an interpretation is largely subjective, caution should be field experience by the authors.
used until it is either supported by data or discounted by data. There are always compromises made when planning each
In many cases, it will be found that considerable work has project. Project objectives, desired site coverage, resolution
already been done at the site including; borings, installation of measurements and spatial sampling will always be bal-
of monitor wells, hydraulic testing, and laboratory work. anced with time, budget, logistical and political (non-
This work is often done by a variety of consultants and con- technical) constraints. However, it can only succeed if the
tractors, sometimes over many years. In such cases, the owner is supportive and there is a small team of dedicated
results must be checked in detail to assure its quality and professionals who are in the field directly involved with the
applicability to the site characterization. site characterization process from beginning to end.
The following sequence of tasks presented in Fig. 12.14 is
12.2.3.1 Errors in Data and Databases an idealized one. There may be changes in the sequence, due
A variety of databases are available from local, state and fed- to technical or non-technical reasons. There may also be
eral agencies. Unfortunately, errors occur in virtually all iterations within a task and between tasks as data is obtained
databases (local, state and federal) as well as consultant and perceived conditions are better defined. All aspects of
reports. For example, in the western US data from water the program must be flexible in order to change direction and
wells are often contained in two databases, one for the well activities as new data is obtained and our understanding of
itself and another for the water use. It is not uncommon to conditions is improved.
find differences in the data for the same well from these two
databases such as well location, sometime listed kilometers
apart. In addition, the nature of databases is that they are a 12.3.1 Project Preparation
compilation of data over time. Over time the people who
gather the data or input the data have likely changed or 12.3.1.1 Defining the Problem
sources of the data has changed. These changes may result in The first step is to clearly define the objectives of the work by
inadvertent errors. For example, how many ways are there to talking with the client and other interested parties. This may
report a water level? Measurements could be reported in include county, state, and federal agency staff who may be
depth below ground surface as a positive or negative number reviewing and approving the project. If specific interests of
or an elevation. We have experienced cases where all three others can be identified and incorporated into the site charac-
were used in the same database. terization effort early, it is more likely that their approval will
An example of this was discovered at a landfill site where be more forthcoming at the end of the project. All too often
groundwater flow and contaminant transport modeling was there will be a wide range of divergent ideas from different
to be completed (Benson and Sharma 1994, 1995). More parties and it is necessary to get all parties in agreement so
than 1,000 monitoring wells were listed in the original data- that clear and realistic objectives can be defined. Good com-
base for the site. After assessing the well data accuracy, it munication is essential at this point.
12.3 Key Steps in the Site Characterization Process 119

Project Preparation
Defining the Problem
Area and Depth
of Interest
Work Plan
Identifying Site
Characterization Team

In Office Field Work

Desk Study Site Walkover
Aerial Photo Adjust Work Plan
Analysis Surface Geophysics
Preliminary Invasive Measurements Conversion of Data to
Conceptual Model Geophysical Logging Useful Information Additional Assessments
and Downhole Interpretation and Risk of Subsidence or Collapse
Measurements Integration of Data Groundwater Vulnerability
Assessment of Large Assessment and Modeling
Final Conceptual Model
Open Voids or Structures
Final Report of Findings
Engineering Measurements
and Monitoring
Hydrologic Characterization
and Monitoring
Dye Trace

Conceptual Model - continually and iteratively being tested and updated

Fig. 12.14 An ideal sequence of activates in a site investigation

During the initial discussions we attempt to obtain the cli- 12.3.1.2 Area and Depth of Interest
ents understanding and objectives for the project. It is also at Part of defining the problem should include defining the area
this point where conditions at a site (whether actual or per- and depth of interest for the investigation. The area of interest
ceived) are presented. It is important to identify what needs to be established from the client’s perspective. In many
assumptions limit our understanding of current site condi- cases, artificial boundary limits are placed on the site charac-
tions. In many cases we find that the client does not have a terization process. It is not uncommon for clients to suggest
good understanding of the importance of the geologic issues that work only be carried out within the immediate confines
and their impact to the project, and the necessary steps to of the site. While the majority of the detailed investigation
satisfactorily resolve them. Our role is to provide under- work will be done within the site boundaries, it should be
standing of the nature and extent of potential karst problems made clear that the regional and local setting should not be
and make recommendations and provide guidance for the ignored since it may provide significant information to
project. It should be clear that the initial definition of the achieve a complete and accurate site characterization.
project objectives might change as the project proceeds and There may be quite a bit of ambiguity in the early stages
preliminary data may reveal unforeseen conditions. The cli- of a project regarding the area and depth of interest. The
ent’s initial expectations must be balanced with the time and project objectives and initial estimates of area and depth of
budget as well as other possible external constraints such as interest along with data coverage, will commonly change as
regulatory issues. This can be accomplished by helping the questions are answered, new data is obtained, and as concep-
client recognize what is feasible and realistic within the lim- tual models are developed during early phases of the
its of technology, budget, time and other constraints. A prac- project.
tical level of reality must be defined prior to planning the site
characterization effort. A balance of risk versus level of 12.3.1.3 The Work Plan
effort (both time and cost) should be discussed to assist the The initial work plan should be focused on meeting project
owner in assessing their goals and level of comfort. objectives as well as challenging the opinions and assump-
120 12 The Strategy

tions of site conditions. The work plan identifies the type of topography. The survey grid can use paint marks on the
data needed to confirm the conceptual model or change it ground (in open areas), surveyor flags or wooden stakes
depending upon findings. In the ESC process, the work plan located at suitable intervals along a line or over a grid. The
is generic and flexible allowing the on-site core team to make survey grid remains in place until all work has been com-
the necessary decisions in the field on where and how to pleted at the site. This step may seem like common sense,
obtain the needed data based upon current findings and however, it is often overlooked. Getting data to correlate or
professional judgment. There will always be unforeseen con- seeing trends in data sets is often based upon a having correct
ditions and new data will be used to guide the fieldwork. and accurate locations for all data.
In addition, a number of other issues need to be addressed

• A site survey grid needs to be established for all tasks so 12.3.2 The Conceptual Model
that spatial correlation between data can be easily
accomplished. The conceptual model is an important building block in our
• Any site clearance or access issues need to be addressed site characterization effort. The term “conceptual model” or
• If there is not an existing land survey of the site, one needs “conceptual framework” is a convenient designation for a
to be obtained. means to visualize the sites expected geologic and hydro-
• If additional aerial photos or topography are needed they logic characteristics. It will typically consist of simple
need to be ordered. graphical presentations, usually cross-sections and plan
• Contractors need to be identified for any unique services views as needed to present the spatial perspective of geologic
not provided by the core team. and hydrologic conditions at a site. The conceptual model
• If needed a test laboratory for the project needs to be provides a means to document the current understanding of
identified. the site geology, hydrology and cultural conditions from
• Any other unique services or supporting team members which we continually test our understanding of these
that may be needed should be identified. conditions.
“The preliminary conceptual model is your first sophis-
Since the site characterization process is one of building ticated guess about the subsurface and what you are likely
and verifying the conceptual model, the results of one or more encounter at the site. This sketch or model is used to stimu-
sets of observations or measurements may suggest additional late an approach to geologic conditions and uncertainties
data is needed in an area or that a different set of data is with ordered thought. There is a significant amount of
needed for correlation. If unforeseen issues arise, the field deductive thinking required to construct the preliminary
plan is modified so that appropriate and adequate data can be conceptual geologic model” (Hatheway 1993). In addition,
acquired to address these new issues. This iterative process Hatheway (1993) suggests to “let the sketch of the prelimi-
continues until a sound conceptual model is developed and nary conceptual model sit for a day or two (if you have that
unknowns and uncertainties are sufficiently minimized. much time), then go through the sketch and explain to your-
Prior to beginning fieldwork a common survey grid must self or a co-worker why you have made your decisions”.
be established for the site. While there are numerous means This provides you with some time to be away from the
to record both data and position digitally using GPS, a local topic and then to come back to it with a fresh look. This
survey grid is needed to provide a reasonable degree of qual- process is recommended for all key issues including inter-
ity control in locating measurements and to allow a wide pretation, integration of data and the final report. On the
range of concurrent activity at the site. The authors have spot, rush interpretations without proper time for thinking
always used a semi-permanent local grid system for any and integration with other data will often lead to improper
project work. This local survey grid allows: conclusions.
A good model of site conditions allows both engineers
• Data to be more easily and accurately collected along a and geologists to understand the spatial relationship and
series of straight parallel, survey lines, interactions of the many geologic and hydrologic
• Preliminary field data to be reviewed and anomalous con- components that make up the site (Hoek 1999). The model
ditions to be readily re-located without the use of GPS, is most effective during these early phases of the investi-
• Different field teams to communicate and integrate data gation by identifying what is known, what is not known,
more clearly whether it is during the same field effort or what is conjectured along with uncertainties and where
months apart. significant gaps in knowledge may lie (Fookes and
Shilston 2001). One of the benefits of the conceptual
A site map should be developed that utilizes this local model is in its ability to help in anticipate conditions
reference grid and includes all pertinent site features and (Fookes 1997).
12.3 Key Steps in the Site Characterization Process 121

during the site characterization process. It should be chal-

3 to 6 m Surficial Aquifer
lenged and changed to incorporate new features and
thick Sand conditions as additional data is obtained in each step of the
site characterization (Fig. 12.17).
about 1 m thick Clay Aquitard The concept of a conceptual model for a site characteriza-
tion is not new. LeGrand and Rosen (2000) discuss the devel-
opment of a preliminary model of regional geologic
Limestone conditions using LeGrand’s 1983 “A Standard System for
Deeper Aquifer
Evaluating Waste Disposal Sites” or DRASTIC by Aller
et al. (1987). Motyka (1998) discusses conceptual models of
hydrologic networks in carbonate rocks and presents various
Fig. 12.15 The preliminary conceptual model may be very simple conceptual models for chalk, limestone and dolomite along
sketch with a brief narrative with their hydrologic parameters.

Geologic 12.3.3 In the Office

Data
The early stage of a site characterization is where the learning
Hydrologic Conceptual curve is steep, with our understanding rapidly increasing.
Data Model There may be a large amount of existing data and informa-
tion available on both regional and local geologic and
Contaminant
hydrologic conditions as well as cultural conditions that may
Data constrain the investigation or impact the site. The steps for
the site characterization at this stage includes:
Fig. 12.16 Geologic, hydrologic and often contaminant data may be
used to develop the conceptual model at a site • A desk study that covers the review of existing regional,
local and available, site-specific literature and data
• Develop a preliminary conceptual model of site condi-
Initially the conceptual model may be very simple sketch tions based upon available information
(Fig. 12.15) with a brief narrative of our early understanding • Acquisition of existing aerial photography and analysis of
of site conditions. It may only be an approximation of reality aerial photos
since we are in the early stages of site characterization and
may have only limited site-specific data available. At a mini- With these steps completed, the fieldwork can begin with
mum, the conceptual model will include both geology and some insight and logical direction. However, the team should
hydrologic features. When dealing with contaminant sites remain flexible and anticipate changes to accommodate the
the conceptual model will also include contaminant data unknowns.
(Fig. 12.16).
In developing the conceptual model of site conditions
we need to keep in mind that the present geologic condi- 12.3.4 The Field Effort
tions at a site are the result of its total geological and geo-
morphological history. This includes the stratigraphy, the The field effort is all about obtaining appropriate, adequate
structure, the past and present climatic conditions (tem- and accurate data. The data should be designed to verify and
perature, precipitation, and vegetation) along with the build upon the preliminary conceptual model, meet project
impact of weathering, glaciation, as well as changes in sea objectives and answer questions regarding the key unknowns
level and uplift or subsidence of the land mass (Fookes at the site. This effort should make up at least half of the total
1997; Fookes et al. 2005). By understanding the geologic site characterization effort. The sequence and number of
history of the site we can develop a reasonable understand- steps will be site-specific to meet project needs. Ideally,
ing of the events that may have controlled what we see these steps include:
today. While we don’t need to make a detailed study of
each of these issues we need to be aware of them and how • An initial site walkover and possibly a flyover
they may have impacted the site. The conceptual model is • The use of surface geophysical measurements to provide
continuously tested and developed through iterations necessary data density
122 12 The Strategy

A starting hypothesis is imbedded into the

Existing Preliminary preliminary conceptual model and is tested,
Data Conceptual Model modified and improved with new data)
New Data

Upgraded
Conceptual Model

New Data
Upgraded
Conceptual Model

New Data
(the objective is to converge
to a Final Conceptual Model Final
as quickly as possible, while Conceptual Model
maximizing confidence and
minimizing assumptions)

Fig. 12.17 The development of the conceptual model will go through many iterative cycles as new data is obtained and as the understanding of
site conditions evolve

• Invasive investigations by direct push measurements, • Integration of the data sets

borings and trenching • Final conceptual model with supporting data
• Downhole geophysical logging and measurements • Reporting and presentation of findings
• Assessment of large open voids or structures, as needed
• Engineering measurements and monitoring
• Hydrologic characterization and measurements (includ-
ing chemistry) 12.3.6 Additional Studies
• Dye tracing
Once the site characterization is completed and a final con-
Throughout the field effort as data is acquired and inter- ceptual model has been developed, additional activities can
preted, the conceptual model is being continually verified, be carried out, if needed. They may include a subsidence or
improved and expanded. Details are added as they become sinkhole collapse risk assessment or groundwater risk assess-
available and the conceptual model is changed if the data ment and modeling. Confidence levels in these activities are
shows unexpected conditions. In addition, the initial work greatly improved by the solid foundation of data and under-
plan should be reviewed and modified as necessary to make standing provided by the site characterization.
sure it is meeting project objectives, as the conceptual model
changes.
12.4 Summary

12.3.5 Conversion of Data to Useful The strategies presented here are embedded throughout the
Information remainder of Part II. The sequence of methodologies pre-
sented follows that of an ideal site characterization effort. In
During the project, as appropriate, adequate and accurate addition, more than half of the site characterization effort
data has been obtained it is checked for completeness and should be devoted to field activities. In practice however, it is
accuracy. Each data set is then interpreted independently and recognized that technical or non-technical factors may mod-
results added to the conceptual model as appropriate. ify the order in which the site characterization is carried out.
However, there remains the process of converting the large The strategies, knowledge and technology, to solve the
amount of data that has been acquired into useful informa- problem of locating, mapping and characterizing geologic
tion. This is an intensive effort of data assessment, analysis, anomalies in karst and pseudokarst are presently avail-
and integration. This part of the site characterization includes: able. The methods range from the traditional boring, sam-
pling, and laboratory testing to dye tracing, mapping of
• Data assessment for completeness and accuracy. caves, the many remote sensing, geophysical and mini-
• Data analysis and interpretation of each independent set mally invasive methods. These “off the shelf” technolo-
of data. gies provide an extremely diverse range of measurements
References 123

and methods for our “tool box” which can be applied to Fookes PG, Shilston DT (2001) Building the geological model: case
site characterization of karst, pseudokarst or other com- study of a rock tunnel in SW England. In: Griffiths JS (ed) Land
surface evaluation for engineering practice, Geologic Society of
plex geologic conditions. England Special Publication 18:123–128
It should be remembered that there is no single, universally Fookes PG et al (2000) Total geological history: a model approach to
applicable method, group of methods, or software that can be the anticipation, observation and understanding of site conditions,
used to achieve a complete and accurate site characterization Proceedings of the international conference on geotechnical and
geological engineering, Melbourne, Australia. Technomic
under all geologic site conditions. While a given method or Publishing Co., Lancaster, pp 370–460
group of methods may be successful in one situation, they Fookes PG et al (eds) (2005) Geomorphology for engineers. Whittles
may not be in another. In addition, most methods will provide Publishing, Caithness
some data, but it will usually take a combination of methods Franklin AG et al (1981) Foundation considerations in sitting of nuclear
facilities in karst terrains and other areas susceptible to ground col-
as well as critical thinking and subjective judgment to put the lapse. Geotechnical Laboratory, U.S. Army Engineers Waterways
pieces of the geologic puzzle in place. There is no easy answer Experiment Station, NUREG/CR- 2062, Vicksburg, MS
and there never will be. The strategy and methods must be Goodman RE (1999) Karl Terzaghi, the engineer as artist. ASCE Press,
selected to meet the specific project needs of each site. Reston
Hatheway A (1993) Preliminary site geologic model: your first sophis-
The many methods available for site characterization are ticated guess about the subsurface, perspectives no. 17. AEG News
discussed in the following sections with examples and mini 36(4):36–39
case histories to illustrate their applications. Each of the Hoek E (1999) Putting numbers to geology – an engineers viewpoint.
methods has its advantages and limitations. Every site char- The second glossop lecture. Q J Eng Geol Hydrogeol 32(1):1–19
Hoek E, Bray JW (1977) Rock slope engineering, 2nd edn. Institute
acterization may use a slightly different set of technologies Mining and Metallurgy, London
and measurements and may apply them in difference ways. Hopkins R (1977) TVA philosophy and methods for site investigation.
The “how to do it details” have been deliberately avoided. In: Butler D (ed) Symposium on cavity detection of subsurface
These details are extensive and are covered by college cavities. US Army Waterways Experiment Station, Vicksburg,
pp 25–26
courses and many other books, along with specialty short Hunt RE (2005) Geotechnical engineering investigation handbook, 2nd
courses. Numerous references are provided to fill this gap. edn. CRC Press, Boca Raton
LeGrand HE (1983) A standardized system for evaluating waste-
disposal sites, 2nd edn. National Water Well Association,
Worthington
References LeGrand HE, Rosen L (2000) Systematic makings of early stage hydro-
geologic conceptual models. Groundwater 38(6):887–893
Aller L et al (1987) DRASTIC: a standardized system for evaluating Motyka J (1998) A conceptual model of hydraulic networks in carbon-
ground water pollution potential using hydrogeologic settings. ate rocks, illustrated by examples from Poland. Hydrogeol J
U.S. Environmental Protection Agency, Robert S Kerr Environmental 6(4):469–482
Research Laboratory, Office of Research and Development Nelson P (1984) Hydrologic properties of rock and rock masses. Short
EPA/600/2-87/035 course notes, p 32
American Society for Testing and Materials (1998) Standard practice Olhoeft GR (1994) Geophysical observations of geological, hydrologi-
for expedited site characterization of vadose and ground water con- cal and geochemical heterogeneity. In: Bell RS, Leper CM (eds)
tamination at hazardous waste contaminated sites, ASTM D 6235- Proceedings of symposium on the application of geophysics to engi-
98a. ASTM International, West Conshohocken neering and environmental problems, vol 1, 27–31 March 1994.
Benson RC, Sharma D (1994) The pitfalls in site characterization and EEGS, Boston, pp 129–141
how to avoid them. Short course presented by Environmental Wilson WL (1995) Sinkhole and buried sinkhole densities and new
Education Enterprises, Salt Lake City Utah sinkhole frequencies in karst of northwest Peninsular Florida. In:
Benson RC, Sharma D (1995) Pitfalls of site characterization and how Beck BF (ed) Proceedings of the 5th multidisciplinary conference
to avoid them. Short course presented by Environmental Education on sinkholes and the engineering and environmental impacts of
Enterprises, Charlotte, NC, November karst: karst geohazards, Gatlinburg, Tennessee. Balkema,
Benson RC et al (1984) Geophysical techniques for sensing buried Rotterdam, pp 79–91
wastes and waste migration. US EPA Environmental Monitoring Worthington SRH (1999) A comprehensive strategy for understanding
Systems Laboratory, Las Vegas, NV. Contract No. 68-03-3050 flow in carbonate aquifers. In: Palmer MV et al (eds) Karst model-
Benson RC et al (1996) Uncertainty in the geologic setting and its ing, Special publication 5. Karst Waters Institute, Leesburg, pp
impact on site characterization. In: Shackelford CD et al (eds) 30–37
Proceedings of uncertainty ’96, uncertainty in the geologic environ- Worthington SRH et al (2000) Matrix, fracture and channel compo-
ment: from theory to practice. ASCE Geotechnical Special nents of storage and flow in a Paleozoic limestone aquifer. In:
Publication No. 58(1):91–103 Sasowsky ID, Wicks CM (eds) Groundwater flow and contaminant
Da Cunha AP (1990a) An introduction to scale effects. In: Da Cunha AP transport in carbonate aquifers. Balkema, Rotterdam, pp 113–128
(ed) Proceedings of the 1st international workshop on scale effects Yuhr L (1998) Managing data and its integrity in the ESC process,
in rock masses. Balkema, Rotterdam, pp 3–27 Symposium on the application of geophysics to environmental and
Da Cunha AP (ed) (1990b) Scale effects in rock masses, Proceedings of engineering problems, EEGS, Chicago, 22–26 March 1998
the 1st international workshop on scale effects in rock masses. Yuhr L et al (1996) Achieving a reasonable level of accuracy in site
Balkema, Rotterdam characterization in the presence of geologic uncertainty. In:
Fookes PG (1997) Geology for engineers: the geological model, Shackelford CD et al (eds) Proceedings of uncertainty ’96, uncer-
prediction and performance. The geological society of London. Q J tainty in the geologic environment: from theory to practice. ASCE
Eng Geol Hydrogeol 30:293–431 Geotechnical Special Publication No. 58(1):195–209
 The Desk Study
13

Abstract
The desk study is where a basic understanding of the site is developed. We also begin to
identify what we know and do not know about the site along with a preliminary conceptual
model for the site. The desk study includes the acquisition of relevant existing data that cov-
ers the regional setting, its geology and geomorphology (the sites geologic history) as well
as local and possibly site-specific conditions. During the desk study, reconnaissance aerial
photos, topographic and geologic maps should be obtained to provide an overview of the
site and surrounding area. Information should be gathered and mined from a variety of
sources including government agencies, sinkhole, cave and well databases, site-specific
reports and even anecdotal information. This information and data is then used to develop
the preliminary conceptual model of site conditions.

13.1 What We Know and Don’t Know • Government agencies (local, county, state or federal
including USGS)
During this phase of work, the process of identifying what we • Publications from geologic and engineering organiza-
know and what we don’t know about the site should begin. tions and their field trips,
Tables 13.1, 13.2, and 13.3 provide an overview of the many • In-house client libraries or files which may contain previ-
questions that may be asked regarding possible geologic, ous work at the site by others such as consulting firms and
hydrologic and karst conditions along with cultural factors that drilling contractors reports
may impact the site. These tables include generic topics, which • Research by local universities
are applicable to most site characterization. These tables are • Newspaper reports
meant to promote a discussion of some of the many features • Observations and experiences of the owner’s, employees
that may impact a site characterization and are not intended to and neighbors.
be inclusive of all possible topics. Not all of these issues will be
included or are applicable on every project. Some of the ques- The wealth of existing data that is easily attainable has
tions may be answered during the desk study, aerial photo increased greatly since data of all sorts have become available on
analysis and site walk over, while other questions will not be the internet. Government agencies often have reports and maps
resolved until further site-specific work has been completed. available on-line and some even have digital maps ready for inte-
gration into geographic information systems (GIS) used for the
project. Field trip reports organized by various professional orga-
13.2 Sources of Existing Information nizations and local geologic groups are often quite useful since
and Data they tend to include unique and critical aspects of the geology for
an area. All existing data should be used in context and any of its
A wide variety of sources for regional and sometimes site- inherent limits of scale or detail should be recognized. In addi-
specific geologic and hydrologic data are available from data- tion, all data should go through a review for errors, this is particu-
bases or published literature. Initial data can be obtained from: larly true for data that will be used for site-specific purposes.

© Springer Science+Business Media Dordrecht 2016 125

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_13
126 13 The Desk Study

Table 13.1 Geologic and karst issues

General geologic and What is the regional and local geologic stratigraphy?
geomorphic issues What rock types will we be dealing with? Soluble rocks such as limestone, dolomite, gypsum, salt, or non-soluble
rocks or possibly some combination
Are there possible outcrops, quarries, and road cuts where rock may be observed?
What is the characteristic of the rock matrix (age, density, porosity, degree of fracturing, etc.)?
What is the nature of the un-weathered intact rock, is it massive rock or is it well jointed, is it thin or thickly bedded strata?
Are there possible fissures or lineaments present?
What is the fracture spacing, if present?
Is there any structure, synclines or anticlines or faults present?
What is the sites geomorphic history (i.e. changes climate, sea level, uplift, and subsidence and its potential impact)?
What is the thickness and character of overburden soils, sediment, or fill material?
Karst specific issues What type of karst features can be expected, their stage of evolution and maturity (see Karst classification Fig. 4.1)?
Identify recent sinkholes, and paleocollapse features at and surrounding the site along with their proximity to the
site using databases and aerial photography
Are there any sinkhole or karst databases or sinkhole maps available?
What type of sinkholes are present (Fig. 3.1)
What is the expected cause of the sinkholes along with their depth of origin?
Are there any spatial trends of sinkholes that are evident?
Is there the potential for buried sinkholes associated with historic sea level stands?
Identify culturally induced sinkholes (pseudokarst) vs natural ones?
What is the nature of the epikarst, the degree of weathering at the top of rock or the top of rock profile (rockhead relief)?
Obtain locations and maps of caves from databases, agencies, NSS or cave grottos
Characterize caves proximity to the site, their depth, classification (fracture, bedding plane or structure) along with
the geologic unit(s) in which they are formed

Table 13.2 Hydrologic and karst issues

General hydrologic Identify the presence of rivers, lakes, springs, and recharge
issues Establish the general direction of surface water flow
What is the seasonal and annual precipitation?
Assess the groundwater system including basin boundaries, depth, flow directions, and velocities
Identify groundwater recharge and discharge points
Determine aquifer characteristics; thickness, porosity, hydraulic conductivity, storativity, isotropy, homogeneity and
flow
Obtain water quality data including; turbidity, color, iron staining, oil sheen, noticeable odor, etc. as well as field
parameters such as specific conductance, temperature, and pH
Evaluate temporal changes; in flow, variation of hydraulic head, and water-quality parameters
Identify the presence of an aquitard, layers of clay or shale
Are there any zones of flooding?
Dry or intermittent creek beds
Identify the presence of dams, lakes, impoundments, and man-made recharge zones
Identify areas of groundwater or other fluid withdrawal (possible subsidence) or injection or spray fields
Are contaminants expected at the site, if so, the type, extent and any details already known?

Karst specific issues What is the depth thickness and nature of the epikarst along with its lateral variability, is it saturated or unsaturated?
Determine the presence and direction of flow in caves and springs
Have any dye traces been carried out?
Assess changes in water levels due to excessive rainfall or drought
Assessment of past sea level changes at the site
Evaluate the past or present fresh water lens and the salt water interface (where appropriate)
Assess changes in water levels due to groundwater withdrawal or dewatering from near by quarries or mines, if present
13.3 Type of Data Available 127

Table 13.3 Cultural issues

Above ground Develop an understanding of past and present land use. This is particularly relevant when working in an area that has been
developed for many decades
Identify quarries and their dewatering
Identify areas of drainage and changes in surface water flow over time
Identify areas of concentrated runoff from structures and paved areas
Identify the presence of dams, lakes, impoundments, and recharge zones
Identify areas of contaminants, leakage or spills
Assess changes in surficial conditions (cut and fill, modification of surface drainage and recharge)
Has there been building stress and settlement due to subsurface conditions?
Identify changes in vegetation and trees
Identify access issues including; roads and trails, along with limitations due to vegetation, wet areas, ponds, lakes and
rivers, fences and property boundaries that may limit or provide access
Identify overhead utilities and metal fences (which may impact access or affect some types of measurements)
Below ground The location of active and abandoned mines as well as mine shafts, tunnels and their dewatering
Identify areas of withdrawal of groundwater from community well fields or private wells, along with the injection of fluids
Locate underground utilities, tanks and other infrastructure that may impact conditions or impact measurements
Identify areas of possible fill, dumping, trash, etc.

13.3 Type of Data Available of data on a local or site-specific basis. In addition, these type
of features are likely to also have a hydrologic impact.
The type of data that may be available from any one of the
above sources is quite varied. A focus should be made on the
following: 13.3.2 Aerial Photos

• Topographic and geologic maps As part of the desk study, aerial photos should be obtained as
• Aerial photos (current and historical) early as possible to provide an overview of the site and sur-
• Geologic and hydrologic reports rounding area. Sources for aerials can range from easily
• Databases of sinkholes and caves or cave maps accessed websites or images with GIS overlays of data from
• Databases of borings, water wells, and monitor wells commercial or government agencies. In the past, a photo
• Maps showing cultural features and changes over time mosaic of the area could be constructed by assembling a
• Anecdotal information sequence of 9″ by 9″ photos into a mosaic (Fig. 13.1). Recent
aerial photos provide the team with further insight and typi-
cally more up to date information than the topographic maps.
Google Earth or Microsoft Virtual Earth provides world-
13.3.1 Topographic and Geologic Maps wide coverage with user friendly GIS overlays with both
vertical and oblique angles along with zoom capabilities.
Topographic maps of the site are one of the most fundamental While not useful for a detailed photo analysis (to be dis-
starting points of the desk study. Topographic maps of various cussed later) they provide an excellent overview of the site
scales are available from the USGS. The most useful are the and surrounding area to familiarize the team with the site and
7.5 min quadrangle maps at a scale of 1:24,000. These maps its setting.
provide topographic data as well as identify hydrologic and Using the topographic maps and the preliminary aerial
cultural features. Sinkholes and springs are often obvious photos we can begin to develop a feel for the level of karst
along with cultural features such as roads, mines, quarries etc. maturity for the site (Table 4.1 and Fig. 4.1). We can also
These maps are now available in both printed or digital for- begin to identify:
mats. These maps provide context for the regional setting and
local conditions, which is essential during the site walkover. • Geologic windows – Areas where rock outcrops can be
Geologic maps tend to cover large areas rather than local observed should be identified. This includes mines,
details. However, it is that regional context that can be impor- quarries, outcrops and road cuts. Time and effort should
tant when working on a complex site characterization. The be allotted in the site walkover to view these outcrops and
presence of faults, synclines, anticlines or overturned beds document conditions found. Geologic field trip guides
while all large-scale features will all impact the interpretation often include these locations as well.
128 13 The Desk Study

surveys. Much of the information available at the beginning

of the site characterization process often lies in the regional
literature. Field trips reports organized by various profes-
sional organizations and local geologic groups are also often
quite useful since they tend to include local or site-specific
aspects of the geology.
While access to the site itself for detailed investigation is
usually not a problem, legal or physical access to some of
the surrounding area may not be available. Should this be
the case, having an understanding of the regional geologic
setting from the desk study will help fill the gap between the
site itself and the surrounding properties.
Sometimes we are lucky and considerable detailed infor-
mation is available in the literature. The following two exam-
ples are from Southwest Texas and the Kansas City area.

13.3.3.1 Southwest Texas

In southwest Texas, Freeman (1968) had observed indica-
tions of elongated collapse ranging in width from a couple of
meters wide and several meters long to as much as 450 m
wide and almost 5 km long. These features were for the most
part aligned in a northeast direction and many had angular
branching suggesting a relation to joint sets. Faults along one
or both sides of the long narrow areas were not uncommon.
Subsidence of 12–15 m was typical and as much as 76 m was
demonstrated.
While this report focused on an area a several kilometers
south of the site we were working on, it alerted us to major
paleokarst features in the area. These features were subse-
quently extrapolated to the site by detailed aerial photo anal-
ysis. This turned out to be a critical aspect of the karst
investigation for the site of interest.

Fig. 13.1 An aerial photo of the site and surrounding area is essential 13.3.3.2 Kansas City Area
early in the project. Such photos can be ordered or constructed from a Within the Kansas City Area, Hasan et al. (1988) have
group of (9 × 9) aerial as a mosaic provided the “Geology of Greater Kansas City” a series of
documents published by the Association of Engineering
Geologists covering many cities of the US. This document
• Hydrologic conditions – Surface water bodies (lakes, covers a broad range of geology and hydrology for the area
rivers, creeks and ponds) should be identified. Some including problems with limestone mines. Another docu-
larger springs may be located on topographic maps and ment published by AEG was also very useful. Gentile (1984)
other locations are found in USGS literature and State discusses deep-seated paleocollapse structures in the Kansas
Geologic Survey reports. Smaller discharge points along City area.
with most recharge points (sinks) will be found during the These documents describe the Pennsylvanian cyclothemes,
site walkover. a sequence of limestone and shale about 152 m thick that
outcrop in the Kansas City area. Each of the shallow limestone
strata is generally thin and do not support the development of
large cavity systems. In contrast, the underlying deeper and
13.3.3 Geologic and Hydrologic Reports massive Mississippian limestones are more than 120 m thick.
Large cavernous systems have developed within the massive
Most areas will have some regional geologic data available Mississippian limestone throughout the Midwest. Some have
in the scientific literature. Geologic and hydrologic reports collapsed in the Kansas City area resulting in visible subsid-
are commonly available from the USGS and state geologic ence at the surface (Gentile 1984). These two documents,
13.3 Type of Data Available 129

Hasan et al. (1988) and Gentile (1984) provided critical 13.3.5 Cave Maps and Databases
support in developing the preliminary conceptual model for
the case history presented in Chap. 25 in Part III. A great deal of information about an area can be obtained
from cave maps and databases (Sect. 3.4). Such data can be
obtained from State agencies that maintain such records and
13.3.4 Sinkhole Databases by consultation with the National Speleological Society
(NSS) and their caving members who have formed local
Sinkhole databases have been developed for some karst areas. groups called grottos (www.cave.org). If cave maps are
For example, the Florida Geologic Survey maintains such a available, the cross section of a cave will allow us to make an
database. These databases can provide the location, date of assessment of minimum and maximum depths of a cave and
occurrence and often contains some details such as size, its relation to the geologic stratigraphy. It will also reveal the
shape and depth, along with the circumstances of collapse. location of any large chambers, if they are present. This
It is often possible to search the database to determine information can be useful in determining preferential depth
the number of sinkholes within a radius of a site or the num- or patterns of dissolution that can support the conceptual
ber of new sinkholes per year in a given area (NSH) (Wilson model for the site.
1995). The number of sinkholes within a given radius of a Cave databases are also available in some areas. This
site will provide an initial indication of nearby sinkhole information can often be searched to provide number of
activity. The NSH can be an initial indication of sinkhole caves near a site under investigation. Figure 13.2 is an example
activity for an area. However, it should also be noted that from West Virginia where a prison facility was being located.
very mature karst (Table 4.1 and Fig. 4.1) can have a small The cave locations were plotted at a 1.6, 3.2 and 4.8 km
number of NSH. Such examples occur in Puerto Rico and radius from the site. A total of 58 caves where located within
the cockpit karst of Jamaica. about a 10 km radius of the site as per communication with
If sinkholes are present, at or near the site, they should be the Virginia Speleological Survey (Lucas, 1997, personal
identified for further follow up as part of a site walkover. communication). Although detailed cave maps were not
Information that may be relevant could include: provided for each cave, this data did provide a first approxi-
mation to cave density within the area of interest.
• Their estimated size (diameter) and depth.
• Are the sinkholes old, recently formed or reactivated
• Are the sinkholes dry, part of a river or lakes 13.3.6 Databases of Borings, Water Wells,
• Sinkhole density (number per unit area) and or typical and Monitor Wells
spacing between sinkholes
• Their geometry (a circular single sinkhole or an elongated Databases are often available for borings, supply wells and
composite of merged smaller sinkholes) monitor wells in many counties and states. These databases
• What is the thickness of overburden may include the well location, geologic data, depth, and
• What is there depth of origin (i.e. near top of rock, or screen interval of the well, along with construction informa-
deeper within the rock matrix) tion. This information is often very useful in providing a
• In what geologic formation do they originate regional perspective. In some case, site-specific data may
• Are there any spatial trends that are evident such as alignment exist. This may include simplified geologic data or depths of
between sinkholes forming linear trends water production zones. This regional geologic and hydrologic
data is often sufficient to develop a preliminary conceptual
In addition, there may be sinkhole risk or susceptibility maps geologic model for the site itself although the data may be
and groundwater vulnerability maps developed for some karst many kilometers from the site (Fig. 13.3).
areas. An early sinkhole risk map of Florida (Fig. 3.10) by
Sinclair and Stewart (1985) is based upon sediment cover thick-
ness and type of sinkholes. Such regional risk maps are useful 13.3.7 Cultural Features and Changes
indicators of the type of sinkhole activity expected at a site.
These regional sinkhole susceptibility maps and sinkhole Development on and around the site may have caused changes
databases provide an excellent overview of sinkhole occur- over time. Changes could be significant such as topography,
rence. Keep in mind that databases tend to list only recent sink- drainage or access. Other changes that may limit access and
hole activity and tend to be biased towards population centers. can impact some of our measurements may include:
Such data may be of interest depending upon its proximity to
the site under investigation; however, it may not necessarily be • Current land use and changes in land use over time
an accurate indicator of site-specific sinkhole activity. • Roads, trails and fencing that can affect access
 N
4.8 km

3.2 km Proposed
Prison
Facility
1.6 km

- cave location

Fig. 13.2 This example shows the locations of known caves based upon a cave database. The cave locations are shown at distances of 1.6, 3.2 and
4.8 km (1, 2, and 3 mile) radius from a site of interest

Stauffer Site Projected

onto cross section
Approximately 5.6 km Approximately 4.8 km
TR 15-2

1072
2560

SSW
NNE
W-12943 Geology Hydrology W-16609

0 Post-Hawthorn Group Undifferentiated Surficial Aquifer 0

Sands and Clays
Tampa Limestone
Approximate Depth in Meters
Approximate Depth in Meters

30 30

Suwannee Limestone Floridan Aquifer

60 60

90 90

Ocala Limestone
120 120

Fig. 13.3 The initial regional geologic and hydrologic data may be many kilometers from the site itself
13.4 Data Mining and Review 131

• Overhead and underground pipelines, power lines, and tion or incorrect construction/depth from that noted in the
other utilities should be identified since these may impact database. The impact of reducing the number of reliable
access and affect our field measurements monitoring wells from 167 to about 45 is obvious when
• Major supply wells, well fields and well injection sites evaluating county-wide water quality conditions or devel-
near the site should be identified oping a groundwater model. But consider the impact of uti-
lizing the database information without realizing these
It is just as important to note the lack of changes at a site. errors (Yuhr 1998).
This could provide some evidence of stability at a site. How It is also possible to identify useful data from unex-
the site has been used over time may also be important. pected sources. This is often possible at a site that has been
Stresses and loads at a site that did not cause any sinkhole in operation for many years and therefore has a large
activity. This information is all part of the site history and amount of site-specific data available. Much of this data is
should be considered when developing a conceptual model. often buried or scattered in old reports and requires a bit of
persistence to find.
An example of this occurred during a site characteriza-
13.3.8 Anecdotal Information tion at a landfill in Independence, Missouri. The landfill
was built upon a shale layer that was to act as a barrier to
Stories of sinkhole activity are often obtained from local news- the groundwater. As the existing data was reviewed, it was
papers and persons living in the area. Caution is recommended found that prior to construction of each new landfill cell a
in believing everything you read and hear. Newspaper stories small number of geotechnical borings were installed and
are just that and are not necessarily reliable sources of technical packer tests completed in the shale layers underlying the
data and quite often the reported sinkholes are due to water or site (upper 3 m). This was part of the regulatory require-
sewer line breaks (pseudokarst) not natural sinkholes. ments for the facility. The data was submitted to the
Observations or anecdotal stories by others must be veri- agency and filed away. During a site characterization
fied because memories may not present a clear picture of the where the objective was to develop a single document
events of one or two decades ago or in some cases only a few summarizing hydrogeologic conditions and review the
weeks ago. However, these sources of information can often groundwater monitoring plan for the site, these data were
provide insight to local or even site-specific conditions that discovered. A map of all existing borings and monitoring
may not be available elsewhere. wells was created. The greatly enhanced density of bor-
ings revealed a portion of a syncline underlying the site,
which impacted the groundwater flow. In addition, packer
13.4 Data Mining and Review tests completed within the shale layer actually showed the
presence of water and a level of transmissivity that was
When a project is initiated at an existing facility or site, the not expected of a barrier layer of shale (aquitard). These
owners will often have a sense of starting with a solid base of under-used data provided a solid base of existing data that
information. There may have been considerable work that could be used at the site, and the initial understanding of
has already been completed at the site by others. However, site conditions was dramatically changed prior to begin-
all too often when the factual data is reviewed, the amount of ning any field efforts.
useful data and its accuracy is often found to be much less At the end of the desk study there is usually a large
than originally expected and perceived conditions change as amount of data available including notes, aerial photos,
a result (Yuhr 1998). maps, topographic sheets, boring logs, portions of data-
As an example, one project included the compilation, bases and GIS documents along with some site-specific
review and physical inspection of a county-wide saltwater reports. This data makes up the initial database for the site
intrusion monitoring well network consisting of 167 wells. and will be added to as the project proceeds. Therefore,
An initial visual inspection to confirm well locations and appropriate standard software and or GIS system should be
access revealed damage or loss of almost half the wells selected for the project. This will facilitate retrieval, com-
(destroyed, paved over, built over, etc.). The remaining half parison, plotting and interpreting data throughout the
of the wells were inspected in more detail and included project.
confirmation of well construction, its total depth and inter- With the project objectives finalized and this initial data a
nal conditions using a downhole camera. Almost 50 % of work plan for field efforts can be developed along with the
the remaining wells were impacted due to collapse, obstruc- preliminary conceptual model of site conditions.
132 13 The Desk Study

13.5 The Preliminary Conceptual Model References

The data obtained during the Desk Study will provide the Freeman VL (1968) Geology of the Comstock-Indian Wells area Val
Verde, Terrell, and Brewster Counties. Texas USGS professional
basis for our preliminary conceptual model of site condi-
Paper 594-K
tions. While a basic understanding of the site has been used Gentile RJ (1984) Paleocollapse structures: longview region, Kansas
during contracting, discussions with the site owner’s and City, Missouri. Bull Assoc Eng Geol 21(2):229–247
developing a work plan, the initial data is now available to Hasan S et al (1988) Geology of greater Kansas City, Missouri and
Kansas, United States of America. Bull Assoc Eng Geol
evaluate what we really know and don’t know about the site.
25(3):281–341
We should be able to include the regional setting and the Sinclair WC, Stewart JW (1985) Sinkhole type, development, and
geomorphology that may impact the site, as well as local and distribution in Florida, Florida Geologic Survey Map Series
possibly even site-specific data. #110, FGS Tallahassee, Florida
Wilson WL (1995) Sinkhole and buried sinkhole densities and new
The preliminary conceptual model will be developed and
sinkhole frequencies in karst of northwest Peninsular Florida. In:
modified throughout the project, as data becomes available. Beck BF (ed) Proceedings of the 5th multidisciplinary conference
Changes and modifications to the conceptual model should on sinkholes and the engineering and environmental impacts of
only be made if they are supported by solid data. When strip- karst: karst geohazards, Gatlinburg, Tennessee. Balkema,
Rotterdam, pp 79–91
ping out unsupported interpretations, it is not unusual to find
Yuhr L (1998) Managing data and its integrity in the ESC process,
that the initial understanding of site conditions was based Symposium on the application of geophysics to environmental and
largely on opinions and assumptions. engineering problems, EEGS, Chicago, 22–26 March 1998
 Aerial Photography and Remote
Sensing Data 14

Abstract
With the desk study completed, the objectives of the site characterization should be well
defined, the area of investigation known and the regional geologic setting established. While
the overview aerial photo or photo mosaic of the area of interest has provided an initial
familiarization of the site, we are now concerned with additional details that can be obtained
from a detailed analysis of the aerial photos.

Aerial photography provides a very cost-effective method of imagery. There is an increasing array of imaging methods
getting a regional and local overview. While topographic and being used for special applications. Many types of aerial
geologic maps are one of the primary sources for this type of photos or satellite imagery are available. They all have
information, not all details can be presented on a single map. advantages and disadvantages (cost, coverage, scale, range
Furthermore, such maps are not frequently revised. It is here, of wavelengths, etc.). Here, the focus is upon the most com-
where aerial photography can provide additional details on monly used data for general geologic and karst
regional and local conditions as well as temporal data often assessment.
over many decades. Aerial photography is available in the US from a number
In addition to aerial photography, selected imaging and of federal agencies such as United States Geological Survey
remote sensing technology has been included in this section. (USGS) and the Soil Conservation Service (SCS). Many
This type of data is generally limited to more regional map- state agencies, such as Department of Transportation (DOT)
ping projects due to the scale of information. However, some also may have aerial photography available as well as private
very useful applications of imaging and remote sensing data sector companies.
have been applied karst projects. For example, aerial photo In more recent years, internet-based access to aerial
and imaging technologies provide both spatial and temporal photography, imagery and satellite data has expanded and
data for developing a karst inventory. Management of karst become quite useful for reconnaissance purposes. Sources
areas has utilized spatial and temporal analysis of terrains for such as Google Earth and Microsoft Virtual Earth provide
planning, development, and zoning. world-wide coverage with user friendly GIS overlays with
Lillesand and Kiefer (1994), Paine and Kiser (2012), and vertical and oblique angles as well as zoom capabilities.
Johnson and Pettersson (1988) provide detailed discussions of The imagery includes scale as well as elevation data and
aerial photography and remote sensing image interpretation. labeling of key roads and features. While not used for a
quantified photo analysis they provide an excellent over-
view for location, planning purposes, and orientation of
14.1 Availability field team members. This is an area that is continually
expanding availability of coverage and features. For exam-
The selection of photography or imagery will depend upon ple, older images are now available on Google Earth to
the specific-site characterization needs, the area involved, look at changes over time.
the availability of existing imagery and the cost of new

© Springer Science+Business Media Dordrecht 2016 133

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_14
134 14 Aerial Photography and Remote Sensing Data

14.2 Scale available to order 9 × 9-in. aerials. These can be used

individually, as stereoscopic pairs for a detailed analysis.
As with all types of measurements, scale of the photos or
imagery is important (Fig. 14.1). Landsat images of
1:500,000–1:1,000,000 provide a synoptic view of the 14.4 Aerial Photos
regional geologic setting over 100’s of km2. High altitude
stereoscopic aerial photographs at scales from 1:60,000 to Aerial photos have been used for site characterization since
1:120:000 are used for intermediate scale analysis. Finally, the early 1900s. Their application can be qualitative or quan-
stereo-paired aerial photo interpretation with 1:10,000 or titative in nature. Black and white aerial photos have been
1:20,000 scale provides a high level of detail for more local the standard for analysis because of their availability and
photo analysis. Real estate/tax assessment aerial photographs ease of qualitative and quantitative interpretation. But more
at a scale of 1:300 are readily accessible for many developed importantly because of their continued acquisition over
areas. Such detailed photos provide local site-specific cul- many decades which provides a temporal record of geologic,
tural details such as roads and buildings. hydrologic, and cultural changes.
Qualitative interpretation consists of simply viewing the
aerial photos and noting obvious features and providing ini-
14.3 Coverage tial familiarization with the site for the early phase of any site
characterization project, this is typically done as part of the
Aerial photo coverage, at a minimum, would normally cover desk study. Initial familiarization may include assessment of
the site itself and the surrounding area. Regional coverage site access and logistical planning which may be impacted
might extend for many kilometers from the site. For exam- by features such as vegetation, lakes, rivers, roads and trails
ple, if the purpose of the site characterization is one of site as well as cultural changes due to development at the site.
selection or if there is complex geomorphology impacting Current land use including property boundaries, fences,
the site. Aerial photography traditionally has been flown on a areas of fill, landfills, structures, major pipelines and trans-
fairly regular basis dating back to about the 1930s. Maps mission lines, along with both quarries and mines can be
showing photo coverage for most, if not all states, are identified. In addition, a review of aerial photos over a period
of time, such as a few decades, will provide insight into land
use changes. One of the most common uses of aerial photos
1:500,000 to
for karst studies is to locate sinkholes including their size
Satellite Data and spacing. By comparing older aerial photos (some go
1:1,000,000
back to the 1930s) with more recent aerials one can deter-
mine sinkhole activity over time.
Historic aerial photography can be used to identify old
1:60,000 to
High Altitude Data sinkholes where more recent land use has caused sinkholes
1:120,000
to be completely covered, partially covered or filled-in.
Panno and Luman (2013), working in a agriculturally-rich
southwestern Illinois, digitized and orthorectified 1940s
black and white aerials. Those older aerial photos allowed
1:20,000 to the near-surface geology to be detected through the mature
Low Altitude Data
1:100,0001 summer crop canopy. Approximately 30 % more sinkholes
are present than were delineated on the USGS 7.5 min topo-
1:300 to 1:6002 Very Low Altitude Data graphic quadrangle maps of the study area.
Quantitative interpretation consists of a detailed analysis
to make a “photogeologic assessment of conditions”.
Quantitative analysis is also often used later in a site charac-
terization when there are very specific questions to be
answered. Quantitative analysis requires the skills of a
trained and experienced photo interpreter. This type of analy-
sis can provide data such as topography and drainage, frac-
1 - commonly used for aerial photo interpretation
ture patterns, geologic mapping and hydrologic conditions.
2 - Used for very detailed data
This is often accomplished by looking at variations in loca-
Fig. 14.1 Range of scales that are commonly used for acquiring aerial tion, size, shape, shadows, tone/color, texture, patterns, etc.
photography and satellite imagery of various features (Clayton et al. 1982, 1995). All of the
14.4 Aerial Photos 135

• Development and use of the land over time can be

determined by a sequence of photographs taken over
decades.
• Special areas of concern can be identified such as areas of
instability (sinkholes and landslides), sources of materials
from quarries and mines, both active and abandoned.
• Site accessibility can be determined to help plan move-
ment of drill rigs or other heavy equipment via roads or
trails.

14.4.1 Fracture Trace and Lineament Analysis

Fracture traces and lineaments are natural linear features

consisting of topographic, vegetation, or soil tonal align-
Fig. 14.2 Two standard stereoscopes for stereo-pair analysis (desk
mirror stereoscope and portable stereoscope) ments, visible primarily on aerial photographs and some-
times on satellite imagery and topographic maps. They are
commonly straight in plan view and are considered surface
aerial photo and imaging methods are responding to the manifestations of vertical or near-vertical zones of fractures.
upper-most surface of the ground. From this we can often Fracture traces are defined as continuous features less than a
make interpretations of deeper geologic and geomorphic kilometer or so and lineaments greater than a couple of kilo-
conditions. Traditionally, this was done manually using meters in length (Parizek 1988). Joints and fracture traces
stereo-pairs (Fig. 14.2). Today, the aerials are digitally can be recognized on 1:20,000 aerial photos and lineaments
acquired and analyzed using computer software programs. are best recognized on 1:500,000 Landsat images. It is best
For example, an array of circular fissures on the ground to use the terms such as “photo fracture trace” or “photo-
surface were initially thought to be due to a large 7 ha col- lineaments” which indicate that the feature was interpreted
lapse within an abandoned limestone mine at a site in the from aerial photos as opposed to other means such as
Kansas City area. The fissures lay directly over the area of the mapped in the field or interpreted from surface geophysical
mine collapse. To confirm this cause and effect relationship, data.
historical aerial photos, prior to mining, were reviewed. Water resource development has utilized fracture trace
Evidence of the surface fissures was confirmed prior to the and lineament analysis to locate wells in fracture zones to
time of mining. It became clear that these fissures were not optimize water recovery. Sinkholes often tend to occur along
due to the mine collapse, but were instead associated with linear patterns of fracture traces or lineaments (Fig. 14.5).
deep-seated paleocollapse, not uncommon in the area (Gentile Littlefield et al. (1984) have suggested that sinkhole proba-
1984). Chapter 25 provides more detail on this case history. bility increases at major lineament intersections and
In general, the type of information that can be obtained decreases as the location moves further from a lineament.
from an analysis of aerial photos might include, but are not Short-term sinkhole development is most probable along the
limited to: largest photolinear features.
Parizek (1988) cites 21 features by which fracture traces
• Topography of the site and the surrounding area. This is and lineaments can be recognized, some of which include:
particularly important when access to the surrounding
area is restricted for any reason. • Long straight river channels are indications of major lin-
• Obvious sinkholes are generally easy to identify on aerial eaments while abrupt and angular changes in stream
photos channels over short distances are indications of local frac-
• Fracture orientation and spacing along with lineament ture trends
analysis (Fig. 14.3) • Sinkholes and depressions that form a linear pattern
• The geology of a site can often be interpreted from land- • Alignment of vegetation
forms, drainage patterns, vegetation and land use • Alignment of soil tonal patterns
(Fig. 14.4).
• Surface water and drainage conditions of the site and sur- Lattman (1958) and Lattman and Parizek (1964) have
rounding area can be assessed including rivers, lakes, described vegetation and soil tonal pattern may also be subtle
ponds, and springs, along with poorly drained ground and indications of lineaments. Greener areas or early growth may
flood zones. be an indication of increased moisture content. Sometimes a
136 14 Aerial Photography and Remote Sensing Data

Fig. 14.3 Fracture pattern map from Florida (Vernon 1951)

linear trend of tree and vegetation growth may be an indicator • Overhead telephone and power lines
of a fracture or fault due to the presence of sediment infilling, • Buried pipelines
weathering and recharge (Norman 1968). A linear trend may
also be caused by cultural features or man’s activity such as While some obvious fracture traces are easily identified,
utilities, roads, and cattle trails. Caution must be used in the only experienced photo-interpreters can aid in characterizing
interpretation of lineaments and they must be field checked to more subtle ones. Detailed systematic examination of stereo-
verify their presence and that they are not related to other pairs is beyond the skill of most geotechnical engineers and
non-geologic causes (Allum 1966). engineering geologists (Clayton et al. 1982, 1995). While
Parizek (1988) lists features which could be mistaken for good use can be made of aerial photos by the inexperienced,
geologic conditions, some of which are cited below: the specialist aerial photo interpreter is necessary to carry out
a detailed analysis.
• Game and livestock trails Photo analysis is normally done on land, but it also has
• Old fence lines been used in shallow water covered areas often up to 10 m
• Field boundaries deep. Finkl has carried out aerial photo analysis of near shore
• Old abandoned railway right of ways conditions and reefs in southeast Florida using experimental
• Old roads and foot paths water penetrating film and Lidar (Finkl and Warner 2005;
• Property lines consisting of trees Finkl et al. 2005).
• Old cultivated fields Photo analysis was carried out between the Florida
• Old wagon and military trails mainland and the upper-most Florida Keys (Key Largo)
14.4 Aerial Photos 137

• An oblique photo includes a greater area when taken from

the same altitude as a vertical photo;
• The view is more natural making features more
recognizable.
• Some objects not visible on vertical photos may be seen
in oblique photos (for example caves or objects concealed
by forest cover).

An oblique historical photo from a superfund site

(Fig. 14.6) shows the location and type of operations at the
site including the large plant facilities, the supporting rail-
way facility, the nearby river and stockpiles of raw materials.
When work started at this site, almost the entire facility has
been decommissioned and removed with only a few remain-
ing buildings. While old site drawings showed the footprint
of the facilities, the oblique aerial photo effectively illus-
trates the magnitude of the operation. This photo also illus-
trated the large piles of materials and railroad track that were
present during operation from the 1940s to 1980s. These fea-
tures would have provided heavy loading and vibrations over
Fig. 14.4 Aerial showing fracture patterns and landforms in southwest a portion of what was later found to be a large zone of paleo-
Texas collapse. The presence of the heavy loads and vibration for
40 years, over this area supported the conclusion that this
where a new bridge was to be constructed in an area that karst feature would be stable if left untouched. Chapter 27
was covered by mangroves and shallow water. This was an provides more detail on this case study.
unusual area to attempt aerial photo interpretation. Two Advancements in aerial photography over the years
independent aerial photo analyses were carried out on the includes going from black and white to color and from ana-
same set of black and white stereo photos. An experienced logue to digital. More recently a method of acquiring low
aerial photo interpreter (Finkl 1994) completed one analy- altitude high-resolution views from multiple oblique angles
sis using a simple stereoscope. Another independent analy- has been developed and is called Pictometry®. This is a pat-
sis was completed with the same photos scanned into a ented process (Pictometry International Corporation 2002),
computer using pattern recognition and edge detection soft- which includes overlapping images (12–20) collected from
ware. It is interesting to note that both independent inter- many directions for any given point on the ground. This
pretations came up with essentially the same results. A “bird’s eye view” allows a more intuitive interpretation when
number of linear trends were identified including a linear different angles of the same feature are reviewed (Alexander
pattern cutting through the mangroves and portions of the et al. 2013).
water-covered area roughly perpendicular to the highway
and centered within the area of interest. By itself this fea-
ture may not have had much credibility, but as other inde- 14.4.3 Aerial Photos and Video from Small
pendent data became available, all data indicated an Unmanned Aircraft
anomalous paleokarst collapse centered on the linear trend
identified by aerial photo interpretation. This is an excellent In the past decade, or so, interest has focused upon the use of
example of the correlation of multiple data sets. Chapter 26 small unmanned aircraft (model airplanes, helicopters and
provides more detail on this case study. drones) to obtain both aerial photography and video. Photos
and video are obtained digitally and, in some cases, are trans-
mitted to the ground real-time as the Unmanned Aerial System
14.4.2 Oblique Aerial Photos (UAS) fly’s its mission. As an example, the Colorado DOT
has used small-model, fixed-wing aircraft and helicopters to
Oblique photos can provide additional perspective and provide a wide range of low altitude, high-resolution photo-
information about site conditions and may be better graphs for both planning construction and monitoring high-
understood by lay people during presentations. Paine and way conditions (Hotchkiss 2007). CDOT’s UAS fleet is
Kiser (2012) discuss oblique aerial photos and point out comprised of three types of remote controlled aircraft includes
benefits: a pusher-type airplane, helicopters and a “hexa-copter”
138 14 Aerial Photography and Remote Sensing Data

Fig. 14.5 Linear trends associated with in sinkholes along a portion US19 in west-central Florida
14.4 Aerial Photos 139

Fig. 14.6 An oblique photo shows many of the site features not easily understood in a vertical photo. This historical photo illustrates the magni-
tude of site activity and diverse uses. Most of these structures had been removed before the field investigation began

Fig. 14.7 Unmanned aerial systems (UAS), fixed-wing aircraft, a helicopter and hexa-copter used to obtain aerial photos and video (Photos
courtesy of the Colorado Department of Transportation)

(Fig. 14.7). The UAS’s have two cameras; one mounted for a These types of small aircraft were used as part of the
pilots view and another one to take the photographs. Photos inspection of a sinkhole collapse at the Corvette Museum
can be vertical to almost any oblique angle needed for the located in Bowling Green, Kentucky. The sinkhole formed
project. under a portion of the museum housing vintage corvettes.
CDOT’s Aerial Reconnaissance Program has and can be The collapse was recorded by building security cameras and
used for project documentation, rockfall and slope stability eight of the corvettes ultimately fell into the sinkhole. The
analysis, landslide analysis, environmental review, archeo- collapse feature was estimated to be 12 m across and about
logical sites, and resource inventory. The general use of the 10 m deep (The National Corvette Museum 2014). Video
UAS’s are applied to inaccessible areas such as canyons, inspection using a drone helicopter was carried out within
where manned aircraft are unable to fly or obtain photographs the collapse by the University of Western Kentucky engi-
at the detail required. neering department (Wall Street Journal 2014).
140 14 Aerial Photography and Remote Sensing Data

These small aircraft provide high resolution, low altitude applied to karst terrain for regional classifications,
photographic documentation at a reasonable cost as com- investigations and monitoring. LiDAR can be used night
pared to manned aircraft. This approach is also easily mobi- or day and penetrates cloud cover, tree cover and
lized with reliable equipment, a quick access to final photos vegetation (Stennett and Grusky 2008; Haneberg 2008).
and flight access to populated areas or small partially
enclosed areas where standard aircraft would have limited or
no access. Disadvantages include weather restrictions, the
same as with standard aircraft, limited visibility due to pre- 14.5.1 Aerial Thermography
cipitation or fog and photos are not easily geo-rectified.
When developing an inventory of site features for watershed
studies in karst terrain locating the various swallets, springs,
14.5 Beyond Black and White Aerial Photos (recharge and discharge) areas can be tedious. Aerial ther-
(Other Formats and Methods) mography can detect variations in temperature as small as
0.05°, allows coverage over large areas and can minimize
While black and white aerial photos will often provide most laborious field efforts to identify these features on the
of the visual information needed for both the regional and ground.
site-specific scale site characterization there will be occa- Aerial thermography surveys have been conducted to
sions where other types of data may be needed. There is a locate sources of springs discharging water on the bottom
wide range of other imaging and remote sensing measure- of lakes, creeks and rivers. Aerial thermography surveys
ments available (Lillesand et al. 2003; Paine and Kiser 2012). provided a means of locating many unknown springs along
Those imaging techniques that are most commonly used for the length of the river system in central Florida (Davis
karst projects include: 2007). The locations of the larger springs were easily estab-
lished, but many of the smaller features had not been
• Color photos can provide additional information of par- located. Figure 14.8 shows a portion of a thermography sur-
ticular use in geologic interpretation. The human eye is vey completed at Silver Springs as part of the Florida
capable of resolving about 100 times more color ranges Springs Initiative. Data was collected with a 2-m pixel reso-
than gray scale values. While color photos are also lim- lution and entered into a GIS system. The thermal anoma-
ited by sun angle and cloud cover they provide additional lies identified by aerial thermography surveys were
details of vegetation and wet areas (Lillesand et al. 2003). ground-truthed in the field by divers. Three main discharge
• Infra-red (false color) film is used in forestry, agriculture points were identified along this section. The thermography
and vegetation studies. Since water totally absorbs infra- survey successfully identified known springs as well as sev-
red radiation, the use of false color infra-red photography eral new discharges throughout Silver Springs area (Munch
is useful for drainage studies including springs and seeps et al. 2006).
(Clayton et al. 1995). While limited by sun angle and
cloud cover, Infra-red provides additional details of veg-
etation and wet areas. 14.5.2 InSAR
• Thermal imagery allows one to map variations in tem-
perature with or without illumination and is usually flown Interferometric Synthetic Aperature Radar (InSAR) mea-
at low altitudes (few 100 m). Applications range from surements are used to measure deformation of the earths
military use to geologic mapping of vents, springs, seeps, surface from a satellite. By subtracting two scans of the
cave openings, etc. on land or water as long as a tempera- same area at different times, it is possible to make precision
ture contrast exists. measurements of changes in elevation. Under ideal condi-
• Synthetic aperature radar interferometry (inSAR) tions it is possible to resolve changes in elevation on the
provides ground deformation data with centimeter preci- order of 10 mm or less. InSAR has been used to map land
sion. Newer methods of processing have enabled ground subsidence associated with geothermal fields, oil and gas
movements to be monitored down to 1 mm per year over fields and aquifer system compaction in California, Nevada
wide areas providing ground vulnerability mapping and Texas (Bawden et al. 2003). InSAR is ideally suited for
(Browitt et al. 2007). Such measurements can be used measurements of regional deformation and can be less
night or day and penetrates cloud cover and vegetation. expensive than obtaining sparse point measurements. The
• LiDAR (light detection and ranging) is a remote sens- regional results can be used to optimize the location of
ing (laser) technology that can be used to map topography point measurements or survey lines at a more local scale
at a very high resolution and accuracy, which has been (Bawden et al. 2003).
14.5 Beyond Black and White Aerial Photos (Other Formats and Methods) 141

Fig. 14.8 Thermography data from Silver Springs, Florida showing three main points of discharge within the river (Munch et al. 2006)

14.5.3 LiDAR 80 % tree coverage then only about one in five shots will
make it to the ground. But with repetition rates of 150,000
LiDAR (light detection and ranging) is a remote sensing per second or more (around 10 shots/m2 and sometimes
technology that can be used to map topography at a very high triple that for helicopter work) reasonable surface cover-
accuracy, covering large areas by aircraft. About 150 systems age can be achieved in tree covered areas (Stennett and
exist throughout the world, in 2008 with about half of these Grusky 2008).
active in the private sector. Aircraft mounted lasers record LiDAR data can provide elevation measurements of the
elevations at a rate of 2,000–5,000 pulses per second and ground surface through most tree cover. As a result, detailed
have a vertical precision of 15 cm (Stennett and Grusky surface topography can be obtained which can reveal the
2008). These measurements can be made in difficult areas of presence of sinkholes and depressions within vegetated
access and highly vegetated areas, and have been used to areas, which do not normally show up in aerial photography.
develop detailed bathymetric data in shallow water depths to One of the many applications of this method is to monitor
10 m (Finkl et al. 2005). This technology can provide such areas and map changes in topography over time. It can also
detailed data that a national effort was initiated to establish be used to locate lineaments and possibly some cave
total coverage of the US using LiDAR. entrances. As an example, the use of LiDAR in this process
One of the myths is that Lidar can see through a canopy has indicated a fourfold increase in the number of sinkholes
of tree cover. Reality is that LIDAR makes measurements mapped for Winona County, Minnesota (Rahimi and
only in the openings or gaps in the tree cover. If there is Alexander 2013).
142 14 Aerial Photography and Remote Sensing Data

LiDAR digital elevation models (DEM’s) aren’t restricted Allum JAE (1966) Photogeology and regional mapping. Pergamon
to sun position on a given day, and can be looked at with an Press Ltd, Oxford
Bawden GW et al (2003) Measuring human-induced land subsidence
infinite number of lighting combinations on the computer. from space. U.S. Geological Survey Fact Sheet 069-03
Therefore, LiDAR–based topography maps and low sun Browitt C et al (2007) Terra not so firma. Geoscientist 17,
angle shaded relief images have the potential to highlight 6(22):18–22
features such as subtle areas of subsidence as well as obvious Clayton CRI et al (1982) Site investigation, a handbook for engineers.
Halsted Press, New York
sinkholes. As with all methods there are inherent limitations. Clayton CRI et al (1995) Site investigations, a handbook for engineers,
Haneberg (2008) cites some of the practical limitations and 2nd edn. Blackwell Science, Cambridge
discusses the accuracies that can be obtained. Davis J (2007) Springs eternal. Prof Surveyor Mag 27(9):8–12
In the best cases, LiDAR can produce gridded DEM’s Doctor D, Young J (2013) An evaluation of automated GIS tools for
delineating karst sinkholes and closed depressions from 1-m
with horizontal spacing on the order of meters (roughly an LiDAR-derived digital elevation data. In: Land L, Doctor DH,
order of magnitude better than conventional USGS 10 m or Stephenson BJ (eds) Proceedings of the 13th multidisciplinary con-
30 m DEM’s) and a nominal vertical accuracy on the order ference on sinkholes and the engineering and environmental impacts
of centimeters. Some LiDAR providers are claiming 7 and of karst, National Cave and Karst Research Institute, Symposium 2,
6–10 May, Carlsbad, pp 449–458
8 cm, more typically only 15 cm is reasonable (Stennett and Finkl C (1994) Report on FDOT Jewfish Creek Bridge. Consultants
Grusky 2008). report, The Coastal Education & Research Foundation, CERF
However, LiDAR can produce DEM’s and shaded relief Remote Sensing Lab, Ft. Lauderdale
images that are difficult to interpret, filled with ambiguous Finkl CW, Warner MT (2005) Morphologic features and morphody-
namic zones along the inner continental shelf of southeastern
data, and processing artifacts. Depending upon the area of Florida: an example of form and process controlled by lithology. J
interest and type of karst terrain the non-karst features may Coast Res Spec Issue 42:79–96
include old building foundations, small quarry pits, cattle Finkl CW et al (2005) Interpretation of seabed geomorphology based
wallows, tree tip-up and in areas of past glaciations ice block on spatial analysis of high-density airborne laser bathymetry (ALB).
J Coast Res 21(3):501–514
melting processes (Alexander et al. 2013). Even with these Gentile RJ (1984) Paleocollapse structures: longview region, Kansas
limitations, this technology can provide an amazing amount City, Missouri. Bull Assoc Eng Geol 21(2):229–247
of detail over large areas, in difficult terrain and with vege- Haneberg W (2008) Elevation errors in a LiDAR digital elevation
tated cover. So while LiDAR is widely being used to map model of west Seattle and their effects on slope stability calcula-
tions. In: Baum RL et al (eds) Landslides and engineering geology
sinkholes and other closed-depressions, standardization for of greater Seattle area, Washington, Geological Society of America,
its use need to be developed so that databases from different Reviews in Engineering Geology, 20, pp 55–65
interpreters or regions can be compared and subjectivity can Hotchkiss A (2007) Low altitude large scale reconnaissance or is this a
be minimized (Doctor and Young 2013). great job or what? 58th annual highway geology conference,
Pennsylvania 16–18 October 2007, pp 60–71
A recent article in Arcnews, published by ESRI discussed Johnson AI, Pettersson CB (1988) Geotechnical applications of remote
efforts of the US Geological Survey who is taking the lead in data transmission. ATSM STP967-EB. American Society for Testing
the nation-wide 3D elevation program (Sugarbaker 2014). Materials, Philadelphia
Both LiDAR and InSAR will be used to acquire three- Lattman LH (1958) Technique of mapping geologic fracture traces and
lineaments on aerial photographs. Photogramm Eng
dimensional elevation data. The USGS has developed the 24(4):568–576
acquisition criteria to be met in order to maintain a uniform, Lattman LH, Parizek RR (1964) Relationship between fracture traces
quality database. These efforts are being met with a network and the occurrence of ground water in carbonate rocks. J Hydrol
of federal, state and tribal partnerships in order to create a 2:73–91
Lillesand TM, Kiefer RW (1994) Remote sensing and image interpreta-
database that is available for use by a variety of government tion, 3rd edn. Wiley, New York
and private industries. For Lidar and InSAR data availability Lillesand TM et al (2003) Remote sensing and image interpretation, 5th
go to www.coast.noaa.gov/inventory. edn. Wiley, New York
Littlefield JR et al (1984) Relationship of modern sinkhole develop-
ment to large scale-photolinear features. In: Beck BF (ed)
Proceedings of the 1st multidisciplinary conference on sinkholes:
References their geology, engineering and environmental impact, Orlando,
Florida. Balkema, Rotterdam, pp 189–195
Alexander S et al (2013) Combining LiDAR, aerial photography and Munch D et al (2006) Fifty-year retrospective study of the ecology of
pictometry® tools for karst features database management. In: Land Silver Springs, Florida. St. Johns River Water Management District;
L, Doctor DH, Stephenson BJ (eds) Proceedings of the 13th multi- University of Florida – Department of Fisheries and Aquatic
disciplinary conference on sinkholes and the engineering and envi- Sciences; and Wetland Solutions, Inc. prepared for Florida
ronmental impacts of karst, National Cave and Karst Research Department of Environmental Protection as part of the Florida
Institute, Symposium 2, 6–10 May, Carlsbad, pp 441–448 Springs Initiative
References 143

Norman JW (1968) Photogeology of linear features in areas covered with ciplinary conference on sinkholes and the engineering and environ-
superficial deposits. Trans Inst Min Metall Appl Earth Sci 77B:60–77 mental impacts of karst, National Cave and Karst Research Institute,
Paine DP, Kiser JD (2012) Aerial photography and image interpreta- Symposium 2, pp 469–480
tion, 3rd edn. Wiley, New York Stennett T, Grusky SW (2008) Lidar: fact or fiction. Professional
Panno SV, Luman DE (2013) Mapping palimpsest karst features on the Surveyor Magazine, April 28(4):16–20
Illinois sinkhole plane using historical aerial photography. Sugarbaker L (2014) US Geological Survey to lead ambitious 3D eleva-
Carbonates Evaporites 28(1–2):201–214 tion program. ArcNews published by ESRI, spring p8 (www.csc.
Parizek R (1988) Lineament analysis: how and why. Engineering & noaa.gov/inventory)
Geology of Karst Terranes, A short course sponsored by the Florida The National Corvette Museum (2014) National corvette museum
Sinkhole Research Institute, Orlando media alert sinkhole collapse in skydome, www.corvettemuseum.
Pictometry International Corporation (2002) US Patent Ser. No. org. Accessed 13 Mar 2014
60/425,275 filed November 8, 2002 Vernon R (1951) Geology of citrus and levy counties, Florida. Florida
Rahimi M, Alexander EC Jr (2013) Locating sinkholes in LiDAR Geologic Survey, Bulletin 33, FGS, Tallahassee, Florida
coverage of a glacio-fluvial karst, Winona County, MN. In: Land L, Wall Street Journal (2014) Drone footage from inside Corvette Museum
Doctor DH, Stephenson BJ (eds) Proceedings of the 13th multidis- Sinkhole. Youtube, February 12, 2014. Accessed 26 Mar 2014
 Site Walkover
15

Abstract
While a site visit may have taken place during the contracting phase of work, these are rapid
tours of a site with little time to stop and observe or take detailed notes. The site walkover
is when we get our first chance to make good observations and an initial assessment of
general site conditions. It provides an opportunity to evaluate and expand our preliminary
understanding of site conditions obtained from the desk study and the review of aerial pho-
tos. This initial effort should allow adequate time for the site to be covered. The observa-
tions made and information gathered from a site walkover should include geologic and
hydrologic observations, inventory of karst features, indications of subsidence or sinkhole
activity as well as an assessment of cultural features that may impact access, safety or a
particular method of measurement used on site.

While the site walkover has many objectives, it should and buried utilities. In addition, site security, along with
progress from a reconnaissance overview to obtain first health and safety aspects can be evaluated.
impressions and then begin to gradually focus upon the site- The objective of the initial site walkover is to leave the
specific details. On-site observations must not be limited to a field with sufficient notes, sketches, and photos to reasonably
single site walkover, but as the field effort proceeds, addi- document surface conditions, which may be of interest for
tional site walkovers and observations should be incorpo- further detailed investigation, or may impact the field efforts.
rated throughout the site characterization process. While the focus is upon the site itself, the surrounding area
should also be inspected to the extent needed and where
access can be obtained.
15.1 The Initial Site Walkover

During the initial site walkover we begin to test the concepts 15.2 Importance of Observations
in our preliminary conceptual model against actual site con-
ditions. It is our first close look at the geologic, hydrologic Observations are one of the most important tools we have in
and karst aspects of the site and the surrounding area. The our tool box and the primary tool used during the site walk-
walkover also provides an opportunity to obtain site-specific over. It is here where experience and judgment aided by
details not previously identified during the desk study, on the skilled observations can make such great contributions.
topographic maps, or the aerial photos. Finally we need to However, some individuals are better observers than others.
identify any site constraints that will limit access or impact For this reason the site walkover field team should consist of
safety of our future field operations. These may include sim- at least two persons, that way, observations and ideas can be
ple access for personnel or heavy equipment. At the same shared and discussed on the spot. “Quite often the most
time we can identify any obstacles for the field investigation, essential services I rendered to my clients had no relation-
including topography, streams, rivers and lakes, as well as ship to the original assignment. They grew out of casual
dense vegetation, fences, structures and both above ground observations I made while inspecting the site” (Terzaghi

© Springer Science+Business Media Dordrecht 2016 145

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_15
146 15 Site Walkover

1957). The authors have also found that observations have seen on our initial visit to an earthen water retention dam
been key to solving numerous engineering geologic puzzles located near Phoenix, Arizona. Small desiccation-like
over the years. cracks were noted perpendicular to the axis of the dam.
Unfortunately field observational skills are becoming a They were subtle and would have been missed unless atten-
lost art and are not easily taught. An early lesson in observa- tion was focused upon surface details. Surveyor flags were
tion was taught to the senior author in an undergraduate geo- then placed at each crack over a distance of approximately
logic field trip to northern Minnesota. As the group gathered 120 m and then the spacing between them was measured. Of
to look at a fault, the professor asked Benson if he would like the 20 survey flags, there spacing was found to be
to comment upon the fault they were here to observe. After a 5.8 m ± 1 m. This suggested that the mass of the earthen
pause and a blank look on Benson’s face the professor sug- material making up the dam was fracturing (due to settle-
gested that the group should notice that Benson had one foot ment and or desiccation) at an interval related to the mass of
on either side of the fault. That was a good early lesson in the material properties.
observation and attention to details. The importance of observation cannot be over empha-
Many karst or pseudokarst features will not be recogniz- sized. “If you do not know what you should be looking for in
able on a topographic map or aerial photos and may only a site investigation, you are not likely to find much of value”
become apparent during the site walkover. This may be due (Fookes 1997). “Only with experience and attention to detail
to their size or being masked by vegetation or cultural fea- can the relative importance and significant of these features
tures. In some cases, there is only subtle evidence of subsid- be interpreted in the field” (Clayton et al. 1982, 1995).
ence or sinkholes present at the surface. Therefore, it will be Observational skills may be the best tool we have available
necessary to locate, map, and evaluate karst features with a to us in the field.
site walkover. For example, of the 535 sinkholes in Winona
County, Minnesota 85 % were not shown on the 7.5-min
topographic maps and were only found by fieldwork, 15.3 Some Tools for the Field
(Dangleish and Alexander 1984). Fewer than 5 % of the
springs in the Mammoth Cave area of Kentucky are shown There are obvious tools such as topographic maps, aerial
on 7.5-min topography maps. The remaining 95 % were photos, rock hammer, a field notebook, camera and a com-
found by fieldwork (Quinlan 1989, 1990). All of these details pass. Limited sampling of soil and rock may be obtained for
provided important clues as to the nature of the site initial analysis and reference. On a new site development, the
conditions. initial area in which a project is to be located can be many
There are many subtle details that can be obtained from km2. While a site walkover is done mostly on foot, larger
careful observations yet can easily be overlooked or simply sites may require an ATV (Fig. 15.1a) or a 4 × 4 pickup truck
dismissed as irrelevant. An example of subtle features was as a means of transportation. A GPS unit and topographic

Fig. 15.1 An ATV is used to cover a large site (a) and field mapping has been aided by the use of GPS and digital field notebooks (b)
15.3 Some Tools for the Field 147

sheets are a must in wide ranging reconnaissance of of excellent examples). The art of field sketching has largely
100’s–1,000’s of hectares particularly over areas with few been academically “pushed off-screen” by the encroachment
landmark or reference points. With the wide-spread use of of other trendy and “more modern” technologies. Few engi-
GPS one might think that pacing would be a lost art, how- neers or geologists today are taught topographic surveying,
ever, pacing remains effective for field reconnaissance in and field sketching is largely neglected in class or camps
wooded areas, underground or in structures. Distances by teaching field geologic mapping. Field sketches can be an
pacing can be better than 2 % on flat smooth surfaces, but important part of the site walkover. The very act of sketching
will be less in rough, hilly, or vegetated terrain. Field map- forces you to stop and observe what site conditions are while
ping can be improved by the use of a GPS system, hand-held they are before your very eyes (Hatheway 1998) (Fig. 15.2).
computers and digital topographic maps and aerial photos Sketches should be made showing features at the site, which
(Fig. 15.1b). are not shown on existing topographic sheets or aerial pho-
Sketches were once standard for professional fieldwork tos. While on-site photographs are a necessary component of
and were well done by many (See Fookes 1997 for a number the site characterization effort, the process of taking a picture

Fig. 15.2 An excavation has exposed a cross section of geology indicating the edge of a paleocollapse. It was photographed (a), sketched (b) and
elevations were obtained along the top of a shale layer to document conditions
148 15 Site Walkover

does not force one to take the time to see the details that out the entire extent of trench from well north of the Love
sketching does. It is often the little details that are most Canal site to the south of it. The exception according to the
important. contractor was the “underground river” he had encountered
While today’s field staff is probably less talented in terms approximately due east at the north end of the site with water
of sketching ability, (that includes the senior author), a reason- flowing away from the Love Canal area. Upon returning to
able alternative to detailed sketching is the use of photo docu- the project office, the team was chastised by the project man-
mentation. Almost anyone can take reasonable good photos in ager for going beyond the site (actually into the next county
the field that can be checked immediately with today’s auto- just a few blocks away). To this manager, the only thing of
mated digital cameras with a wide angle to telephoto lens. interest was the Canal property itself and the regional geo-
However, good observational skills are still necessary in order logic setting was of no interest. Unfortunately this type of
to acquire photo documentation that is useful. Field notes need thinking, or more correctly a lack of understanding of the site
to be taken for each photo, including the direction of the photo characterization process, is all too common.
and a simple sketch to note the key elements of interest. That
way, when the dozens or even hundreds of photos are being
reviewed, at a later time, their location and the points of inter- 15.5 Site Coverage
est can easily be identified and reconstructed.
The initial site walkover is typically at a reconnaissance
effort early in the program and over large areas. However,
15.4 On-Site Walkovers and Off-Site Drives subsequent efforts can approach 100 % site coverage over
key portions of the site, later in the site characterization
The walkover must include the site itself as well as the sur- effort. This may be accomplished gradually as work pro-
rounding area, which may require reconnaissance over many gresses at the site or as a single focused field effort. Either
square kilometers. One should cover enough of the surround- way it is essential that the observations and notes made get
ing area to minimize future surprises, this will usually require summarized and added to the project map or database. The
driving. The area that needs to be covered and detail that is focus and detail that is required will depend upon specific
required will depend upon specific project objectives. In gen- project objectives. For example, a site to be used for an engi-
eral, the further away from the site these features are located, neered structure will be interested in site-specific conditions
the less important the feature would be. However, if a site is that would impact structural integrity. If a site is being evalu-
being evaluated from a hydrologic point of view, particularly ated from a hydrologic point of view, the focus would typi-
if it involved contaminants, the karst inventory would typically cally be basin-wide.
be basin-wide. For these projects, even small features that are
well connected may have an impact at great distances.
Artificial boundary limits are often placed on the site 15.6 Observations and Mapping
characterization process. It is not uncommon for owners to
suggest that work only be carried out within the immediate Geologic and hydrologic mapping can be optimized using
confines of the site (this commonly occurs on sites with the information from the desk study and aerial photos.
environmental contaminant issues). The issue of limiting off- General conditions for a site should be established in a pre-
site access is often driven by legal concerns. This can be an liminary conceptual model at this point, what type of condi-
awkward situation since quite often; critical elements of tions might be expected and therefore what to look for. The
geology are commonly discovered off-site. While the major- initial site walkover would typically only confirm locations
ity of the detailed investigation work will be done within the and document essential information. A separate and focused
site boundaries, it is a serious error to blindly ignore the field effort may be required for specific geologic field map-
regional and local setting. ping in areas with complex geomorphology or when a
A classic example of this occurred during our work at detailed inventory of karst features is necessary.
Love Canal in the early 1980s. The health and safety officer
had not yet shown up at the site and our team’s access was
denied prior to a health and safety briefing. The time was 15.6.1 Geologic Observations
spent driving around the site to familiarize us with the area.
A new sewer line was being installed by a local contractor, Field mapping would be focused upon finding and utilizing
approximately 600 m to the east and parallel to the Love geologic windows or rock exposures in road cuts, rivers,
Canal site. What was of interest was the open trench which creeks, quarries or possible local excavations. Inspection of
exposed a massive clay (a clean clay with the consistency of rock exposures is a key aid in estimating the fabric and char-
modeling clay) which the contractor said extended through- acter of the rock including: joint spacing, and aperture as
15.6 Observations and Mapping 149

Fig. 15.3 Three road cuts illustrate the valuable geologic information available from these geologic windows. (a) Southwest Texas. (b) Puerto
Rico. (c) Kansas City, Kansas

well as degree of weathering and presence of karst features. be adjacent to the site or represent the region surrounding the
In some cases, open fractures or cavities can be found in site. The information provided needs to be put into context
other cases they may be sediment filled. These windows into with relation to the site under investigation so that informa-
the shallow subsurface often provide excellent information. tion is not over extrapolated.
Road cuts are great opportunities to observe rock charac- A road cut along Highway 90 in southwest Texas
teristics that are otherwise not accessible. The road cuts may (Fig. 15.3a) shows two smaller areas of in-filled collapse.
150 15 Site Walkover

side of the quarry at a depth of approximately 36 m below

grade. The largest had a diameter of about 1 m, and the other
two about 0.3 m in diameter (Fig. 15.5b). An area of spring
flow was observed near the upgradient base of the quarry at
its north end (Fig. 15.5c). The cross section of the springs
was approximately 1 m2. This is an example of the relatively
low density of cavity systems within the 223,000 m2 quarry
wall. This illustrates the density of cave systems and the dif-
ficulty of finding them.

15.6.2 Hydrologic Observations

All significant surface hydrologic features should have been

identified during the desk study and they should be verified
during the site walkover. However, many smaller features,
not readily identifiable on aerial photos or topographic maps
such as small ponds, creeks, sinks, springs and seeps, may be
found during the site walkover. The location, elevation and
estimated flow rate need to be added to the site map and data-
base along with the date of the observation and a photograph
Fig. 15.4 The floor of the Missouri Portland Cement Quarry in taken to document the feature. Further data may include the
Independence, Missouri reveals the extent and spacing of fractures in temperature and conductivity data for these waters. This is
the Bethaney Falls Limestone
particularly important, if a dye trace study is part of the pro-
gram since it would help to identify possible injection and
These features are relatively close to one another and they sampling points.
extend from the surface to the roadway indicating that the Streams sometimes become sinking streams and disap-
feature extends deeper well below the road. A road cut in pear, recharging the groundwater. As you follow the stream-
Puerto Rico (Fig. 15.3b) shows a highly weathered and irreg- bed water may reappear, discharging to the surface as a
ular limestone profile. In the Kansas City area, the upper- spring further downstream. The location of a sink or spring
most Argentine Limestone (Fig. 15.3c) is commonly exposed may also move over time as water levels change and systems
along road cuts showing extensive fractures and weathering. become plugged with sediments and debris. In addition,
Quarries are another window of opportunity to observe there are some hydrologically significant features that may
rock characteristics. But because quarries are selected to be only reveal themselves during heavy rainstorm events. Since
in the best rock, they are not necessarily the best places to fieldwork is typically avoided during a heavy rainstorm, a
observe karst. However, when a quarry is available it must be special walkover may be required during such events.
included in the site walkover and offers an opportunity to Temporal changes in hydrologic conditions may occur
observe a major cross section of geologic conditions. The over a short period of time or a longer period of time. Figure
quarry geologist or engineers are usually pleased to show 12.13a shows a small creek adjacent a site under investiga-
you around and can usually provide a wealth of tion. A rainstorm during fieldwork rapidly changed condi-
information. tions over a period of a few hours (Fig. 12.13b). Figure 15.6a
A limestone quarry floor adjacent to a site under investi- is taken from the dry creek bed in the Peace River in west-
gation in Missouri revealed the extent and spacing of vertical central Florida showing the sign for canoeists in wetter sea-
fractures in the Bethaney Falls Limestone (Fig. 15.4). This sons. This dry season occurred over a period of months.
plan view of the rock fracturing was unique and would not Open joints are now revealed as a result of the dry season
have been as readily perceived from any other data or (Fig. 15.6b). Knowing about such unique temporal condi-
observations. tions can provide insight that would otherwise be missed.
A large limestone quarry very near a project site in
Fredrick, Maryland had a perimeter of approximately 15.6.2.1 Flow of Air from Fractures, Caves, Mines,
2,925 m and a maximum depth of approximately 76 m. and Boreholes
Figure 15.5a shows a portion of the western quarry wall. Cool air flowing from the entrance of caves, mines and their
This quarry provided an excellent geologic window. ventilation shafts is common. In addition, air flowing from
However, only three small cavities were seen on the western exposed fractures in rock as well as from well casings whose
15.6 Observations and Mapping 151

Fig. 15.5 A section of the westernmost quarry wall (a), the three cavities identified high on the western wall (b) and the springs (c) located low
on the north wall

screen or open hole has extended into the unsaturated zone. • Bernoulli effect at a tee shape junction between higher
Air is about 1,000 times lighter than water and air currents and lower entrances
require much less energy and have greater diffusion than • Humidity effect due to differences in air density between
water. As a result flowing air can be used as a very sensitive moist and dry cave air, and
tracer (Milanovic 2004). Vertical permeability based upon • Venturi effect of surface winds drawing air out of a cave
air flow in the unsaturated soils has been described by Weeks entrance
(1978, 1987).
Cavers will often follow flowing air within a cave and Some modeling of the fluid-thermal dynamics of air flow
have reported air flow from fissures on the surface which has in caves has been done to simulate two-dimensional air flow
lead to cave systems. Air flow from caves and mines may be in Carlsbad Cavern, NM. This work focuses on buoyancy
constant in one direction or may reverse flow. Lewis (1991, and natural convection due to geothermal heating (Boston
1992, 1996) proposes a number of factors that may cause and et al. 2005).
control air flow from caves including: The flow of cool air from well casings and fractures in
rock exposed at the surface are indicators of interconnection
• Changes in barometric pressure within fractured rock or possibly a cave in the subsurface.
• Surface winds
• A water siphon effect from fast turbulent flow • The authors noted a significant air flow from the casing of
• A water fall in a narrow shaft (water jet pump) a windmill well in southwest Texas which was located in
152 15 Site Walkover

Fig. 15.6 The Peace River in west-central Florida is normally a small quiet stream used by canoeists. However, during periods of low rainfall the river
will dry up completely (a), note the sign on the cypress tree indicating direction for canoeists. When dry, open joints in the rock are exposed (b)

limestone where the water table was about 150–180 m extended about 450 m south of the blast site. A monitor
below grade. Cool air flowed freely from the well casing well about 90 m south of the rim fracture had an obvious
indicating that the borehole had encountered open frac- and continuous air flow from it. Measurements of the per-
tures in the rock, (or possibly a cave system). Other obser- meability have been made at several nuclear chimneys at
vations in the area revealed obvious air flow from fractures the Nevada Test Site (Rosza et al. 1975)
in the surface rock. These were good indicators of the • Thermal imagery of borings used for backfilling of a mine
interconnections of fractures within the rock mass and or in the Kansas City area would quickly indicate which
the possibility of the presence of a cave. Measurements of boreholes had encountered an open mine since cool air
temperature can be made to support flow measurements. would commonly be flowing from them.
• At the Nevada Test Site “Faultless” an underground
atomic bomb of intermediate size (200 k ton–1 M ton) These are examples where attention to detail can pro-
was detonated some 900 m below grade in volcanic tuff. vide additional data regarding the interconnections of frac-
The water table was about 90 m below grade. A signifi- tures within the rock mass and or the possible presence of
cant rim fracture with a displacement of about 3 m a cave or mine.
15.6 Observations and Mapping 153

15.6.3 Inventory of Karst Features • Provide an indication of subsurface activity (or lack of
activity) in an area; and
An inventory of karst features both on and off-site is standard • Establish injection and sampling points for a dye trace
in any site characterization in a karst terrane. This inventory study.
includes possible features identified from Tables 13.1, 13.2,
and 13.3 in the desk study. This effort usually starts at the When sinkholes, sinks, springs or caves are encountered,
regional setting and then moves to the site-specific setting. A their location should be identified on a topographic map,
list of existing or possible karst features would be developed aerial photo, or site map. In some cases, where access is safe
from topographic maps, aerial photos, regional literature, additional insight can be gained by entry. Small caves can
databases and even site-specific reports. During the site often be found that may or may not have been previously
walkover these features should be confirmed and additional identified or mapped.
information on them gathered. The purpose of the karst At a proposed federal prison facility in Virginia, a small
inventory may: cave was encountered during the site walkover (Fig. 15.7a).
This large property had not been previously developed or
• Provide an idea of what type of karst features are present surveyed so there was no existing information on this cave.
in and around a site; Documentation included basic dimensions and geologic
• Reveal patterns of occurrence; formation along with evidence of collapse or presence of

Fig. 15.7 Caves can be encountered during a site walkover. Measurements of a cave were made to establish the cave’s height and span in West
Virginia (a). A small collapse was found at the edge of a house. A quick inspection indicated that this was a small cave (b)
154 15 Site Walkover

flowing water. Although this cave was not within the footprint These features were abandoned and their flags removed. Of
of development, it provided an indication of the type of karst the remaining 42 features, fissures dominated the type of fea-
feature that may be presence elsewhere on the property. ture found and they extended about a meter below grade
While the site walkover starts with an indication of what based upon shallow excavations. Figure 15.8 illustrates some
to expect in the field, an open mind needs to be maintained. of the typical features identified.
A project along the southeast coast of Florida is not neces- When these features were plotted on a site map they were
sarily where you would expect to find a cave. However, a clustered in the southeast portion of the site and indicated a
small cave was found extending under a house (Fig. 15.7b) clear northeast-southwest orientation. The orientation of the
after a collapse occurred as a result of a rain down spout features identified in the site walkover and the aerial photo
concentrating water at the edge of the house. The small cave analysis matched the orientation of the many large paleo-
extended under portions of the house, the swimming pool karst collapse features identified by Freeman (1968) a few
and much of the yard. kilometers south of the site (Fig. 15.9).
An example of a unique karst inventory comes from a One could have easily dismissed these features as local
proposed hazardous waste site in southwest Texas. The ini- desiccation cracks. However, their correlation with the aerial
tial site covered an area of 15 km2, designated by the owner. photos features and alignment with the large paleocollapse
The final area of proposed construction was ultimately nar- features mapped by USGS some 4.8 km to the south indi-
rowed down to 252 ha as the site characterization progressed. cated an increased significance, which were incorporated
However, the reconnaissance walkover (and in this case drive into the conceptual model of the site.
over) was carried out over 165 km2 and focused upon a
known caves, paleokarst collapse features, springs, and road
cuts in the area. 15.6.4 Indications of Subsidence and Sinkhole
Later in the site characterization effort, Jim Quinlan, a Activity
member of the advisory committee for the site, requested a
very detailed site walkover. Based upon the type of facility There are many indicators of subsidence and sinkhole activ-
that was planned and the area of karst in which it was to be ity that occur at the ground surface that can be obtained from
located, a 100 % visual site walkover was completed. This careful observations. Many of these indicators can be quite
effort included the 252 ha as well as a zone of 76 m around obvious while other more subtle indicators can easily be
the entire perimeter. At this point in the project, we knew that overlooked or simply dismissed. Because of the scale of
no obvious karst features were present over the 252 ha. The these features they will not be recognizable on topographic
purpose of the walkover was to identify any subtle signs of maps or aerial photos. Field observations are the only means
subsidence, soil piping, fissures, or depressions at the to locate, map and evaluate these features.
surface. There are a number of precursor indicators of subsidence and
Survey stakes were installed on a 76 by 76 m grid. A team sinkhole development. These indicators can occur naturally or
of 11 people (local high school students) were used to form manifest themselves in man-made structures (Fig. 15.10). The
a line in an east-west direction over the 76 m distance following has been adapted from Newton (1987).
between survey stakes. Prior to the walkover, the team was
briefed and shown examples of surface features they may • Depressions in the soil or pavement commonly resulting
encounter such as localized depressions, soil cracks and fis- in the ponding of water (Fig. 15.10a).
sures and animal borrows. Each person was responsible for • Cracks, fissures in the soil, pavement or concrete
observing the ground surface directly in front of them and (Fig. 15.10b, c).
about 4.5 m to each side. They were instructed to place a • The tilt of trees and fence posts.
surveyor’s flag at any surface feature they observed and if in • Holes in the ground that have formed over short periods
doubt, they were instructed to place a flag. Several profes- of time and which may drain surface water (Fig. 15.10d).
sional staff members walked behind the line to monitor the • Turbidity and mudding of well water or a sudden decline
team’s position, stop the forward movement when necessary, of water level in a well.
to answer questions and to provide experienced observa- • Sudden drainage of a surface water impoundment or a
tions. At the end of each 76 m swath the team would stop and new impoundment designed to hold water.
be debriefed before continuing on to the next 76 m section. • The localized presence of vegetation in the middle of an
A total of 128 features were identified during the 2-day otherwise open field (some plants will flourish with an
walkover. Each feature was mapped and assessed to deter- excess of water pounding at the surface).
mine its true nature. Many of the features were found to be • Localized wilting vegetation due to lack of water, (some
old animal burrows or small areas where water would pond. plants will diminish as moisture drains into a sinkhole).
15.6 Observations and Mapping 155

Fig. 15.8 Three examples from a total of 42 localized features and fissure found during a 100 % site coverage walkover. Excavation of these
fissures by shovel and backhoe revealed that they extended at least a meter below grade

The presence of one or more of these indicators does • Changes in windows and doors sticking and not closing
not necessarily confirm the presence of an impending sink- properly
hole, but it does provide a basis for further investigation. • Walls and floors measurably out of level or plumb
Careful observations may provide clues as to subsurface • Settlement or tilting of a portion of the floor slab
activity and build the conceptual model for the site • Drains and utility services interrupted
characterization. • Broken pipes and utilities
When structures are impacted by subsidence or sinkhole
activity, the initial results are often visible cracks within the Besides a sinkhole, other causes of structural damage to
structure walls, walkways or driveways (Fig. 15.10c). General buildings may include shrink swell clays, a foundation built
indicators of foundation movement have been adapted from on organics or trash, mine subsidence, slope instability, soil
Freeman et al. (2002) and are summarized below: erosion or long-term groundwater withdrawal resulting in
subsidence. These should be kept in mind as one assesses
• Isolated cracks at weak points within the structure structures.
• Cracks taper from top to bottom Freeman et al. (2002) and Audell (1996, 2004) provide
• Cracks occur both externally and internally at the same an excellent description of the cracks in structures using a
location Crack Classification System (CCS). While theses three
• Cracks consistent with a pattern of movement publications do not address subsidence due to sinkholes
156 15 Site Walkover

8 km radius

1.6 km radius 0 1825

Scale in Meters

Aerial Photo Lineaments

Site Aerial Photo Anomalies
Boundary USGS Mapped Karst Features

Fig. 15.9 The location of photo lineaments, the aerial photo anomalies and the larger paleocollapse features mapped by USGS just south of the
site all show the same general orientation

they provide useful data for evaluating cracks in structures locators, or ground penetrating radar. In some cases, utility
due to other causes and can be useful in resolving sinkhole drawing as-built plans may be available. However, caution
claims. should be used since errors are common on such “as-built”
drawings. As-built plans often turn out to be as designed
plans and therefore may not contain deviations that occurred
15.6.5 Cultural Factors, Utilities and Other Site during construction or modifications after construction.
Limitations Surrounding cultural features may also impact a site.
These features may include spray fields, water mains, major
During the initial site walkover observations should be made utility corridors, transmission lines, communication towers,
of possible cultural activity, which can impact the logistics, private and local pumping wells, or well fields. Small com-
safety or possible measurements at a site. This might include munity well fields can result in triggering sinkholes under the
remnants of previous building foundations, drainage chan- right conditions. A single small sinkhole developed in the
nels, debris piles, vegetation, fences and ditches. gypsum beds along a roadway in east-central Michigan where
Above ground and below ground utilities are also an open pit gypsum mining had been carried out. The nearest
important feature that could impact site access and measure- abandoned mine was some 460 m from the sinkhole had been
ment locations. Their presence and locations must be inactive for decades. The only recent change to groundwater
identified. Utility locations on public right of ways will be levels had been the new well field that had been installed
identified by the responsible utility company or marked by a between Lake Huron and the roadway (a distance of about
one call service. On private property utility locations must be 210 m) to supply a new housing development. This was likely
obtained by utility location devices such as pipe and cable the change in water level that triggered the sinkhole.
15.6 Observations and Mapping 157

Fig. 15.10 Early subsidence is commonly indicated by the presence of circular depressions that retain water (a), cracks, fractures or local settle-
ment in the soil (b), cracks in asphalt or concrete driveways (c) and localized depressions or drop-outs (d)

15.6.6 Eye Witness and Anecdotal Information road as you approached the existing bridge that repeatedly
had been filled and patched. This subtle piece of information
Discussions with local workers or residences can often be ended up correlating well with other data that indicated a
quite effective to discover unique site conditions that are not paleocollapse at this location (for more details see Chap. 26).
noted or reported elseware including temporal events. Even Verbal information from persons who have worked at the
if no obvious occurrence of sinkhole collapse has taken site or who live nearby can be helpful. These sources of infor-
place, subtle indications may have been witnessed. For mation can often provide insight to local or even site-specific
example, while working along a bridge alignment, we would conditions that may not be available elsewhere. However,
stop at a local shop for lunch. The owner was curious about caution is called for since people’s memory may be inaccu-
what we were doing. Our conversation revealed a spot in the rate and they may report issues in error or with bias.
158 15 Site Walkover

15.7 Fly Over the site walkover. These will provide the first on-site geologic,
hydrologic and karst data to improve the conceptual model.
As part of the site walkover, it is sometimes useful to make a The following is an example of the early evolution of a
site fly over which can provide a rapid means of familiariza- preliminary conceptual model at a mine collapse site. After
tion with the site itself and its relation to the overall setting. completing a few days of site walkover and a few days inside
A fly over is especially useful over large areas and those with of a mine, we were starting to form some ideas and come to
difficult or limited access on the ground. The fly over may be initial conclusions about a number of issues at the Tobin
done before or after the site walkover depending upon access mine site in Kansas City where a 7 ha mine-roof collapse had
and project needs. occurred. The following morning at breakfast Allen
A fly over is commonly completed using a small fixed Hatheway presented me with a hand written note in which he
wing aircraft or helicopter. A slow speed aircraft with an raised a number of questions about our early thoughts and
overhead wing is the best choice for visibility and obtaining opinions we had started to form about site conditions. The
oblique aerial photos. The senior author has used a two seat note was titled: “Dicks Moment of Truth Exercise” and con-
Piper Cub with an overhead wing to provide the best visibility tained the following:
and with the window open to improve photographic quality.
Helicopters are quite maneuverable but are also noisy and 1. We can’t really show there is paleokarst involved. Yet we want
have a fair but of vibration. The senior author has utilized the to employ such to help explain the mine failure phenomenon.
Goodyear blimp a number of times in the 1970s to monitor 2. The literature on subsidence is complex – no clear answers
dye traces from sewerage outfalls in the ocean and to track 3. We don’t know where the mine water is coming from, or
dye traces of tidal flow in coastal waters and inlets. Today going to, and how much is coming and going.
however, we will be seeing more unmanned aerial systems 4. Why can’t the mine collapse propagate all the way to the ground
(UAS) (Fig. 14.7) being used for such reconnaissance. surface through 51 m of limestone and shale, next week?
Fly over’s often proceed at a rapid pace with many dis- 5. What makes you think the surface fissures in the loess are
tractions including communication with the pilot. During the older than the mine?
flight looking down at your maps for a few seconds and then 6. Why can’t the loess develop arcuate-trace surface fis-
looking out the aircraft window one can easily lose track of sures, or such features oriented perpendicular to the
your location and orientation. Recording audio for notes or ground surface?
video and audio for more complete documentation is recom- 7. Our concepts and present knowledge of mine elevations
mended and good communication and planning with the as tied to major features is lousy.
pilot prior to the flight is essential.
A DOE, Expedited Site Characterization (ESC) program Such discussions between team members are necessary
was being carried out by the Ames Laboratory with the during the site characterization process to keep challenging
assistance of the authors and others at the Central Nevada preconceived ideas. After much more time and data were
Test Area (CNTA). The area of interest extended over hun- collected at this site these issues were all resolved. See Chap.
dreds of hectares and included various disposal areas and a 25 for more details.
large mud pit. The site was mostly an open area with few The revised conceptual model may include improved geo-
reference points. While maps and vertical aerial photos pro- logic cross sections and maps, block diagrams, and photos along
vide an accurate overview of the site, oblique aerial photos with the text necessary to explain them. Possible conditions that
often will provide a better prospective and a feel for access may not have been identified, but might be reasonably expected
and logistics including areas that may be inaccessible. A should be mentioned. The basis for any assumptions should be
site fly over was completed by a key member of the ESC noted along with levels of uncertainty. This model (which is
team which provided perspective of site conditions which based upon knowledge of the site conditions, not guess work)
were difficult to obtain from maps and observations on the will then be used to guide further on-site work.
ground alone.

15.9 Updating the Work Plan

15.8 Updating the Conceptual Model
Now that the site walkover and a possible fly over have been
The site walkover will commonly provide an opportunity to completed, we have a reasonable first approximation of the
resolve a number of the issues identified in Tables 13.1, 13.2, nature of the geology and hydrologic conditions to be
and 13.3 discussed in the desk study. In addition there will expected at the site. We also may have a first approximation
have been other features that will have been discovered during of the karst conditions at the site.
References 159

For example if the focus is on sinkholes, topographic disciplinary conference on sinkholes: their geology, engineering
maps and aerials will have provided data on the type, density and environmental impact, Orlando, FL. Balkema, Rotterdam,
pp 79–85
and patterns of sinkholes and fractures. A review of sinkhole Fookes PG (1997) Geology for engineers: the geological model, predic-
databases along with comparison of old and recent aerial tion and performance. Geol Soc Lond Q J Eng Geol Hydrogeol
photos provides an indication of sinkhole activity. The karst 30(4):293–431
maturity of the site should be able to be determined based Freeman VL (1968) Geology of the Comstock-Indian Wells area Val
Verde, Terrell and Brewster Counties. Texas Geological Survey,
upon the five classes (Fig. 4.1) developed by Waltham and Professional Paper 594-K, USGS Washington, DC
Fookes (2003). Sinkholes on-site should be able to classified Freeman TJ et al (2002) Has your house got cracks? A homeowner’s
(Fig. 3.1). If cover collapse sinkholes are present, we might guide to subsidence and heave damage. Institution of Civil Engineers
can insight into the nature of the void spacing with the rock and Building Research Establishment, 2nd edn. Telford Publishing,
London
based upon their size (volume of material lost). Similar data Hatheway A (1998) “Site sketches” the essential picture, perspectives
can be developed for caves, springs, fracture patterns, etc. no. 32. AEG News 41(1):36–39
While this is based upon existing, readily available data, it Lewis WC (1991) Atmospheric pressure changes and cave airflow: a
should provide a level of focus to the subsequent field efforts. review. NSS Bull 53:1–12
Lewis WC (1992) Natural blowholes and blowing wells. NSS News,
In addition, a reasonable understanding has been estab- October, pp 258–263
lished for the various natural and cultural conditions that Lewis WC (1996) Constant airflow anyone? NSS News, July,
may limit access and control our approach to field work. We pp 181–182
have filled in some of the answers to questions about the site Milanovic T (2004) Water resources engineering in karst. CRC Press,
Boca Raton
and identified other new issues that need resolution. Based Newton JG (1987) Development of sinkholes resulting from man’s
upon this information we can update our work plan for the activities in the eastern United States. USGS Circular 968
subsequent field investigation and confirm that the methods Quinlan JF (1989) Ground-water in karst terranes, recommended proto-
and coverage are all appropriate and adequate. col and implicit assumptions. USEPA Las Vegas Laboratory,
EPA/600/X-89/050
Quinlan JF (1990) Special problems of ground-water monitoring in
karst terranes. In: Neilson DM, Johnson A (eds) Ground water and
References vadose zone monitoring, ASTM special publication 1053, West
Conshohocken
Audell HS (1996) Geotechnical nomenclature and classification system Rosza R et al (1975) Permeability of a nuclear chimney and surface
for crack patterns in buildings. Environ Eng Geosci 2(2):225–248 alluvium, Area 2, ERDA NTS, UCID-16722 Lawrence Livermore
Laboratory, CA
Audell HS (2004) Field guide to crack patterns in buildings, a guide to
Terzaghi K (1957) Consultants, clients and contractors. Presented at the
residential building cracks caused by geologic hazards. Association
Boston Society of Civil Engineers Section of ASCE, Reprinted by
of Engineering Geologists, Southern California Section, Special
Risk Analysis & Research Corporation for the Association of Soil
Publication No. 16
and Foundation Engineers
Boston PJ et al (2005) Micrometeorological modeling of idealized Waltham AC, Fookes PG (2003) Engineering classification of karst
caves: predictions and applications to Carlsbad Cavern, NM, ground conditions. Geol Soc Lond Q J Eng Geol Hydrol
USA. NSS Convention Abstracts. J Cave Karst Stud 36:101–118
Clayton CRI et al (1982) Site investigation, a handbook for engineers. Weeks EP (1978) Field determination of vertical permeability to air
Halsted Press, New York in the unsaturated zone. Geologic Survey Professional Paper
Clayton CRI et al (1995) Site investigations, a handbook for engineers, 1051, Washington, DC
2nd edn. Blackwell Science, Cambridge Weeks EP (1987) Effect of topography on gas flow in unsaturated frac-
Dangleish J, Alexander EC Jr (1984) Sinkhole distribution Winona tured rock, concepts and observations. Am Geophys Union Geophys
County Minnesota. In: Beck BF (ed) Proceedings of the 1st multi- Monogr 42:165–170
 Surface Geophysical Methods
16

Abstract
We now have a work plan that is based upon regional and local existing data and possibly
some site-specific data. A conceptual model has been developed and focuses our attention
on what we know, what we need to know and what we might expect on site. As we begin to
acquire on-site data using surface geophysical methods, we will begin to test the conceptual
model. Their primary benefit is that of greatly increasing the data density over the site,
which improves the definition of background and anomalous conditions. These measure-
ments allow us to more accurately locate invasive measurements based upon data rather
than guesses thereby improving the chances of accurately defining geologic conditions.
This section provides an introduction to the many surface geophysical methods. We have
focused on considerations for their use and how to select methods rather than how to make
the measurements. In addition, a large number of examples showing their application are
provided to show how these methods can be used in a site characterization effort.

16.1 Introduction face hydrogeologic conditions by remote sensing one or

more of its properties.
The geophysical methods encompass a wide range of air- The focus of this section is on applied surface geophysical
borne, surface, marine and downhole measurements methods used on land. Although many of the surface geo-
(Fig. 16.1). Some geophysical measurements can be made physical methods can also be used on water (rivers, lakes and
from an aircraft such as electromagnetic (EM), magnetic, estuaries) or over frozen bodies of water or even underground
thermal imagery, and radiometric measurements. Airborne in tunnels, mines and caves. The various downhole geophys-
geophysical surveys are cost effective when coverage of ical logging and borehole methods are discussed later in the
large areas is required, especially when the area is not easily site characterization process after the borings have been
accessible for ground surveys. Here we will focus upon the located and drilled.
surface geophysical methods, which represent an extremely This section presents an overview of surface geophysics. It
powerful array of tools to help characterize complex subsur- is not intended to provide detailed knowledge of individual
face geologic and hydrologic conditions. While geophysical methods, but rather to introduce the broad array of measure-
methods applied to exploration for oil are looking down ment techniques available and their applications. Numerous
1,000’s of meters, our focus is typically much shallower, examples are provided to illustrate the wide range of methods
from 1 to 30 m or so. Surface geophysics have a wide range and applications to characterize karst or for that matter any
of applications including: engineering, environmental, water other complex geologic condition. Further discussion of geo-
and mineral resources, and even archeological applications. physical methods and their applications can be found in Ward
Its use is referred to as applied engineering or environmental (1990), Butler (2005), Reynolds (2011), Burger (1992),
geophysics but it is all the same thing. Applied geophysics Milsom and Eriksen (2013), Sharma (1997), McCann et al.
can be defined as the measurement and mapping of subsur- (1997), Nazarian and Diehl (2000), and Wightman et al. (2003).

© Springer Science+Business Media Dordrecht 2016 161

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_16
162 16 Surface Geophysical Methods

d
c

Fig. 16.1 Geophysical measurements can be made from the air, on the surface, over water and from within boreholes. (a) Airborne or satellite
remote sensing. (b) Surface geophysics. (c) Marine geophysics. (d) Downhole geophysics

Recall our early discussion of the limited spatial extent • DiMaggio (2007) Commentary: Advancements and
of caves (something on the order of less than 5 % of the Disappointments in Geo-Engineering and Geo-
surface area) and the low probabilities of their detection by Construction – A 30–Year Reflection. “More tools for
borings alone even when a large number of borings are rapid, if not real time monitoring, readout and digital data
used (Fig. 12.2). The great advantage of the surface geo- reduction”.
physical methods is that of greatly increasing the spatial • Mitchell and Kavazanjian (2007) Geoengineering
density of data. Thereby improving our ability to detect and Engineering for the twenty-first Century, “An improved
map spatial changes in complex geologic conditions and ability to see into the earth perhaps the most important
providing a means to accurately locate invasive measure- need in geotechnical engineering”.
ments based upon data rather than guesses or rules of
thumb. This increases accuracy and confidence levels in As a result this section is longer and contains more details
site characterization process. in order to emphasize and illustrate the importance of the
The authors have played a significant role in pioneering geophysical methods.
and promoting the application of surface geophysical tech-
niques over the years. This includes:
16.2 A Brief History of the Surface
• Developing the application of geophysical methods to Geophysical Methods
the site characterization process including application of
the methods to a wide variety of site characterization Geophysical methods are not new; they have been used for
problems (Benson et al. 1984; Benson and Yuhr 1996, many decades in the exploration for oil and gas as well as
1997) mineral and water resource exploration dating back to the
• The development of many of the ASTM guidelines for early 1900s. In the 1950s and 1960s early geotechnical appli-
surface geophysical methods, cations used relatively simple seismic refraction measure-
• The development of DOE’s Expedited Site Characterization ments and analysis to determine top of rock profile and to
program (ESC) in the 1990s as described in Beam et al. measure the P-wave velocity of rock to determine its rippa-
(1997), Benson et al. (1998), Benson (1997), and Yuhr bility. Early engineering seismographs begin to appear
(1998). around the 1960s.
• The development of the ASTM guideline for Expedited The nuclear power plant siting in the 1960s utilized
Site Characterization (ESC) (ASTM 2010a). geophysics to aid in site characterization by adapting seis-
mic reflection methodology to assess the presence of
Others have also emphasized the need for surface geo- faults. In the early 1970s, ground penetrating radar (GPR)
physics in the site characterization process including: systems first became available from Geophysical Survey
Systems, Inc. and the shallow electromagnetic induction
• The National Research Council’s “Seeing into the Earth”, systems became available from Geonics, Ltd. With the
(2000) GPR and EM systems available, we now had two extremely
16.3 An Overview of Surface Geophysics 163

powerful and versatile tools for characterizing shallow 16.3 An Overview of Surface Geophysics
geologic conditions. Both of these methods were capable
of providing continuous high-resolution data along a pro- Table 16.1 lists the common surface geophysical methods
file line. By the early 1970s, sub-bottom profiling (high used for site characterization along with the parameter mea-
resolution seismic reflection on water) and side scan sonar sured and their typical applications. This table provides a
(acoustic imaging of the sea bed) were being applied to general understanding of what can be done with the wide
marine projects. array of methods that are readily available. We recognize that
Although many of the surface geophysical methods are there are other techniques or adaptations of techniques that
capable of making deep measurements, it is the near surface may be applicable and could be considered but have limited
environment, within the upper 30 m or so, that is usually of our discussion to those methods commonly used by the
interest since it is the region most susceptible to human activ- authors and many others. Figure 16.2 illustrates some of the
ity and has the greatest impact on geotechnical and environ- methods and ways in which geophysical data can be acquired.
mental projects. Engineering and environmental applications One of the most critical factors affecting the assessment
of the geophysical methods are relatively new (roughly 1970s) of geologic conditions at any site is the uncertainty of the
and are somewhat different in their application because they data from limited spatial sampling (Fig. 12.2). This is often
are typically used to obtain relatively shallow, higher resolu- the fatal flaw of many site characterization efforts. Achieving
tion data than applications for oil and gas exploration and min- a reasonable statistical sampling of geologic site conditions
ing. The application of the shallow high-resolution geophysical would require borings or other methods of limited spatial
methods to hazardous waste sites (Benson et al. 1984) and sampling to be placed in a close-order grid, which would
critical structure siting of nuclear facilities (Franklin et al. greatly increase time, effort and cost, as well as reduce the
1981) investigations began in the mid to late 1970s and early site to “swiss-cheese”. Yet relying on data from a limited
1980s. With expanding coastal populations and the need for number of borings remains a weak link in many site charac-
additional fresh water in the 1980s and 1990s, there has been terization efforts (Benson et al. 1984; Franklin et al. 1981).
an effort to better characterize the extent of saltwater intrusion Unlike direct sampling and analysis, such as obtaining a
using geophysical measurements. soil, rock or water sample and sending it to a laboratory, the
Today geophysical methods are commonly applied to geophysical methods provide non-invasive, in-situ measure-
geotechnical, environmental and karst problems world- ments of the subsurface geologic conditions with little or no
wide. In the US, the semi-annual Sinkhole Conferences disruption of the surface. The geophysical measurements
(www.sinkholeconference.com) along with the integrate a larger volume of the subsurface compared to bor-
Environmental and Engineering Geophysical Society ings. As such they are averaging much more material into
(EEGS) and their annual Symposium on the Application of their measurement, which has both benefits and disadvan-
Geophysics to Engineering and Environmental Problems tages. Sampling larger volumes increases our chance of
(SAGEEP) conferences have been major factors in promot- detecting changes in geologic conditions. On the other hand
ing the use of the geophysical methods. In 1977, the US if we sample too great of a volume we can lose the sensitivity
Army Engineer Waterways Experiment Station in Vicksburg, and resolution required to detect smaller changes in geologic
Mississippi held a symposium on Detection of Subsurface conditions. Selecting the appropriate method(s) whose scale
Cavities (Butler 1977). During the 1980s and early 1990s, of measurement will optimize our chances of detecting and
the Colorado School of Mines sponsored Tunnel Detection resolving the geologic variability of interest at a specific site
Meetings. The Engineering Waterway Experiment Station is critical.
in Vicksburg, Mississippi and the US Army also carried out The geophysical methods can be used at two different
extensive programs in tunnel and cavity detection. In the stages of the site characterization process.
mid-1990s the Federal Highway Administration (FHWA)
promoted the applications of geophysics to karst related • Early in the site characterization process when our objec-
problems along roadways and infrastructure. An application tive is simply to locate the anomalous geologic conditions
guide on the use of geophysical methods for highway related by increasing sampling density. Here, the use of geophys-
problems, was published by the FWHA (Wightman et al. ics is often qualitative and highly intuitive (Greenhouse
2003). In 2000, the Federal Highway Administration and the et al. 1995).
Missouri Department of Transportation sponsored the First • Later on, in the site characterization process, geophysical
International Conference on the Application of Geophysical data may be used more quantitatively to determine spe-
and Non-Destructive Testing (NDT) Methodologies to cific engineering or hydrologic parameters such as:
Transportation Facilities and Infrastructure. Some of these P-wave velocity for rippability, calculating elastic con-
conferences continue to provide information and examples stants from P-wave and S-wave velocities, locating the
on the application of geophysical methods. fresh-seawater interface, and characterizing the size and
 164

Table 16.1 Surface geophysical methods their measurement and application

Method and common terminology Typical depth of measurement (m) Parameter measured Typical applications Used over water
Ground penetrating radar (GPR) 3–10 m typical, much less in clays, Dielectric constant, electrical High resolution cross sections of soil and rock Fresh water only
(radar) conductive soil, rock or pore fluid conductivity and magnetic showing strata, fractures, voids, cavities and
susceptibility also buried utilities, drums and tanks
Electromagnetic (EM) methods
Frequency domain (FDEM) 1–60 m, depending upon Electrical conductivity and magnetic Detect lateral variations in soil and rock, also Fresh water or waters
instrumentation used susceptibility used to locate underground metal utilities, of low electrical
drums and tanks conductivity
Time domain (TDEM) 5–1,000 m Electrical resistivity Sounding measurements. Top of bedrock. Salt Not normally
water intrusion
Very low frequency (VLF) About 20 m typical Tilt and phase angle of electric field or Tabular conductors such as faults and fracture Not normally
16

field strength, vertical and horizontal zones. Used to locate water and mineral zones
In-phase and quadrature components
Electrical methods
DC resistivity Up to 100 m Electrical resistivity Mapping lateral and vertical resistivity Fresh and salt water
changes in soil and rock
2D imaging Up to 100 m Electrical resistivity 2D profiles mapping lateral and vertical soil Fresh and salt water
and rock changes
Capacitance measurements Up to 12 m Electrical resistivity 2D profiles mapping lateral and vertical soil Not normally
and rock changes
IP/complex resistivity Up to 100 m Electrical resistivity and chargeability 2D profiles mapping lateral and vertical soil Not normally
and rock changes
Surface Geophysical Methods
Self potential (SP) also called Typically shallow <20 m Surface voltages resulting from flowing Seepage of water at dams, impoundments, Fresh and salt water
natural potential or spontaneous water or electrochemical effects in soil flow in conduits or caves, electrochemical
potential or rock effects due to recharge zones
Seismic methods
Seismic refraction 100 m or less Compressional Vp and or shear Vs wave 2D profiles of top of rock profile, weathered Fresh and salt water
velocity zone, rippability
Seismic reflection 300 m or more Compressional Vp and or shear Vs wave 2D profiles of stratigraphy, and structure Fresh and salt water
velocity
Multiple analysis of surface waves 15–45 m Shear wave Vs velocity 2D profiles of variations in stratigraphy and Fresh and salt water
(MASW) for engineering calculations
Microgravity 100 m or more Changes in the local acceleration of Variations in depth to rock, location of buried Not typically used
gravity due to changes in density of valleys, fractures, faults, cavities, cave systems over water
subsurface materials and mines
16.3 An Overview of Surface Geophysics

Magnetic 100 m or more Magnetic susceptibility of soil and rock Lateral variations of geology related to Fresh and salt water
magnetic variations in soil and rock or location
of underground ferrous metal utilities, drums
and tanks
Metal detectors and pipe cable 1–10 m Changes in electrical conductivity or Location of buried utilities and location and Fresh and salt water
locators resistivity of metals, or to induced EM mapping of buried drums, tanks and trash
signal from detector
Thermal imagery At surface Temperature of the surface Location of springs, caves and mine entrances Fresh and salt water
and air flow from wells and fractures
Radiometric At surface Natural gamma radiation Location of fractures, faults and caves along Fresh and salt water
with higher permeability zones within the soil
and rock
165
166 16 Surface Geophysical Methods

Fig. 16.2 A wide range of surface geophysical instrumentation is available to acquire subsurface data
16.3 An Overview of Surface Geophysics 167

dimensions of anomalous conditions such as fractures rock character (ie degree of fractures and bedding thickness)
faults and cave systems. will determine if the rock can be ripped and which size trac-
tor will be required (Caterpillar 1988). Whiteley (1983) pro-
vides a discussion of the relationships between some of the
geologic, hydrologic and engineering parameters and the
16.3.1 Parameter Measured geophysical measurements.
There are also correlations that exist between water qual-
Each of the surface geophysical methods responds to the ity and geophysical measurements. Inorganic contaminants
changes of some physical, chemical, electrical, or radiologi- such as landfill leachate or natural contaminants such as salt-
cal property of the subsurface (Table 16.1). The geophysical water intrusion, can be mapped using electrical/electromag-
methods respond to specific frequencies over a wide range of netic techniques due to their elevated pore fluid conductivity
the seismic/acoustic and electromagnetic spectrum levels (Benson et al. 1985). Where degradation of organic
(Fig. 16.3). Changes in geologic and hydrologic conditions contaminants has occurred an increased electrical conductiv-
will commonly be associated with a change in one or more of ity values have also been noted (Sauck et al. 1998; Cassidy
the measured parameters. The surface geophysical methods et al. 2001; Atekwana et al. 2002). These correlations allow
are able to detect lateral and or vertical changes in a given electrical and electromagnetic measurements to be use to
parameter, which can then be interpreted to provide an map some organic contaminant plumes, identify their bound-
assessment of subsurface geologic or hydrologic conditions. aries and sometimes identify flow pathways.
The geophysical methods do not typically detect the soil/
rock interface itself, but the change in electrical resistivity, den-
sity, P-wave velocity, etc., that occurs at this interface. When 16.3.2 Anomalies
applied to karst, these methods do not typically detect a cavity
itself, but the change in density or electrical resistivity of the The term anomaly is used to describe any unexpected devia-
cavity compared to the surrounding materials. The methods do tion of the data from expected or background conditions. For
not detect the flow of water but the resulting streaming poten- example, an anomaly can represent a change in depth to
tial (voltage) on the surface created by flowing water in a frac- rock, a change in clay content or the presence of a cavity. An
ture or conduit. Regardless of geophysical method selected, anomaly can also represent a cultural change due to a pipe-
there has to be a change or contrast in the measured parameter line, buried tanks or contaminants. When a geologic anomaly
related to the target of interest for the results to be successful. is identified by geophysical measurements, there are a vari-
For example, Caterpillar Tractor Company uses seismic ety of rules of thumb or calculations that can be used to pro-
refraction measurements to obtain compressional wave vide an approximate depth and size of a subsurface target
seismic velocity to establish rock rippability. The combination that resulted in the measured anomaly.
of seismic velocity along with a visual determination of the

1 Hz 10 Hz 102 Hz 103 Hz 104 Hz 105 Hz 106 Hz 107 Hz

Sub-bottom Depth Acoustic Imaging Medical

Seismic
Profilers Sounders Positioning Sonar Ultrasound
Shallow Seismic Methods

Seismic/Acoustic Spectrum

3x10-5 Hz 3x100 Hz 3x105 Hz 3x1010 Hz 3x1015 Hz 3x1020 Hz

GPR
Resistivity VLF Multispectral
Television X-rays
Electromagnetic and Radio Scanners
Gamma rays
Induction Radar and
Th

Passive Cosmic Rays

Vi ted d
R
er

ef Infr

sa

Microwave
m

le

bl
al

e
U
ltr ed
I n
ar

av
f r
e

ar

io
le
t

Electromagnetic Spectrum

Fig. 16.3 The geophysical measurements encompass a wide range of the seismic/acoustic and electromagnetic spectrum
168 16 Surface Geophysical Methods

A geophysical anomaly can be due to many different a cave system (Banwell and Parizek 1986; Upchurch
geologic or hydrologic conditions. For example, an increase et al. 1987; Armstrong and Heemstra 1973; Hansen
in electrical conductivity could be due to inorganic contami- 1975).
nants, increase clay content, increase moisture content,
thicker soil layer, etc. It is our job to have enough data and Another halo effect associated with cave systems may
information about the site to know what conditions to expect. include the stress fractures surrounding the cave and the
With a developing conceptual model in place, based upon many tertiary fractures associated with cave development
existing data, we should be able to narrow down the possible that are too small for exploration and mapping by cavers.
cause of the anomaly. Knowing the location and possible These features surrounding the main cave create “halo
depth of the anomaly we can then locate additional measure- effects” and may increase a gravity anomaly by a factor of
ments such as borings for further characterization. 1.5–2.5 (Butler 1994, 2008; Llopis et al. 2005), making the
cave system more easily detectable.
The use of NSI or halo effects provides a means of indi-
16.3.3 Direct and Indirect Detection rectly detecting the presence of deeper cave systems.This is
of Anomalous Conditions not unlike the use of vegetation stress in aerial photography
to indicate the presence of subsurface lineaments (ie frac-
Direct and indirect detection are two approaches that can be tures, faults and recharge). Depending upon the nature of
used to detect anomalous subsurface conditions using the the project, both direct and indirect methods may be
surface geophysical methods. utilized.

• Direct detection can be accomplished by some of the sur-

face geophysical methods when the feature of interest 16.3.4 Penetration of Measurements
such as a fracture or cave system is large enough, has a
sufficient contrast to its surrounding materials, and is Table 16.1 provides a typical maximum depth of measure-
within the depth range of the specific measurements ment or penetration of the surface geophysical measure-
(Table 16.1). This approach provides a very high level of ments. Again, these are generalized and the selection of
confidence in subsurface conditions. instrumentation, set-up parameters and site-specific condi-
• Indirect detection can be accomplished by the use of near tions will all impact the actual results. For example, both fre-
surface indicators (NSI) or halo effects, which refers to quency domain electromagnetic (FDEM) and GPR
detecting the presence of an anomaly such as a major frac- measurements are limited by high conductivity values of the
ture or cave system without direct detection of the feature soil, rock, and their pore fluids. If the surface sediments are
itself. Very often a cave system at depth (may be beyond the highly resistive (i.e. quartz sands or dry clay-free limestone)
range of our measurement) but will show signs of its pres- GPR depths can be greatly extended. The depth of resistivity
ence and possible activity by NSI or halos within the shallow measurements will be reduced when there is a highly resis-
sediment or rock. These NSI or halo effects may include: tive layer such as clean quartz sand, which would limit ade-
– Soil piping, raveling or dipping strata that may be quate current getting into the ground. These are just a couple
occurring in the shallow sediment as a result of a deep of examples showing the impact of site-specific conditions
collapse, which has not yet reached the surface (Benson on the depth of measurements.
and Yuhr 1992).
– A fracture system over a cave may have caused a local
increase in permeability and provide an area of ground- 16.3.5 Resolution
water recharge resulting in geochemical changes over
the cave. As a result, a local change in electrical con- Resolution is the ability to resolve two geologic conditions
ductivity, resistivity, or SP voltage may provide an or cultural features, which are close together. Both the lateral
indication of the presence of an underlying cavity and vertical resolution of the geophysical method can impact
(Benson and Yuhr 1992). the data and its interpretation. There is a difference between
– Thermal gradients or flowing air may exist at the sur- detecting a subsurface feature and resolving that feature.
face, associated with permeable zones, fractures or Detection does not necessarily include determining the size,
caves within the unsaturated zone (Thompson and thickness or shape of a subsurface feature. As a result the
Marvin 2006; Straley 2008; Bogle and Loy 1995; requirements for detection are less stringent than for resolv-
Campbell and Singer 2001). ing a feature (National Research Council 2000).
– Radiation halos (gamma or radon) may exist at the sur- The resolution is an inherent function of the method
face over areas of increased permeability, fractures or selected, how the method is used, along with the depth,
16.3 An Overview of Surface Geophysics 169

nature and size of the target. As an example, radar can capacitively-coupled resistivity, thermal imaging, radioactive
provide resolution of centimeters when mapping individual measurements, and most of the marine methods.
steel reinforcement in concrete (Fig. 16.4) while resolution By “nearly continuously” we mean that measurements
may be reduced to a meter or so when mapping the top of made by most systems today have a fairly high digital sam-
rock at a depth of 10–20 m. pling rate of 10–30 or more measurements per second. If we
are walking at 4.8 km/h and have a low sample rate of 10
16.3.5.1 Lateral Resolution samples per second then we will have acquired data at 7.2
The lateral resolution of surface geophysical measurements samples/m. Even at speeds of 16 km/h using a vehicle-towed
is an inherent function of the method used, the volume sam- system, the sample rate would be almost 3 samples/m, which
pled by the specific measurement, and the spacing between for all practical geologic purposes is continuous.
measurements. The spacing of measurements along a survey Figure 16.5 shows two sets of frequency domain electro-
line and the distance between survey lines or grids need to be magnetic (EM34) data along the same profile line. The data
close enough to laterally resolve the feature of interest. in Fig. 16.5a was collected at a station spacing of about 60 m,
Selecting the station and line spacing for measurements while the data in Fig. 16.5b was collected nearly continu-
requires insight regarding the likely size, depth and nature of ously. The continuously sampled data provides greater lat-
the feature of interest (Fig. 12.9), the general geologic vari- eral resolution and has clearly defined the fracture patterns
ability and noise conditions along with the physics of the within the gypsum rock buried under about 2 m of
sensing method. This insight only comes from years of sediment.
experience.
Data from some methods may be obtained “nearly con- 16.3.5.2 Vertical Resolution
tinuously” along a survey line, resulting in very high lateral Vertical resolution is the ability to separate the response from
resolution. Such measurements include GPR, FDEM, towed two distinct layers or targets at different depths. Regardless

a
Surface of tunnel wall Rebar
0

Depth in Centimeters
Back of concrete 30
Void Space
Surface of rock
60

b
Void space
between the
concrete tunnel
lining and the
natural rock Rock
Surface

Back of
Concrete
Liner

Fig. 16.4 An example of shallow radar data illustrating the level of behind the concrete (a). A photo behind the concrete tunnel lining (b)
detail that can be achieved shows the thickness of the concrete tunnel shows the variable void space
lining, the reinforcing steel bars within the concrete and the void space
170 16 Surface Geophysical Methods

a 11 Staon Measurements
Approx. 60m apart

Conductivity (mS/m)
0 120 240 360 480
Distance in Meters
Fractures in Gypsum
b
Connuous
Measurements
Conductivity (mS/m)

0 120 240 360 480

Distance in Meters

Fig. 16.5 A comparison of station measurements (a) versus continuous measurements (b) over the same survey line. The continuous electromag-
netic measurements detect a pattern of fractures in gypsum that is undetected in the station measurements

of the method of measurement, the resolution of all surface gravity) require varying degrees of processing and modeling
geophysical measurements decrease with increased depth. to convert the raw field data into useable information. These
As a particular layer of interest or target becomes deeper, a models may include:
greater contrast is necessary between the target and sur-
rounding materials and or its size must be greater in order for • 1D models consisting of a sounding or vertical log that
it to be detected and its size and shape to be resolved. shows variations in the measured parameter with depth
• 2D models consisting of a cross section or contour map
• 3D models consisting of isometric figures combining 1D
16.3.6 Processing and Presentation of Data or 2D models
• 4D models consisting of time-series plots of 1D, 2D or
The results of many surface geophysical methods can be 3D data from the same location
interpreted directly from the field data (Fig. 16.6a) or after
the data have been plotted as a profile or have been con- Geologic conditions exist in a three-dimensional (3D)
toured. This allows preliminary in-field assessments to be world. Hydrologic conditions can change with time and may
made. For example, GPR, frequency domain EM, resistivity be four-dimensional (4D). In many cases, we must simplify
profile data, as well as thermal imaging and radioactive mea- the models by holding some of the variables constant. In
surements can often be interpreted without any processing addition, care should be taken not to over interpret data or
other than for presenting the data. In addition, some marine create more detailed models than the data actually supports.
data (bathymetry, side-scan, and subbottom profiling) can Nowadays, data is often presented in 3D. This is often
also be interpreted without additional processing. accomplished by acquiring data in 2D and processing the
In some cases, simple steps may be carried out to improve data to create a 3D presentation. For example, ground pene-
the data and its presentation. For example, radar data col- trating radar data is gathered as a series of parallel or perpen-
lected along a survey line can have topographic corrections dicular lines and can then be processed to yield 3D block of
applied (Fig. 16.6b). When topographic correction is applied, data. When data is acquired with the intent of processing a
the data is presented in a corrected visual format, which can 3D image, extremely high density data is required (both lat-
improve understanding and interpretation of the data. erally and vertically). Because of the need for such high den-
Many of the methods (such as; resistivity soundings, 2D sity data, three-dimensional data are typically limited to
resistivity, seismic refraction and reflection, MASW, SP, and small areas. This is due to equipment limitations as well as
16.3 An Overview of Surface Geophysics 171

Fault
a

West East
0

Approx. Depth (meters)

5

10

0 100 200 300 400 500 600 700 800 900

Distance (meters)

b
0

Fault

Relative Depth (meters)

10

20

30

0 100 200 300 400 500 600 700 800 900

Distance (meters)

Fig. 16.6 An example of radar data using a 100 MHz antenna without topographic corrections (a) and with topographic corrections (b)

time and cost limitations. Therefore 3D data should be the use of a standard coordinate system for the entire site
strategically used to characterize known anomalous condi- characterization process is necessary. See discussion of sur-
tions which have already been located but require additional vey grid in Sect. 12.3.1 Project Preparation.
detailed data or confirmation and not for reconnaissance Global Positioning Systems (GPS) can now be integrated
purposes. with geophysical measurements and used to record locations
in real-time. However, the sole dependency on GPS systems
can be problematic and does not avoid the need for an estab-
16.3.7 Use of a Survey Grid lished local survey grid system for all data collected on-site.
The survey grid allows spatial correlation of all data, allows
Surface geophysical measurements are commonly made for corrections in positioning errors and aids in orienting
along a survey line, as a series of parallel survey lines or over positions on-site. Some of the geophysical methods, such as
a grid. An important aspect of any geophysical survey is to radar, utilize a time based sampling or distance established
establish a survey line, parallel lines or grid along with a by a counter wheel. In some cases, position will be manually
standard coordinate system. The best data in the world are noted on the data record as one passes a fixed station mark.
useless if no one knows where it came from or cannot easily When errors in positioning occur, and they do, they can be
get back to the specific location. In addition, the geophysical corrected prior to data processing and interpretation. Most
data must be able to be correlated to other data. Therefore, survey work on water will utilize GPS, often incorporating a
172 16 Surface Geophysical Methods

steering system to guide the helmsman. Range markers and • Frequency domain EM systems, can be used to locate
or reference buoys are also used to guide the helmsman. either ferrous or non-ferrous metal utilities as well as
These visible references provide quick visual orientation on trash, conductive contaminants (soils or pore fluids), and
the water. UST’s;
• Magnetometers are use to locate ferrous metal drums,
UST’s and utilities;
16.3.8 Location of Utilities, Buried Drums, • A variety of metal detectors and pipe and cable locators
Tanks and Trash can be used to locate utilities as well as ferrous or non-
ferrous metal trash.
Many of the surface geophysical methods can be utilized to
detect and map such man-made features such as utilities, Figure 16.7 shows the results of selected measurements
buried debris, underground storage tanks (UST’s), drums, over drums, buried debris, utilities and UST’s. Figure 16.7a
etc (Benson et al. 1984). While this is not the focus of this is the response from a metal detector over a trench that con-
book, there are many cases in which these conditions exist at tained 55-gal drums. Figure 16.7b is a contour map of elec-
a site, along with karst or pseudokarst conditions, and can tromagnetic (EM31) data (0–6 m) revealing a conductive
interfere or impact site investigation efforts. The various sur- contaminant plume, buried debris and various utilities.
face geophysical methods used for detecting these man- Figure 16.7c is a set of data that includes electromagnetic
made targets include: (EM31), magnetometer and GPR data over four abandoned
UST’s, each 10,000 gal.
• GPR to locate buried utilities and (UST’s);

Fig. 16.7 Some of the geophysical methods can be used to map man-made features such as buried steel drums (a), utilities and buried debris (b)
and underground storage tanks (c)
16.4 Guidelines for the Selection of the Surface Geophysical Methods 173

16.3.9 Ease of Use fresh or salt water. All of these measurements are made
“continuously” as the survey boat moves through the water
Ease of use refers to the level of effort necessary to make the using a high sampling rate.
geophysical measurements. The level of effort includes There is an array of marine seismic methods commonly
whether the measurements are acquired on a station-by- used in fresh or salt water that includes:
station basis or continuously, the size of the field crew
required for the measurements, or whether the measurements • Seismic refraction and seismic reflection, which provide
require an intimate physical contact with the ground surface. subsurface data in cross section similar to those methods
on land (Fig. 16.8c).
• Some measurements such as, GPR, frequency domain • Bathymetry or echo sounding measurements are made to
EM, capacitively coupled resistivity, magnetic, thermal, determine the depth of water.
and radioactive measurements are relatively easy to make • Sidescan sonar provides an acoustic image of the bottom.
since they can be made by station measurements or by
continuously walking or driving along a survey line.
• Gravity measurements are made on a station-by-station 16.3.11 Limitations
basis without probes driven into the ground. However,
microgravity measurements do require highly accurate All geophysical methods have advantages as well as limita-
(first order) relative elevation data for each gravity station. tions. Some of the surface geophysical methods are limited
• Others measurements such as resistivity, time domain by site-specific conditions. Some of the methods may be
EM, self-potential (SP), and seismic measurements impacted by cultural features, such as nearby metal fences,
require that coils, probes or geophones be placed in con- utilities, or structures, and some are sensitive to ground
tact with the ground along a profile line and cables are vibrations from vehicle traffic, wind, or even remote earth-
deployed and connected. Both resistivity and seismic quakes. Site characterization will often occur in areas that
measurements typically require intimate contact with the have been previously developed, and have been used for a
ground. In some cases a towed cable array may be used to variety of purposes over long periods of time. These sites
speed up data acquisition for some seismic and resistivity may include the presence of buried structures, fill materials,
measurements using a weak ground contact. abandoned infrastructure, utilities, buried debris and con-
taminants which can complicate the investigation. These
These factors need to be considered when addressing site conditions can often eliminate a particular geophysical
conditions and logistics. Site conditions such as heavy veg- method from consideration. The selection and effectiveness
etation, steep slopes, and rocky surface will impact the selec- of a particular surface geophysical method requires not only
tion of methods to be used as well as sources of natural noise identifying the most appropriate method to meet a project
such as wind and storms, cultural noise such as buried utili- objectives but also identifying the method, which will be
ties, vibrations from machinery and transmission lines. successful under site-specific constraints.

16.3.10 Surface Geophysical Data Can 16.4 Guidelines for the Selection
Be Acquired Over Water of the Surface Geophysical Methods

Most of the surface geophysical methods are designed for American Society for Testing and Materials (ASTM) guide-
use on land but many have been adapted for use over bodies lines have been written for many of the surface geophysical
of water such as inland rivers, streams, canals, lagoons, res- methods. The terms guidelines and standards are often used
ervoirs, water-filled quarries, lakes and even ice-covered interchangeably to describe these various documents. These
bodies of water. These water-covered areas often provide documents were not intended to be standards but only to be
additional “geologic windows” into the subsurface, which used as guidelines. These guidelines were developed to pro-
are free from many cultural effects. When water covered vide a common approach for the use of these methods and
areas are adjacent to a site they often provide unique oppor- general information on the selection and application of a par-
tunities to obtain continuous, high resolution geologic data. ticular method for non-geophysical experts. The ASTM stan-
Traditional land-based measurements have been adapted dard guides for surface geophysical methods include:
to be used over water. Both radar and EM can be used over
water, but over fresh water only (Fig. 16.8a). Resistivity • Ground Penetrating Radar Method (ASTM 2011b)
measurements (Fig. 16.8b), spontaneous potential, thermal • Frequency Domain Electromagnetic Method (ASTM
imaging and radioactive measurements can be used over 2008)
174 16 Surface Geophysical Methods

Fig. 16.8 A wide range of surface geophysical measurements can be employed over water. (a) Ground penetrating radar, electromagnetics (fresh
water only), and magnetics. (b) Marine resistivity. (c) Seismic reflection (subbottom) and seismic refraction

• Time Domain Electromagnetic Method (ASTM 2007) In addition, a guideline was developed that covers the
• Direct Current Resistivity Method (ASTM 2010c) selection of surface geophysical methods (ASTM 2011c).
• Seismic Refraction Method (ASTM 2011a) This document includes a table recommending the use of
• Seismic Reflection Method (ASTM 2010d) certain geophysical methods for certain types of problems.
• Gravity Method (ASTM 2010b) All such guidelines are inherently simplified and generalized
• Metal Detection Method (ASTM 2011d) and cannot possibly consider all of the variables involved in
16.5 Application of Surface Geophysical Methods 175

planning a site-specific investigation program. These the data, integrate the geophysical data with other data to
guidelines do not discuss the wide range of options in which arrive at a realistic conceptual geologic model and write
each method can be applied, their limitations, or their possi- the report?
ble applications over water. Without an endless array of foot- • Are there any cultural, access, environmental, or safety
notes, qualifications, limitations and exceptions, such guide- issues that may impact the measurements?
lines must be used with caution and only as a first
approximation guide. However, for someone without any Note that some of these questions deal with non-technical
background in geophysics it is a good place to start, if only issues, which are as critical to the overall success of the pro-
for introducing the wide array of possible techniques avail- gram as are the selection and proper application of the
able for use. Further information on the wide array of surface method(s). No guideline can cover all of these many interwo-
geophysical techniques has been described by Saunders ven issues involved in properly selecting and applying geo-
et al. (1999), Benson and Yuhr 1996, Benson et al. (2003), physical methods. It is not simply a question of selecting a
and McCann et al. (1997). method or methods from an ASTM guideline it is also a
Anderson et al. (2003), Anderson and Ismail (2003), question of how it will be applied, processed, interpreted and
Olhoeft (2003), and Benson et al. (2003) have outlined some integrated with other data along with considerations of the
of the key steps and considerations for the selection and use methods inherent advantages and limitations. The key is to
of the geophysical methods. The following is a composite select those methods and their adaptations, which are most
list from these papers and includes some of the key questions appropriate for the project needs and have the best chance of
and issues to be considered in selecting geophysical methods providing useful results.
and preparing for a site characterization: In practice, someone who has had first-hand experience
with most of the methods should be making the final selec-
• The first step is to clearly understand and define the prob- tion of appropriate methods. This experience should include
lem or project objective. acquisition, processing, interpreting and integration of data.
• What existing geologic and hydrologic data is available Then the site-specific conditions and the project objectives
for the site and or surrounding area must be considered in order to recommend an optimum
• What is the area of interest? And which methods can pro- approach.
vide the necessary coverage in both a technical and cost
effective manner
• What is the target or targets – isolated and discrete, linear 16.5 Application of Surface Geophysical
or planer? Methods
• What is the target’s depth or range of depths?
• What is the required site coverage, spatial sampling and The following section provides examples of a variety of
resolution both laterally and with depth to detect the tar- geophysical methods and illustrates their broad applica-
get and which techniques can provide the desired tions that have been used by the authors. These categories
results? include:
• What are the relevant physical properties of interest to
define the geology and the target and which geophysical • Soil piping and collapse within the sediments
methods will respond to? • Epikarst along with the top of rock profile
• Is there a possibility for NSI or Halo effects from deep- • Fractures and cavities systems within the rock,
seated karst? • Buried sinkholes and paleokarst, and
• What are the site-specific constraints (access, topography, • Pseudokarst conditions.
structures and noise) and which geophysical methods can
perform best under these conditions? The examples provided cover a few of the commonly
• Which techniques (geophysical and non-geophysical) can used geophysical techniques listed in Table 16.1. The site
provide complementary data? locations are mostly within the United States. Each of these
• What data are already available and what other data are examples shows how a surface geophysical method was used
required to interpret and or constrain the interpretation of to acquire a set of data that was critical to the site character-
geophysical data. ization. These examples illustrate the unique insight into
• What are the expected results? subsurface conditions that surface geophysics can provide
• Who will make the decision regarding the methods to be and how they are integrated into the site characterization
used, plan the survey, make the measurements, analyze process.
176 16 Surface Geophysical Methods

16.5.1 Soil Piping and Collapse Within sinkholes. The sinkholes presented a high risk for the trains
the Sediments that provided critical supplies to navy ships in the nearby
port. Titcomb and Keeton (1984) provide early details of
The movement of soils into void spaces within the underlying sinkhole development at this site.
rock (fractures, voids, and cavities) will typically result in The area is covered by quartz sand with increasing clay
surface depressions, small circular soil piping features and content below a depth of 6 m and limestone occurring at a
ultimately a surface collapse. These are often small, nuisance depth of 12 m or more. A conceptual model of site geology
type of features that are triggered by flow of water or changes is shown in Fig. 16.9. The Corps of Engineers had drilled
in water levels. The materials entering the void space are more than 100 borings along 1,200 m of railway to evaluate
either being flushed through the system or in some cases the the problem. The sinkholes develop as small diameter piping
void space is large enough to accommodate the overlying features (<1 m diameter) caused by raveling of the uncon-
material and eventually result in a collapse feature. The solidated sediments (Fig. 8.3a). Subsequently the diameter
examples include: of the surface collapse could enlarge over time (Fig. 8.3b).
The initial soil piping was of such small dimensions that it
• Active cover collapse sinkholes along a military railroad was almost impossible to detect prior to breakthrough at the
in North Carolina surface by a drilling program alone.
• An active cover collapse sinkhole within a water-filled Initial surveys with ground penetrating radar using an
sand quarry in central Florida and 80 MHz antenna resulted in excellent quality data and indi-
• Seepage at an earthen dam in New Mexico. cated that the total depth of the radar penetration was about
10 m. A 1,200 m reconnaissance radar survey was run along
on both sides of the railroad track and identified 14 anoma-
16.5.1.1 Active Collapse Along a Military lous conditions. Each anomaly was then re-surveyed in detail
Railroad (North Carolina) using a series of parallel radar traverses 1 m apart to provide
Sinkhole activity began occurring in 1976 along a critical the exact center location and a detailed cross section of each
military railroad that provided supplies to the Sunny Point anomaly.
Military Ocean Terminal, 27 km southwest of Wilmington, Figure 16.10 shows two examples of radar data over
North Carolina. There had been older sinkhole activity in the anomalous conditions. Figure 16.10a shows the development
area for some time as indicated on topographic maps. Recent of a small diameter piping feature breaking through the silty
sinkhole activity began after construction of a nearby dam to clay layer and migrating upward through the quartz sand.
create an artificial lake for a housing development. This The white area in the radar record is caused by loose material
resulted in a significant increase in hydraulic head near the that provided fewer reflections and indicates the zone of pip-
railroad track triggering the appearance of springs and ing. This piping feature is estimated to be about 1 m in

Typically
<1 meter
diameter

5
Marker bed at
Depth in Meters

top of silty clay

at about 6 m
10

15

20

Fig. 16.9 The geologic conceptual model of conditions based upon borings and reconnaissance radar data. The cavity zone within the limestone
was below the penetration depth of the radar, which was limited to about 10 m
16.5 Application of Surface Geophysical Methods 177

a 0
b
0
1.5 1.5
3 3
Note proximity
of piping to
surface White area indicating

Depth in Meters

Depth in Meters
5 disturbed soils and 5
piping
Narrow white
area indicating 6
soil piping 6
Silty Clay
Silty Clay Marker Bed
Marker Bed
7.5 7.5
Note narrow
weak spot in top
of silty clay 10 Break in top of silty clay
10
with layers dipping
toward center of piping

Fig. 16.10 The radar data from two of the anomalies showing the of soil piping where the silty clay and the sand layers have begun to dip
details of soil piping before it reached the surface, initial soil piping, a downward (b)
narrow feature <1 m in diameter (a), and more advanced development

diameter and is seen to be within 1.5 m from the surface. 16.5.1.2 Sinkhole in a Sand Quarry
Figure 16.10b shows a more developed piping feature. Here (Central Florida)
the top of the silty clay layer has been deflected downward A hydraulic dredge was used to excavate sand at a quarry
and the piping is much more developed than that seen in in central Florida. During operations water levels began to
Fig. 16.10a. This piping feature has also progressed to within drop rapidly due to a sinkhole that developed in the bottom
a meter or so of the surface. The strong reflector at a depth of of the quarry. Water levels had already dropped by more
about 6 m is a silty clay horizon, which was found consis- than 3 m (Fig. 16.11a) before an investigation was initiated
tently throughout the area of investigation and provided an to identify the location of the sinkhole throat and remedi-
excellent marker bed. ate the conditions. Meanwhile dredging operations
The cavities that were causing the sinkholes were located continued.
within the limestone at depths greater than 12 m (beyond the A review of the recent dredging operations records was
site-specific depth of penetration of the radar). Yet, the data used to focus in on the most likely area to start the search. An
clearly show evidence of soil piping and a dipping clay layer echo sounder was used to map the bottom profile of the
in response to the deeper cavity within the limestone. These quarry along a series of profile lines. Figure 16.11b shows
shallow near-surface indicators (NSI) clearly identify the the bottom profile over the sinkhole. A weighted line was
location and activity of the deeper cavity system without used to determine the depth and exact location of the sink-
detecting the cavity itself. hole and a float was used to mark the location of the sink-
The radar data clearly identified the piping features before hole. Once the sinkhole was located, it was sealed using
they reached the surface. These anomalies could then be accu- sediments with a high clay content and cement. This simple,
rately located and a boring could be placed over the center of yet effective survey approach located the throat of the sink-
the small diameter anomaly that was then remediated by hole within about an hour.
grouting. This case history shows very high quality data and Since we had a reliable record of the last areas in which
the level of detail that can sometimes be obtained by radar the dredge had worked, the search was able to quickly focus
surveys. Such data is relatively easy to interpreted from unpro- in on the mostly likely problem area. However, if a large area
cessed records with little independent supporting data. The was to be searched or there was great uncertainty in location,
radar data at this site were of excellent quality due to the pres- other methods would have been more appropriate such as
ence of a clean quartz sand at the surface. The radar survey side scan sonar. Side scan sonar can cover a large area of the
was repeated every 6 months over a two-year period to moni- bottom in a single pass and provide a high resolution image
tor conditions for any collapse activity during construction of of the bottom.
a major railway by-pass system (Benson and Yuhr 1987).
178 16 Surface Geophysical Methods

Note change
in water Level

Original Surface of
Water
b
Water surface at
capability of dredging

time of survey
Maximum depth

0
to top of clay

Depth Below Water Surface (meters)

5
Sand

Clay Horizon
10

Maximum depth of sinkhole collapse

measured at 20 meters by lead line
Sinkhole Collapse in bottom of Quarry
15

distance approx. 60 meters

Fig. 16.11 The quarry shows the original water level along the quarry wall (a) and the bathymetry data that clearly shows the location of the recent
sinkhole in the bottom of the quarry (b)

16.5.1.3 Seepage at an Earthen Dam A visual inspection of the site revealed conditions that
(New Mexico) had not been previously identified, including small periodic
An investigation was initiated due to seepage and small col- fractures perpendicular to the dam’s axis along with small
lapse features at an earthen flood-water retention dam near localized soil piping features. The client had requested the
Carlsbad, New Mexico. There had been four different boring use of ground penetrating radar, which was a new popular
investigations at the site, but none had identified the prob- method at the time and the initial effort focused upon acquir-
lem. The owner was at the point where a decision had to be ing radar data. While radar had limited penetration due to the
made on whether to repair or breach the dam. At the time of silty clayey sediments, it did provide indications of shallow
this work the dam was dry both up and downstream. cavities within the alluvial soil, associated with soil piping.
16.5 Application of Surface Geophysical Methods 179

The radar data (Fig. 16.12a) shows an anomaly about 0.3 m with fractures in the underlying gypsum, which contained
below the surface. Several of the radar anomalies were exca- more moisture. Trenches at the upstream toe of the dam
vated and found to be cavities due to soil piping (Fig. 16.12b). uncovered fractures of 0.15–0.3 m wide in the gypsum that
While the radar penetration was limited to about 2 m it pro- contained some water confirming our interpretation. Not
vided a means of identifying near surface indicators (shallow encountering these very narrow vertical fractures by four
soil piping), which were related to deeper problem areas, but boring programs was understandable.
their cause remained unresolved. Finally a series of survey lines were run parallel to the
An extended site walkover revealed a dry streambed dam both up and downstream. The data (Fig. 16.13) identi-
upstream from the dam that terminated into a small sink- fied localized increases in conductivity on consecutive paral-
hole. Discussion with the local geologist indicated that the lel survey lines showing linear trends upstream, and
area was underlain by gypsum. Because radar penetration downstream of the dam. Note the correlation of some frac-
was limited to 2 m we now considered an alternative method tures over distances of 60 m or more and the spatial changes
and chose electromagnetic (EM34) measurements using a in fracture patterns of 30 m or less. The results clearly identi-
10-m coil spacing providing data to a depth of about 15 m. fied a group of fractures within the gypsum running under
A series of station measurements were made along the the dam. The cause of the seepage had been identified and
upstream toe of the dam, which revealed significant varia- sufficient data was provided to allow a decision to be made
tion in conductivity (Fig. 16.5a). Then a series of continuous on whether to repair or breach the dam.
EM34 measurements were made using a truck mounted sys- This example illustrates the importance of flexibility to
tem. The continuous measurements revealed the presence of switch methods in the field, selecting the appropriate method
localized areas of high conductivity (Fig. 16.5b). These to solve the problem. It also illustrates the benefit of the use
areas of high conductivity were thought to be associated of “continuous” measurements whenever possible.

Fig. 16.12 A shallow radar anomaly was seen within the alluvium (a), a trench was cut exposing a small cavity over a soil fissure (b)
180 16 Surface Geophysical Methods

Approx. 30 meters
100

(distance in feet)
Downstream
50

Crest of Dam

50

100

150
Upstream (distance in feet)
Approx. 150 meters

200

250

300

350

400

450

500
1600 2000 2400 2800 3200 3600 4000 4800
Distance in Feet
Approx 853 meters

Fig. 16.13 Continuous electromagnetic data was plotted as parallel survey lines both up and down stream of a flood retention dam. The linear
trends in the EM data identified fracture zones within the gypsum, which continued under the dam and were verified by trenching
16.5 Application of Surface Geophysical Methods 181

16.5.2 Conditions Within the Epikarst, 16.5.2.2 Landfill Expansion (Florida)

Top of Rock or Rockhead A landfill in west central Florida was undergoing expansion
when a small sinkhole developed in an excavation.
The epikarst or the “skin of the karst” is the dissolutionally Electromagnetic measurements (EM) were used to map vari-
weathered upper portion of the bedrock (Figs. 3.12 and ations in the top of rock. The EM data was acquired using an
3.13). This zone is characterized by its high degree of vari- EM31 instrument that measures bulk electrical conductivity
ability in the depth of dissolution (i.e. its thickness) and the to a depth of about 6 m (Fig. 16.15) and indicated a highly
degree and type of sediment fill within the zones of dissolu- variable bedrock topography. This was similar to that
tion. The top of rock profile (the unweathered rock) or rock- observed at a nearby excavation which had exposed the top
head is a common concern to engineers for excavations, piles of rock. Where the conductive clayey soil is thicker the EM
and foundations. This zone is also of concern when dealing values are higher. The closely spaced changes in conductiv-
with contaminants because of its extreme geologic and ity values are due to clay-filled fractures that occur at an
hydrologic variability. If the top of unweathered rock is rela- approximately spacing of 9–15 m. Another major trend in
tively shallow, less than 15 m or so below grade, there is a the conductivity occurs at approximately 120 m interval and
good chance at defining the epikarst and the top of rock pro- suggests deeper weathered zones or deeper rock where there
file using the surface geophysical methods. The following was more conductive clayey soil present.
are examples of characterization of this zone and include: The original top of rock contour was developed from a
limited number of borings (Fig. 16.16a). An EM survey was
• A diesel fuel spill along a highway in Kentucky, completed using a series of parallel survey lines over a por-
• A landfill expansion in Florida, tion of the landfill. The resulting conductivity contour map
• A site of proposed power plant in Florida, and showed variability across the site. Fourteen boring locations
• An expansion of a power plant in Alabama. were then selected based upon the EM contour map to pro-
vide depths to rock for correlation purposes. This data was
then used to develop a revised top of rock contour map
16.5.2.1 Diesel Fuel Spill Along (Fig. 16.16b). A remedial plan was designed that focused on
Highway (Kentucky) the fractures and deep areas of rock which were grouted to
A diesel fuel spill of 14,000 l occurred along a section of a stabilize conditions and eliminate future subsidence or col-
highway in Kentucky and there was obvious concern about lapse beneath the landfill.
cleanup. Three 2D resistivity survey lines of data were
obtained, one on each side of the divided highway and one in 16.5.2.3 Site of a Proposed Power Plant (Florida)
the median (each about 23 m apart). An electrode spacing of A site was being investigated for a potential power plant in
3 m was used to provide a reasonable degree of resolution. west-central Florida. Seismic refraction survey lines were
Figure 16.14 shows the three parallel lines of 2D resistiv- located through the site based upon aerial photos and physi-
ity cross sections obtained. Unweathered rock with high cal access. The purpose of the investigation was to define
resistivity values on the order of 600 ohm-m or more is seen the top of rock in areas of potential site development.
at depth. Weathered rock had resistivity values between 100 Seismic refraction tomography data was acquired using a
and 600 ohm-m and extends from near the surface to depths geophone spacing of 3 m. The data shows a high degree of
of more than 12 m. Zones of clay and increased moisture variability in the depth to the top of unweathered rock that
with low resistivity values, less than 100 ohm-m, are scat- was interpreted as the 7,000 ft/s contour. Three significant
tered near the surface and in pockets within the weathered lows in the top of unweathered rock are located along the
rock. The resistivity data was successful at this site because line shown in Fig. 16.17. A 30 m deep boring was located
of the significant electrical resistivity contrast between the just north of the survey line near station 410 and did not
overlaying sediments (with low resistivity) and the limestone encounter any rock. These data were used in part to evaluate
rock (with higher resistivity). The top of unweathered rock the suitability of the site and plan locations for the large
varied between 3 to more than 14 m below grade, providing structures associated with a power plant.
a thick and highly variable epikarst zone (Stephenson et al.
2003). 16.5.2.4 Expansion of a Power Plant (Alabama)
This example illustrates the highly complex nature of the At a proposed power plant site, seismic refraction data were
epikarst zone and the difficulty of sampling and remediation used to characterize the top of rock. The site is underlain
within such geologic conditions. With insight from the 2D with relatively hard limestone containing linear weathered
resistivity survey, boring locations and sampling were zones based upon photo-lineaments. A high-resolution seis-
optimized. mic refraction tomography survey was used to define the top
182 16 Surface Geophysical Methods

Unweather Limestone Ground Surface

Low Moisture Content Note 4X vertical exaggeration

Model Data Point

Weathered Limestone Observed Vertical Cavities in Rock

Clay and Highly

Weathered Limestone
Higher Moisture Content
190
Elevation in Meters

180

170
0 30 60 90 120 150 180 210 240
190
Elevation in Meters

180

170
0 30 60 90 120 150 180 210 240
190
Elevation in Meters

180

170
0 30 60 90 120 150 180 210 240
Distance in Meters

Fig. 16.14 Three parallel 2D resistivity cross sections illustrate both the lateral and vertical complexity of the epikarst. The extreme geologic
variability is seen along each survey line as well as between the three survey lines, which are about 23 m apart

of rock and weathered zones within the limestone. Almost (Fig. 16.18), which clearly identified two erosional
5 km of data were acquired over the area of interest, with a lineaments. The resulting data allowed borings to be posi-
geophone spacing of 1.2 m. The resulting data were con- tioned in anomalous and background locations and were
toured to show the depth to top of unweathered rock used to aid in planning development at the site.
 80
West Clay-filled fractures at a East
9 to 15 meter spacing

60

Conductivity (mS/m)
40

Major trend associated with

the depth to top of rock

20

270 330 390 450 480

Horizontal Distance (meters)

Fig. 16.15 Electromagnetic data along one of the profile lines at the base of a landfill excavation shows highly variable conditions associated with
clay-filled fractures within the rock as well as variations in the depth to top of rock

a
90
80

70
N b
10 West
330
East
50
90
80
20 70
60 50
70
50 70 90
50 70 90 270
30

Distance in Meters
Distance about 610 m

40
90
90
70
30
210
20 50
50
50 70
40 90
150

30
270 330 390 450 510
20 Distance in Meters

10
Landfill Boundary

Distance about 450 m

Original top of bedrock contour map Revised top of bedrock contour map based upon
(contours in feet) electromagnetic (EM31) data and new borings
(contours in feet)

Fig. 16.16 A top of rock contour map was based upon widely spaced borings (a), a more detailed top of rock contour map was (b) based upon
electromagnetic data (Fig. 16.15) and 14 new borings (contours in feet)
184 16 Surface Geophysical Methods

Test Boring
Extended to 30 meters
with no rock encountered

0
Depth in Meters

12

18
0 60 120 180 240 300 360 420 480
Distance in Meters

Fig. 16.17 The seismic refraction data indicates extreme lateral changes in the depth to top of unweathered rock based upon the 7,000 ft/s P-wave
velocity contour line

Erosional
Lineaments N
600

30
Distance in Meters

20

300

10

Seismic Refraction
Survey Lines 0
Depth to
Top of Rock
(meters)
0
0 300 600
Distance in Meters

Fig. 16.18 The depth to unweathered bedrock based upon seismic refraction data reveals two major southwest to northeast erosional lineaments

16.5.3 Fractures and Cavities Within the Rock These features are more difficult to detect yet may have a
significant impact to a site. These features can be air-filled,
Below the epikarst zone within the unweathered rock, our water-filled or soil-filled, can be closely or widely spaced,
focus shifts to fractures, cavities, conduits and cave systems. or even random. Larger caves and conduit systems will
16.5 Application of Surface Geophysical Methods 185

typically occupy less than 5 % of the surface area and are used from 1946 through 1969 for storage of residues from
difficult to locate by borings alone (Quinlan J 1999 personal uranium ore processing. Most of these residues were removed
communication). The following examples illustrate how the and building facilities were demolished and buried on-site.
surface geophysical methods can be utilized to locate and The site was then covered with up to 1 m of clean backfill.
characterize these features. These examples include: The state of Missouri is known to have well developed karst
features including many large sinkholes, caves and spring
• Identifying fractures along a new dam alignment in systems. The Missouri Department of Natural Resources
Alabama, (DNR) had raised the issue of potential karst underlying the
• Identifying a fracture zone at a low-level radioactive site site.
in Missouri, The area of investigation encompasses the original 9 ha
• Sinkhole collapse at a military site in Guam, plus surrounding areas for a total of about 67 ha. Previous
• Deep conduits in the Woodville Karst Plain in Northern work at the site had not addressed any aspects of karst. Of
Florida, and the 80 existing borings only 2 extended into limestone a
• Detecting a large cave room under a road widening in short distance, less than 3 m (Fig. 12.11). As part of the
Alabama. DOE Expedited Site Characterization Technology Program,
a site characterization to assess karst conditions was carried
16.5.3.1 New Dam Alignment (Alabama) out but without the benefit of additional borings. This
Anomalous seepage at a dam had been on-going since its included a review of available data, aerial photo analysis, a
construction in the late 1960s. A new dam was being built site walkover and observations of outcrops, microgravity
downstream. A detailed site characterization was being com- data and time domain EM soundings (TDEM), geophysical
pleted for the new dam that included identifying and remedi- and gamma spectrometer logging of monitor wells as well
ating of karst features. An initial resistivity survey identified as soil sampling and analysis.
anomalous areas within the rock and had been excavated One component of the site characterization was a cross
exposing the bare limestone. More detailed resistivity data section created from a single survey line of 50 TDEM sound-
was acquired using two parallel 2D resistivity survey lines ings. The TDEM sounding measurements were spaced 15 m
run over the anomalous area with an electrode spacing of apart. Each 1D sounding was interpreted using a four-layer
1.5 m. The resistivity data (Fig. 16.19) illustrates the location resistivity model. The 50-1D models were then combined to
of a major fracture zone of about 6 m wide dipping to the create a 2D profile across the site (Fig. 16.20). This resistiv-
south. Both 2D resistivity lines indicated similar results. ity profile identified the presence of a major fracture zone.
Lower resistivity values (<400 ohm-m) are probably associ- The TDEM measurements were responding to the contrast
ated with the presence of weathered rock and increased between the thicker clay and moisture within the fractures
moisture within the fracture zone. (low resistivity) verses the surrounding unweathered lime-
stone (higher resistivity). This data spatially correlates with
16.5.3.2 Low Level Radioactive Site (Missouri) the lineament analysis from aerial photos and logging data
The St. Louis Airport Site (SLAPS) is a 9 ha area located on from existing wells and helped to improve the understanding
the north side of the St. Louis airport. The SLAPS site was of site conditions.

Possible
Fracture Zone
159
Elevation in Meters

156

153

150

147

0 30 60
Distance in Meters

Fig. 16.19 2D resistivity data identified a dipping fracture zone along a proposed dam alignment
186 16 Surface Geophysical Methods

Generalized
Geologic
West East
Layer 1
Column
0 Surficial Soils (Layer 1)
Sand (Layer 2)

30 Silty Clays (Layer 3)

Depth in Meters

Fracture
Zone
60

Limestone (Layer 4)

90

120
0 300 600 900
Distance in Meters

Fig. 16.20 Cross section showing 50 four-layer TDEM sounding models defining the top of limestone and a fracture zone within the bedrock

16.5.3.3 Sinkhole Collapse at a Military

Site (Guam)
A high security military facility was located on a cliff (almost
180 m high) on the northwest coast of Guam. The facility
was experiencing localized sinkhole activity (Fig. 16.21).
Areas of surface collapse were typically 1–2 m in diameter
opening into small cavities within the limestone rubble fill
used as soil cover. Concern arose regarding possible subsid-
ence or collapse that could impact the critical military struc-
tures at the site. Radar was selected to survey the site using
an 80 MHz antenna. The survey objectives were to identify
cavities, joints and other areas of concern, which might
impact the structures. The radar data quality was excellent
and provided a depth of penetration to 24 m due to the dry,
clean limestone conditions (Fig. 16.22a). However, the range
was reduced to 12 m to provide higher resolution data within
the shallower zone of concern (Fig. 16.22b). A series of par-
allel radar survey lines 7.5 m apart were run oriented perpen-
dicular to the cliff.
Borings were then located based upon the cavities identi-
fied in the GPR data (hyperbola). Each air rotary boring was Fig. 16.21 One of many sinkholes at a military site on the northwest
geologically logged then a borehole video system was used coast of Guam
to investigate the vertical and lateral extent of any fractures
or cavities encountered. Cavities encountered ranged from
1.8 to 3.6 m in diameter. Small joints (0.3–0.6 m wide) collapse to structures at the facility. The combination of
within the rock appeared to be allowing soil to migrate into radar data, borings and borehole video data provided
the cavities, causing localized collapse features. increased confidence levels in the interpretation of sub-
While large, deep vertical shafts in the limestone are surface conditions. A remedial program was designed and
seen on Guam, this problem was not a deep-seated karst undertaken to stabilize the surface conditions by grouting
problem and there was no eminent hazard of major within the upper 12 m.
16.5 Application of Surface Geophysical Methods 187

a Approximately 60 meters
0

Depth in Meters
12

18

24

b Approximately 60 meters
0

Depth in Meters
6

12
Note hyperbola shaped reflectors indicating cavities

Fig. 16.22 Radar data was initially acquired to a depth of approximately 24 m (a) while it was of excellent quality only the upper 12 m were of
interest for engineering purposes (b). Note the numerous hyperbola (possible cavities) were identified in the upper 12 m of limestone

16.5.3.4 Woodville Karst Plain (Florida) One of the many features being mapped within the WKP
The Woodville Karst Plain (WKP), previously introduced in is the extensive underwater cave and conduit system. Over
Sect. 7.2.2, covers approximately 1,200 km2 and is located in 63 km of caves and conduits have been mapped by cave div-
northern Florida between Tallahassee and the Gulf of ers (Wisenbaker et al. 2007) but it may only represent a small
Mexico. The abundance of sinkholes provides direct connec- portion of the total conduits and caves within the WKP. There
tions between the surface and groundwater making the was a need to better characterize the conduits that feed the
groundwater vulnerable to contamination. Groundwater con- springs in order to develop more accurate hydrologic under-
tamination due to nitrates has been blamed for the declining standing and, eventually, numerical models to predict
water quality conditions in Wakulla Springs (Kincaid et al. groundwater flow and contaminant transport within the
2005). Woodville Karst Plain.
188 16 Surface Geophysical Methods

A variety of geophysical measurements were made at the microgravity data shows a clear response due to the mass
Leon Sinks area in order to evaluate their effectiveness for deficits of the conduits. Both methods also show a broader
mapping large conduits. These measurements included response from the deeper, discrete conduit at station 155,
microgravity and spontaneous potential (SP) measurements indicating a larger area of weathering, fracturing, etc.
as well as several other techniques. Two large conduits were extending to the east.
known to run perpendicular under the road at this site. Their Microgravity measurements were selected as the most
depth and size were known from divers and their location technically appropriate and cost-effective method to identify
had been established by cave radio. This site provided an large deep conduits over a large portion of the WKP area.
excellent location to test the response from various geophys- Over 44 km of microgravity data with station spacing of
ical measurements over known conditions (Yuhr et al. 2008). 30 m were acquired along survey lines upgradient of Wakulla
Figure 16.23 shows both the SP and microgravity data Springs. The purpose of the study was to identify potential
collected at this site. Both of these techniques were used to large conduits not previously identified or mapped.
acquire measurements at a spacing of 6 m and both sets of Significant microgravity anomalies were identified that may
data were successful at this site. The SP data shows a clear be related to large karst conduits 60 m deep or more
response to the flow within the conduits and the (Kaufmann and Dehan 2007).

SP anomalies due to large

conduits and water flow
a

- repeat measurements
0 60 120 180 240
Distance in Meters

Water-Filled Conduit Water-Filled Conduit

5m diameter, 18m deep 9x12m, 60m deep
b West East
60
Residual Gravity (uGals)

40
20
0
-20

-40
-60
0 60 120 180 240
Distance in Meters

Fig. 16.23 SP (a) and microgravity measurements (b) over known the data. Note that the deeper cave system to the right has a broader SP
water-filled conduits at the Leon Sinks site. The triangles show repeat and gravity anomaly
SP measurements as a means to check noise level and repeatability of
16.5 Application of Surface Geophysical Methods 189

An example of the production gravity data (Fig. 16.24a) data shows a physical low in the bottom conditions at each
clearly shows the presence of the three known conduits along spring. The marine 2D resistivity data shows an area of lower
with another deeper conduit, which had not been previously resistivity around each spring indicating a zone of weath-
identified. A cross section (Fig. 16.24b) shows the mapped ered, more permeable rock.
conduits and their dimensions along with an additional large
conduit modeled from the gravity data. The results of these 16.5.3.5 Roadway Widening (Alabama)
surveys aided in the placement of new borings (Yuhr et al. A road was being widened in Madison County, northern
2008) and advance the on-going hydrogeologic characteriza- Alabama. A portion of the expanded roadway crosses over
tion of the karst system to refine the hydrogeologic models Burwell Cave, which had been mapped in the 1960s. The
of the area (Kincaid et al. 2012). cave contains a large room 14 m high and 5.5–6 m wide. The
In addition, as part of the Woodville Karst Plain (WKP) ceiling of the cave was expected to lie about 3 m below the
study, marine geophysical measurements were made in the proposed road expansion. The exact location of this room
Spring Creek area along the coastline to locate springs. The was not known and access to the cave was no longer avail-
Spring Creek area is a tidal stream in Apalachee Bay and able. Therefore, surface geophysical measurements were
contains up to 14 submarine springs within its lower reaches. acquired to determine the location of this large room in the
Figure 16.25 shows a portion of the marine 2D resistivity and cave. The microgravity data (Fig. 16.26) and the 2D resistiv-
echo sounder data over two of the many springs. Both sets of ity data (Fig. 16.27) show the response from the large cave
data show a response from the springs. The echo sounder adjacent to the road. The combination of the microgravity

Fig. 16.24 Microgravity measurements (dots) were made over known known cave systems an additional low gravity anomaly at about station
conduits. An inverse model using the gravity data and known geologic 760 was detected and modeled. (b) shows the depths and dimensions of
conditions (solid line) are shown for comparison (a). In addition to the the known and estimated conduits
190 16 Surface Geophysical Methods

Bottom Depth from

Echo Sounder
Spring Spring
North South
0
Water Depth (m)

12

0 30 60 90 120 150 180 210 240 270 300 330 360

Distance in Meters

Fig. 16.25 The two of the many springs within Spring Creek were located both by echo sounder data (see dots in the upper edge of the image)
and 2D marine resistivity data. The resistivity data suggests a zone of weathered, more permeable rock around each spring

Fig. 16.26 The microgravity data parallel to a proposed roadway expansion detected the location of the large cave room

and resistivity data provided optimum locations for borings 16.5.4 Buried Sinkholes and Paleokarst
to determine exact cave location and depths.
Both microgravity and resistivity measurements were We commonly find paleokarst features on topographic maps
successful because the cave was large and very shallow. The or in historical aerial photos as sinkhole lakes, circular lows
site, while wooded, was relatively flat and required no major in topography or old collapse features. However, paleokarst
topographic corrections for the microgravity data. In addi- features can also include buried sinkholes or collapse fea-
tion, no major utilities, fencing or guardrails were present to tures that have no surface expression. These features may be
provide interference for the resistivity data. stable from a collapse point of view, but they can impact a
16.5 Application of Surface Geophysical Methods 191

Microgravity line
crosses perpendicular

WEST EAST
0
6
Depth in Meters

12

18

24

30
36
0 30 60 90 120 150 180
Distance in Meters

Fig. 16.27 To further assist in the location of the cave a 2D resistivity survey line was run perpendicular to the road and the gravity survey line.
The resistivity data shows the location, lateral extent and the approximate depth of the large cave room about 9 m below grade

site in other ways. These features may present areas of quartz sands. A marker bed of sand with green-gray silt/clay
weaker materials impacting construction or increased per- is seen at a depth of about 6 m. The radar data clearly defines
meability that could impact groundwater flow and future the marker bed dipping downward on both sides of the house
settlement. In addition, there is also a small risk for reactiva- identifying the presence of a paleosinkhole. Radar data is
tion of paleocollapse features. All three of our case histories one of the recommended methods of investigation in the
in Part III include the detection and characterization of a bur- State of Florida (Schmidt 2005).
ied paleocollapse feature with little or no surface expression. In this case, the house happened to be built directly over a
The examples for this section include: paleosinkhole filled with sand that had no surface expression
prior to building. The presence of a cavern within the deeper
• A residence in central Florida, limestone that caused the paleosinkhole is at a depth, beyond
• A proposed hazardous waste site in southwest Texas, the radar data penetration and could be at depths of 15–30 m
• A tunnel boring under a shipping channel in south Florida, or more. The cause of the recent settlement is associated with
and a pumping well located alongside of the house resulting in
• Radiation halo over paleo sinkholes in central Florida. fluctuation of the local water table, possibly aided by con-
centration of surface runoff from the house roof, patios or
driveways. The settlement is clearly related to the paleosink-
16.5.4.1 Residential Subsidence (Florida) hole under the house, however the degree of settlement is
One of the big concerns in Florida is establishing whether or uncertain. It may be a small amount of settlement that would
not a sinkhole is the cause of settlement in order to resolve eventually stabilize, or may be a precursor to a reactivation
insurance claims. When a sinkhole is obvious at the surface of this paleocollapse feature. In either case, the radar data
(Fig. 3.2a) the answer is clear. However, without the pres- has clearly identified the basic issue of whether it is sinkhole
ence of an obvious sinkhole, the cause of subsidence often related or not. Further assessment and monitoring was
becomes a debate between the owner, their insurance com- required in order to determine the potential impact of the
pany, their lawyers and the various consultants. The State of paleosinkhole and its possible remediation.
Florida has attempted to address these complex issues in
regulations (Florida Senate 2010). 16.5.4.2 Proposed Hazardous Waste Site (Texas)
In some cases, the problem can be resolved with the aid of An investigation was being carried out to evaluate a property
surface geophysical methods. Figure 16.28 shows the two regarding its suitability for a hazardous waste site in south-
parallel radar profile lines run along opposite sides of a house west Texas. A previous site about 8 km to the east had been
in which the cause of settlement was in question. Excellent abandoned due to the presence of paleokarst and a large
quality data was obtained to a depth of 9 m due to the clean cave. It had been proposed that this alternate site would not
192 16 Surface Geophysical Methods

N
Residence
Road
a
0
Approximate Depth in Meters

Dipping Marker Bed

(silty to clayey layer)
6

12

b 0
Approximate Depth in Meters

Dipping Marker Bed

(silty to clayey layer)

12

Fig. 16.28 Radar data acquired on either side of a house show clear evidence that the house is located over a paleosinkhole. (a) Radar data run
along west side of residence. (b) Radar data run along east side of residence

have any karst, and the objective at this alternate site was to selected as a rapid and economical means to characterize the
locate areas where the alluvium was thickest to provide alluvium thickness over this large area. Six lines 3,300 m
ample material for excavation and fill. The initial area inves- long were run north-south, spaced 900 m apart. These widely
tigated for the site selection was 3,300 by 4,500 m. spaced lines and measurements provided initial reconnais-
Electromagnetic (EM34) sounding measurements were sance data over this large area.
16.5 Application of Surface Geophysical Methods 193

EM34 measurements were made to depths of 15, 30 and the southeast of the site (Freeman 1968). These features
60 m along each of the survey lines at intervals of 150 m. were clearly identified by upturned rock around their bound-
Figure 16.29a illustrates the EM data from one of the sur- aries (Fig. 16.30) and were as large as 450 m wide and 5 km
vey lines. The interpretation is shown in Fig. 16.29b identi- long (Fig. 15.9).
fying an area of thicker alluvial sediments. The alluvial One of these paleokarst features had been identified in the
sediments were more electrically conductive than the mas- site walkover to the east of the site. The presence of upturned
sive limestone. Where the alluvial sediments become rock at the surface was observed along a railroad cut and
thicker, a higher electrical conductivity is measured by the marked the boundary of the paleokarst feature. There was no
EM34. Ultimately, it was determined that the area mapped other indication of subsidence or collapse at the surface.
on-site with thicker alluvial sediments was a buried stream Surface geophysical measurements (microgravity and elec-
channel. tromagnetic EM34 measurements) were acquired along a
The final site selected was 300 by 750 m. A detailed site single survey line over the feature and a single boring had
characterization was completed and included desk study, site been placed in the center of the collapse. The four sets of
walkover, EM34 measurements, microgravity, extensive bor- independent data (observations, a boring log, and two set of
ings, geophysical logging, borehole video logs and trenching surface geophysical data) provided both the lateral extent
(Yuhr et al. 1993). and depth of the paleocollapse (Fig. 16.31) and were used to
The desk study had identified a USGS report where exten- develop a reliable conceptual model of this paleocollapse
sive paleokarst features had been identified some 5–11 km to (Yuhr et al. 1993).

a 50
Vertical Dipole
10 m coil spacing (0 to 15 m deep)
40
20 m coil spacing (0 to 30 m deep)
Conductivity (mS/m)

40 m coil spacing (0 to 60 m deep)

30

20

10

0
0 600 1200 1800 2400 3000 3600

b
700
Elevation Meters (NGVD)

685
Alluvium

670

0 600 1200 1800 2400 3000 3600

Distance in Meters

Fig. 16.29 Electromagnetic (EM34) station measurements (a) detect an area of increased conductivity. The interpreted model (b) indicates
thicker alluvium centered at station 600, possibly a buried channel or paleokarst
194 16 Surface Geophysical Methods

Fig. 16.30 Frank Woodward (a) observes edges of a paleocollapse feature defined by upturned rock (b)

16.5.4.3 Tunnel Boring Under a Ship cemented materials that may present problems for the
Channel (Florida) tunneling machine.
An underground tunnel linking Interstate 395 on Watson This project was completed in May of 2014. Traffic is
Island and the Port of Miami on Dodge Island in Miami- now flowing through the tunnels which are 1,280 m long
Dade County, Florida had been proposed to alleviate traffic consisting of two parallel tubes 11 m in diameter and dipping
in downtown Miami. The proposed link would consist of two to about 30 m below sea level. This is one of the most expen-
bored tunnels crossing beneath Government Cut, which sive and elaborate transportation projects in south Florida
accommodates both passenger cruise ships and cargo ships history. Asked about technical problems during tunneling,
to the Port of Miami. the project manager said that it was the “softer zones” in the
A suite of marine geophysical methods were used to sea bed. They had to be frozen to 27° below zero so that the
investigate the conditions and determine engineering proper- tunnel boring machine could work effectively (Chardy
ties within Government Cut. The surveys included bathym- 2014). The “soft zones” were probably the areas identified
etry (echo sounder), sub-bottom seismic reflection, and by the geophysical measurements (Fig. 16.32).
seismic refraction. Six lines of survey data were obtained,
two of which were run along the center lines of the two tun- 16.5.4.4 Radiation Halo Over Sinkhole (Central
nels. All three methods provide spatial insight to subbottom Florida)
variability. Seismic refraction data identified variations in Radioactive measurements at the surface are based upon the
compression (P-wave) velocity. Areas of Low P-wave veloc- concept that deeper rocks release uranium and it’s by-
ities measured by the seismic refraction data correlated well products that migrate to the surface via various permeable
with low N-values (blow counts) obtained from existing SPT pathways Armstrong and Heemstra (1973). Upchurch et al.
borings. Figure 16.32 shows the seismic refraction data from (1987) found that gross-alpha radiation is correlated within
one of the survey lines. The top of “hard” rock is defined as 0.3 km of the axis of fracture traces in west-central Florida.
8,000 ft/s. The localized low velocity areas could represent Banwell and Parizek (1986) found similar results for radon
possible zones of sand and sediments in areas between mas- emanating from fracture traces in Pennsylvania. There has
sive coral reefs, or paleosinkholes filled with sediments. In been limited use of radioactive measurements for geotechni-
either case, these areas represent areas of loose or less cal applications and karst. Possible applications include the
16.5 Application of Surface Geophysical Methods 195

a
815.5

Microgravity (milliGals)
Gravity Anomaly
815.0

814.5

814.0
0 150 300 450
b
40
Conductivity (mS/m)

30

20

10
10 m vertical dipole (0 to 15 m)
20 m vertical dipole (0 to 30 m)
0
0 150 300 450 Upturned rock
Distance in Meters defines boundary
Test of paleocollapse
c Borehole

Approx. 45 m
Displacement
Buda Limestone

Segovia Limestone Del Rio Clay

Approx 100 meters Segovia Limestone

Rubble
Fort Terrett Limestone

Fig. 16.31 The boundaries of a paleocollapse feature could be seen on (EM34) measurements (b) were made over the collapse and a boring
the surface adjacent to the railway cut by upturned rock similar to that was drilled in the center, all resulting in the conceptual model (c)
seen in Fig. 16.30. Microgravity measurements (a) and electromagnetic

location of cave openings, and buried sinkholes as well as cent to the two sinkholes (Fig. 16.33b). Radioactive halos
man-made tunnels and mine shafts which provide a more have been found to be present at many other sinkhole sites,
permeable pathway to the surface. buried paleokarst and fracture zones in Florida due to areas
Radiometric measurements were made with a gamma ray of increased permeability, which allow upward migration of
spectrometer along a road passing near the edge of two sink- radioactive materials.
hole depressions in central Florida. Radar data was acquired
along the side of the road shows the edges of two sinkholes
(Fig. 16.33a). Florida sinkholes are commonly sand filled 16.5.5 Pseudokarst Conditions
providing a more permeable pathway for upward migration.
Total count measurements were made along the side of the Pseudokarst can be caused by a wide range of conditions,
road (off of the asphalt) to determine if radioactive halos typically man-induced, such as subsidence due to withdrawal
were detectable over the edges of the two sinkholes. The of fluids, collapse and subsidence associated with mining, or
radioactive measurements were made at 15 m intervals and cavities due to leaking acids. Regardless of the cause, the site
the resulting data showed two areas of elevated values adja- characterization commonly utilizes the same strategy as
196 16 Surface Geophysical Methods

WEST EAST
0
Depth in Meters

6 Water Depth

12

18

24
0 60 120 180 240 300
Distance in Meters
12000
11500
11000
10500
10000
9500
9000
8500
8000
7500
7000
6500
6000
5500
5000

P-wave Velocity
(ft/s)

Fig. 16.32 One of the six seismic refraction survey lines illustrates the variation in seismic velocity and depth of harder rock based on (8,000 ft/s
velocity)

Edges of Two Sinkholes

a
0
100MHz GPR Data
3

Relative Depth (m)

6

12

15
240 300 360 420 480 540 600 660
Distance in Meters
b
400 Elevated total counts over
edges of the two sinkholes
360
Total Counts

320

280

240

200
240 300 360 420 480 540 600 660
Distance in Meters

Fig. 16.33 Ground penetrating radar data using a 100 MHz antenna The total count gamma radiation measurements show elevated values
was acquired along the road, parallel to two sinkholes, the edges of the when passing the two sinkholes (b)
two sinkholes can be seen as dipping strata in the radar data (a).
16.5 Application of Surface Geophysical Methods 197

natural karst conditions. The following examples illustrate success (Kratochvil 1989) and finally resorted to trying
some of the applications of surface geophysics used to char- surface geophysical methods.
acterize pseudokarst conditions and include: In 1988, Kratochvil requested ground penetrating radar
data to detect the fissures. However, radar was not expected
• Fissures on an Air Force runway in California, and to be effective because of the clay content of the runway
• Mapping of an abandoned mine in Ohio. materials. But because the sediments were very uniform
along the runway and the fissures were relatively shallow
16.5.5.1 Fissures on Air Force (within 1 m or so), radar proved to be quite effective.
Runway (California) Figure 16.34b is an example of the radar data using an
Edwards Air Force Base is located north of Los Angles, 80 MHz antenna that for the most part provided an extremely
California, and has long earthen runways, (one as long as uniform radar response over most of the runways. When
11 km) for test fights of advanced aircraft and shuttle land- crossing a fissure the radar responds to the subtle change in
ings. Groundwater withdrawal from a deep aquifer has lead to conditions of void space and an increased moisture level.
the development of extensive surface fissures along the run- Radar was used as a reconnaissance tool and data was
ways (Ward et al. 1992). When the fissures break through the acquired along ten parallel survey lines that ran the length of
surface (Fig. 16.34a) they are repaired. However, when they each runway (Technos 1988).
lay hidden from view just below the surface, they present a The radar survey provided the approximate location of
hazard for aircraft whose wheels may encounter them, caus- anomalous conditions along the long runways (within a
ing breakthrough and possible damage to the aircraft. meter based upon a distance counter on the survey vehicle).
After their mission, the Space Shuttle would sometimes After being located by radar, a special cone penetrometer
land at Edwards Air Force Base. The shuttle’s 86,184 kg hits was used to confirm the presence, width and nature of the
the bare earth desert floor at 340 km/h. The surface had been fissure (more details are provided in Sect. 17.2.2). A reme-
breached on three separate occasions due to the weight of diation team then carried out repairs of the fissures.
aircraft; one of these occurred when the space shuttle landed
(Kratochvil et al. 1992). 16.5.5.2 Mapping Abandoned Mines (Ohio)
Colonel Kratochvil, who had just returned from support- Abandoned underground mines are quite common through-
ing the USA tunnel detection program at the demilitarized out the United States. As infrastructure and development
zone in Korea, was assigned to NASA to evaluate the continue, the location of these abandoned mines becomes a
conditions. He had tried available airborne methods without concern. Maps of abandoned mines are commonly not

Fig. 16.34 A typical surface fissure along the earthen runways at Edwards Air Force Base California (a). Radar data obtained along the earthen
runways at Edward’s Air Force Base shows the location of a typical fissure before it has broken through the surface (b)
198 16 Surface Geophysical Methods

Fig. 16.35 The measured microgravity data (dots) and the modeled response (solid line) over a known abandoned mine (a). The location and
depth of the known mine is about 30 m deep and 3.4 m high based upon existing mine maps (b)

accurate or incomplete and often may not even exist. A high- location of a mine and could model both its depth and height
way realignment in Ohio required the location of a room and with reasonable accuracy (Butler 1980, 1994; Yuhr et al.
pillar gypsum mine to be defined. While old mine maps were 1993).
available and had been pieced together from various mining
companies, the exact location and height of the mines was in
question. Microgravity was used as one of the methods tested References
for its ability to locate the abandoned mines.
Figure 16.35a shows a microgravity profile over a portion Anderson N, Ismail A (2003) A generalized protocol for selecting
appropriate geophysical techniques. Geophysical Technologies for
of a known mine. The depth and thickness of the mine (30 m
detecting Underground Coal Mine Voids Forum, Lexington, KY
deep and about 3.3 m high) was known with a fair degree of Anderson N et al (2003) Selection of appropriate techniques: a general-
certainty and used to estimate what the gravity response ized protocol based on engineering objectives and site characteris-
would be over such conditions. The forward model estimated tics. Presented at the 3rd international conference on applied
geophysics, Orlando, FL, 8–11 Dec 2003
that the gravity survey should detect a large gravity anomaly
Armstrong FE, Heemstra RJ (1973) Radiation halos and hydrocarbon
(>100 microgals), (see solid line in Fig. 16.35a). The actual reservoirs: a review. US Department of the Interior, Bureau of
gravity data measured in the field is also plotted, (see dots in Mines, Information Circular 8579
Fig. 16.35a), and shows very good correlation. Figure 16.35b ASTM (2007) Standard guide for the use of the time domain electro-
magnetic method for subsurface investigation. Designation:
shows a cross section of the known mine based upon existing
D6820-02. American Society for Testing Materials, Philadelphia
mine maps. ASTM (2008) Standard guide for using the frequency domain electro-
This test phase was quite successful in determining that magnetic method for subsurface investigation. Designation:
microgravity measurements could effectively determine the D6639-01. American Society for Testing Materials, Philadelphia
References 199

ASTM (2010a) Standard practice for expedited site characterization of ronmental and engineering problems, EEGS, Wheatridge, CO,
vadose zone and ground water contamination at hazardous waste pp 531–540
contaminated sites. Designation: D6235-04. American Society for Benson RC et al (2003) Some considerations for selection and success-
Testing Materials, Philadelphia ful application of surface geophysical methods. In: 3rd international
ASTM (2010b) Standard guide for using the gravity method for subsur- conference on applied geophysics, geophysics 2003, Orlando, FL,
face investigation. Designation: D6430-99. American Society for pp 8–12
Testing Materials, Philadelphia Bogle FR, Loy K (1995) The application of thermal infrared thermog-
ASTM (2010c) Standard guide for using the direct current resistivity raphy in the identification of submerged springs in Chickamaunga
method for subsurface investigation. Designation: D6431-99. Reservoir, Hamilton County Tennessee. In: Beck BF (ed)
American Society for Testing Materials, Philadelphia Proceedings of the 5th multidisciplinary conference on sinkholes
ASTM (2010d) Standard guide for using the seismic reflection method and the engineering and environmental impacts of karst: karst geo-
for shallow subsurface investigation. Designation: D7128-05. hazards, Gatlinburg, TN. Balkema, Rotterdam, pp 415–424
American Society for Testing Materials, Philadelphia Burger HR (1992) Exploration geophysics of the shallow subsurface.
ASTM (2011a) Standard guide for using the seismic refraction method Prentice Hall, Englewood Cliffs
for subsurface investigation. Designation: D5777-00. American Butler DK (1977) Proceedings of symposium on detection on subsur-
Society for Testing Materials, Philadelphia face cavities. US Army Engineer Waterways Experiment Station,
ASTM (2011b) Standard guide for using the surface ground penetrating Vicksburg
radar method for subsurface investigation. Designation: D6432-11. Butler DK (1980) Microgravity techniques for geotechnical applica-
American Society for Testing Materials, Philadelphia tions. US Army Engineer Waterways Experiment Station,
ASTM (2011c) Standard guide for selecting surface geophysical meth- Miscellaneous Paper GL-80-13
ods. Designation: D6429-99. American Society for Testing Butler DK (1994) Detection and delineation of subsurface cavities, tun-
Materials, Philadelphia nels and mines. Am Soc Civil Eng Ill Sect Geotech Lect Ser
ASTM (2011d) Standard guide for using the metal detection method for 3:1–29
subsurface investigation. Designation: D7046-11. American Society Butler DK (ed) (2005) Near-surface geophysics. Society of Exploration
for Testing Materials, Philadelphia Geophysics, Investigations in geophysics No. 13, Tulsa, OK
Atekwana EA et al (2002) Geophysical investigation of vadose zone Butler DK (2008) Detection and characterization of subsurface cavities,
conductivity anomalies at a hydrocarbon contaminated site: impli- tunnels and abandoned mines. In: Xia J (ed) Near-surface geophys-
cations for the assessment of intrinsic bioremediation. JEEG ics and human activity. Proceedings of the 3rd International
7(3):103–110 Conference on Environmental and Engineering Geophysics
Banwell GM, Parizek RR (1986) Potential for permeability character- (ICEEG). Science Press USA, Monmouth Junction, New Jersey,
ization of crustal fracture zones using helium-4 and radon-222 con- pp 578–584
centrations in groundwater and soil gas. Geologic Society of Campbell CW, Singer M (2001) Application of thermography to
America program with abstracts, Boulder, CO, 18(6)534 groundwater monitoring at Arnold Air Force Base, Tennessee. In:
Beam P et al (1997) Lessons learned, a brief history of site characteriza- Barry BF, Herring JG (eds) Proceedings of the 8th multidisciplinary
tion. HazWaste World/Superfund XVIII, Washington, DC, 2–4 Dec conference on sinkholes and the engineering and environmental
1997 impacts of karst: geotechnical and environmental applications of
Benson RC (1997) ESC: just another new acronym or maybe some- karst geology and hydrology, Louisville, KY, pp 353–358
thing meaningful. Environ Eng Geosci 3(3):453–456 Cassidy DP et al (2001) The effects of LNAPL biodegradation products
Benson RC, Yuhr LB (1987) Assessment and long term monitoring of on electrical conductivity measurements. JEEG 6(1):47–52
localized subsidence using ground penetrating radar. In: Beck BF, Caterpillar (1988) Handbook of ripping, 8th edn. Caterpillar Inc.,
Wilson WL (eds) Proceedings of the 2nd multidisciplinary confer- Peoria
ence on sinkholes and the environmental impacts of karst, karst Chardy A (2014) Port tunnel dream finally becomes reality. Miami
hydrogeology: engineering and environmental applications. Herald, Sunday 18th May 2014
Orlando, FL. Balkema, Rotterdam, pp 161–169 DiMaggio (2007) Commentary: advancements and disappointments in
Benson RC, Yuhr L (1992) A summary of methods for locating and geo-engineering and geo-construction – a 30–year reflection. Geo-
mapping fractures and cavities with emphasis on the geophysical Strata July-August, p 12, 13, 26
methods. In: Proceedings of the symposium on the application of Florida Senate (2010) Interim report 2011-104 December. Committee
geophysics to engineering and environmental problems, EEGS, on banking and insurance, issues relating to sinkhole insurance
Chicago, IL, 26–29 Apr 1992, vol 2, pp 471–486 Franklin AG et al (1981) Foundation considerations in sitting of nuclear
Benson RC, Yuhr L (1996) An introduction to geophysical techniques facilities in karst terrains and other areas susceptible to ground col-
and their applications for engineers and project managers, short lapse. U.S. Army Engineers Waterways Experiment Station
course. Presented at SAGEEP, 28 Apr 1996, Keystone, CO Vicksburg Miss. Prepared for U S Nuclear Regulatory Commission
Benson RC, Yuhr L (1997) An overview of the geophysical methods for NUREG/CR-2062
geotechnical and environmental site characterization. Presented at Freeman VK (1968) Geology of the Comstock-Indian Wells Area, Val
ASCE GeoLogan conference, Logan, UT, July Verde, Terrell and Brewster Counties, Texas. USGS Professional
Benson RC et al (1984) Geophysical techniques for sensing buried Paper 594-k
wastes and waste migration. US EPA Environmental Monitoring Greenhouse J et al (1995) An introduction to application of surface geo-
Systems Laboratory, Las Vegas. Contract No. 68-03-3050 physics in environmental investigations. A short course presented a
Benson R et al (1985) Correlation between field geophysical measure- the symposium on the application of geophysics to environmental
ments and laboratory water sample analysis. Presented at National and engineering problems, environmental and engineering geophys-
Water Well Association and Environmental Protections Agency 2nd ical society, Orlando, FL
national conference on surface and borehole geophysical methods Hansen D (1975) Geological applications manual for portable gamma
in groundwater investigations, San Antonio ray spectrometers. Exploranium, Toronto
Benson RC et al (1998) ESC: how it differs from current state of the Kaufmann RD, DeHan RS (2007) Microgravity mapping of karst con-
practice. In: Symposium on the application of geophysics to envi- duits within the Woodville Karst Plain of North Florida. In:
200 16 Surface Geophysical Methods

Proceedings of the symposium on the application of geophysics to Sharma PV (1997) Enviromental and engineering geophysics.
environmental and engineering problems, environmental and engi- Cambridge University Press, Cambridge
neering geophysical society, April, Denver, CO, pp 1517–1526 Stephenson JB et al (2003) Environmental investigation of a diesel spill
Kincaid T et al (2005) Quantitative groundwater tracing and effective in a karst area of Kentucky. In: Barry BF (ed) Proceedings of the 9th
numerical modeling in karst: an example from the Woodville Karst multidisciplinary conference on sinkholes and the engineering and
Plain of North Florida. In: Beck B (ed) Proceedings of the 10th mul- environmental impacts of karst, ASCE Geo-Institute Geotechnical
tidisciplinary conference on sinkholes and the engineering and envi- Special Publication No. 122, pp 415–425
ronmental impacts of karst, vol 144. ASCE Geo-Institute Straley D (2008) How thermal imaging of caves got its start. NSS News
Geotechnical Special Publication, San Antonio, pp 114–121 66(5):16–17
Kincaid T et al (2012) Demonstrating interconnection between a waste- Technos, Inc. (1988) Final report, ground penetrating radar investiga-
water application facility and a first magnitude spring in a karstic tion of the shuttle landing zone, Rogers Dry Lakebed, Edwards Air
watershed: tracer study of the southeast Farm Wastewater Reuse Force Base, California. Prepared for NASA’s National space trans-
Facility, Tallahassee, FL. Florida Department of Environmental portation system, June, Consultants report
Protection, Florida Geological Survey Report of Investigation No. Thompson J, Marvin M (2006) Experimental research using thermog-
111 raphy to locate heat signatures from caves. In: Proceedings of 17th
Kratochvil GL (1989) Geotechnical evaluation of potential hazards to national cave and karst management symposium, Albany, NY,
shuttle landing operations on Rogers Dry Lakebed Edwards Air November, pp 102–115
Force Base, California. Astronaut Office NASA Lyndon B Johnson Titcomb EP Jr, Keeton JM (1984) Sinkhole activity in the vicinity of the
Space Center, Houston Sunny Point, NC Military Ocean Terminal. In: Beck BF (ed)
Kratochvil G et al (1992) Fishing for fissures. The Military Engineer Proceedings of the 1st multidisciplinary conference on sinkholes:
Jan–Feb, pp 66–68 their geology, engineering and environmental impact, Orlando,
Llopis JL et al (2005) Tunnel detection along the southwest US Border. FL. Balkema, Rotterdam, pp 344–352
In: Proceedings of the symposium on the application of geophysics Upchurch SB et al (1987) Spatial and temporal controls on gross-alpha
to environmental and engineering problems, environmental and radiation, central Florida phosphate district, natural radiation in
engineering geophysical society, pp 430–443 Florida. In: Proceedings of Natural Radiation and Technologically
McCann DM et al (1997) Setting the standard for geophysical surveys Enhanced Natural Radiation in Florida, Clearwater, Florida,
in site investigation. In: McCann DM et al (eds) Modern geophysics pp 50–61
in engineering geology, Geologic Society London, Engineering Ward SW (ed) (1990) Geotechnical and environmental geophysics.
Geology Special Publication No. 12:3–34, London, England Investigations in Geophysics No. 5, Society of Exploration
Milsom J, Eriksen A (2013) Field geophysics, 4th edn, The geological Geophysics Tulsa, Oklahoma
field guide series. Wiley-Blackwell, West Sussex Ward AW et al (1992) Geologic setting of east Antelope Basin with
Mitchell JK, Kavazanjian E Jr (2007) Geoengineering engineering for emphasis on fissuring on Rogers Lake, Edwards Air Force Base,
the 21st century. ASCE Geo-Institute, Geo-Strata July/August, Mojave Desert, California. In: Prince KR et al (eds) US Geological
pp 14–18 Survey Subsidence Interest Group Conference, Edwards Air Force
National Research Council (2000) Seeing into the earth, non-invasive Base, Antelope Valley, California, USGS Open File Report 94-532,
characterization of the shallow subsurface for environmental and pp 35–37
engineering application. National Academy Press, Washington, DC Whiteley RJ (1983) Recent developments in the application of geo-
Nazarian S, Diehl J (ed) (2000) The use of geophysical methods in con- physics to geotechnical investigations. In: Ervin M (ed) In-situ test-
struction. ASCE GeoInstitute Special Publication Number 108, ing for geotechnical investigations. Balkema, Rotterdam, pp 87–110
Reston, Virginia Wightman WE et al (2003) Applications of geophysical methods to
Olhoeft GR (2003) Quality control in geophysics. Geophysical tech- highway related problems. Federal Highway Administration,
nologies for detecting underground coal mine voids form, 28–30 Publication No. FHWA-IF-04-021, Lakewood, Colorado
July 2003, Lexington, KY Wisenbaker M et al (2007) Explorers connect America’s longest under-
Reynolds JM (2011) An introduction to applied and environmental geo- water cave to Wakulla Springs. NSS News 65(12):4–11
physics. Wiley, New York Yuhr L (1998) Managing data and its integrity in the ESC process,
Sauck WA et al (1998) High conductivities associated with an LNAPL Symposium on the application of geophysics to environmental and
plume imaged by integrated geophysical techniques. JEEG engineering problems, EEGS, 22–26 Mar 1998, Chicago
2(3):203–212 Yuhr L et al (1993) Characterization of karst features using electromag-
Saunders W et al (1999) Selecting surface geophysical methods for netics and microgravity: a strategic approach. In: Symposium on the
geological, hydrological, geotechnical, and environmental applications of geophysics to engineering and environmental prob-
investigations: the rational for the ASTM provisional guide, ASTM lems, Environmental and Engineering Geophysical Society,
D 6429-99(2011). In: Durham GM, Marr WA (eds) Proceedings for Englewood, CO, pp 209–228
field instrumentation for soil and rock, ASTM STP 1358, 18–19 Yuhr L et al (2008) Mapping and drilling of conduits within the
June 1998, Atlanta, GA, pp 392–403 Woodville Karst Plain of North Florida. In: Yuhr LB, Alexander EC
Schmidt W (2005) Geological and geotechnical investigation proce- Jr, Beck BF (eds) Proceedings of the 11th multidisciplinary confer-
dures for evaluation of the causes of subsidence damage in Florida. ence on sinkholes and the engineering and environmental impacts of
Florida Geologic Survey Special Publication No. 57, FGS karst. ASCE Geotechnical Special Publication No. 183,
Tallahassee, Florida pp 340–350
 Invasive Methods
17

Abstract
By now the conceptual model has become site-specific with increased details provided by
the results from the surface geophysical data. Both background and anomalous conditions
have been defined along with their lateral extent. Our focus now shifts to obtaining site-
specific geologic details along with engineering, hydrologic and contaminant data. We can
now, with a reasonable degree of certainty, develop an effective invasive program and locate
direct push measurements, borings, sampling points, trenches and excavations in a logical
way so that they provide representative measurements of background and anomalous geo-
logic conditions. Furthermore, by increasing the spatial representativeness of this data, we
can obtain sufficient data using fewer invasive measurements.

17.1 Introduction common practice has been to drill one boring in each corner
and one in the middle of a small site (Fig. 12.10). Regulatory
There are a wide variety of invasive methods to choose from. issues will sometimes tend to dominate the location of moni-
They include minimally invasive methods such as direct push toring wells with some states requiring layouts on a grid pat-
methods and cone penetrometer testing (CPT) to invasive tern while others impose minimum numbers of wells based
drilling methods, along with excavations and trenches upon area (Sara 2003). These are not unreasonable
(Fig. 17.1). These methods can provide a wealth of informa- approaches when no other information is available or likely
tion and data whether for geotechnical, groundwater resources, to be obtained.
or environmental contaminant projects. Recent decades have Without prior geologic knowledge the question of how
seen advancements in direct push technology and drilling many, what spacing, and how deep is not obvious, particu-
along with improved instrumentation for in-situ testing. These larly in karst. In addition, the interpolation or extrapolation
invasive methods and resulting samples provide a high level of of boring data can only be trusted, if the spacing between
detailed data for the site characterization process. Nielsen borings is less than the lateral changes in geologic conditions
(2006) provides more details on these topics. (which is most often an unknown). However, based upon the
Traditionally, the location of invasive measurements and surface geophysical measurements, we should have general
the spacing of measurements along a profile line or spacing understanding of the geologic spatial variability as well as
over a grid is always an issue. Many rules of thumb provided expected geologic features at this point. It is now possible to
in the literature depend upon whether the scope of the inves- locate invasive measurements more effectively.
tigation is an initial exploration or a detailed investigation.
Regularly spaced borings along an alignment or a fixed num-
ber of boreholes for a given area are often used in explora- 17.2 Direct Push Methods
tion. The spacing is then decreased when erratic conditions
are encountered. In any case, all invasive measurements The direct push methods include both percussion (using a rod
should be located in a manner to provide the best possible driven into the ground with blows from a hammer device) and
data needed for the investigation (Hvorselv 1949). Another the cone penetrometer (using a rod pushed into the ground

© Springer Science+Business Media Dordrecht 2016 201

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_17
202 17 Invasive Methods

with hydraulic ram using the weight of the truck). Both of Several ASTM standards have been developed which
these systems are in common use today and provide mini- cover cone penetrometer testing techniques. These standards
mally invasive characterization of subsurface conditions. were not developed for investigating karst sites specifically,
These two techniques are applicable in areas that have an but they do provide general information on the technology.
unconsolidated soil cover in which the probes can penetrate. Sara (2003), Hunt (2005), McCall et al. (2006), and Ruda
and Farrar (2006) all provide further details of these methods
Minimally Invasive Methods
and their use.

a Percussion Driven b Cone Penetrometer

Direct Push Methods 17.2.1 Percussion Driven Direct Push Methods

Percussion driven direct push technology can be used to rap-

idly characterize subsurface conditions in unconsolidated
materials. Direct push measurements are made by driving a
small diameter probe into the sediment using a hammer
device. The direct push percussion equipment is commonly
mounted on the back of a small truck (Fig. 17.2a) and can be
adapted in a variety of ways for access to unique site condi-
Invasive Methods
tions (Fig. 17.2c).
c Drilling d Trenching/Excavation The direct push methods eliminate any drilling fluids or
soil waste from the hole. This can be advantageous when
working at hazardous waste sites. The small diameter holes
can be easily grouted when the probe is extracted. Depths of
9–15 m are commonly obtained and in ideal conditions maxi-
mum depths of up to 30 m or so have been reached. Depth of
penetration is limited by the presence of dense sands, gravel
and cobbles. Penetration rates of 0.5–7 m/min are typical.
The type of data acquired includes the rate of advance-
ment into the sediment (blows/m or speed of penetration).
This is an indicator of the materials resistance to penetration
Fig. 17.1 Methods of minimally invasive (a and b) and invasive inves- by the probe and can be used to help characterize the
tigation (c and d) materials encountered and can often be tied to stratigraphy.

Fig. 17.2 Typical direct push equipment mounted on a truck (a), sediment samples obtained using direct push (b) and a modified portable direct
push system (c)
17.2 Direct Push Methods 203

Soil samples can be obtained (Fig. 17.2b) and an electrical depth of 3–6 m. The presence of this layer, its thickness and
tip can be used at the end of the probe can measure the elec- continuity were critical to the sinkhole risk assessment and
trical conductivity of the material to identify clays or pore contaminant migration. Where this layer was missing or
fluids (Christy et al. 1994; Schulmeister et al. 2003). These dipped deeper was potentially an indication of sinkhole activ-
push technologies are now commonly used for environmen- ity or paleokarst conditions.
tal site characterization, providing sampling of soils, water Direct push technology was used in conjunction with
and contaminants. See ASTM D6001 a guide for direct push ground penetrating radar to characterize the semi-confining
technology (ASTM 2012c). layer. Radar was used to acquire shallow data over the entire
The direct push method using an electrical conductivity site mapping the top of the semi-confining layer identifying
tip was used at the Central Nevada Test Site (CNTA) to char- lows or missing sections. While the radar data provided good
acterize sediments in old drilling mud pits. This area was lateral coverage of the site showing the top of this layer, it did
used in 1968 to test atomic bombs by detonation deep under- not provide details through the semi-confining layer due to its
ground. A large mud pit (about 1.2 ha) was a result of drilling clay content. Direct push technology with an electrical conduc-
a 2.4 m diameter borehole, 975 m deep for a test detonation. tivity tip was then used to provide data through the semi-con-
The mud pit materials were semi-consolidated and would fining layer. The electrical conductivity tip was very effective at
flex as one jumped upon the surface. The pit would have characterizing the semi-confining layer due to its clay content.
likely liquefied with the load of a truck and the vibration of a The direct push measurements provided a confirmation of the
percussion driven sampler. As a result, sampling was carried depth to the top of semi confining layer along with its thick-
out using a modified direct push system that only required ness, relative clay content, and lateral continuity. Figure 17.3
two people using a portable percussion hammer to obtain shows the correlation between the radar data and the direct
samples (Fig. 17.2c). push electrical conductivity data. The radar data shows a local-
In west central Florida the direct push percussion method ized low area in the semi-confining layer. Direct push electrical
using an electrical conductivity tip was used extensively at a conductivity data indicates the presence of clay on both sides
52 ha EPA superfund site. A thin layer of clay over weathered and within the low area. Higher clay content is indicated by
rock formed a semi-confining layer between the sandy surfi- increasing electrical conductivity values (to the right). This
cial aquifer and the underlying limestone Floridan aquifer at a case history is provided in detail in Chap. 27 of this book.
(Top of SCL

(Top of SCL

(Top of SCL
at 5.1 m)
DP-02-44

at 3.7 m)

DP-02-46

at 4.0 m)
DP-02-4

Depth Scale Approximate

Approx. Top
of the SCL
SCL

conductivity 7m
790 820 850
Station Numbering (Distance in Meters)

Fig. 17.3 Direct push electrical conductivity data plotted over radar data used to assess the presence, thickness and clay content of the epikarst
zone which included a semi-confining layer
204 17 Invasive Methods

60
me
te eters
rs 100 m

d conductivity
0 100
c conductivity 0
0 100
b conductivity 0
0 100
0

background
6
Depth in Meters

contaminants
6
contaminants

12
12
12

18
18
18

Fig. 17.4 Contour of electrical conductivity (EM31) data over a buried waste site and direct push electrical conductivity plots showing the vertical
extent of contaminants (b and c) as well as background conditions (d)

At the Savanna River Site (SRS), in southeast South rig being used to evaluate soil conditions at a contaminated
Carolina, a low-level radioactive waste site, an electromag- site in a karst setting.
netic survey using an EM31 instrument mapped the lateral The traditional cone penetrometer data (Fig. 17.6)
extent of buried contaminants over a small area that had a includes sleeve friction, tip pressure, a ratio between the
high conductivity signature (Fig. 17.4a). The direct push sleeve friction and tip pressure, and pore pressure as well as
method with an electrical conductivity tip was then utilized depth, time, and pushing force, which are measured as the
to characterize the vertical extent of the contaminants. Two probe is pushed hydraulically into the ground. Penetration
of the pushes (Fig. 17.4b, c) clearly identified the presence rates of a <1 m per minute are typical and depth of penetra-
and depth of a high conductivity zone (contaminants) within tion depends upon soil conditions but is limited by the pres-
the plume identified by surface electromagnetic measure- ence of dense sands, gravel, and cobbles. McCall et al.
ments. One of the pushes (Fig. 17.4d) clearly identified (2006) provide more details on this topic along with ASTM
background conditions (lack of contaminants) outside of the D-5778 (ASTM 2012b) and ASTM D-6067 (ASTM 2010a).
plume identified by surface EM measurements. A field gas A wide range of measurement options supplement the
chromatography unit was used to analyze soil samples and regular CPT measurements such as:
laboratory analysis of water samples was used to identify
and characterize the contaminants. • Seismic probe to obtain P wave velocity.
• Hydrocarbon detection
• Video
17.2.2 Cone Penetrometer Testing (CPT) • VOC detection
• Soil moisture/resistivity profiles
Cone penetrometer testing is carried out by pushing a small • Soil, gas, and water sampling
diameter probe into the sediment using the weight of the • Gamma radiation profiles
truck and a hydraulic ram. The hydraulic pressure and the
rate of advance into the sediment can be used to characterize At a site in west central Florida, CPT data was an inte-
the materials encountered. Figure 17.5 shows a typical CPT grated part of a site characterization. The site had a thick
17.2 Direct Push Methods 205

Fig. 17.5 A CPT rig is being used to evaluate soil conditions at a contaminated site in a karst setting (a) and a close-up of the probe beneath the
truck (b)

cover of clean quartz sand that provided an ideal application based upon a visual inspection of the sample from the hollow
for ground penetrating radar and CPT methods. The radar stem auger. The SPT acquired during drilling and sampling
data provided a means of rapid cost effective site coverage to indicated loose materials immediately above this clay layer.
identify potential problem areas that showed signs of sink- The natural gamma geophysical log provides essentially con-
hole activity or the presence paleokarst. A cone penetrometer tinuous data throughout the depth of the borehole and a mea-
was then used to further evaluate the potential problem areas sured volume of about 30-cm diameter around the sensor.
by identifying loose materials, voids and changes in This method detects variable clay content from 11 to 17 m
conditions. based upon natural radiation associated with clays. The CPT
As a means of data correlation, a test site was selected in friction ratio data indicates a variable clay content starting at
background conditions without any active or paleokarst a depth of 9 m and continuing to 14 m where refusal was
based upon the radar data. Data obtained included a boring encountered. Note that each set of data was acquired correctly
with standard penetration tests (SPT) values, a geologic log and is accurate, but tells a slightly different story.
and geophysical logs along with CPT measurements made A specialized cone penetrometer was used for a very shal-
immediately adjacent to the borehole. Figure 17.7 shows the low investigation of surface fissures along the earthen run-
results of this test data. ways at Edwards Air Force Base (EAFB) in California. The
This example clearly illustrates the impact of scale on a fissures were a result of groundwater withdrawal from a deep
variety of measurements. The drilling data sampled every aquifer (Fig. 7.10). Radar surveys had been carried out along
1.5 m and describes a 0.6 m thick clay layer at a depth of 14 m the long runways to locate possible fissures before they broke
206 17 Invasive Methods

Fig. 17.6 An example of CPT data showing sleeve friction, tip resistance, and friction ratio

through to the surface (Fig. 16.34). Ten parallel survey lines assessment of subsurface conditions by using the recorded
were run along each of four runways for a total of 362 km of tip pressure in PSI (pounds per square inch). This specialized
radar data. Radar data was used to provide an approximate CPT rig could rapidly identify the lateral boundary of the
location of the fissures (within <1 m). Additional data was fissure (a loose zone of material). The data in Fig. 17.9a
acquired at these locations using a specialized cone pene- shows one of the seven measurements encountering a local-
trometer (Kratochvil 1989; Kratochvil et al. 1992). ized loose zone. This indicated the fissure’s exact location
An automated cone penetrometer, mounted on the front of and that the fissure was narrow, probably in its early stage of
a truck (Fig. 17.8) would push the probe to depth (about development. Figure 17.9b shows a second set of data with a
30–60 cm) at a station, then the probe assembly would move wide range of soil conditions from three of the six pushes,
a short distance and repeat the process. Up to seven measure- indicating that the fissure was wider and more mature. Once
ments could be made over a distance of about 2 m without the location and character of a fissure was identified, repairs
moving the truck. This CPT data provided a detailed were then carried out. Each site was trenched, backfilled, and
17.3 Borings 207

Geologic Standard Penetration Test Natural Gamma Cone Penetrometer

Log blows per foot counts per second Friction Ratio
1 10 100 0 50 100 150 200 0 2 4 6 8 10
0

Sand
6
Depth in Meters

Clay at 9m

9 Clay at 11m

12
Clay at 14m Loose Sand
Clay
15 Sand Increasing Clays

Sandy
Limestone
18
Limestone

21
100% Loss of Circulation

Fig. 17.7 Test site showing geologic log, SPT data and natural gamma log from the sample borehole. The cone penetrometer data is adjacent
borehole about 1 m away. Note 100 % loss of fluid in zone of loose sand

compacted. Khedr et al. (1985) provides details on this auto-

mated cone penetrometer system.

17.3 Borings

Borings have been and continue to be one of the primary

methods for site characterization in geotechnical investiga-
tions and environmental studies. Boring logs remain the
mainstay and are often a critical piece of data in solving the
geologic puzzle for the site characterization process. Borings
provide soil samples and standard penetration tests (SPT)
“N” values, or rock core (RQD) values as well as water lev-
els and samples. Standard driller’s or geologic logs provided
a description of geologic conditions, depth of contacts or
stratigraphic changes. Depending upon the type of drilling
method used, the logs may include rate of drilling, rod drops
or loss of drilling fluid, which can all be indications of karst
conditions. In addition borings also provide access for run-
ning a variety of geophysical logs, making in-situ geotechni-
cal and hydrologic measurements of the soil and rock, and
may eventually be used to install monitoring wells or
piezometers. Borings also provide samples of the soil or rock
Fig. 17.8 Photo of the specialized shallow CPT system used to evalu-
ate the fissures at Edwards Air Force Base (EAFB) for our geologic descriptions and for laboratory testing,
208 17 Invasive Methods

Fig. 17.9 Two examples of CPT data from EAFB. One set of data shows a narrow fissure (a) while another set of data shows a wider fissure (b)
with considerable variation
17.3 Borings 209

providing the engineer with much of the data necessary for Always avoid drilling a hole or placing a well unless you are
design. reasonably sure of its location and purpose. A field plan for
Borings are often the first method to be used to determine drilling should be based upon all prior data and the resulting
top of rock profile or rockhead. Hencher (2012) discusses the conceptual model. The drilling plan should identify not only
problem of determining the rockhead. This approach is prob- locations and depths of borings but the nature of the data
lematic in karst due to the highly variable conditions often required and possible anomalous conditions expected along
encountered (Fookes 1997). It is here that we find unpredict- with other long-term use of the boring and subsequent data.
able variations in the depth of rock, its degree of weathering, Selection of the most appropriate drilling method should be
frequency of fissuring, buried pinnacles, the extent of iso- made by the engineer or geologist along with a knowledgeable
lated loose blocks of rock and undercut pinnacles supported drilling contractor and must take into consideration site condi-
only by the surrounding soil. It is often difficult to determine tions, objectives, availability of equipment, personnel, and
the difference between a large boulder, a pinnacle, and the costs. Often, the best approach when selecting a drilling
upper surface of continuous in-situ massive rock (Waltham method is to talk to drilling firms who are experienced with the
and Fookes 2003). local geologic conditions. Sometimes choosing a lesser level
of drilling technology is better if one can find a driller who is
familiar with the local geology and has proven track record.
17.3.1 A Drilling Plan Drilling is, in part, an art and its success is dependent upon
good practice and skill of the driller. An experienced driller is
Ideally drilling would not be done until the desk study and site an invaluable asset to the site investigation effort.
walkover had been completed, a preliminary conceptual model In all cases, flexibility must be maintained in the drilling
developed and the surface geophysical investigation carried program to accommodate changes in conditions that are dis-
out. There are always exceptions. In an area where there is covered as the site characterization effort proceeds. This
little or no geologic data available (a most unusual situation includes changing locations of borings and adding additional
since there will be at least regional data in most areas) it may borings where unusual conditions have been identified.
be necessary to drill a few borings early in the field program to These decisions must be the responsibility of the site charac-
establish the general geologic conditions. In such cases, loca- terization team who are in the field.
tions are based upon the best available data at the time.
Drilling can be an expensive procedure and the informa-
tion obtained from each borehole must be maximized. The 17.3.2 Drilling Methods
objective of an effective geotechnical or environmental site
characterization should be to minimize the number of bore- There are a wide variety of drilling methods (Table 17.1).
holes, piezometers and wells while maximizing their repre- They range from simple hand probes or soil augers to a wide
sentation of background and anomalous geologic conditions. variety of jetting, percussion and rotary drilling methods.

Table 17.1 Common direct push and drilling methods

Method Use in soil or rock Use of casing Use of drilling fluids Geologic samples
Direct push methods
Direct push (percussion) Soils, soft rock Possible N Undisturbed
Cone penetrometer testing (CPT) Soils, soft rock Possible N Undisturbed
Jetting or percussion methods
Wash boring Soils N Water None
Cable tool Soils and rock Y Water Chips
Downhole hammer Rock Y Air, water, foam Chips
Auger methods
Hand probes and augers Soils N No Disturbed
Solid stem augers Soils, soft rock N No Disturbed
Hollow stem augers Soils, soft rock Y-the auger itself No Disturbed
Rotary methods
Direct rotary Soils and rock Optional Air, water, mud, foam Chips
Reverse circulation Soils and rock Optional Air, water mud foam Chips
Dual tube reverse circulation Soils and rock Y Air, water mud foam Chips
Rotosonic Soils and rock Y Air, water, or mud optional Undisturbed
210 17 Invasive Methods

With any group of methods, they all have their advantages line of wash borings was completed from the pond edge to the
and limitations. Some drilling methods work well in pond center. The probes could extend to a maximum depth of
unconsolidated materials and others in solid rock. Some of 6 m. Refusal was encountered at the edge of the pond and
the drilling methods provide rapid penetration, but are progressively got deeper toward the center. The combination
destructive, disturbing the sample as a result of the drilling of the wash borings and radar data confirmed our interpreta-
process. Other drilling methods are relatively slow but can tion that this was a sinkhole. See case history in Chap. 27.
recover relatively undisturbed samples of soil or rock.
Drilling fluids (air or water and sometimes additives) are 17.3.2.2 Percussion Methods
commonly used to carry cuttings to the surface and maintain A variety of percussion methods of drilling are available
a clean, open borehole. including repeatedly dropping a cutting tool (cable tool) or by
In some cases, the borehole can be advanced without con- the use of a surface or a downhole hammer. These are destruc-
cern of collapse within the borehole. In other cases borehole tive drilling methods where only chips are recovered. However,
collapse is a common problem (such as with thick unconsoli- they are generally a fast and efficient means of drilling.
dated sands) and must be considered when selecting the drill- For example, small sinkholes were occurring at a military
ing method. An unstable borehole is maintained by the use of site on the cliffs of Guam near critical radar structures. The
drilling mud or by advancing casing as the hole is drilled. site was located at an elevation of about 180 m above sea
Casing effectively stabilizes the hole from collapse, but limits level with a water table near sea level. A detailed radar sur-
access to the formation for direct tests (ie packer tests and vey was done to identify the lateral locations and distribution
certain geophysical logs), and observations (by video, acous- of cavities at the site. See Sect. 16.5.3.3 for further discus-
tic or optical televiewer), or sampling of fluids and flow. sion along with Figs. 16.21 and 16.22. A percussion drill rig
Environmental aspects of installing piezometers and was used to investigate selected radar anomalies. The drilling
monitor wells, geotechnical measurements and instrumenta- method was quick and effective since no samples were nec-
tion as well sampling of water and contaminants will often essary. Voids were indicated due to loss of air circulation and
tend to dominate drilling requirements. In some cases the rod drops. Each boring was then logged with a borehole
method of drilling will be dictated by what equipment is video camera to visually document the cavities encountered.
available. These factors and more need to be considered These cavities were typically 0.5–2 m in diameter with a few
when selecting the appropriate method for a given geologic larger ones of 2.5–3.5 m in diameter. Once identified the
conditions and project objective. cavities were grouted up.
Poor drilling procedures, casing installation, installation
of instrumentation, can lead to subsequent problems with the 17.3.2.3 Auger Methods
data acquisition, and monitoring. While specifications are Auger drilling rigs are commonly available and are used in
easily written on paper, it is another thing to carry out the many site investigation efforts. Augers can range from hand
necessary details in the field and to verify them once in place. augers (Fig. 17.10a) to small portable augers (Fig. 17.10b) to
Sara (2003), Hunt (2005), and Ruda and Farrar (2006) large truck mounted augers with SPT capabilities
provide further details on drilling methods and their use. (Fig. 17.10c).
Several ASTM standards have been developed which cover Solid stem augers provide a rapid means of drilling and
hollow stem augering, SPT and split barrel sampling as sampling of unconsolidated materials. Sampling of disturbed
well as rock core drilling and sampling techniques (ASTM soil is done as the auger spins-up the sample or by pulling the
2006, 2011, 2014). auger out of the ground to extract the sample on the auger.
Hollow stem augers are one of the most commonly used and
17.3.2.1 Jetting Methods readily available methods for drilling and sampling in uncon-
Wash boring or probing with a hand probe and water under solidated material. Standard penetrating tests (SPT)
pressure is a simple means of probing shallow soil condi- “N”-values are often obtained as part of the drilling and sam-
tions without a drill rig and is readily adapted for work in pling process. Augers are unable to penetrate very dense
shallow water. The method has been used to define top of compacted materials and boulders. No fluids are used and
clay or rock, and probe mud thickness of sinkhole ponds and the auger remains in place to maintain the open borehole.
lakes. A small centrifugal pump provides a jet of water from More details on auger drilling are provided in various ASTM
a 2 cm steel pipe, which is pushed into the sediments until it guidelines (ASTM 2006, 2009, 2011).
encounters resistance or rock. Large diameter augers provide a means of developing
A simple wash boring probe was used to probe the bottom holes up to 1 m in diameter. Besides their common use for
of a circular pond that appeared to be a sinkhole. Ground pen- poured in-place piles, these larger holes are sometimes used
etrating radar data obtained up to the edge of the pond indi- to provide direct access for a person to observe and log geo-
cated strata dipping toward the center of the pond. A profile logic conditions.
17.3 Borings 211

Fig. 17.10 A variety of auger devices are used. A hand auger (a), a small portable auger rig (b) and a common hollow stem auger (HSA) drill rig
(c) which are commonly used for drilling in unconsolidated material

17.3.2.4 Rotary Methods geologic log based upon chips from drilling and the
Rotary drilling using air or water as a drilling fluid can be used continuous geophysical logs (natural gamma, gamma-
in all types of unconsolidated materials and rock. Air rotary gamma density, and neutron porosity) that can be run through
drilling can function in areas in which a supply of water is not steel casing providing additional geologic data from the bor-
available. Rotary percussion drilling uses both a rotary and ing for correlation.
percussion of the drill bit to fracture the material. Geologic
sampling is by cuttings, which are carried upward by return air Rotosonic (Vibratory) Methods
or water, which flows between the borehole wall and the drill Rotosonic drilling uses vibration of 50–50 Hz of the drill
pipe and aids in maintaining a stable borehole wall. Caving in stem in addition to downward pressure and rotation of the bit
unconsolidated materials can be avoided by the use of drilling drill to rapidly drill through most materials. The sonic
mud or by advancing casing as the hole is drilled. method has proven fast and effective in a wide variety of
geologic environments. Drilling can be done in most uncon-
Dual Tube Reverse-Circulation solidated materials and softer bedrock (sandstone, limestone,
Dual-tube reverse circulation rotary provides very rapid shale, and slate) and is very rapid. Harder rock can be drilled
drilling in both unconsolidated materials and rock forma- but at increased cost.
tions (Fig. 17.11a). This method of drilling is useful in diffi- Long continuous cores (typically 3 m) with core size of
cult conditions and reducing lost circulation by providing a 7.5–25 cm can be obtained. The core sample is removed from
casing as part of the drilling process. The inner tube rotates the core barrel into a plastic sleeve or stainless steel tray with
the drill bit and the outer tube functions as the casing. Drilling a minimum of disturbance. Samples may also be collected
fluid flows down between the two tubes, out around the drill with clear plastic liners or stainless steel split liners inside of
bit and up the inner tube. As a result the method is useful in the core barrel. The unconsolidated sample may be somewhat
fractured rock, boulders, and cavernous conditions because it disturbed (by vibration and handling) but are essentially con-
minimizes problems with lost circulation and borehole sta- tinuous. Figure 17.12 shows a soil core being removed from
bility. Sampling is by chips returned to the surface by air, the drill rig at the Savanna River Site (SRS), screened for
water or mud (Fig. 17.11b, c). radiation using a Geiger counter and then boxed for storage.
Dual tube reverse circulation drilling is commonly used An outer casing provides a means to keep to borehole
where thick sands or loose sediments are a problem. In many from collapsing while the core is retrieved and minimizes
areas of Florida, thick loose sand and sediments are known cross contamination during drilling, if fluids are used. It also
to collapse into the borehole when casing is removed. Where makes the installation of monitor wells or geotechnical
such problems are known to occur, the borehole can be instrumentation very efficient. Drilling can be done with or
drilled by the dual tube method and with the casing still in without fluids, (sometimes just water is used) and there are
the ground selected geophysical logs can acquire data no drilling wastes. For more details see ASTM D6914 on
through the steel casing. In this manner, we have both the sonic drilling (ASTM 2010c).
Fig. 17.11 A typical dual tube air rotary rig (a) commonly used when to catch the cuttings (b) is changed at regular intervals in order to iden-
drilling in loose sediment. The outer tube acts as casing to maintain an tify cuttings from a certain depth range (c)
open hole in loose sediments that are commonly found in Florida. A pan

Fig. 17.12 A sediment sample from a rotosonic drill rig is being placed into a plastic sleeve (a), which is then monitored for radioactivity (b). The
core is then labeled and stored (c)
17.3 Borings 213

17.3.3 Indications of Karst When Drilling 17.3.3.2 Drilling in Rock

When drilling in rock, poor core recovery, rod drops and
17.3.3.1 Drilling in Unconsolidated Materials fluid loss are a good indication of the presence of highly
There are several indicators of subsidence or sinkhole activity weathered zones, open fractures, voids or cavities. These
when drilling in unconsolidated materials. These include low types of conditions are not uncommon and if they are
SPT or blow counts along with rod drop and fluid losses. SPT small and isolated may not be of concern. However, a spa-
results are commonly found in driller’s logs and provide a tial concentration of such indications laterally or at a par-
means of measuring the resistance of the unconsolidated mate- ticular depth may be significant and should not be
rials to the penetration of the sampler using the blow count of a disregarded. It is then necessary to verify the significance
standard hammer. The “N” values (blows per foot) from SPT of this data by other independent data such as geophysical
are useful for identifying loose or raveling zones in areas of logs or borehole imaging methods. See ASTM guideline
suspected subsidence. More details on SPT measurements are D2113 for details on rock core drilling and sampling
provided in an ASTM guideline D1586 (ASTM 2011). (ASTM 2014).
A slurry wall was being designed along the north side of For example, initial drilling had been carried out for a
a landfill in Florida. Drilling was completed along the pro- new bridge into the Florida Keys. Borings were spaced at
posed alignment with a spacing interval of about 15 m and approximately 75 m intervals along 2.6 km to depths of
went to depth of refusal. SPT data was collected and used 20–47 m with the majority extending to 30 m. The 34 boring
along with soil samples to identify the top of a hard clay logs had been plotted on a large sheet and indicated rela-
layer in which the bottom of the slurry wall was to be placed. tively uniform geologic conditions. Upon close review of
This was a data set that appeared to have excellent data den- details, 27 % of all rod drop and fluid loss had been noted in
sity. However, due to concerns with potential geologic vari- two adjacent boreholes. Two additional nearby borings
ability in this karst setting, a surface geophysical survey was accounted for an additional 34 % of all fluid loss, but this
completed along the alignment using MASW method (map- was not mentioned in the geotechnical report. This turned
ping shear wave velocities). Figure 17.13 shows a contour of out to be one of the many indicators for an area of a large
shear-wave velocities from the MASW data along with the paleokarst collapse zone. This was a case of boring encoun-
corresponding SPT data from the drilling. The SPT data and tering indications of potential problems yet the data was bur-
the MASW data showed remarkable correlation. Sharp ied within a large data set and went unnoticed. If these
increases in blow counts (from less than 10 to >20) corre- anomalies had occurred in widely spaced borings they prob-
sponded to an increase in shear wave velocity, roughly at ably would have little significance, but the fact that the data
900 ft/s contour. While both sets of data are valid, the MASW clustered spatially is what caught our attention. This case
data shows some of the variability not detected by the SPT history is presented Chap. 26.
data. For example, see stations 125 and 157 in Fig. 17.13.
58

59

61

63

66

67
60

65

68
62

64

0
-3
Depth in Meters

-6
-9
-12
-15
-18
-21
15 30 45 60 75 90 105 120 135 150 165 180
Top of Blue-Green Clay Layer Distance in Meters
based upon boring data
SPT Blow Counts from Test Borings
300

400

500

600

700

800

900

1000

1100

1200

1300

(Scale 0 to 100)
Shear-wave Velocity
(ft/sec)

Fig. 17.13 SPT data and shear wave velocity data (MASW data) is used to define top of clay layer for a slurry wall
214 17 Invasive Methods

17.3.3.3 Geologic Sampling pieces longer than 10 cm versus the total length of core run,
Most of the engineering and hydrologic properties of soils, to provide an evaluation of the rock quality. These RQD val-
and rock have been determined from measurements made ues range from 0 to 100 with 0 being poor quality and 100
within a borehole along with laboratory measurements from being excellent quality rock. ASTM D5878 describes rock
borehole samples. There is a long history of laboratory mea- mass classification systems (ASTM 2008). Bell (2004) and
surements with highly quantified results. The results of Hunt (2005) provide further discussion of the RQD designa-
localized field tests and laboratory analysis on small samples tions along with the properties of soil and rock for engineer-
(with dimensions of <0.3 m) are commonly used to represent ing purposes and methods of testing. Hencher (2012)
conditions of a large mass of soil and rock (with dimensions describes various methods of rock characterization.
of ten’s to hundreds of meters). Unfortunately, it is the better Many rotary methods are destructive when drilling and
quality intact samples that are usually sent to the laboratory result in chips of rock brought to the surface by drilling flu-
for testing and the worse samples are never retrieved or ids. How the chips are collected, how often, who is describ-
tested. This leads to biasing of the data and assumes that the ing them and if they are saved in chip trays will all affect the
intact samples recovered are representative of the actual site quality of data retrieved. At a Superfund site in west central
conditions. This practice is not acceptable in the heteroge- Florida, a rotary drilling technique was used for monitoring
neous conditions found in fractured rock and karst. See well installation. Samples of the chips were acquired every
ASTM D5434 for details on logging of soil and rock (ASTM 0.75–1.5 m, described and used to created geologic logs by a
2012a). geologist and saved in chip trays (Fig. 17.15a) for further
documentation. The chip trays were used to create a visual
Soil Samples cross section (Fig. 17.15b) through an anomalous area (a
Split spoon samples provide a majority of the data in uncon- buried paleocollapse) at the site. The change in materials
solidated materials. While somewhat disturbed, they provide from background conditions consisting of clean, light-col-
reasonably good samples for geologic and engineering pur- ored sands, gravels and limestone to anomalous conditions
pose. It is not unusual to see samples obtained at intervals of consisting of variable, dark organic and peat materials can
1.5–3 m. However, continuous sampling should be used clearly be seen in the chip trays. This visual cross section
when dealing with any complex geologic conditions, such as provided a very effective way to illustrate conditions at a
karst. Once soil samples are collected there are a long list of public meeting. See the full case history in Chap. 27.
laboratory tests that can be performed such as grain size
analysis, permeability testing and moisture content. See
ASTM guide D1586 for STP and split-barrel sampling of 17.3.4 Special Considerations When Drilling
soils (ASTM 2011). or Using Drilling Data

Rock Samples 17.3.4.1 The Need for Angle Borings

Rock core is the standard for assessing conditions in rock. While the great majority of borings are vertical, inclined or
Rock core samples may be obtained by a variety of rotary angle borings often become necessary to characterize frac-
drilling and core sampling methods in a variety of diameters. tured rock, fault zones and to characterize complex geologic
A variety of tests on the rock core can be run such as com- conditions. When strata are close to horizontal, joints will
pression tests. ASTM D2113 provides information on the generally be close to vertical. Dissolution will often occur
practice of rock core drilling and sampling of rock (ASTM along these joints in soluble rock. Encountering these verti-
2014). cal joints with vertical borings becomes an almost impossi-
The amount of core or percent recovery is an indication of ble task based upon the probability of encountering the joint.
rock quality or the presence of fractures/voids. Figure 17.14 As a result angle borings are necessary.
shows core collected at two different karst sites. Figure 17.14a Sowers (1996) discusses an example where TVA carried
was acquired in the Kansas City area and shows fairly good out a site characterization at a nuclear plant site in central
core recovery of limestone and shale with lots of near hori- Tennessee. A very closely spaced grid of vertical borings
zontal fracturing. Figure 17.14b was acquired in the south was used to investigate the horizontal bedded strata. The
Florida area and shows poor core recovery and a highly investigation included several hundred vertical borings fol-
porous relatively soft limestone. lowed by geophysical logging. No angle borings were made
Core recovery or percent recovery compares the total vol- because the drilling equipment was limited to vertical holes.
ume recovered to the total volume of the core run. Rock Later excavations for a cooling water canal encountered sev-
Quality Designation (RQD) values are a more rigorous eval- eral dissolution enlarged joints 0.6–1.2 m wide and 1.2–2 m
uation for rock core than simply a core recovery. This clas- high at a depth of 12 m in the limestone. These features had
sification uses a percentage based upon total length of core been completely missed by the vertical drilling.
17.3 Borings 215

Fig. 17.14 Core samples from a Kansas site (a) indicate almost total core recovery while core samples from south Florida (b) indicate poor
recovery and weather core

Angle borings were successfully used to confirm the pres- acid leak. Angle borings were then used to confirm the
ence and evaluate a fracture zone at a site in Kansas City, interpretation of the microgravity data and to provide further
Kansas. A small soil piping feature was observed within a data to define the exact boundaries of the cavity. These same
surface fissure in the loess soils at the site. The loess soils borings were then used for grouting of the cavity.
were removed revealing an obvious fracture zone in the rock Figure 17.17b shows the drilling plan to assess conditions
(Fig. 17.16a). To evaluate the fracture zone, angle borings under a large structure within a well-developed industrial
were made to intersect the fracture. Borehole video located setting.
fractures in the shale and large quantities of water (1,000 gal
at 10 gpm) were poured into these two angle holes without 17.3.4.2 Drilling May Sometimes Trigger
being rejected (Fig. 17.16b). The exposed fracture was then a Sinkhole
used as an injection point for a dye trace to evaluate the In some cases, drilling can be done without concern of col-
connection of the fracture zone into the underlying mine. See lapse. In other cases drilling can trigger a collapse. The ham-
the complete case history in Chap. 25. mer blows, drilling pressure and drilling fluids (both water
Inclined borings are also used to provide access under and air along with excessive pressure) can be enough to trig-
existing structures. Figure 17.17a shows a drill rig squeezed ger already unstable conditions. The process of drilling by
between two buildings at a resort in the Florida Keys drilling itself does not create a cavity, but it can be the final event to
at an angle to obtain data under an existing building. At trigger a collapse by simply encountering an existing geo-
another site, microgravity measurements indicated a large logic weak zone. This is discussed with examples in Part I,
cavity zone had developed under a large structure due to an Sect. 8.1.2.
216 17 Invasive Methods

Fig. 17.15 Chip samples are placed in chip trays for storage (a). Chip trays from eight borings were laid out from north to south to illustrate the
change in geologic conditions at the site (b)

17.3.4.3 Flexibility When Encountering of this opportunity to determine the nature and orientation of
Anomalous Conditions the cavity. This could have been done using a video inspec-
All voids in limestone were interconnected at one time tion or optical televiewer log (above the water table) or an
because they were formed by the flow of water (Waltham and acoustic televiewer log (below the water table). This infor-
Fookes 2003). When a void is encountered within the rock mation would have helped to determine a possible zone of
mass it should not be assumed to be an isolated feature, but cavity development so that further investigation efforts could
it should be considered part of a conduit system developed have been focused upon likely zones of cavity development.
by groundwater flow. An effort should be made to determine This is similar to the strategy used by TVA (Hopkins 1977).
its lateral dimensions and its orientation, assuming it is a part
of a fracture or cavity system. For example, an 2.4 m thick 17.3.4.4 Drill Rig Access
cavity was encountered at a depth of 33 m while drilling a In many cases, borings have been located where the drill rig
borehole in limestone at a proposed nuclear power plant site. had easy access rather than where the problem lies. Rex
This was a bit of fracture or surprise to the on-site staff and Morey, the developer of ground penetrating radar, was
they promptly grouted it, which was their standard proce- demonstrating the use of radar to the Florida Department of
dure. The better approach would have been to take advantage Transportation (FDOT). The top of rock had been identified
17.3 Borings 217

Fig. 17.16 A fissure in the loess soil cover was excavated and exposed a major opening into top of rock (a). Two angle borings indicated fractures
that would not retain water (b)

as well as a number of possible isolated cavities near the top radar data and the boring data was impossible because the
of rock within a ditch parallel to a road. The FDOT drillers drill rig was not located along the line of radar data.
were sent to the field to verify conditions, and reported find-
ing no cavities. This was most unusual since the radar data 17.3.4.5 Borehole Deviation
was quite clear showing the presence of many cavities. Much The driller will not usually be aware of possible deviation of
later, a discussion with the senior FDOT geologist revealed the borehole from vertical (Fig. 17.18), yet it can be critical
that the drill rig was actually located at the edge of the road when making subsequent measurements from that boring.
because of easy drill rig access. This was 6 m or more from Deep water level measurements where there is little gradient
the line of radar data in the ditch where recharge was can be significantly affected by borings that are not vertical.
occurring. In this case, a one to one correlation between the Some geotechnical and hole-to-hole geophysical
218 17 Invasive Methods

Fig. 17.17 Drilling of angle borings allows access to conditions under a structure (a) and may require a combination of vertical and angle bore-
holes may be required to sufficiently assess conditions under a structure (b)

measurements are also impacted by errors introduced by the concrete floor. Settlement and cracking of the concrete
borehole deviation. See the example of borehole deviation floor was observed at a number of locations. New equipment
(Fig. 18.13) in Chap. 18 geophysical logging section. was to be installed and there was concern that the failure of
the floor could cause damage to the production line.
17.3.4.6 Drilling Inside Structures A detailed assessment of existing conditions of the build-
It is not uncommon to find industrial facilities, in karst set- ing floor was made including the time history of equipment
tings, that have been impacted by the presence of voids or installation dates, and subsequent leaks as well as areas of
cavities. While deeper features may be investigated through visible subsidence and cracks in the floor. Then a ground
angle borings from outside of the building, smaller, shal- penetrating radar survey was carried out to identify voids
lower features may require coring through the concrete floor. under the concrete floor and their lateral extent.
Acid leaks had occurred at an electroplating plant in These data were used to determine specific locations for
southwestern Missouri, and voids were known to exist under direct sampling. Core holes were drilled through the concrete
17.3 Borings 219

and the depth of the void could be directly measured. An of voids. All of this data was used as the basis for providing
auger was used to remove soil samples to test for the pres- recommendations for remediation.
ence of acid (Fig. 17.19a). A side looking borehole video
camera (Fig. 17.19b) was used to evaluate the lateral extent 17.3.4.7 Misinterpretation of Data
While the boring data itself may be accurate, the interpretation
may not be Sara (2003). The example in Fig. 17.20 illustrates
how drilling data can be used to create a cross section for a site.
In this case, the drillers or geologic log may be totally correct
by itself but the data may be wrongly interpreted. Widely
spaced borings and or interpreting pinnacles and a floating
boulder as the top of rock or low areas between pinnacles as the
top of rock will result in an inaccurate top of unweathered rock
profile. Sometimes the true details are only discovered when
the site has been excavated. In addition, cross sections devel-
oped from boring data will often have boring data projected
over large distances onto the cross section. In areas of karst,
where there is a high degree of spatial variability, this practice
can be misleading. The location of all borings on the cross sec-
tion should be shown on a map, so that the data used in the
perceived development of the cross section can be clearly evaluated.
actual At a site in Kansas, both borings and seismic refraction
data were used to determine the depth to top of rock. The
seismic measurements resulted in depths much deeper than
that determined by the boring data for the same locations. It
was only after persistent questioning that the reason for this
discrepancy was identified. The top of rock had been defined
Fig. 17.18 Borehole deviations from vertical can impact certain mea- by lawyers as the “first trace of rock” in the boring logs and
surements yet the driller will not be aware of the deviation included isolated boulders well above the top of massive

Fig. 17.19 A cored borehole was cut through the concrete floor of an electroplating plant to assess suspected voids under the floor and acquire
samples (a). A borehole video camera was used to assess the extent of the void under the floor (b)
220 17 Invasive Methods

Fig. 17.20 A cross section can be easily misinterpreted from boring data alone. In both cases (a and b) the driller’s or geologic log may be totally
correct by itself but the overall interpretation may be wrong

in-place rock. These results, although technically incorrect, Drillers logs (made by the driller) or geologic logs (made
were now accepted as the legal definition of top of rock and by an engineer or geologist) include descriptions of the soil
were not allowed to be changed. The seismic refraction data and rock samples, as well as water levels, the rate of drilling,
was measuring the top of massive in-situ unweathered rock. rod drops, fluid loss and other related data. It is interesting to
While both sets of data were acquired accurately, the misin- consider the manner in which boring logs are developed.
terpretation of the driller’s logs by the lawyers forced the While we have many ASTM guidelines and standards for the
seismic data to be in question rather than resolve the errone- field aspects of drilling, and recording of geologic data, these
ous decision of the legal definition. guidelines or standards alone do not assure quality results.
It’s the experience and dedication of people that provide
17.3.4.8 Ground Truth and Hard Data quality results and we don’t have a standard for that.
For decades we have seen borings referred to as the ultimate It is almost too much to ask of the driller and his helper to
means of “ground truth” or “hard data” and are the standard manage the drilling operation, keeping an accurate drillers
from which other data and interpretations must be refer- log, recording all of the details and changes in drilling condi-
enced. However, boring logs whether done by a driller, geol- tions. There is considerable multitasking (recording drilling
ogist, or engineer, (like any other source of data) are rate, fluid loss, rod drops, SPT measurements, and sampling
subjective and are also subject to a variety of errors, even and packaging of sediments or labeling of core) taking place
with experienced personnel. while operating the drill rig (including safety issues) all of
17.3 Borings 221

which can easily lead to gaps in the data. Considering that require monitoring wells or piezometers to be installed. The
most drilling is contracted on a unit basis (and often at low field plan should maximize the data from each of the borings.
bid) under such conditions the best drillers and the drilling For example, installing small diameter piezometers will limit
companies are forced to compromise. Furthermore the work the amount of data that can be obtained, simply due to the
is often carried out under difficult weather conditions and the diameter of the casing. The plan should also consider the
field logs may become wet and muddy and some portions of longevity of each boring. Existing boreholes can provide the
a page may not even be legible. Then back in the office the opportunity for additional data, if needed, avoid grouting the
logs are often handed to someone (often other than the per- borings too quickly.
son in the field) to be typed, and details get lost. It is here
where the potential for transcription errors, misinterpreta-
tions and omissions to occur. 17.3.5 An Optimum Approach for Drilling
Unfortunately, it has become common to think of the task and Sampling
of drilling, logging of the borehole and acquiring samples as
a trivial one, which can be done by almost anyone if given The following is a proposed strategy to optimize borings,
some guidelines. In fact, it is as difficult a task as any of geophysically logging and sampling. It is based upon the
those in the site characterization process. site characterization strategies presented and assumes that a
There can be significant differences due to training and solid set of data are used to properly select boring loca-
experience of the persons doing the interpretation of geo- tions. This approach maximizes the amount of data from
logic conditions and record keeping. Unfortunately all too each boring and at the same time minimizes the number of
often the geologist or engineer sent to the field are the most borings. This strategy also minimizes the number of sam-
junior staff with little experience and often no formal train- ples that need to be obtained and stored. It provides inde-
ing in drilling or logging of a borehole in soil or rock. Most pendent sets of data (boring logs and geophysical logs) to
universities do not teach the fundamentals associated with improve confidence levels. It follows the same philosophy
drilling and geologic logging. In many situations today we that TVA had used for the extensive work in the construc-
find the most junior professional staff assigned to drilling tion of dams and nuclear power plants, (Hopkins 1977),
operations. We have seen cases when a young geologist with and further defined by DOE ESC recommendations in
a master’s degree was assigned to a drill rig for the first time. ASTM D6235 (2010b).
He had no training in college on the aspects of geologic log-
ging and was placed in the position of logging borings with • Initially two or three borings are drilled to establish on-site
little more than the companies reference document in hand geologic conditions. These would include detailed drillers
and “good luck, the driller will help you”. logs and continuous sediment and core samples. The con-
Another example was from a project that specified the tinuous soil samples and rock cores can be examined in
requirement for a geologist to log the core (but did not define detail and tested for various geotechnical parameters.
the level of experience required). The drilling company did • Then each of these initial borings would be geophysi-
not have a geologist on staff and hired a recent graduate part cally logged with an appropriate suite of measurements
time for the job in order to comply with the project specifica- to provide an independent set of geologic data for corre-
tions. Unfortunately the geologist had no training or experi- lation with the driller’s logs. This combination of two
ence in such work. Luckily the driller was an old hand who independent sets of data (the geologic log from drilling
had spent many years drilling and logging in the specific for- and the geophysical logs) provides a geologic reference
mations, including extensive work on the specific site. The for the site.
driller provided the descriptions to the new geologist who • Subsequent drilling can be done destructively and all bor-
then made the field notes. The project succeeded because of ings logged with the same suite of geophysical tools. This
the experienced driller and the inexperienced geologist had a set of data can then be compared to the reference borings
great learning experience. In some cases a driller who has previously completed to aid in their interpretation.
considerable experience in a particular geologic setting will • If and when geologic conditions are seen to change a few
produce better quality drillers logs than an inexperienced additional reference borings may be required (continu-
geologist or engineer. ously cored, geophysically logged, and samples sent for
testing, as needed).
17.3.4.9 Maximize the Data from Boreholes
Some of the borings completed on-site will be used to acquire All too often core samples are missing in zones of weath-
additional data such as pump tests, engineering properties, ered rock and voids, which is of particular concern when
water levels or water quality samples over time. Some of working in karst. In addition, the need to acquire and safely
these tests or measurements require open boreholes, some store large amounts of core (and oriented core) can be
222 17 Invasive Methods

logistically difficult. Both of these issues can be resolved by trench will allow safe entry. Federal (OSHA) and State
obtaining a high resolution 360° image of the borehole using regulations cover the safety aspects of trenching.
an acoustic televiewer (ATV) in water-filled boreholes or an The success of trenching depends upon getting the trench
optical televiewer (OTV) in dry boreholes. Both the acoustic in the right location, sufficiently deep to expose the area of
and optical televiewer provide an excellent means of obtain- interest, and the skills necessary to provide detailed docu-
ing a continuous and detailed in-situ image of geologic con- mentation of the conditions using cross section or face maps.
ditions within an open borehole even in areas where samples A thorough cleaning of the trench walls and a detailed grid or
cannot be recovered. The ATV and OTV data is also digital string lines must be established over the surface of the trench
and can be presented as a virtual core sample, which can be wall. The trench wall must be carefully cleaned (much as
rotated for viewing. A discussion of geophysical logging is with an archeological excavation) and many more features
provided in the following section. and detail should be mapped on the face map, then should be
little left to conjecture (Hatheway 1982). This requires a high
level of observational and documentation skills. Age dating
17.4 Excavations and Trenches methods may include petrographic analysis, clay mineral-
ogic analysis and radiometric techniques. An objective
Outcrops and core from borings usually provide only a frag- trench log attempts to portray equally all physical features in
mented picture of soil and rock conditions. Excavations and an impartial manner without regard to relative importance
trenching offer one of the most definitive of all invasive and subjective interpretation is minimized during the record-
exploratory methods since it provides a means of direct ing process.
observation of geologic conditions over a larger area. Although photographs and video may be obtained, a
Trenches are necessary when attempting to assess details of detailed sketch along with a detailed descriptive analysis of
geologic conditions, assessing fractures and fissures, settle- the geology is generally the primary means of documenta-
ment, subsidence, or the dating of the movement induced by tion. Wide excavations might consider the use of electronic
an earthquake. distance meter (EDM), total station, photographic or video
Trenches can range from small shallow shovel cut trench documentation, (Keaton 2009; Haneberg 2009).
by hand to backhoe (Fig. 17.21) and large dozer cut trenches. Excavation for landfill expansion provided a large exposed
Safety becomes an issue if the trench is narrow and deep, geologic cut through an area being studied in Kansas City,
then bracing to prevent cave-in is required. A properly braced Kansas. The excavation was first photographed (Fig. 17.22a)

Fig. 17.21 Shallow trenches whether by backhoe (a and c) or dug by hand (b) can provide critical observations in the assessment of subsurface
conditions
17.4 Excavations and Trenches 223

Fig. 17.22 An exposed section of limestone and shale at the edge of a paleocollapse. The top of the shale was marked and elevations surveyed (a).
Cleaning the face of the dozer cut further exposed fractures in the shale (b)

and then sketched in detail. Then elevations were established design. See Part III, Chap. 25 for a more detailed description
across the entire face of the excavation along a distinct geo- of this project.
logic unit. This revealed a low area of 3 m in the center of the At another site in Pennsylvania, excavations were utilized
167 m wide excavation. This was the northernmost edge of a to assess the investigation of geologic pathways that might
large 7 ha paleocollapse at the site. Localized areas were fur- be controlling movements of organic contaminants from a
ther excavated using shovels and hand picks and exposed military base. Excavation revealed the complex top of rock
open fractures within the underlying shale (Fig. 17.22b). conditions (Fig. 17.23). The layers of limestone can be seen
This geologic window was unexpected and invaluable since dipping which played an important part in controlling the
it provided cross section through the edge of an obvious movement of contaminants at this site.
paleocollapse. Furthermore, it exposed open fractures in the An example of a very large trench is shown in Fig. 17.24.
uppermost shale at the site, which impacted on-going landfill This is from a site being characterized for a proposed
224 17 Invasive Methods

Fig. 17.23 An excavation clearly shows the complex nature of the top of limestone

Fig. 17.24 A large trench is being excavated to expose the top of rock and assess possible buried channel or paleokarst conditions

hazardous waste facility in southwest Texas. The site had a more than 15 m wide and about 11 m deep, which exposed
thick alluvium over limestone. Some anomalous areas (pos- the top of limestone. The large trench allowed detailed map-
sible buried channels or paleosinkholes) had been identified ping to be completed and evaluated possible karst conditions
by surface geophysical measurements. This site was a large at the top of rock.
open area, and heavy excavating equipment was readily Further details of trench mapping are provided in
available on-site so trenching was an easy and effective Hatheway and Leighton (1979), Hatheway and McClure
option to provide more detailed data. This large trench was (1979), and McCalpin and Shlemon (1996).
References 225

References Hatheway AW, Leighton FB (1979) Trenching as an exploratory

method. In: Hatheway AW, McClure CR Jr (eds) Geologic society
of America, geology in the siting of nuclear power plants. Rev Eng
ASTM (2006) Standard guide for use of hollow-stem augers for geoen-
Geol 4:169–195
vironmental exploration and the installation of subsurface water-
Hatheway AW, McClure CR Jr (eds) (1979) Geology in the siting of
quality monitoring devices. D5784-95(2006). American Society for
nuclear power plants. Geol Soc Am Rev Eng Geol 4:169–195
Testing and Materials, West Conshohocken
Hencher S (2012) Practical engineering geology, applied geotechnics,
ASTM (2008) Standard guide for using rock-mass classification sys-
vol 4. Spoon Press, London
tems for engineering purposes. D5878-08. American Society for
Hopkins R (1977) TVA philosophy and methods for site investigation.
Testing and Materials, West Conshohocken
In: Butler D (ed) Symposium on cavity detection of subsurface cavi-
ASTM (2009) Standard practice for soil exploration and sampling by
ties. U.S. Army Engineers Waterways Experiment Station
auger borings. D1452-09. American Society for Testing and
Vicksburg, Vicksburg, pp 25–26
Materials, West Conshohocken
Hunt RE (2005) Geotechnical engineering investigation handbook, 2nd
ASTM (2010a) Standard practice for using the electronic piezocone
edn. CRC Press/Taylor and Francis, Boca Raton
penetrometer tests for environmental site characterization.
Hvorselv MJ (1949) Surface exploration and sampling of soils for civil
D6067-10. American Society for Testing and Materials, West
engineering purposes. The Waterways Experiment Station, Vicksburg
Conshohocken
Keaton JR (2009) Close-range terrestrial photogrammetry for geotech-
ASTM (2010b) Standard practice for expedited site characterization of
nical characterization and visualization. Geo-Strata, March/April,
vadose and ground water contamination at hazardous waste con-
pp 18–20
taminated sites. D6235-04(2010). American Society for Testing and
Khedr SA et al (1985) Automated cone penetrometer a non-destructive
Materials, West Conshohocken
field test for subgrade evaluation. Annual transportation research
ASTM (2010c) Standard practice for sonic drilling for site character-
board meeting, National Academy of Sciences, Washington, DC
ization and the installation of subsurface monitoring devices.
Kratochvil G (1989) Geological evaluation of potential hazards to shut-
D6914-04(2010). American Society for Testing and Materials, West
tle landing operations on Rodgers Dry Lakebed. Astronaut Office
Conshohocken
NASA Lyndon B Johnson Space Center Houston Texas, Edwards
ASTM (2011) Standard test method for standard penetration test (STP)
Air Force Base
and split-barrel sampling of soils. D1586-11. American Society for
Kratochvil G et al (1992) Fishing for fissures. The Military Engineer,
Testing and Materials, West Conshohocken
Jan–Feb, pp 66–68
ASTM (2012a) Standard guide for field logging of subsurface explora-
McCall W et al (2006) Use of direct push technologies in environmen-
tions of soil and rock. D5434-12. American Society for Testing and
tal site characterization and ground-water monitoring. In: Nielsen
Materials, West Conshohocken
DM (ed) Practical handbook of environmental site characterization
ASTM (2012b) Standard test method for electronic friction cone and
and ground-water monitoring, 2nd edn. CRC Taylor & Francis,
piezocone penetration testing of soils. D5778-12. American Society
Boca Raton, pp 345–471
for Testing and Materials, West Conshohocken
McCalpin JP, Shlemon RJ (1996) Trenching as a tool in geologic inves-
ASTM (2012c) Standard guide for direct-push groundwater sampling
tigations. Prof Geol 33(12):19–22
for environmental site characterization. D6001-05(2012). American
Nielsen D (ed) (2006) Practical handbook of environmental site charac-
Society for Testing and Materials, West Conshohocken
terization and ground water monitoring, 2nd edn. CRC Taylor &
ASTM (2014) Standard practice for rock core drilling and sampling of
Francis, Boca Raton
rock for site investigation. D2113-14. American Society for Testing
Ruda T, Farrar J (2006) Environmental drilling for soil sampling, rock
and Materials, West Conshohocken
coring, borehole logging, and monitoring well installation. In:
Bell FG (2004) Engineering geology and construction. Spoon Press,
Nielsen DM (ed) Practical handbook of environmental site charac-
London
terization and ground water monitoring, 2nd edn. CRC Taylor &
Christy CD et al (1994) A percussion probing tool for the direct sensing
Francis, Boca Raton, pp 297–344
of soil conductivity, Geoprobe technical paper no. 94-100, Geoprobe
Sara MN (2003) Site assessment and remediation handbook, 2nd edn.
Salinas, Kansas
Lewis Publishers, New York
Fookes PG (1997) Geology for engineers: the geological model,
Schulmeister MK et al (2003) Direct-push electrical conductivity log-
prediction and performance. Geol Soc Q J Eng Geol 30(4):
ging for high-resolution hydrostratigraphic characterization.
293–424
Ground Water Monit Remediat Summer 23(3):52–62
Haneberg WC (2009) Rapid prototyping of computer models to charac-
Sowers GF (1996) Building on sinkholes, design and construction of
terize discontinuous rock masses. Geo-Strata, March/April, pp 22,
foundations in karst terrain. ASCE Press, Reston (out of print)
24–26
Waltham AC, Fookes PG (2003) Engineering classification of karst
Hatheway AW (1982) Trench, shaft, and tunnel mapping. Bull Assoc
ground conditions. Geol Soc Lond Q J Eng Geol 36(2):101–118
Eng Geol 19(2):173–180
 Geophysical Logging
18

Abstract
At this stage in a site characterization, we have refined the conceptual model based upon the
surface geophysics and located new borings or monitor wells with insight. While drilling
these boreholes has provided geologic logs and samples at some interval, there is additional
information that can be obtained. This is where geophysical logging techniques are applied.
This suite of measurements provides detailed, often continuous, vertical data within open
boreholes and cased wells.
In addition, there may be existing borings, monitoring wells, supply wells on-site that
provide opportunities for gathering additional detailed geologic and hydrologic data. This
chapter provides a brief introduction to geophysical logging along with many consider-
ations for using this type of measurement. Examples are also provided showing a variety of
the geophysical logging measurements and their application.

18.1 Introduction tool in an open or cased borehole. The digital data is dis-
played on the computer screen and recorded on a computer.
Geophysical logging provides measurements of a wide vari- The raw data can also be printed as a hard copy in the field.
ety of parameters down the length of an open borehole or The logging tools provide in-situ measurements of one or
cased monitor well, similar to the surface geophysical meth- more parameters of borehole fluids and undisturbed soil and
ods (Fig. 18.1). Geophysical logging is commonly used in rock conditions surrounding the borehole. Most logging
the oil and gas industry as part of exploration and develop- tools provide essentially continuous measurements at a
ment. This logging is done by specialty logging contractors sample rate of 0.3 cm as the probe is moved up or down the
with large dedicated trucks and sophisticated logging tools, hole. Some logging tools use station measurements that are
which may be many cm in diameter and several meters long. acquired as the probe is held at a specific depth over some
Here, we are discussing applied geotechnical and hydrologic time interval. The depth or penetration of the measurement
logging in small diameter boreholes and wells. varies from the fluid within the borehole to the borehole
Figure 18.2 is a typical small logging system, from Mount wall up to a meter into the soil or rock depending upon the
Sopris, used for characterizing conditions within open bore- particular log being used.
holes and cased monitoring wells. The systems consist of a While drillers and geologic logs are subjective based
laptop computer, system control unit, winch (Fig. 18.2a), tri- upon personal experience, quality of samples retrieved and
pod over the borehole or well (Fig. 18.2b) and a variety of visual interpretation, the geophysical logging measurements
logging probes that are lowered into a borehole to acquire are very non-subjective. The geophysical logging measure-
data (Fig. 18.2c). This system has been installed in a vehicle, ments provide in situ measurements and are virtually 100 %
however, some systems are small enough to be carried to repeatable. It is the interpretation of the logging data that can
remote locations. Most logs can be run in holes as small as be subjective.
5 cm diameter and therefore can be run in most cased monitor The objective of geophysical logging is to obtain more
wells. Logging is carried out by lowering or raising a logging detailed and less subjective measurements from a borehole

© Springer Science+Business Media Dordrecht 2016 227

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_18
228 18 Geophysical Logging

or well than can be obtained from drilling and sampling mation along the length of a hole. These data can be used to
alone. Drilling any kind of a borehole or well is an added characterize geologic and hydrologic and engineering
cost to the project and the borehole provides data at just one properties of the unsaturated and saturated zones within open
point over the site. Therefore, the data from each borehole or cased wells and help to constrain the vertical geohydro-
or well should be maximized with the addition of geophysi- logic details in our conceptual model of the site.
cal logging. The combination of a driller’s log, or geologic Further details about logging can be found in Hearst et al.
log, with geophysical logs provides an increased confidence (2000), Chapellien (1992), and Killeen (1986).
level in our interpretations of soil and rock conditions.
While the resolution of all surface geophysical measure-
ments decreases with depth, geophysical logging measure- 18.2 Geophysical Logging Measurements
ments maintain their resolution independent of depth
providing high resolution, nearly continuous, in-situ infor- A wide range of geophysical logs are available that can be
used to gather different types of data for determining the
geologic and hydrologic conditions. The purposes for log-
ging may include:

• acquiring detailed geohydrologic data after new borings

have been properly located,
• recovering geohydrologic data from existing borings and
• evaluating the condition of existing wells.
Increasing Depth

Geophysical logging data provides an independent means

of obtaining geologic and hydrologic data from old and new
borings. This provides a consistent set of data that allows the
geologic log and the geophysical logs to be compared within
a single borehole as well as between boreholes. The geo-
physical logs can also be used as a means of quality control,
provide multiple data sets for correlation and to aid to
interpretation.
Table 18.1 lists the more commonly available logs along
Variations in
with the parameter that is measured or calculated, and the
Parameter
Measured applications in which the logs may be used. The most generic
name for the particular log is used avoiding names applied
Fig. 18.1 Sketch of typical downhole logging measurements by specific manufacturers and service companies. The logs

Fig. 18.2 Photos of typical logging system in operation. (a) Control unit, laptop and winch. (b) Tripod over well. (c) Probe being lowered into
well
18.2 Geophysical Logging Measurements 229

Table 18.1 Summary of common geophysical logs

Saturated or
Parameter measured or Open hole, PVC unsaturated Radius of measurement
Log calculated or steel borehole (cm and m)
Electrical/EM logs Induction Bulk electrical conductivity Open hole or PVC Both About 75 cm
of soil, rock, and pore fluids
and magnetic susceptibility
of soils and rock
Resistivity Electrical resistivity of soil, Open hole Saturated Up to a meter or so
rock and pore fluids
Single point resistance Electrical resistance of soil, Open hole Saturated Small/variable
rock and pore fluids
Spontaneous potential Voltage – responds to Open hole Saturated Small/variable
(SP) electrochemical effects of
differences in borehole
fluids, oxygen reduction of
minerals and streaming
potential due to movement
of pore fluids
Fluid logs Conductivity Electrical conductivity of Open hole Saturated Within the borehole
borehole fluids
Spinner flowmeter Flow rate of fluid within the Open hole Saturated Within the borehole
borehole
Heat pulse flowmeter Flow rate and direction of Open hole Saturated Within the borehole
fluid within the borehole
fluid
Temperature Temperature of borehole Open hole Saturated Within the borehole
fluid
Chemical sensors Selected chemical Open hole Saturated Within the borehole
parameters of the borehole
fluid
Mechanical logs Caliper Diameter of borehole or Typically open Both Borehole wall or casing
casing hole
Deviation Inclination and direction of Open hole or PVC Both Borehole
borehole (deviation from Steel using gyro
vertical) system
Visual/imaging logs Television Visual image of borehole or Typically open Both View range a few cm to
casing hole about a meter
Optical televiewer A 360° optical image of the Typically open Both View range up to about
borehole wall (or casing) hole 30 cm
Acoustic televiewer A 360° acoustic image of Typically open Saturated View range 10 cm
the borehole wall (or casing) hole
Nuclear logs Natural gamma Total count rate of natural Any Both 15–30 cm
gamma radiation
Natural gamma- Identification and Any Both 15–30 cm
spectrometry quantitative analysis of the
radioisotopes that
contribute to the total
natural gamma count rate
Gamma-gamma Relative density or true Any Both 15 cm
density, if calibrated
Neutron Relative moisture/fluid Any Both 15 cm
content in unsaturated zone,
porosity in saturated zone
(continued)
230 18 Geophysical Logging

Table 18.1 (continued)

Saturated or
Parameter measured or Open hole, PVC unsaturated Radius of measurement
Log calculated or steel borehole (cm and m)
Other logs Radar Provides EM reflection data Open hole or PVC Both About 3–15 m
at greater distance from the
borehole
Gravity Provides changes in density Any Both Up to 30 m or more
at greater distance from the
borehole
Full waveform sonic Travel time and magnitude Open hole Saturated Up to a meter
of seismic signal in soil and
rock

have been divided into six groups based upon the way they borehole while the optical log can operate in either a dry
make measurements. or water-filled borehole.
Nuclear Logs – nuclear logs are measuring either natural
Electrical/Electromagnetic Logs – These logs measure elec- radiation or back scattered radiation from a radioactive
trical properties of the subsurface [conductivity, resistivity, source used in the logging tool. The natural gamma log
resistance or spontaneous potential (voltage)]. They aid in measures natural radiation, and is a more commonly used
characterizing lithology/stratigraphy, identifying fractures log to map lithologies such as clays, shales and weathered
and cavities, detecting inorganic pore fluids (ie. salt water zones. A spectral gamma log also measures natural radia-
intrusion, inorganic contaminants, etc.) and in some cases tion and can differentiate between radioactive isotopes.
aid in detecting flow. These logs typically require an open The logs utilizing a radioactive source are much less com-
borehole with the exception of the electromagnetic log, monly available but are very useful when mapping varia-
which can be used in a PVC-cased borehole. tions in density (gamma-gamma log) and porosity
Fluid Logs – The fluid logs measure a property of the borehole (neutron log). All of these nuclear logs can be used in
fluid such as fluid conductivity, temperature, flow of fluid PVC or steel-cased wells.
(spinner flowmeter or heat pulse logs), and tracers within Other Specialty Logs – there are a few specialty logs such as
the fluid or specific chemistry of the fluid. These measure- sonic waveform, ground penetrating radar and gravity. Sonic
ments require an open borehole in which the fluids are in logs can provide an in situ measurement of P and S wave
equilibrium in order to be representative and are made by seismic velocity within a borehole. Borehole radar measure-
slowly lowering the probe down the borehole, while most ments can be made from a single borehole with a transmitter
other logs are made coming up the hole. If there is fluid and receiver antenna located in the same borehole or between
flow through fractures in the borehole wall this flow can be holes with the transmitter and receiver in different boreholes.
detected by a flowmeter or changes in temperature or fluid The radar range is dependent upon soil or rock type. In resis-
conductivity to measure the presence and rate of flow. tive rocks ranges of 10–40 m can be achieved. However, in
Mechanical Logs – the mechanical logs include caliper and conductive, clay-rich or silty rock ranges can be limited to
deviation logs. The caliper log is mechanically measuring 5 m or less. Borehole radar has been used for a variety of
the borehole diameter allowing identification of variations applications including bedrock fractures, voids, and tracing
caused by fractures, cavities, wash-outs, etc. It can also be tests. Singha et al. (2000) provides a brief summary of bore-
used in cased boreholes to confirm casing diameter, where hole radar theory and its application.
casing ends and open borehole begins, or where damage to
a casing has occurred. The deviation tool measures the devi- Unlike conventional nuclear density logs, borehole gravity
ation of the borehole from vertical, providing corrections measurements provide density measurement for large vol-
for other measurements made within the same borehole. umes of rock surrounding the borehole and are not affected
Video and Imaging Logs – video logs provide an image by borehole conditions. The effective radius of investigation
down the axis of the borehole or the borehole wall. This is about 5 times the spacing between measurements and can
can be used to identify lithology as well as fractures, cavi- be 10’s of meters or more. Detection of structure, faults, cav-
ties in an open hole or conditions within a cased well. The erns and mines can be detected at significant distance from
imaging logs (acoustic televiewer and optical televiewer) the borehole. More information can be obtained from the
provide a digital 360° image of the borehole wall that can manufacturer, Scintrex Ltd. (www.scintrexltd.com).
be used for quantitative analysis of strike, dip and aper- While most logging tools are measuring one parameter at
ture of fractures. They can also be used to inspect cased a time, there are a few that measure two parameters at a time
wells for damage. The acoustic log requires a water-filled such as resistance/spontaneous potential or temperature/fluid
18.2 Geophysical Logging Measurements 231

Fig. 18.3 Typical presentation of geophysical logs including the drillers log and well construction

conductivity. In addition, some manufacturers sell probes that No significant geologic variations were noted in the drill-
can be combined so that one pass in the borehole can measure er’s log. However, the geophysical logs show a distinct
multiple parameters. A common combination includes natu- change in conditions between 24 and 27 m deep. This area is
ral gamma, resistance and spontaneous potential. within the limestone and shows the presence of clay and has
Figure 18.3 is a typical presentation of geophysical log- a lower density than the rock above and below indicating a
ging data along with the corresponding geologic log and well possible filled cavity, fracture zone, etc. It is this ability to
construction information. The site is in northern Alabama detect these subtle conditions that makes an array of geo-
where there is a clay residuum overlying limestone and is physical logging tools such an asset to the site characteriza-
unsaturated. The project included the logging of all existing tion process.
wells. The purpose was to provide a uniform set of data in
which to assess geologic conditions and to identify possible
karst conditions. Three geophysical logs were used that all 18.2.1 Key Aspects of Geophysical Logging
provide data within steel cased wells and included natural
gamma, neutron (porosity), and gamma-gamma (density). In The selection of logs to be used for a project will primarily
this example, the existing well had a steel surface casing to be based upon the objectives of the project. However, logis-
about 12 m deep. The three logs are all responding to relative tics and site-specific conditions will also impact the selec-
variations in three different properties: tion. For example, if you are working in newly drilled
boreholes, will the open borehole be stable or is a temporary
• natural gamma log – variation in clay content, casing needed, and are the boreholes saturated or dry. If you
• neutron log – variations in porosity, and are logging existing monitoring wells, what are their diame-
• gamma-gamma log – variations in density. ter and casing materials? There are a wide variety of logging
232 18 Geophysical Logging

Borehole Fluid - temperature, fluid conductivity,

flow, various chemical sensors
Borehole Wall - caliper, video log, acoustic
and optical televiewer

15 to 45-cm radius - nuclear logs, some electrical logs

0.75 m radius or more - induction, full waveform sonic

Fig. 18.4 The radius of measurement of some geophysical logs

measurements that can be made in a variety of borehole and If logging open boreholes is not feasible, and all measure-
well conditions. Although many logging measurements ments must be made through casing then you are limited to
require an open borehole, some limited measurements can be the possible logs to choose from which includes:
made from within steel or PVC cased wells with a minimum
diameter of 5 cm. Table 18.1 includes some of the generic • Natural gamma log, if the fractures are clay filled (PVC or
conditions under which each of the logs can be used. There steel casing),
are other considerations when incorporating geophysical • Induction conductivity log if the fractures are clay filled
logging into a site characterization. Some of these key or have slightly conductive pore fluids (PVC casing only),
aspects of the geophysical logging methods that should be • Gamma-gamma log to detect an decrease in density (PVC
considered when selecting appropriate logs include: or steel casing), or
• Neutron log to detect an increase in porosity (PVC or
• Parameter measured, steel casing).
• Radius of measurement,
• Borehole conditions,
• Vertical resolution and speed of logging, 18.2.1.2 Radius of Measurement
• Sequence of logging, and Most geophysical logs measure 360° from the logging tool in
• Processing and presentation of data. all directions over some distance. The radius of sampling (or
depth of penetration) ranges from measuring the fluid in the
borehole, to the borehole wall, to a few centimeters into the
18.2.1.1 Parameter Measured borehole wall to a meter or more into the formation (Fig. 18.4).
Each of the borehole logging measurements are responding to The measurements of in situ soil and rock properties become
changes of some physical, electrical, or chemical property of more representative as the volume of material sampled by
the borehole, its fluids or geologic conditions (Table 18.1). geophysical logging increases beyond the borehole wall.
Changes in geologic and hydrologic conditions will be asso- While not as commonly used, full waveform sonic can pro-
ciated with a change in one or more of the measured param- vide data up to a meter from the borehole, but typically much
eters. For example, if detecting fractures is the objective you less. Ground penetrating radar can provide data up to several
could possibly use one or a combination of logs depending meters from the borehole and gravity measurements can pro-
upon site-specific conditions. If an open borehole is available, vide data to tens of meters or more from the borehole.
then caliper logs may be your first choice as a simple straight-
forward method. If more information on the fracture is neces- 18.2.1.3 Borehole Conditions
sary, then you might use downhole video or imaging using Variations in borehole conditions can affect or even limit the
acoustic or optical televiewer logs. If determining flow along use of some logs. Some logs will allow measurements to be
the fracture is of interest then the options include spontaneous made in the unsaturated zone as well as the saturated zone,
potential, temperature and fluid conductivity which all may while others can only be run in the saturated zone. Most logs
provide an indication of flow. If you want to quantify the flow require an open hole while some logs will provide measure-
then the options are a spinner flowmeter or heat pulse logs. ments from inside plastic or steel casing (Table 18.1).
18.2 Geophysical Logging Measurements 233

The borehole or well diameter has to be large enough to conditions, which may cause tools to be stuck in the
allow the logging tool to pass smoothly up and down the borehole.
borehole or well. The longer the logging tool, the more sus- In general, all logs except fluid properties and video
ceptible it is to bends in the casing or deviations in the bore- should be run with the probe moving up the borehole to
hole. Therefore, you need a borehole diameter larger than the reduce depth errors. Most logs are run continuously through
logging tool and large enough to accommodate those bends the length of the borehole. Some measurements such as (P
or deviation without causing the tool to get stuck. However, and S waves and gravity) are made at discrete intervals along
it is also possible to have a borehole diameter that is too the borehole.
large. If a logging tool only measures out a few cm in radius Nuclear logs with radioactive sources such as gamma-
then a larger borehole would reduce the effectiveness of the gamma (density) logs and neutron (moisture/porosity) logs
measurements. In a very large diameter borehole, the log- present special issues. The terms of licensing requires the
ging tool can be constrained by a spring up against one side operators to minimize the chances of a radioactive source
of the borehole (using a decentralizer spring) to obtain more being stuck down a borehole. Because of the consequences
effective measurements. of loosing a tool with a radioactive source, unstable bore-
holes should not be logged with radioactive probes. An
18.2.1.4 Resolution and Speed of Logging approach used by the authors is to only run nuclear logs
Vertical resolution will depend upon the digital sample rate through the drill stem, after final installation of the well or to
and the speed at which the logging tool is moved through the have the borehole temporally cased with a PVC casing in
borehole. Typical sampling for most logs is on the order of which to run the nuclear logs. If using a temporary casing, it
every 0.3 cm with a typical logging speed of 5 m/min. This can then be removed and other logs can be acquired in an
provides essentially continuous data, for geologic purposes, open borehole. If a nuclear log is run in an open hole, the
over the length of the borehole. The radial resolution of the hole must be run with a video log first to assure a clear path
geophysical logs is constant with depth and is dictated by the and stability of the borehole.
specific logging tools penetration of measurement.
While a typical logging speed for many logs is about 5 m/ 18.2.1.6 Processing and Presentation of Data
min, some logs require a much slower speed of logging. The typical geophysical logs are printed as raw data versus
Video logs, which are recording images, utilize a much depth (Fig. 18.3). Many of the logs are used to identify rela-
slower speed of logging so that the images can be viewed tive changes and can be interpreted directly from the field
effectively. Temperature and fluid conductivity logs are also data once the data have been plotted. A few of the logs such
run much slower (about 1 m/min) to allow the sensors to as full waveform sonic, ground penetrating radar and gravity
respond to changing conditions and minimize the distur- will require different levels of processing and presentation.
bance of the borehole fluids. Geophysical logs can be used in a qualitative or quantita-
tive manner, depending upon project objectives. For example,
18.2.1.5 The Sequence of Logging a gamma-gamma (density) log can be used to indicate that one
The sequence of logging or the order the tools are used may zone of rock is more or less dense than another zone. This
affect the quality of the data depending upon the logs selected same data can be expressed in density units if the probe has
for a project. been calibrated to known conditions or samples from the bore-
Measurements within the water column require the bore- hole have been tested in the laboratory for comparison. While
hole fluids to be in equilibrium with ambient conditions. quantitative results can be obtained from most geophysical
After drilling, and possibly installing a well, sufficient time logs, caution must be used in over-interpreting and quantify-
must be allowed so that the borehole can return to equilib- ing of data, if proper calibration to site-specific conditions or
rium. This may require from 10 to 100 times the duration of correlation to laboratory analysis has not been properly made.
the drilling. Even the logging operation mixes the borehole ASTM has developed a guide for planning and conduct-
fluid. Therefore, logs measuring fluid properties (tempera- ing borehole geophysical logging measurements D5753-95
ture, fluid resistivity, and fluid sampling) should be run prior (ASTM 2010a) along with a few guidelines for specific geo-
to other logs to minimize mixing of borehole fluids. physical logging measurements. They include:
Consideration should also be given to when video logs are
run because some logging tools may degrade borehole con- • ASTM D6274-98 for gamma logging (ASTM 2010b),
ditions and borehole fluid clarity. For example, caliper logs • ASTM D6726-01 for electromagnetic induction logging
or those using bow-springs that contact the borehole wall (ASTM 2007a),
should be run late in the logging sequence. The contact with • ASTM D6727-01 for neutron logging (ASTM 2007b), and
the borehole wall creates a greater possibility of material • ASTM D6167-97 for mechanical caliper logging (ASTM
falling into the borehole, degrading visibility and creating 2011).
234 18 Geophysical Logging

18.3 Various Applications 18.3.2 Low Density Zones, Fractures

for Geophysical Logs and Cavities

There are a wide variety of applications for the geophysical Development on Grand Cayman, Cayman Islands in the
logging methods. The application requires a borehole or Caribbean, required an injection borehole. A suite of geophys-
existing well and a contrast in the parameter being measured ical logs was run in a borehole in order to identify an optimum
by the specific logging tool. The applications can generally depth for injection. The objective was to identify a more per-
be categorized into six areas and include: meable zone with fractures or cavities. The limestone of Grand
Cayman is known for its dissolution features, so cavities were
• Assessment of natural geologic and hydrogeologic expected as well as low density weathered rock and small to
conditions; large fractures. Figure 18.7 shows a portion of the geophysical
• Measurement of soil and rock properties for engineering logs acquired within this borehole in the saturated zone. The
applications; suite of logs all indicated a broad zone of cavities extending
• Detection and mapping of contaminant plumes and spills; from 63 to 69 m deep at the contact between two formations.
• Evaluation of construction and conditions within existing Each of the logs responded to this cavity zone:
monitoring wells and piezometers;
• Evaluation of man-made soil, rock or concrete • The caliper log indicated an increase in borehole
structures; and diameter;
• Groundwater and mineral exploration and evaluation. • The resistance log indicated a decrease in resistance asso-
ciated with water-filled cavities;
The following examples illustrate some of the wide range • The spontaneous potential indicated an increase in volt-
of applications for the geophysical logs as applied to geol- age (due to possible flowing water);
ogy, hydrology and karst problems. These are but a few • The gamma-gamma (density) log indicated a decrease in
examples to highlight some of the logging methods. There density (an increase in cps); and
are a wide variety of logging tools and variations on their • The neutron (porosity) log indicated an increase in poros-
application that are available. ity (a decrease in cps).

This is an excellent example in which multiple logs all

18.3.1 Mapping Stratigraphy indicated a zone of increased fractures, porosity and cavities.
The use of multiple logs also improved confidence levels in
Geophysical logging was used effectively to map the lime- the interpretation.
stone and shale layers at a site in Kansas City, Kansas. The Another example of logging data within a complex hydro-
site was a landfill that was expanding over an abandoned geologic system is from the Woodville Karst Plain (WKP) in
limestone mine (see case history in Part III, Chap. 25 for northern Florida. After a test phase utilizing several geo-
more details). The natural gamma log provided an extremely physical techniques, microgravity was selected as the most
repeatable response to the limestone and shale layers as well effective means for regional surveying within the WKP in
as strong responses to three key marker beds of shale order to detect large mass deficits associated with conduits
(Fig. 18.5). The limestone is clearly shown by the lower and caves (see Part II, Sect. 16.5.3 for more details on this
gamma values and certain shale beds by higher gamma project). A total of 43 km of gravity data was obtained. Four
values. of the large regional microgravity anomalies (70–90 uGals)
The three obvious shale beds in the natural gamma log indicating mass deficits within the subsurface were drilled
were used as marker beds for correlation between boreholes. and geophysically logged (Yuhr et al. 2008a).
In this case, the consistency of the gamma logs between The field plan for the geophysical logging required two
boreholes (Fig. 18.5) is a measure of uniform geologic con- phases. The first phase was a suite of logs that were acquired
ditions over the site. Thirty-nine natural gamma logs were through a temporary 5 cm PVC casing because radioactive
acquired across the site using existing mine fillholes and old source logs were being used. This included the nuclear logs
mine vent holes. Surface elevation for each borehole was natural gamma, gamma-gamma, neutron-neutron and an
established and the elevation of the uppermost shale layer induction log. The temporary casing was then removed and
seen in the natural gamma log was used to develop the con- fluids within the boreholes were allowed to stabilize for 24 h.
tour map of this shale layer. The contour map clearly shows The second round of logging included temperature, fluid
a distinct depression in the shale layer, as much as 5 m conductivity, and caliper. Downhole video was added in all
(Fig. 18.6). This data helped define a paleocollapse zone, boreholes after the caliper log got stuck in the first borehole
which originated 180–240 m below the mine. at a depth of 75 m.
18.3 Various Applications for Geophysical Logs 235

Fig. 18.5 Correlation of stratigraphy using geophysical logs. Note the good correlation between boreholes using the shale marker bed

Figure 18.8 shows the complete suite of logs acquired in of 90 m. The first phase of logging was completed through the
one of the four boreholes (B6-1). An open borehole was diffi- PVC casing, then it was removed. The fluid logs (temperature
cult to maintain during drilling due to the numerous fractures and fluid conductivity) were run at the bottom of the open
and cavities encountered. A steel surface casing extended to borehole (70–90 m). The steel casing was then removed and
about 70 m and a temporary PVC casing was used inside the the fluid logs were then run in the upper 70 m. The caliper log
steel casing extending to the bottom of the borehole at a depth was the last log to be run and could not be lowered to the
236 18 Geophysical Logging

Fig. 18.6 A contour of the uppermost shale bed identified in the geophysical logs (Fig. 18.5) from 39 boreholes shows a depression of about 5 m
that corresponds to a paleocollapse. Note contour lines are elevation in feet

bottom of the borehole due to an obstruction at 70 m. Therefore of borehole geophysical measurements to characterize the
data is only from the upper 70 m. This borehole is an example geology and possible brine cavities from the dissolution min-
of some of the difficulties that are often encountered when ing of salt. Data included 13 borings, core analysis, geophysi-
drilling and logging in karst conditions. cal logs (natural gamma, density, ATV, and deviation logs),
Figure 18.9 shows the gamma-gamma (density) and cali- vertical seismic profiling (VSP), crosshole and borehole grav-
per logs from all four boreholes. This data clearly show the ity measurements. The borehole gravity measurements were
broad low-density zones within each of the boreholes. There effective in identifying depth and thickness of possible brine
were also discrete cavity zones indicated in three of the bor- cavities (Michigan Department of Transportation 2008).
ings. One of the boreholes B2-1 was not run with a caliper Further details of the wide range of borehole based geo-
due to the tool getting stuck and not being able to maintain physical techniques intended for investigation of deep waste
an open uncased hole. depositories along with world-wide examples including the
The downhole video revealed turbulent flow at discrete cavi- Detroit River Crossing Bridge are found in Williams et al.
ties and upward flow in two of the borings (B2-1 and B6-1). (2009). These same methods are also applicable to site char-
This was based upon movement of particulate matter in the acterization of karst and pseudokarst conditions.
water column. The other two borings (B5-1 and B5-2) showed
little to no particulate movement, however observations of a
small crawfish-like creature in one of these borings (B5-1) indi- 18.3.3 An Alternate to Core Samples or
cate some level of connectivity to a more open conduit system. Oriented Core
The logging data confirmed the presence of low-density
zones mapped by microgravity data. This then supported the Core sampling is a standard approach for detailed geologic
use of the microgravity data to estimate the locations and assessment of rock and necessary in order to provide samples
frequency of higher porosity zones within the WKP beyond for engineering tests and laboratory analysis. Oriented core
those mapped by cave divers and dye tracing. This data could has been used to determine not only the presence of fractures,
also be used to support a groundwater model of the WKP. but their orientation, strike and dip. However, if detailed min-
Another example is from an investigation of the proposed eralogical diagnostics or geotechnical laboratory analysis are
new Detroit River Crossing Bridge that utilized a wide range not required from the core then an alternative approach is
18.3 Various Applications for Geophysical Logs 237

Increasing
Density

60

Cayman
Formation
63
gray to tan

Low Density Zone

dense

High Porosity and

fossiliferous
dolomite
Cavities
Depth in Meters

66
Brac
Formation
white to tan
69
very dense
fossiliferous
dolomite
and brown
limestone
72

75
0 10 20 30 40 50
Caliper Log (cm)

Increasing
Porosity

Fig. 18.7 A suite of geophysical logs from the Cayman Islands identifying a porous/cavity zone for injection of fluids

available. Both the acoustic televiewer (ATV) and optical developing groundwater models, establishing vulnerability
televiewer (OTV) provide an excellent means of obtaining a mapping or assessing contaminant transport and distribution.
continuous and detailed in-situ image of geologic conditions Some of the geophysical logging tools can indicate flow by
within an open borehole similar to a core (Fig. 18.10). This changes in a temperature or fluid conductivity log. Other
data was obtained by Colog, who was part of the site charac- geophysical logging tools such as impeller or heat pulse flow
terization team at Marshall Space Flight Center in Huntsville, logs can measure flow directly.
Alabama. The level of detail is as good or better than that Another method that has been developed is called
obtained from oriented core. In addition, the log does not HydroPhysical logging (Colog 2005). This technique uses
induce fractures that might occur during coring and the log an open borehole filled with deionized water. Time series
provides data in weaker zones where core may not retrieved. fluid electrical conductivity logs are then run. Changes in
Furthermore, the 360° data allows a pseudo-image of the core the fluid conductivity are measured as the borehole returns
to be developed. Quantitative structural analysis can be done to ambient conditions and natural groundwater moves both
such as fracture aperture, strike, dip, rose diagrams and ste- into the borehole, displacing the deionized water.
reo-net plots. Since the logs are digitally recorded, the storage Figure 18.11a is a plot of the time series fluid conductivity
and maintenance of core is avoided. logs in a single borehole. The data has been annotated
showing the flow into and out of the borehole. The changes
measured in the fluid electrical conductivity logs over time
18.3.4 Groundwater Flow and Contaminants can be used to calculate various hydrologic parameters for
specific intervals or a discrete feature. These calculations
Mapping of geologic features such as cavities, fractures, include flow rates, hydraulic conductivity and transmissiv-
void, etc. is only one part of the investigation. The connectiv- ity. These tests can be run under ambient or pumping con-
ity and flow associated with these features is required when ditions. This is particularly useful when hydrologic
238 18 Geophysical Logging

Natural Gamma (cps) Gamma-Gamma (cps) Caliper (inches)

2000

4000
1000

3000
100
20

40

60

80

12

16

20
0

8
0
Increasing Density

30
Gamma

Caliper
Depth in Meters

Neutron

60

Cavity Zone
Temperature

Increasing Porosity
90
21.1

21.2

21.3

21.4

21.5
21
0

200

300
100

400

Neutron Temperature (C)

Fig. 18.8 A suite of geophysical logs from the Woodville Karst Plain in northern Florida showing a cavity zone

conditions or connectivity of karst conditions are of information was integrated with a diverse suite of existing
concern. data and then used to guide the location of twenty-eight (28)
A geologic and karst characterization was completed at deep bedrock borings to assess groundwater flow and
the Marshall Space Flight Center (MSFC) in northern quality.
Alabama. It was a multi-phase project that incorporated mul- The integration of surface geophysical data with existing
tiple surface geophysical methods, drilling and borehole data proved to be highly beneficial when selecting the loca-
geophysical logging methods. The primary objective of the tions for bedrock wells to intercept groundwater flow.
surface geophysical investigation was to evaluate the extent, However, the drilling and logging program indicated that,
occurrence and hydrogeologic significance of the transition even in the most significant anomalous areas, no major karst
(epikarst) zone and deeper karst features that may provide features were encountered (e.g. old collapse features, large
preferential groundwater flow paths (Yuhr et al. 2005). This dissolutioned zones, caves, etc.). Instead, small discrete
18.3 Various Applications for Geophysical Logs 239

Fig. 18.9 The caliper and gamma-gamma logs from the four boreholes in Woodville Karst Plain. The caliper is plotted as a single line and the
density log is shaded with density decreasing as values increase to the right

fractures or fracture zones were detected and were pervasive formed by acid leaks into the underlying limestone at an industrial
throughout the depths drilled (75–90 m deep). Figure 18.11b facility. This is type of pseudokarst can rapidly dissolve the lime-
plots the hydraulic conductivity of the flow zones (calculated stone creating problems over short periods of time (tens of years).
from the HydroPhysical™ logging) versus their depth and Two small adjacent sinkholes had developed along side of
thickness to illustrate the spatial distribution of the flow a large manufacturing facility in Puerto Rico. They were
zones. The highest flow was focused within the upper 22 m. each less than a meter in diameter and one was about 3 m
This upper 22 m high flow zone corresponds to the weath- deep while the other was about 11 m deep. An initial assess-
ered limestone in our conceptual model (Yuhr et al. 2008b). ment was made with a borehole video camera lowered into
The small discrete fractures and limited fracture zones that the deeper sinkhole. There appeared to be a problem with the
were detected appear to be the key factor in controlling camera as it shut down when it was lowered into the deeper
groundwater flow. sinkhole. While examining the camera, it became apparent
that it was hot to the touch and likely overheating. After
examining the camera, which appeared to be operational, the
18.3.5 Pseudokarst Due to Acid Leaks problem was found to be an increase in temperature with
depth in the subsurface that was overheating the camera.
Logging can be very effective whether evaluating naturally occur- While the site is located in a karst terrain, it was becoming
ring karst or pseudokarst. This example of pseudokarst was obvious that these sinkholes were not naturally occurring.
240 18 Geophysical Logging

Fig. 18.10 Examples of ATV log (a) and OTV log (b) illustrating their ability to characterize fractures (Colog 2005)

The elevated temperatures and unusual odor within the sink- A conceptual model was developed for the void space and
holes indicated that there might be acid leaking from the a grouting plan was designed based upon this investigation.
plant and actively dissolving the limestone. These two small A total of 400 m3 of grout were injected in the upper 23–30 m
sinkholes were an early indication of possibly larger voids, to stabilize the structure.
which could critically impact the structure.
Microgravity measurements were made in and around the
building to assess the potential mass deficit in the subsurface 18.3.6 Corrections Due to Borehole Deviation
due to dissolution. The gravity survey indentified a 100-micro-
Gal anomaly centered at the two sinkholes. The mass deficit Most borings are drilled vertical and we assume that the
calculated from the gravity data was approximately 650 m3, borehole is vertical throughout its total depth. However, it is
much larger than observed at the two sinkholes. A drilling not uncommon for a borehole to deviate significantly from
program was then initiated using both vertical and angled the vertical. The impact of borehole deviation can be negli-
boreholes along with geophysical logging, which confirmed gible in many applications, but can induce significant errors
significant dissolution and voids to a depth of 45 m. in certain measurements, such as deep water level measure-
Two geophysical logs were used in the borings, a natural ments with a relatively flat water table, or hole-to-hole seis-
gamma log to indicate variations in clay content and an elec- mic measurements and other geotechnical measurements.
tromagnetic induction log to indicate variation in conductiv- Therefore, for those critical measurements, we need to know
ity. Figure 18.12 shows an example of these logs from three if the borehole deviates from vertical and the extent of the
of the vertical borings. The consistent low values of deviation.
10–20 cps in the natural gamma log suggest that there was no A commonly used deviation log (or inclinometer probe)
clay present. Background conductivity values for limestone utilizes magnetic and accelerometers sensors, and provide
are typically 10–20 mS/m. The induction logs measured con- continuous data. But these probes are not effective within
ductivity of 50 mS/m with values increasing to greater than steel casing. For steel cased holes or areas with strong mag-
200 mS/m. The combination of no clay indicated by the nat- netic soil or rock a gyro-based system can be used. The older
ural gamma logs, high conductivities measured in the induc- Fotobar deviation probes, which use bubble levels and a
tion log, and high temperatures in the subsurface all indicated 16 mm camera are still in use and provide data at discrete
the presence of extensive acids. depths.
18.3 Various Applications for Geophysical Logs 241

a
0

Flow
into

Flow
into

Depth in Meters

Flow out of borehole

Flow
into
Flow
into
73

Example of HydrophysicalTM logging data

b

4.2
m)
ne(
Zo
low
0
Cen o fF
ter D ess
epth kn
of Flo ic
w Zo Th
ne (m 0
)
76
Plot of flow zones hydraulic conductivity (taken from
hydrophysical data) versus depth and thickness

Fig. 18.11 An example of HydroPhysical logging (Colog 2005) (a) and the spatial distribution of flow zones identified by the HydroPhysical data (b)
242 18 Geophysical Logging

Aproximate Center of
Gravity Anomaly
Mass Deficit Indicated
South North

Two Small
Sinkholes

BH-1

BH-2

BH-3
Natural Natural Natural
Conductivity Gamma Conductivity Gamma Conductivity Gamma
(mS/m) (cps) (mS/m) (cps) (mS/m) (cps)
100 200 0 40 100 200 0 40 100 200 0 40
00 0 0

12
Area of Acids

18
Depth in Meters

Area of Acids

24
Zone of Muck

Area of Acids
30

36

42

48

54
0 40 0 40 0 40
o o o
Temperature ( C) Temperature ( C) Temperature ( C)

Temperature Circulation
Cavities
Measurement Lost

Fig. 18.12 Natural gamma, electromagnetic and temperature data indicating the presence and distribution of acids in three borings

A situation in which borehole deviation may need to be It was then discovered that they were entangled with each
considered is that of determining water levels. If the water other indicating that the boreholes were not parallel as
table is relatively flat and deep (more than 60 m) borehole assumed, but deviated significantly intersecting one another
deviation alone can result in significant errors in the mea- (Hopkins, 2013, personal communication).
surement of water level. For example, water table mea-
surements at a depth of 76–82 m in northwest Puerto Rico
were affected by borehole deviation. Deviation logs were 18.3.7 Reconstructing Geohydrologic and Well
acquired in 16 deep wells that all had elevation errors in Construction Data
the range of 0.25–1.44 m due to deviation. Figure 18.13
shows the borehole deviation in one deeper borehole in There are many sites where existing borings or wells are
Puerto Rico. available. These borings are often installed over time, by dif-
An extreme example of deviation was noted when hole to ferent contractors, using different criteria for construction. In
hole geophysical measurements were being made by these situations, there are a number of reasons to acquire
Tennessee Valley Authority (TVA). The two probes became geophysical logs in these wells (Turner and Benson 1986).
stuck at a depth of about 30 m in two boreholes 10 m apart. For example:
18.3 Various Applications for Geophysical Logs 243

Location of
Borehole at surface

eters
Depth in M
75
0
Ea 12
sti
ng 6
(m 6 Location of borehole
)
)
hing (m at depth approx 3.8 m
12 0 Nort
offset from surface
location

Fig. 18.13 A deviation log showing the distance from vertical a deep borehole was actually drilled

• If there is no geologic log of a borehole or well, or where used to develop a contour map of the semi-confining layer at
there are questions about such data, geophysical logging the site, evaluate the epikarst and identify the presence of
can often provide the necessary geologic information. karst features. For more information see the case study in
Even if a boring log is available, subtle variations in the Part III, Chap. 27.
geology or hydrologic conditions may be of interest and
the geophysical logs may be used to verify the original 18.3.7.2 Reconstructing Geologic Data
geologic logs and provide additional data. Brine water associated with oil exploration wells was col-
• If correlation between new and existing wells needs to be lected in above ground storage tanks in southwest Florida.
established, using a consistent suite of geophysical logs in During operations at one site, brine was spilled on the sur-
all wells can be useful. face creating a significant contaminant plume. A suite of
• If well construction details need to be determined or veri- monitoring wells had been installed over the site, but
fied, such as, casing material, depth of casing, diameter, chemical sampling did not reveal elevated values or a
screen interval and type of screen. clear plume boundary based upon chloride data. The con-
• If an evaluation of existing conditions is required to assess ductive brine contaminants provided an excellent target in
damage or deterioration or other changes in cased or open which to use both surface and downhole electrical mea-
hole conditions geophysical logs can provide the needed surements. An investigation using surface geophysical
data. electromagnetic methods to map the lateral extent of the
plume and downhole electromagnetic induction logging
to map the vertical extent of the plume as well as a thor-
18.3.7.1 Consistent Data Set for Geologic ough review of regional literature and all chemical data
Evaluation was completed.
A Superfund site in west-central Florida, had been in opera- The review of existing literature identified an aquitard,
tion for about 40 years as a phosphate processing plant. They which separated the surficial and deeper aquifers. The review
had numerous existing wells on-site that included piezome- of on-site drilling data showed that the majority of the moni-
ters, monitoring wells, and deeper supply wells. As part of toring wells were shallow and completed well above this
the site characterization, geophysical logging was completed aquitard. The few wells that were drilled below the aquitard
in all existing wells and new borings in order to provide a were sampled at a 3 m interval when drilling and had missed
consistent data set that was used to aid in the evaluation of identifying the aquitard. The deeper wells had their screens
the geology at the site. This data set consisted of natural installed just below the aquitard.
gamma (clay content), gamma-gamma (density), neutron The results of the surface electromagnetic conductivity
(porosity) and induction (conductivity) logs. By logging the data showed a very clear high conductivity plume surround-
existing wells, we took advantage of having these vertical ing the above ground tank. Spatially the wells were located
geologic windows and added 35 spatial data points to sup- within the high conductivity plume and should have defined
port the conceptual model. This information was specifically it based upon chloride data. However, the chloride values
244 18 Geophysical Logging

Conceptual Model
Agricultural Supply Well Monitor Well
with screen both above screened below
Conductivity (mS/m) and below the aquitard the aquitard Conductivity (mS/m)
0 500 1000 0 500 1000
0 0

15 15
Depth in Meters

Highest
Concentration of Brine
30 (based upon 30
induction logs)
aquitard

Deeper Extension of Brine

via fully screened irrigation
45 45
well (note high conductivities
below aquitard)

Fig. 18.14 Electromagnetic induction logs identified brine contaminant distribution

were all relatively low and did not indicate a coherent plume 18.3.7.3 Monitoring Well Audit
of contaminants. Assessment of existing borings, wells and monitor wells can
Induction logging was completed in all monitoring wells provide basic geologic and hydrologic data. This type of
on-site. While in the field we found another well associated assessment can also indicate whether well conditions area
with supplying water for agricultural purposes. This well as designed or even have deteriorated causing any data
was not part of the monitoring system and extended much acquired from them to be invalid (Turner and Benson 1986).
deeper. This well was also logged with the induction logging A landfill located in northern Indiana initially began in 1979
tool. and was in the process of permit expansion in 1990. The
Figure 18.14 shows two of the electromagnetic induction 55 ha site is located in an area of glacial till (22 m thick),
logs from the site that were located near the center of the overlying a sand aquifer. Depth to bedrock is approximately
high conductivity plume. The induction logs indicated very 50 m.
high conductivity (greater than 1,000 mS/m) centered at a As part of the landfill expansion, a comprehensive hydro-
depth of about 30 m just above the aquitard. These very high geologic report was being put together. Much of the data and
conductivity values confirm the presence of the brine and information came from existing site-specific reports, borings
show that the dense brine fluid settled down on top of the and monitoring wells. Five (5) monitoring wells had been
aquitard. installed and sampled quarterly since 1979. The landfill was
The reason the chemical sampling data from the wells did well designed with an engineered liner consisting of on-site
not define the plume is because the shallow wells were glacial till and leachate sumps, which were pumped on a
screened above the plume and the deeper wells were sam- regular basis. No problems had been detected in any of the
pling just below the aquitard, also missing the plume. Once wells.
the vertical extent of the brine was identified, screened inter- A monitoring well audit was completed at the site that
vals and chemical data from the existing monitoring wells included a surface inspection of the well and geophysical
could be put into proper perspective. logging using a natural gamma and induction logs as well as
18.4 Downhole, Crosshole and Tomographic Measurements 245

a b

Axial lens looking down Radial lens looking at the

the axis of well showing sidewall of well confirming
damaged PVC casing screen interval and condition

Fig. 18.15 Downhole video camera logging provides for assessment of damaged casing (a) and the presence of well screen (b)

a video camera. A simple check of the well elevations indi- site-specific geologic conditions had been defined so that
cated a 1.5-m error in one of the five wells. This had a signifi- measurement locations can be selected to meet very specific
cant impact on the original water table map. Natural gamma objectives. While surface geophysical measurements are
and induction logs of existing and new wells indicated the very effective in locating anomalous conditions, they often
presence of two aquifers, not one, based upon the presence of lack resolution or penetration required for deeper foundation
an aquitard. A downhole video camera was used with both design. It is here where downhole and crosshole measure-
axial (Fig. 18.15a) and radial (Fig. 18.15b) lens. The down- ments are necessary to obtain the detailed engineering data
hole camera in the wells showed that the formal construction often needed for foundation design.
diagrams for the wells were completely in error. The wells Figure 18.16 illustrates these types of borehole measure-
had broken casing (Fig. 18.15a) and screened intervals in ments. Measurements of P and S wave velocities using seis-
some wells were created using a few hacksaw cuts. The mic methods can be made between the surface and a
screened intervals were also found to be interconnecting the borehole (downhole measurements – Fig. 18.16a). These
two aquifers. provide in-situ measurements of undisturbed soil and rock
The monitor well audit revealed a number of issues that conditions from about 3 m or so away from a borehole.
made the data from the original monitoring network useless. Measurements can also be made between two or more bore-
All five monitoring wells were replaced with a cluster of two holes (crosshole measurements – Fig. 18.16b). The distance
wells (one in the shallow aquifer and one in the deep aquifer) between boreholes is typically 3–6 m. These measurements
and new water table and potentiometric maps were devel- of P and S wave velocities made at discrete depths within
oped along with a new site-specific hydrogeologic report, the borehole are used to calculate elastic modulus. A variety
which was all submitted to the state. of measurements such as seismic, resistivity and radar can
also be made between two or more boreholes (Fig. 18.16c)
to provide topographic imaging of conditions between
18.4 Downhole, Crosshole boreholes.
and Tomographic Measurements

Besides traditional geophysical logging in a single borehole 18.4.1 Downhole and Uphole Measurements
there are a range of other measurements that can be made
between the ground surface and a borehole (downhole) or Many foundation and soil-structure interaction problems
between boreholes (crosshole). These methods significantly require values of elastic moduli. Determinations of elastic
expand the volume of material that is sampled from the bore- moduli can be made using laboratory tests on cores or in-situ
hole. These measurements would normally be made after measurements of seismic compressional and shear wave
246 18 Geophysical Logging

a
Source
b c

Receiver Source

Raypaths
Receiver
Source

Raypaths
Raypaths

Fig. 18.16 The types of measurements that can be made from the borehole to the surface or between boreholes. (a) Surface to borehole or bore-
hole to surface (downhole measurements). (b) Hole-to-hole or crosshole. (c) Tomography

velocities (P- and S-waves). Measurements from surface to installed both within the paleocollapse as well as in back-
hole, hole-to-hole, or crosshole methods not only provide ground conditions. Surface to borehole seismic measure-
data on in situ conditions but also sample a large volume of ments were completed in a number of test borings across the
material. Values of elastic parameters such as shear modulus, site. Figure 18.7c, d show plots of both the P-wave and the
Young’s modulus and Poisson’s ratio can then be determined S-wave velocity data. It is interesting to note that there was
by the measurement of P- and S-wave velocities, provided very little variation in P-wave velocities across the site
the density of the soil is known or can be reliably estimated regardless of location. However, the shear-wave velocities
(Anderson and Woods 1975; Ballard and McLean 1975; show a slight but clear decrease in velocities within the
Stokoe and Abdel-Razzak 1975; Wair et al. 2012). paleocollapse feature due to the type of infill material. The
Seismic measurements made in a surface to borehole results suggested that the paleocollapse was stable with
configuration are commonly referred to as Vertical Seismic reasonable strength characteristics. For more detailed on this
Profiling (VSP). The velocity data are also used to model case history see Part III, Chap. 27.
and interpret seismic refraction and reflection data. These
measurements have been commonly made using a sledge
hammer and plank at the surface as an energy source. By 18.4.2 Crosshole (Hole to Hole) Measurements
striking a vertical blow (Fig. 18.17a) compressional waves
(P-waves) are developed. By striking the opposite ends of Crosshole or hole-to-hole measurements (Fig. 18.16b) have
the plank (held in place by the wheels of a vehicle) opposing been commonplace for years, making its entrance into the
horizontal shear waves (S-waves) are developed engineering profession in the early 1970s. Hole-to-hole mea-
(Fig. 18.17b). The seismic energy is recorded as a function surements of seismic velocity will generally provide better
of depth by geophones clamped into place in a dry borehole vertical resolution of both P and S-wave measurements, than
or hydrophones placed within a water-filled borehole at var- that obtained from either down or uphole measurements.
ious depth intervals. These measurements will provide seis- Measurements are commonly made between two or three
mic velocities of both P and S-waves versus depth at a boreholes in a line. Measurements are made by placing a
borehole location. See ASTM D7400-14 standard test meth- source in one borehole and receivers in the other borehole or
ods for downhole seismic testing for more information boreholes (Fig. 18.18a) with a spacing of 6–7 m or less
(ASTM 2014b). between boreholes. Measurements are made at some inter-
Hole to surface measurements simply reverses the proce- val moving down the borehole and both P and S-wave
dures with the hammer or source in the borehole and the geo- velocities can be calculated from the each measurement.
phones at the surface. Because the energy source is downhole, The data in Fig. 18.18b was acquired every 1.5 m and shows
there is an increase in frequency of the signal and improved much more detail than that presented in Fig. 18.17c, d.
resolution. Receivers can be geophones in dry boreholes or hydro-
An example of surface to borehole seismic measurements phones in water-filled boreholes. A source that can generate
comes from the karst investigation at the Superfund site on both P and S-waves is used. ASTM D4428 M-14 provides
the west coast of central Florida. The location of a paleocol- guidance for the spacing and construction of boreholes used
lapse feature had already been identified and its stability and for P and S-wave measurements for use in hole-to-hole seis-
strength was part of the assessment. Boreholes were already mic testing (ASTM 2014a). These same concepts are appli-
18.4 Downhole, Crosshole and Tomographic Measurements 247

Fig. 18.17 Downhole seismic velocities were obtained using a hammer to generate compression P-waves (a) and S-waves (b) using a plank on
the surface and geophones in the borehole. P-wave velocity (c) and S-wave velocity (d) are plotted versus depth for nine boreholes

cable to casing installations for other geotechnical hole mode where the transmitter antenna is located in one
measurement applications. borehole and the receiver antenna is located in another hole
Other measurements can be made hole-to-hole including some distance away (Fig. 18.19a). These systems can deter-
borehole ground penetrating radar (Haeni et al. 2002) and mine conditions between two boreholes and provide detection
borehole resistivity. Borehole radar has been used in the reflec- of fractures (Fig. 18.19b), cavities and other discontinuities
tion mode where the transmitter and receiver antennas are such as pipelines and tunnels, which are not intersected by the
located in the same borehole. Radar is also used in the cross- borehole (Singha et al. 2000; Williams et al. 2009).
248 18 Geophysical Logging

Fig. 18.18 Hole-to-hole seismic measurements between two boreholes and resulting velocity data. (a) Hole-to-hole seismic measurements. (b)
Crosshole seismic data

18.4.3 Tomographic (Imaging) Measurements between two boreholes by measuring the seismic travel-
times between a source (in one hole) and a receiver (in the
Tomographic measurements can be used to image conditions other hole) at different combinations of source-receiver
between two or more boreholes (Fig. 18.16c). When mea- depths. In theory, a single source and single receiver can be
surements between two boreholes are made at a number of used to make tomographic measurements by moving the
angles the data can be used to develop detailed 2D cross sec- transmitter and receiver to provide a combination of mea-
tions between two boreholes. Three-dimensional cross sec- surements. However, in practice, a series of receivers in one
tions can be developed by making measurements between borehole are used with a transmitter in another borehole.
three or more boreholes. When the borings are installed This approach saves considerable time.
around the perimeter of an anomalous area, the resulting data Both seismic and radar have been used in crosshole and
can provide an image of anomalous conditions between the tomographic measurements. Corin et al. (1997) describe the
boreholes. Seismic measurements are the most commonly use of radar tomography to assess karst conditions between
used tomographic measurements, however, a number of pairs of piers along a viaduct. The boreholes were 20–40 m
other methods, including radar, EM and resistivity can also deep and the distance between boreholes ranged from 12 to
be used. 20 m. They were able to resolve areas of massive limestone,
For example, seismic crosshole tomography can be used weathered limestone and voids filled with sand based upon
to obtain a 2D cross section of seismic P-wave velocities changes in the radar velocity due to water content.
References 249

Borehole radar being used in angled boreholes extending beneath building

b0

10
Depth in Meters

20

30

0 10 20
Radial Distance in Meters
Example of 60 MHz borehole radar data showing
planar fractures (Singha et al. 2000)

Fig. 18.19 Borehole radar data being acquired between two boreholes (a) and an example of radar data showing fractures (b) (Singha et al. 2000)

ASTM (2010a) Standard guide for planning and conducting borehole

References geophysical logging D5753-05(2010). American Society for Testing
and Materials, West Conshohocken
Anderson DG, Woods RD (1975) Comparison of field and laboratory ASTM (2010b) Standard guide for conducting borehole geophysical
shear moduli. In: ASCE proceedings of the conference on in situ logging-gamma, D6274-10. American Society for Testing and
measurement of soil properties, volume 1 in situ measurement of Materials, West Conshohocken
soil properties. North Carolina State University, June 1975. ASCE, ASTM (2011) Standard guide for conducting borehole geophysical
New York, pp 69–92 logging-mechanical caliper, D6167-11. American Society for
ASTM (2007a) Standard guide for conducting borehole geophysical Testing and Materials, West Conshohocken
logging-electromagnetic induction, D6726-01(2007). American ASTM (2014a) Standard test methods for crosshole seismic testing,
Society for Testing and Materials, West Conshohocken D4428M-14. American Society for Testing and Materials, West
ASTM (2007b) Standard guide for conducting borehole geophysical Conshohocken
logging-Neutron, D6727-01(2007). American Society for Testing ASTM (2014b) Standard test methods for downhole seismic testing,
and Materials, West Conshohocken D7400-14. American Society for Testing and Materials, West
Conshohocken
250 18 Geophysical Logging

Ballard RF Jr, McLean FG (1975) Seismic field methods for in situ ment of soil properties, volume 1 in situ measurement of soil prop-
moduli. In: ASCE proceedings of the conference on in situ measure- erties. North Carolina State University, June 1975. ASCE, New York,
ment of soil properties, volume 1 in situ measurement of soil prop- pp 422–449
erties. North Carolina State University, June 1975. ASCE, New York, Turner MS, Benson RC (1986) Auditing existing monitor wells. In:
pp 121–150 Focus on southeastern ground water issues conference. National
Chapellien D (1992) Well logging in hydrogeology. Balkema, Water Well Association, Westerville, Ohio
Rotterdam Wair BR et al (2012) Guidelines for estimation of shear wave velocity
Colog (2005) Hydrogeophysical™ and geophysical logging results profiles, PEER report 2012/08. Pacific Earthquake Engineering
Marshall Space Flight Center, Huntsville, AL. Consultants report Research Center, Berkley, 68p
June 20 Williams MME et al (2009) Review of borehole based geophysical site
Corin L et al (1997) Radar tomography applied to foundation design in a evaluation tools and techniques, Nuclear Waste Management
karstic environment. In: McCann DM et al (eds) Modern geophysics Organization NWMO TR-2009-25
in engineering geology. Geologic Society engineering geology special Yuhr L et al (2005) An integrated geophysical approach for a karst char-
publication no. 12. The Geological Society, London, pp 167–173 acterization of the Marshall Space Flight Center. In: Beck BF (ed)
Haeni FP et al (2002) Detection and mapping of fractures and cavities Proceedings of the 10th multidisciplinary conference on sinkholes
using borehole radar. In: Proceedings of fractured rock, National and the engineering and environmental impacts of karst. ASCE
Ground Water Association, Denver, CO, p 4 Geo-Institute. Geotechnical Special Publication No. 144,
Hearst JR et al (2000) Well logging for physical properties, a handbook for pp 572–579
geophysicists, geologists and engineers, 2nd edn. Wiley, West Sussex Yuhr L et al (2008a) Mapping and drilling of conduits within the
Killeen PG (ed) (1986) Borehole geophysics for mining and geotechni- Woodville Karst Plain of north Florida. In: Yuhr LB et al (eds)
cal applications. Geological survey of Canada paper 85–27 Proceedings of the 11th multidisciplinary conference, sinkholes and
Michigan Department of Transportation (2008) Detroit river interna- the engineering and environmental impacts of karst. ASCE
tional crossing study. Brine well cavity investigation program tech- Geotechnical Special Publication No. 183, pp 340–350
nical report, February, Detroit, Michigan Yuhr L et al (2008b) Karst characterization of the Marshall Space Flight
Singha K et al (2000) Borehole-radar methods: tools for characteriza- Center: two years later. In: Beck BF (ed) Proceedings of the 11th
tion of fractured rock. USGS Fact Sheet 054-00 May multidisciplinary conference on sinkholes and the engineering and
Stokoe KH, Abdel-razzak KG (1975) Shear moduli of two compacted environmental impacts of karst. ASCE Geo-Institute. Geotechnical
fills. In: ASCE proceedings of the conference on in situ measure- Special Publication No. 183, pp 98–109
 Assessment of Larger Open Voids
and Structures 19

Abstract
On occasion, larger open voids, caves, conduits and structures such as large pipes, tunnels,
shafts and mines will be a part of the site characterization. These features may be dry or
water-filled. While direct visual inspection and photographic or video documentation would
be the preferred approach to these openings or structures, it is not always possible. In addi-
tion, there may be a need to obtain more quantified measurements and data. There are
however, a range of tools that can be used whether access is easily obtained or limited
through a borehole or crevice.

19.1 A Variety of Methods water-filled spaces and can carry a variety of sensors and
tools.
When faced with a situation where a space (natural or
man-made) requires inspection as part of the site character- Table 19.1 summarizes the typical conditions and limita-
ization a unique approach may be necessary. There are a tions in which these methods can be used to assess dimen-
variety of methods that can be employed to image and evalu- sions and conditions of large voids or openings.
ate such conditions, all of which are fairly common and
available. The range of methods for assessment include:
19.2 Visual Inspection
• Direct visual inspection – when conditions permit direct
access safely, visual inspection, supplemented with The most powerful tool we have available is one of observa-
sketches and photographic or video documentation tions. Whenever possible, visual inspections should be made
provides the most comprehensive assessment of such along with sketches, photo or video documentation, but par-
conditions. ticularly when unusual conditions present themselves. Peck
• Photographs and video – can provide conditions of water- (1972) warns that observations made by an intelligent person
filled and air-filled open spaces. are often overlooked as the instrument of choice. Only when
• Cave mapping systems – can provide coordinates and we cannot make direct observations should more specialized
dimensions of air-filled caves as well as mines and other instruments be employed.
open spaces. Visual inspections and observational skills have been
• Laser measurements – can provide size, shape and condi- stressed throughout the site characterization process. Here
tions of large air-filled open spaces. we are addressing more unique physical areas or conditions
• Sonar measurements – can provide size, shape and condi- that require “thinking outside the box”. Figure 19.1 shows
tions in large water-filled open spaces. some of the unique conditions in which direct observations
• Robotic and autonomous vehicles equipped with video, can be made in a borehole, water intake pipe to a nuclear
laser, or sonar – can provide a means to access air-filled or power plant or underwater cave.

© Springer Science+Business Media Dordrecht 2016 251

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_19
252 19 Assessment of Larger Open Voids and Structures

Table 19.1 Conditions and limitations of methods for assessing large voids or openings
Minimum diameter
Method Conditions opening for access Useful distance Limitations Comments
Visual Dry About 60 cm for Typically 15–30 m Size of opening. Safe physical Visual observations
human access or so without access. Limited visibility by supported by
special lighting dust or fog in air sketches, photos, or
video
Water filled (by divers) About 60 cm or 0–10 m Size of opening. Safe physical Visual observations
more for human access. Limited visibility by supported by
access particulate matter in water sketches, photos, or
video
Photos or video Dry 5 cm diameter Up to 30 m Limited by dripping water and Visibility can
documentation for borehole depending upon debris in a borehole, or by dust sometimes be
systems lighting used and fog in large open voids or extended by using
mines lighting in separate
boreholes further
from camera
Water filled 5 cm diameter 0–10 m Limited visibility due to Visibility can
for borehole particulate matter sometimes be
systems improved by using
flocculants
Laser Dry 7–45 cm, From 1 to >50 m Type of surface may limit Laser scanning
depending upon reflections system can provide
equipment 3D models
Sonar Water filled 7–30 cm From 1 to 100 m Sonar image quality is At higher frequency
depending upon function of frequency and short range
equipment images can approach
photographic quality
Robotic vehicle Air filled or water filled Typical diameter Limited by cable Limited visibility without Can be equipped with
(ROV) depending upon vehicle opening for a length (up to sonar. Cable drag and snags. camera, video, sonar,
sewer inspection 300 m) Knowledge of accurate or laser
system 20 cm. A position
small ROV about
40 cm opening
Autonomous Air filled or water filled Most AUV’s will 16–160 km or Limited by obstructions and Can be equipped with
vehicles (AUV) depending upon vehicle require an more battery life camera, video, and or
opening of about sonar
1 m or more

Fig. 19.1 First-hand visual observations under unusual conditions. (a) Large borehole. (b) Intake pipe. (c) Underwater cave (Photo a courtesy of
Dick Woods)
19.2 Visual Inspection 253

19.2.1 Concerns for Deep Foundation Piles ers are necessary to make observations. In the late 1960s
cave divers begin to explore the many sinkholes and caves in
The use of 60 cm bucket auger borings and downhole visual Florida, first as a hobby, then slowly shifting to serious div-
geologic inspection has been widely used in California to ing technology, cave mapping and scientific measurements.
characterize landslide problems. Critical to this approach are Their original equipment was pretty much standard SCUBA
the regulatory and safety issues in a borehole, which are dis- gear using compressed air along with a few homemade
cussed by Scullin (1994). The same concept can be adapted pieces of specialized equipment such as dive lights. As time
to inspection of geologic conditions within vertical mine passed equipment became more advanced including:
shafts, vertical pits into caves, etc. rebreathers, special gas mixtures and scooters enabling them
A new manufacturing facility was being constructed to make longer and deeper penetrations into the cave
which was to contain vibration sensitive equipment. The systems.
design required 2.5 m diameter piers to be founded on While the authors were never directly involved with cave
unweathered shale to provide adequate stiffness. Surface diving, many of their friends were, including Tom Mount,
wave seismic measurements indicated about 6–8 m of clay Steve Cawthon, Rick Freeze and Bob Friedmann. These
with nearly constant shear wave velocity over a thin stratum early cave diving pioneers from the Miami area provided the
of weathered shale over intact shale. The precise depth and authors with maps of many of the caves they explored in the
thickness of the weathered zone was not uniform over the late 1960s. Here Cawthon is making a reconnaissance
site, so at each caisson, the quality of the rock needed to be inspection using an underwater scooter (Fig. 19.1c).
determined. If the shale was not un-weathered or stiff More recently, extensive studies have been completed to
enough, the caisson would be excavated deeper and the rock characterize the Woodville Karst Plain (WKP) in northern
retested. Visual inspection and testing of the rock quality Florida. There are two major springs in the area, the
using a hand-held Housel Penetrometer was seen as the best Wakulla Spring and the Spring Creek system further south
option. Dr. Richard Woods was the vibration consultant to at the edge of the Gulf of Mexico (Kincaid et al. 2012).
the owner and performed both the surface wave measure- Cave divers have made tremendous contributions to the
ments and caisson rock testing. He is seen in the photo descriptions and connections of the underwater network of
descending into a caisson for inspection of rock quality at the caves within the WKP. Cave divers began exploring the
bottom of the borehole (Fig. 19.1a). sinkholes and cave systems in the Leon Sinks area and
Wakulla area in 1970s. After years of exploration it was
thought that the two areas were connected. That connection
19.2.2 Power Plant Ocean Water Intake System became the focus of two divers and their support team. The
connection was finally made between Turner Sink (the
Sometimes site characterization is not a geotechnical issue. southern most end of the Leon sinks area) and Wakulla
The 2.5 m diameter ocean water intake line at the St Lucie Springs by Jarrod Jablonski and Casey McKinlay in July
Florida nuclear power plant was being restricted by rapid 2007. It was a 6.5 h dive followed by 14 h of decompres-
barnacle growth. The ocean intake line was inspected using sion. Their support team consisted of approximately 50
divers and photographs to monitor the rate of growth on dif- members.
ferent antifouling materials. The intake canal was inspected This is the longest underwater cave system in the United
using an echo sounder and bottom sediment sampling, and States (a total of 45 km of mapped surveyed passages) and
the power plant intake was visually inspected and photo- the fourth largest in the world. The depth ranged from 15 to
graphed including the grizzly intakes and piping up to the 85 m, the five largest mapped caves range from >2 km to
condenser (Fig. 19.1b) to evaluate both biofouling and >23 km in length and conduit diameters ranged from less
corrosion. than 2 m to greater than 30 m averaging approximately
10–15 m. The diver’s amazing story is told by Wisenbaker
(2006a, b), Wisenbaker and McKinlay (2006) and
19.2.3 Mapping by Cavers and Cave Divers Wisenbaker et al. (2007).

Cavers have provided much of the data on cave locations and

maps. Most caves will have an opening to access the cave, 19.2.4 Tarpon Springs Bridge Failure
however, one caver has a more aggressive approach. John
Ackerman has excavated sinkholes to uncover the entrance Several sections of the northbound US19 bridge over the
to a caves (Faulkner 2008). When caves are underwater Anclote River in west central Florida collapsed in the late
direct observations becomes more complicated and cave div- afternoon in 1968. One person was killed and five were
254 19 Assessment of Larger Open Voids and Structures

Fig. 19.2 A geologist prepares to make a dive to inspect conditions of a bridge collapse (Photo courtesy of Paul Beam)

injured in the incident. Rescue divers at the site reported that This is another case where curiosity and observations pro-
the bridge pilings of the collapsed bridge sections went vided the answer to a problem. The combination of extensive
straight down into the river bottom. Officials from the Florida drilling data under the supervision of a professional geolo-
Department of Transportation (FDOT) stated that the cause gist as well as first hand underwater observations by the
of collapse was believed to be a large sinkhole, which are same geologist provided a solid basis for the interpretation of
common in the area. The bridge was built in 1946 and no the cause of the bridge failure.
structural problems had been previously noted.
Paul Beam, a young geologist, from the FDOT Research
and Testing Laboratory supervised the investigation of the 19.2.5 Road Widening Adjacent Sinkhole
collapse. Three drill rigs (one on a barge) were used to sam-
ple subsurface conditions. After 4 weeks of subsurface core Another example of observations is taken from a water-filled
sampling by the three drill rigs, Beam reported that there sinkhole only 30 m away from US19 in the town of Hudson
was no evidence of sinkhole activity. Being professionally in west central Florida near the Gulf of Mexico. In 1980,
curious, Beam used SCUBA gear and an underwater cam- plans to widen the highway had created concern about the
era to further examine the area (Fig. 19.2). The bridge pil- stability of the road widening near the edge of the sinkhole.
ings were steel H-beams with a concrete jacket around An investigation of the sinkhole was made by divers. They
them to protect them from sea water corrosion. However, measured the depth and shape of the sinkhole using tape and
tidal action and river flow had scoured the river bottom compass measurements (Figs. 19.3 and 19.4). They also
beneath the concrete jacket allowing corrosion to occur on obtained vertical profile measurements of conductivity and
the now exposed H-beam piles. Mr. Beam discovered soft- temperature and photographed key geologic conditions
ball-sized holes in some of the pilings of the still standing within the sinkhole. Ground penetrating radar measurements
sections of the bridge. Based upon drilling data indicating were made on the surface parallel to both sides of US19.
no sinkholes and his personal observations he concluded The Suwannee Limestone formation outcrops at the sur-
that the bridge collapse was not caused by sinkhole activity face and is approximately 45 m thick. The Suwannee
but by corrosion of the steel piles (Beam, 1984, personal Limestone is underlain unconformably by the Crystal River
communication). formation. Much of the cave development in the area appears
19.2 Visual Inspection 255

Fig. 19.3 A diver enters water-filled sinkhole to map, photograph conditions and measure conductivity/temperature (a). Note block fall photo-
graphed at a depth of 42 m (b)

a Plan View b Cross Section

Southbound Lanes US 19 A A’ Surface 0

A Chimney
C 12

Depth in Meters
Surface Fresh Water
(38 m dia.) F1 24
Saltwater

Chimney F1 Debris Pile 36

C
(18 m dia.)
F3
F2 F3 & F4
F4 48
A’ c Degrees Celsius Millimhos/cm (X1000)
Bottom 0
16 20 24 28 0 1 2 3 4 5
Water Depth in Meters

12
N

F1 & F2: Vertical Fractures in Wall

0 15 F3 & F4: Bottom Fissures
24
Approx. Scale in Meters C: Small “cave”

36
Temperature Conductivity

Temperature and Conductivity Profiles

Fig. 19.4 Plan view (a) and cross section (b) of mapped sinkhole along with temperature and conductivity profiles (c) measured within the
sinkhole
256 19 Assessment of Larger Open Voids and Structures

to have developed near this contact and sinkholes are com- During the monitoring of collapse conditions in a
mon in the area. Cave divers have noted both fracture and limestone mine (1989–1993), which was partially flooded,
bedding control within many of the local sinkholes (Freeman, a Nikonos V underwater camera using ASA 400 film and a
1980, personal communication). SunPak 355AF flash were used to photograph collapse con-
The sinkhole has a classic bell shape typical of many ditions (see Figs. 25.18 through 25.21 in Part III). These pho-
sinkholes in Florida (Fig. 19.4b). It is 38 m in diameter at the tos provided adequate documentation of individual collapse
surface, about 18 m in diameter at the chimney. The major areas where distances between pillars was approximately
conduit from which the sinkhole developed lies at a depth of 18 m. Details beyond a distance of about 15–22 m were not
about 45 m and is oriented in the NE to SW direction coinci- clearly illuminated due to limited lighting from the single
dent with the reported fracture patterns in the area. A 1 m flash system. However, the use of a million candle power
wide fissure, 5 m high was seen in the southwest wall of the 12 V handheld spotlight provided localized viewing to more
sink at a depth of 22–27 m. An undercut of the wall and a than 30 m, well beyond the range of the flash system, but the
fallen block was noted at the bottom of the sinkhole on the field of view is limited by the width of the narrow spotlight
southwest side of the sink (Fig. 19.3b). No recent signs of beam (Benson and Hatheway 1994). Photographic documen-
collapse activity were noted by the two divers, one of whom tation was effectively used to document both stable collapse
was an experienced cave diver, the other a geologist. areas as well as new collapse areas as they occurred. The
A ground penetrating radar survey was carried out description of conditions to site owners and regulators was
between the sinkhole and US19. There were no anomalous easily conveyed using this method.
conditions seen in the radar data. However, further east of Inexpensive chemical light sticks (cyalume) were used as
US19, about 1.2 km, sinkholes were detected by radar and an effective means to mark the boundaries of areas of interest
confirmed by drilling. prior to taking photographs. They were also effective as
Vertical profile measurements of conductivity and tem- safety markers to mark our trail so that we could quickly ori-
perature were made. This data indicates that the fresh-salt ent ourselves and as a safety means to find our way out of a
water interface is at a depth of about 21 m below grade mine or cave should all lighting fail.
(Fig. 19.4c). The fresh-salt water interface (the zone of dis- Many cavers have obtained extraordinarily well-
solution) now occurs well above the original depth of the illuminated photos of large caverns over distances of much
cave, which was developed at a time of lower sea level. greater than 30 m (Fig. 9.1). This is accomplished by the use
While the eastern toe of the sink extends almost to the of multiple lighting locations, special lighting sources, tri-
edge of the southbound lane of US19 (Fig. 19.4a), there is pods and long exposures. This is something not normally
more than 30 m of massive limestone above the easternmost available for on-site inspections and would require an expe-
extent of the sinkhole. There was also no indication of sink- rienced photographic team. For more details on how this was
hole activity in the radar data between the sinkhole and high- accomplished see Shooting the Big Room of Camp’s Gulf
way. The conclusions were that this paleosinkhole appears to (Anderson 2000).
be dormant and the risk of collapse is virtually zero, if no Many of the cavities or caves requiring assessment are not
other factors impact this area. accessible and must be accessed using equipment through
boreholes, small crevices or collapse areas.
Borehole video camera systems are typically intended for
19.3 Photographic and Video viewing borehole conditions within a few centimeters to a
Documentation meter or so from the borehole. The authors have used a
downhole video camera to assess conditions within dry
Both cameras and video systems provide one of the simplest mines and have found that effective viewing distances range
and most available means of documenting conditions. A dis- from 15 to 20 m using 500 to 2,000 W of omni directional
tinct advantage of the digital systems is that the results can lighting from the same hole as the video camera (Fig. 19.5a).
be viewed immediately so that one can retake the photo, if Where additional boreholes are available, lighting can be
needed. Some systems are water-proof or at least splash lowered into boreholes further from the video camera to
resistant which is necessary in certain environments. increase the distance of viewing (Fig. 19.5b).
There are inherent distance limitations when using cam- Besides those video camera systems that are used in bore-
eras or video to determine conditions within a large, dark holes, there are a wide range of other off-the-shelf video sys-
cave or mine. In an air-filled system, limitations include dirt tems available which can be adapted to unique conditions for
and dripping water on the lenses, as well as dust and fog, and inspection and documentation. One system referred to as a
the inherent distance of the lighting system. In a water-filled snake, and is used by plumbers to inspect sewer lines. The
space visibility may be as little as a 30 cm or so due to par- video camera is located at the end of a long flexible rod,
ticulate matter. In clear water visibility can be 3 m or more. which can be used for inspection in small diameter pipes, as
19.3 Photographic and Video Documentation 257

a b Approximate
Limit of Images

Remote borehole
with lighting

Mine Pillars

Mine Pillars

Approximate
Limit of Images

20 m 20 m

Borehole with camera Borehole with camera

and lighting and lighting

Setup using a single borehole Set up using additional

for both the camera and lighting boreholes for remote lighting

Fig. 19.5 The range of visibility for borehole video camera and lighting in the same borehole is limited to about 23 m (a). The distance can be
extended considerably by locating additional lighting in a boring further from the camera system (b)

small as 3 cm in diameter. These can also be used to probe and a few radar survey lines indicated a void from 3 to 4 m
into small openings within a cave or mine. Another system below grade in the area of the collapse. A borehole video
mounts the video camera at the end of a ridged rod, which camera was lowered down two of the boreholes, which had
can be used to probe into inaccessible or dangerous areas. encountered a void and revealed the possibility of a small
These systems are used for inspecting shallow confined areas cave system extending away from the house. A local resident
and sewer lines and have been used to aid in search and res- reported that that he and his friends used to play in this cave
cue missions. when they were kids many years ago. The entrance was in
Sewer inspection video systems are in common use by the neighbor’s yard but had been filled with debris. The cave
most municipalities. These systems are typically available as had gone unnoticed by the builder and the problem did not
truck mounted systems but are also configured as portable sys- appear until a downspout at one corner of the house had
tems. The video cameras are mounted on wheels or tracked focused enough run-off to erode sufficient material into the
platforms, which are steerable systems to provide directional cave so that a local collapse occurred (Fig. 19.6a).
control of the platform for inspection of sewer lines and tun- A light was lowered into one of the boreholes and it
nels. The cameras commonly have pan and tilt capabilities and exposed the larger portion of the cave (Fig. 19.6b) in which
offer better lighting for viewing longer distances than most kids had played. Then a video camera was then attached to
borehole video cameras. While intended for inspection in hor- the end of a long pole and from within the collapse area adja-
izontal sewer lines, the authors have lowered such systems cent to the house (Fig. 19.6a) it was used to probe into areas
down a borehole or shaft to inspect caves, tunnels, and mines. that were physically inaccessible. The void extended under a
A small collapse had occurred at the corner of a residence portion of the house 0.3–1 m in height (estimated at about
located in southeast Florida (Fig. 19.6a). The house was 45 m3) and under the swimming pool, but also expanded into
located on the coastal ridge facing the Gulf Stream. Six shal- a larger cave system away from the house where kids had
low borings and a ground penetrating radar survey were played. This larger area was estimated to be about 3 m wide
completed at the site to evaluate the extent of any voids and 1–2 m in height and its volume was estimated to be
around the building and over the property. Two of the borings (approximately 141 m3). A grouting program was then
258 19 Assessment of Larger Open Voids and Structures

Fig. 19.6 The video camera was attached to the end of a long pole in order to provide visual inspection of inaccessible areas (a). A light lowered
into one of the boreholes provided a partial illumination of the larger portion of the cave (b)

designed to stabilize the portion of the void space under the been installed along with a large pump about 1 m tall, which
house and the larger cave under the swimming pool. was painted red. Years later we were required to relocate this
Although the boring and GPR data provided initial insight specific area of the mine from the surface, to verify its loca-
to the problem, the details only became clear as we acquired tion for backfilling. Given the mine maps and surveyors
direct observations, photos and video data which enabled us coordinate system we were able to locate the area on the sur-
to piece together bits of data into a final sketch of the cave face. Once a borehole had been drilled into the open mine
along with estimates of its dimensions. space, a standard sewer inspection camera was lowered
down the borehole to verify our location. Observing the
pump with the video camera would provide confirmation
19.3.1 Its Not Always Easy that we were in the right location. It took about 4 h of tedious
field work (tweaking with horizontal scanning, along with
Detailed maps of a mine in the Kansas City area along with the pan and tilt system and adjusting video system variables)
extensive visual and photo documentation had been acquired. to obtain a fuzzy image of the pump which was approxi-
At one location near the mine portal entrance a sump had mately 15 m from the borehole. Problems with dripping
19.4 Cave Mapping Systems 259

water and small particles and a slight fog as well as compass and tape and to provide a location on the surface to
insufficient lighting prevented a clear image of the pump. drill into the cave system for sampling and monitoring.
Only by repeated panning of the area with a variety of optical A cave radio consists of a transmitter and a receiver coil.
changes were we finally able to identify the pump and verify The coils are less than a meter in diameter. The system oper-
its location with a vague video image. ates at very low frequency in the kilohertz range. The trans-
mitter coil is leveled on the cave floor so that the center axis
of the coil is vertical. Then the receiver coil is moved over
19.4 Cave Mapping Systems the ground surface until an electromagnetic null is achieved.
By establishing a null at two or more locations on the sur-
Cavers and cave divers have made great contributions to face, the intersections of the lines will be directly over the
locating and mapping both dry and underwater cave sys- transmitter coil in the cave (Fig. 19.7a). By measuring the tilt
tems. Cave maps have been traditionally developed using angle and off-set distance from the location, the depth of the
tape, compass and clinometer. The care taken by cavers has cave can be estimated (Fig. 19.7b). Typical location accuracy
generally produced reasonably accurate maps. Digital of better than 0.5 m is obtained under ideal conditions (away
advancements include laser distance measuring devices
instead of the tape measure along with digital compass and
clinometer. Data is now entered directly to a hand-held com- a
puter instead of written in a notebook. Special cave survey
software and data management software including GIS is The receiver coil
now available and being used. This equipment and software moved about on
results in faster measurements, accurate recording and pro- the surface to
locate a null
cessing providing nearly real-time digital cave maps to the
field team similar technology can be utilized in mapping
abandoned mines.
Electromagnetic
field produced
by transmitter
19.4.1 Inertial Navigation Systems (INS)
coil

GPS is now the choice for most positioning in open areas

not restricted by tree cover or tall buildings. There are alter-
The transmitter coil
natives for use inside of structures and underground. An must be leveled on
Inertial Navigation Systems (INS) utilizes the data from the cave floor
accelerometers and gyros feed into a computer system. Determining the location
After establishing known control points to reference the sys-
tem, one can walk about with the INS in a back pack to map b
3D space and the INS will keep track of your position rela- The depth can be
tive to the control points. By stopping at each survey point estimated by the
for a few seconds the position of that point is recorded. At tilt angle of the
receiver coil and its
the end of the survey the data is referenced to the control distance from the
points. A professor from Missouri State University has used center null point
Inertial Navigation Systems (INS) to map caves (Vickery
2010).

19.4.2 Cave Radio

If a cave is accessible by cavers or cave divers then a “cave

radio” (more correctly a “cave location” system) can be
used. A cave radio can be used to locate and determine an
Determining the depth
approximate depth of a point within the cave by measure-
ments on the surface. Radio location is used to provide a Fig. 19.7 The cave radio location system set up to provide a single
means to verify the locations of long traverses made with location (a) as well as to estimate the depth of the cave (b)
260 19 Assessment of Larger Open Voids and Structures

from buildings, metal fences, and power lines). Measurements 19.5.1 Lasers
to a few hundred meters depth can be made and measure-
ments have been made in both air caves and under water There a few different type of laser distance measurements. A
caves. simple single laser distance measurements can provide the
Under optimum conditions it should be possible to deter- distance to an object. Small hand-held laser range finders
mine depth within 3 %. As with most technology, one should have working ranges from 1 to a few 1,000 m with accura-
seek out those with field experience, who have mastered the cies of ±1 mm to 1 m. A pair of laser beams from different
skill of radio location. Further details on cave radios and boreholes can be used to triangulate location. Collimated
communication can be obtained from the US based NSS sec- laser beams can be projected onto an object or fracture so
tion for cave electronics (caves.org/section/commelect) or that an accurate dimension of a feature can be obtained.
from the UK based cave radio and electronics group (CREG) More sophisticated laser scanning systems can be used to
(www.bcra.org.uk/creg/). provide a 3D digital image of the shape of caves or mines.
A multi-electrode 2D resistivity survey had identified a
significant anomaly in an area where new sinkholes had
19.5 Laser and Sonar Systems occurred each year. One borehole was drilled into the
approximate center of a suspected cavity based upon a 2D
It is often difficult to accurately determine the distance and resistivity survey made at the surface. A 2 m high void was
size of an object such as the wall of a cave, a mine wall or encountered below approximately 7 m of bedrock. The cave
pillars, or the aperature of fissures from photos or video interior was roughly estimated based upon video observa-
alone. Accurately determining distances and shape can be tions from the one borehole. Then four other boreholes were
accomplished by the use of lasers (in dry environments) or drilled into the cave around its perimeter based upon the 2D
by sonar (in water-filled environments). They can be used to resistivity survey and the results of a downhole video camera
map and define the size and shape of larger open spaces such in the first hole.
as caves, mines, tunnels and shafts. However, some means of A simple off-the-shelf laser distance measurement device
access is required such as a shaft or a borehole. Figure 19.8 was modified and attached to the end of a ridged rod, which
illustrated how a laser or sonar device can be used to make was lowered down a borehole and was supported by a tripod.
such detailed measurements. The laser was placed at a given depth and could be rotated

Rigid Rod
Opening into
Void Space

Transducer

Raise and lower to scan

to scan at different depths

Fig. 19.8 A laser or sonar system can be used to develop a 360° scan of a void space at a given depth. Repeating this process at different depths
will develop a complete 3D image
19.6 Remotely Operated and Autonomous Vehicles for Inspection 261

360°. The rotation angle was determined at the surface by provide longer ranges of 600 m with lower resolution. Higher
dial plate on top of the tripod. By making measurements frequencies of 1 MHz have less range of 50 m but result in an
around 360° the perimeter of the cave could be accurately image with much greater resolution.
determined. The lasers were then lowered a small distance Sonar is commonly used as an echosounder on a boat,
(about 0.5 m) and the 360° scan is repeated. The process was which is used to determine the depth of water as well as the
repeated through the entire depth of the cavity. The approxi- presence of fish. Simple adaptations have been made using
mate shape of the cavity could then be reconstructed from the small sonar transducer from a simple fisherman’s echo-
this set of data. The cave was found to be 2 m high and sounder or fish finder mounted on the end of a pole and low-
approximately 4 m long and 1 and 0.5 m wide (Roth et al. ered down a borehole or shaft to develop a vertical profile of
2004). a void. An alternative is to mount the transducer horizontally
This is another example of some fairly simple measure- and rotate 360° to produce a cross section of the opening or
ments with off-the-shelf hardware being applied to define the void. The sensor is then lowered some distance and the 360°
size and shape of a cave with a reasonable level of detail and scan is repeated. The process can be repeated through the
accuracy. depth of the opening to provide a number of 2D cross section
The use of 3D mapping by LiDAR (light detection and of the opening or void (Fig. 19.8).
ranging) is now commonly used to create detailed cloud Special systems have the motorized capability to provide
maps of buildings and complex structures. The use of 3D 360° scanning in a horizontal plain by electronic control
LiDAR and digital photography was used to map the interior from the surface. Travel time of the acoustic signal is mea-
of Devil’s Sinkhole in Texas. The combination of the two sured and converted to distance to create an image of the area
methods provides a 3D photo realistic model with very high being scanned. A variation of the standard scanning sonar is
resolution (Neubert et al. 2008). a pan-and-tilt version. In addition to the standard 360° hori-
A laser system called the Dry Ferret was developed by zontal scan, the sonar head also tilts to provide a profile from
Workhorse Technology, (a spin off from Carnegie Mellon horizontal to vertical.
University). The system utilizes a laser device to produce Scanning sonar is commonly used in the ocean industry
three dimensional views of dry underground spaces and can for inspection of oil rigs and pipelines and hazards such as
be deployed in a 20 cm borehole. The system is gyro and wrecks. The oil and gas industry has used scanning sonar to
compass stabilized and also includes a low-light color video determine the size of large cavities around a borehole. The
system (www.workhorsetech.com). navy and research organizations make extensive use of sonar.
The device is lowered down a borehole and scans 360° The Wet Ferret, a scanning sonar, was also developed by
then is lowered some distance and makes another scan. This Workhorse Technology. The system utilizes sonar to produce
process is continued in order to acquire data from the ceiling two-dimensional views of water-filled underground spaces
to the floor of the mine, cave or other open area. The system and can be deployed in a 15 cm borehole. The system is gyro
also has pan and tilt capabilities, which can image an area and compass stabilized and also includes a low-light color
from one position. This extensive set of point cloud data is video system. The data is acquired digitally and software is
then processed to create the 3D Mesh Model of the shape and available to develop 2D or 3D views with a resolution of
boundaries of the void space. 10 mm at distances up to 100 m (www.workhorsetech.com).
This system was used to aid in assessing open areas Figure 19.9 is a single 360° sonar scan of an abandoned
remaining during a mine backfilling project. Once the back- mine showing the location and spacing of pillars.
fill limited access into the mine, the only way to confirm the
extent of fill, was to drill boreholes and view conditions
using a downhole video camera. The Dry Ferret laser system 19.6 Remotely Operated and Autonomous
was also used to evaluate these areas. Vehicles for Inspection

There are a number of remotely operated and autonomous

19.5.2 Sonar vehicles that can be used for inspection and mapping in both
dry and water filled environments. These systems can be
Extending the distance of measurements beyond that of cam- used as platforms for a variety of instrumentation.
eras and video in water-filled environments can be accom-
plished by the use of sonar. Depending upon the application,
sonar can provide data from a few tens of meters to a few 19.6.1 Remotely Operated Vehicles (ROVs)
hundred meters or more with resolution from a few centime-
ters to a few meters. The frequency determines the range and There are a number of small and inexpensive mini-ROV sys-
resolution of the sonar system. A lower frequency of 100 KHz tems available, which are referred to as observational class
262 19 Assessment of Larger Open Voids and Structures

Fig. 19.9 A single sonar scan made with the Wet Ferret in a water-filled mine. The sonar probe is located between two pillars. Line of sight reflec-
tions are occurring from pillars up to 67 m away. Pillar spacing is known to be about 20 m

Fig. 19.10 Two small mini-ROVs, one dry mini-ROV was used in a mine to inspect an area of active roof-collapse (a) and the other was used to
inspect conditions in a water-filled sinkhole (b)

ROV systems. Figure 19.10 shows two such systems used by robotic arm and the ability to retrieve samples. They include
the authors, one is a land based system (Fig. 19.10a) and one intuitive controls along with information displays and auto
is underwater system (Fig. 19.10b). Many of the features functions. These small ROV’s can be deployed from a small
found on the larger commercial ROV systems used in the boat by hand without major rigging or large winches. They
offshore energy business have been adapted to these mini- are used to inspect pipelines and structures in power plants
ROV systems such as powerful thrusters and maneuverabil- and dams and to enter and inspect remote, inaccessible, and
ity, high quality video and a large field of view, as well as a dangerous areas in both wet and dry conditions. They are
References 263

also being used in security, and search and recovery 19.6.2 Autonomous Vehicles
operations. One marine ROV manufacturer has added the
capabilities for crawler mode by mounting the ROV on a Untethered autonomous underwater vehicles (AUV’s) have
four-wheel drive platform. become more common in the oceanographic industry over
Besides the common problem of visibility, particularly the past decade (Merrill 2007). Autonomous vehicles are
underwater, knowing where you are is one of the major prob- now routinely used in oceanographic surveys and have also
lems in using a small tethered robotic vehicle. Most ROV’s been used for inspection of flooded tunnels. An AUV
are equipped with depth sensors, compass and a number of equipped with video cameras was used to inspect a 24 km
turns indicator. However, after changing direction a number water supply tunnel 2.25 m in diameter for damage. The con-
of times it can become difficult to keep track of your position crete tunnel, in Germany, had not been inspected visually in
within a large sinkhole, cave or mine. In some cases, acoustic more than 40 years (Kalwa 2012).
pingers can be used to add direction, range or position data Workhorse Technology has also developed an AUV called
by triangulation. For further details on small observational the Cave Crawler that provides a means of inspecting and
ROV’s and their use see Christ and Wernli (2007). documenting conditions in abandoned mines and areas to
dangerous for human entry. The Cave Crawler can be oper-
19.6.1.1 Using an ROV to Inspect Conditions ated from a joystick control or may be completely autono-
in a Flooded Mine mous within air-filled voids with walk-in access. The vehicle
In 2006, a uranium ore mining operation in Canada suffered weighs 180 kg and is about the size of a small “Bobcat” con-
a roof collapse approximately 450 m below grade resulting struction vehicle. Its positional accuracy is 5 cm and it is
in flooding of the mine at rates of 15,000–20,000 gal per equipped with an inertial positioning system with gyro and
minute. By the time flooding had stabilized, water had com- compass. Other mapping sensors may include a 3D laser,
pletely flooded all of the mine and water level had risen to low-light video, high resolution still camera, radar and ther-
within 17 m below grade. mal image system depending upon mission objectives (www.
Snyder and Cook (2008) summarize the approach used in workhorsetech.com).
the inspection of this flooded mine. They cite the operation Deep Phreatic Thermal Explorer (DEPTHX) is an
of an ROV system within the low visibility and complex unmanned autonomous robotic vehicle that was a NASA
mine tunnels with unexpected obstructions to provide an funded effort using an autonomous vehicle to explore deep
inspection of conditions prior to remediation. Access was cenotes in the state of Tamaulipas, Mexico. The project was
only available from the vertical shaft where the ROV was a 4-year effort to map and obtain chemical data along with
lowered to the 450 m level to explore the lateral tunnel and to water and rock samples from four large cenotes.
inspect conditions of a jammed high-pressure bulkhead door The DEPTHX vehicle is approximately 1.5 m in height
that had failed to close. by 1.9 m in length and width with a dry weight of 1,500 kg.
Accurate maps of the mine were available and a 3D map The vehicle can move at speed of about 0.2 m/s. It is equipped
and model of the mine along with known conditions such as with sonar, depth sensors, and a navigation system for map-
audits, location of heavy equipment and utilities were incor- ping, as well as multi-parameter water quality measure-
porated into the model. As a result there was a good deal of ments, water samplers rock samplers video and still cameras.
data and maps available prior to launching the ROV. The The final test of DEPTHX was carried out in Zacaton, the
ROV was equipped with a variety of high resolution sonar deepest underwater vertical shaft and the second deepest
systems along with pan and tilt cameras and lighting. The underwater cave in the world (319 m) (Gary et al. 2008).
issue of real time navigation required an inertial navigation
system supported by a doppler-velocity log, a gyroscope,
and accelerometers in three dimensions. High resolution References
sonar on the ROV provided avoidance of unknown obstacles
(Snyder and Cook 2008). Anderson C (2000) Shooting the big room of Camps Gulf. NSS News
58(7):201
The combination of an accurate 3D map and model of the
Benson RC, Hatheway AW (1994) Underground lighting. NSS News
mine along with the sophisticated navigation system enabled 52(7):214–215
the ROV operator at the surface to maneuver through the Christ RD, Wernli RL Sr (2007) The ROV manual: a user guide for
mine passages. It is interesting to note that this work over- observation-class remotely operated vehicles. Elsevier, New York
Faulkner S (2008) In the media. NSS News 66(5):18–20
came limitations and was carried out in this very complex
Gary M et al (2008) 3D mapping and characterization of Sistema
environment by incorporating off-the-shelf hardware and Zacaton from DEPTHX (Deep Phreatic Thermal eXplorer). In: Yuhr
software including the ROV. L et al (eds) Proceedings of the 11th multidisciplinary conference on
264 19 Assessment of Larger Open Voids and Structures

sinkholes and the engineering and environmental impacts of karst, on the Application of Geophysics to Engineering and Environmental
ASCE Geo-Institute, Geotechnical Special Publication No. 183 Problems, Environmental and Engineering Geophysical Society,
Kalwa J (2012) Sea cat AUV inspects water supply tunnel. Sea Technol pp 1307–1314
53:43–44 Scullin CM (1994) Subsurface exploration using bucket auger borings
Kincaid T et al (2012) Demonstrating interconnection between a wastewa- and down-hole geologic inspection. Bull Assoc Eng Geol
ter application facility and a first magnitude spring in a karstic water- 31(1):91–105
shed: tracer study of the southeast Farm Wastewater Reuse Facility, Snyder JZ, Cook M (2008) ROV ops in low visibility, 3-D models and
Tallahassee, Florida. Florida Department of Environmental Protection, inertial navigation. Marine Technology Reporter, May,
Florida Geological Survey Report of Investigation No. 111 51(4):24–31
Merrill ML (2007) Unmanned untethered submersible technology. Vickery K (2010) Inertial navigation systems (INS) not just for land
Marine Technology Reporter, October, pp 38–41 anymore. Professional Surveyor Magazine, May, pp 43–45
Neubert B et al (2008) 3D photo real modeling of Devil’s Sinkhole- Wisenbaker M (2006a) A history of exploration in America’s longest
Rocksprings, Texas. In: Yuhr L et al (eds) Proceedings of the 11th underwater cave, Part 1. NSS News, June, pp 4–7, 30
multidisciplinary conference on sinkholes and the engineering and Wisenbaker M (2006b) A history of exploration in America’s longest
environmental impacts of karst, ASCE Geo-Institute, Geotechnical underwater cave, Part 2. NSS News, July, pp 4–7, 30
Special Publication No. 183 Wisenbaker M, McKinlay C (2006) A history of exploration in
Peck RB (1972) Observation and instrumentation: some elementary con- America’s longest underwater cave, Part 3. NSS News, August,
siderations. US Dep Transp Fed Highw Adm Highw Focus 4(2):1–5 p 13
Roth et al (2004) Measuring cave dimension remotely using laser Wisenbaker M et al (2007) Explores connect America’s longest under-
pointers and a downhole camera. In: Proceedings of the symposium water cave to Wakulla Springs. NSS News 65(12):2–11
 Engineering Measurements
and Monitoring
20

Abstract
At this stage of the site characterization some geotechnical data has already been obtain
which may include soil and rock samples, STP blow counts, and RQD values. Geophysical
logging data has provided additional geotechnical details from borings. Some laboratory
analysis may have been made on selected samples from borings. With a refined conceptual
model of conditions in hand, the engineering measurements and monitoring is now focused
upon obtaining very site-specific measurements and data to address specific engineering
questions and quantify additional engineering parameters. This is a broad range of measure-
ment types and data. Within this chapter, we will focus specifically on the engineering
measurements and monitoring of subsidence and collapse. We summarize some of the common
methods for measuring and monitoring land subsidence and potential collapse, cover the
general applications, some considerations and provide a few examples. As always, each
technique or measurement has an associated scale of measurement, advantages and limitations.
While the many details for making these geotechnical measurements are beyond the
scope of this book, further details can be found in Dunicliff (1988), Hunt (2005) and by
specialty conferences (DiMaggio and Osborn 2007) who all provide a solid foundation for
these types of measurements. Due to the advances that continue to occur in geotechnical
instrumentation this information should be supplemented with recent manufacturers litera-
ture and recommendations.

20.1 In-Situ Geotechnical Measurements ers to their book for the many details of geotechnical instru-
and Monitoring mentation, measurements and monitoring which are beyond
the scope of this book. The topics covered in their book include
In their book, Geotechnical Instrumentation for Monitoring measurement of groundwater pressure, total stress in soil,
Field Performance Dunnicliff (1988), with the assistance of stress changes in rock, deformation, load and strain in struc-
Gordon Green, provide an excellent coverage of the topics of tural members and temperature. They also emphasize the
measurement and monitoring soil and rock conditions for many pitfalls in acquiring such data and refer to the “25 links
engineering projects. While their book is more than three in the chain” required to achieve reliable results.
decades old it remains one of the best references for such mea- Selection of measurement type and instrumentation requires
surements and provides numerous guidelines for achieving consideration of scale as well as sensitivity. Some of the mea-
quality results. The details of the instrumentation will have surements can be quite local while others are regional in nature.
changed due to rapidly changing technology but the basic Figure 20.1 and Table 20.1 summarize the range of these type
principles of measurement remain the same. Both Dunnicliff of measurements. In addition, sensitivity of the instrumentation
and Green are practicing experts in the field and have also needs to fit the expected results, which may require some guess
conducted excellent short courses on the topics. We refer read- work. Instrumentation for such measurements may include:

© Springer Science+Business Media Dordrecht 2016 265

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_20
266 20 Engineering Measurements and Monitoring

b
Survey for location
and elevation c
Measure of
Rotation or Tilt

d
e

Fig. 20.1 Some of the many measurements used to measure and moni- (c) Tilt meters or Inclinometers for monitoring rotation or movement on
tor subsidence. (a) LiDAR (from airplane) or InSAR (from satellite) for multiple planes. (d) TDR or OTDR for monitoring horizontal or vertical
mapping surface deformation over large areas. (b) Benchmarks for movement. (e) Strain gage type of measurements, very localized hori-
monitoring horizontal or vertical movements by optical survey or GPS. zontal or vertical movement

• Level survey (by first order land survey procedures) • Tilt meters (used to monitor changes in tilt of a
• Piezometers (a variety of devices for measurement of structure).
pore-water pressure and water levels). • Borehole extensometers (to monitor settlement, heave,
• Inclinometers (to measure lateral earth movements or and lateral deformation in soil and rock).
deflection of retaining walls and piles under load and to • Time domain reflectometry (TDR) and optical domain
provide a means of measuring settlement of embank- reflectometry (OTDR) measurements (provide a means of
ments foundations and other structures)
20.2 Monitoring Subsidence 267

Table 20.1 Comparison of land movement and subsidence measuring techniques

Method Displacement Resolution (millimeters) Spatial scale of measurement
InSAR Range (distance to ground) 5–10 Regional
Lidar Range (distance to ground) 70–150 Regional
Land survey (sprit level) to Vertical and horizontal 0.1–1 Local line or network
benchmark
Land survey (geodimeter) Horizontal 1 Regional line or network
Borehole extensometer Vertical or any angle 0.01–0.1 Point
TDR and OTDR Vertical or horizontal or any angle Varies with distance, 75 mm at Line
60 m distance
GPS Vertical and horizontal 20 vertical 5 horizontal Network, usually local
Tilit meter Rotation about a fixed point 10 arc sec Single point
Inclinometer Horizontal movement from the ±6 mm Length of the borehole
vertical per 25 m
Strain gauges Any orientation ±5 μ strain Localized on structure

monitoring changes in moisture or water level and is also Whenever instrumentation is to be installed and measure-
used to monitor subsidence). ments made in boreholes the method of drilling, casing
• Acoustic emissions used to detect abrupt movement of installation along with other factors need to be of special
soil and rock due to failure. concern.
• Strain gauges and load cells (provide a means of measur-
ing strain on steel and concrete structures). • A clean, stable, and open borehole without the use of
drilling mud is necessary for some geotechnical instru-
Again, we must be aware of scale effects in our measure- mentation and measurements.
ments and how they can impact the results. DaCunha (1990) • Borehole deviation may be of concern with certain
discusses the scale effects and the concept of representative measurements.
elementary volume (REV) associated with determination of
deformability, strength, internal stresses in rock masses as Detailed specifications for drilling and installation need
well as hydraulic properties. to be developed and executed in the field in order to assure
Some of the applicable engineering measurements are quality measurements. In his forward to Dunnicliff’s (1988)
also discussed in the chapters on surface geophysics book, Ralph Peck states that even with the utmost care and
(Chap. 16) and geophysical logging (Chap. 18). For example, diligence following the manuals, equipment and their result-
seismic measurements of compression (p) wave and shear (s) ing measurements can be unusable if a critical step is over-
wave velocities are used in calculation of elastic modulus. looked. There are also unique conditions where the manual
The measurement of velocity can be accomplished by should be purposely set aside. In both of these cases, a solid
surface seismic refraction measurements, MASW mea- geotechnical background and experience with the particular
surements, downhole or by hole-to-hole measurements equipment are necessary for success.
(ASTM 2014a, b and Wair et al. 2012).

20.2 Monitoring Subsidence

20.1.1 Drilling and the Installation
of Instrumentation Monitoring of subsidence may be required as a warning of
potential collapse where known cavities, caves or mines are
Many of the engineering measurements are made from located. Areas with signs of active subsidence may be moni-
within a borehole. Measurements made from within a bore- tored to confirm subsidence and assess subsidence rates.
hole are affected by both the location and the volume sam- There are a number of different ways to measure and monitor
pled (REV). An essential prerequisite to the efficient design land subsidence ranging from regional surveys to local site-
and monitoring of a valid instrumentation program is having specific monitoring (Fig. 20.1). The complexity of the mea-
an understanding of the geology and the proposed measure- surements can range from
ment. This improves the chances that the data are representa-
tive of the overall geologic system including both best and • simple observations and documentation (photographic/
worst case conditions. spray paint/etc.)
268 20 Engineering Measurements and Monitoring

• repetitive surveying measurements – horizontal and vertical control stations are used for horizontal deformation
changes (Dunnicliff 1988). A benchmark typically consists of a rod
• In place instrumentation for monitoring, some instrumen- or pipe anchored into rock at depth. The rod is isolated from
tation can also have data transferred directly to the office. contact with the surrounding soil and rock by being sur-
rounded by a larger diameter pipe. Since shallow datum
The strategy, as well as the technology to measure and points are subject to a variety of factors that can render them
monitor subsidence and collapse in karst and over mines, is unstable, deep benchmarks should be considered (Figs. 20.1b
currently available (Benson 2001). Table 20.1 summarizes and 20.2). Elevation at the top of rod is established and moni-
some of the common methods for measuring and monitoring tored over time in order to detect subsidence.
land subsidence. The methods are relatively simple in con- Benchmarks are commonly used to monitor subsidence
cept, but can be difficult in practice and as always each tech- related to sinkholes, mines, ground subsidence due to fluid
nique or measurement has advantages and limitations. All of withdrawal and monitoring dam stability and other critical
the many details for making these geotechnical measure- structures. An optical deformation survey provides a direct
ments are beyond the scope of this book. Advances continue measurement of the movement of a surface monument in
to occur in geotechnical instrumentation including measure- both the vertical direction and sometimes in the horizontal
ment technology, fiber optics, along with digital data acquisi- direction. By repeating measurements at regular intervals, a
tion and transfer. For further details, see Dunnicliff (1988) time series model of structural movements can be developed
supplemented by recent manufacture’s literature and providing accurate data for failure prediction. See Dunnicliff
recommendations. (1988) for categories of instruments for measuring deforma-
tion and the surveying methods that can be used.
A subsidence or deformation analysis requires the design,
20.2.1 Regional Subsidence observation and adjustment of a high accuracy survey net-
work. Surveyors who have performed first order geodetic
Regional subsidence monitoring requires very detailed mea- surveys are familiar with the basic principles. However, a
surements over a large area. This type of measurement can deformation survey usually involves even stricter accuracy
only be effectively obtained from airborne or satellite data requirements and requires virtually all aspects of the survey
such as InSAR or LiDAR (Fig. 20.1a). These techniques are to be rigidly controlled. A vertical network can be expected
summarized in Bawden et al. (2003) and have been discussed to detect vertical displacements on the order of 0.00021 m
in Sects. 14.5.2 and 14.5.3. using first order measurements (Boston Survey Consultants
These are valuable tools for developing regional subsid- 1990).
ence maps or showing frequency of new sinkhole develop- A system of telescoping benchmarks (TBM) was utilized
ment. Many papers are being written about the use of these in South Africa as a result of a devastating loss of lives and
methods along with their inherent limitations. It has been structures in the gold mining area (Brink 1984). This system
shown that field checking results, even on a random basis, is of TBMs was utilized to act as a warning device to monitor
necessary for accuracy (Doctor and Doctor 2012; Doctor the development of a sinkhole beneath existing structures.
et al. 2008). Each TBM consisted of a number of pipes, one within
another, with their bottoms cemented into the rock at differ-
ent depths. If a sinkhole or subsidence began to occur, the
20.2.2 Site-Specific Subsidence deepest rod would detect movement before the shallower
rods providing an early warning of subsidence.
Monitoring for site-specific subsidence or potential collapse In another example, benchmark monuments were installed
may be required in population areas to minimize loss of life over an abandoned limestone mine in the Kansas City area to
and property. There are a number of options available that determine if any surface subsidence was occurring associ-
can be used for developing a monitoring network over a large ated with a 7 ha collapse within the mine. The possibility of
area or more limited monitoring along a single survey line surface subsidence due to mine collapse was determined to
over a very specific area of concern. be highly unlikely by all experts who had inspected the mine
conditions. However, surface fissures that coincided with the
20.2.2.1 Reference Datum and Benchmarks area of mine collapse caused the question of mine stability to
For a site-wide assessment of subsidence, benchmarks or remain a concern by the state.
monuments can be installed and surveyed at intervals over The monument design was modified from that in
time. Measurements can be relative or absolute when tied to Dunnicliff (1988). Each monument consisted of a 2 cm steel
a stable reference datum. Benchmarks are the reference rebar with bolt on coupling anchored in a minimum of 1.4 m
datums used to detect vertical deformation while horizontal into the uppermost rock (Argentine Limestone) with concrete.
20.2 Monitoring Subsidence 269

Top of #6 rebar extends

above opened protective
casing

10 cm square protective
Ground casing with locking cap
Surface

#6 Rebar (2 cm)

5 cm diameter
PVC casing

15 cm auger hole
Overburden

Sand fill

Grout installed after

PVC placement

Grout installed prior to

PVC and rebar placement

Cemented a minimum
of 1.4 m into bedrock

Fig. 20.2 A benchmark consists of a rod or pipe anchored at depth

The rebar was encased within a 5 cm PVC casing, which was The deeper monuments were more than 30 m deep anchored
filled with loose sand. The outside of the PVC casing was in a massive limestone.
then grouted in place (Fig. 20.2). A protective steel casing Figure 20.4 shows examples of the subsidence data
was installed over the PVC casing. Three of the monuments obtained from this site. Three of the shallow monuments
were equipped with thermocouples at 30 cm intervals so that both on and off of the paleocollapse area are presented in
thermal expansion/contraction of the subsidence monuments Fig. 20.4a that showed little or no movement. Figure 20.4b
could be accounted for over seasonal measurements. The shows three of the monuments that did show subsidence.
vertical network could be expected to detect vertical dis- Monuments 6 and 10 are both shallow, monument 10 showed
placements of the order of 0.00021 m using first order verti- a very gradual subsidence while monument 6 had an abrupt
cal measurements. increase in subsidence. Monument 13 was a deep monument
A total of 11 shallow monuments were initially installed installed into the massive limestone after subsidence was
embedded in the uppermost rock (Fig. 20.3). Nine of the detected in the shallow monuments. It is very close to monu-
monuments were over the mine and two were off the mine ment 6 and has a similar rate of subsidence indicating that
and provided stable reference points. As a result of early both the shallow and deep rock is showing subsidence.
measurements that showed subsidence activity two addi- Further details are discussed in Part III, Chap. 25 case
tional shallow and two deeper monuments were installed. histories.
270 20 Engineering Measurements and Monitoring

Southwest Northeast

4 Additional Monuments
Original 11 Surface Fissures
Subsidence
Monuments 14/15 12/13
100 2 3 4 5 11 6 7 8 10 9 282

270
Argentine Limestone

Lane Shale 258

Elevation in Meters
Chanute Shale 246

Westerville Limestone
Wea Shale 234

Winterset Limestone
222
Bethany Falls Limestone Mine
Hushpuckney Shale
210
Central Collapse Area

Fig. 20.3 The locations of subsidence monuments are shown over an area of mine collapse

This project required extreme quality control measures in An example of this type of instrumentation comes from a
order to assure the accuracy of the data required. The team site in southeastern New Mexico where large surface col-
consisted of the authors who designed and installed the mon- lapse due to man-made cavities is not uncommon. The injec-
uments, Boston Survey Consultants who checked our monu- tion of fresh water into salt formations creates brine, which
ment design and layout, provided checks of data during the is pumped out and used as drilling fluids for the oil industry.
survey sequence and carried out an independent survey as a In the process of creating this brine fluid, the natural salt for-
means of quality control. A local survey firm who had expe- mation is dissolved leaving cavities of substantial size. The
rience with highly accurate survey measurements made peri- resulting surface collapse is often on the order of 100 m in
odic measurements and provided the data. This entire group diameter or more (Fig. 7.11). These features are typically in
worked closely throughout the program to minimize pitfalls remote areas with little direct impact to communities and
cited by Dunnicliff (1988) when carrying out such work. infrastructure. However, two brine wells were located in the
city of Carlsbad, New Mexico within 50 m of each other.
20.2.2.2 Inclinometers and Tilt Meters This is a populated area at the intersection of two major high-
Inclinometers and tilt meters are another means of measur- ways, a railroad track, a drainage canal and a trailer park
ing changes in the surface or subsurface movement immediately adjacent. Collapse in this area would have a
(Fig. 20.1c). Inclinometers typically monitor horizontal major impact on life and property (Land and Veni 2012).
movement from the vertical. Instrumentation consists a Brine production has ceased and an investigation to esti-
probe that runs along the length of a specialized designed mate cavity size and extent was completed. One of the inves-
track that is installed in a borehole. Time series measure- tigations utilized a 2D seismic reflection survey (Goodman
ments can detect changes in any lateral movement in multi- et al. 2009) and another utilized an electrical resistivity sur-
ple planes through the length of the borehole. Tilt meters vey (Land and Veni 2012). High-resolution tilt meters were
monitor the change in rotation of a fixed point on the surface, installed in the two borings as a monitoring system over the
within a borehole or on a structure. cavity. This early-warning system should provide indications
20.2 Monitoring Subsidence 271

a
0.0

Elevation Change in Centimeters

-1.5

-3.0
Monument 2 (shallow - off collapse area)
Monument 4 (shallow)
Monument 7 (shallow)
-4.5
0 200 400 600
Time in Days
b
0.0
Elevation Change in Centimeters

-1.5

-3.0

Monument 6 (shallow)
Monument 10 (shallow)
Monument 13 (deep)
-4.5
0 200 400 600
Time in Days

Fig. 20.4 Example of subsidence data over a paleocollapse area with monuments showing little to no movement (a) and small but abrupt move-
ment (b)

of subsurface subsidence that may precede a collapse. tion), the location of the break or distortion in the cable can
Surface subsidence is also being monitored directly over the be obtained. Optical domain reflectometry (OTDR) is simi-
cavern by optical survey measurements. In addition, two lar to TDR but uses a fiber optic cable and light rather than
wells were instrumented with pressure transducers and data an electric pulse (O’Conner and Dowding 1999).
loggers to detect sudden changes in water levels (Goodman These methods are being used in geotechnical and envi-
et al. 2009). ronmental applications to monitor changes in deformation of
soil and rock as well as changes in moisture content, water
20.2.2.3 Time Domain Reflectometry (TDR) level and density of soil (ASTM 2012). The cable may be
and Optical Domain Reflectometry installed vertically in a borehole or horizontally in a trench
(OTDR) Measurements (Fig. 20.1d) or at any angle from the vertical to the horizon-
Time domain reflectometry (TDR) measurements use an tal. TDR measurements can also be used within reinforced
electrical impulse sent down the length of an electrical cable. concrete to monitor strain and cracking. Prior to installation
When a significant distortion or a break in the cable is the TDR cable can be deformed (by crimping the cable) at
encountered a reflection occurs and is detected by the instru- intervals to provide a reflection from a known distance for
ment. By measuring the travel time of the impulse (reflec- calibrating measurements. A significant benefit of these
272 20 Engineering Measurements and Monitoring

Fig. 20.5 An example of a strain gauge mounted across a crack in a structure (a) to monitor any change in movement, (b) shows a close up of the
strain gauge

measurements is their ability to be continuously monitored 20.2.3 Localized Settlement or Subsidence

having a telemetry network transfer data to an office. This
enables TDR to be readily adapted to monitoring of karst or Local site-specific subsidence is typically focused on a spe-
mine related subsidence. Examples of TDR measurements to cific structure to address specific concerns often after signs
monitor soil and rock slopes are given by Anderson and of subsidence are noted such as cracks in a concrete struc-
Welch (2000). ture, warping of a sidewalk or localized subsidence in the
The details of the field installation become critically ground surface. A variety of strain gauges and transducers
important. The borehole diameter should be small to maxi- are available for monitoring these very localized conditions.
mize coupling between the soil or rock and the TDR cable. They are discussed in detail along with their application in
The grout strength should be less than the bearing capacity Dunnicliff (1988).
of the soil so that any movement of the soil or rock is coupled Strain gauges essentially measure the changes in the
to the TDR cable. width of a fracture or crack. They can be employed to moni-
An example of TDR measurements is from a longwall tor movement of cracks on structures (Fig. 20.1e). This type
coal mine that was advancing at a depth of approximately of measurement is quite simple, very localized and can be
150 m beneath I-70 in Pennsylvania. TDR measurements used in the horizontal or vertical direction (Fig. 20.5). This
were made in seven boreholes along with 32 tilt meters to type of measurement is typically used to monitor existing
detect the subsidence as the mine face advanced (O’Conner cracks to assess if they are old and stable or new and active.
et al. 2001). Precursor shear deformation was detected by
TDR when the face of the mine was more than 760 m from
the TDR cable. The rate of deformation increased when the References
mine face was within 60 m of a TDR cable. Tilt meter
response began as the mine face moved underneath the high- Anderson N, Welch D (2000) Practical applications of time domain
reflectometry (TDR) to monitor and analyze soil and rock slopes.
way and reached a peak value as the mine face moved past.
In: Marr WA (ed) Geotechnical measurements lab and field, ASCE
Dowding and O’Conner (2000) have compared TDR and GeoDenver, Reston, Virginia, pp 65–79
inclinometer measurements for slope monitoring. As with all ASTM (2012) Standard test method for water content and density of
measurements there are advantages and disadvantages, both soil in situ by time domain reflectometry (TDR) D6780/
D6780M-12. American Society for Testing and Materials, West
methods provide useful information. The TDR measure-
Conshohocken
ments respond best to localized shear and inclinometers are ASTM (2014a) Standard test methods for crosshole seismic testing,
especially sensitive to gradual changes in inclination or soils D4428M-14. American Society for Testing and Materials, West
undergoing general shear. Conshohocken
References 273

ASTM (2014b) Standard test methods for downhole seismic testing, Doctor DH et al (2008) Bedrock structural controls on the occurrence
D7400-14. American Society for Testing and Materials, West of sinkholes and springs in the northern Great Valley Karst,
Conshohocken Virginia and West Virginia. In: Yuhr L, Alexander CE Jr, Beck BF
Bawden GW et al (2003) Measuring human-induced land subsidence (eds) Proceedings of the 11th multidisciplinary conference on
from space. US Geological Survey Fact Sheet 069-03, December sinkholes and the engineering and environmental impacts of karst,
Benson RC (2001) Strategies of site characterization and risk manage- ASCE Geo-Institute, Geotechnical Special Publication No. 183,
ment. A presentation to National Research Council, committee on pp 12–22
coal waste impoundment failures subcommittee on impoundment Dowding CH, O’Connor KM (2000) Comparison of TDR and incli-
site characterization, National Research Council, Washington, DC nometers for slope monitoring. ASCE GeoDenver, Colorado
Boston Survey Consultants (BSC) (1990) Design and analysis of the Dunnicliff J (1988) Geotechnical instrumentation for monitoring field
subsidence monitoring network at forest view landfill. Consultants performance. Wiley, New York
report, Boston, Massachusetts Goodman WM et al (2009) Two-dimensional seismic evaluation of the
Brink ABA (1984) A brief review of the South African sinkhole prob- I&W brine cavern, Carlsbad, New Mexico. RESPEC Consulting
lem. In: Beck BF (ed) Proceedings of the 1st multidisciplinary con- and Services, Topical Report RSI-2083. http://www.emnrd.state.
ference on sinkholes: their geology, engineering and environmental nm.us/OCD/documents/RSI2083.pdf
impact, Orlando, FL. Balkema, Rotterdam, pp 123–127 Hunt RE (2005) Geotechnical engineering investigation handbook, 2nd
Da Cunha AP (ed) (1990) Scale effects in rock masses. In: Proceedings edn. CRC Press, Boca Raton
of the 1st international workshop on scale effects in rock masses. Land L, Veni G (2012) Electrical resistivity surveys of anthropogenic karst
Balkema, Rotterdam, pp 3–27 phenomena, southeastern New Mexico. Geology 34(4):117–125
DiMaggio J, Osborn P (eds) (2007) Field measurements in geomechan- O’Connor KM, Dowding CH (1999) Geo measurements by pulsing
ics. In: Proceedings of the 7th international symposium, Boston, TDR cables and probes. CRC Press, Boca Raton
MA, 24–27 September 2007, ASCE Geo-Institute, Geotechnical O’Connor KM et al (2001) Real time monitoring of subsidence along
Special Publication No. 175 I-70 in Washington and Pennsylvania. Transportation Research
Doctor DH, Doctor KZ (2012) Spatial analysis of geologic and hydro- Record Journal of the Transportation Research Board No. 1772,
logic features relating to sinkhole occurrence in Jefferson County, pp 32–39
West Virginia. Carbonates and evaporites, Special Issue on Wair BR et al (2012) Guidelines for estimation of shear wave velocity
Integrating Science and Engineering to Solve Karst Problems Part I profiles, PEER report 2012/08. Pacific Earthquake Engineering
June 27(2):143–152 Research Center, Berkley, 68p
 Hydrologic Characterization
and Measurements 21

Abstract
This chapter focuses attention upon the unique hydrologic characteristics of fractured and
karst aquifers and the wide range of measurements that can be utilized to characterize them.
At this stage of the characterization we should have a well-developed geologic conceptual
model of conditions with many of the questions about the site already answered (Tables
13.1, 13.2 and 13.3). We should already know the type of surface hydrologic features at the
site; the surface water and groundwater boundaries, in general if not in detail; the depth,
thickness and level of complexity of the epikarst; whether the aquifer is confined or uncon-
fined, and whether there is more than one. In addition, we will have preliminary data from
existing monitoring wells or initial borings on water levels, potentiometric surfaces and
flow directions. The objectives for this phase of work will focus upon the details to quantify
and support our conceptual model. This is obviously important for projects focused on
groundwater resources or contaminant remediation. But surface and groundwater in a frac-
tured or karst setting will also impact most engineered structures, if not managed well they
can result in unstable conditions.

21.1 A Complex System range of scales and that can be sampled over time in order to
accurately characterize them.
The hydrologic cycle is a large integrated system. Yet, the In a karst setting, there are a number of potential impacts
topics of surface water and groundwater have migrated into on the hydrology (water levels, flow, basin boundaries, etc.)
separate topics in text books, courses and practice. In a karst that can come from a number of sources. A site-specific char-
setting the rapid transmission of surface water into the aqui- acterization must include consideration of these variables
fer through sinkholes or sinking streams, results in surface that may include:
and groundwater being an intimately coupled system and
must be treated as such by practicing professionals. • Karst heterogeneities themselves including fractures,
Even though it is an integrated system we must consider conduits, sinkholes, and springs
the hydrologic parameters that may be associated with each • Structural features that may modify groundwater flow
zone of the complex karst system from the surface down include anticlines, synclines, and faults. These features
(Fig. 21.1). An individual feature such as a spring may be dry can concentrate groundwater flow or provide barriers to
unless heavy rainfall occurs. The epikarst layer may act as a groundwater flow
unit that both stores fluid and yet transmits it quickly at • Temporal variations such as recharge from rainfall,
discrete points. Each of the saturated zones may have a flow drought conditions, tidal effects, etc.
component that is different from those units (layers) above • Man-made changes can affect water levels or recharge
and below or even adjacent features. This wide range of such as dams/reservoirs, supply wells, concentrated run-
hydrologic conditions require a variety of measurements at a off, etc.

© Springer Science+Business Media Dordrecht 2016 275

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_21
276 21 Hydrologic Characterization and Measurements

inputs outputs
Surface Water

Vadose Zone flow direction

and rate
Epikarst

Surficial Aquifer

Aquaclude/Aquitard

Lower Aquifer

Fig. 21.1 A conceptual cross section with water flow options that may specific features such as fractures, conduits, weathered zones that will
need to be considered or measured within the vadose zone, the epikarst, also need to be considered and measured
any aquifers, and any aquitards. Within each of these areas will be

Therefore, there are a multitude of methods for field and Others support these ideas. The devil is in the details
laboratory measurement of hydrologic properties and sam- including the issues of scale and obtaining sufficient appro-
pling of groundwater and contaminants. For details, the priate and adequate data for this immensely complex
reader is referred to Freeze and Cheery (1979), Sara (2003), hydrologic system (Palmer 2006; Kovacs and Sauter 2007).
Nielsen (2006), and Hunt (2005). Sasowky (2000), and Efforts for groundwater modeling would benefit greatly from
Palmer (2006) also provide a practical discussion of ground- a more complete and accurate site characterization.
water hydraulics and its measurement specific to karst. Also
see ASTM D5717, a standard guide for developing ground-
water monitoring systems specifically for karst and 21.2 Karst Is a Multiple Porosity System
fractured-rock aquifers (ASTM 1998).
The next and often necessary step in hydrologic charac- Porosity is a primary factor when describing karst hydrology:
terization is that of groundwater modeling. However, the
details of groundwater modeling in karst are well beyond • Primary porosity is that associated with the original rock
the scope of this book. Palmer et al. (1999) is a publication as it was deposited and lithified.
specifically devoted to modeling of groundwater in karst. • Secondary porosity is associated with post deposition
It provides an excellent overview of the many aspects of activity caused by strain and stress resulting in fractures.
modeling karst flow. Topics include the various models • Tertiary porosity is caused by enlargement of secondary
that are employed, along with papers on the acquisition porosity features by dissolution. These tertiary porosity
and application of field data. Palmer suggests that the zones are commonly found as localized zones along frac-
acquisition of valid and sufficient field data is by far the tures or bedding planes. The size of these openings can
most limiting factor in successful modeling of groundwa- range from a few centimeters to many tens of meters.
ter flow in karst. In addition, they point out fatal flaws that Here again the issue of scale becomes important.
prevent successful modeling of groundwater flow in karst
include: Although the rock mass can provide most of the storage
of groundwater, flow in the rock mass is only a small per-
• The lack of approaching the karst system in its entirety, and centage of the volume of water discharging from a karst
• The limited acquisition of valid and sufficient field data. aquifer. Flow is primarily through fractures and conduits. A
21.3 Lets Revisit the Issue of Scale 277

major factor controlling the exchange of water between the In addition, the flow of groundwater through a
rock matrix and a conduit is the hydraulic gradient (Fig. 21.2). well-integrated network of fractures or conduits commonly
During low flow conditions, the water can flow from the rock transports water at rates several orders of magnitude greater
matrix to the conduit. During high flow conditions, the flow than those encountered in non-karst groundwater systems.
between the rock matrix and the conduit may reverse with Aley (2008) summarizes the mean straight-line groundwa-
water flowing from the conduit to the rock matrix. ter velocities from ten tracer tests, which ranged from 234
to 5,840 m/day. These rates are four to six orders of magni-
tude greater than groundwater flow rates commonly
reported in the literature for most aquifers. Such data dem-
onstrates that karst groundwater flow can rapidly transport
contaminants for considerable distances. However, flow
may need to be considered in all the hydrologic units, from
the surface, through the unsaturated zone and the epikarst
into the saturated zones (Fig. 21.1). Hydraulic conductivity
may vary by many orders of magnitude when going from a
relatively tight rock matrix to essentially pipe flow in cav-
ernous limestone. In most cases tertiary porosity will domi-
nate the flow.

21.3 Lets Revisit the Issue of Scale

The issue of scale is important to our hydrologic measure-

ments and their interpretation. The concept of REV, pre-
sented in Chap. 12, can be applied to measurements of both
hydrologic and engineering properties (Da Cunha 1990).
Benson and Yuhr (1993) present a discussion of spatial sam-
pling considerations in site characterization and its impor-
tance in characterizing fractured rock and karst systems. For
example, measurements from the same location (i.e., one
borehole) but taken at different scales (volumes measured)
will often yield very different values of hydraulic conductivi-
ties of fractured rock. A laboratory permeability test run on
an intact piece of core will generally yield lower hydraulic
conductivities than an in-situ slug test. A slug test will typi-
cally yield lower values than a packer test and a packer test
will usually yield lower values than a pump test. Each of
these measurements is integrating a larger volume of rock. If
the values from these measurements are all similar, we can
then assume that the system is uniformly fractured at a scale
similar to the smallest volume measured. On the other hand,
if the measurements disagree then we have identified scale
differences in the fracture system.
Rules of thumb suggest that to obtain representative
hydraulic conductivity values from fractured rock requires
that measurements be made on a volume of rock whose
dimensions are ten times the fracture spacing (Nelson 1984).
Da Cunha (1990) shows a variation of seven orders of mag-
nitude variation in hydraulic conductivity as the scale of tests
Fig. 21.2 Flow directions and water levels respond differently with range from laboratory to basin scale.
low, moderate and high recharge (a) Low rainfall and limited inflow
from upgradient (b) Moderate rainfall and flow from upgradient (c)
The scale of measurements can be divided into three lev-
High rainfall and large flow from upgradient (Adapted from Sowers els, regional, intermediate (or site-specific) and local near-
1996) field details. A regional scale of measurement can cover
278 21 Hydrologic Characterization and Measurements

many square kilometers and encompass the entire site and its Adequate temporal measurements must be made to avoid
surroundings. These measurements would include basin- gaps and aliasing of the data. Since observations and many
wide observations of hydrologic features along with measur- measurements are usually intermittent in nature they can eas-
ing discharge and recharge areas and dye trace testing. ily miss anomalous temporal conditions. To detect and record
Intermediate scale (or site-specific scale) of measurements anomalous temporal conditions, continuous recording instru-
would include pump tests, local dye tracing and surface geo- ments are necessary so that the rare events are detected and
physical surveys. Local scale within the site would include can be quantified.
boreholes, sampling and measurements such as slug, packer Two examples illustrating the need for temporal measure-
and pump tests, geophysical logging along with laboratory ments have been previously presented and represent changes
testing of core samples (Fig. 21.3). in surface water conditions. One example was the small
creek that one could step across without getting your feet wet
(Fig. 12.13a), which turned into a raging torrent (Fig. 12.13b)
21.4 Temporal Aspects that would sweep you away as a result of heavy rainfall in the
nearby mountains. Had we not been on-site during this event,
In addition to the spatial aspects of measurements, the tem- the speed and magnitude of this change in conditions would
poral aspects of measurements must also be considered. have been missed. As it was, our staff gauge measuring water
Temporal changes can include both natural (rainfall, tides, levels in the creek was insufficient for documenting this
drought) and man induced activities (pumping supply wells, event. The second example comes from the Peace River in
discharges from reservoirs). These changes can also occur southwest Florida. This area is normally a small quiet stream
rapidly or gradually. Dramatic changes in conditions over used by canoeists, but on occasion will become dry (Fig.
very short periods of time are common in karst hydrology. 15.6). While this is not a rapid change in conditions,

b
a
c

slug

transducer

d
Wells Monitoring
Water Level Responses Sample for Dye
Pumping Well e
Input Dye

Fig. 21.3 Some of the wide range of measurements that can be used to hole) (c) Slug test or Packer test measuring conditions surrounding the
characterize hydrologic conditions, each is associated with an increas- screened or packed interval (up to a few meters) (d) Pump test measur-
ing volume of measurement from detail to regional (a) Laboratory tests ing conditions over network of wells (to tens of m2) (e) Dye trace tests
on core sample (few tens of cm) (b) Logging tools measuring condi- measuring conditions over large distances (to hundreds of meters)
tions within the borehole or well (immediately surrounding the bore-
21.5 Hydrologic Measurements 279

depending upon the timing of work in this area, the hydrologic In order to optimize any hydrologic measurements,
impact and assessment of this river in its dry condition could particularly those using borings or wells, the locations need
have easily been missed. to be put into context at the site. Are they in the right loca-
Temporal changes in groundwater conditions also occur tion, drilled and screened at the correct depth to be represen-
and need to be anticipated. Fractures and conduits respond tative of conditions at the site? Conditions of existing wells
quickly to recharge events and may spill over into higher lay- should have been checked to confirm that they were con-
ing conduits or fractures (Fig. 21.2). In fractured and karst structed properly and have not deteriorated which could
rock it is not uncommon to see large variations in water lev- impact any data from them. An excellent example was pro-
els in response to recharge events. The degree of response in vided in Sect. 18.2.7 where a monitoring well audit revealed
a given well is dependent upon the size and interconnection that all five existing monitoring wells at a landfill were
between fractures and conduits encountered by the well and invalid. There were a number of construction errors along
their connections to the surface. Monitor wells in fractured with deterioration and they were installed assuming a single
rock and karst conditions may see large variations in water aquifer when there were actually two. These problems invali-
levels in immediate response to recharge events. dated all data from these wells (water levels and water qual-
Changes in water level data due to recharge from rainfall ity) that had been collected over many years.
can add information regarding the connectivity between the Figure 21.3 illustrates a few common hydrologic mea-
surface and groundwater to the conceptual model. In surements and is presented from the smallest volume of mea-
fractured-rock and karst aquifers with minimal overburden, surement to the largest. The types of measurements can
recharge tends to be rapid. Milanovic (2004) cites changes in provide specific hydrologic parameters at various scales to
water level of 90 m in 10 h and a maximum change of 312 m evaluate and characterize various hydrologic conditions.
in 183 days in eastern Herzegovina. These changes in water Some of these measurements include:
level elevations can cause flow to move in different direc-
tions on both a short term and seasonal basis resulting in • Laboratory testing (Fig. 21.3a) includes tests of water
significant temporal changes. A fractured rock or karst aqui- quality as well as soil or rock samples.
fer with a thick overburden may have a long temporal lag • Staff gauges and water level meters are very commonly
similar to that of a granular aquifer. Recharge may be distrib- used to monitor water levels at the surface or in a well.
uted through an aerially extensive network of fractures or • Flumes and weirs are structures placed in surface water
through soil (dispersed recharge), or may be concentrated at bodies to direct a portion of the flow through an opening
points that connect directly to the aquifer (point recharge via that can be instrumented to measure flow.
a sinkhole). The percentage of point recharge strongly influ- • Geophysical logging measurements (Fig. 21.3b) can mea-
ences the character and variability of its discharge and water sure flow within a borehole directly (impeller flow meters)
quality. These factors become critical when dealing with or indirectly (temperature logs). See Chap. 18 for more
contaminants. details on geophysical logging.
• Single well tests such as slug test or packer tests
(Fig. 21.3c) utilize man-induced changes in water levels
21.5 Hydrologic Measurements from wells or borings which can allow us to measure
recovery times and calculate various aquifer
The need for identifying and quantifying certain hydrologic parameters.
conditions will vary depending upon project objectives. As • Pump tests (Fig. 21.3d) utilize a single, often centrally
with all types of measurements, hydrologic measurements located well to be pumped to create a drawdown of the
should start with the simple and move to the more complex. water level, surrounding wells are then monitored for
These measurements should address the project needs changes in water levels.
such as: • Dye tracing (Fig. 21.3e) uses special chemical dyes intro-
duced into surface or groundwater at one point and sam-
• Water levels and potentiometric surfaces – spatial and pled at other points to confirm connections and travel
temporal impacts or changes times. The topic of dye tracing is often a critical compo-
• Flow – directions (laterally and vertically), rates (under nent for a site characterization in karst, therefore, we have
various rainfall events or seasonal conditions) included a separate chapter (Chap. 22) addressing dye
• Interconnections between surface, epikarst and trace measurements.
groundwater. • Thermal infrared and optical time domain reflectometry
• Aquitards, barriers and impermeable layers or zone have been mentioned for application in the sections on
• Water quality – natural conditions and contaminants, if surface geophysics (Chap. 16) and engineering measure-
needed ments (Chap. 20). These techniques can also provide
280 21 Hydrologic Characterization and Measurements

effective hydrologic measurement options in unusual The use of optical time domain (OTDR) has been used to
conditions. detect the flow of spring waters into rivers and lakes. Using
• Surface geophysical measurements can sometimes be fiber optic cable, Lane et al. (2008) have demonstrated the
used to measure the presence of moisture/water or indi- application of this method for stream-aquifer interaction and
cate flow by measuring the conductive component of the estuary–aquifer interaction. They suggest that the method pro-
groundwater associated with inorganic contaminants. vides a means to observe temperature over large areas with
fiber optic cables as long as 30 km, temperature resolution
from 0.1 to 0.01 °C, a spatial resolution of 1 m and temporal
resolution on the order of seconds to minutes. This type of data
21.6 Surface Water provides both the location as well as change in temperature
and data can be sent to a central monitoring station.
Surface water flow is often the easiest to measure since it can This method of measurement provides extremely good
be seen and is accessible. Direct measurements of surface lateral resolution along the length of the fiber optic cable. It
flow can be made by weirs, flumes, flowmeters or by moni- can also provide a response off to the side of the cable if the
toring a staff gauge (Palmer 2007). Surface water is often cooler spring water is forming a plume of some size.
affected the most by temporal changes (rainfall, evaporation, However, these measurements are inherently limited by the
runoff, discharges). Within the surface water body, localized specific location of the cable since the extent of the tempera-
flow can be coming from a spring or going into a sinkhole. ture plume may be off to one side of the cable, beyond the
These submerged features are often difficult to detect but can range of the cables sensitivity.
be a major contributor of flow to or from an aquifer. Off-shore spring flow is quite common in Florida and
Simple observations of surface water flow can often many of their general locations are known by fishermen and
reveal low areas or locations and patterns of groundwater divers. An example of detecting a near-shore spring using
recharge that were previously overlooked. Identifying areas marine resistivity has been presented in Chap. 16, Fig. 16.25.
of concentrated surface water that may be flowing into the This spring provided an electrical contrast in water quality at
subsurface can often trigger soil raveling and ultimately col- the point of discharge, which was detected in the marine
lapse features. This is particularly important in developed resistivity data.
areas that have changed the natural conditions (parking lots, In southeast Florida, Finkl and Krupa (2003) have mea-
stormwater drainage systems or stormwater ponds) (Fig. sured the flow and water quality from off-shore seeps and
8.1). springs using seepage meters. Flow was measured and water
Surface water is also often where contaminants are intro- quality samples acquired. These localized data were used to
duced into the system. The sources can range from street estimate total flows and nutrients levels being discharged
run-off to wastewater spray fields to agricultural use of fertil- off-shore via submerged seeps and springs.
izer/pesticides to accidental contaminants spills. These can
be point sources or broader areas of infiltration.
21.7 The Unsaturated Zone

21.6.1 Submerged Spring Flow Within Rivers, Seepage of water and flow within the unsaturated zone can
Lakes and Off-Shore be complex and extremely difficult to determine with any
level of confidence. The unsaturated zone may include soil
Springs with active flow on the surface are relatively easy to overburden, the epikarst or bedrock. Subtleties such as
locate and measure (Fig. 3.11b). To locate those hidden slightly cemented layers or variations in clay/silt content and
sources of spring flow within rivers and lakes or off-shore are root structures can greatly affect flow direction and speed
more problematic. An early method was to tow a tempera- within the shallow unsaturated soil zones. The epikarst can
ture and conductivity sensor (or array of sensors) through the act as a trap for groundwater or provide localized pathways
water to detect changes in temperature or conductivity asso- to the saturated zone. Unsaturated bedrock containing frac-
ciated with flow from springs. However this required that the tures, bedding planes, localized cavities and conduits can be
sensor pass within the area of flow, which was problematic. open or soil-filled impacting their connectivity. These are
The use of airborne thermal imaging (infrared camera) to very local details, which are difficult to measure and are eas-
located submerged spring flow has been very successful and ily overlooked.
provides a means to cover larger areas improving the chance Flow within the unsaturated soil zone and the epikarst are
of encountering these features (see Sect. 14.5.1 and Fig. usually of little concern unless there is a contaminant spill or
14.8). Anderson (2005) discusses the topic of assessing flow, raveling of soils into openings creating unstable surface con-
using heat as a tracer. ditions. An example of the complexity of the epikarst zone is
21.7 The Unsaturated Zone 281

illustrated by the 2D resistivity data acquired along a highway cut through the area of the paleocollapse feature, a major
in Kentucky (Fig. 16.14). The importance of the epikarst is mine roof collapse of 7 ha occurred and was referred to as
emphasized by the detailed study of the semi-confining layer the central collapse area (CCA) (Benson et al. 1994).
(SCL) at the EPA Superfund case history presented in Chap. The initial hydrologic conceptual model for the site
27. In this case history the continuity of the SCL was critical included:
to the prevention of shallow sinkholes as well as contaminant
migration. The details of measurements and sampling within • Unsaturated bedrock between the mine and the surface,
the unsaturated zone can be found in Wilson et al. (1994) and 52 m or more.
Nielsen and Johnson (1990). • The abandoned limestone mine was filling with water.
There may be indications of connection and potential
pathways for flow in the unsaturated zone indicated by air These issues impacted the development of the groundwa-
flow. Cool air flow is commonly noticed at the entrance of ter monitoring plan for the site.
caves, mines and their ventilation shafts. In addition, obvious
air flow has been noted at many sites coming from fractures 21.7.1.1 Unsaturated Bedrock
in the rock as well as open boreholes and well casings (see Based upon existing monitor wells and extensive drilling
Sect. 15.6.2 for further discussion). This can be an indication into the mine it was determined that the bedrock above the
that the borehole had encountered open fractures in the rock mine (52 m or more) was unsaturated except for limited
or possibly a cave system. quantities of isolated perched water (Fig. 21.4).
Existing monitoring wells were mostly dry, only a few
had limited amounts of water. There were many borings and
21.7.1 Groundwater Monitoring for a Landfill mine backfilling holes available. Geophysical logs and
in Karst downhole video camera had been run in many of these open
borings and indicated that the majority of the strata overlay-
A groundwater monitoring plan was required for the expan- ing the mine, (limestone and shale) were dry. While many
sion of a landfill over an abandoned underground mine in the fractures were identified, only one fracture with flowing
Kansas City area known as the Tobin mine (see Chap. 25 for water was observed.
more details of this case history). A paleokarst collapse Eight new wells drilled into the Raytown-Paola Limestone
structure due to deep-seated cavities about 182 m below the encountered small quantities of water, while none of the nine
mine floor had induced fractures in the overlying rock and new wells drilled into the Drum Limestone encountered
resulted in fissures within the loess soil. When the mine was water. Downhole video camera inspection of the borings

Fig. 21.4 Conceptual model for groundwater monitoring plan at a landfill overlying unsaturated bedrock and an abandoned limestone mine
282 21 Hydrologic Characterization and Measurements

showed isolated fracture flow within the Raytown-Paola enough lateral hydraulic conductivity to serve as a pathway
Limestone. Additional drilling, hydrologic tests (core analy- for significant water flow into the Tobin mine.
sis and packer tests) and even hydrofracturing was completed In addition, the Kansas River is located approximately
to assess optimum locations for monitoring. Permeability 213 m to the south of the Tobin mine. The elevation of the
testing in both limestones indicated values of 10−6 cm/s. mine floor in the southwest corner of the mine is about 7.6 m
Hydrofracturing of this zone improved the connection by below that of the Kansas River. A piezometer was installed
two orders of magnitude. The purpose of the hydrofracturing between the Kansas River and the southwest corner of the
was to improve the chance of the monitoring wells capturing mine. Water level differences of about 7.6 m were measured
any groundwater entering the bedrock. between the river and the mine. River water would flow
through the alluvium underlying the river into the permeable
21.7.1.2 The Source of Mine Water zone underlying the mine. This was a relatively simple solu-
The abandoned limestone mine was gradually filling with tion to determine the source of the water filling the mine. The
water, but the source of the water was unclear and little, if source for the mine water also supported the concept that the
any, water was found in the bedrock units between the sur- rock overlying the mine was essentially unsaturated.
face and the mine. The former mine superintendent,
Nicholson (1988 personal communication), indicated that 21.7.1.3 Results of the Dye Trace
the mine was generally dry until mining extended to the Although the rock overlying the mine was essentially unsat-
southwest. Then a truck of water about 3,700 l was removed urated, a dye trace test was carried out to evaluate the possi-
every few days. As the mine extended further westward, a ble connection from surface to the mine by the fractures
number of sumps were installed below the mine floor and associated with the paleocollapse. Section 17.2.4 discusses
pumped as needed to control water. After mining operations the excavation of a surface fracture and its evaluation using
ended in 1980 and without continued pumping, the mine angle borings (Fig. 17.16). This fracture then was used as the
began to slowly fill with water from the southwest. dye injection point. A total of 106,000 l of water was used to
Monitoring of water level within the mine was started in pre-wet and post flush the dye.
late 1987 using a digital water level recorder. A staff gauge This water along with dye was found to rapidly migrate
within the mine provided back-up data in case the electronic latterly in the Raytown-Paola and Drum limestone forma-
equipment failed and was read on each trip into the mine. tions up to 152 m from the injection site. This identified two
Measurements indicated that water was increasing at about zones within the unsaturated strata where monitor wells
3.3 cm/month or about 40 cm/year. At that time there was an could be located. These two strata were also the uppermost
estimated 227 × 106 l of water in the mine increasing at an limestone strata that would be continuous under the landfill.
estimated rate of approximately 1.3 × 106 l/month.
Possible sources of water entering the mine had been sug- 21.7.1.4 The Proposed Groundwater
gested by others including: Monitoring Plan
The proposed groundwater monitoring plan (Fig. 21.4) was
• Seepage from the portal wall (about 45 m of exposed based upon monitoring for possible landfill leachate at four
bedrock), points:
• Surface runoff into the mine from the portal road, and
• Seepage from the mine-roof. • The leachate collection system at the base of the landfill,
• Monitoring wells within the Raytown-Paola and Drum
None of these could account for the large amount of water Limestones which are the first continuous limestone units
entering the mine. below the landfill, and
Considerable work had been carried out at two other land- • A monitoring well within the mine.
fills and a large mine and quarry some 29 km east in Jackson
County, Missouri. The geology there was identical to that of
the Tobin mine (Fig. 21.4). The bedrock units immediately
below the mined Bethany Falls Limestone, were thought to 21.8 The Saturated Zone
be impermeable. These units have a total thickness of up to
3 m and include the Hushpuckney Shale, Middle Creek Measurements within the saturated zone measure flow
Limestone and Ladore Shale. However, lateral flow within directly. The more traditional measurements include using
these units had been observed at a number of locations at the either a single borehole or multiple boreholes for slug, packer
mine and quarry due to the secondary porosity related to or pump tests. Such measurements assume that the screen or
weathering and swelling of the shale (Camp 1988 personal open interval of the piezometer, well or borehole is represen-
communication). This suggests that this zone may have tative of hydrologic conditions. Several geophysical logging
21.8 The Saturated Zone 283

techniques may also be used within a borehole or well, they • Topographic maps (regional scale)
can provide an indication of flow or are able to quantify the • Aerial Photos (regional scale)
flow (see Chap. 18). • Regional geology and on-site borings (regional and site-
Traditional testing methods, such as slug tests, are limited specific detail)
to the immediate vicinity of the test well and its screen inter- • Field observations of outcrops and measurements along
val (Fig. 21.3c). Packer tests can be utilized in an open hole adjacent river and in a nearby quarry (regional and site-
to measure hydraulic conductivity over specific intervals of specific scale)
the borehole. Pump tests are a way to obtain values of • A cave map of Mystery Cave (site-specific scale off-site)
hydraulic conductivity over a much larger volume of the site • Groundwater flow from monitor wells on-site (site-
and are therefore more representative of conditions further specific scale)
from the borehole (Fig. 21.3d). These measurements are usu- • Water quality measurements from monitor wells on-site
ally limited to site-specific or detailed measurements. When (site-specific scale)
data over larger areas are needed dye tracing is generally • The results of pump tests and (site-specific scale)
used. See Chap. 22 for more details on dye tracing. • Electrical resistivity soundings (site-specific scale)

21.8.1 Example of Fracture Flow 21.8.1.1 Strong Linear Trends and Fracture
in Southeastern Minnesota Patterns
The topographic maps and aerial photo analysis (Fig. 21.5)
A hydrogeologic investigation was completed at a landfill indicated major linear trends and fracture patterns over the
located within a karst area of southeast Minnesota. The regional setting with linear trends up to 300 m long. Abrupt
investigation included data from a number of different mea- changes in the direction of the Root River bed including
surements at a range of scales including: numerous 90° bends and two 180° bends are a strong indication

Fig. 21.5 Topographic map indicates a number of linear trends (linea- over distances of tens of meters within the river along with a 180° turn
ments) over distances of hundreds of meters along the river with numer- just south of the site (USGS 7.5 min Quadrangles Spring Valley,
ous fracture trends (major valleys). A number of 90° bends are seen Wykoff, Ostrander and Cherry Grove, Minnesota, 1965)
284 21 Hydrologic Characterization and Measurements

of the fracture pattern at a scale of 30 m or so. Many deep, 21.8.1.2 Indications of Flow
linear drainage valleys are seen adjacent to the Root River. Water level data from monitor wells indicated the direction
A cross section of the site geology, within the Cedar of groundwater flow to the southwest of the site (Fig. 21.8).
Valley, Maquoketa, Dubuque, and the Gelena Formations is Specific conductance values from these monitor wells also
shown in Fig. 21.6. Observations of outcrops in the Cedar indicated the conductive plume from the landfill flowing in a
Valley Formation were made along 3 km of the Root River direction southwest of the site (Barr Engineering 1984).
downgradient of the site. Observations were also made of the Early results of a pump test initially indicated an elon-
Maquoketa, Dubuque, and the Gelena Formations in a nearby gated drawdown after one hour of pumping suggesting frac-
quarry. Both the outcrops along the river and those in the ture flow in a roughly north to south orientation (Fig. 21.9a),
quarry indicated closely spaced fracture patterns with spac- after 7.5 h the drawdown became elliptical (Fig. 21.9b), and
ing of a meter to 10 m. after 20 h became circular (Fig. 21.9c). This change in draw-
Rose diagrams were developed using the linear trends and down over time suggests that localized fracture flow initially
fracture orientation observed from aerials and topographic maps dominates but becomes more uniform “matrix-like” flow as
as well as those fractures observed along the Root River and in the time of the test increases (Barr Engineering 1984).
the nearby quarry. The features observed in the aerials and topo- Nine resistivity soundings (to depths of more than 30 m)
graphic maps are at a regional or site-specific scale. The rose were made around the perimeter of the landfill including
diagrams were plotted for features observed in the Cedar Valley upgradient and downgradient in order to detect the inorganic
and the Maquoketa Formations (Fig. 21.7a). Both of these contaminant plume, which has a distinct lower electrical resis-
formations indicate that these features have very distinct, well tivity signature. Figure 21.10 shows the locations of the resis-
defined orthogonal orientations (NW-SE and NE-SW). The data tivity soundings and an averaged electrical resistivity value at
from along the river and in a nearby quarry are more representa- depth. Multiple soundings were completed at each location,
tive of site-specific (30 m or more) and local scales (3 m or so). varying the orientation in order to detect any impact or control
These rose diagrams clearly show highly fractured formations due to the fractures/joints. The resistivity data at each location
with a more diverse and wide ranging orientation than those was repeatable, with no evidence of fracture control and
observed from the aerials and topographic maps Fig. 21.7b). clearly indicated a contrast in the groundwater quality within
The general location of Mystery Cave (the largest in the limestone with relatively high background values upgradi-
Minnesota) is located about 6 km to the east of the site within ent of the site and uniform lower values under and downgradi-
the Dubuque and Galena Formations (Figs. 21.5 and 21.6). ent of the landfill due to the presence of the leachate plume.
Its passageways are highly joint and fracture controlled. A The results of these resistivity soundings indicate that the
main east-west fracture system is superimposed with a NE to plume from the landfill tends to be uniform throughout the
SW and NW to SE fracture system. Many short passageways underlying limestone based upon resistivity values and has
are on the order of 30 m or so (Alexander et al. 2006). probably moved below the nearby Root River to the south.

WEST
South Branch EAST
411 Root River

Site
Windrow Formation
(sand/gravels)
Cedar Valley Formation
Elevation (m)

396
(dolomitic limestone) South Branch
Root River

Maquoketa (shaly dolomitic limestone) and

Dubuque (limestone/shale) Formations Mystery
381 Cave

Galena Formation
(dolomitic limestone/shale)
365
0 2 4 6 8
Distance in Kilometers

Fig. 21.6 Cross section shows the site location, the geologic section, the Root River valley and the nearby mystery Cave some 5.5–7 km east of
the site
21.8 The Saturated Zone 285

a N N

W EW E

S S
Cedar Valley Formation Maquoketa Formation

b N N

W E W E

S S
Cedar Valley Formation Maquoketa Formation

Fig. 21.7 Rose diagrams shows the fracture orientations obtained from aerials and topographic maps as compared to those obtained from the
outcrops along the Root River and from the nearby quarry (a) Data from aerials and topographic maps (b) Data from outcrops and nearby quarry

21.8.1.3 Scale Effects on Flow • The resistivity soundings to a depth of 30 m suggested

Strong major fracture patterns with a distinct and limited uniform flow though a highly fractured system of lime-
orientation were identified in both the topographic maps and stone with a coherent plume.
aerials photos. These strong fracture patterns were also seen
in the map of Mystery Cave. However, the more localized While each set of independent data by itself may not be
fracture patterns measured at outcrops along the river and in totally convincing, the correlation of these independent sets
a nearby quarry in both the Cedar Valley and Maquoketa of data over a wide range of scale provided a possible basis
Formations show many more fractures in a more diverse for interpreting the fracture flow at this location as equivalent
range of orientations. porous media (EPM). In this example the interpretation of
Additional data from localized measurements suggest a more uniform flow can be applied to the scale of the site
more uniform groundwater flow: itself, but may not be applicable to small areas within a site
depending upon the degree of local fractures and their
• The water levels from the monitor wells show a gradual interconnections.
flow to the southwest
• Pump test indicated that the fractures were well intercon-
nected and flow was uniform over a larger scale (longer 21.8.2 An Example of Flow
pumping time). from an Artesian Well
• The specific conductance values from the monitor wells
indicate leachate migrating to the southwest. A 30 cm oil exploration well had been drilled to a depth of
about 380 m in 1944 in southeast Florida. The well was
286 21 Hydrologic Characterization and Measurements

Location of
900
Pump Test
13
00
N

1290

12
790

95
650 600

12
85
12

600 1170
1020
80
Distance in Meters

490
850
940
830
650 430
650
800
800
300

Various wells
0 450 Specific Conductance (umhos/cm)
Groundwater Contour Levels

0 300 600 900 1200

Distance in Meters

Fig. 21.8 The direction of groundwater flow was determined from water levels in monitor wells along with specific conductance values (a) After
1 h of pumping (b) After 8 h of pumping (c) After 20 h of pumping (Note the sharp turns in the Root River west and south of the site)

artesian from the deeper Floridan Aquifer and flowed with The shallow conductivity contours (0–7.5 m deep) indicate
brackish water for 35 years before it was recognized as a an 11.5 km long plume extending from the well, 1.5 to 3 km
groundwater problem for the surficial Biscayne Aquifer. wide headed southeast toward Homestead Florida’s wellfield
The Grossman Hammock well is located within the (Fig. 21.12). Conductivity values of 8 milliseimens/m (mS/m)
Everglades National Park surrounded by a hardwood ham- represent the outermost measurable boundary of the plume
mock currently referred to as Chekika Day Use Area. Initial using the EM34 measurements. The deeper conductivity data
artesian flow was 7,500 l/min when the well was drilled in (0–15 m deep) shows a very similar plume shape, although
1944 and 3,750 l/min when the well was capped in 1995. slightly smaller in lateral extent than the shallower conductiv-
That would be roughly 15 × 1010 l for the 50 years of flow. ity data. The resistivity sounding data indicated that the plume
The Biscayne Aquifer lies within the upper 15 m, consists of is very uniform with depth. This was confirmed by wells at
young porous limestone, (Miami Oolite and the Fort various depths, showing specific conductance was also very
Thomson Formation) and provides a sole-source of potable uniform with depth. Both the shallow and deep data indicate
water for the southeastern Florida communities (Fig. 21.11) an off-shoot of the main plume extending to the east. However,
(Waller 1982). the boundaries in this area are based upon more limited data
The specific conductance of the artesian water (4,500– due to access and therefore are not well defined. The cause of
5,000 mg/l) was an order of magnitude greater than specific this portion of the plume is unknown but may be due to a
conductance levels of the surficial water (300–400 mg/l). porous zone developed at lower sea level stand.
This electrical contrast was used to map the plume using sur- Figure 21.13 is a plot of conductivity data crossing the
face electrical geophysical techniques (electromagnetic and plume. This cross section also runs along a canal where spe-
resistivity). EM34 electrical conductivity station measure- cific conductance measurements were made and are also
ments were made at the surface to map the lateral extent of plotted. There is a strong correlation between the conductiv-
the plume. A total of 133 station measurements were made ity measured by the surface geophysical methods and the
over 30 km2. Two measurements were made at each station water quality data of the canal. Both sets of data show sharp
to provide shallow data (0–7.5 m) and deeper data (0–15 m). boundary conditions.
21.9 Groundwater Contaminants 287

In this case, the limestone is a relatively young, highly

a BR6 0.02 porous matrix, the water table is very shallow (<3 m bls) and

BR1 0.
4 N brackish water flowed from the well for almost 40 years. The
scale of the brackish water plume is about 12 km long and
0.00
BR5 1.5–3 km wide, minimizing the impact of localized hetero-
0.00 geneities. In this case, flow can be characterized as equiva-
PW1
0.2
0.8
PW2 0.02 lent porous media (EPM).
0.6
0.90

0.64 BR4
BR2 0.00
BR3
PW3 0.00 21.8.3 Equivalent Porous Media
0.00

These previous two examples illustrate the application of

b EPM. In some limited cases the concept of Equivalent Porous
BR6 0.28
Media (EPM) can be used to evaluate flow in fractured rock
and limestone. When flow occurs through a young porous
BR1
0.25 limestone such as the Miami Limestone or a well fracture
BR5 media in which the fractures are closely spaced and uni-
0.03
PW1 formly distributed the system may be considered analogous
PW2 0.81
2.42 to porous media EPM (at an appropriate scale). For EPM
2.0
0.5
1.0
1.5

BR4 conditions to apply the observed vertical and horizontal frac-

BR2 0.03 tures should be numerous, the distance between the fractures
BR3
PW3 2.13 0.20 should be orders of magnitude smaller than the size of the
0.38
site under investigation, and the fractures should show appre-
ciable interconnection. When these conditions exist, then
c BR6 0.54 equivalent porous media (EPM) concepts can sometimes be
applied to the overall setting.
BR1 While the EPM approach is not generally appropriate
0.41
BR5 for modeling flow in most karst aquifers, it may be applied
0.03
1.0

PW1 with certain restrictions Kovacs and Sauter (2007).

2.0
3.0

PW2 2.52
4.0

4.72 Alexander (2005) cautions us about the use of Darcey’s

Law and the use of EPM in characterization of groundwater
BR4
BR2 0.04 flow in karst.
BR3
PW3 2.47 0.38
0.65

0 30
21.9 Groundwater Contaminants

approx horizontal There is often the need to determine the flow or map the
scale in meters extent of various contaminants as part of a site characteriza-
Fig. 21.9 The results of a pump test over time showing the water table
tion. The type of contaminants needs to be known so that
drawdown at 1, 8, and 20 h of pumping their interaction with the geologic and cultural conditions
can be anticipated. Are the contaminants inorganic or
Once the plume boundaries were established, resistivity organic? If they are organics, are they floaters, mixers or
soundings were made to establish the thickness of the plume sinkers? Would they be bound up in the soils or trapped in
and the base of the Biscayne aquifer (upper Tamiami sand pockets or perched on clay lenses? Are they coming
Formation). Data from existing test wells provided further from a point source or being dispersed over large areas? The
information on water quality for correlation with the geo- answers to these questions are necessary in order for the geo-
physical measurements. logic conceptual model to be developed for the site and
The limestone of the Biscayne Aquifer, a couple of meters appropriate sample points to be determined.
below the water table has enlarged porosity as high as 75 %
(Sowers 1996). The coherent nature of the plume, its sharp
boundaries and uniform conditions with depth indicate that 21.9.1 Inorganic Contaminants
the movement of the brackish water plume within the
Biscayne Aquifer can be described as equivalent porous Inorganic contaminants typically have a higher electrical
media as it flows towards the wellfield to the south. conductivity than groundwater. When there is an electrical
288 21 Hydrologic Characterization and Measurements

N R-7
R-8
900 ohm-m
R-9
1050 ohm-m
900 925 ohm-m

Estimated Extend of
Low Resistivity in the
Maquoketa Formation
600 R-6
Distance in Meters

800 ohm-m

R-1
300 ohm-m
300 R-5
550 ohm-m
R-3
800 ohm-m R-2
1000 ohm-m
R-4
550 ohm-m
0

Resistivity Sounding Location

0 300 600 900 1200

Distance in Meters

Fig. 21.10 Resistivity soundings R-1, R-4 and R-5 measured lower resistivity values (<600 ohm-m) indicating the presence of landfill leachate.
Background values range from 800 to 1050 ohm-m

B B’

Grossman Well Well

Hammock Cluster Cluster
Artesian Well #1 #2 Well #3 Well#4
0
Approximate Depth in Meters

Miami Oolite

6 Biscayne Aquifer
Fort Thompson Formation
12
Base of Biscayne Aquifer
18 Upper Tamiami Formation
Casing to 134m
Drilled depth 380m
24

0 1 2
Horizontal
Scale in Kilometers

Fig. 21.11 The geologic section and monitor wells used to characterize the plume from the Grossman Hammock artesian well
21.9 Groundwater Contaminants 289

A
Grossman
Hammock
Artesian
Well

X Grossman Drive
10
mS
/m

?
8 mS
/m

2
SW 237 Ave

Y
10 mS/m Hainlin Drive
A’

Z 3

N 0 1 2
4
Horizontal
S/m

Scale in Kilometers
8m
Country Club Road

1 to 4 Monitoring Wells
X to Z Resistivity Sounding

10 mS/m contour

8 mS/m contour - outer limit

of detection

Fig. 21.12 A 11.5 km long plume headed southeast toward a wellfield based upon electromagnetic measurements (0–7.5 m deep)

conductivity contrast between the contaminant and measurements. The borehole geophysical methods tradi-
surrounding pore fluids, soil and rock, surface geophysics tionally include resistivity/resistance and electromagnetic
can often be used to map the lateral extent and flow of measurements.
the contaminants while borehole geophysics can be used to The survey results often provide insight into the plume
characterize the vertical extent of contaminants. The behavior whether it is extending in the direction of the gen-
surface geophysical methods traditionally include eral groundwater flow or flowing along dominant fracture
resistivity, electromagnetic (EM31 and EM34) and radar trends. Benson et al. (1985) have compared the results of
 290 21 Hydrologic Characterization and Measurements

North South

20
A A’ 2000

18

16 1600
Electrical Conductivity mS/m

14

Specific Conductance mg/L

12 1200

10 Background conductivity

8 800

4 400
Shallow electrical conductivity (0 to 7.5 m)
2 Deep electrical conductivity (0 to 15 m)
Specific conductance from canal waters (<3 m)
0 0
0 3 6 9
Distance in Kilometers
Fig. 21.13 The plume had very sharp boundaries defined by surface geophysical electromagnetic measurements and specific conductance mea-
surements within canal waters

electrical geophysical survey methods and the water quality • Floaters are low density hydrocarbons which tend to float
data from numerous landfills, using specific conductance as at the top of the water table,
a leachate indicator parameter. A correlation of 90 % at the • Mixers are water soluble organic compounds which tend
95 % confidence level was found with both electrical resis- to migrate through an aquifer along with the groundwater
tivity and electromagnetic geophysical techniques. flow, and
For example, acid leaks had occurred at an electroplating • Sinkers are high-density halogenated hydrocarbons
plant in southwestern Missouri. Surface geophysical whose movement is strongly influenced by gravity.
electromagnetic measurements were acquired at two different
depths (EM31: 0–6 m and EM34: 0–15 m). The electrical Some organic contaminants such as petroleum products
contrast of the acids to background conditions provided a will degrade resulting in an increase in electrical conductiv-
parameter for mapping their extent. Figure 21.14 shows both ity (Sauck et al. 1998; Cassidy et al. 2001), which can then
the shallow and deep contoured electromagnetic data. The be mapped using electrical geophysical methods. Knowing
shallower data (Fig. 21.14a) shows a broader plume within how the contaminants were introduced into the subsurface
the soils (yellow to red). The deeper data (Fig. 21.14b) shows and how they may behave within the groundwater needs to
the contaminants (yellow to red) within the bedrock, largely be considered along with the geologic and cultural condi-
being controlled by fractures. tions present at a site.

21.9.2 Organic Contaminants 21.10 Aquitards and Barriers

The detection and mapping of organic contaminants is a In the 1950s and 1960s clay soils were thought to be imper-
more difficult task. In very general terms organic contami- meable (Daniel 1993a). Our understanding of clays has
nants can be broken into three categories: changed significantly over the past few decades. Subtle vari-
21.9 Groundwater Contaminants 291

a
Former Lagoons Easting (ft)
0 200 400 600 800 1000 1200 1400 1600
800 800
40 10
20
20

600 50 600
about 200 meters

50
Northing (ft)

Northing (ft)
40 30 30
50

400 Former Acid Pit 400

30
200
20 200

0 200 400 600 800 1000 1200 1400 1600

Easting (ft)

about 500 meters

b
Former Lagoons Easting (ft)
0 200 400 600 800 1000 1200 1400 1600
800 800
30

20
600 20 10 600
about 200 meters

30
Northing (ft)

Northing (ft)
400 40 400
Former Acid Pit
20
200 200
50

0 200 400 600 800 1000 1200 1400 1600

Easting (ft)
about 500 meters

Fig. 21.14 Electromagnetic surveys were used to map inorganic contaminants (acids), which were being controlled by fractures (a) Electromagnetic
survey (EM31: 0–6 m) (b) Electromagnetic survey (EM34: 0–15 m). Note: warmer colors indicate higher conductivity areas

ations in the clay/silt content or weakly cemented layers ers to various organic compounds may be 10–1,000 times
within soils can greatly affect unsaturated groundwater flow greater than water (Daniel 1985).
as well as contaminant transport. Clay liners that are used for Natural aquitards are often described as being present
landfills have leakage rates 10–100 times higher than pre- under an entire site and become the basis for design of land-
dicted by laboratory tests (Daniel 1984, 1993b). Furthermore, fills, groundwater modeling, risk analysis, and remediation
laboratory tests have shown that the permeability of clay lin- plans. In subsequent work it is not unusual to find that the
292 21 Hydrologic Characterization and Measurements

aquitard is sometimes found to be discontinuous or without The SCL had been identified in regional literature and during
a significant decrease in hydraulic conductivity to be classi- on-site investigations. This semi-confining layer is described
fied as an aquitard. Two issues need to be addressed in char- as a low permeability unit that retards vertical movement of
acterizing an aquitard. First, is there a sufficient change in groundwater between the two aquifers. It consists of a sandy-
hydraulic conductivity to establish an aquitard, and second, clay layer that grade into a weathered clayey limestone (the
is the lateral extent of the aquitard continuous across the upper portion of the Tampa Limestone) (Fig. 21.15).
site. Without such knowledge our conceptual models are The semi-confining layer was laterally and vertically
often flawed and subsequent design, groundwater flow mapped by several independent methods (drilling, direct
modeling and remediation may be in error. Pay attention to push geoprobe with a electrical conductivity tip, geophysical
the details. logging, ground penetrating radar and 2D resistivity). It was
For example, at the Savannah River Site (SRS), a “green found that the semi-confining layer was continuous over the
clay” which underlies a disposal facility had been described entire site with the exception of two large paleocollapse fea-
as an impermeable aquitard. The continuity and imperme- tures that were present. It is on average 2.3 m thick, but var-
ability of this aquitard was the basis for limiting the depth of ies from 0.3 to 7 m thick. Chemical sampling in paired wells
contaminant investigations, risk analysis and subsequent showed no chemicals migrating downward from the surficial
remediation. However, a sieve analysis of samples taken to the deeper Floridan aquifer. Visual observations of sam-
from this aquitard indicated it to be mostly coarse to very ples showed that the lower portion of this layer was typically
fine sand with only 10 % silts. Two electrical conductivity hard and dry. However, its ability to confine or impede flow
logs obtained with a geoprobe within 6 m of one another between the aquifers was in question. Therefore, additional
indicated that the properties of this aquitard are both verti- laboratory testing on samples of this layer were completed.
cally and laterally variable (Technos 1995). This analysis was completed on both the upper and lower
In some cases, the presence of an aquitard may not have portions of the semi-confining layer and included:
been identified. The example presented in Sect. 18.2.7
showed where an aquitard was missed due to widely spaced • Permeameter testing for hydraulic conductivity (15
samples during drilling and no review of regional literature. samples across the site), and
This aquitard played an important role in the vertical migra- • Evaluation of the silt and clay content through sieve anal-
tion of a brine contaminant plume. Because the aquitard and ysis (33 samples), hydrometer analysis (15 samples) and
its impact were unknown, monitoring wells were inappropri- x-ray diffraction (7 samples).
ately screened, missing the contaminants completely (Fig.
18.14). Not knowing an aquitard was present, clearly affected The results of the permeameter analysis showed that
the understanding of conditions at this site. hydraulic conductivities ranged from 2.84 × 10−4 to
Our case history from the Superfund site in Part III, Chap. <1 × 10−8 cm/s. The sieve and hydrometer testing all indi-
27 included an aquitard across the site referred to as the cated variability in the semi-confining layer, but in general
semi-confining layer (SCL) separating the surficial aquifer showed a high silt and clay content. The lower hydraulic
and the Floridan Aquifer in Tarpon Springs, Florida. The conductivities correlated well to high percentages of parti-
SCL played a critical role in limiting contaminant migration, cles passing the #200 sieve. However, the x-ray diffraction
indicating the presence of paleocollapse and developing a showed that some of the clay-sized particles were actually
subsidence risk assessment for the site (Yuhr et al. 2003). quartz and calcite. While this unit was variable, it did have

Hydraulic Conductivity
-2 -4
4.4 x 10 cm/sec to 3.5 x 10 cm/sec (slug test)
-2
3.8 x 10 to 3.0 x 10-3 cm/sec (pump test)
-4 -8
2.84 x 10 to <1 x 10 cm/sec
(lab permeameter tests)
-1 -1
1.5 x 10 to 3.0 x 10 (slug test)

Fig. 21.15 The conceptual model of the semi confining layer (SCL) and the surficial and Florida aquifers along with measured hydraulic conduc-
tivities for each layer
References 293

hydraulic conductivities that were one to six orders of mag- Daniel DE (1993b) Clay liners. In: Daniel DE (ed) Geotechnical prac-
nitude lower than the sands above or limestone below. This tice for waste disposal. Chapman & Hall, London, pp 137–163
Finkl CW, Krupa S (2003) Environmental impacts of coastal-plain
successfully limited contaminant migration into the underly- activities on sandy beach systems: hazards, perception and mitiga-
ing Floridan aquifer. Its lateral continuity across the site pro- tion. J Coast Res SI 35 (Proceedings of the Brazilian Symposium on
vided an indication of stability with regard to contaminant Sandy Beaches: Morphodynamics, Ecology, Uses, Hazards and
migration and sinkhole or subsidence activity. Where the Management):132–150. Itajai, SC–Brazil, ISSN 0749-0208
Freeze RA, Cherry JA (1979) Groundwater. Prentice Hall, Englewood
SCL was absent correlated to two paleocollapse areas. Cliffs
Hunt RE (2005) Geotechnical engineering investigation handbook, 2nd
edn. CRC Press, New York
Kovacs A, Sauter M (2007) Modeling karst hydrodynamics. In:
References Goldscheider N, Drew D (eds) Methods in karst hydrogeology,
International contributions to hydrogeology 26, International Assoc.
Alexander EC Jr (2005) Karst hydrogeology and the nature of reality of Hydrogeologists. Taylor & Francis, New York, pp 201–222
revisited: philosophical musings of a less frustrated curmudgeon. Lane J et al (2008) Fiber-optic distributed temperature sensing: a new
In: Beck B (ed) Proceedings of the 10th multidisciplinary confer- tool for assessment and monitoring of hydrologic processes. In:
ence on sinkholes and the engineering and environmental impacts of Proceeding of the symposium on the application of geophysics to
karst, San Antonio, Texas September 2005. ASCE Geo-Institute engineering and environmental problems, Environmental and
Geotechnical Special Publication 144 Engineering Geophysical Society, Apr 6–10, 2008, Philadelphia, PA
Alexander EC Jr et al (2006) The karst geology of southeastern Milanovic PT (2004) Water resources engineering in karst. CRC Press,
Minnesota. AIPG National Conference Field Trip, Sept 26 Boca Raton
Aley T (2008) Chapter 15 tracer tests—dye. In: Weight WD (ed) Neilsen D (2006) Practical handbook of environmental site character-
Hydrogeology field manual, 2nd edn. McGraw-Hill, New York, ization and ground-water monitoring. CRC Press, Taylor & Francis
pp 675–698 Group, New York
Anderson MP (2005) Heat as a ground water tracer. Ground Water Nelson P (1984) Short course notes for hydrologic properties of rock
43(6):951–968 and rock masses. University of Texas, Austin, Texas
ASTM (1998) Standard guide for design of ground-water monitoring Nielsen DM, Johnson A (eds) (1990) Groundwater and vadose zone mon-
systems in Karst and fractured-rock aquifers D5717–95e1 itoring. ASTM STP 1053, ASTM, West Conshohocken, Pennsylvania
(withdrawn 2005), ASTM International, West Conshohocken, Palmer AN (2006) Digital modeling of karst aquifers – successes, fail-
Pennsylvania ures, and promises. In: Harmon RS, Wicks CM (eds) Perspectives on
Barr Engineering (1984) Remediation of the Ironwood Landfill. karst geomorphology, hydrology and geochemistry – a tribute vol-
Consultants report ume to Derek C. Ford and William B White: Geological Society of
Benson RC, Yuhr L (1993) Spatial sampling considerations and their America, Special Paper 404, GSA, Boulder, Colorado, pp 243–250
applications to characterizing fractured rock and karst systems. In: Palmer AN (2007) Cave geology. Cave books, an affiliate of the cave
Beck BF (ed) Proceedings of the 4th multidisciplinary conference research foundation. Cave Books, Dayton
on sinkholes and the engineering and environmental impacts of Palmer AN et al (eds) (1999) Karst modeling. Special publication 5.
karst: applied karst geology, Panama City, Florida. Balkema, Karst Waters Institute, Charles Town, West Virginia
Rotterdam Sara MN (2003) Site assessment and remediation handbook, 2nd edn.
Benson RC et al (1985) Correlation between field geophysical mea- Lewis Publisher, New York
surements and laboratory water sample analysis. In: Proceedings of Sasowsky ID (2000) Carbonate aquifers: a review of thoughts and
the conference on surface and borehole geophysical methods in methods. In: Sasowsky W (ed) Groundwater flow and contaminant
ground water investigations, National Water Well Association, San transport in carbonate aquifers. Balkema, Rotterdam, pp 1–14
Antonio, TX, NWWA, Westerville, Ohio, pp 178–197 Sauck WA et al (1998) High conductivities associated with an LNAPL
Benson RC et al (1994) Ground water monitoring in unsaturated and plume imaged by integrated geophysical techniques. JEEG
saturated zones at a site with a paleokarst collapse structure. Am 2(3):203–212
Inst Hydrol Hydrol Sci Technol 10(1–4):31–46 Sowers GF (1996) Building on sinkholes, design and construction of
Cassidy DP et al (2001) The effects of LNAPL biodegradation products foundations in karst terrain. ASCE Press, Reston (out of print)
on electrical conductivity measurements. JEEG 6(1):47–52 Technos Inc. (1995) Final report, geologic, hydrologic and geophysical
Da Cunha AP (ed) (1990) Scale effects in rock masses. In: Proceedings work at the D-area seepage basin, Savannah River Site Plant,
of the 1st international workshop on scale effects in rock masses. Savannah Georgia for Ames Laboratory Technology Integration
Balkema, Rotterdam, pp 3–27 Program, Ames, Iowa. Consultant Report. Dec 7
Daniel DE (1984) Predicting hydraulic conductivity of clay liners. J Waller BG (1982) Aerial extent of a plume of mineralized water from a
Geotech Eng 110(2):285–300 flowing artisian well in Dade County. USGS Water Resources
Daniel DE (1985) Effects of organic chemicals on clay liners. In: Investigations Report 82–20, Tallahassee, Florida
Proceedings of the 7th symposium on management of uranium mill Wilson LG et al (ed) (1994) Handbook of vadose zone characterization
tailings, low-level waste and hazardous waste, Ft. Collins CO, and monitoring. Lewis Publishers, CRC Press, New York
pp 373–390 Yuhr L et al (2003) A case history of a large karst investigation pre-
Daniel DE (1993a) Introduction. In: Daniel DE (ed) Geotechnical prac- sented at the 3rd international conference on applied Geophysics,
tice for waste disposal. Chapman & Hall, London, pp 3–14 Orlando
 Dye Tracing
22

Abstract
Dye tracing is a valuable method to characterize flow in fractured-rock and karst aquifers.
It is the method of choice for regional and intermediate flow measurements but can also be
applied to some local measurements including the epikarst zone. In this chapter, we review
some considerations for dye tracing, including types of dyes, along with sampling and
analysis methods. A few examples are provided showing how dye tracing has been applied
to answer some very specific questions including flow under a dam site, vulnerability of a
cave habitat, confirmation of contaminant source and evaluation of capture zones for a
major spring. A special thanks to Tom Aley of the Ozark Underground Laboratory, who
provided assistance with this chapter as well as project examples based upon his lengthy
experience with dye tracing.

22.1 Introduction introduction and sampling points, developing a sampling

program (intervals, methods of sampling and methods of
A dye trace study is a way to track water flow using man- analysis) as well as a clear objective and realistic expecta-
made or natural tracers. Dye trace studies can be used to: tions from the data. A dye trace could take a few hours for a
short distance or be a basin-wide study that requires hundreds
• Prove a direct connection between two points, of samples and many months or years of time. In any case,
• Determine groundwater flow directions and rates, the planning is critical to assure optimum results.
• Identify basin boundaries, and While dye trace tests are useful for determining flow
• Determine optimum monitoring frequency for groundwa- connections and velocities in a system, they do not character-
ter quality. ize the size of the conduits or provide the locations of a cave
system, which are necessary for geotechnical investigations.
Our good friend, Jim Quinlan, was fond of repeating the Further details for dye tracing can be found in Aley (2002)
following phrase. “One well-designed tracer test, properly Groundwater Tracing Handbook which can be downloaded
done and correctly interpreted, is worth 1,000 expert opinions from his website and Alexander and Quinlan (1992) Practical
or 100 computer models of groundwater flow”. He had it Tracing of Groundwater with Emphasis on Karst Terranes, a
printed on his business card. The truth in this statement lies short course manual. In addition Smart (2005) discusses
in the fact that all appropriate field measurements, which are further details, strategy and errors of dye tracing.
properly carried out and interpreted are always better than
opinions and models. This is true not only of dye tracing but
of all methods of measurement and monitoring. 22.2 Considerations for Dye Tracing
A well designed dye trace test needs to consider a variety
of conditions and variables prior to injecting any dye or A tracer is a substance intentionally added to groundwater to
tracer and should involve someone experienced in this field. give it a distinctive signature that makes the groundwater
This should include mapping karst features, assessing dye recognizable elsewhere. Numerous types of tracers are used,

© Springer Science+Business Media Dordrecht 2016 295

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_22
296 22 Dye Tracing

Fig. 22.1 Fluorescein dye being introduced by Tom Aley into a sinking stream in Nevada (Photo courtesy of T. Aley and G. Baker)

both man-made and natural. The most common man-made

dyes include fluorescein, eosine, sulforhodamine B, and rho-
damine WT (Aley 2002, 2008). Natural tracers might include
salts or temperature.
Dye tracing is a point-to-point method that is dependent
upon the location of the dye introduction and sampling
points. This would typically require an inventory of karst
features as well as other natural and man-made windows into
the groundwater system. There are many different ways in
which dyes can be introduced into the groundwater system.
Sites for the introduction of dyes include; sinkholes, surface
streams that sink into the subsurface (Fig. 22.1), and directly
into cave streams. In some cases wells (Fig. 22.2), borings,
trenches, or ditches along a road may be used.
Selection of dye introduction and sampling points should
be optimized to assure a successful dye trace. A saturated
karst feature such as a sinking stream or water-filled sinkhole
would provide a greater probability of getting dye into the
subsurface karst system, than a dry, topographic depression
that may lie well above groundwater. Sometimes dyes are
placed into dry locations where they will be taken into
solution by the first flow of water. This is very useful for Fig. 22.2 Barry Beck introducing dye into a monitor well at a site in
stormwater management studies and in remote areas. West Central Florida to determine its possible connection to a recent
However, a saturated karst feature will usually result in the large sinkhole
shortest travel time and will also have the greatest chance to
yield a positive result. The exact location of dye introduction
will depend upon the objective of the dye trace. These In the case of a dry introduction site such as a dry well,
selections must also take into consideration the logistical dry fracture, or dry sinkhole the introduction area should be
aspects of both dye introduction and sampling. flooded with water prior to the introduction of the dye to
22.2 Considerations for Dye Tracing 297

Fig. 22.3 A large water truck used to pre-flush and post-flush the dye introduction site in an unsaturated system

saturate the system. After the introduction of the dye, the 22.2.2 Estimating Quantity of Dye
system should be flushed again with large quantities of water
to assist in flushing the dye through the fracture or karst system. Many factors affect the quantity of dye to be used. These
This usually requires a large quantity of water. Figure 22.3 include:
shows a water truck used for flushing dry systems.
The entire process of dye tracing in the epikarst is • Analytical protocol. Laboratory analysis with a spectro-
considerably different than in the karst groundwater sys- fluorophotometer operated under a synchronous scan
tem and requires additional knowledge and experience. protocol is the most sensitive and reliable method. This is
Ground water tracing in the epikarst should be considered especially true at waste sites and in industrialized areas
at waste sites or spills where the majority of contaminants where multiple fluorescent compounds may be present.
are localized within the epikarst zone with the potential to Other analytical approaches may require the use of more
be dispersed laterally and vertically through this zone. dye for credible results.
See Aley (1997) for more details regarding dye tracing • Factors along the flow route that may degrade the dye or
within the epikarst. absorb it onto earth materials. Different dyes have different
properties and are not all equally suited to all site-specific
conditions.
22.2.1 Water Sampling Prior • The amount of dye (called the dye equivalent) in the dye
to Introducing Dye mixture being used can vary. As an example, Aley (2002)
found dye equivalents in commercially available fluores-
There are numerous sources of dyes and related chemicals, cein dye mixtures varied from 2 to 80 %. If you do not
which are used in a variety of manufactured products that know the dye equivalent you have no knowledge of how
find their way into the surface and groundwater system. In much dye was introduced.
addition, it is not uncommon to find that dye trace testing has • The intensity of fluorescence for the particular dye. As an
already been done in a karst area, which may impact the example, after adjusting for dye equivalent percentages,
selection of dye. fluorescein emission fluorescence peaks in water are about
An adequate quantitative sampling of all sample locations 13 times larger than rhodamine WT peaks (Aley 2002).
prior to dye introduction is essential to establish the natural • Travel distances and volumes of water to be dyed.
background water quality and possible sources of interfer- • Nature of the flow system. Traces from a sinkhole to a
ence. Ideally, in order to assess background conditions at a spring routinely require much less dye than traces from
monitoring point, it should be monitored before and for most other features in the landscape. Traces to wells
several days or weeks, after a major recharge events. routinely require more dye than traces to springs.
298 22 Dye Tracing

Aley (2013, personal communication) indicates that less

than 0.5 kg of dye is seldom used. In most karst areas 0.5–
2.5 kg of dye is used. However, experience is the most effec-
tive method for determining appropriate dye quantities.

22.2.2.1 A Case of Too Much Dye

Rock quarries in western Dade County, Florida have operated
since the 1950s providing rock and cement to Florida’s con-
struction industry. The quarries encompass an area of about
200 km2 called the lake belt. After the mining, the quarries are
left as open fresh water lakes. These quarries allow the surface
water to be intimately connected to the groundwater over large
areas.
The quarries lie just west of the northwest wellfield that
provides drinking water to Dade County, Florida and draws
water from the sole-source Biscayne Aquifer. The wellfield
consists of 15 large wells providing water for residential and
commercial use.
Concerns arose about possible contamination by
Cryptosporidium, a parasite living in human and animal intes-
tines, or from terrorist acts contaminating a quarry. A 800 m no
mining protection buffer zone had been established around the
wellfield based upon chemical and bacteria contamination, but
more recent USGS studies have indicated rapid flow through Fig. 22.4 Mixing of fluorescein dye prior to its introduction into a
the limestone of the Biscayne Aquifer. The present buffer major fracture
zone is now thought to be insufficient to protect against
Cryptosporidium. No Cryptosporidium or giardina have ever
been detected in the lakes or the wells or the drinking water at vehicles, hotel rooms, along with eating facilities and may
the plant. Unfortunately Cryptosporidium is not controlled by not be detected. Minimizing all sources of cross contamina-
chlorine, which is commonly used to treat groundwater. tion between transporting and handling of dye as well as
USGS carried out a dye tracing test by introducing sampling procedures are critical to maintaining a reliable
approximately 50 kg of rhodamine dye in a well about 100 m analysis of all dye tracing efforts, yet little is written on this
upgradient from one of the large (7,900 gpm) pumping wells. critical subject.
The dye was expected to reach the well in 2–3 days, instead However discussions with Jim Quinlan on this subject
it arrived in 4–6 h after dye was introduced (Renken et al. revealed these guidelines. The individual responsible for
2005). The operator of the water treatment plant noticed a transporting, mixing, handling, and injecting dye should be
bright red color to the water and called for the well to be shut isolated from other personnel and sampling points to avoid
down. The following newspaper accounts of the event inadvertently cross contamination. If powder dyes are used,
reported it as “faucets flowed pink within hours after a dye small amounts can be windborne leading to wide spread
was introduced into a nearby test well”. Another reported undetected cross contamination. Those handling and intro-
“Beneath the Pink Underwear” because clothes in the wash ducing the dye should use separate vehicles, hotel rooms and
machines turned red that day (Dudley 2003). eating facilities. These precautionary procedures were used
This is a good example in which the planning phase of the in a dye trace assessing flow into the abandoned Tobin
dye trace work was missing some of the geologic and hydro- limestone mine (see Part III Chap. 25). Figure 22.4 shows
logic pieces of the puzzle (i.e. groundwater flow rate due to the mixing of the dye prior to its introduction. The person
the large pumping wells 100 m away) and the quantity of dye handling the dye used proper protective clothing and after
had been grossly over-estimated. introduction of the dye, they were removed from further
activities at the site to minimize possible cross
contamination.
22.2.3 Handling of Dye and Avoiding Cross Those collecting samples must use extreme care and be
Contamination trained in proper sampling techniques to avoid contamina-
tion. If such practice is not followed the results of dye traces
Small amounts of dye powder are easily lost to the air. Both might be questioned due to possible cross contamination
powder and liquids can easily contaminate personnel, clothing, with the dye used.
22.2 Considerations for Dye Tracing 299

Fig. 22.6 A gumdrop sampler (a) made famous by the late Dr. Jim
Fig. 22.5 A Turner fluorometer with water being continuously pumped
Quinlan. The gumdrop sampler suspends an activated charcoal sampler
through it to sample for the presence of any dye as the boat moves
above a streambed by a few centimeters out of the bottom sediments.
through the water within the Tobin mine. The results were recorded on
Here Jim installs a gumdrop sampler in the abandoned Tobin mine
a strip chart recorder
using a cement block as a base (b)

22.2.4 Sampling and Analysis Methods

Sampling for dye can be accomplished in a number of ways

depending upon the specific needs.

• In some cases the presence of dye can be determined

visually and no sample is taken.
• Discrete water samples may be obtained at selected time
intervals.
• Sampling and analysis can be done using a field fluorom-
eter. This provides a means of obtaining a continuous
sample and analysis in real time (Fig. 22.5).
• The use of charcoal packets, called ‘bugs’, are commonly
employed to acquire an integrated water sample over a
period of time (Fig. 22.6a). The bugs are removed and
replaced at selected sample intervals. Fig. 22.7 An automated water sampler can be used to obtain discrete
• Automated water sampling can be used at selected intervals samples at programmable intervals
using an automated pump and sampler system (Fig. 22.7).

Sample intervals would be selected to meet project intervals. This provides a means of obtaining water samples
objectives. However, typical sampling is more frequent without having to have personal on-site thereby reducing the
during and immediately after the introduction of the dye and cost of sampling.
becomes less frequent with the increase in distance of travel Dye detection techniques have a range of over eight orders
and time after dye introduction. Use of a programmable of magnitude from visual detection to the use of a spectrofluo-
automated water sampler (Fig. 22.7) such as manufactured rophotometer. Dye detection limits for four commonly used
by ISCO can be used to obtain discrete samples at periodic dyes are shown in Table 22.1. Values are based upon fluores-
300 22 Dye Tracing

Table 22.1 Dye detection methods (Aley 2013 personal communication)

Fluorescein (75 % Eosine (75 % dye Rhodamine WT (75 % Sulforhodamine B (20 %
Analysis method dye equivalent) ug/L equivalent) ug/L dye equivalent) ug/L dye equivalent) ug/L
Water sample analysis by 0.002 0.015 0.015 0.008
spectrofluoro-photometera
Elutant from carbon sampler and analysis 0.025 0.050 0.017 0.080
by spectrofluorophotometerb
Dark room with flashlight beam; visually 2 10 50 5
inspection by an experienced person
Visually in the field by an experienced 7 135 125 50
person
Visually in the field by the general public 140 13,500 2,500 1,000
a
Synchronous scan protocol with excitation slit of 3.0 nm and emission slit of 1.5 nm
b
Synchronous scan protocol with excitation slit of 5.0 nm and emission slit of 3.0 nm

cein, eosine, and sulforhodamine B mixtures with 75 % dye location and visually identifying it at another location. For
equivalents and rhodamine mixtures with 20 % dye equiva- example the authors have used dye tracing to determine
lent. All values micrograms/l (parts per billion). The concen- localized flow and dispersion characteristics of a sewage
tration of dye eluted from an activated carbon sampler left in outfall into the ocean off of Miami, to monitor the exchange
place for a week will be at least two orders of magnitude of surface waters within ocean inlets and to evaluate
greater than the mean dye concentration in the water. exchanges within marina basins and the Florida Bay.
Quantitative tracer studies require the introduction of
22.2.4.1 Charcoal Bugs known quantities of tracer and measuring tracer concentra-
Activated carbon samplers, referred to as charcoal bugs, are tions downgradient as they increase above background levels
commonly used in groundwater tracing studies. Activated char- and then return to background levels. This data is used to
coal is enclosed in a plastic mesh and placed at the sample loca- develop a breakthrough curve. Figure 22.8 is a very simplified
tion (Fig. 22.6a). These samplers adsorb and accumulate most breakthrough curve where dye concentrations are plotted
dyes providing an integrated sample over time at a sample point. against time relative to dye introduction.
An activated carbon sampler is left submerged for a week Breakthrough curves can be much more detailed and
where there is at least a slightly visible current. The concentra- allow for complex analysis (Fig. 22.9). This more detailed
tion of dye eluted from the sampler is routinely two orders of data provides an increase level of confidence and accuracy
magnitude greater or more than the mean concentration in the for determining:
water. The concentration factor in samplers placed in wells
with minimal amounts of water circulation is typically less, but • Identification of introduced tracer
is routinely at least an order of magnitude greater than in water • Measurement of groundwater velocities
samples (Aley 2013 personal communication). The charcoal • Estimating hydraulic properties, and
bugs are particularly applicable when working at very low con- • Assessment of hydraulic conditions.
centrations of dye or when sampling at large time intervals.
A water sample should be retrieved along with the char-
coal sampler and a new charcoal sampler would be put in its
place. The water samples provide independent data on dye 22.3.1 Dye Traces Confirming Connections
concentrations in the water at the time of sampling as a at a Dam Site
means of quality control. While the charcoal sampler is inte-
grating the dye over the interval between sampling, the water Crawford et al. (2005) describe the use of dye to deter-
sample is representing a single point in time. mine the connection between upstream sinkholes and
downstream springs and boils at a dam site. A hydro-
graphic survey using sonar and video had identified five
22.3 Results and Analysis of Dye Trace sinkholes in a lake upstream of the dam. There was a boil
Studies in the river about 550 m downstream and a spring 240 m
downstream on land. Three dyes were introduced into
Dye trace studies can be qualitative or quantitative. Simple three of the upstream sinkholes and samples collected
dye trace studies to assess near field connections are often downstream with an ISCO automatic sampler and acti-
used. These can be a simple as introducing the dye in one vated charcoal bugs.
22.3 Results and Analysis of Dye Trace Studies 301

Dye Detection Breakthrough Curve

12
Sample Point #5
Sample Point #7
10 Sample Point #2

Concentration (ppb)
8

4 Time of
Dye
Introduction
2

0
-6 -1 0 1 2 5 10 22 44 128 285 429 438
Day

Fig. 22.8 A simple breakthrough curve from the Tobin mine showing time and concentration of dye at three sample points within the mine (Note:
the time scale is not linear)

The dye trace clearly showed the connection between the was installed on the cave stream and programmed to pull
upstream sinkholes and the downstream spring and the boil. water samples every 2 h and make composites of these
The actual route taken by the water was unknown. Two con- samples every 8 h. Figure 22.10 shows the dye breakthrough
ceptual models were proposed to account for the flow curve resulting from this groundwater trace.
pathways. The details of the geologic setting and the dye test Water first flowed through the culvert at about 0500 h on
are provided in Crawford et al. (2005). November 15, 2006 (7 days after completion of the chip
sealing and 35 h after the dye was introduced). Precipitation
on November 15 and 16 equaled 4.6 cm. Dye first reached
22.3.2 Assessing Vulnerability of a Cave the sampler in the cave stream about 57 or 58 h after the first
Habitat flow through the culvert. This water sample, and the subse-
quent water sample that had the highest detected dye concen-
The following example was provided by Thomas Aley at the tration, were both analyzed for TPH. Both TPH concentrations
Ozark Underground Laboratory. A segment of federal high- were <100 ug/L, which was the detection limit. The straight-
way in Missouri was to be repaved by chip-sealing. The line travel distance through the karst groundwater system
highway segment was known to contribute water to Tumbling was about 5,100 m from one culvert and about 5,800 m from
Creek Cave, which provides habitat for an aquatic snail that the other. Using an average straight-line travel distance of
is on the federal endangered species list. There was concern 5,450 m and a travel time of 57 h equals a mean straight-line
that the process of chip sealing might yield detectable petro- groundwater velocity of 95 m/h. The resulting data were use-
leum hydrocarbons in the first stormwater runoff event after ful for evaluating potential TPH contamination of waters in
the chip seal application, and that this might adversely the cave stream but also valuable for planning emergency
impact the snail in the cave stream. Missouri Department of actions in the event of a highway spill in the studied area.
Transportation wanted to analyze water samples from the
cave stream for Total Purgeable Hydrocarbons (TPH) that
might be derived from the first flush of stormwater runoff 22.3.3 Dye Traces in the Woodville Karst Plain
after the chip sealing.
About 0.5 kg of fluorescein dye mixture containing 75 % Wakulla Spring is a first magnitude spring that lies within the
dye equivalent was placed in each of two road culverts in the Woodville Karst Plain (WKP) in north Florida. The WKP
repaved area. The culverts were dry at the time of dye place- was introduced in Part I, Sect. 7.2.2 and discussed in both the
ment, but the dye was placed where it would be taken into surface geophysical Sect. 16.5.3 and the geophysical logging
solution by the first storm flow. An automatic water sampler Sect. 18.3.2.
302 22 Dye Tracing

Fig. 22.9 The results from the analysis from a carbon sampler elutant 9.60 ug/L; the center peak is eosine at a concentration of 10.6 ug/L, and
analyzed with a spectrofluorophotometer. The curve shows three dyes the peak on the right is sulforhodamine B at a concentration of 14.1 ug/L
present in elutant from an activated carbon sampler in place for 14 days. (Figure courtesy of Aley)
The peak on the left represents fluorescein with a concentration of

The geology of the Woodville Karst Plain (WKP) consists Discharge from the Floridan aquifer under the WKP is
of a thin veneer of unconsolidated Pleistocene quartz sand and through springs in the southern part of the region including
shell beds overlaying a thick sequence of relatively horizontal Wakulla Spring, and submarine springs in Spring Creek
carbonate rocks that comprise the upper Floridan aquifer. Group at the edge of the Gulf of Mexico. Wakulla Spring is
More than 1,000 sinkholes have been estimated to exist across the largest inland spring in the WKP with an average discharge
the entire WKP, which is about 1,200 km2 (Kincaid et al. of 11 m3/s, and the third largest spring in Florida. The Spring
2012). The regional recharge for the area has been estimated to Creek group, which includes at least 14 underwater vents
cover 2,500 km2, (Gerami 1994; Davis 1996). Recharge to the along the Apalachee Bay in the Gulf of Mexico is listed as
Floridan aquifer in the Woodville Karst Plain occurs by: the largest spring in Florida (Scott et al. 2002).
Nitrate has been blamed for the decreasing water clarity
1. Sinking streams, in Wakulla Spring by fueling growth of hydrilla and algae.
2. Direct infiltration of precipitation through sinkholes, The City of Tallahassee’s Southeast Farm Wastewater Reuse
3. Infiltration through the variable thick sands and soils Facility lies within the recharge area of the WKP. This facil-
overlaying the aquifer ity sprays secondary treated wastewater onto the land surface
22.3 Results and Analysis of Dye Trace Studies 303

Fig. 22.10 A Fluorescien dye breakthrough curve at a weir in Tumbling Creek Cave. A rainfall event did not occur and create a flow rate until
35 h after the dye was introduced (Figure courtesy of Aley)

and was suspected as being a contributing source of nitrates 22.3.4 Evaluation of 2, 10 and 100 Year
at Wakulla Spring (Chelette et al. 2002). Capture Zone for Silver Springs, Florida
A number of dye traces have been carried out within the
WKP (Kincaid et al. 2005). The primary objectives of the The following example was provided by T Aley at the Ozark
most recent dye trace study was to show the connection Underground Laboratory. Silver Springs spring group is one
between the wastewater facility south of Tallahassee and the of the largest first magnitude springs in Florida with
springs and watersheds to the south, including: discharges from 30 named springs and spring vent clusters
located along the upper 1.2 km of the Silver River in Ocala,
• Developing tracer recovery curves, Florida. In 2004, the St. Johns River Water Management
• Evaluating hydraulic properties and mechanisms to be District delineated 2, 10, 100, and 1,000-year capture zones
used in subsequent numeric modeling, for Silver Springs based on a particle track simulation
• Estimate mass recoveries for specific pathways, and (MODPATH). One objective of the dye trace program was to
• Evaluate the effectiveness of in situ tracer detection evaluate the credibility of the delineated capture zones.
devices for longer term studies. An extensive dye trace program that lasted almost 17 months
and was a major component of the study (McGurk et al. 2012;
Dyes were injected in and around the wastewater URS et al. 2011). The traces were designed to identify domi-
sprayfield. Automatic water sampling was used at the nant groundwater pathways and travel times between specific
sampling points with high probability of detection, while locations and the springs. Dye sampling placed primary reli-
periodic grab samples were used at lower probability ance on activated carbon samplers since almost all dye concen-
of detection points. Samples were analyzed using a trations at sampling stations were expected to be below the
spectrofluorophotometer. detection limits in water samples. These detection limits were
The results of the dye trace test indicated a clear con- 0.002 ppb for the fluorescein mixture, 0.015 ppb for the eosine
nection from the wastewater facility to Wakulla Spring, a mixture, 0.015 for the rhodamine WT mixture, and 0.008 ppb
distance of about 16 km, at a rate of between 204 and for the sulforhodamine B mixture.
297 m/day (Fig. 22.11). The details of the dyes used, Four separate dyes were used, one in each of four injec-
injection locations, sample locations and time, along with tion sites (numbered 1–4). Fig. 22.12 shows the locations of
the results of analysis are described in Kincaid et al. the dye introduction points with respect to the springs as well
(2012). as the dye detection points.
304 22 Dye Tracing

Fig. 22.11 The results of numerous dye traces within the Woodville Karst Plain from sinkhole to sinkhole as well as from the City of Tallahassee’s
waste water disposal spray field to Wakulla Spring (Kincaid et al. 2012)

1. Ocala Civic Theatre: 9 kg of rhodamine WT dye mixture 2. Tuscawilla Park: 13.6 kg of eosine dye mixture with a
with a 20 % dye equivalent was introduced into a recently 75 % dye equivalent was introduced into a drainage well
formed sinkhole in a drainage retention area (at the Ocala 8.2 km from the nearest spring in the Silver Springs
Civic Theatre) 2.2 km from the nearest outlet for Silver Group at Tuscawilla Park. The first arrival time for the
Springs. This dye was detected at a number of the springs dye at this spring was between 295 and 312 days after
in the Silver Springs group. First arrival times for the dye dye introduction. This represented a mean straight-line
ranged from 5 to 10 days at two of the closer springs and averaged groundwater velocity of 27 m/day. The dye
39–45 days at one of the more distant springs. Based introduction point was within the estimated 10-year cap-
upon first dye arrival times mean straight-line groundwa- ture zone for Silver Springs and about 1,036 m from the
ter velocities averaged 321 m/day. The dye introduction outer edge of the 2-year capture zone.
point was within the estimated 2-year capture zone for 3. Pontiac Sink: 22.6 kg of sulforhodamine B dye mixture
Silver Springs. with a 75 % dye equivalent was introduced into a large
22.3 Results and Analysis of Dye Trace Studies 305

Orange Lake
I-75
4
N
Orange Lake to well
avg. 711 m/day

Orange Lake to well

avg. 76 m/day Silver Spring
Ocala Civic Theatre to Springs
avg. 321 m/day

1
2 Tusawilla Park to Springs
avg. 27 m/day
3
0 3 Est. 2 Yr
Capture Zone
Approx Scale in Kilometers Pontiac Sink to Well
avg. 45 m/day
1 - Ocala Civic Theatre
2 - Tuscawilla Park
3 - Pontiac Sink Est. 10 Yr
4 - Orange Lake Capture Zone
Dye Injection Point
Dye Sample Point Positive
Dye Sample Point Negative
Est. 100 Yr
Capture Zone

Fig. 22.12 Map of the Silver Springs spring group showing the 2, 10, and 100 year capture zones along with the points of dye introduction, points
of detection, their travel times and flow velocity (McGurk et al. 2012)

sinkhole adjacent to a constructed wetland for treating Distances from the sinkhole to the wells ranged from
stormwater runoff (Pontiac Sink). The sinkhole is 10 km about 3.4 to 14 km. First arrival times for the dye ranged
from the nearest outlet for Silver Springs. This dye was from 6 to 13 days at wells 6.75 km or less from the
detected at three public water supply wells roughly located sinkhole to 181 to 194 days at a well 14 km from the sink-
between the sinkhole and Silver Springs. Distances from hole. Based upon first dye arrival times mean straight-
the sinkhole to the wells ranged from 4 to 7 km. First arrival line groundwater velocities ranged from 76 to 711 m/day.
times for the dye ranged from 50 to 57 days at the nearest The dye introduction point is about 4 km outside of
well to 147–164 days at the furthest well. Based upon first the boundary of the estimated 100-year capture zone for
dye arrival times at the furthest distance mean straight-line Silver Springs.
groundwater velocities averaging 45 m/day. The dye intro-
duction point is about in the middle of the estimated The groundwater travel rates determined by the dye trace
10-year capture zone for Silver Springs. study indicate that Silver Springs is fed by a multiple porosity
4. Orange Lake: 13.6 kg of fluorescein dye mixture with a system that includes both conduits and large pore flow.
75 % dye equivalent was introduced into the Heagy-Burry Groundwater travel rates were routinely faster than model esti-
sinkhole in the edge of Orange Lake. Water routinely mates that were based upon porous media assumptions. The
drains from the lake into this sinkhole. Fluorescein dye most notable example of this was provided by the fluorescein
from this introduction was detected at four public water introduction at Orange Lake. This dye introduction point was
supply wells between the sinkhole and Silver Springs. 4 km outside of the boundary of the 100-year capture zone for
306 22 Dye Tracing

Silver Springs, yet it was first detected about half a year later at Chelette A et al (2002) Nitrate loading as an indicator of nonpoint
a well almost midway between the dye introduction point and source pollution in the lower S. Marks-Wakulla Rivers watershed:
Water Resources Special Report 02-1, Northwest Florida Water
Silver Springs. Aley (2008) reported a similar orders of magni- management District, Havana, Florida
tude differences between modeled travel times and those dem- Crawford NC et al (2005) Hydrologic investigation of leakage through
onstrated by dye tracing in a karst aquifer in Maryland. sinkholes in the bed of Lake Seminole to springs located down-
stream from Jim Woodruff Dam. In: Beck B (ed) Proceedings of the
10th multidisciplinary conference on sinkholes and the engineering
and environmental impacts of karst, San Antonio, Texas Sept 2005.
22.4 Limitations of Dye Traces ASCE Geo-Institute Geotechnical Special Publication 144,
pp 489–499
Even the best planned dye tract tests can encounter problems Davis H (1996) Hydrologic investigation and simulation of ground-
water flow in the upper Floridan Aquifer of north-central Florida
with varying levels of impact. Some of these problems may and delination of contributing areas for selected city of Tallahassee,
include: Florida water supply wells. USGS Water-Resources Investigation
Report 95-4296, USGS, Denver, Colorado
• Background fluorescence can mask (in part) the tracer Dudley S (2003) Beneath the pink underwear. The Miami News Times,
5 June 2003
signature Gerarmi A (1994) Hydrology of the St. Marks river basin, northwest
• Very dark tannic waters may impact measurement Florida. MS Thesis Florida State University, Tallahassee Florida
accuracy of tracer concentrations Kincaid T et al (2005) Quantitative groundwater tracing and effective
• Tracers discharge to unsampled points numerical modeling in karst: an example from the Woodville Karst
Plain of North Florida: In: Beck B (ed) Proceedings of the 10th
• Tracer travel times are longer than the sampling period multidisciplinary conference on sinkholes and the engineering and
environmental impacts of karst, San Antonio, Texas Sept 2005.
Some of these problems can be avoided or minimized ASCE Geo-Institute Geotechnical Special Publication 144,
with a well-planned and executed dye trace program. Other p 114–121
Kincaid T et al (2012) Demonstrating interconnection between a waste-
problems are simply those conditions that are unexpected water application facility and a first magnitude spring in a karstic
and cannot be anticipated. watershed: tracer study of the southeast Farm Wastewater Reuse
Facility, Tallahassee, Florida. Florida Department of Environmental
Protection, Florida Geological Survey Report of Investigation No.
111
References McGurk BE et al (2012) Silver Springs nutrient pathway characteriza-
tion project, final report. St. Johns River Water Management District
Alexander EC, Quinlan JF (1992) Practical tracing of groundwater with Special Publication SJ 2012-SP3
emphasis on karst terranes, 2nd edition. A short course presented at Renken RA et al (2005) Assessing the vulnerability of a municipal well
the annual meeting of the Geologic Society of America, Boulder, field to contamination in a karst aquifer. GSA Environ Eng Geosci
CO 11(4):319–331
Aley T (1997) Groundwater tracing in the epikarst. In: Beck B, Scott TM et al (2002) First magnitude springs of Florida. Florida
Stephenson JB (eds) Proceedings of the 6th Multidisciplinary Geological Survey Open File Report 85
Conference on Sinkholes and the Engineering and Environmental Smart C (2005) Error and techniques in fluorescent dye tracing. In:
Impacts of Karst: the engineering geology and hydrogeology of Beck B (ed) Proceedings of the 10th multidisciplinary conference
Karst Terranes. Springfield, Missouri Balkema, Rotterdam, on sinkholes and the engineering and environmental impacts of
pp 207–211 karst, San Antonio, Texas Sept 2005. ASCE Geo-Institute
Aley T (2002) Groundwater tracing handbook. Ozark Underground Geotechnical Special Publication 144, p 500–509
Laboratory available on-line from www.ozarkundergroundlab.com URS et al (2011) Hydrogeologic evaluation to support nutrient ground-
Aley T (2008) Chapter 15 tracer tests—dye. In: Weight WD (ed) water pathway delineation near Silver Springs, Marion County,
Hydrogeology field manual, 2nd edn. McGraw-Hill, New York, Florida. Final report with extended sampling prepared for St. Johns
pp 675–698 River Water Management District, Consultants Report
 The Conversion of Data to Useful
Information 23

Abstract
As the site characterization has progressed, many of the steps for processing, plotting,
checking and interpreting the individual data sets have already been carried out. In addition,
the conceptual model has been iteratively improved based upon these various sets of data.
At this point we will have a wide range of data from different sources (our own as well as
others). Much of this data is in different formats and scales. Now is the time for a final
review and fresh look at all of the data when we have it all before us and the rush of field-
work is over. The large amounts of data, gathered by the site characterization process, by
themselves are not that useful. A formidable part of the site characterization process is hav-
ing engineers, geologists and hydrologists assemble, conduct interpretations and integrate
this diverse set of data. The conversion of raw data to useful information is a value-added
process, which is achieved by careful professional analyses (Sharma 1994 personal com-
munication). As the raw data is converted to useful information, a conceptual model or
models are developed. The ultimate objective of the site characterization is to develop the
most accurate conceptual models of site conditions possible within the time and budget
constraints of the project. This is done by progressively minimizing the envelope of uncer-
tainties in the interpretation of the data.

23.1 An Assessment of All Data them from the active data set so that they cannot impact
further interpretations or calculations.
All of the data, both existing and newly acquired should
already have been assessed in order to minimize the uncer- After the data is acquired, a re-assessment of the data
tainties in the data. The assessment should include steps to: should be made. It is possible to meet all three criteria in a
data set, yet the data is not useable, in part, due to unexpected
• Make sure the data acquired is appropriate for the project conditions that caused noise in the data, interference that
needs and specific to the site itself. masks useable results, or impossible results. Situations like
• Determine that adequate data has been obtained both spa- these should be identified.
tially (laterally and with depth) and temporally to provide Once these three criteria are met, then and only then, can
sufficient coverage over the site and at different scales of we begin the assembly, interpretation and integration of dis-
measurements. parate sources of data. This allows us to proceed with confi-
• Confirm the accuracy of the data with thorough reviews dence in using the data to develop an accurate final conceptual
and analysis. This includes identifying any errors, gaps, model of the geologic and hydrologic site conditions (Yuhr
or impossible data. Correct them, if possible, or remove 1998; Yuhr et al. 1996).

© Springer Science+Business Media Dordrecht 2016 307

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_23
308 23 The Conversion of Data to Useful Information

23.2 Managing the Data review graphic plots of data to the same scale and format
than to review the same data in tabular form or other formats.
A proper site characterization will encompass a wide range The eye is a very powerful tool and can see spatial variations
of professional disciplines and generate an enormous amount in data if it is put in graphic form.
of data from a wide range of sources. Each data set may have Looking at a single piece of data at a time is not tremen-
a different scale of measurement, inherent bias, limitations dously useful or effective. Reviewing massive amounts of
and possible errors. Therefore, it should be recognized that printed data from different sources and in different formats is
the data are never totally complete or accurate. The line difficult and is often avoided. For example, looking at page
between factual data and interpretations and opinions is cer- after page of tabulated laboratory geotechnical measure-
tainly a gray one. In addition it should be recognized that we ments or chemical data can be confusing and may not readily
cannot and will not know everything about a site (Fig. 11.1). identify subtle changes or trends, errors or gaps in the data.
However, the site characterization strategy presented in Another example is the process of reviewing 50 geologic
Chap. 12 will have optimized the data acquired and recog- boring logs done by different firms and individuals with dif-
nizes the different levels of uncertainty that arise in both data ferent formats and level of detail. Both of these examples are
values as assembled and in their interpretations. difficult and confusing tasks. On the other hand graphic plots
There are no magic tricks involved to improve confidence of the printed laboratory results or 50 boring logs plotted to
levels and minimize uncertainties, just common sense, perse- the same scale in the same format can provide a more rapid
verance and attention to detail. These steps began before data and accurate review of such data.
was even acquired, by selecting the appropriate methods that The use of graphics allows subtle trends, errors, and gaps
would meet objectives and overcome site-specific con- in the data or conflicts between disparate sets of data to be
straints. Adequate spatial or temporal sampling for each identified and resolved. When appropriate graphic formats
method or measurement was optimized and the data was are utilized, data can be much more readily checked, inte-
checked for its accuracy. These three criteria (appropriate, grated and interpreted with confidence (Benson and Sharma
adequate and accurate) were also used to evaluate any exist- 1994, 1995). The issue of visualization of data is an important
ing data and databases. one, both from the data interpreter’s point of view and presen-
The next steps include a final assessment, processing, tation of information to others. Presentation of the data with
interpretation and integration of all the data as a complete consistent scales and complete labeling is one of the objec-
group. However, many of these steps are overlooked or tives of software graphics (Benson and Sharma 1994).
thought of as busy work for lower-level personnel. Yet, it is
here, where the professionals with experience are able to fit
the puzzle pieces together, identify the spatial trends and 23.3.2 Selecting Scales
make correlations between diverse data sets, which is criti-
cally important. The task of data assessment and interpreta- A key step is to assemble all of the data into a common
tion is not at all an easy one and there are a number of steps graphics format, with drawings to a common scale, when-
in the process that require considerable skill, experience and ever possible. As a rule, maps and cross sections should be
persistence. shown to a consistent scale, when possible, so that they may
be directly and easily compared. However, a comprehensive
site characterization will encompass data that represents a
23.3 Assembly of Data wide range of scales including regional data (kilometers),
site-specific data (hundreds of meters) and detailed data
The site characterization process is one of understanding (meters). As a result three different scales will commonly be
spatial and sometimes temporal variations in geologic and required for a site characterization. In some cases, more than
hydrologic conditions. Being able to put data or results of three scales may be necessary. Selecting the wrong scale
data into spatial and/or chronological perspective is essen- may hide subtle variations in some data that could have sig-
tial. However, integrating data from a wide variety of sources nificance or may magnify variations in some data that are
can be daunting. unimportant. This is where our professional judgment is
utilized.

23.3.1 The Use of Graphics

23.3.3 Developing Graphics
Almost all of the processes used for interpretation and analy-
sis of data can be summarized as comparing, contrasting, and Contour maps, cross sections, and fence diagrams are widely
correlating data spatially or with time. It is here where graph- used formats for presentation and interpretation of geologic
ics is an important tool since it is much quicker and easier to and hydrologic conditions. Regardless of the graphics being
23.3 Assembly of Data 309

a
West East

36 m
760 m
Southern cross section of the site
(all borings located along thel ine)

b
West East
Creek

36 m
760m
760 m
Northern cross section of the site
(some of the borings located far north off of the line)

Fig. 23.1 Two cross sections based upon boring data. The southern tances of 60–180 m. In addition, the presence, and thickness of the
cross section (a) is based upon a line of borings all on the cross section clayey aquitard along with the low in the top of rock underlying it, had
while the northern cross section (b) has borings projected over dis- been extrapolated onto the northern cross section

used, each figure must be supported by the data upon which An example is shown in Fig. 23.1 where existing geologic
it is based. For example, when a geologic cross section is data was obtained from two lines of borings at a low-level radio-
based upon borings data the locations of borings should be active disposal site north of the St Louis airport. The southern
shown on a plan view map. This will provide a means of cross section utilized a line of borings that were all located along
assessing the reliability of the cross section. If all borings lie the line of the cross section (Fig. 23.1a). The northern cross sec-
on or reasonably close to the cross section, there will be a tion (Fig. 23.1b) utilized some borings that were off-line by
higher level of confidence in the cross section compared with 60–200 m. Only one boring along the south line was used to
a cross section in which the borings are projected from long define the low in the top of rock seen in both cross sections.
distances away. In addition to the boring locations, their Based upon the geologic logs alone, it was not clear what other
depths and possibly a summary of the actual geologic log data, if any, were used to interpret a bedrock valley across the
may be needed. This will allow the cross section interpreta- site. Later in the project, gravity data confirmed the presence of
tion to be readily evaluated. the buried valley seen in the cross section.
310 23 The Conversion of Data to Useful Information

23.4 Processing of Data algorithm used for gridding and contouring and how much
one needs to honor the raw data.
Some data requires little, if any, processing such as our notes, Figure 23.2 shows two contours of the same data. The
sketches, and photographs. Other data will simply require data being contoured is the elevation of a shale layer from 39
transfer from our field worksheets to a standard format borings over about 40 ha. The site has fairly flat-lying hori-
(water levels or geologic logs). Some data (surface geophysi- zontal strata based upon a desk study, on-site observations
cal data, geophysical logs, pump tests, dye tests etc.) will and measurements. The contour shown in Fig. 23.2a used an
require some level of processing and result in a graphic and/ inverse distance algorithm while the contour in Fig. 23.2b
or a quantitative value. The processing of data requires close used a minimum curvature algorithm. Both of the contours in
attention to detail and a variety of diagnostic tools. Fig. 23.2 used default variables and are mathematically
The proper software can greatly facilitate the processing, correct but provide a very different presentation of the data.
managing and integrating the data. There is a wide range of The contour using inverse distance algorithm is honoring the
software available, each with its own advantages and limita- original data more effectively, but as a result creates a highly
tions. It is also important to be aware of what is being done variable contours. The contour using the minimum curvature
to the data by processing, the inherent assumptions in the algorithm is honoring the original data less effectively and is
software and the variables that are being selected. When data smoothing out the variations in the data. The contour using
is processed strange things can sometimes happen such as the minimum curvature algorithm (Fig. 23.2b) is the most
errors in data entry, errors in data transfer from program to appropriate since it illustrates the general trend of localized
program, choice of inappropriate processing software or subsidence over a paleocollapse feature, which was the
algorithms, inappropriate variables selected, over or under objective of this data.
filtering of the data, etc. As such, processing of the data As shown in Fig. 23.2, contouring can greatly affect the
should be an integral part of the site characterization team’s presentation of data and its ultimate interpretation. This
effort with proper control and assessment of the results by includes how closely the raw data is honored by the con-
the team members to assure the output is consistent with rea- tours, the contour interval selected, whether the interval is
sonable geologic and hydrologic site conditions. The team linear or logarithmic and how color or shading is used. When
members themselves must be the ones processing and man- using contours, the data point locations should be plotted on
aging the data thereby minimizing any opportunities for such a contour to show the spatial extent of the data that is being
problems. contoured. When possible, raw data values should be plotted
A simple and common example is one of contouring of on the contour as well.
data. Contouring is a practical way for displaying data to
show the third dimension on a map. Contours are easy to
interpret, and a skill widely used in the geologic and engi- 23.5 The Final Interpretation
neering communities. Drawing of contour maps and inter- and Conceptual Model
preting contour maps used to be universally taught to all
students of geology and engineering. In recent years, com- 23.5.1 Integration of Independent Data Sets
puter programs for contour data have become readily avail-
able and are sometimes used without any understanding of Throughout Part II of this book, we have repeatedly empha-
the impact to the data and its presentation. sized the need to make multiple independent observations
There are a wide variety of gridding and contouring and measurements of geologic and hydrologic conditions at
algorithms available in the different computer programs. different scales so that we can compare the results of one
They are each designed to effectively handle different types independent set of data to another. Once each individual set
of spatial data. Needless to say that by the time we see the of data has been interpreted by itself, the resulting interpre-
final contour, the spatial aspects of the raw data can be tation should then be checked against other sets of data and
changed in many ways. The basic concepts and limitations their interpretation to see if there is correlation or discrep-
of contouring and spatial interpretation and analysis of point ancies. The correlation of two or more independent sets of
data are presented in two papers by (Schreuder 1997, 1999) data is extremely powerful to reduce uncertainty. When the
and are well worth reading by all those using contouring results from different independent measurements agree, a
programs. higher level of confidence in our interpretation is achieved
In addition, one must have an understanding of the type of (Fig. 23.3). In addition, the resulting conceptual model is
data being contoured. Is the contour to represent a smoothly improved and uncertainties assumptions and opinions are
varying surface topography or highly variable top of rock? minimized. At the same time, others can easily and confi-
An understanding of the data will help determine the dently arrive at the same conclusions based upon the data.
23.5 The Final Interpretation and Conceptual Model 311

Fig. 23.2 Comparison of two different computer algorithms used for contouring the same data resulting in very different presentations

On the other hand, if it is found that the individual data karst investigation along a bridge alignment (Chap. 26) had
sets are in disagreement or don’t support the conceptual seven distinct sets of data, which showed excellent spatial
model for the site, something is likely wrong with our data, correlation (Fig. 23.4). These data included:
our interpretation or our conceptual model and additional
data or analysis may be required along with revised 1. Spatial correlation of rod drops and fluid loss in the
interpretations. original set of borings logs,
The integration and correlation of individual data sets is 2. An area of roadway that required frequent repaving due to
essential for building a complete and accurate understanding settlement,
of the site conditions. The three case histories presented in 3. An aerial photo lineament passing through the center of
Part III utilize diverse data, which show spatial correlation in the problem area,
order to build a final conceptual model. Each individual set 4. A large microgravity anomaly indicating a low-density
of data presents a piece of the puzzle. As an example, the zone,
312 23 The Conversion of Data to Useful Information

be supported by adequate site-specific data and observations

to provide a convincing basis for the assessment.
An example of this occurred at a landfill in Missouri. The
landfill had been constructed above an abandoned limestone
MEDIUM mine. A review of hydrogeologic conditions and a ground-
MEDIUM water monitoring plan was needed. The initial conceptual
LOW LOW model of the site (Fig. 23.5a) showed that existing monitor-
ing wells screened in the limestone and located closely
HIGH
HIGH around the perimeter of the landfill contained little, if any,
water. The limestone mine under the landfill contained about
a meter or so of water and was underlain by a impermeable
MEDIUM
MEDIUM MEDIUM
MEDIUM shale layer. So how was there water in the limestone mine
and not in the wells?
A detailed site characterization confirmed that the lime-
stone was unsaturated, but found that the underlying shale
LOW layer was actually quite permeable. There was also a subtle
dip within the mine to the west, and the water was found to
be flowing into the mine through the shale layer immediately
under the mine floor at this low area (Fig. 23.5b). The moni-
tor wells were only drilled into the limestone and not into the
Fig. 23.3 The circles represent independent data sets. When two sets deeper shale, which accounted for them being dry.
of data correlate, confidence levels increase from low to medium, when
all three sets of data correlate, confidence levels are higher Again, it is the attention to details and the willingness to
put the pieces of the geologic puzzle together that can clarify
conditions. In this case, the facts were correct, there was
water in the mine and the surrounding wells were dry, but the
5. Marine seismic reflection data indicating dipping strata at original conceptual model (Fig. 23.5a) was wrong. The
depth, assumption of shale always being impermeable was incor-
6. Additional borings confirming zones of rod drops and rect and the degradation by sulfides under the mine floor
fluid loss had been ignored even though it had been cited in the local
7. Geophysical logging showing areas of low-density literature (Hasan et al. 1988).
materials.

Some of these independent data sets by themselves may 23.5.3 Final Conceptual Model
not seem very significant. However, the spatial correlation of
all data sets clearly increases our confidence level in both the A preliminary conceptual model of geologic conditions is
individual interpretations as well as their correlation and used to guide early work at the site. This preliminary concep-
integration. An additional piece of data included independent tual model developed at the end of the desk study has been
research by USGS off-shore that identified a 600 m diameter continually upgraded throughout the site characterization
paleocollapse feature. The presence of such a feature shows process as new data has become available Fig. 12.17. This
what might be expected in the area and lends support to our process requires the conceptual model to be continually
interpretation of a large paleocollapse at the site. questioned, tested, verified, and modified as additional site-
specific data is acquired to minimizing uncertainties and
maximizing confidence levels (Yuhr 1998).
23.5.2 Final Interpretation As each individual data set is evaluated and interpreted it
is integrated into the conceptual model. Ideally each piece of
While many of our engineering and hydrologic methods of data fits into the conceptual model supporting other data and
measurement are highly quantitative, in many cases our final expanding our understanding of conditions. But what happens
interpretations will be subjective in which the judgment of when the data doesn’t fit into the conceptual model?
the investigator is used to quantify the results. Judgment is The first step would be to re-check the data to confirm that
based on inductive reasoning that incorporates site-specific it meets the criteria of appropriate, adequate and accurate
data, observations and professional opinion, experience and data and make sure that the data’s limitations are understood.
integrity of the site characterization team into an assessment. If this step indicates that the data is solid then check the indi-
As part of the data analysis, subjective data and results must vidual data interpretation. Is there more than one interpretation
23.5 The Final Interpretation and Conceptual Model 313

Existing Data

29% of rod Frequent

drops and fluid repaving of Independent
loss occurred roadway due research off-shore
in two adjacent to settlement revealed a 600 m
borings dia. paleocollapse

Conceptual Model
of a Large
Paleocollapse
Zone

A 100 uGal Seismic Geophysical

Microgravity Reflection Data Logging
Anomaly Revealed Indicated Low
Dipping Strata Density Zone

Aerial Photo Additional

Analysis Borings
Identifying Major With Fluid Losses
Lineament and Rod Drops

Newly Acquired Data

Fig. 23.4 A combination of existing and newly acquired data were used to develop and support the final conceptual model for the presence of a
large paleocollapse

that can be made? Is there one that fits the conceptual model conditions which incorporates all of the essential features
better? If the data and interpretation are not in question, then of the physical system. This final conceptual model must
the conceptual model needs to be checked. be supported by the available data including maps and
Initial conceptual models are often based upon early opin- cross sections of sufficient detail, specific properties, and
ions rather than data or on data that has not yet been checked. calculations or models.
While the conceptual models should be tested and modified A final conceptual model of karst conditions is shown in
throughout the site characterization process, early interpreta- Fig. 23.6 from the low-level radioactive disposal site
tion without sufficient data can easily lead to an erroneous (SLAPS) just north of the St Louis Airport. Supporting data
understanding of site conditions and an incorrect conceptual for the conceptual model included:
model.
However, at this point, the integration of data over a wide • The geologic cross sections along with 81 borings over
range of scales and correlation of multiple data sets should the site (Fig. 23.1)
have minimized the geologic uncertainties and increase the • Geology and Geomorphology
accuracy of the overall site characterization to an acceptable • Analysis of topographic maps and aerial photos for linea-
confidence level. By now we have the data to develop a rea- ment analysis
sonably, accurate regional and local conceptual model of site • Karst literature and cave maps
314 23 The Conversion of Data to Useful Information

a Monitor Wells
East Perimeter Monitoring showed water in
Wells were all dry mine West

Landfill

Impermeable Abandoned Mine

Shale

Initial conceptual model based upon available data

b
Perimeter Monitoring Wells
East were not drilling into the shale West

Some Water Probably Infiltrated

Through Cracks in the Landfill

Sump continually
pumped to dewater
adjacent quarry

The shale underlying the mine was very permeable

and mine floor was dipping to the west
Final conceptual model with additional data

Fig. 23.5 The initial conceptual model for a site shows conflicting the mine floor and permeability of the shale layer underlying the mine,
issues regarding the presence of water in the mine versus dry wells both of which allows water to enter the mine while the monitoring wells
around the mine (a). An updated conceptual model illustrates the dip in in the limestone remains dry (b)

• Mapping outcrops joints and bedding planes. The eastern boundary of 3M clay
• Microgravity and TDEM resistivity measurements layer had also been defined along with where it was missing
• Geophysical logging from the eastern portion of the site. However, the lack of borings
into the bedrock (Fig. 12.11) has limited any detailed assess-
The final conceptual models included a plan view and ment of conditions or contaminants within the limestone.
cross section for the site. The buried valley was confirmed As complete as a conceptual model may be, it remains a
running south to north through the site. Within the Florissant simplified hypothesis of site conditions and must be under-
Basin, which underlies the site the limestone had been stood as such. It is also expected that these conceptual mod-
exposed allowing extensive weathering and dissolution. The els will be supported by more detailed technical data such as
area was then filled with glacial material. Nearby outcrops a contour map for a particular unit. However, we must con-
confirmed open fissures were likely to be present in the upper stantly be aware that even a sophisticated conceptual model
most limestone. Significant weathering at the top of rock was does not represent a totally complete or accurate description
confirmed in addition to preferential dissolutioning along of a site.
23.6 Visualization and Presentation of Data 315

a N Area of
Missing
3M Unit

railroad
Buried
k
Channel Cree

SLAPS
Site

b Distance in Meters
0 500 1000
0

Clayey silt/silty clay

Depth in Meters

20 3M Unit (Clay) Shale

40
Limestone

60

Fig. 23.6 The final conceptual model of geologic conditions for the SLAPS site. The extent of the buried channel and where the 3M clay unit is
missing is shown using a plan view map (a), while the fissures in the upper most portion of the limestone is shown in a cross section (b)

23.6 Visualization and Presentation Graphics used for interpretation analysis and the report
of Data must contain supporting data upon which the figure is based
and may be somewhat cluttered with this additional data.
There are a wide range of graphics and presentation software Graphics used for presentation purposes are intended to pro-
that can be utilized for presentation of data along with many vide a summary of conditions and do not need to include the
options for contours, cross sections and other formats. Even multitude of details needed for purposes of data interpreta-
a simple sketch (Fig. 23.7) can be used to illustrate the tion, correlation and the report. Computer graphics can often
conceptual relationship between observed surface fissures be used to make complex geologic conditions clear to inter-
and a major mine collapse. This figure was the basis of our ested professionals and non-technical audiences.
early conceptual model of conditions in our case history pre- Geology and hydrology is a three-dimensional system
sented in Chap. 25, which was later supported by a wide thus it is best understood and presented where possible in
range of additional data. three-dimensions (3D). The use of 3D sketches is advocated
The use of software and particularly graphics software by (Fookes 1997; Fookes et al. 2005). Such art skills seldom
meets two critical project requirements; one is to aid the exist with most professionals today having been replaced by
analysis and interpretation of data, the other is in the effec- photographs and computer graphics. In many cases, simple
tive visual presentation of data. The graphics used for techni- lines of 2D graphics can be shown together as fence dia-
cal analysis, interpretation and reports may be different than grams to provide a 3D representation of data. Figure 23.8 is
graphics used for general presentations. an example of such data. This presentation illustrates the
316 23 The Conversion of Data to Useful Information

Surface Fissure in Loess

Loess Soil

Major Fracture
Exposed in Argentine-Frisbie
Excavation Limestone
50

Lane Shale

40

Possible
Paleofracture

Meters Above Mine Floor

30

Massive Westerville
Limestone
20
Wea Shale

Block Limestone Break in dip, likely

Fontana Shale associated with collapse

Winter set
Limestone Mine is bulked shut 10
around perimeter of
Stark-Galesburg 7 hectare collapse
Shale
Bethany Falls
Limestone Interior of the
7 hectare collapse

Fig. 23.7 A simple sketch of an initial conceptual model illustrates the possible connection between surface fissures seen in the loess soil, the
uppermost rock and the major collapse within the mine

trends in the data from line to line and avoids unnecessary the 3D imaging has been promoted to integrate many sets of
and unsupported extrapolation of data. measurements into one or more of the final images. The
Three-dimensional graphics developed by computer may problem is that there are inherently increasing levels of inter-
be pleasing to the eye but may be technically defective. polations and extrapolations from the raw data points in the
Experienced professionals need to monitor the process and process of developing such complex 3D images. These
control results so that accuracy is maintained. As the use of images often leave reality behind creating artifacts and pos-
computers and software continues to expand, there is a ten- sibly misrepresenting the actual data. Graphics should be
dency to promote greater use of 3D graphics. In some cases used to convey information not to impress.
References 317

Fig. 23.8 Three lines of 2D MASW (surface wave) data show correlation between lines and illustrates their 3D relationship

23.7 Documentation: A Final Report The report should step a technical reader through the site
characterization process so that the basis of the interpreta-
The final report of findings should be based upon a solid tions are clear. The report should be written so that other
foundation of data and should include sufficiently complete professionals can easily reach the same conclusions.
information so that the owner and other professionals can In most cases, it is recommended that a technical presen-
understand the purpose, technical approach, and the results tation of the site characterization and its results be made.
of the investigation along with any limitations. The report All too often busy professionals may only read the execu-
should include: tive summary and scan the report. A verbal presentation
provides the focused review and opportunity for questions
• The strategy for the site characterization effort along with and discussion.
its goals and with any limitations in scope,
• The methodologies used along with their advantages and
limitations, References
• Individual data sets should be discussed and then inte-
grated with other data, Benson RC, Sharma D (1994) The pitfalls in site characterization and
• A final conceptual model or models of site conditions how to avoid them. Short course presented by Environmental
Education Enterprises, Salt Lake City
along with a supporting data,
Benson RC, Sharma D (1995) Pitfalls of site characterization and how
• The results of modeling or risk assessment (if needed), and to avoid them. Short course presented by Environmental Education
• Any inherent limitations of the site characterization effort. Enterprises, Charlotte, November
318 23 The Conversion of Data to Useful Information

Fookes PG (1997) Geology for engineers: the geological model, prediction Yuhr L (1998) Managing data and its integrity in the ESC process.
and performance. Q J Eng Geol Hydrogeol 30:293–431 Symposium on the application of geophysics to environmental and
Fookes PG et al (eds) (2005) Geomorphology for engineers. Whittles engineering problems, EEGS, 22–26 March 1998, EEGS, Denver,
Publishing, Caithness Colorado
Hasan S et al (1988) Geology of greater Kansas City, Missouri and Kansas, Yuhr L et al (1996) Achieving a reasonable level of accuracy in site
United States of America. Bull Assoc Eng Geol 25(3):281–341 characterization in the presence of geologic uncertainty. In:
Schreuder WA (1997) On contouring surfaces. Principia Mathematica Shackelford CD et al (eds) Proceedings of uncertainty ‘96, uncer-
Inc. PM-TN-1997-1, Lakewood, Colorado tainty in the geologic environment: from theory to practice. ASCE
Schreuder WA (1999) Spatial interpretation and analysis of point data. Geotechnical Special Publication No. 58(1), pp 195–209, ASCE
Principia Mathematica Inc. PM-TN-1999-1, Lakewood, Colorado Reston, Virginia
 Risk Assessment
24

Abstract
Once the site characterization is complete, there is often the necessity of completing a risk
assessment. Here the focus is twofold. The risk of sinkhole or collapse associated with natu-
ral voids in soluble rock, abandoned mines, other open structures within the ground or the
vulnerability of groundwater in karst terrains. A brief introduction is provided to risk assess-
ment including objective and subjective approaches. Risk assessment can be of a regional
nature providing general trends or can be site-specific. A few examples are provided for
both.

24.1 Definition of Risk approaches for dealing with general geologic risk assess-
ment. Haneberg (2000) provides a mathematical approach
The terms hazard and risk are used in various ways and often using deterministic and probabilistic analysis in geotechnical
interchangeably leading to confusion. The following defini- engineering. Vick (2002) discusses both objective and sub-
tions are derived from Sowers (1996) and Waltham et al. jective approaches to risk assessment. We have presented a
(2005), who specifically address karst as well as Vick (2002) non-mathematical approach to risk assessment, which is
who provides an overview of risk for engineering purposes. simple in concept but can be difficult in practice. Our
approach is based upon having a complete and accurate site
Hazard – A condition with the potential for causing an unde- characterization in which to base any assessment of subsidence
sirable consequence. Hazard is the event itself such as risk or groundwater vulnerability.
subsidence or collapse. Its magnitude or size is a descrip- For further information there are a variety of books on
tion of its impact. For example, subsidence is generally geologic hazards that include: Ross (1984), Bolt et al.
slow and often occurs over a larger area while collapse (1977), Scheidegger (1975), Shuirman and Slosson (1992),
can be quite rapid and is generally localized. and Hunt (2007). Most of these books deal with earth-
Risk – A measure of the probability and severity of an quakes, landslides, or floods, only a few mention the haz-
adverse effect to health, property or the environment. ard of subsidence, collapse or that of groundwater
Risk is the combination of the likelihood or probability of vulnerability.
the hazard occurring and its magnitude or consequences.
Risk is usually described with terms such as unlikely,
probable or highly likely or by a numeric scale of 1–10, 24.2 Objective and Subjective Methods
with 10 being most likely. for Risk Assessment

There have been numerous approaches proposed for risk Risk assessments are typically approached in one of two
assessment. Beck (1991) had presented an approach for sink- ways, either objective (quantitative) or subjective (qualita-
hole risk assessment based upon available statistical data on tive). An objective approach relies on being able to quantify
sinkholes within the area. Engineers have proposed various variables. A subjective approach relies on the professional

© Springer Science+Business Media Dordrecht 2016 319

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_24
320 24 Risk Assessment

experience and the interpretation of conditions. Benson and ous geologic conditions and or historic trends from databases
Sharma (1994, 1995) have proposed: over an area of concern. For groundwater concerns a variety of
The assessment of risks in a quantitative manner is technically programs have been developed which superimposed the pres-
feasible for most man-made structures, because such structural ence of aquifers and their properties along with cultural factors
and engineered materials under a wide variety of circumstances to determine areas of groundwater vulnerability.
can be expressed through mathematical equations. Mathematical The following are some examples taken from work primar-
equations can range in complexity from simple algebraic equa-
tions to partial differential equations, but involve variables that ily in the State of Florida. However, other states or countries
can be analyzed through quantitative, statistical or stochastic dealing with karst issues have developed similar databases,
means. Consequently, the influence of each variable upon the regional risk maps, or software programs to evaluate risks.
engineering decision or action can often be identified as a quan-
titative risk factor that may then be expressed in terms of some
probability of occurrence.
However, assessment of risks for most natural phenomena, such 24.3.1 Sinkhole Databases
as the occurrence of sinkholes, is not tractable to systematic math-
ematical analysis in the same manner. While the broad range of Many states have developed databases on sinkhole occur-
factors that influence such phenomena are generally known, their
specific inter-relationships in a given site or region are often the rence and have maps showing sinkhole prone areas. These
topic of conjecture. While the resulting subjective reasoning behind data provide an indication of areas of sinkhole risk on a
assigning risk factors may well be grounded in sound practical regional basis and provides a means of identifying potential
experience along with site-specific observations and measurements, risk in relation to a specific area of concern. This information
they are not easily cast into mathematical expressions that then
allow risks, or even probabilities of occurrences, to be quantified is often used to develop the regional setting for a site-specific
through analysis. For these reasons risk assessments of subsidence investigation. This has been discussed in Chap. 13 and an
or collapse are commonly based upon subjective probabilities. example of this is provided in Chap. 27. The sinkholes occur-
ring within 1.6, 5 and 16 km from the site under investigation
Risk can be developed using theory and numerical calcula- were identified in a sinkhole database on the west coast of
tions (objective method). It is the process of using historical data Florida. Such data helps to set the stage for what might be
(past occurrences) to predict future. Examples of this type of expected at the site under investigation.
risk assessment include how the risk of earthquakes is assessed These databases typically include new sinkhole data as it
for a particular area at a particular magnitude. However, all too is reported but not the older historic sinkhole data. The historic
often, geologic hazards such as sinkhole collapse cannot be sinkholes provide the degree of maturity of the karst. The his-
accurately quantified with such statistical methods. If the hazard toric sinkhole data is obtained from a review of topographic
does not have a defined history of occurrence, if the geologic sheets and older aerial photography, which is commonly
conditions are modified by an engineered structure, or if local- available back to the 1930s (see Chap. 14). The presence of
ized anomalous geologic conditions are present but unknown, a both historic sinkholes as well as recent sinkholes is needed
subjective method of characterizing risk is needed. While these to identify the complete nature of sinkhole prone areas.
subjective methods are largely based upon the experience of the In 1995, Bill Wilson, then with the Florida Sinkhole
person assessing the risk, they must be based upon solid data Research Institute developed a method to describe sinkhole
and observations in order to support any conclusions. This is occurrence based upon their statewide database. His results
why a site characterization is needed for such efforts. were presented as the number of new sinkholes per km2 per
Vick (2002) points out that a subjective assessment of risk is year (NSH). This numeric evaluation provides an indication
equally as valid as an objective assessment of risk, and they are of the degree of active sinkhole development within an area
not mutually exclusive. Both the objective and subjective thereby improving the regional risk assessment based upon
approach to risk assessments provides a valid approach if based the data in a sinkhole database (Wilson 1995).
upon a reasonable amount of data and experience. These
approaches are often used in combination to develop a hybrid
approach. For example, where data for one variable is in abun- 24.3.2 Regional Sinkhole Risk Maps
dance and can be readily quantified while other variables have
limited data and rely more on professional judgment. Sinclair and Stewart (1985) developed one of the early sink-
hole risk maps for the state of Florida (Fig. 3.10). It was
developed based upon the type of overlying sediment (sand
24.3 Regional Risk Assessments versus clayey material) and its thickness. Based upon these
variables they categorized areas as follows:
Regional risk assessments have been developed for subsidence
and collapse as well as groundwater concerns. Subsidence and • Thinly covered limestone (<10 m of sediment) – sinkholes
sinkhole collapse risk assessments are made based upon vari- are few, shallow broad and develop gradually
24.3 Regional Risk Assessments 321

• 10–60 m of sand covered limestone – sinkholes are few, that would cause or trigger a collapse, the groundwater
shallow, small diameter and develop gradually vulnerability focuses on variables that would affect the trans-
• 10–60 m of clayey covered limestone – sinkholes are port of contaminants (thin soil cover, thick epikarst, high
numerous, vary in size, develop abruptly hydraulic conductivities). A number of approaches to assess
• Greater than >60 m of sediment covered limestone – sink- the regional groundwater vulnerability have been developed.
hole are very few, large diameter, and deep sinkholes do occur These are in essence risk maps for the potential of ground-
water contamination.
This approach provided a reasonably accurate risk assess- LeGrand (1983) developed a standardized system for evalu-
ment on a regional basis. ating areas suitable for waste disposal sites. Later, a group with
Gao and Alexander (2003) used GIS mapping to develop the National Water Well Association adapted the LeGrand
sinkhole probability map for southeastern Minnesota. They System for broader hydrologic settings and not just waste dis-
used a risk-based decision tree, which considered factors posal sites, called DRASTIC (Aller et al. 1987). This was a sys-
such as type of bedrock, depth to bedrock, sinkhole density, tematic means used to evaluate the groundwater pollution
and the distance to nearest sinkhole. Six zones of risk were potential for any hydrogeologic setting within the United States.
characterized: These were “screening tools” to be used with existing informa-
tion for the assessment of potential groundwater vulnerability to
• No Sinkhole Risk (non-carbonate areas). pollution based upon a calculated index. The method could be
• Low Risk (areas of carbonate rocks with no sinkholes applied to a wide range of geologic and hydrologic settings.
observed). DRASTIC recognized karst limestone settings, but did not uti-
• Low To Moderate Risk (widely scattered sinkholes or lize any site-specific details in its evaluation.
clusters of sinkholes). A modified version of DRASTIC was developed by the
• Moderate To High Risk (areas where sinkholes are a rou- Florida Geologic Survey to accommodate karst conditions
tine part of the landscape). and is used as an Aquifer Vulnerability Model referred to as
• High Risk (minimum distance to the nearest sinkhole is the Florida Aquifer Vulnerability Assessment (FAVA)
700 m and sinkholes are common with densities exceed- (Arthur et al. 2005). This is a GIS-based modeling technique.
ing 10/km2). The relative vulnerability is calculated based upon seven key
• Sinkhole Plains (where sinkholes are the dominant land- factors that focus on aquifer thickness and hydraulic conduc-
scape feature, minimum distance to the nearest sinkhole tivity along with effective karst features.
is 400 m and sinkholes are common with densities exceed-
ing 1/km2). Relative vulnerability = (Ts / Ks + Teg / Keg + Tconf / Kconf )
+ Kf
The primary controls on sinkhole development were the
type of bedrock geology and the thickness of sediment cover Ts – soil thickness
along with structure. A final risk factor was the nearest Teg – environmental geology (vadose) thickness
neighbor effect. The majority of sinkholes in southeastern Tconf – confinement thickness
Minnesota were found to form in concentrated zones. Gao Ks – soil conductivity (weighted average)
(2002) found that 95 % of the sinkholes in the Devonian Keg – environmental geology (vadose) hydraulic
karst, and 99 % of the sinkholes in the Galena-Maquoketa conductivity
karst, are less than 400 m away from their nearest neighbor. Kconf – confinement hydraulic conductivity
In the Prairie du Chien karst, sinkholes are less than 700 m Kf – effective karst features (the presence and proximity of
away from their nearest neighbor. closed depressions)
Both of these examples have taken existing data on key
variables and developed maps indicating what conditions As a tool, FAVA provides a scientifically defensible water
may be expected for a given location. The example from resources management and protection tool. It is based upon
Minnesota is also an excellent example of a hybrid approach pre-development data (natural systems with no man-made
to risk assessment effectively blending both objective and alterations). Results are presented as three levels of site clas-
subjective data. sifications, less vulnerable, vulnerable or, more vulnerable.
The purpose is to facilitate regional planning and develop-
ment to minimize adverse impacts on groundwater quality.
24.3.3 Groundwater Vulnerability in Karst Another method called EPIK has been developed to
define protection zones for karst aquifers in Switzerland
Karst terrains provide some of the more complex settings for (Doerfliger and Zwahlen 1995). Groundwater vulnerability
evaluating groundwater, its flow and potential for contamina- maps based on this method can provide guidance for
tion. Where the risk of sinkholes focuses on the variables establishing groundwater protection zones within karst
322 24 Risk Assessment

catchment areas. The EPIK vulnerability rating is based on Triggering Mechanisms

Void Space
four hydrogeological parameters: (spatial and temporal)

1. E: The presence and character of the epikarst conditions,

2. P: The presence and character of the protective soil cover, Size, Type, Depth, Natural Man-Made
3. I: The rate of infiltration conditions and Distribution, etc.
4. K: The level of development of the karst network including
the extent and interconnectability of the conduit system.

These parameters are used to assess specific aspects of the

karst flow regime. The EPIK method is implemented in three
stages. A semi-quantitative evaluation and field mapping of Risk Assessment
the four parameters is completed. Calculation of a protection
index is completed by combining and weighting of the values Fig. 24.1 Key factors in determining risk of subsidence or collapse
of the four parameters for each unit area in the catchment.
The cartographic representation is developed of the distribu-
tion of the protection index for the entire catchment. Benson and Hatheway 2001). It has also been used to assess
mine subsidence and collapse (Hatheway et al. 2000).
In concept you must have two conditions in order to have
24.3.4 Advantages and Limitations subsidence or collapse potential, a void space for material to
of Regional Methods move into and a triggering mechanism to precipitate the
movement of the material (Fig. 24.1). The natural dissolution
All of these regional risk assessments utilize a group of key of limestone is a slow process, occurring at a rate of about
variables that are evaluated individually then correlated spa- 1–3 cm/1,000 years. As such, further natural dissolution of
tially in order to assign a level of risk or vulnerability. These rock by itself is not generally of concern regarding the life of
regional risk assessments have been used effectively for engineered structures unless acid spills or leaks are present.
regional planning purposes, for insurance actuaries and even The concern then lies with the existing void space within the
groundwater resource protection. In some cases, regional rock. Once the void space is identified, the possible trigger-
data can be used to improve the design of critical structures ing mechanisms must be identified. The triggering mecha-
and infrastructure and determine set back distances from nisms can be natural, man induced or a combination of both.
sinkholes, protected water bodies or wellfields. Regional Without these two conditions (a void space and a triggering
data, risk maps or vulnerability maps can also be useful as a mechanism), subsidence or collapse would not necessarily
starting point for a site-specific risk assessment. be possible.
The strategy is summarized in the flow diagram
(Fig. 24.2). A strong base of site-specific data must support
24.4 Site-Specific Risk Assessment the development of a risk assessment and begins with a site
characterization. Therefore, with the completion of a site
Risk assessment at the regional scale primarily utilizes his- characterization the basic geologic and hydrologic data
torical data to predict future occurrences. A risk assessment should be available to support the risk assessment.
for subsidence or collapse at the site-specific scale requires From the site characterization effort we should have iden-
that more details be considered which impact the occurrence tified the types of sinkholes present (Fig. 3.1), the level of
of subsidence or collapse. Therefore, when a site-specific risk maturity at the site (Fig. 4.1) and the size or sizes of the sink-
assessment for subsidence or collapse is required, it should be holes that might develop at the site (Fig. 10.3). This informa-
an integral part of the site characterization process. tion should have been refined to be very site-specific. We
The risk assessment strategy summarized by Benson et al. should have the information in order to determine where sub-
(2003) is based upon having completed a site characteriza- sidence or collapse may occur.
tion and having a good understanding of site conditions
along with a solid base of supporting data. However, a risk • The presence of geologic and hydrologic conditions at a
assessment is ultimately often very subjective and largely site, which may be susceptible to subsidence or collapse.
based upon professional experience. The authors have suc- Where is the void space, how big is it and could it accept
cessfully applied the strategy for site-specific risk assess- large quantities of materials? See Chap. 10 for the con-
ment for the past few decades in areas of collapse, subsidence ceptual models of cover collapse sinkholes.
and paleokarst (Benson et al. 2002, 2003; Benson 2001;
24.4 Site-Specific Risk Assessment 323

DATA ACTION

Regional Geology
Identify Potential Karst
Cave/Sinkhole Databases Related Problems from
Regional Data
Regional and Existing
Boring/Well Data

Typical Site Characterization Efforts

Aerial Photography Develop Initial
Conceptual Model of
Site Walkover & Observations On-site Karst Conditions

Surface Geophysical Data

Minimally-Invasive
Push Technology

Borings, Sampling
and Trenches
Identify Site-Specific
Geophysical Logging, Karst Features Using
Downhole and Hole-to-Hole Appropriate & Adequate
Measurements Data

Engineering Data

Hydrogeologic Data Integrate Data and

Refine Conceptual Model
Dye Traces

History of Subsidence
or Collapse

Typical Subsidence or Collapse

Surface Water Movement

Groundwater Usage Risk Assessment Efforts

Identify Site-Specific
Climate Conditions
Mechanisms of Collapse
Other Site-Specific Factors
such as cultural development,
etc. Refine Conceptual Model
and Integrate Data
Current Site Use

Potential Future Site Use Assign a Risk Level for

Entire Site and/or
Specific Areas

Fig. 24.2 A strategy for risk assessment in karst

• The spatial and temporal conditions at a site, which are gering mechanisms that may cause subsidence or collapse.
likely to trigger subsidence or collapse (with a focus upon These issues may not have been a part of the site character-
surface water flow and associated hydrologic conditions). ization and are critical to a complete assessment of risk.
The completeness and accuracy of a risk assessment is a
However, additional site-specific data will commonly be function of our ability to identify and understand all of
necessary to provide the additional details regarding the trig- the risk factors that may cause subsidence or collapse at a
324 24 Risk Assessment

particular site. This may include current and potential future • Low: surface subsidence or collapse is not likely to occur
use of the site such as development, construction, remedia- over lifetime of project.
tion, etc. that may change conditions and trigger subsidence • No Risk
or collapse.
Sinkholes do occur naturally without any man-made trig- Additional examples of risk assessment are included with
gering mechanisms. However, in culturally developed areas, it each of the three case histories in Part III. There will be situ-
is usually the interaction of man-made changes in site condi- ations where some issues may not be clearly resolved by
tions that increase risk and trigger subsidence or collapse (see data. In these cases, subjective judgment and opinions will
Chap. 8 for a discussion of triggering mechanisms). Conditions be supported by limited data. The examples provided within
which are likely to trigger subsidence or collapse include areas each of the case histories include characterization of paleo-
of concentrated surface water runoff (large roof areas, paved collapse features with a strong set of data. However, the
parking lots, storm water basins, etc.), heavy rainfall or assessment of risk of future subsidence or collapse related to
drought conditions, drilling, construction, and blasting, etc. or these deep paleocollapse are supported by more limited data
even broken utility lines (sewers or water lines). In some cases and subjective opinion.
the triggering factors can be compounded, where two or more Attempts to further quantify, model risks or predict time
factors may interact to accelerate collapse. For example, heavy of collapse are generally inappropriate since detailed tempo-
rains after a period of drought can often trigger soil raveling ral data are commonly unavailable. However, if a site is
and ultimate collapse. It is the spatial distribution, interaction properly instrumented and monitored, warnings of collapse
and correlation of all these factors, which will allow us to may be reasonably estimated based upon a variety of data.
identify an area as having a higher or lower risk. Subtle precursors can often provide warnings of an eminent
Assessment of risk should be relatively straight forward, collapse. Examples of sites with monitoring instrumentation
in concept, if appropriate and adequate data has been include:
acquired during our site characterization. This would include
identifying the site-specific potential for weak rock or void • South Africa’s use of telescoping bench marks
spaces and triggering mechanisms. Then and only then can a • Time domain reflectometry monitoring a road in Florida
realistic estimate of subsidence or collapse risk can be made • Tilt meters in boreholes monitoring a brine cavern in
for site-specific conditions. While the concept is simple, its Carlsbad, New Mexico
execution requires extensive experience and extreme atten-
tion to details. Further discussion of these monitoring efforts is provided
Risk assessments are commonly based upon subjective in Chap. 20, “Engineering Measurements and Monitoring”.
interpretations. The dilemma faced in subjective risk assess-
ment is that there are often bold statements made (opinions)
promoting either low or high risk with a minimum amount of 24.4.1 Examples of Site-Specific Risk
site-specific data to back them up. In that case, the level of Assessment
confidence in the risk assessment is more or less based upon
the qualifications of the person or persons providing the risk Sowers (1996) offered a strategy for reducing risk for the
assessment. Therefore, a risk analysis must clearly identify foundation design of critical structures (hospitals, police and
all interpretations and assumptions along with supporting fire departments) in karst. “It may be prudent to design foun-
data upon which it is based. This supporting data minimizes dations to tolerate a large dome collapse, which is assumed
concerns of professional integrity and increases confidence to occur at any random location beneath the structure. Since,
levels. However, it should also be recognized that while the even in sinkhole prone areas simultaneous, closely spaced
risk of sinkholes may be minimized by a proper site charac- dropouts are rare. Therefore, a design for a one occurrence at
terization along with management of triggering mechanisms, a time at a random location beneath the structure is a reason-
it is not likely to be totally eliminated. able presumption”. This strategy is only effective if the
Subjective risk predictions for subsidence or collapse nature of the event can be estimated in terms of its size so
must be made inherently general and limited, using terms that an adequate design can be made.
such as: The following examples illustrate approaches to site-
specific risk assessment. The factors used to assess the risk
• Very High; surface subsidence or collapse is underway or vary and are site-specific.
is eminent
• High: surface subsidence or collapse is actively occurring 24.4.1.1 Highway I-70 Near Frederick, Maryland
or is very likely to occur; A risk assessment was made along a section of Highway I-70
• Medium: surface subsidence or collapse may occur if fur- near Frederick, Maryland (Zhou et al. 2003). The spatial dis-
ther aggravated by on-site mechanisms of collapse; tribution and density of 138 sinkholes along 8 km2 of road-
24.4 Site-Specific Risk Assessment 325

way were used to develop a risk assessment model. More posed upon the gravity data were two localized microgravity
than 70 % of the sinkholes were within 300 m of the high- anomalies of 300–500 microgals. These two anomalies indi-
way. Eight factors were considered including: cated significant mass deficits in the limestone along the
highway, likely due to the presence of cavities and caves.
• Topography, Therefore, these areas have a high level of risk for sinkholes
• Proximity to topographic depressions, to occur. The areas of mass deficit were modeled and super-
• Proximity to existing sinkholes, imposed on the geologic cross section (Fig. 24.4b). The
• Rock type, westernmost anomaly was estimated to be a lower density
• Soil type, zone of about 180 m wide and 75 m deep. The easternmost
• The presence of geophysical anomalies, anomaly was estimated to be a lower density zone of about
• Proximity to geologic structure and 450 m wide and 30–45 m deep.
• Thickness of overburden. One boring was drilled within each of these anomalies
along with a corresponding borings within adjacent back-
A scoring system was then used to predict the occurrence ground conditions. These borings confirmed our interpreta-
of new sinkholes along the highway itself. A relative rating tion. The borings within the anomalies showed a distinct
of 1–7 was developed for every 30 m segment of the highway increase in the thickness of highly weathered and fractured
with higher values indicating a relatively higher risk of new rock as well as in the number of voids (Fig. 24.4b). A few
sinkholes occurring. Where multiple variables indicated risk, months after completion of field work a large 30 m diameter
the risk was higher. Where fewer variables indicated risk, the and 18 m deep sinkhole occurred south of the roadway
risk was lower. within the western most microgravity anomaly (Fig. 24.3).
During an early assessment of this site by the authors in This is an example of how site-specific data can be used to
1996, sinkhole activity had been observed along highway improve site-specific risk assessment (Benson et al. 1997).
I-70 and to the north about 150 m (Fig. 24.3). As part of our
investigation, a microgravity survey had been carried out 24.4.1.2 Spatial Distribution of Sinkholes
along the highway. The microgravity data clearly indicated Both Gao and Alexander (2003) and Zhou et al. (2003) have
the boundary between the Lime Kiln and the Grove commented upon the concentration or proximity of sink-
Formations (up to 800 microgals) due to a density contrast holes. Zhou et al. (2003) suggested that new sinkholes tend
between the formations (Fig. 24.4a). In addition, superim- to develop in the vicinity of previous ones because geologic

West
Areas Surveyed N East

269+00 249+00 229+00

209+00
I-70 Westbound
195+00
I-70 Eastbound
Recent
Sinkhole
Collapse Grove Limestone
Lime Kiln
Limestone Lime Kiln
Quarry Limestone

Area of Sinkhole Activity

Projected Lime Kiln/Grove Contact
Gravity Anomalies
0 150 300
Scale in Meters

Fig. 24.3 Map of study area along I-70 in Maryland

326 24 Risk Assessment

Boundaries of the
a Grove Formation
WEST EAST
Microgravity (microGals)

0
Two Large Gravity Anomalies
-200

-400

-600
Lime Kiln Lime Kiln
-800

-1000
26500 25500 24500 23500 22500 21500 20500 19500
Station Location (feet)
Approx. 2300 meters

b Cavity Zones
Location of recent modeled from gravity data
sinkhole collapse
Soil
0

30

60
Approximate Depth in Meters

90 Sandy
Layer

120

Grove Limestone
150
Boring Data
Overburden Soils

180 Highly Weathered Rock

Lime Kiln
Weathered/Fractured Rock
Limestone Unweathered Rock
Voids
210

Fig. 24.4 Microgravity data, geologic model and boring results along I-70 in Maryland (a) Microgravity data acquired along I-70 (b) Modeled
microgravity data and boring results

and hydrologic conditions are favorable for sinkhole col- This indicated that the risk of additional sinkholes is very
lapse in an area. Risk assessments have often looked at the high within 30 m of an existing sinkhole.
spatial proximity of sinkholes and have deduced that there is This example illustrates that there is a spatial relationship
a higher risk in closer proximity to existing sinkhole activity. to sinkhole density. However, looking at the spatial distribu-
Analysis of the sinkhole collapse along Interstate-70 near tion of sinkholes without consideration of geologic condi-
Fredrick, Maryland indicated that the radius of influence of a tions such as the depth of origin and the nature of the void
sinkhole was approximately 30 m at that site (Zhou et al. 2003). space within the rock does not provide a complete picture.
24.4 Site-Specific Risk Assessment 327

a b

Within the uppermost rock

Deep within the massive rock

Numerous isolated shallow A deep conduit system
cavities are found in the epikarst with a localized cavern are
or the top of rock at intervals of a often found at intervals of a
meter to a few meters or so few 100 meters or more

Fig. 24.5 Conceptual models of sinkhole development from epikarst or top of shallow rock are more pervasive (a), while large caverns formed
deeper within the massive rock are more widely spaced (b)

This concept uses two of the five simplified conceptual 24.4.1.3 EPA Superfund Site
models presented in Chap. 10. For example, a simplified An extensive site characterization was completed at the
conceptual model of cover collapse sinkhole development Superfund site in Tarpon Springs, Florida (Part III, Chap.
that incorporated the sinkhole size and spacing includes: 27). This site was a former phosphate ore processing area.
The site is in an area of historic and current sinkhole activity.
• Smaller sinkholes tend to originate from the epikarst or As part of this project, a risk assessment was completed at
top of rock where many smaller isolated voids typically the site. The risk assessment was to address the site in gen-
occur (Fig. 24.5a). This will result in the spacing between eral but also specifically address the planned remediation
sinkholes to be relatively close (on the order of a few and its impact on potential sinkhole development.
meters to a few tens of meters) and their locations are The site characterization was completed at an unusual level
more random. of detail with shallow geologic data density approaching
• Very large sinkholes are generally associated with the 100 %. As a result we knew the shallow geology and hydrol-
presence of a large cavern. A cavern will typically develop ogy at the site extremely well. This level of data also provided
at weak points such as at the intersection of major fracture a high level of confidence in which to base a risk assessment.
systems or at a critical structural feature (Fig. 24.5b). We were able to determine that there were three zones in
These sinkholes will typically occur at larger intervals of which dissolution had occurred and where void space or
on the order of a few 100 m or so and result in very large caves had developed (Fig. 24.7).
sinkholes.
• Zone 1: shallow, small voids occur within the top of the
These conceptual models (Fig. 24.5) are very simplified Tampa Limestone just below a semi-confining layer, gen-
models of two major factors (depth to rock and nature of the erally less than 7.5 m deep.
void space), which control the size and spatial distribution of • Zone 2: intermediate voids or conduits at the unconfor-
sinkholes. This information can play a critical role in assessing mity between the Tampa Limestone and the Suwannee
risk. Such data may provide technical background for deter- Limestone at a depth of about 18 m.
mination of setbacks from very large sinkholes or for engi- • Zone 3: deeper conduits or caverns had developed at a
neering measures to improve the site conditions (Fig. 24.6). depth of about 55–90 m at the lower third to half of the
328 24 Risk Assessment

Set
Back

Low Risk of Low Risk of

Very Large
Another Very Another Very
Large Sinkhole Sinkhole Large Sinkhole

Plan view of a very large sinkhole

b
Low Risk of Another Very Large Sinkhole

Large cavern
collapse

Cross section of a very large sinkhole

Fig. 24.6 Since large sinkholes are typically controlled by major frac- tively low (a) Plan view of a very large sinkhole (b) Cross section of a
ture system structure within rock they occur further apart, therefore risk very large sinkhole
near the large sinkhole assuming a reasonable set back may be rela-

Suwannee Limestone (this was the probable source of the Zone 1

large paleocollapse features on-site). All remediation efforts were expected to be fairly shallow,
focused in the sands of the surficial aquifer, above Zone 1.
Groundwater levels at the site are controlled by the adja- No deeper excavation, drilling, etc. would or should be tak-
cent river and the Gulf of Mexico and remain relatively con- ing place in or below the semi-confining layer that would
stant. There is only a small difference in head between the impact sinkhole development. However, triggering mecha-
surficial and Floridan aquifers. Tidal changes are the most nisms for sinkhole development associated with remediation
significant factor in changing groundwater levels and pri- and construction include:
marily influence the Floridan Aquifer. There is no excessive
pumping of groundwater from either the surficial or Floridan • Breaching of the semi-confining layer by excavation or
aquifer within the property. The risk of sinkhole collapse at drilling
this site is very low, if the site were left by itself, unaffected • Creating an artificially higher head of water in the surficial
by further activity. aquifer which might tend to breach the semi-confining
layer
References 329

Zone 1 Zone 2 Zone 3

Shallow, Small Deeper, Re-activation of
Collapse Larger Collapse Paleocollapse

Risk of Very Low, if SCL Virtually Zero Virtually Zero

Occurrence is not breached

Surficial
Sands
SCL
Tampa
Small isolated voids Tampa
Limestone
in epikarst at top of Limestone
rock, typically <7.5 m
Suwannee
Voids at unconformity Limestone
about 18 m deep

not to scale
Porous Zones and Caves
in the lower half to third of
Suwannee Limestone
(est. 55 to 90 m deep)
Ocala Limestone

Fig. 24.7 Risk assessment at EPA Superfund site showing void development at three depths and their associated risk of sinkhole development or
reactivation

• Over pressuring the Floridan aquifer by excessive drilling them including loading from material piles, vibration due to
pressures when drilling into the Tampa Limestone, which trains and local traffic of heavy trucks. As a result, we con-
might breach the semi-confining layer from below. cluded that the risk of subsidence or collapse due to reactiva-
tion of these features at their depths of origin is considered
Any of these actions could cause a sinkhole to develop by virtually zero.
breaching the semi-confining layer and allowing sands to
ravel into the voids within the upper Tampa Limestone.
However, this worse-case scenario would likely create a References
small (3–6 m in diameter) localized sinkhole.
Aller L et al (1987) DRASTIC: a standardized system for evaluating
ground water pollution potential using hydrogeologic settings. US
Zone 2
Environmental Protection Agency, Robert S Kerr Environmental
The probability of development of small shallow cover sub- Research Laboratory, Office of Research and Development
sidence sinkholes originating from 18 m deep (Zone 2) is EPA/600/2-87/035
considered to be virtually zero. Arthur J et al (2005) Florida aquifer vulnerability assessment (FAVA):
contamination potential of Florida’s principal aquifer systems.
Florida Geological Survey Report for FDEP 21 Mar
Zone 3 Beck BF (1991) On calculating the risk of sinkhole collapse. In: Kastning
Approximately 8 % of the site (in two areas) contained EH, Kastning KM (eds) Proceedings of the Appalachian karst sympo-
paleocollapse features, which originated from a depth of sium. National Speleological Society, Radford, pp 231–236
Benson RC (2001) Strategies of site characterization and risk manage-
55–90 m bls (the dissolution at these depths occurred at
ment presented to National Research Council, Committee on coal
lower sea level stands). The depth of origin is well beyond waste impoundment failures subcommittee on impoundment site
the depths of any remediation efforts. These two areas of the characterization, Washington, DC
site were mapped in detail and were dated to 48,000 years Benson RC, Hatheway AW (2001) A systematic approach for evaluat-
ing subsidence risk. AEG/AIPG Annual Meeting St. Louis
ago using C-14 methods. These features were considered
Benson RC, Sharma D (1994) The pitfalls in site characterization and
stable since there was no surface expression (topographic how to avoid them. Short course presented by Environmental
low), 40 years of facility operations occurring over or near Education Enterprises, Salt Lake City
330 24 Risk Assessment

Benson RC, Sharma D (1995) Pitfalls of site characterization and how Hunt RE (2007) Geologic hazards, a field guide for geotechnical
to avoid them. Short course presented by Environmental Education engineers. CRC Press, New York
Enterprises, Charlotte, November LeGrand HE (1983) A standardized system for evaluating waste-
Benson RC et al (1997) Assessment and prediction of karst conditions. disposal sites. National Water Well Association, Worthington
Presented at the Federal Highway Administration Southeastern Ross SS (1984) Construction disasters, design failures, causes, and pre-
Regional Meeting, 27–29 Oct 1997, Chattanooga vention, Engineering news record series. McGraw-Hill, Columbus
Benson RC et al (2002) Assessing the risk of karst subsidence AEG/ Scheidegger AE (1975) Physical aspects of natural catastrophes.
AIPG Joint Annual Meeting Reno, pp 22–29 Elsevier, New York
Benson RC et al (2003) Assessing the risk of karst subsidence and col- Shuirman G, Slosson JE (1992) Forensic engineering, environmental
lapse. In: Barry BF (ed) Proceedings of the 9th multidisciplinary case histories for civil engineers and geologists. Academic, San
conference on sinkholes and the engineering and environmental Diego
impacts of karst, Huntsville, ASCE Geo-Institute Geotechnical Sinclair WC, Stewart JW (1985) Sinkhole type, development and distri-
Special Publication No. 122, ASCE Reston, Virginia, pp 31–39 bution in Florida, Florida Geologic Survey, Map Series #110,
Bolt BA et al (1977) Geologic hazards, 2nd edn. Springer, New York Tallahassee
Doerfliger N, Zwahlen F (1995) EPIK: a new method for outlining of Sowers GF (1996) Building on sinkholes, design and construction of
protection areas in karst environment. In: Gunay G, Johnson I (eds) foundations in karst terrain. ASCE Press, Reston (out of print)
Proceedings of the 5th International symposium and field seminar Vick SG (2002) Degrees of belief subjective probability and engineer-
on karst waters and environmental impacts, Antalya, September. ing judgment. ASCE Press, Reston
Balkema, Rotterdam, pp 117–123 Waltham AC et al (2005) Sinkholes and subsidence, karst and cavern-
Gao Y (2002) Karst feature distribution in southern Minnesota: extend- ous rocks in engineering and construction. Springer-Praxis,
ing GIS-based database for spatial analysis and resource manage- Chichester
ment. Ph.D. thesis, University of Minnesota Wilson WL (1995) Sinkhole and buried sinkhole densities and new
Gao Y, Alexander EC Jr (2003) A mathematical model for a map of sinkhole frequencies in karst of northwest Peninsular Florida. In:
relative sinkhole risk in Fillmore County, Minnesota. In: Barry BF Beck BF (ed) Proceedings of the 5th multidisciplinary conference
(ed) Proceedings of the 9th multidisciplinary conference on sink- on sinkholes and the engineering and environmental impacts of
holes and the engineering and environmental impacts of karst, karst: karst geohazards, Gatlinburg, Tennessee. Balkema,
Huntsville, Alabama. ASCE Geo-Institute Geotechnical Special Rotterdam, pp 79–91
Publication No. 122, ASCE Reston, Virginia, pp 439–449 Zhou W et al (2003) Sinkhole risk assessment along Highway I-70
Haneberg WC (2000) Deterministic and probabilistic approaches to near Frederick, Maryland. In: Barry BF (ed) Proceedings of the
geologic hazard assessment. Environ Eng Geosci 6(3):209–226 9th multidisciplinary conference on sinkholes and the engineer-
Hatheway AW et al (2000) Geotechnical risk assessment of Route 169 ing and environmental impacts of karst. ASCE Geo-Institute
over the Briarcliff Mine, Kansas City, Clay County, Missouri. Geotechnical Special Publication No. 122, ASCE Reston,
Consultant Report to Kansas Dept. of Transportation Virginia, pp 591–601
 Part III
Case Histories

Most site characterizations are carried out over a long period of time with a variety of objectives
using different contractors, consultants and even owners. Even with the best of intentions on the
part of owners and consultants it is difficult to complete such a site characterization in a mean-
ingful way and to achieve an accurate understanding of site conditions. Such an approach results
in a marginal site characterization effort at best.
The three case histories included here are those in which the authors have had complete
cooperation of the owner throughout the site characterizations process. These case histories
represent a range of conditions including the total time of the project, the strategy used, and
the methods employed. Most important to the project was the fact that the senior staff was
intimately involved from the beginning to the end of the project. This continuity has enabled
us to achieve a high level of confidence in the site characterization effort for each of these
three projects.
 The Development of a Landfill over
an Abandoned Limestone Mine 25

Abstract
The first case history deals with the expansion of a landfill over a 56 ha abandoned room
and pillar limestone mine in Kansas City area. A massive 7 ha mine-roof collapse had
occurred in the mine and fissures were discovered on the surface. The State of Kansas
required that the mine be backfilled to avoid any surface subsidence. This project did not
follow our standard site characterization strategy. Work at this site began with the single
task of using a borehole video camera to evaluate the extent of mine collapse and the effec-
tiveness of fly ash being used to backfill the mine. The project slowly evolved into a sub-
stantial site characterization effort that included an assessment of mine collapse, monitoring
mine-roof collapse conditions over time, a groundwater monitoring plan for the landfill and
a subsidence risk assessment along with many other tasks. The project was carried out over
a period of 7 years from 1987 to 1993 with minor efforts in mine backfilling continuing
through 2003. About half of the project effort was completed underground in the mine.

25.1 Background of pleistocene-aged loess overlays much of the site. The mine
operated from 1960 to 1980. Early mining began at the portal
There has been more than a century of extensive under- (Fig. 25.1) and extended eastward about 425 m, then mining
ground mining of the limestone within Greater Kansas City extended west of the portal for about 730 m. The Bethany Falls
(Kansas and Missouri) area. The Bethany Falls Limestone limestone was mined to height of 4.2 m. The mine occupies an
provides an excellent limestone and is typically a one level area of approximately 56 ha of which the westernmost 34 ha
room-and-pillar mine within the relatively flat-lying uniform would underlie the proposed landfill (Fig. 25.1).
geology. Kansas City limestone mines are almost always dry A major mine-roof collapse occurred in the western portion
and stable giving the city the distinction of being number one of the mine in the early 1970s. This collapse encompassed an
in the world in terms of human use and occupancy of under- area of up to 7 ha, which is referred to as the Central Collapse
ground space after mining. More than 200 businesses are Area (CCA) (Fig. 25.3). The collapse is reported to have notice-
located underground occupying more than 484 ha as of 1983 ably shook the ground, broke windows in the trucks within the
(Hasan et al. 1988). In some cases, commercial buildings are mine and was recorded on seismographs in Lawrence, Kansas
developed over abandoned mines. State agencies now require some 40 km away. After this large mine-roof failure of the
that mines be backfilled prior to developing facilities over CCA, inspection revealed a number of fissures within the loess
them to avoid possible subsidence. at the surface (Fig. 25.3). The fissures occurred in a circular
The Tobin limestone mine is located in Wyandotte County, pattern within in the thick loess sediment and were centered
Kansas City, Kansas, west of highway 635 and north of inter- over the CCA in the mine. The conclusion by the owner and a
state I-70 just north of the Kansas River (Fig. 25.1). The mine is local consultant was that the mine-roof failure had resulted in
overlain by a stratigraphic sequence of about 52 m of alternating the surface fissures. This was a reasonable conclusion based
shale and limestone (Pennsylvanian age) (Fig. 25.2). Up to 18 m upon limited information available at the time.

© Springer Science+Business Media Dordrecht 2016 333

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_25
334 25 The Development of a Landfill over an Abandoned Limestone Mine

Site
Location

N Kansas

Mine Boundary
(dashed outline)

Medical facilities
over eastern end
of the Tobin Mine
Mine Portal

I-70

Extent of Landfill
Kansa Development
s Rive (white solid outline)
r

0 365

Approximate Scale in Meters

Fig. 25.1 A 1987 aerial photo shows the site just north of the Kansas River and the initial landfill can be seen in the northern most portion of the
mine

Prior to the collapse of the CCA, pillar size and spacing They concluded that failure of the pillars would be very
were not uniform and distances were determined by pacing. unlikely even at 90 % extraction ratios.
After the CCA occurred, mining operations west of the CCA In March of 1987, a localized but significant roof-fall of
began using a uniform pillar layout of pillars of about 7.6 by about 0.12 ha occurred about 60 m from the west portal
7.6 m on 18 m centers (Nicholson 1988), an extraction ratio entrance. The owners considered the mine unsafe and further
of about 82 %. access to the mine was denied based upon recommendations
As the mining ended, the owner at the time had started a from their attorneys. The authors were then engaged to
landfill over the mine. As a requirement of the Kansas State observe general conditions within the mine and evaluate the
permitting process, the mine had to be backfilled to prevent extent of fly-ash backfill using a downhole video camera.
possible surface subsidence. Initially, fly-ash from local Observations by borehole camera were a very slow, incom-
power companies was being used as backfill material. The plete, and a tedious process.
efforts included dry fly-ash, cast in place columns of fly ash In June 1987, the authors gained physical access into the
and a mixture of fly ash that was mixed in a cement truck and mine and the formal site characterization process began to
dumped into the mine via 20 cm uncased boreholes. characterize mine conditions and assess subsidence risk.
In July of 1986, an engineering firm’s mining staff made This work extended over a period of 7 years from 1987 to
a 1-day inspection of mine conditions. It was concluded, “the 1993. Minor efforts continued through 2003 to support mine
mine workings and overlaying rock are stable and not likely backfilling efforts. As time passed other tasks associated
to cause significant movement or fracturing at the ground with the project were included such as developing and
surface” (Golder Assoc. 1986). They evaluated the pillar sta- installing a groundwater monitoring plan for the landfill and
bility in three ways: providing quality control monitoring of the mine backfilling
process. Work included hundreds of hours in the mine and
• Estimated overburden stress of the rock and proposed endless hours of observations and measurements at the
landfill. surface.
• Pillar strength based upon compressive strength of 82.7– Figure 25.4 summarizes the main tasks for site character-
137.9 MPa for the Bethany Falls Limestone ization and monitoring associated with this project. Many of
• Possible pillar punching into the weaker Hushpuckney the tasks were carried out over time and interacted with one
Shale below the mine floor. another.
25.1 Background 335

Loess

Argentine-Frisbee
Limestone

Lane Shale
Raytown-Paola
Limestone

Approximately 52 m
Chanute Shale

Drum Limestone
Quivira Shale
Westerville
Limestone

Wea Shale

West East Block Limestone

Portal Fontana Shale
Portal

Winterset Limestone

Stark Shale
Galesburg Shale
Bethaney Falls
Limestone (mined layer)
Hushpuckney Shale
Middle Creek Limestone
Center Portal Ladore Shale
with sump and pump Note: contacts are estimated visually

Fig. 25.2 The geologic section along with a photo of the mine portal entrances

N
Mine Boundary
Boundary of (limit of rock extraction)
Mine Roof
Collapse
(the CCA)

Surface Fissures
in Loess Soil

0 120
Scale in Meters

Fig. 25.3 The mine boundary along with the extent of the central collapse area (CCA) and location of the surface fissures shown in dashed lines
(a). The photo shows a typical surface fissure (b)
336 25 The Development of a Landfill over an Abandoned Limestone Mine

Preliminary
Conceptual
Model
An Assessment of the CCA
and the Surface Fissures
A Subsidence Risk Assessment

An Assessment of the
Mine Conditions
Mine Backfilling Program
Final
Groundwater Monitoring Plan Conceptual
Model

Fig. 25.4 The work tasks at this site included a wide range of activities over a 7 year period. These tasks supported the development of a concep-
tual model of site conditions along with a subsidence risk assessment

25.2 An Assessment of the CCA The Bethany Falls Limestone is considered the best quality
and the Surface Fissures limestone in the Kansas City area and is found in thickness
of 4.2–6.4 m (Fig. 25.2). Two of the limestone strata above the
The initial site characterization effort focused upon develop- mine are significant because of their thickness. The Winterset
ing an understanding of the mine-roof collapse and the rela- Limestone is about 8.8 m thick and is found about 2.4–3 m
tionship between the CCA and the surface fissures. The tasks above the Bethany Falls Limestone. The Westerville Limestone
completed to meet these objectives included: is massive, about 6 m thick, and found about 18 m above the
Bethany Falls Limestone (Hasan et al. 1988).
• A review of regional and local geology and Two geologic aspects are known to impact the mine sta-
geomorphology bility in the area (Hasan et al. 1988). The first is the presence
• Mapping of the surface fissures and selected trenching of an anomalous zone of rock in the upper part of the Bethany
• Developing a preliminary conceptual model of the CCA Falls Limestone known locally as the “Rubble Zone”,
and their relation to the surface fissures “Buckshot”, or “Peanut Rock” by the local limestone min-
• An aerial photo analysis ers. The second is the Hushpuckney Shale, which lies imme-
• Site-specific geology (field mapping, limited drilling, and diately beneath the Bethany Falls Limestone and contains
geophysical logging all new and existing boreholes) sulfides which can lead to swelling and floor heave.
• Confirmation of the preliminary conceptual model The “Rubble Zone” (Fig. 25.5) is a carbonate-nodular,
and highly over-consolidated clay which is friable and typi-
cally 0.45–1 m thick. This zone appears to occur as channels
that are much thicker at some locations and inadvertently
25.2.1 Review of Regional and Local Geology exposed by mining operations. The rubble zone is known to
deteriorate rapidly when exposed due to increased moisture,
The geology of the area is well known and has been sum- leading to localized roof failure (Hasan et al. 1988). This is
marized in “Geology of Greater Kansas City” (Hasan et al. the initial mechanism of most roof-failure of mines in the
1988), which also provides an excellent insight and overview Kansas City area.
of the limestone mines in the area along with a discussion on The Hushpuckney Shale occurs below the Bethany Falls
the commercial use of underground space. The subsurface Limestone (Fig. 25.5) and contains 5–6 % (by volume) sulfides
strata are relatively flat-laying and uniform. The in the form of pyrite, sphalerite and chalcopyrite. Conversion
Pennsylvanian age bedrock overlaying the mine is made up of sulfide minerals into gypsum results in a volume increase
of an alternating sequence of limestone and shale, about of six to eight times, generating high stresses resulting in
52 m thick that is covered by a thick loess of up to 18 m or swelling and floor heave. Noticeable heave of the floor can
more (Fig. 25.2). The deeper Mississippian limestone and occur between 2 and 5 years after extraction of the Bethany
dolomite beds lie disconformably below the Pennsylvanian Falls Limestone and total amount of heave (ranging from 2.5 to
bedrock, and are over 120 m thick (Gentile 1984). 20 cm), may occur over a 10 year period (Hasan et al. 1988).
25.2 An Assessment of the CCA and the Surface Fissures 337

Winterset Limestone b

Starke Shale
Galesburg Shale
Rubble Rock

Bethany Zone of Mined

Falls Rock, Typically c
Limestone 4.2 m thick

Hushpuckney Shale
Middle Creek Limestone

Upper Ladore Shale

Fig. 25.5 The two geologic conditions affecting mine stability include the rubble rock zone at the top of the Bethany Falls Limestone and the
Hushpuckney Shale at the base of the Bethany Falls Limestone

25.2.2 Regional Geomorphology 25.2.3 Mapping and Trenching of the Surface

Fissures
The site itself is located on the northeastern flank of the
Shawnee Syncline, which causes the bedrock at the site to A total of 23 surface fissures within the loess were located
dip to the southwest. As a result, the mine floor dips slightly and surveyed. The main pattern of the surface fissures formed
to the southwest (up to a few degrees). a set of large “concentric circles” overlaying the CCA
Gentile (1984) has discussed the extensive paleokarst that (Fig. 25.3a). The larger set of fissures has a diameter of about
had developed in the deeper Mississippian rocks which has 180 m. A partial set of concentric fissures was also located to
been recognized throughout the mid-continent. He has pro- the northeast. The surface fissures within the loess were 7.6–
posed that collapse occurring within the Mississippi lime- 76 m long, 0.3–1.2 m wide and 0.3–1.2 m deep (Fig. 25.3b).
stone had propagated to the surface in the Kansas City area Attention was focused upon the major set of concentric fis-
(Fig. 9.7a). These paleocollapse structures range in size from sures that coincided with the center of the CCA.
a few hectare or so to a couple of km2 and include large Seven trenches up to 7.6 m long and up to 3.3 m deep
blocks of bedrock that have moved downward a few meters were cut perpendicular to the surface fissures at different
along high angle, normal faults. They were formed during an locations. Each was examined in detail and it was clear that
earlier interval of geologic time and are covered by a loess the fissures had not occurred recently and in fact were quite
regolith. Consequently, they are commonly unnoticed until old, likely occurring before the mine collapse in early 1970s.
uncovered in excavations. Observations of these paleocollapse However, this was just an opinion based upon our observations
zones have been made at road cuts in the Kansas City area by and needed further verification. In addition to the fissures, there
Gentile (1984) and by our site characterization team (Fig. 9.7b). were numerous small depressions, sinkhole like in appearance.
338 25 The Development of a Landfill over an Abandoned Limestone Mine

These were quickly identified as being associated with the fissures were simply part of the original paleocollapse activity
large open fissures within the Argentine Limestone where prior to the development of the mine. This preliminary con-
the loess cover was thin. ceptual model of the relationship between the mine collapse of
the CCA and the surface fissures was developed based upon:

25.2.4 Developing a Preliminary • Observations in the mine by a senior mining experts

Conceptual Model (Golder Associates 1986);
• Direct observations in seven trenches;
The preliminary conceptual model (Fig. 25.6) for the CCA • Regional mapping and the conceptual model from Gentile
and surface fissures assumes that a paleocollapse zone origi- (1984), and
nating within the deeper Mississippian rocks existed prior to • Observations of paleokarst collapse along road cuts noted
mining. The mine intersected 7 ha of weakened rock due to the by Gentile (1984) and subsequently observed by our site
paleocollapse when it was extended to the west. The surface characterization team (Fig. 9.7b).

Vee-shaped
Surface Fissures

+60

Based upon
site-specific
observations Mine Roof
and Hasan (1988) Collapse
Mine 0

Pleasanton

Approximate Depth in Meters from the mine

Group
Pennsylvanian

-60
Marmaton
System

Group

Paleo-Fractures
Gentile (1984)

Cherokee Group
Based upon

and -120
Atokan Series

-180
Mississippian
System

Paleo-Solution
Cavity

Fig. 25.6 A preliminary conceptual model illustrating the relationship between the surface fissures, the CCA and the paleocollapse within the
deeper Mississippian bedrock
25.2 An Assessment of the CCA and the Surface Fissures 339

While we were reasonably certain of our preliminary 25.2.5.2 Topography

conceptual model (Fig. 25.6) at this point it was merely an While detailed topographic maps of the site had been devel-
opinion and would need further verification. oped, there had been extensive excavations and modification
of surface topography as a result of on-going landfill devel-
opment. The combination of a USGS topographic map and
25.2.5 Aerial Photo Analysis older aerial photography (1954) were used to evaluate origi-
nal surface drainage and the depth of valleys over the site.
An aerial photo analysis was carried out using both stereo- There is considerable topographic variation across the
scopic and monoscopic photos (Beccasio 1988). The aerial site. Figure 25.7 utilizes the USGS 7.5 min Shawnee
photography ranged from 1988, 1985, 1971, 1964, and 1954. Quadrangle (1975) to illustrate the variations in topography
This coverage provided information from pre-mining, at the site. The elevation to the north is 292 m and decreases
throughout the years of mining (from 1960 to 1980) and to the south to approximately 240 m, then decreases further
post-mining. The objectives for the aerial photography to the Kansas River that is approximately 210 m south of the
analysis included fracture trace analysis, pre-development site. The east to west hill and valley terrain has elevation
topography and observing the presence of surface fissures. changes of up to 45 m.
There are two deeply incised north to south valleys over
25.2.5.1 Fracture Traces the mine (Fig. 25.7). The westernmost valley is located about
The 1954 air photos provided the best means for mapping 100 m west of the CCA and it extends downward to the top
fractures in the project area since there was minimal cultural of the Westerville Limestone at the southern side of the mine
development at that time. The primary fracture trend is (Fig. 25.8). The easternmost valley is located south of the
north-northeast, which is reflected in the strong topographic CCA and it extends downward to the top of the Quivira
and drainage alignments on-site. A secondary northwest Shale at the south side of the mine. These incised valleys
fracture trend was also observed (Beccasio 1988). are probably associated with fractures that could extend

Fig. 25.7 The topographic map from the site reveals extensive topographic variation over the western portion of the mine. Two deep valleys can
be seen across the site (USGS 7.5 Minute Quadrangle Shawnee, Kansas revised 1975)
340 25 The Development of a Landfill over an Abandoned Limestone Mine

West East

Loess

275

Quindaro Glacial Till

Shale
Argentine Limestone

Frisbie
Limestone Lane Shale

Raytown-Paola Limestone

Chanute Shale

Elevation in Meters
Drum Limestone

Quivira Shale 250

Westerville Limestone

Block Wea Shale

Limestone

Fontana
Shale
Winterset Limestone

225
Starke Shale
Galesburg Shale
Bethany Falls Limestone
Hushpuckney Shale
Middle Creek Limestone
Ladore Shale

Fig. 25.8 A cross section through the westernmost incised valley associated with fractures that extend deeper into the rock, possibly to
(shown in Fig. 25.7) that extends to the top of the Westerville Limestone the mine resulting in a zone of weak rock and may be a potential
at the southern side of the mine. These surface features are probably geologic flaw at the site

deeper into the mine. They could provide a zone in which aerial photos, the older 1954 pre-mining photos were then
surface water can more easily enter the underlying rock, and reviewed. Although limited by vegetation cover (mid-July
over long periods of time will have weakened the rock. with maximum foliage coverage) traces of similar concen-
However, there was no evidence of such fractures from tric, curvilinear patterns were identified in the 1954 photos
within the mine. (Beccasio 1988). Identifying traces of the surface fissures
before mining began, in the 1954 aerial photos, was a major
25.2.5.3 Evidence of Surface Fissures factor in confirming our preliminary conceptual model and
in the Loess Sediments the relationship between the CCA and the surface fissures.
The surface fissures were first identified on the 1988 and This provided verification that the surface fissures were not a
1985 air photos, characterized by a series of subtle furrows result of subsidence due to the mine-roof collapse. These fis-
arranged in a concentric or curvilinear pattern based upon sures were present before mining and likely a result of a
field mapping. Having identified the fissures in the recent paleocollapse that had taken place long ago.
25.2 An Assessment of the CCA and the Surface Fissures 341

25.2.6 Site-Specific Geology responses from three shale layers. From top to bottom
they include:
At the beginning of our investigation there were about 40
mine backfilling holes immediately to the north and west of • The Muncie Creek Shale a thin layer within the Raytown
the CCA. These boreholes were used for the purpose of Paola Limestone sequence,
injecting fly-ash backfill material into the mine. They were • The Quivira Shale over the massive Westerville
open boreholes that were destructively drilled. There were Limestone, and
also five existing cased vent holes into the mine and a few • The Stark Shale located just below the Westerville
older piezometers and monitor wells that had been installed Limestone.
by others.
While the regional geology was well described in the lit-
erature (Hasan et al. 1988) there were no core samples or 25.2.6.1 Excavation During Landfill
recent site-specific geologic data available. To provide site- Development
specific geologic data one new borehole was drilled as a geo- As the landfill was being developed, clearing of the land and
logic reference hole. NX-core was obtained during drilling excavation exposed a section of the Argentine Limestone
and then the hole was geophysically logged with natural (the uppermost limestone at the site) and the underlying
gamma, induction, gamma-gamma (density) and neutron Lane Shale across the western portion of the CCA. This
(porosity) logs to provide a detailed reference for geologic unexpected window into the geology provided an opportu-
conditions at this site. Since these logs can make measure- nity to gain additional understanding of site conditions. This
ments within open and PVC-cased boreholes, they were run excavation was over 180 m long and extended over the
in all available fillholes, vent holes and monitoring wells western portion of the CCA. Figure 25.9a shows photograph
on-site. These logs provided very repeatable data indicating of the excavation and a portion of the sketched excavation
fairly uniform geologic conditions across the site. (Fig. 25.9b) to document conditions.
Figure 18.5 shows the natural gamma logs from three The top of the Lane Shale was marked at intervals along
boreholes at the site. The natural gamma logs were found to the entire excavation and the elevations surveyed. The loca-
be very repeatable from hole to hole and provided distinct tion of each survey point along the excavation is shown in

a Northern Portion
(Figure 15.2 - Site Walkover) Extent of Excavation
North South

b
Q U
I J K M N O P
L
Elevation in Meters (MSL)

269

265
Argentine
Limestone 262

259
Lane Shale

Fig. 25.9 A photograph of the north to south excavation that cuts across the western side of the CCA (a) and a partial sketch of the excavation (b)
342 25 The Development of a Landfill over an Abandoned Limestone Mine

a
Northern

N
Mine Boundary

1 through 8

Photo and Sketch Haul Road

Figure 15.2

A through Q

Central
Collapse
Area

Southern
Mine Boundary

0 150
Scale in Meters
Northeast Southwest
b
266
Stations
1 to 8 Stations A to Q

265
Elevation in Meters

264

263

262

0 30 60 90 120 150 180 210

Horizontal Distance in Meters

Fig. 25.10 A map of the excavated section (a) along with the profile of the top of the Lane Shale (b), which revealed the presence of a
depression

Fig. 25.10a. The resulting elevation profile of the Lane Shale was a bit of a surprise. These features are quite weath-
Shale is shown in Fig. 25.10b. While the site appears to ered and appear much older than the time of the CCA less
have flat-lying, uniform geology, the elevation profile along than 17 years earlier.
the excavation clearly indicates some unusual conditions.
This exposure revealed the presence of a depression about 25.2.6.2 Identifying a Circular Depression Over
180 m wide with a displacement of as much as 4.5 m. the CCA
Numerous local fractures, open voids and areas of high stress A contour map of the bedrock was developed over the CCA
were seen in the Lane Shale along the face of this cut and its surrounding area. This contour map incorporated
(Fig. 25.11). The extent of open fissures within the Lane data from the 30 geophysical logs acquired over and around
25.2 An Assessment of the CCA and the Surface Fissures 343

Fig. 25.11 Local fractures and open voids along with areas of high than the CCA less than 17 years earlier (a) Open fracture (b) Evidence
stress are seen in the Lane Shale along the face of the excavation. of stress along fractures (c) Typical evidence for water in joint of Lane
These features are quite weathered and are thought to be much older Shale

the CCA. The Muncie Creek shale marker bed (the 25.2.7 Confirmation of the Preliminary
uppermost of the three key marker beds) identified in the Conceptual Model
natural gamma logs (Fig. 18.5) was used as a reference.
In addition, the elevation measurements made along the Many of our early opinions and assumptions presented in the
cross section through the excavation western portion of preliminary conceptual model had now been supported by
the CCA (Fig. 25.10b) and strata within the mine portal this additional data thereby improving our confidence level
(Fig. 25.2) were all used to develop the contour map of in the conceptual model. Additional data included:
the area.
The contour map (Fig. 25.12) indicated the presence of a • The results of the aerial photo analysis indicating that the
local circular area with up to 4.5 m of displacement. It is surface fissures existed prior to the mining of the Tobin mine.
reasonable to assume that this local depression is related to • The elevation profile developed along the surface excava-
the paleocollapse feature. Gentile (1984, 1988 personal com- tion through the western portion of the CCA verified a
munication) reported similar features in the area. These fea- collapse
tures can have 1–1.5 m of vertical displacement and extend a • The local circular depression seen in the contour map
couple of hundred meters. over the CCA further verifying a collapse
344 25 The Development of a Landfill over an Abandoned Limestone Mine

CCA
0 120

Scale in Meters

Elevations for the Lane Shale

Fig. 25.12 A contour map was developed showing a local depression centered over the CCA (note contours in feet)

25.2.8 Hydraulic Connection were probably similar to those observed in the excavation.
of the Paleocollapse Fractures In addition, it was found that these angle boreholes would
not hold water.
As the conceptual model at the site developed, the hydraulic A dye trace study to assess the potential for flow from
connection of the paleocollapse fractures extending from the the surface into the mine via the paleocollapse fracture was
surface fissures downward into the mine was of concern. The carried out utilizing this open joint in the Argentine
older portions of the landfill had used a clay liner at the base Limestone as the dye injection point. Fourteen points were
of the landfill. Now, a fabric liner and leachate collection selected in the mine around the perimeter of the CCA to
system was being installed under newer portions of the land- sample mine-water for the presence of dye using charcoal
fill. However, It was unknown whether any fractures associ- bugs (Fig. 25.13a). Jim Quinlan assisted with the design of the
ated with the paleocollapse would provide a direct pathway dye survey and the placement of charcoal bugs (Fig. 22.6b).
for leachate migration from the landfill into the mine. In addition, charcoal bugs were located within a number
One of the surface fissures in the loess had a small hole at of piezometers and monitor wells and within two ponds on
its base. More than 3,785 l of water was allowed to flow into the surface. Background water samples were obtained from
this hole at a rate of more than 37 l/min without causing the all stations within the mine as well as a number of piezome-
hole to overflow. The loess was then excavated to expose the ters, monitor wells and within two ponds on the surface.
surface of the Argentine Limestone where a significant fis- The city water and tanker truck used to provide water for the
sure in the rock was observed (Fig. 17.16a). This fissure is dye trace was also sampled. Quinlan (1991) and Aley (1991)
located over the highest mine-roof collapse found in the provided analysis of all water and dye samples using a
CCA and was suspected to be a fracture associated with the scanning spectrofluorometer.
deeper paleocollapse. Prior to injecting the dye into the large fracture (Fig.
Two angle borings were drilled through the exposed frac- 17.16a) 15,140 l of water was injected to pre-wet the fracture
ture. Observations were then made within the open boreholes system. Then 0.45 kg of fluorescein dye was mixed with
with a borehole video camera (Fig. 17.16b). A number of 170 l of water and was allowed to flow into the injection site.
fractures and open voids were observed near the projected The dye was followed by 90,840 l of water as a chaser at a
location of the paleofracture all within the Lane Shale. These flow of approximately 340 l/min.
25.2 An Assessment of the CCA and the Surface Fissures 345

Dye Detection Breakthrough Curve

12
Sample Point #5
Sample Point #7
N 10 Sample Point #2

Concentration (ppb)
8

4
#1
#2 2
#3
#4
0
-6 -1 0 1 2 5 10 22 44 128 285 429 438
#5
Central b Day
Collapse #14
Area
#6
#8 #13
#9
#7 0 150 300
#10 #12
#11 Scale in Meters
Dye Injection Point

Mine Sampling Point - dye not detected

Hwy I-70
Mine Sampling Point - dye detected
a First three dye detection points in the mine

Fig. 25.13 Dye sampling was located at 14 points around the perimeter of the CCA (a) and results from the dye trace indicated the presence of
dye in the mine at three locations between 22 and 44 days after injection (b)

At this point in the site characterization, the mine had • Sample station #2 north of the CCA and about 137 m
flooded to the point that movement within the mine was north of the dye injection point,
accomplished by small boat. Prior to injection of dye and dur- • Sample station #7 at the south mine wall and about 180 m
ing the first 10 days after injection of dye, sampling was car- southwest of the dye injection point.
ried out in the mine using a Turner portable flourometer with
continuous recording (Fig. 22.5). Water was pumped continu- The presence of dye at station #5 suggests a possible
ously through the flourometer as the boat traversed through a hydraulic connection between the surface and the mine via
common course within the mine passing by each charcoal the paleofracture. However, the other two sampling stations
bug sample station. No dye was detected by the Turner flou- are located farther from a paleofracture suggesting that dye
rometer during the first 10 days of sampling. The charcoal traveled there via means other than the paleofractures.
bugs in the mine were scheduled to be replaced and analyzed There are three possible pathways for dye to enter the
at approximately 1, 2, 5, 10, 20, 50, 100, and 500 days. mine (Fig. 25.14).
Samples of water with dye were recovered at two older
monitor wells, about 22 and 152 m north of the injection site 1. Via the paleofracture associated with the deep-seated
1 day after dye injection. These wells were screened within collapse in the Mississippian Limestone (Fig. 25.6).
the Raytown-Paola and Drum Limestone. This detection of The paleofracture was observed at three locations within
dye was unexpected but indicated the potential for rapid lat- the mine. The vertical fracture was so tight that thin
eral flow within these formations. This rapid flow may only knife blade could not be pushed into an opening of
occur within the stressed and fractured rock overlying about 0.10 cm.
the CCA. 2. Via the open fissures, fractures and voids that had been
Between 22 and 44 days, dye began to appear at three observed visually in the Argentine Limestone and the
locations in the mine (Fig. 25.13b) and included: Lane Shale (Fig. 25.11). It might be reasonable to assume
that the entire column of rock over the CCA contains
• Sample station #5 at the west edge of the CCA and about similar open fissures, fractures and voids. If so, these could
45 m north-northwest of the dye injection point, be possible pathways for dye to have migrated into the mine.
346 25 The Development of a Landfill over an Abandoned Limestone Mine

Dye Injection into Open

Argentine-Frisbee Fillhole
Limestone into mine

Loess

Argentine-Frisbee Open fractures and

Limestone bedding planes observed in
roadcuts and borehole video
Open Fissures
Lane Shale observed in excavation
(Figure 25.11)
Raytown-Paola
Limestone
Chanute Shale
Chanute Shale swells to close off
20 cm boreholes
in 8 to 10 months
Drum Limestone
Dye reached
Quivira Shale Raytown-Paola
Westerville and Drum Limestone
Limestone within 1 day of injection

Wea Shale Paleofracture associated

Block Limestone with collapse within the
Fontana Shale deep Mississippian Limestone

Winterset Limestone

Stark Shale Paleofracture

Galesburg Shale observed in the mine
less than 0.10 cm
Bethaney Falls
Limestone Mine
Hushpuckney Shale

Fig. 25.14 Rapid lateral movement of dye within the Raytown-Paola and Drum Limestones could have intersected either fractured rock within
the CCA or mine fillholes that had not been grouted and provided direct access to the mine

3. Via the mine fillholes that may not have been grouted While the dye trace did not definitively answer the ques-
prior to the dye injection. The lateral movement of water tion of a direct connection for flow from the surface to the
within the Raytown-Paola and Drum Limestones could have mine there was information gained. There was also the unex-
flowed down open boreholes being used to fill the mine. pected lateral flow within the Raytown-Paola and Drum
Limestones within 1 day after injection of the dye. However,
The later two options are the most likely pathways into we do not know if this rapid flow was limited to the area over
the mine: by the open fissures, fractures and voids that had the highly fractured rock within the CCA or was common to
been observed within the stressed column of rock over the the entire site. While this was not the focus of the dye trace
CCA or via mine fillholes. study, it did ultimately impact the groundwater monitoring
Additional backfilling of the mine by the rock slurry plan later in the project.
method began about 133 days into the dye tracer test.
Backfilling water was being pumped from the mine and
re-injected with the crushed rock backfill at about 5,677 l/min. 25.2.9 Subsidence Measurements
After the mine backfilling began the amount of dye detected
in the mine rapidly increased to 10 ppb or more and was Through extensive characterization within the mine, it had
found along the entire west side of the CCA where backfill been determined that the possibility of surface subsidence
material was being placed (Fig. 25.13a). Because of this due to mine collapse was highly unlikely. Three outside
further dye sampling was discontinued. experts Golder (1986), Tein (1988), and Brink (1989) who
25.2 An Assessment of the CCA and the Surface Fissures 347

had inspected the mine conditions had also indicated that Argentine Limestone a minimum of 1.2 m. A network of 11
mine conditions were stable and would not result in surface subsidence monuments were initially installed from south to
subsidence. However, the question of mine stability still north across the CCA (Fig. 25.15):
remained a concern by the state and subsidence measure-
ments were undertaken. • Two monuments M-100 and M-9 were located off of the
Early on we recognized that such measurements were dif- mine, one to the south and the other to the north to pro-
ficult and would require an extreme level of expertise along vide stable reference points.
with quality assurance (QA) and quality control (QC). To • Two monuments M-2 and M-3 were located over the
accomplish this we engaged a survey firm who were experts mine, but off of the CCA.
in subsidence measurements to oversee the program and a • Seven other monuments M-4, M-5, M-6, M-7, M-8,
local survey firm with experience in such measurements. The M-10, and M-11 were located over the CCA.
design and layout of the monument network was evaluated
and it was estimated that the network could be expected to A series of baseline subsidence measurements were made
detect vertical displacements on the order of (0.021 cm using 1 week and again 1 month after the monuments were installed
first order vertical measurements (Boston Survey Consultants to assess stability and provide baseline data. Then a sequence
(BSC) 1990). of subsidence measurements was carried out about every
The subsidence monument design was modified from that 3–4 weeks by Anderson Survey co. (1990).
in (Dunnicliff and Green 1988) (Fig. 20.2). Monuments were Because of heavy truck and equipment traffic on-site
located along a clear visual path for easy access of the survey during the week, survey measurements were made on
crew. The monuments were anchored in the uppermost weekends when there was little or no traffic present on-site.

Northern
Mine Boundary

N
M-9
Surface Excavation

M-8
M-7
Point of M-6 M-10
Dye Injection M-12 M-13
M-11
M-14 Boundary of CCA
M-5
M-4 M-15
M-3

M-2

M-100

Southern
Mine Boundary

0 150
Scale in Meters

Shallow subsidence monument anchored in Argentine Limestone

Deeper subsidence monument anchored in Westerville Limestone

Fig. 25.15 The location of subsidence monuments that were installed across the CCA
348 25 The Development of a Landfill over an Abandoned Limestone Mine

South North
a 0 b 0
Subsidence in Centimeters

Subsidence in Centimeters
M-5 shallow
M-11 shallow
-1.5 M-14 shallow -1.5
M-15 deep

-3.0 -3.0

Time of Dye
Nearby
Injection
Excavation
-4.5 -4.5

M-10
M-11

M-9
M-100

M-2

M-5

M-6
M-3
M-4

M-8
400

M-7
0 200 600
Time (days)
CCA
Rate of subsidence over time for monuments Maximum subsidence at each of the
showing the maximum subsidence original subsidence monuments

Fig. 25.16 Abrupt changes in subsidence are seen in two shallow Westerville Limestone. Note that subsidence seems to be slowing
monuments anchored in the shallow Argentine Limestone. A similar and stopping with time (a). The maximum subsidence of all initial 11
response is seen in the newly installed monuments one anchored in the monuments is shown in profile (b)
shallow Argentine Limestone and the other anchored in the deeper

BSC reviewed the monument design, field survey procedures Then the subsidence of all the monuments appears to
and the data acquired by Anderson Survey Co. BSC also decrease and subsidence seems to be stopping by day 600.
carried out an independent set of measurements as a means of Figure 25.16b shows the maximum subsidence profile
quality control. The subsidence monument design and across the initial eleven monuments from south to north.
installation, the field measurements by two experienced The maximum subsidence measured was almost 4.5 cm.
firms, along with the QA/QC procedures have resulted in the This was measured at two adjacent monument locations
subsidence data being the highest possible quality. (M-5 and M-11). While these two adjacent monuments pro-
vided very similar data, there was considerable differential in
25.2.9.1 The Results of Subsidence both the rate and amount of settlement taking place along the
Measurements line of monuments.
The initial subsidence monuments indicate little, if any, After 600 days and a total of 26 sets of measurements,
significant subsidence. Some of the movement is due to further efforts were discontinued due to excavation for the
noise or error in measurement since some values indicate a expanding landfill impacting the monuments. No obvious
movement upward. changes in mine roof collapse were observed during this time.
Figure 25.16a shows the data from the two of the monu- Figure 25.15 shows the location of subsidence monu-
ments showing the greatest subsidence (M-5 and M-11). The ments along with the excavation west of the CCA and the
subsidence begins slowly, but after the 13th set of measure- point of dye injection. The excavation is location within
ments (about day 340) an abrupt increase in subsidence 45 m of the monuments and was excavated a few months
occurs. As a result of this rapid subsidence, four additional prior to the rapid change in settlement of shallow monu-
subsidence monuments were installed near those monuments ments. As a result, the monuments which were anchored in
that showed an increased rate of subsidence. Two shallow the shallow Argentine Limestone (a highly fractured rock)
monuments (M-12 and M-14) were installed in the Argentine could have been affected by the lateral stress relief of the
Limestone and two deep monuments (M-13 and M-15) were excavation which had removed as much as 10 m of rock.
installed in the massive Westerville Limestone at a depth of The introduction of 105,980 l of water associated with the
more than 30 m to evaluate whether the rapid subsidence was dye trace, 9 months before the rapid change in settlement has
due to shallow or deeper cause. Figure 20.3 shows a cross likely contributed to the movement of both shallow and deep
section of all monuments both shallow and deep. monuments over the CCA. This was an exceptional amount
The newly installed monuments near M-5, both shallow and of water rapidly injected into an otherwise dry system.
deep (M-14 and m-15), immediately began to show the same The combination of the excavation within 45 m from the
pattern of subsidence as the nearby monuments (Fig. 25.16a). line of monuments and the 105,980 l of water as part of the
25.3 An Assessment of the Mine Conditions 349

dye trace study seem to be the most likely cause of subsid- • Estimating the extent of upward collapse within
ence based upon known geologic and hydrologic conditions. the CCA;
We sometimes encounter results from our observations and • Identifying the source of water filling mine;
measurements that are highly variable and are difficult to • Evaluating mine water quality and
interpret with a high degree of certainty. While the subsid- • Assessing the potential subsidence over the CCA.
ence data itself is considered quite accurate, our interpreta-
tions of the causes of subsidence still contain a level of
assumptions and opinions that are not backed up by site-
specific detailed data. 25.3.1 Development of a Detailed Mine Map

Access into the mine required a detailed mine safety plan.

25.3 An Assessment of the Mine Part of the safety plan included a detailed working map of
Conditions the mine and was one of the first priorities. The map pro-
vided a means of safely carrying out routine inspections
Assessing mine conditions was focused upon the western within the mine as well as providing a means to document
portion of the mine over which a landfill was being devel- the location of conditions within the mine. A mine map was
oped. Observations in the eastern portion of the mine were available from mining operations (Tuttle, Ayers, and
limited to a 100 m or so around the portal entrance and under Woodward Co. 1979) reference points had been marked on
the mine portal road. The assessment of mine conditions was pillars and on the ceiling, but the map was quite cluttered
important to the project and addressed concerns of the state and not easy to interpret. However, it was used as a basis for
and ultimately aided in the assessment of subsidence risk. developing a new detailed mine map that incorporated the
This work had been on-going since the early part of the areas of mine-roof collapse (Fig. 25.17). The working mine
project and included: map was started as soon as access to the mine was obtained
in June of 1987. It focused upon documenting a detailed
• Developing a detailed mine map boundary of the CCA and other collapse areas, as well as
• Monitoring any changes in mine conditions for 3 years; identifying a routine pathway to be used in traveling within
• Determining the mechanism and types of mine collapse; the mine. The working mine map evolved over time as we
• Acquiring bulking data accessed various parts of the mine.

Fig. 25.17 A detailed map of the mine including the CCA and other collapse areas. The mine map also identifies the main routes used to travel
within the mine
350 25 The Development of a Landfill over an Abandoned Limestone Mine

25.3.2 Monitoring of Changes in Mine strength of the Bethany Falls Limestone is

Conditions over 3 Years 82.7–137.9 MPa.
• No major roof collapse similar to the CCA had occurred
As part of developing the mine map, conditions within the within any of the oldest portion of the mine to the east
mine were monitored. Monitoring of mine conditions was where randomly spaced pillars were used.
focused upon the CCA and other isolated collapse areas but
also included general conditions of the pillars, floor and ceil- The general conditions within the mine, outside the CCA,
ing as well as water levels. were considered good. The collective opinions of those
Monitoring of collapse areas consisted of visual inspec- experienced experts who had entered the mine strongly sup-
tions and photographing each collapse site every 3–4 months ported the fact that the surface fissures were a result of a
over a period of 3 years. Each area of collapse was identified collapse originating deep within the Mississippian Limestone
using a large white sheet of plastic with black numbers and and propagated to the surface, long before mining occurred.
photographed over time providing long-term monitoring of Recommendations by Brink (1989) included:
any changes. A total of 125 collapse sites were identified for
monitoring. The photographs from each inspection was • Extension of the landfill over the CCA should be permitted
reviewed and compared to the previous photographs. This provided that the surface paleocollapse fissures are
inspection process provided a means to document any chang- exposed and sealed.
ing conditions at the large number of collapse sites. Note that • That subsidence monitoring be made over the mine and
the photos included here do not have numbers as these were • A groundwater monitoring program should be pursued.
the very first set of photos.
The CCA had not expanded laterally since the major col- Although scattered roof failures can be expected in the
lapse in the early 1970s. This was based upon roof stability future due to the exposure of the rubble zone. They will be of
warning devices and support beams that had been installed limited vertical extent and will not cause subsidence or
at the south edge of the CCA after the initial collapse. fracturing at the ground surface. This was also the conclusion
Furthermore over more than 3 years of monitoring the CCA by Hasan et al. (1988). Furthermore, no other mine in the
there had not been any significant lateral or vertical expan- greater Kansas City area had encountered an instantaneous
sion of the CCA. collapse of the size of the CCA.
About 60 % of the stations monitored (around the CCA
and other isolated collapse areas) over the 3-year period have
had some small level of additional collapse activity. The 25.3.3 Determining the Types of Mine
additional collapse were all small, localized failures, typi- Collapse
cally about 1 m3 associated with further degradation of the
rubble zone or edges of the Starke-Galesburg Shale. The mine-roof collapse areas were classified into four types
Two independent mine experts were engaged to help eval- to provide a convenient means of characterizing the nature
uate mine conditions. Both experts, Tein (1988) from the of collapse.
Rolla School of Mines and Brink (1989) from South Africa
provided similar and supporting observations and conclusions 25.3.3.1 Type A Isolated Collapse
to that of an earlier consultant (Golder Associates 1986). Type A collapse are isolated, occur within individual rooms
(centered between pillars) and are limited to about 0.04 ha
• The pillars show no signs of spalling, punching, or stress (Fig. 25.18a). The Bethany Falls Limestone was mined to
related failure. within 0.3–0.6 m from its top. Type A collapse extend
• No evidence of roof failure due to cracking or spalling upward into the Galesburg Shale, (about 2 m thick) and then
was apparent in the areas inspected around the perimeter into the Stark Shale (about 0.6 m thick) to the base of the
of the CCA or in general. Winterset Limestone. In a few cases, collapse would extend
• Generally the mine floor appears to be excellent, although further upward into the thick Winterset Limestone
floor heave was observed south of the CCA in an area (Fig. 25.18b). Type A collapse are the most typical of the
where water was present on the mine floor. Floor heave is roof collapse and are open so that the upward extent of col-
due to degradation of the Hushpuckney Shale as a result lapse can be directly observed.
of the mine filling with water. About 15–30 cm of floor
heave was noted at three locations. 25.3.3.2 Type B Coalesced Collapse
• Engineered pillars of 7.6 by 7.6 m on 18 m centers would Type B collapse are coalesced collapse, where adjacent
yield about 82 % extraction. However failure of pillars is isolated Type A collapse areas have merged to result in a
unlikely even at an extraction of 90 %. Compression collapse with a larger aerial extent between two or more
25.3 An Assessment of the Mine Conditions 351

Fig. 25.18 Type A collapse are local, isolated collapse between pillars and would have a typical area of about 334 m2 (a). Collapse would extend
upward into the Stark Galesburg Shale and in some cases into the Winterset Limestone (b)

rooms with pillars intact. This type of collapse has occurred Since these three mine-roof collapse were all located
at an estimated 10–30 % of sites. They are open so that the along the outer perimeter of the CCA they are thought to
upward extent can be directly observed (Fig. 25.19). occur along the paleofractures associated with the CCA. A
Two areas in particular illustrate extensive Type B col- fracture was observed within the mine at each of these three
lapse. They are located east and are separate from the CCA locations within the Stark Galesburg Shale. No displace-
and occupy an area of about 0.2–0.4 ha (Fig. 25.17). In July ment was observed at the crack but strata dipped downward
1990, a newly discovered but older long sinuous roof col- at an angle of 10–20° on the inside of the collapse. In each
lapse was discovered in an area east of the mine portal road case, the paleofracture was within a few degrees of vertical and
(Fig. 25.17). This roof collapse is more than 304 m in length. was tight and would not allow a thin knife blade to enter.
Here, the rubble zone is as much as 2.1 m thick compared to Access to one of these three sites was obtained by crawl-
its typical 0.45–1 m thickness. Roof failure extended upward ing upward on the pile of rubble (Fig. 25.20a). Upon reach-
to the bottom of the Winterset Limestone. The collapse fol- ing the top of the collapsed rock pile observations were made
lows a sinuous path much like a meandering stream channel. of what is believed to be the uppermost extent of rock failure
This is a good example of a deep paleochannel of the rubble within the CCA. Overhead we could see the base of the mas-
zone that had been inadvertently exposed by mining sive Westerville limestone (Fig. 25.20b) approximately 18 m
operations. above the mine-roof. The collapse tapered inward and
became smaller as it extended upward (Fig. 25.20c). This is
25.3.3.3 Type C Isolated High Collapse the uppermost extent of collapse observed within the CCA.
Type C collapse are isolated high collapse which have
extended upward beyond the Winterset Limestone 25.3.3.4 Type D Closed Collapse
(Fig. 25.20). Only three Type C collapse were discovered. Approximately 10 % of the collapse sites observed are
However, there may be others that are blocked off and are Type D collapse. They are closed collapse (blocked off from
inaccessible. These three areas around the perimeter of the access) in which the upward extent of collapse cannot be
CCA provided physical access up into the collapse zone. seen or accessed (Fig. 25.21). This type of collapse is typi-
cally found around the perimeter of the CCA.
352 25 The Development of a Landfill over an Abandoned Limestone Mine

Fig. 25.19 Type B coalesced collapse extend latterly from room to room encompassing larger areas (a) and extend upward into the Stark
Galesburg Shale similar to Type A (b)

25.3.4 Determining the Mechanism of Mine much as 18 m above the mine (Fig. 25.20c). The paleofracture
Collapse zone has resulted in a local circular depression with subsid-
ence of up to 4.5 m over the CCA (Fig. 25.12). Rock over the
Two geologic aspects appear to impact mine stability and CCA is extensively fractured and fissured as seen in the Lane
cause roof collapse: Shale exposed at the excavation (Fig. 25.11) and also noted
in various video logs.
• The presence of paleofractures that are associated with The zone of rubble rock (Fig. 25.5) is often found in what
the CCA and extend through the rock column from below appears to be channels that vary in thickness. In thicker
the mine upward to the surface. channels, the channel bottom can inadvertently be exposed
• The presence of a localized thicker zone of rubble rock in by mining operations. The rubble rock is the cause of the
the upper part of the Bethany Falls Limestone. Type A, B and D collapse areas (Figs. 25.18, 25.19 and
25.20). One of the longest channels of rubble rock with
The paleofractures have weakened the rock and are thickness up to 2.1 m was found east of the mine portals
caused by the collapse of a large deep-seated cavern within (Fig. 25.17). It was more than 304 m long. In addition, the
the Mississippian Limestone which lies more than 180 m rubble zone is known to deteriorate rapidly when exposed
below the mine (Fig. 25.6). This deep-seated collapse had to moisture, which leads to localized roof failure (Hasan
occurred well before mining had taken place. The paleofrac- et al. 1988). The outer perimeter of the CCA contains many
tures (Fig. 25.6) have had the greatest impact in causing collapse initiated by the rubble rock. In addition, all of the
mine-roof collapse. These features allowed the single large collapse areas outside of the CCA have been initiated by the
collapse event to occur forming the CCA (Fig. 25.3) and also rubble rock. This mechanism is responsible for the majority
appears to allow the highest upward migration of collapse, as of roof failures in this and other mines in the area.
25.3 An Assessment of the Mine Conditions 353

Fig. 25.20 Type C collapse are those in which roof collapse has assess conditions (a). Looking up we could see the base of the massive
extended upward higher than the Winterset Limestone up to the Westerville Limestone some 18 m above the mine roof (b) the cross
Westerville Limestone One of them allowed physical access upward to section shows the detail (c)

25.3.5 Bulking Measurements Hasan et al. (1988) indicate that bulking of failed roof rock
over limestone mines in the Kansas City area is about 30 %.
Fallen rock will occupy a greater volume than it did when the The amount of bulking that occurs can vary considerably
rock was in place. This increase in volume is referred to as depending upon:
bulking or swell factor and has been discussed in Sect. 9.4.1.
The bulking of fallen rock is one of the factors to be assessed • The size and shape of the broken rock
when evaluating whether a mine collapse could potentially • The relative strength of the rock
reach the surface. If the rock overlaying a mine (or cave) is • The geometry of the open cave or mine.
thick enough, the volume of bulking rock will be more than
the remaining void space and will eventually fill the void Bulking will decrease as the opening becomes filled up
space preventing further upward migration of roof failure, and the fall distance is reduced, then there will be less
before it reaches the surface (Fig. 9.5). Rock that falls in rotation and breakage of the rock. The bulk rock volume
smaller pieces (Fig. 9.6a) result in a higher bulking factor may also decrease over time due to compaction of the rock
than large intact slabs with little rotation (Fig. 9.6b). fragments and the presence of water.
354 25 The Development of a Landfill over an Abandoned Limestone Mine

Fig. 25.21 Type D collapse are choked-off where fallen rock has prevented visual or physical access up into the collapse area (a). The vertical
extent is unknown, but probably extends into the Winterset Limestone or higher (b)

Extensive direct measurements and photo documentation ings indicate open voids ranging from 0.6 to 2.4 m, averaging
of bulking were made in the Tobin Mine. Measurements were 1.5 m and estimated bulking factors averaging 36 %
made of the volume of the roof void created by the fallen rock (Fig. 25.22). The highest extent of collapse from these seven
and the volume of the rock rubble pile on the mine floor. At borings was in TB-8, about 9 m above the mine roof into the
some locations the entire perimeter of the rock fall area was Winterset Limestone. The result from these seven borings
accessible enabling very accurate measurements to be made. also suggests that the roof collapse over most of the CCA,
In other locations access was only available on one side of the is typically Type A and B collapse seen within the mine
collapse requiring an estimate to be made. These bulking mea- (having a vertical extent limited to the base of or within the
surements resulted in an average bulking factor of 42 %. A Winterset Limestone) (Fig. 25.18 and 25.19). Based upon
conservative bulking factor as low as 30 % would limit upward these borings into the CCA collapse an average bulking
collapse to below the base of the Westerville Limestone about factor of 36 % was estimated.
18 m above the mine (Fig. 25.2). Furthermore, the Westerville
Limestone, which is massive and about 6 m thick, would likely
limit any further upward collapse. 25.3.6 Further Support of the CCA
Because the interior of the CCA was inaccessible, a ques- Conceptual Model
tion remained as to the height of collapse and the degree of
bulking within the CCA. Seven borings were drilled into the Our preliminary conceptual model of mine conditions
CCA and logged with a borehole video camera. These bor- and the relationship of the CCA and surface fissures has
25.3 An Assessment of the Mine Conditions 355

TB-2 TB-3 TB-4 TB-8 TB-10 TB-11 P2A

216

Void

Winterset
219 Collapse Limestone
Elevation in Meters

Rubble
Void
Void Void
Collapse Void
222 Rubble
Collapse Stark/
Void Collapse
Rubble Galesburg Shale
Void Rubble Collapse
Rubble
Collapse
//////////// Rubble
Zone
225 Rubble Collapse
Rubble
Bethany
Falls
Limestone
228 Mine

High angle fractures

observed in 5 of the 7 boreholes
Top of collapse rubble
observed in the boreholes

Fig. 25.22 Seven borings were used to evaluate the upward extent of collapse and bulking factor within the CCA. The results indicated an average
bulking of 36 %

been supported by further direct evidence. A high level of • Direct observations at one point which exposed what was
confidence can now be placed on the conceptual model thought to be the uppermost zone of collapse within the
based upon: CCA, about 18 m above the mine roof at the base of the
massive Westerville Limestone.
• Our extensive direct observations and photo documenta-
tion around the perimeter of the CCA.
• Observations were supported by three independent mine
experts, Golder (1986), Tein (1988), and Brink (1989) 25.3.7 Determining Sources of Water Filling
who also thought that the mine was stable. This was also the Mine
the conclusion by Hasan et al. (1988).
• Extensive bulking measurements throughout the mine The former mine superintendent (Nicholson 1988), indicated
ranging from 35 to 42 % suggesting that worse case col- that the mine was generally dry until mining extended south-
lapse scenario will not reach the surface. west of the CCA by 90–120 m. Then a truck of water, about
• Seven borings into the CCA along with video logs to 3,700 l, was removed every few days. As the mine extended
measure the height of roof collapse and the height of the further westward, a number of sumps 0.9–1.5 m deep were
rubble pile with an average bulking factor of 36 %. installed below the mine floor and pumped as needed to con-
• Direct observations of the vertical paleofracture at three trol water. These sumps were installed within the Hushpuckney
locations within the CCA. The fracture was tight with a Shale, the Middle Creek Limestone, and the Ladore Shale,
dip of 10–20° to the inside of the collapse. which is up to 3 m thick (Fig. 25.5). A large pump had been
356 25 The Development of a Landfill over an Abandoned Limestone Mine

installed in the main sump at the center portal to remove runoff While no measurements were made in the Hushpuckney
water from the portal road (Fig. 25.17). After mining opera- Shale at the Tobin mine, considerable work had been carried
tions ended in 1980 and without continued pumping, the mine out at two other landfills and the large Missouri Portland
began to slowly flood from the southwest (Nicholson 1988 Cement Co. mine and quarry some 29 km east in Jackson
personal communication). Subsequently a well and pump was County Missouri. The geology there is identical to that of the
installed in the extreme southwestern corner of the mine to Tobin mine.
provide water for daily dust control along the landfill roads. Lateral flow within the Hushpuckney Shale had been
Initial access to the mine in 1987 was by foot. During our observed at a number of locations at the Missouri Portland
first trip into the mine the mine floor around the CCA was Cement Co. mine and quarry. A section of the exposed
mostly dry. Monitoring of water level within the mine was Hushpuckney Shale, Middle Creek Limestone and Ladore
started in late 1987 using a digital water level recorder. Shale can be seen in outcrops. These observations along with
A staff gauge was used to provide back-up data in case the discussion with the geologist at the Missouri Portland
electronic equipment failed and was read on each trip into Cement Co. indicated that the zone of rock immediately
the mine. Measurements indicated that water was increasing below the Bethany Falls Limestone was a permeable zone
at about 3.3 cm/month or about 40 cm/year. At that time that transmitted water latterly quite rapidly (Camp 1988
there was an estimated 227× 106 l of water in the mine personal communication).
increasing at an estimated rate of approximately 1.3 × 106 l/ Most of the mines in the Kansas City area, to the east, are
month. well above the levels of the Kansas and the Missouri Rivers.
By April 1990 water levels had extended eastward to At the Tobin mine the elevation of the mine floor in the
encircle the CCA. The extreme northernmost portion of the southwest corner of the mine is about 7.6 m below that of the
mine was dry at that time. As time passed, access to the north- Kansas River, which is approximately 210 m to the south of
ernmost portion of the mine to obtain water quality samples the mine (Fig. 21.4). A piezometer was installed between the
became more difficult due to raising water levels and finally Kansas River and the southwest corner of the mine and
became impossible. On our last trip into the northernmost showed that water from the Kansas River was flowing into
portion of the mine to sample mine water in early 1991, the the mine. Water from the Kansas River flows through a
clearance between the water and the mine roof west of the thickness of 12–21 m of alluvium capable of yielding from
CCA was about 45 cm requiring us to lie back in the boat and 567 to 3,785 l/min (O’Connor 1971). This water with a head
push on the mine roof by hand to obtain passage. of approximately 7.6 m or more, then flows through the
Possible sources of water entering the mine had been sug- permeable zone about 213 m into the mine. The issue of the
gested by others and included: source of mine water had now been resolved.
When the water level in the Kansas River and the mine are
• Seepage from the portal wall, in equilibrium the mine will be flooded over most if not all of
• Surface runoff into the mine from the portal road, and the CCA. Water level measurements were discontinued in
• Seepage from the mine roof including the paleocollapse 1995, but observations by downhole video camera in 2003 at
fracture zone. the sump area near the portal entrance indicated water levels
were above the mine floor level at the portal entrance.
The combination of water seeping out of the northeastern
side of the portal wall and the water flowing down the portal
road were both associated with intermittent rainfall and 25.3.8 Water Quality Measurements
could not possibly account for the large volume of water
entering the mine. In addition, the hundreds of hours of As part of a groundwater monitoring plan a mine-water sam-
observations in the mine revealed no flow or significant pling program was initiated to characterize background qual-
seepage into the mine along fractures in the overlaying rock ity of mine-water. Field parameters measured included pH,
or along the paleofracture associated with the CCA. (9.7–12.4 near areas with fly-ash fill and 7.6–9.1 away from
These possible options for the source of water flowing the fly-ash fill) specific conductance (2,260–7,340 umhos/
into the mine could not account for the large quantity of cm) and temperature (12.9–16.4 °C). Laboratory analyses
water entering the mine, at an estimated 1.3 × 106 l/month. were made for those expected parameters to be required for
The focus now shifted to other possibilities for flow into the Subtitle D – Phase I, U.S.-EPA solid waste landfill monitor-
mine by a permeable zone below the Bethany Falls Limestone ing. In addition, the samples were analyzed for several pesti-
consisting of the Hushpuckney Shale, the Middle Creek cides and herbicides. Geochemical analyses for major ions
Limestone, and the Ladore Shale (Fig. 25.5). This zone is up indicate that the mine-water is predominantly sodium sulfate
to 3 m thick (Hasan et al. 1988). in character.
25.5 Subsidence Risk Assessment 357

25.3.8.1 Possible Further Dissolution The results of the dye trace study indicated rapid lateral
of Limestone flow within the Raytown Paola and Drum Limestones (up to
There was some concern by the state about the possible dis- 150 m in 1 day). The Raytown Paola Limestone was also
solution of the remaining mine pillars as well as the crushed- the first continuous layer of limestone under the landfill.
rock limestone that was being used as mine backfill material. These two zones were selected as the uppermost units to be
Natural dissolution of limestone is caused by fresh rain water monitored. In addition the mine water would be monitored
entering the limestone groundwater system. Since rain water (Fig. 21.4).
is slightly acidic due to CO2 in the atmosphere and humic
acids, the slightly acid water slowly dissolves the limestone.
Ford and Williams (2007) indicate that dissolution rates are 25.5 Subsidence Risk Assessment
on the order of 2.5 cm/1,000 years.
The water entering the mine is dominantly from the All to often risk assessments are based upon an overview of
Kansas River, which is flowing through fractured rock, conditions (opinions and assumptions) with little site-
including limestone, into the mine. River water flowing specific data to back them up. Because of the extensive work
through the limestone on the way to the mine becomes par- at this mine site and others within the area, the possible fac-
tially or totally saturated with calcium. The calcite saturation tors associated with mine collapse are well understood.
index (SI) was calculated for several mine water samples. The strategy used to predict the risk of surface subsid-
The SI for a given sample is an indicator of whether the sam- ence is simple, but highly dependent upon first identifying
ple is saturated with respect to calcite (i.e. whether it has the the key geologic, hydrologic and cultural factors which
capacity to dissolve any limestone). A positive SI value indi- may possibly contribute to surface subsidence. Then,
cates that the water is oversaturated and will not dissolve obtaining appropriate and adequate site-specific data to
calcite. The SI values were calculated using procedures characterize these factors and carrying out a detailed
developed by Hem (1989). assessment of them. This strategy also helps to minimize
The high pH measured in the mine-water (7.6–9.1) and an the use of opinions and assumptions. Subsidence risk pre-
average positive saturation index of (0.63) calculated from diction is then relatively straightforward in concept and a
geochemical analyses, indicate that geochemical conditions realistic estimate of subsidence risk can be made (Benson
are oversaturated and dissolution of the pillars or crushed and Hatheway 2001).
limestone backfill will not occur. After water level in the
mine is in equilibrium with that of the Kansas River, there
will be no further flow into the mine and the water chemistry 25.5.1 Mine Stability Assessment
will come to equilibrium and any dissolution of limestone,
however small, will cease. Mine conditions were assessed by Benson and Hatheway
over a period of more than 3 years. In addition, three inde-
pendent mine experts Golder (1986), Tein (1988), and Brink
25.4 A Groundwater Monitoring Plan (1989) had made 1-day inspections of the mine. All con-
cluded that the mine itself is stable and would not cause
Very early portions of the landfill had been placed upon the surface subsidence.
top of thick loess or clay layers. Later when mine backfilling The key factors associated with possible mine collapse
occurred the Argentine Limestone (the uppermost rock at can be divide up into three zones:
the site) was removed and crushed for use as backfilling
material. The landfill was then placed upon a plastic liner • Zone 1 Conditions above the open mine,
and a leachate collection system was installed. A detailed • Zone 2 Conditions within the mine and,
groundwater monitoring plan was developed for the site • Zone 3 Conditions below the mine
(Benson et al. 1991b) and is presented in Sect 21.7.1 and
summarized here. Zone 1: Conditions above the open mine, include two
Sixteen of the open boreholes had been logged with a potential problem areas:
borehole video camera. The video logs were used to identify
fractures and zones of water. Based upon the video logs, 1. The two deeply incised valleys (Figs. 25.7) one of which
joints were common in the Raytown-Paola and Drum reaches the massive Westerville Limestone. These frac-
Limestone and limited quantities of water were encountered. ture zones may extend to the mine level allowing surface
Little, if any water, was found in the majority of bedrock water to seep into the underlying strata thereby reducing
overlying the mine. In addition, the Chanute Shale was seen its strength. However, there were no indications of frac-
to swell and seal off 20 cm diameter open boreholes that tures of mine roof failure within the mine at these two
could then hold water in less than 10–12 months. locations
358 25 The Development of a Landfill over an Abandoned Limestone Mine

2. A variety of data clearly shows the unique conditions of • A thick layer of rock over the mine (52 m) of limestone
subsidence along with fractures and open fissures within and shale)
the rock over the CCA. These conditions were further • Arguments of mine stability by the authors and other
verified by the rapid lateral flow of dye trace water within mine experts
the Raytown-Paola and Drum Limestones along with the • A bulking factor of 30 % would have stopped collapse at
results of subsidence measurements and the observations or below the Westerville Limestone
of high angle fractures (Fig. 25.22). These conditions of • The presence of the massive and thick (6 m) Westerville
inherently weakened rock within the area of the CCA Limestone
could result in further settlement. However, any such set-
tlement would likely be small (probably only a few cm), However surprising things do happen! The following are
see subsidence measurements (Fig. 25.16). two examples of subsidence in which we have had the oppor-
tunity to make observations of the cause and effects of sur-
Zone 2: Conditions within the mine include the rubble face subsidence in the Kansas City Area. In both cases, the
rock on top of the Bethany Falls Limestone and the overlaying strata are identical to that in the Tobin mine,
Hushpuckney Shale directly beneath the Bethany Falls approximately 52 m of a limestone and shale. Under normal
Limestone (Fig. 25.5). Increasing water levels would nega- hydrogeologic conditions we would not have expected sur-
tively impact both of these conditions. face subsidence or collapse to occur at these two sites due to
The mine water level will continue to increase until it the thickness of rock over the mine and the associated bulk-
comes into equilibrium with that of the Kansas River. As of ing factor using a similar analysis to that of the Tobin mine.
2003, water was over the mine floor at the portal entrance.
Water will eventually fill the westernmost portion of the 25.5.2.1 The Inland Storage and Distribution
mine and the entire CCA. This will cause further degradation Center
of the rubble rock, and the Hushpuckney Shale. The Inland Storage and Distribution Center is located a few
Further degradation of the rubble rock will initiate Type A kilometers to the south of the Tobin mine and had railway
and B collapse. These are localized collapse, which extend access into a mine along with refrigerated storage. A local-
upward to or into the Winterset Limestone. Further degrada- ized portion of the mine-roof had collapsed and resulted in
tion of the Hushpuckney Shale could also lead to pillar surface subsidence. The mine collapse had choked-off and
puncturing of a few centimeters so. However, this would be was not accessible to determine the upper extent of roof fail-
uniform over the mine. If this occurred it would impact the ure similar to a Type D collapse in the Tobin mine (Fig. 25.21).
entire western half of the mine by uniform subsidence (on Surface evidence of subsidence was seen as open fractures of
the order of a few cm). a few centimeters. Surface observations along with topo-
Zone 3: Conditions below the mine would include reacti- graphic maps revealed that the collapse had occurred along a
vation of the original paleocollapse associated with the deep deeply incised valley. It is likely that this valley had focused
cavern system within the Mississippian Limestone some surface water into a major fracture system and water had
180 m or more below the mine (Fig. 25.6). This is highly migrated downward into the rock over a long period of time
unlikely, since the original collapse of the cavern within the resulting in a zone of weakened rock. This zone of weakened
Mississippian Limestone occurred in past geologic time. rock allowed the mine-roof collapse to propagate to the
These paleocollapse features are believed to be structurally surface.
stable and there have been no known reactivations of similar
collapse zones in the Kansas City area (Gentile 1988 per- 25.5.2.2 The Eastern Portion of the Tobin Mine
sonal communication). A medical facility consisting of two buildings and large
If the worse case conditions in Zone 1 and 2 were to paved parking lots had been constructed over the eastern end
occur, any subsequent surface subsidence would be small on of the Tobin mine (Fig. 25.1). The area of the medical facility
the order of a few centimeters, and would be minimized by including the buildings and paved parking lots covered
the presence of the mine backfill material. approximately 1.6 ha.
While this was an area of the mine that had not employed
uniform engineered pillar design, no significant roof failure
25.5.2 Two Examples of Subsidence had been reported by the miners or noted by the surveyors in
the eastern half of the mine. Furthermore, the mine superin-
The following four factors have been emphasized as an tendent, (Nicholson 1988) indicated that the Bethany Falls
argument against subsidence reaching the surface at the Limestone was generally better quality on the east side of
Tobin mine site: the mine and that the eastern portion of the mine did not have
25.6 The Mine Backfilling Program 359

Fig. 25.23 Evidence of surface subsidence up to 2 m is seen in the paved parking lots of the two medical office buildings that were built over the
easternmost extent of the Tobin Mine

a water problem. There had been a few small, localized roof The presence of the lineaments (fracture zones) alone
collapse of Type A, similar to those in the western portion of may not have been significant. However, the combination of
the mine, but there was no large area of collapse such as the the lineaments along with the concentration of runoff from
CCA. There was no reason to suspect surface subsidence 1.6 ha created the mechanism to trigger collapse and subsid-
would occur. However, subsidence did occur over much of the ence at this site.
1.6 ha, most obvious in the paved parking lots resulting in a In cases where we have had the opportunity to review
undulating surface with as much 2 m of subsidence (Fig. 25.23). site-specific conditions, a geologic flaw is commonly found
The medical offices were abandoned and removed. to be present along with possible cultural features which
An analysis of the aerial photos (Beccasio 1988) indi- impact site conditions. When such conditions are identified,
cated the presence of a major north-northwest lineament they can often be avoided or managed by proper engineering.
crossing the northwest portion of the medical complex and a
minor northeast lineament crossing the center of the medical
complex. These lineaments intersected at the location of the 25.6 The Mine Backfilling Program
medical facility buildings.
Three factors were identified that in combination likely Although the mine had been assessed as being stable, the
contributed to triggering the collapse and subsequent surface responsible State agency had required that the mine be
subsidence at this location. backfilled to insure the long-term stability of the mine. Their
specifications included:
• First the site was located over or near the intersection of
two photolineaments (fracture systems) which most likely • Filling the mine 30 m beyond landfill boundary.
resulted in a zone of weakened rock which could have • Filling the mine to at least 90 % full (or 3.8 m of the 4.2 m
extended downward to the level of the mine. high mine)
• Second the two buildings and their associated parking lots • Fill material was to have a strength of 1.5 MPa.
covered an area of approximately 1.6 ha. This resulted in
a large amount of surface water runoff from the building
roofs and paved parking lots being concentrated over
these fracture zones of weakened rock. 25.6.1 Initial Efforts with Fly-Ash
• Third, there was little or no loess soil cover at this site to
slow infiltration of rainfall and runoff. The highly frac- A major effort of mine backfilling with fly ash began in late
tured Argentine Limestone was at the surface providing a 1987. The mixture consisted of approximately 816 kg of type
permeable zone for rainfall and runoff to penetrate. F fly ash, 90 kg of kiln dust, 90 kg of bottom ash, and 360 l
360 25 The Development of a Landfill over an Abandoned Limestone Mine

of water. This slurry was created at an on-site facility using velocity required to keep the solid particles in suspension
material waste from local power companies. It was mixed in and to transport them away from the borehole once they had
cement trucks and dumped into the mine via 20 cm open entered the mine. Furthermore, the crushed rock slurry
boreholes. method did not set up and block the fillhole as did the fly ash.
While such a fly-ash mixture is an excellent backfilling This allowed the filling operation to be conducted during a
material, it will harden in a short period of time much as 10-h shift per day (Goodson Associates 1988).
cement. The fly-ash would set up and harden over evenings As the slurry first enters the open mine from the injection
and weekends and would, at some fillhole sites, form a zone hole, its velocity drops rapidly and solid particles settle out
of conical fill under the borehole choking off the borehole and near the borehole, forming a doughnut-shaped mound on the
preventing further filling of the mine. A variety of alternatives mine floor (Fig. 10.1a). As the height of the mound
were considered including a 24 h per day batch plant to mix approaches the mine roof, the velocity of the slurry increases
and transport the fly ash to the mine via surface piping. through a narrow channel, and solid particles are transported
to the outer diameter of the mound. Here the velocity
decreases abruptly and solids are deposited.
25.6.2 Crushed Rock Slurry Backfill As the mine backfill material begins to fill the mine to
the roof, the flow of water and rock cut a new path to the
An experienced mine consultant, Goodson Associates, Inc., perimeter of the fill. At some point friction begins to increase
of Denver, Colorado recommended the process of using a and a new channel is cut. This process of cutting a new
crushed rock slurry. Based upon favorable economic factors, channel occurs about every 20 min and continues in a
the owner elected to employ the crushed rock slurry backfill random fashion around the perimeter of the fillhole placing
program, which provided three benefits: rock to the roof of the mine until an area of about 22–30 m
radius around the fillhole was backfilled to the mine-roof
• First, it removed the highly weathered and fractured (100 % full) (Fig. 10.1). At this point, the friction loss of the
uppermost limestone, the Argentine-Frisbie Limestone, rock slurry between the rock fill and the mine-roof became
from immediately beneath the landfill. greater than the head of slurry being injected and further
• Second, the on-site Argentine-Frisbie Limestone was an filling ceases.
existing on-site resource of backfill material. Fillholes were spaced approximately 45–60 m apart in
• Third, by removal of about 7.6 m thick Argentine-Frisbie order to provide 90–100 % filling of the mine. The amount of
Limestone, the volume available for the landfill was sig- fill placed in each hole would vary from 7,600 to more than
nificantly increased and the cost benefits essentially paid 15,300 m3 assuming a density of 1.4 tons/yd3) depending
for the backfilling of the mine. upon the location of nearby mine walls and previous fill
material. The slope of the fill (angle of repose) ranged from
The initial base of the landfill was designed to lie upon the 4 to 9° and averaged 6°.
loess soil or on a clay cover on top of the Argentine-Frisbie
Limestone. Once the Argentine Frisbie-Limestone was
removed, the landfill cells were then paced on a plastic liner 25.6.3 A QA/QC Program for Mine Backfilling
over the 10.6 m thick Lane Shale (Fig. 25.2).
The upper layer of highly fractured and weathered A Quality Assurance Audit program and Quality Control
Argentine-Frisbie Limestone (about 7.6 m thick) was procedures were developed and included:
removed using a D9-L caterpillar impact ripper with a single
tooth. The rock was delivered to a crusher that provided • Direct observation, measurements along with photos and
1.25 cm rock, which was then mixed with water and pumped video documentation of the early backfilling.
via piping to a mine fillhole and injected as mine backfill • Quality control monitoring included mapping the perim-
material. The large volume of water used was obtained from eter of the fill after each backfilling sequence. A contour
the mine and was re-circulated as rock and water were map of fill was developed based upon these observations
injected into the mine. and measurements. Photographs were taken of the entire
The crushed rock pumped slurry method differs from accessible perimeter of backfill to document conditions.
open-gravity feed methods previously used with fly-ash. The • Quantities of fill were recorded daily and
pumping energy is used to achieve a dynamic suspension of • Physical property measurements of the mine backfill
solid crushed limestone in water and the system is essentially material were made and included sieve analysis, Los
closed from the point of slurry mixing to the bottom of the Angeles (LA) abrasion test, minimum index density,
injection borehole at 5,677 l/min. The energy provided by maximum index density, and a California bearing ratio
the pump and the static head in the borehole provided the test.
25.7 Conclusions 361

Fig. 25.24 In-situ compression tests were made on the crushed rock fill material

25.6.3.1 In-Situ Compression Tests of Fill • The crushed rock backfill strength is more than 15 times
Material the amount required to resist compression in the theoreti-
In-situ compression tests were carried out on both fly-ash cal worst-case instance of a completely failed rock
and crushed rock backfill which had been placed within the overburden;
mine. The State required a minimum of 1.5 MPa for the • The fly-ash backfill strength is more than 3.5 times that
backfill materials. Figure 25.24 shows the acquisition of required.
compression tests over crushed rock. The average in-situ
value for the tests on crushed rock was 24.7 MPa with an In either case, the backfill material has adequate strength
average vertical displacement of 0.0155 cm and no indica- and an unusually high factor of safety against resistance to
tion of failure. A strong limestone typically has unconfined worst-case (but unanticipated) mine stability conditions.
compressive strength of 103.4–137.9 MPa. The average The crushed rock slurry program proved successful and
maximum in-situ values for the two tests on fly-ash was provided a dramatic increase in the degree of lateral and ver-
5.5 MPa with an average displacement of 1.7 cm. These tests tical filling to achieve closure of the open mine space. Once
on fly-ash and crushed rock provided results well above the the permanent roof-void space was reduced to near-zero, the
requirements of the 1.5 MPa by the State. perceived risk of further subsidence was terminated. While
both fly-ash and rock slurry proved to be good materials,
from a strength point of view, the economics of the addi-
25.6.4 Summary of Mine Backfilling tional landfill air space gained by removal of 7.6 m of lime-
stone made the crushed rock slurry not only a technical
The backfilling of the mine eliminated the State agency’s success, but an economic one as well. Further details are
concerns for long-term stability of the mine. While the state given in by Benson et al. (2000).
requirements were to backfill the mine to 90 %, the crushed
rock slurry method provided essentially 100 % backfill to the
mine roof over a radius of 22–30 m and fill to 90 % to the 25.7 Conclusions
mine-roof over distances up to 68 m from the fillhole.
A fill material strength of 1.5 MPa was required to sup- The work at this site was carried out over a period of 7 years
port the weight of a theoretically completely failed rock and provided abundant site-specific assessment of geology,
overburden (a maximum of 54 m of rock and a maximum hydrology and mine conditions over time. Hundreds of hours
landfill thickness of approximately 36 m). were spent both above ground and underground in the mine.
362 25 The Development of a Landfill over an Abandoned Limestone Mine

At most sites we will not have nearly the time or data as we that significant subsidence is highly unlikely and that the
have at this site. backfilling of the mine will prevent any significant surface
During this 7 year period there had been four different site subsidence.
engineers involved and the ownership had changed three This case history uses knowledge from extensive investi-
times resulting in a loss of corporate memory as well as data gation of the Tobin mine and other mines in the Kansas City
files and reports. However, the same senior staff was involved area. A series of papers have been presented on the work at
with the site characterization over the entire 7 year period. this site and others in the area. They include:
This continuity was a key factor in acquiring the data and
integrating the many pieces of the geologic puzzle at this • Hatheway et al. (1990)
site. • Benson et al. (1991a)
None of the issues identified in this case history required • Benson et al. (1991b)
a high level of analysis. Simple observations and measure- • Benson et al. (2000)
ments along with an assessment and integration of a variety • Benson (2001)
of data were used to piece together the geologic and hydro- • Benson and Hatheway (2001)
logic puzzle of the site conditions. • Hatheway et al. (2001)
Existing literature from Hasan et al. (1988) provided a
solid basis for understanding of the local geology hydrology The landfill operations have since been completed and the
of the region and specific mine conditions. Gentile (1984) landfill has been covered over. In addition, the mine portal
had reported the presence of paleocollapse from within the road has been filled with debris and the portals covered over.
deeper Mississippi Limestone, which became the basis of
our conceptual model. The contributions of other mining
experts Golder Assoc. (1986), Tein (1988), Brink (1989), References
Colaizzi (1990 personal communication), and Camp (1988
personal communication) also provided support for this Aley T (1991) Results of spectrofluorometer analysis of dye samples
work. from charcoal bugs. Consultant report
Anderson Survey Co. (1990) Subsidence monument measurements.
A preliminary conceptual model of what we feel caused Consultant report
the surface fissure and subsequent mine collapse is shown in Beccasio D (1988) Photogeologic interpretation of the Forest View
Fig. 25.6. This conceptual model was developed based upon Landfill Site, Kansas City Kansas. Consultants report
work by Gentile (1984) along with observations from trench- Benson RC (2001) Strategies of site characterization and risk manage-
ment. An invited presentation to the National Research Council,
ing and within the mine. Subsequently the preliminary con- committee on coal waste impoundment failures, subcommittee on
ceptual model was supported by impoundment site characterization, Washington DC
Benson RB, Hatheway AW (2001) A systematic approach for evaluat-
• Aerial photo interpretation that identified evidence of fis- ing subsidence risk. Presented at the AEG/AIPG annual meeting,
St. Louis, MO
sures in the 1954 photos indicating that the surface fis- Benson RC et al (1991a) Site characterization for a landfill located over
sures were present prior to mining. a mine with superimposed paleokarst collapse, a case history. In:
• The circular depression with up to 4.5 m of elevation Proceedings of the 3rd conference on hydrology, ecology, monitor-
change located approximately centered over the CCA ing and management of groundwater in karst terranes, Nashville,
4–6 Dec, pp 131–145
(Fig. 25.12). Benson RC et al (1991b) Ground water monitoring in unsaturated and
• Exposure of old fissures in the Lane Shale at an excava- saturated zones at a site with paleocollapse structures. Presented at
tion that cut across the west portion of the CCA the annual meeting and 10th anniversary of the American Institute of
(Fig. 25.11). Hydrology International conference on hydrology and hydrogeol-
ogy, 3–7 November 1991, Orlando, FL
Benson RC et al (2000) Stabilization against geologic instability –
There is sufficient data, based upon on-site and in mine underground documentation of a mine backfilling case history.
measurements, observations, along with opinions of inde- ASCE Geo-Denver Conference, Colorado, 5–8 Aug
pendent experts to provide a reasonable degree of confidence Boston Survey Consultants (BSC) (1990) Design and analysis of the
subsidence monitoring network at Forest View Landfill. Consultant
in the conceptual model of the mine and surface conditions. report
Two worse case scenarios are presented where subsidence Brink T (1989) Inspection of the Tobin mine. Consultant report
has occurred under identical geologic conditions. When such Dunnicliff J, Green GE (1988) Geotechnical instrumentation for moni-
subsidence occurs under circumstances in which it would not toring field performance. Wiley, New York
Ford D, Williams P (2007) Karst hydrogeology and geomorphology.
have normally been be predicted, it is a combination of a Wiley, Chichester
geologic flaw and possibly the impact of cultural conditions Gentile RJ (1984) Paleocollapse structures: Longview Region, Kansas
that have caused the subsidence. A risk assessment indicated City, Missouri. Bull Assoc Eng Geol 21(2):229–247
References 363

Golder Associates (1986) Mine backfilling requirements for the Forest Consultants report for the Missouri Department of Transportation,
View Landfill Site, Kansas City, Kansas. Consultant report unpublished
Goodson Associates, Inc. (1988) Proposal for mine backfilling using Hem JD (1989) Study and interpretation of the chemical characteristics
crushed limestone. Consultant report of natural water. US Geologic Survey, Water Supply Paper 2254
Hasan S et al (1988) Geology of greater Kansas City, Missouri and Nicholson M (1988) Former Tobin mine superintendent. Notes from per-
Kansas, United States of America. Bull Assoc Eng Geol sonal interview with Hatheway AW at Forest View Landfill, July 12
25(3):281–341 O’Connor HG (1971) Geology and ground water resources of Johnson
Hatheway AW et al (1990) Geologic factors responsible for mine County, northeastern Kansas. Kansas Geological Survey Bulletin 203
collapse under a portion of the Forest View Landfill Site, Kansas Quinlan J (1991) Results of spectrofluorometer analysis of water and
City, Kansas. In: 33rd annual meeting of the Association of dye samples. Consultant report
Engineering Geologists, Engineering Geology for the 90’s, Tein JC (1988) Mine stability at Forest View Landfill. Consultant Report
Pittsburgh, PA, 1–5 Oct Tuttle, Ayers, and Woodward Co. (1979) Plat map of J.A. Tobin mine
Hatheway AW et al (2001) Geotechnical risk assessment of US Route showing time development of mine boundaries. Consultants report.
169 over the Briarcliff Mine, Kansas City, Clay County, Missouri. Kansas City, Missouri
 Site Characterization Along Bridge
Alignment 26

Abstract
The second case history involves a site characterization for a new bridge into the Florida
Keys. There was concern that there may be paleosinkholes present because of the alignment
of four circular shaped lakes. The presence of paleosinkholes could impact the foundation
for the new larger and higher bridge over Jewfish Creek between the Florida mainland and
Key Largo in the upper Florida Keys. The site characterization effort was limited to the
existing two-lane road and the surrounding areas that were covered in water. A major drill-
ing and sampling effort had already been completed as part of the Florida Department of
Transportation (FDOT) geotechnical investigation. The project was carried out in 1994 over
a period of 6 months in three phases following our standard site characterization strategy.

26.1 Background Lake Surprise. The drilled shaft foundations will extend to
depths of about 18 m.
A new bridge was being planned by the Florida Department There had been virtually no sinkhole activity reported in
of Transportation to ease traffic flow into and out of the this area of the Keys, and initially there was no reason to
Florida Keys, particularly during evacuations due to a pend- suspect karst in the upper Florida Keys. However, a review of
ing hurricane. Access to the Florida Keys has been by two marine charts and aerial photos clearly shows Lake Surprise
2-lane narrow roads crossing from the lower tip of Florida’s as one of four circular lakes in alignment along a northeast–
mainland across extensive bays and mangrove covered areas. southwest trend (Fig. 26.1). Such an alignment is common
Card Sound road to the northeast has a high bridge crossing for sinkholes throughout central Florida often indicating the
the Florida Intracoastal Waterway and leads to the northern presence of an underground cave system. A project manager
portion of Key Largo. The Florida Keys Scenic Highway, with the Federal Highway Administration (FWHA) had seen
US1 crosses the bascule bridge constructed in 1944 at the linear trend of circular lakes on maps and had raised the
Jewfish Creek and then runs across Lake Surprise entering issue of possible sinkhole problems at the site. However in
the middle of Key Largo (Fig. 26.1). The majority of traffic the Florida Bay area and the Keys such patterns are often
into or out of the Florida Keys uses US1. During a hurricane referred to as mangrove lakes and are not thought to be asso-
all traffic must exit the Keys by one of these two roads. ciated with sinkholes (Shinn 1994a).
The planned new bridge must span Jewfish Creek, at the US-1 is a narrow two-lane road constructed on fill through
existing bascule bridge and cross Lake Surprise, a circular mangroves and Lake Surprise (Fig. 26.1). There are limited
lake with US-1 passing though its center for about 1,066 m. areas of developed land both sides of Jewfish Creek, which
The new bridge will begin northwest of the present bascule are associated with two marinas and a condominium. Other
bridge at Jewfish Creek, then extend southeast to Key Largo than the narrow two-lane US-1, the remainder of the site
for a distance of more than 2,400 m (Fig. 26.1). Drilled shaft consists of either open water that is accessible by boat, or
foundations for the bridge will be placed over Jewfish Creek covered by mangroves with no practical access. Local wild-
(the highest portion of the bridge) and through the center of life, such as the saltwater crocodile, were also of concern.

© Springer Science+Business Media Dordrecht 2016 365

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_26
366 26 Site Characterization Along Bridge Alignment

State of
Florida
Alignment of four
circular lakes

Site Location

US
1

Old Bridge
e
ris
rp
Su
k

Approximate
ke
ee

La

Extent of
Cr

New Bridge
ish
wf
Je

Location of large
sinkhole within the
Florida reefs
8 km to southeast
go
ar
yL
Ke

0 600
Approximate
Scale in Meters

Fig. 26.1 Site location map showing the extent of new bridge construction over Jewfish Creek and Lake Surprise along with the alignment of four
circular lakes
26.3 The Approach 367

26.2 An Initial Site Assessment extensive road maintenance was routinely carried out by
DOT at this location, (personal communication with the
Prior to beginning the site characterization effort there was a manager of Gilberts Marina and the Jewfish Creek bridge
review of findings by USGS, and an initial site visit. tender, 1994).
A meeting with the Florida Department of Transportation
(FDOT) engineers indicated that extensive road maintenance
26.2.1 Findings by USGS due to continuing subsidence had been carried out at that
location over a number of years. The FDOT engineers had
The United States Geological Survey (USGS) had located a explained the subsidence was due to decaying timbers, which
large 594 m diameter paleokarst sinkhole collapse filled with had been encountered by shallow borings (Miro 1994 per-
unconsolidated materials among the reefs off of Key Largo, sonal communication). The timbers were used in the con-
8 km to the southeast of Jewfish Creek (Fig. 26.1). The sink- struction of Flagler’s overseas railway to the Florida Keys
hole is located in about 6 m of water within the Key Largo completed in 1912.
National Marine Sanctuary. USGS has cored to a depth of Given no other data or information one would consider
32 m and wash bored to a depth of 55 m finding only uncon- this simply as a routine road maintenance problem. The prob-
solidated fill and no solid rock. The oldest C14 age from the lem with the explanation was that wooden piles and timbers
jet probe was 5,650 +/−90 year before present (BP) (Shinn had been used all along this area on US 1 to construct Flagler’s
et al. 1994b). The 594 m diameter and minimum depth of overseas railway. However, this 15 m long area was the only
55 m suggest a very large paleosinkhole. The proximity of location where such subsidence was occurring, there were no
such a large paleosinkhole within the Florida Keys implies other such maintenance problems in the area along the old
the possibility of other deep cave systems in the area. railway line form the Florida mainland to Key Largo.

26.2.2 Initial Site Visit 26.3 The Approach

During the initial site visit to establish boat launching facili- The FHWA, who provided guidance and funding for the
ties, an obvious patch about 15 m long was noted in the project, initiated a karst investigation of the site. The purpose
southbound lane of US-1 about 243 m south of Jewfish Creek of this investigation was to evaluate the possible presence
bridge (Fig. 26.2). The Jewfish Creek bridge tender and the and impact of significant paleokarst (old sinkholes or
nearby business owner at Gilberts Marina indicated that cavities) along the alignment of the proposed bridge to be

Fig. 26.2 A patch about 15 m long was noted in the southbound lane of US-1. The Jewfish Creek lift bridge is seen in the top of the photograph.
The view is towards the northwest
368 26 Site Characterization Along Bridge Alignment

Initial Site Assessment

PHASE 1
Reconnaissance Efforts
(cover large area to detect any
indications of unusual conditions)

PHASE 2
Confirmatory Measurements
(unusual conditions indicated, confirm
existence and estimate lateral extent,
if possible)

PHASE 3
Detailed Measurements
(anomaly confirmed, identify possible
cause and potential impact)

Risk Assessment
and Recommendations

Preliminary Conceptual Model Final

Conceptual Model (supported by Conceptual Model
additional data) (supported by multiple
sets of independent data)

Fig. 26.3 The overall organization of tasks for the three-phased project

built over Lake Surprise and Jewfish Creek. The initial focus 26.4 Phase I Reconnaissance Investigation
was upon Lake Surprise that was thought to be a possible
old sinkhole. Phase I reconnaissance efforts consisted of:
Two critical questions were asked at the beginning of
this study: • A review of geologic literature and site geomorphology;
• A site fly over to assess site access;
• Can an investigation fully identify unstable geology, such • Reconnaissance marine seismic reflection through Lake
that bridge decks can be placed or lengthened to safely Surprise;
span the critical geologic conditions? • A microgravity survey along US1 from Key Largo across
• Can the bridge as designed, be stable for at least 100 years Lake Surprise and extending northwest of Jewfish creek.
in the given geologic environment? • An aerial photo lineament analysis; and
• A detailed review of existing boring data for possible
Figure 26.3 shows the organization and flow of the overall indications of karst conditions.
project. A phased strategy was developed to investigate the
possible karst conditions at this site. The investigation would
consist of a three-phased approach:
26.4.1 Review of Regional and Local Geology
• Phase I was a reconnaissance phase to assess whether or
not potential karst problems existed; Sources of information for regional geologic conditions
• Phase II was a confirmation phase, to verify and expand came from work by Parker et al. (1955), Lane (1981), and
findings of anomalous conditions that might be found in Scott (1988). Additional information was obtained from
Phase I; deeper regional investigations were being completed in
• Phase III was a detailed investigation to specifically assess southeast Florida and the Florida Keys by the University of
the impact of any karst conditions found. Miami (McNeill et al. 1995).
26.4 Phase I Reconnaissance Investigation 369

Layer 1
Organics, peats, silts and fill
(varies from 0 to 6 m thick, avg 3 m)
0
Layer 2 - Limestone
Key Largo Limestone (typically 21 m deep) Based upon site
specific boring data
Layer 3 - Hard Pan Layer (Law Eng 1994)

Layer 4 - Sand

Long Key Member of the Peace River Formation

100 (firm fine sand with some calcareous fragments Confining
Beds
and occasional sandstone)
Approximate Depth in Meters

Based upon
regional data
(depths are
approximate)

Limestone

200 Arcadian Formation Floridan

of the Hawthorne Group Aquifer
(the principle component
of the Floridan Aquifer)

Zone of possible
cavern development

300
Suwannee Limestone Formation

Fig. 26.4 The geologic setting for the site

Local geology was provided by the drilling data acquired • Layer 2 is a vuggy limestone, which averages about 21 m
for the bridge design. Prior to beginning our work, Law deep over much of the site. This layer is equivalent to the
Engineering (1994) had completed a series of 34 borings Key Largo Limestone which is a coralline reef rock rang-
over a distance of 2,621 m. The borings were spaced on aver- ing from hard and dense to soft with open voids. The
age 76 m apart along the bridge right-of-way and had been spaces between and around the coral heads are filled with
drilled to depths of 20–47 m with the majority extending limestone debris, cemented carbonate sands and unce-
to 30 m. mented carbonate sands.
The stratigraphic sequence at the site is shown in Fig. 26.4. • Layer 3 is a significant hard-pan layer present beneath the
Four geologic layers had been described as follows: Key Largo Limestone at a depth of about 21–30 m. This
hard-pan is the Caloosahatchee equivalent (Scott 1995
• Layer 1 consists of water, organics, peat, silt and fill mate- personal communication). It consists of sandy marls, clay,
rials that varies from 0 to 6 m thick, but is typically less silt, sand, and shell beds and generally has low
than 3 m thick. permeability.
370 26 Site Characterization Along Bridge Alignment

• Layer 4 consists of a firm, fine quartz sand with some • The proximity of the large sinkhole identified by USGS
calcareous fragments and occasional sandstone. Only 4 of only 8 km from Lake Surprise (Shinn et al. 1994b) clearly
the 34 existing borings were drilled deep enough to implies the potential presence of other large deep cave
encounter this layer. This layer extends from approxi- systems and paleocollapse in the area.
mately 30–150 m deep; and is the Long Key Member of
the Peace River Formation (McNeill et al. 1995). While both regional and local evidence suggests dissolu-
tion has occurred in the area, the karst features typically
Below Layer 4 between approximately 150 and 290 m, expressed at the surface such as subsidence and collapse are
the Arcadian Formation of the Hawthorne Group limestone not readily visible in South Florida. This is due to the infill-
occurs (Scott 1988) (Fig. 26.4). Deep borings have encoun- ing of karst features with sediments and their stability due to
tered the Arcadia Formation both north and south of Jewfish the very shallow water table in the area. Surface elevation on
Creek. This is the first limestone in which a large cave sys- Key Largo is about 2.4 m above MSL.
tem could develop. The Suwannee (limestone) Formation
lies below the Arcadian Formation (McNeill et al. 1995).
26.4.3 A Site Fly Over

26.4.2 The Regional Geomorphology A site flyover was made prior to starting fieldwork. The aer-
ial observations provided insight to access and limitations in
Large surficial karst features or deeper caves have not been and around the mangroves and identified any roads, trails, or
recorded in south Florida and in the Florida Keys as they are areas of dry land that may be accessible which are not indi-
in central Florida where active subsidence and collapse is cated on charts or topographic maps of the area. While stan-
on-going and readily noticed. On a regional scale, the nearest dard marine charts provided a feel for the access by boat and
large open sinkhole occurs in the Sarasota County 290 km topographic maps identified a few dirt roads and trails, the
northwest from Lake Surprise. Warm Salt Spring (in Sarasota fly over provided a much better feel for the limitations and
County) is well known as an archaeological site (Ferguson areas of access including dirt roads and trails within the man-
et al. 1947). Other examples include the numerous Blueholes groves as well as small channels not shown on the marine
on Andros Island in the Bahamas 290 km to the east. Dean’s chart or topographic map. The oblique photos acquired were
Bluehole on Long Island about 442 km to the southeast in the also useful for describing conditions to other members of the
Bahamas is at least 202 m deep and is described as the deep- field investigation team and in planning fieldwork.
est blue hole in the world by Wilson (1994).
From a review of regional geomorphology we find that
sea levels had been lower by as much as 130 m (Balsillie and 26.4.4 Reconnaissance Marine Seismic
Donoghue 2004). Dissolution and cave development is Reflection Survey
known to occur at the base of the fresh water lens (the fresh
to salt water contact) (Ford and Williams 2007). The combi- The Phase I reconnaissance efforts with seismic reflection
nation of much lower sea levels and a thick fresh water lens were focused within Lake Surprise since it was thought to be
would have created an environment prone to the dissolution a possible paleosinkhole lake. A total of 13.8 km of recon-
of limestone and development of caves at depths of 180– naissance seismic reflection data was obtained from both
240 m or more below present sea level. north and south of US-1 on Lake Surprise (Fig. 26.1).
On a much more local scale, there are a number of factors Data was obtained from a small boat using an EG&G
which indicate that dissolution has occurred in the area. Model 230 Uniboom system. Reflected signals were received
by a single channel eight element hydrophone array. The
• Numerous springs have been identified over 100 years data was recorded with a record length of 60 ms for a depth
ago within Biscayne Bay and offshore of Miami just of about 63 m based upon an estimated velocity of 2,133 m/s.
64 km north-northeast of Lake Surprise. The presence of Although there is some lateral coverage to each side of the
springs indicates conduits with much greater porosity survey line, from a practical point of view the seismic reflec-
than the surrounding materials. Some of these appear to tion data can be considered to be from directly under the
be in-filled sinkholes or blue holes (Wanless 1994 per- boat.
sonal communication). The seismic reflection data over Lake Surprise indicates
• Evidence of cavities and voids taking considerable grout horizontally bedded sediments. No obvious anomalies that
has been encountered in the borings drilled for the geo- might indicate the presence of karst conditions were identi-
technical foundations of tall buildings and bridges in the fied in this data. While the 13.8 km of survey line provided a
Miami-Ft. Lauderdale area (Berkovitz 1994 personal reasonable coverage to detect possible dipping strata associated
communication). with a sinkhole, the throat of a sinkhole is quite localized and
26.4 Phase I Reconnaissance Investigation 371

200 Northwest Southeast

150
Lake Surprise
Jewfish Creek
100
10 - 25 uGal Center of Anomaly 10 - 25 uGal 45 uGal
Gravity in microGals

50
anomaly (station 1221+00) anomalies anomaly

-50

-100
Large 100 microGal
Large 175 microGal
Gravity anomaly
-150 gravity anomaly

-200

Reference Stations (distance in feet)

0 300

Distance in Meters

Fig. 26.5 The microgravity data along a 2,621 m survey line along US1 is shown in profile, with each station value shown as a dot

can lie anywhere within the sinkhole lake and might have There is a large negative gravity anomaly of 175 microGals,
been missed by this reconnaissance seismic survey. that extends more than 900 m. This anomaly begins near the
northwest edge of Lake Surprise and extends northwest of
the Jewfish Creek bridge (Fig. 26.5). This anomaly consists
26.4.5 Microgravity Data of two independent anomalies, which are superimposed
upon one another.
A microgravity survey was run along the edge of the north- The large gravity anomaly of 175 microGals centered at
bound lane of US-1. The survey extended from Key Largo Jewfish Creek can be explained by the increasing elevations
across Lake Surprise and northwest of the Jewfish Creek of gravity stations along the built-up roadway approaching
Bridge. This covered the entire extent of the proposed new the bridge, plus the existence of the deeper channel at Jewfish
construction (a total of 2,621 m), however the focus was Creek Bridge. Therefore, this large anomaly is a function of
upon Lake Surprise. Gravity measurements were obtained at local changes in topography and can be disregarded. After
intervals of 7.6 m over Lake Surprise and intervals of 15 m removing the topographic effects, a smaller negative gravity
beyond Lake Surprise for a total of 245 gravity measure- anomaly of 100 microgals remains (Fig. 26.5). This very sig-
ments. Gravity measurements were made at night to mini- nificant 100 microgal anomaly is centered at Station 1221+00
mize the noise from traffic along the narrow two-lane road. (FDOT station numbers in feet) and is the only significant
A Lacoste and Romberg Model G microgal gravimeter with anomaly identified by the microgravity survey.
electronic readout by EDCON was used to acquire the data. The simplest possible cause of the gravity anomaly would
The microgravity data for the Phase I survey is shown in be a localized increase in depth to rock with organics or
Fig. 26.5. There is a level of geologic and vehicle noise pres- loose sediments above rock. But ten boreholes within the
ent in the microgravity data. This noise level is estimated to large 100 microgal gravity anomaly do not indicate a deep-
be about ±10 microGals. Therefore, only anomalies above ening of the top of rock. In addition, these ten boreholes
20 microGals and with a consistent trend are considered to show no sign of a major void or other significant geologic
be significant. feature that could account for a gravity anomaly of this mag-
There are two small gravity anomalies of 10–25 micro- nitude, indicating that the source of the anomaly is deeper
Gals over Lake Surprise and one negative anomaly of than 30 m, the depth of the borings.
45 microGals at the southeast end of Lake Surprise. Another
small gravity anomaly of 10–25 microGals is seen at the
northwest end of the survey line. Each of these can be 26.4.6 Aerial Photo Analysis
explained by variations in the depth to rock seen in the bor-
ing logs and the presence of peat (a lower density material) While photo analysis normally uses aerial images over land,
within these low zones and as a result they can be ignored. images over shallow water-covered areas can also be used,
372 26 Site Characterization Along Bridge Alignment

often to depths of 10 m (Finkl and Warner 2005). In this case, detailed review of the existing boring data was made to
the aerial photo analysis was carried out over an area that is assess indications of karst conditions. This data included the
covered by a combination of shallow water, from 1.2 to 3 m 34 existing borings by Law Engineering (1994). Core boxes
deep, and by mangroves. were made available and selected cores were inspected. The
An analysis of aerial photos was carried out to detect the geologic conditions seem to be reasonably well defined, but
presence of photo-lineaments (Finkl 1994). The IDRISI at this point the data had not been interpreted in terms of
image processing – raster GIS analysis program was used. karst conditions.
Lineaments are significant linear features within the land-
scape that may be structurally controlled (Billings 1954) and 26.4.7.1 Standard Penetration Tests
are usually associated with one or a combination of joints, Standard penetration tests (SPT) were run in all 34 borings
fractures, dikes, faults, geologic boundaries, or depositional and were reviewed for patterns and trends within each of the
history or geomorphology (Bagdley 1960, 1965). In the case geologic layers (Fig. 26.6).
of karst, a photo-lineament is a near surface indication of a SPT values within Layer 2, the Key Largo Limestone,
possible fracture zone along which a cave system may be range from 2 to 98 with an average of 28. They appear to be
developed. Existing black and white stereo-paired aerial randomly distributed both vertically and laterally over the 34
photographs (2-12-1991) were obtained from the Florida borings. The base of the Key Largo Limestone is found at an
Department of Transportation for the general area. Two inde- average depth of about 21 m.
pendent aerial photo analyses were carried out, one manually STP values within Layer 3, the hard-pan (Caloosahatchee
using a stereoscope by an experienced aerial photo inter- equivalent) are typically 50 blows per about 7 cm of penetra-
preter and another by computer analysis. Both analyses iden- tion defining a very hard layer. These SPT values are quite
tified a major NE-SW lineament centered at station 1224+00, uniformly distributed both vertically and laterally over the
about 100 m northwest of the center of the gravity anomaly. 34 borings. The base of Layer 3 is found at an average depth
The focus now shifts to the narrow channel between Lake of about 30 m.
Surprise and Jewfish Creek based upon the microgravity STP values within Layer 4 (fine quartz sand) range from
anomaly, and the location of the photo-lineament. 5 to 70 and averaging about 30. Only 4 of the 34 borings had
been drilled deep enough to encounter Layer 4.

26.4.7 A Detailed Review of Existing 26.4.7.2 Fluid Loss

Boring Data Fluid loss was noted in the drill logs as it occurred in each
borehole and indicated very porous zones or cavities. In their
As a result of the microgravity data indicating anomalous report, the drillers noted that circulation losses were gener-
conditions between Lake Surprise and Jewfish Creek, a ally abrupt. In many cases, total loss of circulation occurred

Major photo-lineament and

patch in southbound lane of US1
Center of Existing
Northwest Jewfish Creek Bridge Southeast
Center of Gravity Anomaly
PB-12

PB-10

PB-8
PB-7

0
Layer 2
Key Largo Limestone
SPT’s 2 to 98, avg 28 15
Depth in Meters

Layer 3 - Hard Pan Layer

SPT’s avg 50/about 7 cm
30

Layer 4 - Sands
very limited SPT data 45
5 to 70, avg 30

60
1260+00 1250+00 1240+00 1230+00 1220+00 1210+00 1200+00 1190+00 1180+00
FDOT Stations (Distance in feet)
- Indication of Fluid Loss

0 300
Scale in Meters

Fig. 26.6 The locations and depth of the original 34 borings along with averaged SPTs and the locations of fluid losses
26.4 Phase I Reconnaissance Investigation 373

even when using a thick drilling fluid (Law Engineering • Layer 4 the sands of the Long Key Member of the Peace
1994). The fluid losses are shown in Fig. 26.6. River Formation has three fluid losses in only one boring
Two adjacent boreholes, PB-7, with eight fluid losses and between 27 and 37 m.
PB-8, with ten fluid losses accounted for 27 % of all the
fluid losses noted in the 34 borings. Two near-by borings
PB-10 near the centerline of the Jewfish Creek Bridge and 26.4.7.3 Core Recovery and RQD Values
PB-12 just northwest of the Jewfish Creek Bridge account NX cores were obtained from 11 of the 34 borings at selected
for an additional 34 % of all the fluid losses noted in the 34 depths. Core recovery ranged from 0 to 100 % and averaged
borings. Sixty one percent of all fluid losses in the 34 borings 53 %. RQD values ranged from 0 to 79 % and averaged 20.
are concentrated within 4 borings PB-7, PB-8, PB-10, and RQD values indicated very poor to good quality rock in
PB-12 that lie between stations 1217+50 and 1234+00. those limited areas and depths sampled.
This is a distance of 503 m between the northwestern edge
of Lake Surprise and the northwestern edge of the Jewfish
creek bridge. 26.4.8 The Correlation of Anomalous
These fluid losses are distributed as follows: Conditions

• Layer 2 the Key Largo Limestone whose base occurs at a The shallow seismic reflection data and gravity data did not
depth of about 21 m has 21 fluid losses. indicate anomalous conditions within Lake Surprise.
• Layer 3 the hard-pan (Caloosahatchee equivalent) However, the nature and spatial correlation of the following
whose base occurs at a depth of about 30 m has 16 fluid four independent sets of site-specific data (Fig. 26.7) pro-
losses. vided strong evidence for potential karst conditions between

Lake Surprise

Patch in Road
Borings PB-7 & PB-8

Center of the
Borings PB-10 & PB-12 100 uGal gravity anomaly

Aerial Photo Lineament

Lake Surprise

0 200 400
Scale in Meters

Note: FDOT reference stationing

along road is in feet

Fig. 26.7 The location of four anomalous conditions identified in Phase I show strong spatial correlation
374 26 Site Characterization Along Bridge Alignment

Lake Surprise and Jewfish Creek. These four independent 26.5 Phase II Confirmation Phase
sets of data include:
The purpose of Phase II was to verify and expand findings of
• The wide 100 microGal gravity anomaly between Lake anomalous conditions identified in Phase I. The area for
Surprise and Jewfish Creek. This was the first indication additional measurements was focused upon the narrow chan-
of possible significant karst conditions at the site because nel about 38 m wide that connected Lake Surprise and
of the implied mass deficit (voids or low-density zones in Jewfish Creek (Fig. 26.8). Phase II confirmation phase con-
the subsurface or the possibility of a paleocollapse cavern sisted of:
at depth).
• The lineament identified in the aerial photo interpretation • Marine seismic reflection data between Lake Surprise and
that passes between the center of the gravity anomaly and Jewfish Creek which would aid in the interpretation of the
Jewfish Creek. microgravity data
• The two boreholes near the center of the gravity anomaly • Microgravity data was obtained along the south side of
PB7 and PB8 and two boreholes PB10 and PB12 near US 1 parallel to the Phase I microgravity data. Additional
Jewfish Creek that account for 61 % of all the fluid losses gravity data was obtained to the south within accessible
noted in the 34 borings. areas.
• The patch in the road due to continued subsidence and
maintenance along the southbound lane of US1 indicates This data would confirm the original findings and help to
anomalous conditions and lies about 45 m northwest of develop a conceptual model of site conditions.
the center of the gravity anomaly.

This preliminary data begins to show spatial correlation of 26.5.1 Additional Marine Seismic
four independent sets of data focusing our attention upon the Reflection Data
channel between Lake Surprise and Jewfish Creek. In addi-
tion, USGS had identified a large sediment-filled sinkhole just Four closely spaced parallel lines (about 7.6 m apart) of seis-
8 km to the southeast (Shinn et al. 1994b). These preliminary mic reflection data were obtained along the narrow channel
indications of karst conditions lead to Phase II of the project. between Lake Surprise and Jewfish Creek. Data was obtained

Fig. 26.8 Narrow channel between Lake Surprise and Jewfish Creek used for detailed marine seismic reflection measurements. The drilling barge
is seen in the background. Lake Surprise lies beyond the drilling barge. The view is towards the southeast
26.5 Phase II Confirmation Phase 375

Aerial Photo
Lineament at 1224+00

Center of
Jewfish Creek Gravity Anomaly
Northwest Lake Surprise
Obvious zone of
Southeast
Patch on Road sagging and dipping strata

1216+00
1217+00
1221+00

1218+00
1222+00

1219+00
1223+00

1220+00
0
0
Two-way travel time in milliseconds

Approximate Depth in Meters

20

40

60 64
FDOT reference station in feet
0 300

Approx. Horizontal
Scale in Meters

Fig. 26.9 A portion of one of the four shallow seismic reflection lines between Lake Surprise and Jewfish Creek showing dipping strata

with the same equipment and settings used in Phase I. Each confirming the presence and magnitude of the gravity anom-
survey line was about 396 m long and ran parallel to the aly. Additional gravity data off of the road to the southwest
microgravity survey line along US-1. The positioning of the extended the gravity anomaly at least 76 m off of US 1 to the
survey boat was tightly controlled by an array of marker southwest.
buoys that were referenced to the project survey grid. Each
of the four lines clearly indicated the presence of an anoma- 26.5.2.1 Analysis of the Microgravity Data
lous zone with sagging and dipping strata indicating subsid- The 100 microgal gravity anomaly clearly indicates a defi-
ence (Fig. 26.9). ciency in subsurface mass. A gravity anomaly can be caused
by a number of geologic conditions and we must therefore,
estimate possible causes based upon:
26.5.2 Additional Microgravity Data
• Rules of thumb;
During Phase II an additional 116 gravity measurements • Forward gravity models of simple geologic and karst
were made to confirm the anomaly and estimate its lateral geometries that are reasonable for the area; and
extent. Gravity measurements were made parallel to the ini- • Additional data which would constrain interpretations.
tial survey line within the anomalous zone along the
southbound side of US1 and within the limited area of ser- A simple rule of thumb that can be used to estimate the
vice roads and trails immediately southwest of US1 where possible depth of the condition that is causing the gravity
there was solid ground. anomaly is the half-width rule (Butler 1980). The half-width
Results from the gravity profile line on the south side of depth for this gravity anomaly assuming a spherical cavity is
the road clearly repeat the data obtained from Phase I line, about a depth of 172 m.
376 26 Site Characterization Along Bridge Alignment

Gravity in microgals
Sphere 173 m deep
Field Measurements and radius of 45 m
Modeled Response water-filled sphere
100

Fig. 26.10 Simplified geologic conditions can be modeled to fit the gravity data in order to estimate the cause and depth of the gravity anomaly

As a further refinement we can run inverse gravity models sonable conceptual model of the depth of origin of collapse
(matching modeled responses to actual field measurement). at this site. Since the first limestone in which a cave could
A model was calculated using a spherical water-filled cavity develop is the Arcadian Formation, it is reasonable to con-
at depth (Fig. 26.10). The model indicates that a spherical clude that deeper geologic conditions should be
water-filled cavity about 173 m deep with a radius of 45 m investigated.
could be a possible cause of the mass deficit. Note that an
open, water-filled sphere presents the worse case scenario in
terms of potential karst features. A collapsed sphere, filled 26.6 Phase III Detailed Investigation
with broken rock and sediments would have to be of larger
dimensions in order to match the gravity data acquired at the As a result of the additional gravity and seismic reflection
site. This is a more likely scenario based upon the overlying data in Phase II, further detailed work within the narrow
subsidence and dipping strata. No further models were run at channel between Lake Surprise and Jewfish Creek was pur-
the time. sued to provide additional data to constrain our interpreta-
Based upon regional geology a large cave system that had tion. Phase III detailed investigation consisted of:
collapsed must be located in Layer 5, the Arcadian Formation
(Scott 1988). This is the first limestone (Fig. 26.4) in which • Deeper seismic reflection data over the anomalous area,
a large cave system could develop. The top of the enlarged • Measurements of seismic velocity
cavern would occur at a minimum depth of 150 m (the top of • Drilling of four boreholes within the anomalous zone to
the Arcadian Formation but could have developed at a depth depths of 30–60 m with continuous sampling, and
somewhere between 180 and 240 m or more. This is consis- • Geophysical logging in the four new boreholes
tent with the depth of other known cave systems, such as the
Bluehole on Long Island in the southern Bahamas which is
about 198 m deep (Wilson 1994) and lower sea levels of as
much as 130 m (Balsillie and Donoghue 2004). 26.6.1 Deeper Seismic Reflection Data

The shallow seismic reflection data clearly indicated the

26.5.3 A Conceptual Model presence of a paleokarst collapse zone by dipping strata, but
the seismic data only extended to a depth of about 60 m.
A conceptual model was developed (Fig. 26.11) and is sup- Seismic reflection data was obtained in Phase III to provide
ported by the additional dipping strata to depths of 63 m seen deeper data. A single survey line (approximately 600 m
in the shallow seismic reflection data between Lake Surprise long) was run in the channel between Lake Surprise and
and Jewfish Creek (Fig. 26.9) and the depths obtained from Jewfish Creek.
modeling of the gravity data (Fig. 26.10). Data was obtained with an EG&G model 267 three-
The seismic data indicated subsidence and dipping strata element sparker array and the signal was received by a single
to a depth of 60 m or more indicating that the source or cause channel eight-element hydrophone. A record length of
of this anomaly was probably deeper than 60 m. The com- 250 ms was used at a sample rate of 125 μs. The depth scale
bined estimated depth of 173 m from the gravity models is based upon an estimated average seismic velocity of
(Fig. 26.10) along with geologic interpretations (Fig. 26.4) 2,133 m per second and indicates a maximum depth to the
and regional geomorphology can be used to establish a rea- seismic data of 266 m.
26.6 Phase III Detailed Investigation 377

0
Key Largo Limestone

Hard Pan
Approximate depth
of bridge pilings

Sands
Subsidence, dissolution enlarged
joints, fractures or open voids
100
Estimated Depth in Meters

(Long Key Member of the

Peace River Formation)
Roof of the large cavern fails
resulting in collapse and subsidence
of overlying strata

Large Large cavern

Cavern (approximated by sphere)
200
Arcadian Limestone Formation

Cave Cave

300

Suwannee Limestone Formation

Fig. 26.11 A conceptual model can now be developed indicating a collapsed cavern probably within the Arcadian Formation at a depth greater
than 150 m. This is the first section of limestone thick enough to support development of a cave system and an enlarged cavern

The deep seismic data clearly indicated subsidence and vided regional velocity measurements on other core samples
dipping strata over a distance of approximately 488 m (from from the area. These measurements provided an average
stations 1213+00 to 1229+00). Relatively horizontal strata velocity of 2,133 m/s that was used to establish the depth
are seen on either end of the seismic record away from the scale for the seismic data. Based upon this velocity, the
paleocollapse zone. The deepest and most obvious feature of shallow seismic reflection data (Fig. 26.9) has a maximum
dipping strata is centered near 1215+00 that indicates possi- depth of approximately 63 m and the deeper seismic data
ble collapse to a depth of 240 m or more. An example of the (Fig. 26.12) has a maximum depth of approximately 267 m.
deep seismic data is shown in Fig. 26.12. This data suggests
that the origin of collapse lies within the Arcadian (lime-
stone) Formation or possibly the deeper Suwannee (lime- 26.6.2 Drilling of Four Additional Boreholes
stone) Formation (Fig. 26.4). The depth of collapse is in
keeping with dramatically lower sea levels (Balsillie and Four additional borings were drilled in Phase III. These were
Donoghue 2004) and a thick fresh water lens. located 76–120 m southeast of the center of the gravity
anomaly. Their locations were selected based upon the shal-
26.6.1.1 Seismic Velocity Measurements low seismic reflection data obtained in Phase II (Fig. 26.9).
Up to now, the depth scales for the seismic data were based TB-1, and TB-2 were drilled to 30 m and TB-3, and TB-4
upon an assumed velocity of 2,133 m/s. To determine the were drilled to 60 m. Drilling was done from a barge and
actual depth of the seismic record the seismic velocity of the continuous core was obtained from each of the four borings
P wave was measured. Two independent velocity measure- along with RQD values. The core showed considerable vari-
ments were made, one by surface seismic refraction mea- ation in RQD (Law Engineering 1995).
surements over a distance of 300 m and another by downhole These four additional borings indicate similar overall
measurements to a depth of 45 m. McNeill et al. (1995) pro- geologic conditions to those found in the original 34 borings.
378 26 Site Characterization Along Bridge Alignment

Interpreted Seismic

1212+00

1213+00

1214+00

1215+00

1217+00
1216+00
Data (centered at
Station 1215+25)
0
Layer 1: Water and Mud
Layer 2: Key Largo Limstone
Layer 3: Hard Pan

Layer 4
Approx. Depth in Meters

Uncemented Sands
Long Key Member of the
Peace River Formation

Limestone
Arcadia Formation
Of the
Hawthorne Group
(Scott 1988)

267
Approx 200 m
Fig. 26.12 A portion of the deep seismic reflection data to a depth of about 267 m is shown along with its interpretation and geologic section

However, dramatic differences in porosity were evident.

Figure 26.13 shows two short lengths of core obtained from
the Key Largo Limestone. One of the cores was obtained
over Lake Surprise (outside of the paleocollapse zone) from
PB-5 and is considered to be representative of the core from
the initial 34 borings (Fig. 26.13a). The second core is from
TB-4 and shows extensive dissolution and increased poros-
ity. This core represents worse case conditions from the last
four borings drilled within the paleocollapse zone
(Fig. 26.13b).

26.6.3 Geophysical Logs from the Four

Boreholes

A suite of three geophysical logs was run through the steel

casing for each of the four boreholes. They included natural
gamma, gamma-gamma (density) and neutron-neutron
(porosity). For simplicity, we have shown only the density
logs (Fig. 26.14). Higher measured values indicate lower
densities. The density logs in all four borings show a number
of zones of low density, some of which are open water-filled Fig. 26.13 A sample of limestone core outside of the anomalous area
voids up to 2 m thick. Based upon these four logs, a dominant within Lake Surprise (a) and within the anomalous area (b)
26.6 Phase III Detailed Investigation 379

Data obtained through

water column

TB-1 TB-2 TB-3 TB-4

Gamma-Gamma Gamma-Gamma Gamma-Gamma Gamma-Gamma
(density) (density) (density) (density)
0 0

6 6
Depth in Feet

Layer 2
12 12 Key Largo
Limestone
18 18

24 24 Layer 3
Hard Pan

30 30

36

Layer 4
42 Sand

48

54

60

Fig. 26.14 The geophysical density logs from the four additional borings within the anomalous area (see Fig. 26.15 for locations)

zone of low-density occurs at a depth between 6 and 15 m 60 m deep. As a result, the available data regarding porous
(within Layer 2, the Key Largo Limestone). It is this zone zones based upon the density logs and fluid losses is limited
that probably account for subsidence in US-1 (Fig. 26.2). both laterally and vertically.
However some low densities are seen to a depth of 21 m, the
base of the Key Largo Limestone. The hard-pan layer (Layer
3) defined by high SPT values, is also seen as a high density 26.6.4 Support for the Final Conceptual Model
zone (lower values) in the density logs.
Another deeper low-density zone is seen to occur in Our original focus was upon Lake Surprise itself as a possi-
Borings TB-3 and TB-4 that extends to 60 m. These low- ble paleosinkhole. However, there are no significant anoma-
density zones occur in the upper portion of (Layer 4) the lies in the 14 borings, in the gravity data along US-1 or in the
sandy Long Key Member of the Peace River Formation, 13.8 km of reconnaissance subbottom profile lines through
between 30 and 45 m. These are possible zones of sediment Lake Surprise. Based upon these three sets of data, Lake
piping. Surprise itself is probably not a large filled paleokarst col-
It must be pointed out that only 5 of the 34 original bor- lapse sinkhole. However, it is not uncommon for a large
ings were drilled to depths below 30 m. The four additional sinkhole lake to have a small sinkhole located off to one side
borings were all centered near PB-7 and drilled to 30 and of the lake, which could have been missed.
380 26 Site Characterization Along Bridge Alignment

Table 26.1 Summary of site-specific findings

Phase I – reconnaissance phase Key findings
A site visit Noted limited access for investigation and identified patch in the southbound
lane of US1
A site fly over Provide further assessment of limited access
USGS findings Provided evidence of deep-seated paleocollapse within the Florida Keys reef
8 km from site
Microgravity survey Identified a 100 microGal anomaly between Jewfish Creek and Lake Surprise
centered at station 1221+00 indicating mass deficit
Marine seismic reflection survey through Lake Surprise Did not locate any indications of karst within Lake Surprise
Aerial photo-lineament analysis Identified a major NE-SW lineament through site at station 1224+00
A detailed review of existing 34 original borings Identified 61 % of all fluid losses concentrated within four borings which lie
between stations 1217+50 and 1234+00
Phase II – confirmation phase Key findings
Detailed seismic reflection survey in the channel between Identified subsidence and dipping layers extending to a depth of 60 m
Lake Surprise and Jewfish Creek suggesting causes are deeper
Additional microgravity data Confirmed original findings and expanded anomaly to the southwest
Modeling of microgravity data Suggested deep-seated cavern more than 150 m deep
Phase III – detailed investigation Key findings
Deep seismic reflection data in the channel between Lake Identified subsidence and dipping strata between stations 1213+00 and
Surprise and Jewfish Creek 1229+00 extending to depths of 267 m or more
Drill four additional boreholes from 30 to 60 m with Cores indicated highly variable porosity
continuous core
Run geophysical logs in 4 boreholes Multiple low-density and open zones identified between 6 and 15 m in the
Key Largo Limestone and between 30 and 45 m in the sandy Caloosahatchee
equivalent
Measure seismic P wave velocity Provided seismic velocity to calculate depth of seismic reflection data

Multiple sets of site-specific independent data (Table 26.1) sets of data that provide overwhelming evidence for the pres-
were used to develop the final conceptual model for this site. ence of a large paleocollapse originating at depth. Because
All sets of independent data have identified anomalous con- multiple sets of independent data correlate, (Fig. 26.15) we
ditions between Lake Surprise and Jewfish Creek, all within can have a high degree of confidence in our interpretations
a distance of approximately 480 m. The general location and and conceptual model (Fig. 26.12).
spatial correlation of these anomalous conditions are shown Additional support for deep-seated karst in the area is pro-
in Fig. 26.15. vided by:
The 100 microgal gravity anomaly was the first piece of
data which focused attention away from Lake Surprise and • Lower sea levels by as much as 130 m (Balsillie and
onto the narrow channel between Lake Surprise and Jewfish Donoghue 2004) and a fresh water lens; and
Creek. While rules of thumb and inverse modeling indicated • The presence of the large sinkhole identified by USGS
a source depth of 150 m or more, it was the deep seismic (Shinn et al. 1994b) in the reefs off of Key Largo just
reflection data (Fig. 26.12) that provided indications of sub- 8 km to the southeast.
sidence and dipping strata to depths of 267 m or more that
spatially correlated with the microgravity data. This indi-
cated the presence of an old paleocollapse feature whose ori-
gin lies within the Arcadian Limestone Formation or possibly 26.7 Risk Assessment
in the Suwannee Formation. Four additional borings and
geophysical logs to depths of 30 and 60 m (Fig. 26.14) con- When a karst feature does occur, it often can have significant
firm that the anomalous area is more porous with zones of impact upon structures. The area of concern extends from
open voids. stations 1214+00–1226+00, which includes the location of
Some of the individual sets of data identified in Table 26.1 the piers for the highest portion of the bridge over Jewfish
such as loss of drilling fluid, repeated repair of subsidence in Creek. Prior to this site characterization effort the only evi-
the road, and a major photo-lineament through the site by dence of subsidence at the site was the small area of local-
themselves do not provide conclusive evidence of a major ized settlement seen on the southbound lane of US1 that
karst feature. It is the spatial correlation of these independent required periodic maintenance (Fig. 26.2).
26.7 Risk Assessment 381

FDOT Reference Stations in Feet

Anomalous Area
in Deep Seismic

Gravity Anomaly

Maintenance patch
Aerial Photo on road
Northwest Lineament Southeast

TB-3 & 4
PB-8
PB-7
TB-1
TB-2
PB-12 PB-10

Lake Surprise
0 Layer 1 - Organics, peats, silts and
fill with variable depths

Jewfish
Creek
Layer 2
12 Key Largo
Approximate Depth in Meters

Limestone

24 Layer 3
Hard Pan
Fluid Circulation Loss

Low-density zone (density logs)

36 Dipping strata from

Layer 4 seismic reflection
Sand
0 300

Horizontal
48 Scale in Meters

Fig. 26.15 Spatial correlation of anomalous conditions identified by independent data sets provides an increased confidence level in the interpre-
tation and our conceptual model of geologic conditions (Fig. 26.11)

There are three geologic factors that could result in sub- 1. Some localized zones of increased permeability and open
sidence or collapse at this site, each are highly unlikely, but voids were found within the Key Largo Limestone (Layer
possible, and may impact construction and or long-term 2) at depths of 6–15 m based upon the original 34 borings
operation of the structure. (Fig. 26.6) and the geophysical logs in the additional 4
borings (Fig. 26.14). These features could impact the con-
1. The most probable condition to impact the site is the struction or long-term maintenance.
localized low-density zones or open voids within the Key
Largo Limestone at a depth of 6–15 m (Fig. 26.14) The bridge piers will be founded on cast in-place drilled
2. Localized low-density zones or open voids within the shafts, which will extend to depths of about 18 m in the Key
unconsolidated sands beneath the Key Largo Limestone Largo Limestone within the area of concern. By using drilled
at a depth of 30–45 m (Fig. 26.14) and poured shafts most of the load is carried by skin friction
3. Larger open voids associated with a deep-seated cavern- (compared to driven piles where most of the load is carried
ous system from which the paleocollapse is thought to by the pile tip).
occur at depths of 180–240 m or more (Fig. 26.11)
382 26 Site Characterization Along Bridge Alignment

There is a hard-pan zone (Layer 3) at the base of the Key Several factors exist to suggest that further collapse or
Largo Limestone between 21 and 30 m (Fig. 26.4). It has reactivation of this deep zone is unlikely to virtually zero.
very high SPT values of 50 per about 7 cm. In areas where
bridge piers are to be located within the anomalous zone, a • Dissolution has ceased.
pilot boring should be extended to establish the presence of • The seismic data suggests that extensive subsidence has
this hard-pan layer beneath each bridge pier. Then grouting occurred over the collapse indicating that the large cavern
the pilot borehole to design depth. has likely been filled with sediments
Geologic conditions should be monitored closely during • These deep collapse are quite old and has been stable for
construction by careful inspection of each pier boring within a long time. Data from Shinn et al. (1994b) indicates C14
the anomalous zone for evidence of any voids, monitoring of ages of 5,650 ± 90 years before present in the Florida
fluid loss and rod drop while drilling and monitoring of any Keys and data from Yuhr et al. (2003) indicates ages of
concrete loss when pouring piers. If any unusual conditions 40,000 and 48,800 years before present in the Tarpon
are encountered an alternative would be to extend the bridge Springs Florida area.
piers to rest upon the hard-pan layer at the base of the Key • In addition, all construction and other further activities at
Largo Limestone. the site would not impact the geology at these depths.
Long-term maintenance concerns within this depth are
associated with disposal of surface runoff from the bridge
deck. The proposed approach for disposal of surface runoff
from the bridge deck is by injection wells drilled into the first 26.7.1 Limitations
permeable zone. This practice should be avoided within the
anomalous zone. The discharge of large quantities of surface The limitation of this site characterization is the limited site-
water runoff directly adjacent or near bridge piles, within specific spatial data acquired at this site. The data that indi-
this zone, could possibly flush materials away creating larger cated increased porosity and open voids in Layer 2, the Key
voids. While the probability of inducing subsidence or col- Largo Limestone was laterally limited to the line of borings
lapse due to disposal of runoff within this depth is low, it (34 original and 4 additional). This data density along this
should be avoided. survey line is extensive, but does not provide any data off of
this line. The data that indicated increase porosity and open
2. Localized low-density zones and open voids within the voids in Layer 4 (quartz sand) was very limited. Only four of
sands of Layer 4 at a depth of 30–45 m have been identi- the original borings extended less than 18 m into this layer.
fied and could allow soil raveling or piping. The largest of Two of the four additional borings extended 30 m into this
those identified is about 1.2 m high and all lie below the layer. The limited number of borings and geophysical logs
hard-pan, Layer 3. The risk of these voids impacting the clearly provided an indication of the depth of the zones of
bridge piles is low since they tend to be relatively small, increased porosity and open voids, the actual lateral and ver-
randomly located and occur below the hard-pan, Layer 3. tical extent of these zones has not been established. However,
it is reasonable to assume that these zones extend laterally
However, the lateral extent and depth of these low-density along the line of borings and off of the line of borings some
zones within Layer 4 is unknown. Only 4 of the original 34 distance within the anomalous zone of the paleocollapse.
borings and 2 of the additional borings were drilled to lim-
ited depths into the sands of Layer 4. All borings were
located within the anomalous zone of the paleocollapse, 26.8 Conclusions
which is 360 m wide. The lateral extent of these features is
unknown. The approach to this site characterization includes many of
the key concepts for dealing with geologic uncertainty in the
3. The origin of the large, deep-seated cavernous system in site characterization process. They include:
which the paleocollapse occurred is believed to lie at a
depth of 180–240 m or more below grade. The deep seis- • A variety of site-specific measurements appropriate to
mic data (Fig. 26.12) shows extensive subsidence and site conditions;
dipping strata throughout the zone of paleocollapse sug- • Measurements with a wide range of scale;
gesting that the large cavern has collapsed and is likely • Correlation of a number of independent measurements to
filled in with sediments (Fig. 26.11). Further dissolution improve confidence levels in the interpretation; and
of the deeper cave systems has ceased and the cave sys- • Development of a conceptual model supported by solid
tem is now totally saturated by seawater. data.
References 383

It is interesting to note that prior to beginning the site The results of this study provided sufficient data to allow
characterization there were two independent pieces of data the bridge to be constructed with a minimum risk. This deep-
indicating potential karst conditions between Lake surprise seated paleocollapse feature appears to be stable and well
and Jewfish Creek. These two pieces of data included the below the depth of impact due to construction. While
patch in the southbound lane of US-1 and the location of four increased porosities and small voids could impact the proj-
borings that accounted for 61 % of all fluid losses in the 34 ect, design and construction recommendations are provided
original borings. However, these were subtle indications that to minimize the potential long-term impact of geologic con-
alone would not have indicated a serious problem. It was not ditions and subsidence collapse of the bridge structure.
until the gravity anomaly was detected in Phase I that we Technical reports and papers have been presented about
began to focus attention upon this area. this work include:
All of the data that has been used to solve the geologic
puzzle and to develop an accurate conceptual model of • Technos, Inc. (1994)
geologic conditions at this site is summarized in Table 26.1. • Benson et al. (1995a)
The data includes multiple sets of independent and diverse • Benson et al. (1995b)
data with different scales of measurement and depth of mea-
surement. No single set of measurements would have given a
clear understanding of geologic conditions at this site.
However, these independent sets of site-specific data corre- References
lated spatially (Fig. 26.15) to provide a solid basis for the
final conceptual model (Fig. 26.11). Badgley PC (1960) The analysis of structural patterns in bedrock.
American Institute Mining and Metallurgical Engineers, S. M.
The data subsequently obtained indicated the presence of
E. Trans 225:381–389
a large anomalous area (360 m) along the bridge alignment Badgley PC (1965) Structural and tectonic principles. Harper and Row,
between stations 1214+00 and 1226+00. This anomalous New York
zone is located at the highest portion of the new bridge. The Balsillie JH, Donoghue JF (2004) High resolution sea–level history for
the Gulf of Mexico since the last glacial maximum, Florida Geologic
mass deficit indicated by the microgravity data and overlay-
Survey Report of Investigation 103
ing subsidence and dipping strata seen in the deep seismic Benson RC et al (1995a) Assessment of potential karst conditions for a
data and provided the evidence for a large cavernous zone new bridge in the Florida Keys. Symposium on the application of
originating deep within the Arcadian Formation at depths of geophysics to engineering & environmental problems, Orlando,
23–26 Apr 1995, pp 22–26
180–290 m or possibly deeper within the Suwannee
Benson RC et al (1995b) Subsurface investigation of possible karst con-
Formation. ditions at the Jewfish Creek bridge replacement, Key Largo, Florida.
The possible presence of the deep cavern system was sup- In: Beck BF (ed) Proceedings of the 5th multidisciplinary confer-
ported by conditions for dissolution created by a lower sea ence on sinkholes and the engineering and environmental impacts of
karst: karst geohazards, Gatlinburg, Tennessee. Balkema,
level of up to 130 m (Balsillie and Donoghue 2004) and a
Rotterdam, pp 409–414
thick fresh water lens (Ford and Williams 2007) resulting in Billings MP (1954) Structural geology. Prentice-Hall, Englewood
development of caves at significant depths below present sea Cliffs
level. Dean’s Bluehole on Long Island in the Bahamas is at Butler DK (1980) Microgravimetric techniques for geotechnical appli-
cations. U.S. Army Engineer Waterway Experiment Station,
least 200 m deep and is an example of karst formed at these
Miscellaneous Paper GL-80-13, Vicksburg, Mississippi
depths. Further support for the presence of deep-seated Ferguson GE et al (1947) Springs of Florida. Florida Geological Survey,
paleokarst is provided by the presence of a large paleocol- Geological Bulletin No. 31
lapse sinkhole 8 km to the southeast within the Florida reefs Finkl C (1994) Report on FDOT Jewfish Creek Bridge. Consultants
report, The Coastal Education & Research Foundation, CERF
(Shinn et al. 1994b).
Remote Sensing Lab, Ft. Lauderdale, Florida
Based upon the data acquired and conceptual model Finkl CW, Warner MT (2005) Morphologic features and morphody-
developed, the two critical questions that were asked at the namic zones along the inner continental shelf of southeastern
beginning of this study can be addressed: Florida: an example of form and process controlled by lithology. J
Coast Res 42:79–96, Special Issue
Ford D, Williams P (2007) Karst hydrogeology and geomorphology.
1. Can an investigation fully identify unstable geology, such Wiley, New York
that bridge decks can be placed or lengthened to safely Lane E (1981) Environmental geology series, Miami sheet, Map Series
span the critical geologic conditions? 101. Florida Geological Survey, Tallahassee
Law Engineering and Environmental Services (1994) Report of geo-
2. Can the bridge as designed, be stable for at least 100 years
technical exploration, Jewfish Creek Bridge Replacement, Volumes
in the given geologic environment? I and II
384 26 Site Characterization Along Bridge Alignment

Law Engineering and Environmental Services (1995) Presentation of Shinn EA et al (1994b) A giant sediment trap in the Florida keys. US
test boring records, Jewfish Creek – Supplemental Drilling, 8 Aug Geological Survey Center for Coastal Geology, St. Petersburg
McNeill D et al (1995) Core workshop: late tertiary-quaternary cores Technos, Inc (1994) Subsurface investigation of possible karst condi-
from the Florida keys and everglades. Sponsored by the Florida geo- tions at Jewfish Creek bridge replacement Monroe County, Florida.
logical survey. University of Miami, RSMAS/MGG, Miami Consultant Report for Stienman Boynton Gronquist & Birdsall and
Parker GG et al (1955) Water resources of southeastern Florida. USGS Florida Department of Transportation, 16 Nov
Geological Survey Water-Supply Paper 1255 Wilson WL (1994) Morphology and hydrology of the deepest know cave in
Scott TM (1988) The lithostratigraphy of the Hawthorn Group the Bahamas: Dean’s blue hole, long island. Bahamas J Sci 2(1):10–14
(Miocene) of Florida. Florida Geological Survey, Bulletin No. 59 Yuhr L et al (2003) A case history of a large karst investigation.
Shinn EA (1994a) Letter to Mr. David C. Miro, Florida Department of Presented at the 3rd international conference on applied geophysics,
Transportation Dec 21 Geophysics 2003 Orlando 8–12 Dec 2003
 EPA Superfund Site
27

Abstract
This third case history was a 52 ha EPA Superfund site located in Tarpon Springs north of
Tampa, Florida that had processed phosphate ore to produce elemental phosphorus. While
the site had been investigated and a Record Of Decision (ROD) had been completed from
EPA regarding the remediation, there remained concerns with the site. These concerns
included the development of sinkholes on the site, as well as the presence of buried drums
of elemental phosphorus waste material, and the potential for groundwater contamination.
The project was carried out between 2001 and 2003 over a period of 30 months following
our standard site characterization strategy. Because this was a superfund site, there were a
number of stakeholders involved from beginning to end. The project included extensive
meetings with local interests, county, state and federal agencies with EPA oversight. The
client wanted this project completed right the first time with no questions left unanswered.
Therefore, a large number of methods were utilized with heavy emphasis on surface, bore-
hole and marine geophysical methods. Some of these methods were used to provide 100 %
coverage of the property. The work at this site is probably one of the most complete karst
site characterizations carried out.

27.1 Background area, and a system of settling ponds. Figure 14.6 is a historical
aerial photo that shows the plant and general operations at
The site is located approximately 24 km northwest of Tampa, the site including storage piles of raw phosphate ore and
Florida, in the northwest corner of Pinellas County. The site railway operations. The view is toward the Anclote River to
is adjacent to the Anclote River, which flows into the Gulf of the southwest.
Mexico approximately 3 km west of the site. Figure 27.1 The site is divided into two areas, a North and South Parcel
shows the regional setting for the Stauffer site from the Gulf (Fig. 27.2). The production facility, a series of settling and
of Mexico to the west, to alternate highway US-19 and disposal ponds were located on the South Parcel. The North
US-19 to the east and the Anclote River to the south. Parcel was used for slag processing, storage and a waste
Production began in 1947 at the Stauffer Management pond. Two byproducts came from the furnace: slag and fer-
Company site. The principal raw materials were phosphate rophosphorus. The calcium silicate slag was drained periodi-
ore (primarily tricalcium phosphate), along with coke and cally from the furnace into the adjacent slag pit. The slag was
silica rock that are extensively mined in the west-central then cooled and solidified prior to being trucked to the slag
Florida area. Essentially all raw materials, products and processing area in the North Parcel for crushing to produce
byproducts were received and shipped by rail. A high tem- aggregate that was sold and used off-site. The ferrophospho-
perature electric arc furnace melted the raw materials, and rus was also removed from the furnace as a liquid, allowed to
elemental phosphorus evolved as a gas, which was con- cooled and solidified for sale to the steel industry. Some of the
densed under water and further processed to remove impuri- waste phosphorus was placed in drums and covered with
ties. The site included production facilities, slag processing water then disposed of in on-site ponds on the South Parcel.

© Springer Science+Business Media Dordrecht 2016 385

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9_27
386 27 EPA Superfund Site

Fig. 27.1 Site map (USGS 7.5 min Series Tarpon Springs Quadrangle, 1995 Elfers Quadrangle, 1998)

After the plant was shut down in 1981, the plant produc- 27.1.1 Record of Decision (ROD)
tion facilities and most of the buildings were decontaminated,
demolished and removed from the site. The facility perma- A Record of Decision (ROD) was issued by US-EPA (1998)
nently closed in 1983. The only buildings remaining on-site that outlined the remediation for the property. The remedy
at the time of this investigation were the main office build- selected by the EPA includes the following components:
ing, an equipment building, a guard building, and a lunch-
room building located on the northeast corner of the South • Limited excavation of contaminated material and soil that
Parcel (Fig. 27.2). exceeds residential cleanup standards.
The southeastern most portion of the South Parcel was • Consolidation of contaminated material and soil in the main
purchased in 1981 as a buffer along the southeastern edge of pond area, slag processing area, and other areas on-site.
the ponds. The Corps of Engineers had previous used the • In-situ solidification and stabilization of pond material
property for placement of dredge spoils from the Anclote and contaminated soil below the water table in the
River in 1972 (Fig. 27.2). No production-related operations consolidation areas.
took place in this area. • Clay caps will be placed over the remaining consolidation
The site topography is generally flat, with a slight slope areas.
southward towards the Anclote River. The average land sur-
face elevation is about 3 m above sea level, but can be locally In addition, institutional controls on the site including
higher in the pond areas reaching as much as 10 m. These are deed restrictions such as land use ordinances, physical barriers,
areas where ponds have been filled with disposal materials. and restrictions for water supply wells were put in place.
27.2 Objectives of the Overall Investigation 387

Fig. 27.2 Aerial photo of the site showing the North and South Parcels

• A solidification/stabilization treatability study on pond

27.2 Objectives of the Overall materials and soils was carried out by O’Brien & Gere
Investigation Engineers, Inc. (OBG), the lead engineering team and
responsible for the remediation of the site.
On August 25, 2000 Stauffer Management Company and the
EPA entered into an Agreement to conduct certain additional For this case history, the focus is upon the assessment of
studies to evaluate whether the remedy selected for the geologic and karst conditions, along with the site hydrogeol-
source control will provide protection of human health and ogy, as they impact the site and its proposed remediation.
the environment throughout the life of the remedy (US EPA
2000a, b). Three major additional studies were undertaken to
verify previous findings and resolve all remaining issues. 27.2.1 Objectives of the Geologic and Karst
These studies included: Investigation

• A geologic and karst characterization as well as a sink- The objectives of the geologic and karst studies carried out
hole risk assessment that was carried out by the authors by Technos, Inc. are summarized as:
and their firm, Technos Inc.
• A site-wide groundwater characterization and contami- • Evaluate whether or not sinkholes are present within the
nant study that was carried out by Parsons Engineering site and the probability that sinkholes or karst features
Science, Inc. (Parsons). will form over the required life of the remedy.
388 27 EPA Superfund Site

• Evaluate the nature and extent of the intermittent semi- that we don’t have to come back and do more work or do
confining clay layer (SCL) existing between the surficial things over”.
aquifer and Floridan aquifers to the extent that it impacts
the effectiveness of the selected remedy.
• Delineate the horizontal extent of the ponds and estimate 27.2.4 Review and Oversight Committee
the depths of the pond materials.
• Evaluate the presence of buried drums, storage tanks, or A technical oversight committee was formed to review and
other potential sources of contamination at the site that provide input on the work plan, as well as the data and its
have not already been identified. interpretation at key milestone meetings throughout the proj-
• Evaluate the ability of the underlying geology to support ect. This committee consisted of representatives from the
the proposed remedy over the life of the remedy as US-EPA, the Florida Department of Environmental
required under 40 CFR Part 192 under reasonably antici- Protection (FDEP), United States Geological Survey
pated site conditions. (USGS) and the Corps of Engineers as well as the Pinellas
County Department of Environmental Management (DEM),
Pinellas County Health Department, Dr. Mark Stewart of
the University of South Florida as well as private sector
27.2.2 Objectives of the Hydrologic interest groups.
Investigation The purpose of this committee was to establish consensus
on the technical issues as the project moved forward and
The objectives of Parson’s hydrologic investigation were to address any concerns or questions as they were encountered.
evaluate the surficial aquifer (to approximately 6 m deep), They provided comments and approval of the project strat-
the semi-confining layer at the base of the surficial aquifer, egy and work plan. As the project proceeded, the oversight
and the Upper Floridan aquifer (approximately 6–15 m committee was presented each data set and its interpretation.
deep). To accomplish this, the following work elements were Any concerns or questions were addressed and interpreta-
completed: tions agreed upon. This allowed the project to move forward
in a smooth and methodical manner with a consensus on all
• Assess the flow direction and estimate the lateral hydrau- technical matters.
lic gradients within the surficial and Upper Floridan aqui-
fers and estimate the vertical gradient between these two
aquifers. 27.3 Technical Approach
• Evaluate the hydrologic nature of the semi-confining
layer (SCL) between the surficial and Upper Floridan The strategy used for the project is that presented in Chap.
aquifers. 12 which is based upon the ASTM D6235 Standard Practice
• Establish the locations and concentrations of groundwater for Expedited Site Characterization (ASTM 1998). Two key
constituents in both aquifers as well as the general geo- components of the strategy were utilizing a core team and
chemistry of the groundwater. diverse methodologies.
• Characterize the effects of tides on the aquifers and evalu-
ate whether the tides in the Anclote River will affect the
proposed remedy. 27.3.1 Core Team

One of the key aspects leading to the success of this project

was the core group of companies and individuals that were
27.2.3 The Owner’s Goals used throughout the project from the beginning to end.
This included key team members as well as subcontractors
Obtaining appropriate and adequate site-specific data is (surveyors, drillers, and laboratories). This consistency
often a significant problem with budget and or time con- allowed the team to develop an understanding of the site and
straints along with many other factors will limit this effort. allowed the project to move forward in a coherent manner.
Numerous studies had already been completed at the site The geologic and karst studies as well as the hydrologic
over a period of years, however, none had been comprehen- studies were conducted concurrently in conformance with the
sive or had answered all of the questions about the sites Work Plans for the Additional Studies Program. In addition,
geology, karst conditions, groundwater and contaminant field efforts and data were coordinated and shared, so that data
issues. As a result the owners instructions at this site were was not duplicated and the geologic and hydrologic puzzle
clear, “Do whatever it takes to do it right the first time so could be analyzed and integrated as efficiently as possible.
27.4 Site Preparation 389

A quality control and quality assurance (QC/QA) program The surface geophysical techniques utilized for assess-
was an integral part of the site characterization effort because ment of the shallow zone included ground penetrating radar
of the critical nature of the project from both a political and (GPR), frequency domain electromagnetics (EM), magnet-
technical perspective. The fieldwork was initiated February ics, and multi-frequency electromagnetics (MFEM). The
2002 and final reports were submitted June 2003. assessment of the deeper zone utilized 2D resistivity imaging
and microgravity measurements on-site and marine geo-
physical techniques in the adjacent river. In addition, the
27.3.2 Methodologies investigation included:

The issue of resolving karst conditions is one of adequate spa- • invasive measurements (direct push electrical conductiv-
tial sampling, since it is quite easy to miss the spatially variable ity (DPEC), rotosonic drilling, hollow-stem augers and
karst conditions. The measurements used and data acquired wash borings),
must not only be appropriate for the site-specific geologic and • geophysical logging in all borings and wells (existing and
cultural conditions but must be spatially adequate, both laterally new);
and vertically, to detect any geologic anomalies of interest. • marine geophysical measurements to take advantage of
Therefore, to resolve the karst issues with a high level of reso- data from within the adjacent Anclote river,
lution and confidence a wide range of surface and borehole • in-situ monitoring, testing and laboratory analysis for a
geophysical surveys were utilized. These sets of data provided range of hydrologic and geologic parameters, and
adequate and sometimes total site coverage and allowed • additional geophysical measurements to obtain compres-
background and anomalous conditions to be readily identified. sional (P) and shear (S) wave velocities for calculating
Detailed invasive data could then be located and acquired elastic modulus.
using traditional drilling, sampling and laboratory analysis.
The methods selected provided a range of measurements Figure 27.4 illustrates the overall strategy used on this
with different independently measured parameters, along project and follows the strategy presented in Chap. 12. The
with a wide range of scale (regional, local and site-specific project started with getting the site ready for fieldwork and
details) and depths of measurements. The combination of getting all the background data in order. This included a
these diverse measurements and their integration provided a desk study, aerial photo analysis and site walkover. A pre-
level of redundancy in the data obtained. If different measure- liminary conceptual model was developed based upon this
ment techniques result in a similar interpretation, their corre- information. When fieldwork began, each data set was inde-
lation, or redundancy, provides an increase in the level of pendently analyzed and interpreted before being integrated
confidence in the data and its interpretation. It is recognized with other data. Once integrated, it was checked against the
that the range of measurements and their density is unusual conceptual model and the model was upgraded or modified
and unique to this project. This was driven by the client’s as necessary. This iterative process allowed the project to
objective to “do whatever it takes to do it right the first time”. address issues as they were identified and kept the project
Table 27.1 summarizes the methods used at this site for moving forward.
the purposes of assessing geologic, hydrologic, and karst
conditions; including the depth of measurement, and the site
coverage provided by each measurement. This large set of 27.4 Site Preparation
on-site measurements provided a solid foundation of data
from which to build the site conceptual model and reach con- Maximum site availability was required for this investigation
clusions that minimize opinions and assumptions about on- to allow a high density of measurements to be made. This
site conditions (Yuhr et al. 1996). required debris piles to be removed or consolidated and over-
The geologic and karst issues were primarily divided into grown brush and small trees to be cleared. An extensive
two zones, a shallow and deeper zone. Figure 27.3 summa- clearing operation was completed to enable easy access and
rizes these measurements and their depth: allow continuous survey lines of geophysical data to be
acquired by a small four wheel drive vehicle. This site clear-
• The shallow zone including the sands and the semi con- ance was a major effort on-site and was completed by a
fining layer SCL which range from about 3 m on the north subcontractor prior to any field measurements.
side of the property to about 6 m deep near the Anclote All new data acquired by the different team members
River and the surficial aquifer that lies within this zone; and required coordination and integration. Therefore, a standard
• The deeper zone below the SCL that included the Tampa reference grid was established for everyone to use. The use
and Suwannee Limestone that contains the upper Floridan of a fixed survey grid facilitated the spatial integration of all
aquifer. old and new data. The grid also enabled anyone to quickly
390 27 EPA Superfund Site

Table 27.1 Summary of measurement techniques used

Approximate depth of
Method Parameter measured measurement Data quantity
Surface and borehole geophysical measurements
Electromagnetics (EM31) Conductivity and in-phase 0–6 m Continuous 3 m line spacing, 177 km
(metals) of data
Ground penetrating radar (GPR) with Complex dielectric constant 0–7.3 m Continuous 3 m line spacing, 145 km
100 MHz antenna of data
Multi-frequency electromagnetics Conductivity and in-phase Up to 6 m Continuous 1.5 m line spacing over
(MFEM) metals at multiple frequencies selected areas, 80 km of data
Magnetic Magnetic susceptibility Up to 6 m for a single Continuous 3 m line spacing over
steel drum selected areas, 18 km of data
2D resistivity dipole-dipole array Electrical resistivity 0–24 m 6 m electrode spacing run along lines
spaced 30 m apart, 16 km of data
Microgravity Density >30 m 30 m line spacing with 12 m station
spacing. 6 m station spacing south
portion of South Parcel, a total of
1,826 stations
Geophysical logging Natural gamma, density, Limited to depth of well 74 borings and wells
porosity, and electrical or boring max 48 m
conductivity
Invasive measurements
Direct push electrical conductivity Electrical conductivity 12 m 110 DPEC pushes
(DPEC)
Direct push soil cores Sediments (visual) 12 m 18 soil cores
Hollow stem auger SPT values and soil and rock 3.6–15 m 23 shallow and 20 deep monitoring
samples wells
Rotosonic test borings Soil and rock samples 23–48 m 15 borings
Wash boring Refusal of penetration Up to 4.5 m 14 in sinkhole pond
Hydrologic measurements
Slug test Hydraulic conductivity Over well screen Interval 9 tests
Pump test Draw down hydraulic Depth of pump and well 1 in surficial aquifer and 1 in
conductivity screen Floridan aquifer
Water levels Water level elevations 3.6–15 m 39 shallow and 20 deep wells
monitored for 17 months
Tidal effects Water level changes versus time 0–15 m Every 15 min for 5 days in 21 wells
plus tide gauge
Laboratory testing
Water quality and testing 23 metals 3.6–15 m 28 shallow wells and 20 deep wells
9 inorganics
5 radiological
Permeability testing Hydraulic conductivity Depth of sample 15 samples from 7 locations
Particle size analysis – sieve analysis Grain size Depth of sample 18 samples from 9 test borings
Particle size analysis – hydrometer Grain size Depth of sample 15 samples from 7 locations
analysis
C-14 dating Age 15.8 and 24.6 m 3 samples from 2 boring locations
X-ray diffraction measurements Mineral constituents Within the SCL 6 samples from 6 locations within
the SCL
Marine geophysical measurements
Bathymetry Bottom profile River bottom 1.6 km in Meyers cove
Sidescan sonar Bottom image River bottom 16 km of data in the Anclote River
Subbottom profiling (seismic Seismic coefficient of reflection 45 m below river bottom 19 km in the Anclote River
reflection)
Geotechnical field measurements
Seismic refraction Velocity of P-waves 12 m 1.5 m geophone spacing, 2 test lines
Hole to hole seismic measurements Velocity of P and S-waves Up to 23 m Between 3 boreholes 4.5 m apart, 2
(crosshole) sets of measurements
Vertical seismic profiling (VSP) Velocity of P and S-waves 23–48 m 9 test borings
27.4 Site Preparation 391

gs
rin
l

Bo
on m A oil c rica
co opro r Side tion)

tos oring ger res)

t
d M ste nd s elect

We Tes

gs
o
Ge Son ry & eflec
ing

rin
s
m

ole
l
Bo

ole
gg

cr

Ho ctivity PEC
(se otto

u
Lo

sh
mi

nh
st
bb

a
ty

Te
ut be D
s

ros
it y

ow
al

i
tivi

filin Su

a
Geologic

cs
sic

an et
v

ic C
ic D
gra

t
sis

i
Hydrologic

Sc ym

on
i
eti

Pro rine
g
hy

an llow
EM
Section

Re
31

ism
cro
R

gn

ism
op
Section

nd
Ma
GP

Ba
EM

MF
(near the river)

Ma
2D

Ro
Ge
Mi

Se
Se

Geohydrology
0

Limited to River Bottom

Surficial

Shallow
Sands Aquifer
6
Semi-Confining

Limited to borehole depth

23 to 48 meters

23 to 48 meters
Layer

12 Tampa
Limestone

Geohydrology
Floridan

Deeper
18
Aquifer
Depth in Meters

24

30
Suwannee
Limestone
36

42

48

Fig. 27.3 Type and depth of measurements used to meet project objectives

Site Prep
- Brush Clearing
- Survey Grid Preliminary
Conceptual
Desk Study Models
Iteratively Upgraded as Information

Aerial Photo Analysis

Site Walkover
and Data Become Available
Geologic/Hydrologic/Karst
Final Conceptual Models

Interpretation
of each
data set
Acquisition of New
Geologic and
Hydrologic Data
using a Identify
wide Anomalies
range of
measurements
(see Table 27.1) Detailed Integrate
Characterization Results
of Anomalies of data sets

Sinkhole Risk Analysis

Fig. 27.4 Overall project strategy

392 27 EPA Superfund Site

walk to and reoccupy any location within the entire 52 ha at the site (Beck and Sayed 1991). These three borings indi-
without use of a GPS system. cate some variation in depth and thickness of the Tampa and
All existing site features including buildings, borings, Suwannee limestone, but generally support the regional
wells, disposal ponds and piles of waste materials were ref- trends.
erenced to this standard grid. Current aerial photography of The geologic section (Fig. 13.3) defines four lithostrati-
the site was also obtained after brush clearing and was used graphic units. From shallow to deeper, these units are:
to develop a topographic map with a 0.3 m contour interval.
1. Post-Hawthorn Group Undifferentiated Sands and Clays
2. Hawthorn Group consisting of the Tampa Member of the
27.5 The Desk Study Arcadia Formation, herein referred to as the Tampa
Limestone
While the site was being prepared for fieldwork (brush clear- 3. Suwannee Limestone
ance, surveying and a topographic map) a desk study was 4. Ocala Limestone.
carried out. An abundance of geologic and hydrologic infor-
mation was available from the Florida Geologic Survey bul- Only the Tampa and Suwannee Limestone are of interest
letins and reports. Karst data included cave maps, spring here. The deeper Ocala Limestone is not expected to have
surveys, county sinkhole databases, state sinkhole databases karst features that would impact the remediation at the site.
and general karst reports provided by a variety of sources. The following descriptions are primarily from Green et al.
These reports provided information on the stratigraphy, (1995).
hydrologic setting, sinkhole distribution and their frequency
of occurrence, lineaments, depth of known cave systems in 27.5.2.1 Post-Hawthorn Group Undifferentiated
the area, and distribution of springs. The information pro- Sands and Clays
vided in these reports formed the basis for the initial under- The Post-Hawthorn Group sediments are primarily com-
standing of the regional site conditions and included general posed of varying proportions of quartz sand, shell and clay
geologic and hydrologic information. with variable amounts of organics and naturally occurring
re-worked phosphate. Much of the county is covered by
unconsolidated sediment referred to as sand. Elevations at
27.5.1 The General Site Setting the site are approximately 3 m MSL with a thin 3–6 m cover
of sand (Beck and Sayed 1991).
The Stauffer site is located near the town of Tarpon Springs, in
Pinellas County, on the west central coast of the Florida penin- 27.5.2.2 The Hawthorn Group
sula. The site itself lies on the north side of the Anclote River The Hawthorn Group Consisting of the Tampa Member of
(Fig. 27.1) and is underlain by hundreds of meters of lime- the Arcadia Formation and is referred to as the Tampa
stones and dolomites, which compose the Floridan Aquifer. For Limestone. In Pinellas County, the Hawthorn Group sedi-
purposes of evaluating sinkhole hazards and the karst topogra- ments lie unconformably above the Suwannee Limestone.
phy of Pinellas County, our focus was the geology of the upper- Only the Tampa Member (Tampa Limestone) of the Arcadia
most limestones (about 60 m) and the overlaying clastic strata. Formation is present in the study area.
The Tampa Limestone is white to yellowish gray in color,
and ranges from a limestone, quartz sand and clay (Green
27.5.2 Regional Geology et al. 1995; Scott 1988). The Tampa Limestone commonly
contains 10–30 % quartz sand. Based upon projections of
A regional geologic section for the Tarpon Springs area was the regional cross-section to the site (Fig. 13.3), the top of
constructed using five existing borings. Two deep borings, the Tampa Limestone at the Stauffer site was expected to be
W-16609 located approximately 4.8 km north-northeast of about 5–15 m below land surface (bls) and approximately
the site, and W-12943 located approximately 5.6 km south- 10–15 m thick.
southwest of the site. These two borings are referenced in
cross-sections included in publications by the Florida 27.5.2.3 Suwannee Limestone
Geologic Survey (Arthur et al. 2001; Green et al. 1995) and The Hawthorn Group sediments, which include the Tampa
show a slight southward dip to the formations. Three addi- Limestone, overlie the Suwannee Limestone unconformably.
tional borings in Pinellas County east of the site (TR15-2, In some cases, the upper Suwannee lithologies appear to
2560 and 1072) are added to the regional cross-section and grade upward into the Hawthorn Group (Green et al. 1995;
provide more localized depth and thickness of the Tampa Beck and Sayed 1991). The lithology of the Suwannee
Limestone and the depth of the top of Suwannee Limestone Limestone ranges from a light-gray to yellowish-gray packestone
27.5 The Desk Study 393

to grainstone. These carbonates are variably moldic with Suwannee Limestones (Hickey 1982). The high productivity
trace amounts of sand and clay within the upper portions. is due to the high primary porosity along with the karst char-
Trace amounts of chert and organics occur throughout the acter of these limestones. The Floridan aquifer provides
unit. A dolostone or dolomitic limestone layer, approxi- some of the public water supply for Pinellas County, although
mately 3–6 m thick, commonly occurs within the lower one- it has limited use within the western portion of the county
third of the unit in the study area (Green et al. 1995). Based due to saltwater intrusion along the coast.
upon projections of the regional cross-section to the site (Fig. Saltwater intrusion is known to occur along the Anclote
13.3) the top of the Suwannee Limestone at the Stauffer site River. In 1963, chloride content in and around the Anclote
is expected to be about 15–24 m bls (Beck and Sayed 1991) River near the site was shown as more than 250 mg/L (maxi-
with a thickness of about 60 m. mum for drinking water) at depths of 0–38 m (Cherry 1966).
The chloride concentration at selected locations within the
Anclote River near the site in 1969 ranged from 3,100 mg/L
27.5.3 Regional Hydrology at low tide to 15,200 mg/L at high tide (Reichenbaugh 1972).

The hydrogeologic section in Pinellas County consists of an

unconfined surficial aquifer in the surficial sands and the 27.5.4 Regional Geomorphology
karstified limestones of the Floridan Aquifer, consisting of
the Tampa and Suwannee Limestones (Fig. 13.3). The Tampa The karst character of the limestone aquifer in the county is
Limestone is hydrogeologically connected with the underly- well-demonstrated by the existence of paleokarst sinkhole
ing Suwannee Formation and is included in the Upper lakes on land, numerous springs and numerous bayous
Floridan aquifer (Beck and Sayed 1991; Wetterhall 1964). within the Anclote River. A bayou is a circular body of water
The surficial aquifer is thin, heterogeneous, has low val- within the Anclote River thought to be associated with a
ues of transmissivity and low yields to pumping wells. The paleocollapse. The karst character of the limestone aquifer in
surficial aquifer in Pinellas County has only limited use as a the immediate area is well documented by the cavernous
supplemental or alternative source of water (Southwest connection between Lake Tarpon and Tarpon Spring
Florida Water Management District 1988). The thin nature (Fig. 27.1). Lake Tarpon drained through a sinkhole, via a
of the surficial aquifer in the vicinity of the site limits its deep cavern and discharged at Tarpon Spring located in
usefulness as a drinking water supply; however, the aquifer Spring Bayou (Wetterhall 1965).
is used for small irrigation purposes. The surficial aquifer at From a review of regional geomorphology it was found
the site flows toward the south and discharges to the Anclote that sea levels had been lower by as much as 130 m (Balsillie
River. The surficial aquifer is separated from the Floridan and Donoghue 2004). The combination of much lower sea
aquifer by a semi-confining layer (SCL) of clay, sandy clay, levels and a thick fresh water lens would have resulted in the
marl, and limestone from the base of the Post-Hawthorn dissolution of limestone and the development of caves at
Group and upper weathered surface of the Tampa Limestone. significant depths of 240 m or more. This model can account
Regionally, the water table in the surficial aquifer is gen- for the many large and therefore deeper-seated paleokarst
erally 1.5–3 m below the land surface and the potentiometric seen in the area.
surface of the confined Upper Floridan aquifer is also rela-
tively flat and is also generally 1.5–3 m above sea level. In 27.5.4.1 Aerial Photo Analysis
the Tarpon Springs area, where surface elevations are low, Three types of aerial photos were obtained for this project
the head difference between the surficial and Floridan aqui- and they include stereo-pairs, individual aerial photographs,
fers is very small (Beck and Sayed 1991; Wetterhall 1964). and various historic oblique aerial photographs of the site.
Further inland to the east, beneath higher ridges, the Stereo-pair photos were acquired and reviewed for the fol-
piezometric surface of the Upper Floridan aquifer is much lowing years 1926, 1942, 1951, 1962, 1967, 1997 and 2001.
higher. There is a significant head difference between these These photos typically covered an area of 1.6 km radius or
two aquifers in these areas. The regional flow direction more around the Stauffer site.
within the Upper Floridan aquifer is shown to be westward to A review of these aerial photos was made to identify obvi-
the Gulf of Mexico. However, within the areas north and ous sinkholes in and around the site area. Each sinkhole
south of the Anclote River, flow is toward the river indicating identified in the latest 2001 aerial photography was also seen
that the river serves as a discharge point for the Upper in the aerial photography from 1926, prior to the construc-
Floridan aquifer (Coble 1973). tion of the plant. A number of obvious paleosinkholes due to
The most productive aquifer in Pinellas County is the their circular shape are seen in the aerial photos including
Upper Floridan aquifer and the most productive zone of this some bayous within the adjacent Anclote River and its
aquifer is in the uppermost limestones, the Tampa and branches (Fig. 27.5). Two of these were located along the
394 27 EPA Superfund Site

Fig. 27.5 Aerial photo of the site setting including numerous sinkholes and the Anclote River with its bayous (FDOT PD 4805-4-019;
2-18-2001)
27.5 The Desk Study 395

western perimeter of the South Parcel of the site (Fig. 27.2). cavity system that collapsed to form this paleosinkhole is
They are the Sinkhole Pond and Meyers Cove. All of these at a depth of 40–43 m bls within the lower third of the
features appear in the oldest aerial photography from 1926 Suwannee Limestone at that location.
indicating that they are old and not due to recent sinkhole
activity. These examples indicate that large cave conduits would
Photo lineaments are seen on aerial photographs as align- be expected to occur within the lower half to lower third of
ments of structure, joints, sinkholes, sinkhole lakes, stream- the Suwannee Limestone at depths of approximately
beds, vegetation trends, and other surface features thought to 30–60 m below grade and even deeper. Wetterhall (1965)
be controlled by subsurface conditions. Lineaments that are indicates that within Pasco County (the county immediately
manifestations of subsurface geologic conditions in lime- north of Pinellas) most domestic and many irrigation wells
stone are often the result of dissolution of rock along joints, produce from the lower part of the Suwannee Limestone.
with major karst depressions centered where two sets of This data can be used as an indication of where significant
joints intersect (Parizek 1976). dissolution could be expected at the Stauffer site.
Regional lineaments in the State of Florida generally
trend northeast-southwest and northwest-southeast (Vernon 27.5.4.3 Springs
1951). A number of local lineaments were identified from Springs in the area are discussed in Wetterhall (1965) and
the aerial photos and USGS topographic maps. The Anclote Scott et al. (2004). There are numerous springs located in
River and its bayous are to a large extent fracture controlled Pinellas county and Pasco county to the north. The largest
and contain a number of circular bayous, which are pale- spring in the area is a tourist attraction called Weeki Wachee
osinkholes. Figure 27.6 shows three significant lineaments in Springs about 45 km to the north. This is a first magnitude
proximity to the Stauffer Site along with other features of spring (Scott et al. 2004). The spring flows from near the
interest. contact between the Suwannee and the Ocala Limestone at a
depth of 42 m deep. Its source of water is probably below the
• The most obvious local lineament is the stretch of the Ocala in the Avon Park Limestone.
Anclote River south of the Stauffer Facility that forms a Tarpon Spring is located about 2.7 km southeast of the
long lineament (oriented approximately 45°W). This is site and is located within Spring Bayou at a depth of about
coincident with the connection between Tarpon Spring in 38 m deep (Fig. 27.1). Lake Tarpon Sink is located on the
Spring Bayou and Lake Tarpon sink in Lake Tarpon about western shore of Lake Tarpon further to the southeast
5 km to the southeast (Fig. 27.1) and described by (Fig. 27.1). The sink has a depth of 35 m below MSL and is
Wetterhall (1965). the eastern end of a deep conduit system providing a connec-
• Another local lineament is seen extending 10°W through tion between Lake Tarpon Sink and Tarpon Springs. The
the eastern part of the Stauffer site and Kreamer Bayou. connecting conduit lies within the Suwannee Limestone
This lineament becomes very subtle north of site, where it (Wetterhall 1965).
is only based upon a linear trend in vegetation. Based upon local knowledge, (discussions with fishermen
• A possible lineament along the west side of the site and divers) there are several springs within the bayous of the
extends through Meyers Cove and the sinkhole pond Anclote River as well as in the area near shore in the Gulf of
extending 10°W. Mexico. The assessment of possible local springs was made
visually at low tide. The nearest spring to the site is located
in the Port Tarpon Marina about 0.5 km southeast of the site
27.5.4.2 Cave Systems Within the Region (Fig. 27.6) and is in the Tampa Limestone.
The caves in the area are all under water and are based upon These local springs, such as the one identified in the Port
maps developed by cave divers. In addition to the cave con- Tarpon marina are of smaller magnitude (probably less than
necting Lake Tarpon and Tarpon Spring. A number of caves a few hundred GPM) and flow only for a short duration at
frequented by cave divers exist to the north of the site and low tide. This indicates a very small head difference between
include: the river and the groundwater source for the spring. A small
head difference would account for the low flow noticed at
• Beaconwoods Cave System is located about 19 km north many local springs only at low tides.
of the site underlying a large area of Bayonet Point,
Florida. This is an extensive cave system about 3.2 km in
length at a depth of approximate 38–45 m bls (about mid- 27.5.5 Reported Sinkholes and Sinkhole
way within the Suwannee Limestone at that location) Trends in the Region
(Wilson, 2001, personal communication).
• A large open paleosinkhole along US19 in Hudson, A wide range of databases, reports, and newspaper clippings
Florida is located about 22.5 km north of the site. The were available to provide sinkhole data. While none are
396 27 EPA Superfund Site

Fig. 27.6 The Stauffer site with locations of three photo-lineaments, karst features, two dredge disposal sites and two springs shown
(USGS 7.5 min Series Tarpon Springs Quadrangle, 1995)
27.5 The Desk Study 397

complete and there is redundancy, they are representative of 27.5.5.3 The Sinkhole Database
reported sinkhole activity. The following summarizes the Bill Wisner and Steve Denahan began a sinkhole database in
data obtained from these sources. Unfortunately, collapse the 1960s, who were with the Florida Department of
caused by broken water or sewer lines are also often reported Transportation (FDOT) at the time. Bill Wilson then
as naturally occurring sinkholes. expanded this database, when he was with the Florida
Sinkhole Research Institute, and was later continued by Bill
27.5.5.1 Sinclair and Stewart’s Sinkhole Map at his own firm Subsurface Exploration Inc. (SEI). This data-
Sinclair and Stewart (1985) used the concept that the thick- base has been used as a means of estimating sinkhole loca-
ness of cover exerts a significant control on sinkhole devel- tions, densities, and their temporal trends, new sinkholes per
opment and have used this concept to develop their sinkhole km2 per year (NSH) and to provide sinkhole risk assessments
probability map showing the expected types of sinkholes (Wilson 1995).
based upon cover thickness (Fig. 3.10). The Stauffer site lies A plot of historic sinkhole occurrence was provided by
within an area classified as “Bare or thinly covered limestone SEI for the sinkhole occurrence within 1.6, 8 and 16 km
where the sinkholes are few, generally shallow and broad, radius from the Stauffer site (Fig. 27.8). At the date of the
and develop gradually”. They are referred to as cover subsid- study (June 2002):
ence sinkholes where the sandy soils (about 3–6 m thick)
migrate downward into local cavities. Sinkholes reported by • Seventy-eight sinkholes have occurred within the 16 km
newspapers in the vicinity are typically 3–6 m in diameter radius
and 3–6 m deep, which supports the Sinclair and Stewart • Thirty-five sinkholes have occurred within the 8 km
(1985) sinkhole map. radius
• Three sinkholes have occurred within the 1.6 km radius
27.5.5.2 Corps of Engineers Dredge Disposal
Sites Note that only three sinkholes were identified in the data-
In 1972 the Corps of Engineers used a parcel of land imme- base within 1.6 km of the site and are located approximately
diately southeast and adjacent to the Stauffer site for dredge 1.4 km east-southeast of the site where elevations start to
disposal (Fig. 27.6). The elevation at this site was about 3 m become higher and there is considerable development and
above sea level and the top of rock averaged about 6 m below concentration of surface runoff.
grade and was covered by a layer of clay (the SCL). The site Based upon the results of the database (Fig. 27.8) alone one
was adjacent to the Anclote River and both the surficial and could easily reach the conclusion that the site was susceptible
Floridan groundwater levels are shallow with little difference to sinkhole activity. Yet, on site, and within the immediate sur-
in head. Both geologic and hydrologic conditions provided a rounding area there had been no recent sinkhole activity based
relatively stable environment. No sinkholes developed at this upon aerial photo analysis back to 1926, corporate records and
dredge disposal site that is directly adjacent to the southern discussion with former workers at the facility.
series of settling ponds at the Stauffer site (Fig. 27.6). The database had recorded sinkholes that were all located
Stauffer purchased this property in 1981 as a buffer to the east, inland and upgradient of the site by almost a kilometer.
settling ponds. They occurred in the more developed areas at slightly higher
In 1999, the US Army Corps of Engineers utilized a par- elevations. Many of the sinkholes in the database have been
cel of land for disposal of Anclote River dredge material caused by cultural conditions such as those associated with
located about 915 m due east of the Stauffer site (Fig. 27.6). concentrated drainage of surface water, leaky sewers, water
The area had been excavated to the top of rock removing the line breaks or where construction was taking place. Note that
surficial sands and the semi-confining clay layer and a berm this database did not include the sinkholes from the Corps of
was built around the perimeter. Dredge spoil was pumped Engineers dredge disposal site in 1999 just 914 m east of the
into the disposal area with large volumes of water increasing site (Figs. 27.6 and 27.7).
the local head of water.
A few days after dredging started, four small sinkholes
(6–9 m across) reportedly developed in and around the site. 27.5.6 Review of Corporate Files
Figure 27.7 shows two of these sinkholes. These sinkholes and Interviews with Previous Workers
most likely developed as a result of sediment raveling into
localized voids within the upper portion of the Tampa On-site existing data and information was obtained from
Limestone due to the missing semi-confining layer and the Stauffer’s project files and included some site-specific geo-
large quantities of water associated with the dredged materi- logic data, information on the ponds/disposal areas (loca-
als. These are typical of the small, shallow sinkholes that tions and depths) and possible drum disposal areas. The
occur in the area (Sinclair and Stewart 1985). results of work by previous consultants all had a very specific
398 27 EPA Superfund Site

Fig. 27.7 Sinkholes at Corps of Engineers dredge disposal site in spoil area. ( b ) Repaired sinkhole within bermed dredge spoil area
1999. The location of these sinkholes is due east of the Stauffer (Photos courtesy of B. Bergen, Pinellas County Health
site and noted on Fig. 27.6 . ( a ) Sinkhole north of bermed dredge Department)

focus and were not intended to be comprehensive. Where 27.6 The Preliminary Conceptual Models
possible, this information was used to develop the beginning of
an on-site database as well as preliminary conceptual model The general geologic and hydrologic conditions were rea-
and included site-specific geologic data from drilling. sonably well defined by the regional literature and site-
Communication with retired workers from the site indicated specific reports prior to beginning any work at the site
that there were no incidences of subsidence or sinkholes during (Fig. 27.9). Subsequent efforts would be focused upon
the period of plant operations (1947–1981). This is a strong verifying these data and providing further detailed infor-
indication of geologic stability over the site under operating mation of site-specific conditions. Then, as additional data
conditions which included variable static and dynamic loading were acquired, the conceptual model was verified or
from piles of stored materials and railway traffic (Fig. 14.6). revised.
27.6 The Preliminary Conceptual Models 399

10-mile radius

5-mile radius

Stauffer Site

1.6 km radius 3 sinkholes

8 km radius

16 km radius

0 4 8

Approx. Scale in Kilometers

Fig. 27.8 The locations of sinkholes within a 1.6, 8 and 16 km radius of the site based upon a sinkhole database (SEI 2002)

27.6.1 Geologic Conceptual Model and silt with limestone fragments. Although the limestone
contains some clays and silts, the overall lithology is
The preliminary conceptual model of site geology competent and difficult to drill through.
(Fig. 27.9a) was based upon five regional borings extending • Zones of dissolution, high permeability and cave develop-
24–137 m below grade. These borings located the depths of ment were also cited in the literature
the Hawthorn Group, the Tampa Formation and the contact
with the Suwannee Formation. Site-specific data came from
on-site existing borings and site-specific reports and
indicated: 27.6.2 Hydrologic Conceptual Model

• a surficial layer of sand was found to be 3–6 m thick The site hydrogeology is a relatively flat, low flow system
becoming thicker toward the Anclote River, flowing to the south or southwest with ultimate discharge to
• a semi-confining clayey layer (SCL) ranges from 0.3 to the Anclote River by both the surficial and Upper Floridan
2.4 m thick, aquifers (Fig. 27.9b). The flat gradients, in combination with
• the Tampa Limestone was found to be about 15 m thick the relatively low hydraulic conductivities of both aquifers, are
over the site and lies unconformably over the Suwannee indicative of a low flow velocity groundwater system. Water
Limestone. These formations were found to be hard, levels in both aquifers are maintained relatively constant due
cemented layers interspersed with layers of soft, lime clay to the adjacent Anclote River and nearby Gulf of Mexico.
400 27 EPA Superfund Site

a Geology b Hydrology Avg water level in both

the Surficial Aquifer and
Sand avg 4 m thick Upper Floridan Aquifer
surficial aquifer was 1.8 to 3 m bls
Clays 0.3 to 2.4 m
SCL Semi- Confining Layer
Weathered Top of
Tampa Limestone

Isolated Voids in the Upper Tampa Limestone

Tampa Limestone Avg 15 m thick

Low -Density Zone at

~20 m
Unconformity

Upper
Suwannee Limestone
Avg 71.6 m Floridan
aquifer

Zone of Dissolution
lower half to lower third
(about 55 to 90 m)

~90 m Unconformity Note not to scale

Ocala Limestone

Fig. 27.9 Preliminary conceptual model of geologic conditions (a) and hydrologic conditions (b)

27.6.3 Karst Conceptual Model 1999 (Fig. 27.6). This appears to be due to the removal of the
SCL and the disposal of dredge material containing large
There was clearly sinkhole activity in the region both old and quantities of water. Furthermore, consider that the Corps of
new. Paleosinkholes were numerous in the aerial photo analysis Engineers dredge disposal site used in 1972, immediately
with two identified along the west side of the South Parcel, adjacent to the property, did not have any sinkholes develop
the Sinkhole Pond and Meyers Cove (Fig. 27.2). Both of where the SCL was left intact.
these features were identified in the earliest aerial photos Two photo lineaments had been identified extending
(1926) indicating that they were old and stable. through the eastern and western sides of the South Parcel,
The new sinkhole activity as reported in the databases and both at about 10°W (Fig. 27.6). The lineament to the west
local newspapers is relatively small and likely due to shallow aligns with Meyers Cove and Sinkhole Pond, both of which
voids in the top of rock. In addition, this activity lies to the east are old features. In addition, there are no recent sinkholes or
at a somewhat greater elevation, with thicker sand cover, in an current sinkhole activity associated with the lineament to
area with extensive development, and away from the stabilizing the east.
influence of the high water table of the nearby Anclote River
and the Gulf of Mexico. No active sinkholes were noted near
the Stauffer site in any of the data including the aerial photos 27.7 Shallow Geohydrologic Conditions
dating back to 1926. Furthermore discussions with individuals
who had worked at the Stauffer site for years had indicated that The shallow hydrogeologic conditions consist of the surficial
there had never been any sinkhole occurrence at the site. sandy aquifer and the semi-confining layer (SCL). The SCL
The nearest sinkhole activity occurred off-site 914 m to between the surficial aquifer and the Upper Florida aquifer
the east at the Corps of Engineers dredge disposal site in (Fig. 27.9) is critical to the site for two main reasons:
27.7 Shallow Geohydrologic Conditions 401

• Limiting contaminant transport downward into the metals. The GPR data did not provided indications of any
Floridan Aquifer and anomalous conditions within the sands over the SCL such as
• Limiting the downward movement of sands into local dipping strata or piping within the sands that might indicate
voids within the upper portion of the Tampa Limestone sinkhole activity.
resulting in sinkholes. In addition, slug test and pumping test results in the surfi-
cial aquifer indicated a range of hydraulic conductivities that
In addition, the head or pressure difference [static and are relatively low and consistent with that of a fine sand aqui-
dynamic (tidal)] between shallow and deeper aquifers could fer. The hydraulic conductivity of the surficial aquifer sands
impact a weak zone within the SCL by breaching this layer. ranged from 0.3 to 37.8 m/day based on both slug and pump-
For example, an upward head may lead to a local spring, and ing tests, and the surficial aquifer transmissivity ranged from
a downward head may aid in moving unconsolidated mate- 75 to 85 m2/day based on the pumping test results (Parsons
rial downward through weakened zones in the SCL resulting 1999, 2003).
in subsidence and possibly allow contaminants to flow into The general groundwater geochemistry in the surficial
the Floridan Aquifer. In any case, significant (man-induced) aquifer (sodium, chloride, salinity levels, etc.) indicates that
changes in head or breaching of the SCL could result in con- with the exception of proximity to the Anclote River, the
taminant transport and sinkhole collapse. groundwater in the proposed remediation area is considered
In addition, the lateral and vertical extent of the ponds fresh water and there appears to be little geochemical vari-
were assessed as well as the potential for buried drums, ability that would affect the proposed remedy (Parsons 1999,
tanks or other sources of contamination. These additional 2003).
objectives fall within the shallow hydrogeologic conditions
and will be the focus of the remediation.
27.7.2 The Semi-confining Layer

27.7.1 The Sands and Surficial Aquifer Various techniques were used to characterize the continuity,
thickness, and content of the semi-confining layer. Ground
The shallow geology was densely sampled using two con- penetrating radar typically mapped the lateral extent and
tinuous surface geophysical techniques, electromagnetic variations in depth of this layer, but not its thickness. While
conductivity measurements and ground penetrating radar resistivity imaging was primarily used for assessing deeper
(GPR). The electromagnetic measurements were completed geologic conditions, it also provided data on the SCL show-
using an EM31 which provide a composite conductivity ing its continuity, depth and relative thickness across the site
value to a depth of approximate 5.5–6 m. In addition, a sec- (Fig. 27.11). Detailed information on the SCL was provided
ondary measurement (in-phase data) indicative of the pres- by direct push electrical conductivity (DPEC). This Geoprobe
ence of buried metals was obtained simultaneously with the technique was used throughout the site and locations were
EM31. The GPR measurements were completed using a guided by the surface geophysical data. This technique pro-
100 MHz antenna and time window of 150 ns, which pro- vided information on the top of the SCL, the variability
vided a 2D image of the subsurface to a depth of approxi- within the SCL (clay content) and the bottom of the SCL
mate 7.3 m. Both of these measurements were obtained based upon electrical conductivity measurements as well as
along east to west parallel survey lines spaced 3 m apart. allowed core samples to be collected at select locations.
This provided approximately 100 % coverage over the entire Figure 17.3 shows an example of GPR over a low area in the
site with the exception of buildings, pond areas and piles of top of the SCL along with DPEC data. The DPEC data mea-
debris materials. sures variations in electrical conductivity, higher conductiv-
Electromagnetic measurements at the site (EM31) indi- ity indicates more clay content. In addition, geophysical logs
cate that the surficial sands are low conductivity materials, (natural gamma and induction) were obtained in all borings
about 10 mS/m, that are generally uniform both laterally and and provided very detailed profiles through the SCL showing
vertically. Existing geologic logs as well as subsequent mea- its depth, variability in thickness and variability of clay con-
surements and sampling on-site indicated that the surficial tent within the SCL.
sands at the site are fine to medium-grained, tan to orange in The semi-confining layer is composed of a clayey layer
color with traces of silt or clays and are found consistently that grades into a weathered, clayey limestone that is a part
across the site ranging in thickness from 2 to 8.5 m thick, of the upper Tampa Limestone. A zone of clay and weath-
averaging 4.5 m. Figure 27.10 is a contour map of the electrical ered limestone marks the top of the Tampa Limestone and is
conductivity across the site. Values of electrical conductivity considered to be the bottom of the SCL. The weathered zone
higher than background can be accounted for due to utilities, is typically harder and dryer, with limestone fragments that
building foundations, slag debris, settling ponds or buried distinguish it from the clayey layer above. The clayey layer
402 27 EPA Superfund Site

Fig. 27.10 EM conductivity contour map

is thought to be an unconformity at the top of the Tampa • Localized lows in the top of the SCL were identified and
Limestone, which is an erosional surface. are thought to be erosional surfaces
Figure 27.12 shows a contour of the top of the SCL based • The SCL is laterally continuous throughout the entire site,
upon DPEC data, all new and existing borings and geophysical with the exception of one area along the eastern edge of
logs. The semi-confining layer: the South Parcel.

• is described in the geologic logs as moist to wet sands, The nature of the semi-confining layer was further
silty clay grading with depth to very dry, very hard sandy, defined by permeability testing, that indicate hydraulic
silty clay and limestone conductivities of 2.84 × 10−4 to <1.0 × 10−8 cm/s which are
• ranges in depth from 3 to 8 m bls, dipping to the south and typically two or more orders of magnitude lower than the
has a range of elevations from almost +1.5 m MSL in the hydraulic conductivities of the surficial aquifer 10−2 to
north to −3 m MSL in the south near the river 10−4 cm/s. The contrast in hydraulic conductivities indicates
• has a variable thickness which ranges from 0.3 to 7.3 m, that the SCL would minimize the cross-connection between
with an average of about 2.3 m the surficial and Upper Floridan aquifers at the site
• varies in clay content based upon variations in 110 DPEC (Parsons 2003).
electrical measurements and 74 geophysical logs The chemical data collected directly beneath or adjacent
• is variable, but generally high in silt and clay content to settling ponds confirm that the semi-confining layer is
based upon X-ray diffraction, sieve analysis and hydrometer impeding the flow between the surficial aquifer and Upper
analysis Floridan aquifer. Chemical data show that no primary drinking
27.7 Shallow Geohydrologic Conditions 403

Distance in Meters
609 640 670 701 731 762 792 823 853 883 914 944 3200
0

Sand DIRECT-PUSH SAND

-3
620
540
-6 460
DIRECT-PUSH CLAY
380
300
260
-9
SCL Clays DIRECT-PUSH REFUSAL
220
180
140
Depth (m)

-12 120
100
80
60
-15 50
Limestone 40
30
-18 20
15
10
5
-21 0

Resistivity
-24
(ohm-m)
Brackish Water Center of model resistivity block

2000 2100 2200 2300 2400 2500 2600 2700 2800 2900 3000 3100 3200
Station Easting (ft)
0 30 60
Horizontal
Scale in Meters

Fig. 27.11 Example of 2D resistivity data along with two DPEC pushes identifying the top and bottom of the SCL

water standards were exceeded in the Upper Floridan aquifer have little or no effect on the proposed remedy in the surficial
(Parsons 2003). aquifer. The presence of the adjacent Anclote River and the
Pumping tests conducted in both aquifers showed no nearby Gulf of Mexico provide consistency to the water lev-
measurable effect on the non-pumping aquifer. The lack of els at the site. In addition, the tidal change is relatively small
measurable drawdown in the non-pumping aquifer during (about 0.76 m semi-diurnal tide from tidal tables) and the
the aquifer pumping tests indicates that there is not a strong head between the surficial aquifer and the Floridan aquifer is
hydraulic connection between the two aquifers. The vertical very small (fractions of a centimeter). These factors provide
gradient between aquifers is generally flat and varies from a high degree of stability and reduced stress on the SCL. This
slightly downward to slightly upward. The lack of a strong stability minimizes the opportunity for rupture of the SCL
vertical gradient would tend to minimize the potential for a leading to any contaminant flow from the surficial aquifer
weak zone or breach within the SCL layer. Therefore, the into the Floridan aquifer or the flow of sand into voids in the
potential for contaminants in the surficial aquifer to migrate uppermost Tampa Limestone.
toward the Upper Floridan aquifer or to enable the down-
ward movement of sands into local voids within the upper
portion of the Tampa Limestone resulting in sinkholes is also 27.7.3 Pond Boundaries
minimized by the SCL (Parsons 2003).
Water level data was collected over several tidal cycles. The pond disposal areas (Fig. 27.3) were characterized to
Tidal influence in the surficial aquifer was hardly discernible determine their boundaries, depths, and possible migration
beyond 150 m from the shoreline. Therefore, tides should of contaminants. Because the material in the ponds was elec-
404 27 EPA Superfund Site

Fig. 27.12 Elevation contour map of the top of the SCL based upon DPEC data, existing and new borings as well as geophysical logs

trically conductivity, the electromagnetic survey was the pri- A combination of DPEC and soil cores were used to confirm
mary technique used to meet this objective. Figure 27.10 the depth of the pond materials which range from 1.2 to
shows the conductivity contour map from the electromag- 6.4 m bls. Based upon this data the depths of the pond mate-
netic survey. rials do not appear to breach the semi-confining layer.
A sharp electrical conductivity contrast (from 10 to All ponds were found to lie within the surficial aquifer
greater than 30 mS/m) between the natural soil and pond sands and did not breach the SCL and no contamination was
materials allowed the pond boundaries to be mapped using found below the SCL in the upper Floridan Aquifer.
electromagnetic techniques (EM31). Other areas of elevated
conductivities are due to buried building foundations, utili-
ties and remnants of slag piles. Based upon electromagnetic 27.7.4 Buried Drums
measurements, the locations and boundaries of the ponds are
all in general agreement with previous maps produced by Concern was built around the rumor of a large cache of
other site investigators, with the exception of Pond 39 located drums being buried in one or more location at the site. Both
in the center of the North Parcel (Fig. 27.10). electromagnetic and ground penetrating radar provided
Pond 39 appears to be approximately 30 % longer, extend- essentially 100 % site coverage and included the areas where
ing to the north, and may be divided into two sections. buried drums were suspected. These techniques were use to
27.8 Deeper Geohydrologic Conditions 405

focus more detailed surveys for buried drums using multi- above the SDWS. Elevated sodium and chloride levels were
frequency EM system (MFEM) which responds to both fer- found in wells near the shoreline, a result of the proximity of
rous and non-ferrous metal and magnetometer data which the river and saltwater intrusion. One surficial well in a source
responds to only ferrous metal. area in the South Parcel contained elemental phosphorus and
Based upon electromagnetic and MFEM measurements, two contained gross alpha above the PDWS.
five areas were identified that potentially contain buried
metal debris (Fig. 27.10). While some ferrous metal was
indicated in the magnetometer data there was no indications 27.8 Deeper Geohydrologic Conditions
of the presence of a large cache of drums in one or more
locations at the site. The deeper geology consists of the Tampa and Suwannee
The corrosion of buried steel drums is highly variable Limestone, and the Upper Floridan aquifer. While no reme-
depending upon soil type, moisture, and chemistry, as well as diation will be taking place within this deeper zone, it was
the condition of the drum and its coatings. The drums of included in the investigation to determine the presence or
phosphorous waste would have been disposed of in water risk of contaminants as well as the risk of sinkholes occur-
filled containers then disposed of below the water table to ring within this zone.
prevent ignition of the phosphorous. Below the weathered zone that is considered part of the
The wall thickness of a 55-gal drum can range from 50 to SCL, the Tampa Limestone consists of layers of fine grained
75 mils. Based upon uniform corrosion rates of 1 to 10 mil limestones, fine to coarse-grained sands, gravelly limestones
per year (Uhlig 1965; Romanoff 1957) it would take as long and silty/sandy stiff clays. The contact between the Tampa
as 50–75 years or as short as 5–7.5 year for a drum to totally Limestone and the Suwannee Limestone is an unconformity
corrode. However, localized pitting corrosion can be much and difficult to identify. However, there is typically a high
more rapid leading to leakage in a much shorter time. Data production zone at this unconformity between the Tampa
from Tuthill and Schillmoller (1966) indicates a rate of and the Suwannee (Hickey 1982). Below this zone, the
15–30 mils per year or more for carbon steel in quiet seawa- Suwannee Limestone is highly variable, ranging in composi-
ter. At this rate pitting could penetrate a drum in tion from fine grained to sandy to gravelly limestone, with
2.5–3.3 years. layers of silty/sandy clays and clayey marls. Cave develop-
Drums that were disposed of in the ponds have likely ment is known to occur within the lower half to lower third
corroded, if not totally, then nearly so over time (more of the Suwannee Limestone.
than 20 years since the plant closed in 1983). This would The deeper geologic conditions were characterized using
reduce the probability of their detection by magnetometer 2D resistivity imaging and microgravity measurements.
measurement and make a successful barrel recovery program These two sets of geophysical data provided the capability of
unlikely. detecting changes in geologic conditions such as open cavity
systems and paleocollapse within the deeper strata. Both of
these measurements integrate a larger volume of the subsur-
27.7.5 Contaminants face and were acquired using survey lines spaced 30 m apart.
The 2D resistivity imaging data was obtained using a 6 m
Groundwater contamination in the surficial aquifer appears electrode spacing and a dipole-dipole array providing data to
as “hot spots” at pond areas in the North and South Parcels. a depth of 24 m. The resistivity imaging measurements pro-
Metals of concern at the site that exceed the Primary Drinking vide a 2D cross section of variations in electrical resistivity.
Water Standards (PDWS) (EPA 2000) include antimony, The data at this site consisted of three geologic layers, the
arsenic, cadmium, nickel and thallium. Most of the wells shallow sands, the SCL and the upper Tampa Limestone
containing these constituents only had one or two constitu- which all had difference electrical properties. Figure 27.11 is
ents each, while one South Parcel monitoring well had four an example of the resistivity data from the site showing these
of the constituents and another had three. These localized layers. A fourth layer, at the bottom of the data represents
“hot spots” are closely associated with source areas and brackish water intrusion from the Anclote River.
large areal plumes of these constituents do not exist at Microgravity measurements provided a means of charac-
the site. terizing possible karst features such as large voids, a cavern
Metal constituents exceeding Secondary Drinking Water system or areas of paleocollapse originating within the deeper
Standards (SDWS) (EPA 2000) in either aquifer include Tampa and Suwannee Limestone due to changes in density.
aluminum, iron, manganese, and zinc. Inorganic constituents While the microgravity measurement results in a single value
exceeding the SDWS include chloride, fluoride and sulfate. of relative gravity for each station measurement, it is the trend
Fluoride is the most widespread constituent at the site, being of these values along the survey line that indicate the presence
found in 18 surficial and 4 Upper Floridan wells at levels of karst conditions based upon changes in subsurface density.
406 27 EPA Superfund Site

Microgravity measurements were made at station intervals of able effect on the non-pumping aquifer. The lack of measur-
12 m across most of the site, but a spacing of 6 m was used able drawdown in the non-pumping aquifer during the
on the south half of the South Parcel, for a total of 1,826 aquifer pumping tests indicates that there is not a strong
measurements over the 52-ha. The tighter station spacing hydraulic connection between the two aquifers.
was requested by Dr. Stewart, from the University of South Tidal influences were noted in all Upper Floridan aquifer
Florida, who was part of the project oversight committee, wells including those in the North Parcel. Lag times in
and wanted to see more detailed data in the southern half of response to tidal fluctuations were generally shorter in the
the South Parcel where the critical site features were located. Upper Floridan aquifer than in the surficial. This shows that
In addition, drilling, sampling and geophysical logging the tides in the Anclote River had a greater affect on the
added details to the deeper geohydrologic assessment. Upper Floridan aquifer.
Geophysical logs were acquired in all existing and new
wells, piezometers and test borings. The suite of geophysical
logs included natural gamma, induction (electrical conduc- 27.8.1 Anomalous Areas
tivity), gamma, gamma (density), and neutron (porosity)
logs. The geophysical logs provided continuous vertical The microgravity data identified changes in subsurface den-
measurements of in-situ properties and were obtained sity and areas of possible deeper karst conditions (Fig. 27.13).
through PVC or steel casing. These logs provided high- This method was not susceptible to cultural interference like
resolution vertical data at the site. the 2D resistivity data and provided high quality data
The transmissivity of the Upper Floridan aquifer ranged throughout the site. The microgravity data indicated some
from 11 to 26 m2/day, based on the pumping test. Pumping distinct variations in subsurface density across the site. Three
tests conducted in the Florida aquifers showed no measur- anomalous areas of low gravity values occur and include:

Fig. 27.13 Contour map of microgravity data indicating variations in subsurface density
27.8 Deeper Geohydrologic Conditions 407

• An area of low gravity values occurs along the northern was acquired north of the site and indicated that the low-
edge of the North Parcel. gravity values extended further to the north.
• A broad band of low gravity values is observed across the It appears very likely that the high electrical conductiv-
eastern side of the South Parcel. ity and low gravity values are spatially coincident, but may
• Extremely low gravity values occur adjacent to the have separate causes. The high conductivity area along the
Anclote River and can be ignored since they are due to the northern fence line appears to be due to the localized pres-
mass deficit associated with the river. ence of shallow, high conductivity materials associated
with waste materials from plant operations. The area of
low-gravity values is related to a dissolutioned zone or
27.8.1.1 Northern Edge of the North Parcel paleocollapse feature that is centered to the north, off of the
A high electrical conductivity anomaly (Fig. 27.10) corre- Stauffer site.
lates spatially with the low gravity anomaly detected along
the northernmost fence line of the site in the North Parcel 27.8.1.2 Eastern Edge of the South Parcel
(Fig. 27.13). A review of historical aerial photographs indi- The eastern edge of the South Parcel contains a large linear
cated that previous on-site activity was conducted in this south-to-north trend of lower gravity values (about −20 to
area. In addition, the location of the west-east road that runs −30 microgals). Within this linear trend are two localized
along the northern edge of the site was actually moved south areas of even lower gravity values (−60 to −80 microgals)
by about 30 m in the 1970s. DPEC measurements were com- (Fig. 27.13). This linear trend coincides with a major photo-
pleted in the area and a localized area of high conductivity lineament through the area (Fig. 27.6) and the SCL is miss-
materials were detected within the upper 2.4–3 m over the ing in the GPR data throughout this area. DPEC measurements
SCL layer. A deep test boring was completed within the area were used to evaluate whether this gravity low was due to a
of the low-gravity values and geophysically logged. Low- deeper SCL and top of bedrock.
densities were detected between 6 and 18 m based upon the Figure 27.14 shows the linear low gravity values along
gamma-gamma (density) logs. Additional microgravity data the eastern edge of the site and the corresponding top of SCL

Fig. 27.14 Correlation between gravity anomaly along eastern edge of the South Parcel (a) and top of semi-confining layer (b). Seven rotosonic
boring locations are also shown in (a)
408 27 EPA Superfund Site

contour based upon DPEC data that had a maximum depth of Additional supporting data include:
penetration of 12 m. The DPEC data indicated that the SCL
dips inward around the perimeter of the low gravity values • The cave systems and springs in and around this region all
and is absent within the center of the low gravity values. indicate that major dissolution has occurred within the
There was a direct correlation between the missing SCL Suwannee Limestone some 45–60 m bls.
layer and the low gravity values. • The lower sea levels of up to 130 m (Balsillie and
Rotosonic borings were then used to assess deeper geo- Donoghue 2004) support the development of cave sys-
logic strata across the site. Borings to depths of 23–48 m tems at depth.
were made at 22 locations with the rotosonic drilling method. • An obvious lineament was noted extending from the
Continuous core samples were obtained and geologically Kreamer Bayou in the Anclote River northward through
logged. Rotosonic borings were made in both background the east side of the Stauffer site at about 10°W (Fig. 27.6).
and anomalous areas defined by the microgravity This lineament passes through the linear paleocollapse
measurements. area and crosses through the microgravity anomaly at the
Six rotosonic borings (SB-10 through SB-15) were north edge of the north parcel.
located within the areas of lowest gravity values along the
eastern edge of the site (Fig. 27.14). Boring SB-9 was located These supporting data are not conclusive on their own,
in a background area south of this low-gravity area. The bor- but when combined with the other data they provide a coher-
ings within the area of low gravity indicated that the SCL ent understanding of geologic conditions at this site.
and underlying limestone was missing. The materials The areas of low gravity values (Fig. 27.13) exhibit no
encountered were generally unconsolidated, interbedded visible surface evidence of settlement, nor are there any indi-
sand and clays with some limestone fragments. In addition, cations on historical aerial photos. Due to the age of the fill
thin bands of organic materials (wood, peat, etc) were materials, it is believed that these paleocollapse zones are
observed intermittently over a large depth interval. These and will remain stable for the following reasons:
unconsolidated layers along with organic materials were
encountered where the Tampa and Suwannee Limestone • Their age from the carbon dating suggest that they have
should have been. been in place for >40,000 years
Samples of organic material from test borings were sub- • Large piles of materials have been in place on-site at one
mitted for Carbon-14 age dating. These organic samples time or another resulting in long periods of static loading
from 15.8 to 25 m were carbon dated to 40,000 and over much of this paleocollapse area.
48,800 years before present. The lower Tampa Limestone • In addition, dynamic loading has occurred due to railroad
was deposited more than 24 million years ago and the Upper car movement over decades with no impact.
Suwannee Limestone was deposited before that. The age of • The north to south county Anclote road over and adjacent
the organic material indicates that the paleocollapse occurred to this area has been in place for decades. Continued traf-
long after the Tampa and Suwannee were deposited. fic on this road includes heavy cement trucks and trailers
Geophysical logs obtained in all of the rotosonic borings carrying reinforced concrete structures. These trucks have
within the gravity anomaly confirmed that the semi-confining created significant loads with low frequency vibrations
layer was missing and no limestone was encountered to the and have traveled over this area for many decades with no
maximum depth of borings. Figure 27.15 shows a cross section impact or indications of subsidence.
of the gamma-gamma (density) logs from background in the
south through the gravity low to background to the north. Additional surface and borehole geophysical measure-
These data also show that the paleocollapse feature has an ments were made to provide in-situ values of P and S wave
abundance of low-density zones at random depths. Note that seismic velocities to enable bulk modulus values of the sands
all of the low-density zones are at a depth of 12 m or more and limestone to be calculated. These measurements were
within the paleocollapse area. This suggests a degree of located in both background and anomalous areas for com-
stability within the upper 12 m of sediment. parison and included:
The integration and correlation of these independent data
(microgravity, DPEC pushes, core obtained from rotosonic • Surface seismic refraction measurements to provide rep-
borings, geophysical logs, and carbon dating) clearly provide resentative P-wave velocities within the Tampa Limestone.
a solid case for interpretation of the area as a paleocollapse • Vertical seismic profiling (VSP) measurements to depths
feature. The spatial extent, geologic character, and age of of 23–48 m to provide representative P and S wave seis-
the paleocollapse have been defined (Fig. 27.16) along with mic velocities within the Tampa and Suwannee Limestone.
its depth of origin at more than 48 m below grade (the limit • Hole to hole (P and Shear wave) measurements to 23 m
of drilling data acquired). deep were made in background conditions to provide
27.8 Deeper Geohydrologic Conditions 409

South North

12
Depth Below Land Surface (meters)

Background Conditions

18

24

30

36

42

48

Low Gravity Centered Low Gravity

at Pond 49B Centered at 2000N

Fig. 27.15 South to north cross-section of seven geophysical density logs through gravity anomaly along eastern edge of the South Parcel.
Location of the borings is shown on Fig. 27.14

representative P and S wave seismic velocities within the 27.8.2 The Sinkhole Pond
surficial sands and the upper Tampa Limestone.
The sinkhole pond located on the northwestern edge of the
It was found that the P-wave velocities across the site southern parcel is a shallow circular depression of approxi-
were quite similar in both the background and anomalous mately 42 m in diameter (Fig. 27.2). It is water-filled much
areas. The shear-wave velocities from the background areas of the year but becomes mostly dry at times. Ground pene-
were 40–60 % higher than shear-wave velocities measured trating radar data indicate that the top semi-confining layer
within the unconsolidated materials of the paleocollapse becomes deeper towards the pond (dipping to the west). This
zones (see Sect. 18.4, Figs. 18.17 and 18.18). This implies dip in the semi-confining layer was confirmed with DPEC
the presence of weaker, less consolidated materials within data at the edge of the sinkhole. The data indicate that the
the paleocollapse features. However, no surface evidence of SCL is approximately 4.5–5.5 m bls immediately east of the
unstable conditions has been noted over the paleocollapse sinkhole pond.
area throughout the plant operations of 40 years or during A series of wash borings were used in the pond to detect
our site characterization. the presence of the SCL and top of rock. Wash borings indi-
410 27 EPA Superfund Site

South Linear Zone of Paleocollapse North

Localized Localized
LowGravity LowGravity
Values Values

Sand
0
SCL

Tampa Limestone

12
Elevation in Meters (MSL)

24
Suwannee Limestone

36
Intermittent Limestone
Organics Present

Low Densities Measured

48
? ? ? ?
120 240 360 480 600 720 840 960
Distance in Meters

Fig. 27.16 South to north conceptual model of the paleocollapse feature based upon microgravity, DPEC data, borings, geophysical logs, the
presence of organic materials and C-14 dating

cate that the sediment thickness increases toward the center feature is at least 45–60 m below MSL placing it in the mid
of the pond indicating the presence of a sinkhole. The sink- to lower Suwannee Limestone.
hole pond is not in an area of potential remediation and
should not have any impact on the proposed remediation
activity. 27.8.4 Within the Anclote River Adjacent
to the Site

27.8.3 Meyers Cove The Anclote River lies immediately south of the site and has
numerous circular areas (bayous) indicating the presence of
Meyers Cove is located on the southwestern edge of the paleokarst activity (Fig. 27.1). Marine geophysical surveys
southern parcel just south of the sinkhole pond (Fig. 27.2). It were run in the river to assess karst conditions immediately
is about 100 m east to west and 210 m north to south, con- adjacent to the site. The measurements included bathymetry,
tains boat docks for the adjoining housing development and side scan sonar and seismic reflection data. Positioning was
is one of many large bayous seen within the Anclote River. accomplished by GPS.
Seismic reflection data were acquired in Meyers Cove along Bathymetry and side scan imaging were obtained to look
multiple parallel survey lines spaced 6 m apart. Several fea- for indications of local depressions in the river bottom, which
tures in the seismic reflection data (hyperbolas and dipping may indicate the presence of sinkhole activity. Seismic
strata) indicate past paleocollapse activity in this area as reflection data provided 2D cross sections and was used to
would be expected (Fig. 27.17). Based upon the seismic evaluate the presence of paleokarst sinkhole activity below
reflection data, the depth of origin for this paleocollapse the river bottom.
27.9 The Conceptual Model for the Site 411

South North River as circular features (bayous), along with Meyers Cove
Northing (feet) and the sinkhole pond. These features can be seen on the
900 1000 1100 1200 1300 1400 1500 1600 earliest aerial photos (1926) and therefore are quite old and
0 appear to be stable (Fig. 27.5). Three photolineaments, a
nearby spring, and small sinkholes at a number of locations
were all found within the area.
Two additional karst features have been identified on site:
11

• One feature is located at the very north edge of the north

parcel (Fig. 27.13). This appears to be the southern edge
21
of a paleocollapse that extends further north off-site and
was not of further interest to this site characterization due
Estimated Depth Meters

to its off-site location.

• The second feature was located along the eastern edge of
32 the South Parcel. It was a large linear paleocollapse (about
4 ha). It extends along a linear south-to-north trend with a
portion extending under an area of settling ponds and off-
site to the east (Fig. 27.13). There was no evidence of this
43 paleocollapse seen at the surface and it does not show up
on the oldest aerial photos from 1926.

The photolineament (N10°W) along the eastern side of

53 the site (Fig. 27.6) intersects both of these paleocollapse
areas.

64 27.9 The Conceptual Model for the Site

274 Distance in Meters 487 Key features such as detailed lithology and karst condi-
tions along with the vertical and horizontal hydraulic gra-
Fig. 27.17 Marine seismic reflection data from Meyers Cove showing dients, hydraulic conductivities, the direction of
dipping strata and voids (hyperbola) evidence of paleokarst to depths of groundwater flow, the effects of tides, and contaminant
60 m or more
distributions have been established. Based on these fac-
tors, a conceptual model of the geologic, hydrologic and
karst conditions has been developed and is illustrated in
Seismic reflection data from the Anclote River southwest Figs. 27.9, 27.16 and 27.18.
of the Stauffer property identified two paleokarst collapse
features. Figure 27.6 shows the location of these anomalies. • These studies found that 92 % of the Stauffer site
These features were observed in data collected on both the (about 48 ha) do not contain areas of deep-seated
south and north sides of the channel, which is more than 30 m paleocollapse karst or any obvious indications of shal-
wide. This data was similar to that found in Meyers Cove low subsidence activity and will easily support the pro-
(Fig. 27.17) suggesting a depth of origin of about 45–60 m. posed remedy.
• Two buried paleokarst features were found over the site
(about 8 % or 4 ha) (Fig. 27.13). One is located on the
27.8.5 Summary of Karst Conditions north edge of the North Parcel and extends off site to the
north. It is not of concern since only its southern edge
The assessment of subsidence, sinkholes or paleocollapse extends onto the site. Furthermore, no remediation will
activity ranged from a regional assessment to a detailed site- take place there. Only one large paleokarst feature on the
specific investigation using a wide variety of observations eastern side of the site is of concern (Fig. 27.16). The ori-
and measurements. The Stauffer site lies in an area where gin of the paleocollapse features appears to be in the
karst conditions are known to exist. Many karst features lower portion of the Suwannee Limestone based upon
were identified in the area surrounding the site including; rotosonic borings. Based upon C-14 dating this karst fea-
larger, paleosinkholes, which are seen throughout the Anclote ture collapsed more than 40,000 years ago and then begin
412 27 EPA Superfund Site

South North
Approx.Extent of Saltwater Intrusion Surficial and
Anclote River within the Surficial aquifer Floridan Aquifer
Settling Ponds Groundwater Flow

Surficial
Sands water table for both aquifers
aquifer
0
Semi-Confining Layer
Elevation in Meters (MSL)

Tampa Limestone

15
Floridan
Suwannee Limestone
Low-Density Zone aquifer
at Unconformity

30

Approx. Extent of Saltwater Intrusion

0 100
within the Floridan Aquifer
Approx. Scale in Meters
Note: vertical exaggeration of 10:1

Fig. 27.18 South to north conceptual cross section of site geology and hydrology

to fill with sediments. This paleocollapse area appears to • There are no groundwater factors at the site that should
be very stable with an extremely low risk of reactivation affect the implementation or the long-term effectiveness
over the life of the remedy. of the remedy presented in the Record of Decision.
• The semi-confining layer is continuous over about 92 % • The disposal pond boundaries are well defined and there
of the site. Its thickness ranges from 0.3 to 7.3 m, with is no evidence that the pond material extends beyond the
an average thickness of 2.3 m. This semi-confining confines of the pond basins. The depth of pond materials
layer is minimizing cross contamination between the does not appear to breach the semi-confining layer.
surficial and Upper Floridan aquifers and preventing • Areas of buried metal debris have been identified;
sinkhole development caused by voids in the upper however, the data does not support the presence of a large
Tampa Limestone. cache of drums remaining in one or more locations at
• The site hydrogeology is nearly flat, low flow system with the site. It is very likely that the steel drums, if they were
ultimate discharge to the Anclote River by both the surfi- present, have corroded in the two or more decades since
cial and Upper Floridan aquifers. the plant has shut down.
• The vertical head between the Surficial aquifer and
Upper Floridan aquifer is minimal; therefore there is little
driving force for the downward migration of contaminants,
even in those areas of the site where the SCL is thin or 27.10 Sinkhole Risk Assessment
absent. In addition there is little driving force for the
downward migration of sands and development of shallow The natural dissolution of limestone is a slow process, occur-
sinkholes. ring at a rate of about 2.5 cm/1,000 years. As such, further
• Groundwater contamination for metals above the Primary dissolution of rock by itself is not a problem to the life of
Drinking Water Standard is generally localized near engineered structures (Sowers 1996). The concern then lies
source areas in the surficial aquifer. with existing dissolution features and the risk of sinkholes
• The areas of brackish water intrusion in the surficial aqui- developing due to these dissolution features. Three zones of
fer are generally restricted to the shoreline. dissolution were identified at the site and briefly discussed in
• Tidal affects on the surficial aquifer are minimal and Sect. 24.4.1 Risk Assessments (Fig. 24.8).
should have little or no effect on the proposed remedy.
• Off-site drinking water wells are not impacted by the site Zone 1 is a shallow zone of dissolution and weathering that
because the groundwater flow in both aquifers discharges exists at the top of the Tampa Limestone just beneath the
into the Anclote River. SCL. This zone contains small isolated voids or loose
27.11 The Ability of Geology to Support the Proposed Remedy 413

materials. This zone lies approximately 3–9 m bls 27.10.1 Summary of Risk Assessment
(dipping from north to south). The presence of the semi-
confining layer acts as the barrier, preventing the overlying The water levels in both the Surficial and Upper Floridan
sands or contaminants from migrating into these localized aquifers are shallow with flow towards the Anclote River and
voids within the upper portion of the Tampa Limestone. there are no significant vertical gradients. These conditions
Zone 2 is the unconformity between the Tampa and the are moderated by the presence of the river and will aid in
Suwannee Limestone (about 18 m bls) that was identified preventing any extreme groundwater conditions to occur on-
in both the regional literature and the geophysical logs as site. If the site were left by itself, unaffected by further activ-
typically being a low-density zone. ity, the probability of development of small shallow cover
Zone 3 is a deeper zone within the Suwannee Limestone subsidence sinkholes below the SCL in Zone 1 or redevelop-
some 45–60 m bls in which cave systems and springs ment of a paleocollapse sinkhole or development of a new
have developed throughout the region based upon regional deep-seated cover collapse sinkhole within the Tampa or
literature, and is further verified by the seismic reflection Suwannee Limestone (Zones 2 and 3) is considered to be
data in Meyers Cove and the Anclote River. virtually zero (Fig. 24.8).

Only Zone 1 at the top of the Tampa Limestone just beneath

the SCL is of any concern (Figs. 24.7 and 27.9a), and then 27.11 The Ability of Geology to Support
only if the SCL is breached by remediation or subsequent the Proposed Remedy
activity on site. A worse case scenario would result in a small,
localized sinkhole 6–9 m in diameter, similar to those at the One of the objectives of this investigation was to evaluate the
Corps of Engineers dredge disposal site (Figs. 24.7 and 27.9a). ability of the sites geology and hydrology to support the pro-
The other two deeper zones of potential karst development, posed remedy over the life of the remedy as required under
Zones 2 and 3 (Figs. 24.7 and 27.9a) are at depths well below reasonably anticipated site conditions. The following com-
any possible influence on remediation efforts at the site. ments were provided at the completion of the site character-
In addition, there is the paleocollapse (Zone 3), which is ization investigation and are based upon the proposed
located on the eastern side of the South Parcel a portion of remediation of soils at that time.
which is under the pond areas (Pond 49A and 49B) Most of the Stauffer site, about 92 % or 48 ha, was found
(Fig. 27.13). Concern here lies with the potential for reacti- to consist of background geohydrologic conditions. Evidence
vation of this paleocollapse feature. of past sinkhole activity is provided by the sinkhole pond and
Many factors may trigger sinkhole collapse including both Meyers Cove along the western edge of the South Parcel.
natural and man-made activities. The dominant factors in These two features were present in the 1926 aerials indicat-
triggering collapse are water related such as heavy prolonged ing that they are quite old. In addition, two areas of buried
rainfall and or drought, concentration of surface water run- paleocollapse were identified, one along the northern edge of
off, or changes in groundwater levels by heavy pumping. the site and one along the eastern edge of the site. Only the
In addition changes in surface loads, vibration, blasting, and eastern edge paleocollapse is of concern due to its location at
excessive drilling fluid or grouting pressures may trigger col- the edge of a remediation area.
lapse (Benson and Hussin 2002; Benson et al. 2003). The proposed remedy consists of:
Groundwater levels at the site are controlled by the adja-
cent river and the nearby Gulf of Mexico and remain rela- • Excavation of certain materials
tively constant. Tidal changes are the most significant factor • Consolidation of materials
in changing groundwater levels and primarily influence the • Capping of materials with an impermeable layer
Floridan aquifer. There is only a very small difference in • In-situ stabilization of pond materials
head between the surficial and Floridan aquifers (thou-
sandths of a meter). There is no excessive pumping of Figure 27.19 illustrates the conceptual cross section of the
groundwater from either aquifer within the property. proposed remedy and its potential impact on the geologic
Dynamic loading and vibrations are potential triggering hydrologic and karst conditions. All remediation will occur
mechanisms that have existed at the Stauffer site during above the semi-confining layer and above Zone 1 (the epikarst
plant operations (1947–1981). They were associated with and top of Tampa Limestone). It is critical to avoid breaching
railway operations along the eastern portions of the site, as the SCL during remediation or by long-term site use.
well as static loading from large piles of coke, ferrophos- Excavation of materials will decrease loading within the
phorous and slag (Fig. 14.6). These conditions existed excavated area. This is not considered to be a major risk fac-
directly over or near the area of the linear paleocollapse tor. However, any topographic low created due to excavation
feature along the eastern boundary of the South Parcel will tend to concentrate surface water run-off. Storm water
with no effect. will have to be managed during and after construction so that
414 27 EPA Superfund Site

1) Cap over
Rainfall
Consolidated Materials
Prevents Infiltration
Storm Water
Management
Ground
Surface Feature 2) Material
Above Water
Table
Remains Dry

3) Material
Surficial Below Water
aquifer
Table is
Solidified

4) Semi Confining
Upper Layer Present
Floridan over 92% of the site
aquifer

Fig. 27.19 Conceptual model of remediation factors and their impact to the shallow geology

large quantities are not concentrated in a local area. bilization must also be controlled to avoid breach the
The materials to be excavated may extend to the top of the semi-confining layer.
semi-confining layer. Therefore, care will be necessary to
avoid breaching the semi-confining layer during excavation. Portions of Ponds 49A and 49B (Fig. 27.13) that require
Recall that two sinkholes 914 m east of the site resulted remediation lie over the southern portion of the paleocol-
when the Corps of Engineers excavated the SCL and used lapse area covering a total of less than 0.2 ha. The SCL as
the area for disposal of dredge material. well as the Tampa and Suwannee limestones are missing
within this area of paleocollapse. Geophysical logs within
Consolidation of materials will increase loading in selected the paleocollapse zone indicate localized low-density zones
areas. Both static loading as well as dynamic loading has likely associated with the random sediments filling the paleo-
occurred at the site during plant operations and was much collapse. These low-density zones (Fig. 27.15) are all deeper
greater than that expected due to consolidation. The con- than 12 m and deeper than remediation activities. The paleo-
solidation itself and its effects are not considered to be a collapse feature itself appears to be stable from its origin at
risk factor. depth within Zone 3 or deeper. Regardless of the very low
Capping of materials with an impermeable layer will poten- risk, monitoring of conditions (vibration and or pressure)
tially change the surface water run-off characteristics. during remediation of Ponds 49A and 49B will be necessary
Storm water run-off will have to be managed so that it is to minimize any potential impact.
not concentrated in a local area. The capping system will Once the stabilized material hardens groundwater flow
be designed to control run-off and avoid concentration or patterns may be changed slightly. It is expected that the
build up of surface water run-off. Therefore, capping is stabilized material would have reduced permeability, and will
not considered to be a major risk factor. act as an obstruction to groundwater flow. Because no pond
In-situ stabilization is to be focused on pond materials and material appears to extend below the semi-confining layer,
is carried out by blending a cement or binding mixture this change in groundwater flow should be limited to the
with in-place soils. Some vibrations and/or pressures are surficial aquifer and is expected to have negligible impact to
created during the mixing associated with the stabiliza- the sites hydrology.
tion process. The amplitude of vibration and its frequency
or pressures is not known at this time. Excessive vibra-
tions and or pressures could breach the SCL allowing 27.12 About the Site Characterization
sands to ravel into small isolated voids within the surface Strategy
of the Tampa Limestone and contaminants to migrate into
the Florida aquifer. Measures will be incorporated into This project was unique in several ways, but the two that had
the design and implemented during construction to moni- the greatest impact was the involvement and technical under-
tor and control vibrations and pressures. The depth of sta- standing by the client to get the project done right the first
References 415

time as well as a budget to allow the team to get things com- Beck BF, Sayed S (1991) The sinkhole hazard in Pinellas County: a
pleted in a timely manner. In addition, the project followed geologic summary for planning purposes. Florida Sinkhole Research
Institute and Pinellas County Department of Public Works. Report
many of the core site characterization strategies imbedded 90-91-1
within the ESC process that included: Benson RC, Hussin J (2002) Assessment of karst, pseudokarst and its
remediation. Workshop presented at Geophysics 2002, 2nd annual
• An active team of experienced professionals who were conference on the application of geophysical and NDT methodolo-
gies to transportation facilities, Federal Highway Administration,
involved from the beginning to the end of the project in 15–19 Apr 2002, Los Angeles, CA
the data acquisition, interpretation and integration. This Benson RC et al (2003) Assessing the risk of karst subsidence. In:
experience and continuity of the project team created a Proceedings from the AEG/AIPS 2002 joint annual meeting, 22–29
high level of confidence in the project results. Sept 2003, Reno, NV
Cherry RN (1966) Chloride content of ground water in Pinellas County,
• Accurate data was obtained by having hands-on experi- Florida in 1950 and 1963. Florida Geologic Survey, Map Series No. 20
enced professionals in the field, who developed standard Coble RW (1973) The Anclote and Pithlachascotee Rivers as water-
operating procedures (SOP) for the project and utilized a supply sources. USGS, Map Series No. 61
strong quality control program throughout the project. Environmental Protection Agency (2000) National primary and second-
ary drinking water regulations 40 CFR Parts 9, 141, 142 and 143.
• Because no single method or measurement will uniquely Fed Regist Vol. 65 No. 87, 4 May
define subsurface conditions, a wide range of independent Green R et al (1995) Lithostratigraphic and hydrostratigraphic cross-
measurements responding to different parameters and vol- sections through Pinellas and Hillsborough Counties, southwest
ume of measurement (scale) were used and integrated. Florida. Florida Geological Survey, Tallahassee, Open File Report 61
Hickey JJ (1982) Hydrogeology and results of injection tests at waste-
• The correlation of independent measurements provided a injection test sites in Pinellas County, Florida. USGS, Water Supply
significant increase in the level of confidence in the inter- Paper 2183
pretation and final conceptual model. Parizek RR (1976) On the nature and significance of fracture traces
• Active involvement and consensus building of all stake- and lineaments in carbonate and other terranes. In: Yevjevich V (ed)
Karst hydrology and water resources, Vol. 1, karst hydrology,
holders from the beginning and throughout the project Ft. Collins, CO, Water Resources Publication 3-1/3-108
allowed technical questions or concerns to be addressed Parsons Engineering Science, Inc. (1999) Baseline groundwater moni-
and resolved as they occurred. toring report. Stauffer Management Company, Tarpon Springs,
Consultants report
Parsons Engineering Science, Inc. (2003) Draft groundwater investiga-
While this project was unique, the basic approach can be tion of the Stauffer Site. Stauffer Management Company, Tarpon
scaled down and successfully applied to other karst site Springs, Consultants report
characterization. Reichenbaugh RC (1972) Sea-water intrusion in the upper part of the
The final investigation results were submitted by Parsons Floridan aquifer in coastal Pasco County, Florida, 1969. Florida
Geologic Survey, Map Series No. 47
Engineering Science (2003) and Technos, Inc. under the Romanoff M (1957) Underground corrosion. National Bureau of
cover of O’Brien and Gere (2003). Papers describing this Standards. Circular 579
project were presented by Yuhr (2003); Yuhr et al. (2003). Scott TM (1988) The lithostratigraphy of the Hawthorn Group
The once EPA Superfund site has been cleaned up and (Miocene) of Florida. Florida Geological Survey Bulletin No. 59
Scott TM et al (2004) Springs of Florida. Florida Geologic Bulletin No. 66
awaits redevelopment. It is one of the largest undeveloped Sinclair WC, Stewart JW (1985) Sinkhole type, development, and dis-
tracts of land along the Anclote River left in Pinellas County. tribution in Florida. US Geological Survey, Map Series No. 110
The contaminated soil was capped beneath a layer of dirt and Southwest Florida Water Management (SWFWMD) (1988) Ground-
sod. More than $21 million was spent for remediation efforts water resource availability inventory: Pinellas County, Florida.
Southwest Florida Water Management District, Brooksville
that were completed in 2011. Because polluted soil remains Sowers GF (1996) Building on sinkholes, design and construction of
on site homes can never be built there. Local legislators are foundations in karst terrain. American Society of Civil Engineers,
trying to attract industrial and manufacturing business to the New York (out of print)
site. The state of Florida has budgeted funds for dredging, Technos, Inc. (2003) Geophysical studies report for Stauffer
Management Company, Tarpon Springs, Florida. Submitted to
wharf stabilization and road improvements at the site. One O’Brien & Gere Engineers, Inc. June, Consultant Report
manufacturing client is seriously considering the property Telford S (2014) Cleaned up Stauffer Chemical plant site awaits rede-
(Telford 2014). velopment. Florida Specifier, September, Goldenrod, Florida
Tuthill AH, Schillmoller CM (1966) Guidelines for selection of marine
materials. International Nickel Company, New York
Uhlig HH (1965) Corrosion and corrosion control an introduction to
References corrosion science and engineering. Wiley, New York
US Environmental Protection Agency (1998) Record of decision, the
Arthur JD et al (2001) Lithostratigraphic and hydrostratigraphic cross- decision summary, operable unit 1. Stauffer Chemical Tarpon
sections through Levy-Marion to Pasco Counties, southwest Florida. Springs Site, Tarpon Springs
Florida Geological Survey, Tallahassee. Open File Report 81 US Environmental Protection Agency (2000a) Agreement to conduct
Balsillie JH, Donoghue JF (2004) High resolution sea–level history for studies to include: geophysical studies, treatability studies, and
the Gulf of Mexico since the last glacial maximum. Florida Geologic groundwater characterization studies. Stauffer Chemical Tarpon
Survey Report of Investigation 103 Springs Site, Tarpon Springs
416 27 EPA Superfund Site

US Environmental Protection Agency (2000b) Scope of work, agree- on sinkholes and the engineering and environmental impacts of
ment to conduct studies to include: geophysical studies, treatability karst: karst geohazards, Gatlinburg, TN. Balkema, Rotterdam,
studies, and groundwater characterization studies. Stauffer pp 79–91
Chemical Company Superfund Site, Tarpon Springs Yuhr L (2003) A case history of a comprehensive karst investigation in
Vernon RO (1951) Geology of Citrus and Levy Counties, Florida. Tarpon Springs, Florida. Presented at the 18th annual regional phos-
Florida Geological Survey, Bulletin No. 33 phate conference Lakeland, FL, 16–17 Oct 2003
Wetterhall WS (1964) Geohydrologic reconnaissance of Pasco and Yuhr L et al (1996) Achieving a reasonable level of accuracy in site
southern Hernando Counties, Florida. Florida Geological Survey, characterization in the presence of geologic uncertainly. In:
Report of Investigations No. 34 Shackelford CD, Nelson PP, Roth MJS (eds) Proceedings of uncer-
Wetterhall WS (1965) Reconnaissance of springs and sinks in west-central tainty 1996 – uncertainty in the geologic environment: from theory
Florida. Florida Geological Survey, Report of Investigations No. 39 to practice. ASCE Geotechnical Special Publication No. 68
Wilson WL (1995) Sinkhole and buried sinkhole densities and new Yuhr L et al (2003) A case history of a large karst investigation.
sinkhole frequencies in karst of northwest Peninsular Florida. In: Presented at the 3rd international conference on applied geophysics,
Beck BF (ed) Proceedings of the 5th multidisciplinary conference geophysics 2003 Orlando, FL, 8–12 Dec 2003
 Index

A Buried sinkholes, 19, 20, 54–55, 108, 126, 175, 190–191, 195
Acid leaks, 4, 8, 59, 195, 215, 218, 239–240, 290, 322 Buried stream channel, 193
Aerial photos/photography
analysis, 358, 372
coverage, 110, 134, 148 C
fracture trace, 135, 136, 194 Capture zones for Silver Springs, 303
infra-red (false color), 140 Carbon-14 age dating, 11, 408
lineament analysis, 112, 185, 313, 368, 380 Carbonates
oblique aerial photos, 137, 158, 393 chalk, 7, 10, 14, 62, 78, 121
photo lineaments, 109, 135, 156, 181, 185, 311, 368, 372, 380, dolomite, 7, 8, 10, 13, 14, 31, 63, 82, 121, 126, 237, 336, 392
395, 400 limestone, 3, 7, 27, 41, 52, 99, 117, 125, 133, 145, 201, 227, 251,
qualitative interpretation, 134 265, 275, 295, 307, 319, 333, 365, 385
quantitative interpretation, 134 Cave crawler, 263
scale of photos, 134, 154 Cave divers, 17, 20, 29, 31, 45, 57, 71, 81, 187, 236, 253, 256,
Air flow, 150–152, 165, 281 259, 395
Airport runways, 50, 60, 197 Cave mapping system
Alignment of Circular lakes, 366, 367 cave radio (cave location), 32, 129, 130, 188, 190, 253, 259–260
Anecdotal information, 127, 131, 157 inertial navigation system (INS), 259, 263
Aquitards and barriers, 28, 92, 121, 126, 131, 243–245, 275, 276, 279, Cavern collapse, 75–84
290–293, 309, 386, 413 Caves
Areas affected cave conduits, 30, 395
US, 41, 42 databases, 109, 129, 130
worldwide, 41–42 geometry and densities, 30–31
Artesian well, 285–289 large cave systems, 9, 31, 92, 370, 376
Assembly of data, 308–309 management, 32
Assessment of data, 109, 308 maps, 29–30, 109, 127, 129, 251, 253, 259, 283, 313, 392
Automatic water samples, 301, 303 types of caves, 31–32
Autonomous vehicles (AUV) Cave system, 4, 9, 12, 14, 17, 23, 29–32, 45, 46, 56, 71, 75, 77, 80, 81,
deep phreatic thermal explorer (DEPTHX), 263 83, 90–92, 94, 95, 109, 150–152, 165, 167, 168, 184, 188,
189, 253, 257, 259, 281, 295, 365, 367, 370, 372, 376, 378,
382, 392, 395, 408, 413
B Cavity/cavities, 8, 10, 14, 21, 31, 41, 52, 54, 55, 59, 60, 62, 64, 68, 80,
Blueholes, 370, 376, 383 82, 90, 92, 101, 102, 108, 128, 129, 149–151, 163–165,
Borehole deviation, 217–219, 240, 242, 243 167, 168, 175–179, 182, 184–190, 195, 210, 213, 215–218,
Borehole radar, 230, 247, 249 230, 231, 234–238, 247, 256, 260, 261, 267, 270, 280, 281,
Borehole video camera, 210, 219, 239, 256, 257, 344, 354, 357 325, 327, 367, 370, 372, 375, 376, 395, 397, 405
Breakdown Circular depression, 157, 342–343, 352, 362, 409
caves, 78, 80 Climate, 7, 11, 13, 27, 42, 47, 58, 126, 323
domes, 75–78 Collapse from great depths, 82–84
mines, 80 Conceptual model
rock, 81 final, 122, 308, 311–317, 368, 379–380, 383, 415
Breakthrough curve, 300, 301, 303, 345 geologic, 176, 287, 399
Bridge hydrologic, 281, 399–400
bridge alignment, 157, 311, 365–383 karst, 400
bridge failure, 253 preliminary, 103, 118, 120–122, 129, 131, 132, 145, 148, 158,
Brine wells, 61, 270 209, 312, 336, 338–339, 340, 343–344, 354–355, 362,
Bulking 389, 398–400
factor, 81, 82, 353–355, 358 Conduits, 3, 7, 8, 24, 29, 30, 36, 55, 57, 69, 71, 73, 78, 91, 167, 184,
measurements, 81, 353–355 185, 187–189, 216, 234, 236, 253, 256, 275–277, 279, 280,
Buried drums, 165, 172, 388, 401, 404–405 295, 305, 322, 327, 370, 395
Buried paleokarst, 54, 114, 190–191, 195, 411, 413 Confirming connections, 300–301

© Springer Science+Business Media Dordrecht 2016 417

R.C. Benson, L.B. Yuhr, Site Characterization in Karst and Pseudokarst Terraines,
DOI 10.1007/978-94-017-9924-9
418 Index

Contaminants rotary, 71, 186, 209, 211, 212, 214

organics, 167, 168, 223, 230, 280, 287, 289, 290 rotosonic, 209, 211, 212, 389, 390, 407, 408, 411
plume, 108, 167, 172, 204, 234, 243, 244, 280, 284, 292 washing borings, 51
Conversion of data, 122, 307–317 Dye trace/tracing
Core analysis of, 11, 300–306
recovery, 213–215, 373 charcoal bugs, 300, 344, 345
rock quality designation (RQD), 207, 214, 373, 377 considerations for, 295–300
Core team, 103, 120, 388–389 detection, 299–300, 305
Cover collapse sinkholes estimating quantity of dye, 297–298
conceptual models, 90–95 handling of dye, 298
scale models, 92 introduction, 295
Crosshole/hole-to-hole measurements, 103, 236, 240, 245–249 sampling and analysis methods, 163, 185, 299
Cultural factors, 125, 156, 357 water sampling, 163, 204, 263, 297, 299–301, 303, 344, 356, 357
Cultural features, 127, 129–131, 136, 146, 156, 168, 173, 359

E
D Edwards Aquifer, 46
Damaging impact, 49–64 Engineering measurements
Dams, 8, 10, 50–52, 55, 67, 99, 102, 146, 176, 178–180, 185, 221, installation of instrumentation, 210, 267
262, 268, 275, 300–301 EPA superfund site, 203, 327–329, 385–415
Data Epigenic karst, 8, 9
accurate, 109–118, 121, 122, 173, 220, 312, 415 Epikarst, 3, 27–29, 55, 68, 71, 90, 91, 94, 126, 175, 181–184, 203,
adequate 238, 243, 275–277, 279–281, 297, 321, 322, 327, 329
vs. depth, 233 Equivalent porous media (EPM), 285, 287
lateral, 113 Erosional features, 58
appropriate, 109–118 Evaporite karst, 60, 95
errors in, 118, 171, 310 Evaporites
temporal, 115, 133, 324 anhydrite, 7, 42, 60
Data mining, 131 gypsum, 3, 7, 10, 42, 51, 55–57, 60, 62, 78, 94, 126, 156, 169, 170,
Deep foundation piles, 253 179, 180, 198, 336
Desk Study, 121, 125–131, 134, 145, 148, 150, 153, 158, 193, 209, salt, 3, 7, 9–11, 13, 31, 42, 45, 55, 60–64, 70, 83, 92, 94, 126, 131,
312, 389, 392–398 163, 164, 167, 173, 230, 236, 256, 270, 296, 365, 370,
Detection dilemma 393, 405, 412
direct detection, 108–109 Excavations, 55, 69, 101, 147, 148, 154, 155, 181, 183, 192, 201, 214,
indirect detection, 109 222–224, 282, 328, 337, 339, 341–343, 346–348, 362,
probability of detection, 108, 109, 114, 303 386, 413, 414
statistical methods, 109, 320 Excessive grouting, 54
Development of karst, 3, 7–14 Existing boring data, 185, 228, 244, 368, 372–373, 399
Direct Push Expedited site characterization (ESC)
cone penetrometer testing (CPT), 201, 202, 204–208 strategy, 103, 120, 162, 185, 221, 388, 415
electrical conductivity (DPEC), 389, 401–404, 407–410 Exposed geologic cut, 222
percussion driven, 202–204 Expressway failure, 52–54
Disposal ponds, 385, 392, 412
Dissolution
non-soluble rocks, 10 F
rate of, 8 Final conceptual model, 122, 312–315, 379–380
Doline, 17, 19 Final interpretation of data, 310–315
Downhole and uphole measurements, 245–246 Fissures, 58–61, 336–347
Drilling Flooded mine, 263
angle borings, 56, 214–215, 218, 282, 344 Floridan aquifer, 46–47, 57, 71, 203, 286, 292, 302, 328, 329, 369,
an optimum approach, 175, 221–222 388–390, 392, 393, 399, 401–404, 406, 412, 413
borehole deviation, 217–219, 240, 242, 267 Flow
drilling plan, 209, 215 fracture, 283–285
drill rig access, 216–217 indications of, 284–285
fluid losses, 213, 220, 311, 372–374, 379, 380, 382, 383 Fluid loss, 372–373
ground truth and hard data, 220–221 Fly over, 158, 370
indications of karst, 102, 207, 213, 368, 372, 374, 380 Fracture
inside structures, 218–219 flow, 283–285
misinterpretation of data, 219–220 patterns, 82, 136, 137, 169, 179, 256, 283–285
rod drops, 207, 210, 213, 220, 311, 382 system, 71, 83, 91, 168, 277, 284, 344, 359
triggering a sinkhole, 67–73, 75, 156, 176, 328 trace, 135–137, 194, 339
Drilling methods zone, 83, 109, 114, 135, 180, 185, 186, 195, 215, 231, 239, 352,
auger, 205, 209–211, 219, 253, 389 356, 359, 372
dual tube reverse circulation, 209, 211 Fresh water lenses
jetting, 209, 210 Ghyben–Herberg, 10
percussion, 102, 201–204, 209–211 Further dissolution, 4, 9–10, 61, 75, 357, 382, 412
Index 419

G K
Geologic maps, 102, 110, 112, 127, 133, 134, 140, 146 Karren karst, 27
Geologic sampling Karst
chip trays, 214, 216 benefits to karst, 45–47
core recovery, 213–215, 373 development of karst, 3, 7–14
rock quality designation (RQD), 207, 214, 373, 377 features, 1, 3, 8, 14, 17–32, 35, 36, 38, 45, 46, 50, 58, 102, 126,
Geomorphology, 10–13, 30, 35, 102, 110, 132, 134, 148, 313, 336, 146, 148–149, 153–154, 185, 238, 243, 295, 321, 370, 376,
337, 370, 372, 376, 393 387, 392, 396, 405, 411
Geophysical logging karst maturity, 22, 35–39, 95, 127, 159
parameter measured, 167, 228–230, 232, 390
radius of measurement, 232
resolution and speed, 233 L
sequence of logging, 232, 233 Landfill, 50, 89, 102, 118, 131, 134, 167, 181, 183, 213, 222, 223,
tools, 237 234, 244, 279, 281–284, 288–291, 312, 333–362
Geophysical logging methods Large cavern, 30, 31, 61, 75–78, 81, 82, 90, 92, 93, 95, 128, 256,
acoustic televiewer (ATV), 146, 216, 221–222, 229, 230, 236, 327, 382, 383
237, 240 Larger open voids, 251–263, 381
alternate to core samples, 236–237 Laser, 140, 141, 251, 252, 259–261, 263
borehole conditions, 230, 232–233, 236 Light detection and ranging (LiDAR), 136, 140–142, 261, 266–268
caliper log, 230, 232–235 Limestone mines, 78, 80, 81, 87–89, 128, 135, 234, 256, 268, 281,
electrical/electromagnetic, 230 282, 298, 312, 333–362
fluid logs, 229, 230 Low-density zones, 236, 311, 374, 379, 381, 382, 408, 413, 414
gamma-gamma (density) log, 230–234, 236, 239, 243, 341, 378,
407, 408
mechanical logs, 229, 230 M
natural gamma log, 207, 230–234, 240, 341, 343 Managing data, 103, 105, 259, 307, 309
nuclear logs, 229, 230, 232–234 Man-made lakes, 55, 67, 68
optical televiewer (OTV), 210, 216, 222, 229, 230, 232, 237, 240 Mechanical erosion, 8, 10, 31
specialty logs, 227, 230 Mechanics of cavern breakdown, 78, 80
video and imaging logs, 230 Mine backfilling
Graphics compression tests, 360–361
selecting scales, 308 crushed rock, 88–90, 346, 357, 360, 361
Groundwater contamination, 23, 55–58, 287–293, 405, 412 fly-ash, 88, 89, 334, 341, 356, 359–361
Groundwater flow, 237–239 QA/QC program, 360–361
Groundwater monitoring, 131, 281–282, 334, 346, 350, 356, 357 Mine conditions, 268, 334, , 347, 349–357, 361
Groundwater resources, 46, 201, 275, 322 Mine map, 198, 258, 349
Groundwater vulnerability Mineral resources, 45, 47, 161
DRASTIC, 121, 321 Mine roof fail/failures/collapse
EPIK, 320–321 mechanism of collapse, 352–353
Florida Aquifer Vulnerability Assessment (FAVA), 321 monitoring of, 281, 357–359
Groundwater withdrawl, 197, 205 types of collapse, 350–353
Gypstacks, 56, 94 Minimally invasive methods
Gypsum, 3, 6, 7, 10, 51, 55–57, 60, 64, 78, 94, 126, 156, 170, 179, cone penetrometer testing (CPT), 201, 204–208
180, 198, 336 direct push, 122, 201–207, 209, 389, 390
Mining wastes, 55–57
Mixing dissolution, 9–10
H Modes of breakdown
Horizontal borings and tunneling, 59–60 cantilever beam, 78, 79
Housing development, 50, 52, 55, 72, 156, 176, 410 fixed beam, 78, 79
Hushpuckney Shale, 282, 334–337, 350, 355, 356, 358 Monitoring well audit, 244–245, 279
Hydraulic gradients, 12, 55, 68, 277 Multiple porosity system
Hydrology primary, 8, 14, 276, 393
characterization, 315 secondary, 8, 27, 30, 276, 277, 282
measurements, 391 tertiary, 8, 11, 14, 71, 276
HydroPhysical logging, 237, 239, 241
Hypogenic karst, 8, 9
N
Near surface indictors (NSI), 30, 109, 168, 175, 177, 179, 372
I New sinkholes (NSH), 19, 23, 35, 36, 54, 68, 109, 129, 260, 268, 320,
Inclinometers, 81, 240, 266–272 325, 326, 397, 400
Integration of data, 120, 122, 175, 310–312, 313 Nitrates, 26, 47, 57, 187, 302–303
Interferometric Synthetic Aperature Radar (InSAR),
140–142, 266–268
Invasive measurements, 110, 111, 119, 161, 162, 201, 389, 390 O
Inventory of karst features, 133, 148, 153–154, 296 Observational method, 101, 103, 118
Issue of scale, 27–28, 112 Observational skills, 101, 146, 148, 251
420 Index

Observations definitions of, 322

geologic, 148–150 EPA Superfund site, 327–329
hydrologic, 150–152 objective and subjective methods, 322, 324
Ocean water intake, 253 regional assessment, 57, 411
Oil and gas fields, 47, 140 risk factor, 320, 321, 323–324
Optical time domain (OTD), 279–280 scoring system, 325
Other soluble rocks, 7, 10, 41 sinkhole probability map, 321, 397
Oversight committee, 388, 406 site-specific assessment, 322–329
subjective risk predictions, 324
Road widening, 185, 254–256
P Rockhead, 27, 28, 36, 126, 181, 209
Paleocollapse Rock thicknesss, 78, 79, 81
fractures, 344–346, 356 ROD. See Record of decision (ROD)
Paleokarst, 3–4, 13, 41, 42, 47, 54, 101, 128, 137, 154, 158, 175, Rose diagrams, 237, 284, 285
190–195, 203, 205, 213, 224, 281, 322, 337, 338, 367–368, ROVs. See Remotely operated vehicles (ROVs)
376, 379, 383, 393, 410, 411 Rubble zone/rock, 80, 336–337, 350–352
Paleosinkhole, 12, 19, 47, 114, 191, 192, 194, 224, 256, 367, 370, 379,
393, 395, 400, 411
Perched water, 281 S
Permeameter, 292 Salt mine, 63–64, 83, 92, 94
Petroleum industry activities, 60–62 Saltwater interface, 10, 13, 31
Phased strategy Saturated zone, 27, 228, 229, 232, 234, 275, 277, 280, 282–287
confirmation, 368, 374–376, 380 Scale
detailed data, 134 impact of, 112–113, 205
reconnaissance, 370–371 observations and measurements, 87–89, 103, 110, 112, 120
Photographic or video documentation, 222, 251, 252, 256–259, 360 Sea levels (changes), 11, 12, 45, 308, 376, 377, 380, 393, 408
Physical properties of limestone Seepage at earthen dam, 176, 178–180
porosity, 13, 14 Seismic crosshole tomography, 248
strength, 13, 14 Seismic velocity measurements
Pitfalls hole-to-hole, 103, 217–218, 240, 246–248, 267
impact of computers, 105 vertical seismic profiling (VSP), 246, 408
interdisciplinary approach, 105 Semi-confining layer (SCL), 28, 281, 292, 388–390, 393, 397,
non-technical, 104, 105 399–405, 407–410, 412–414
Plume boundaries, 243, 287 Shear deformation, 272
Pond boundaries, 403–404, 412 Sinkhole lake, 17, 20, 23, 54–55, 129, 190, 370–371, 379, 393, 395
Preliminary conceptual model, 103, 118, 120, 121, 128–129, 131, 132, Sinkhole risk assessment, 203, 397, 412–413
145, 148, 158, 209, 312, 336, 338–340, 343–344, 354–355, Sinkhole risk maps, 23, 25, 129, 320–321
362, 389, 398–400 Sinkholes
Presentation of data, 170–171, 232, 233, 310, 315–317 databases, 23, 129, 159, 320, 392, 397, 399
Processing of data, 170–171, 232, 233, 310 densities/distribution, 22–23, 129, 321, 326, 392
Pseudokarst linear trends, 22–23, 109, 129, 365, 395
analogous karst, 4 maps, 23, 126
false karst, 4 range of sizes/size of a sinkhole, 18–22, 322
Pseudokarst impacts, 58–64 spatial proximity, 326
Pump/pumping tests, 8, 42, 60, 61, 63, 67, 68, 70, 89, 112, 151, 156, speed of a sinkhole, 71–73
191, 210, 221, 237, 244, 258, 259, 270, 277–279, 282–287, susceptibility maps, 23, 129
298, 299, 310, 328, 345, 346, 355, 360, 390, 393, 397, 401, types of sinkholes, 17, 18, 322, 397
403, 406, 413 Sinkhole throat, 177
Sinking streams, 3, 23–26, 46, 57, 150, 275, 296, 302
Site characterization
R defining the problem, 99
Reconstructing geological data, 243–244 geologic conditions, 101, 102, 104
Record of decision (ROD), 386–387, 412 interdisciplinary approach, 105
Reference grid, 120, 389 karst conditions, 101, 102
Regional key steps in, 118–122
geology, 283, 341, 376, 392 pitfalls of, 104–106
geomorphology, 337, 370, 376, 393–395 sequence of work, 110–112
hydrology, 392 strategies for, 101–103
subsidence, 58–60, 268 survey grid, 113, 120, 171–172
Release of acids, 55, 59 team, 99, 102–104
Remediation of soils, 413 uncertainties, 100
Remotely operated vehicles (ROVs), 252, 261–263 Site coverage, 114, 118, 148, 155, 175, 205, 389, 404
Remote sensing data, 133, 142 Site preparation, 389–392
Representative elementary volume (REV), 112, 267, 277 Site visit, 102, 367, 380
Risk assessment Site walkover, 121, 127–129, 145–159
along a section of highway, 324–325 Sketches, 101, 102, 145, 147, 222–223, 251, 252, 310, 315, 341
Index 421

Slug tests, 112, 277–279, 283, 390, 401 T

Soil piping, 55, 58, 67, 71, 77, 154, 168, 175–180 TDR. See Time domain reflectometry (TDR)
Sonar, 163, 173, 177, 251, 252, 260–263, 300, 390, 410 Tectonic uplift, 12
Sources of mine water, 282 Telescoping benchmarks (TBMs), 268
Spatial correlation, 171, 311, 312, 372–374, 380, 381 Temporal aspects, 278–279
Spectrofluorophotometer, 297, 300, 302, 303 Thermal imaging/thermal imagery, 140, 152, 161, 165, 169, 170,
Springs 173, 280
submarine springs, 24, 25, 189, 302 Thermography, 140, 141
submerged spring flow, 280 Tilt meters, 266, 270–272, 324
Standard penetration tests (SPT), 194, 205, 207, 210, 213, 220, 372, Time domain reflectometry (TDR), 266–267, 271–272
379, 382, 390 Tomographic measurements, 245–249
Strain gauges, 267, 272 Top of rock, 3, 20, 27, 28, 69, 71, 90–92, 114, 115, 126, 129, 162,
Strategy, 87, 100–103, 107–123, 195, 216, 221, 268, 295, 307, 317, 164, 169, 175, 181–184, 209, 216–217, 219–220, 223, 224,
322–324, 331, 357, 368, 388, 389, 391 309, 310, 314, 327, 371, 397, 400, 409–410
Stratigraphy, 30, 110, 121, 126, 129, 164, 201–202, 230, 234, 235, 392 Top of unweathered rock, 181, 182, 184, 219
Structural impacts, 49–55, 148 Topographic maps, 23, 24, 127, 128, 135, 145–147, 150, 153, 154,
Subsidence 159, 176, 283–285, 313, 339, 358, 370, 392
examples of, 64, 272, 358–359 Topography, 113–114, 120, 129, 134, 135, 140–142, 145, 146, 175,
measurements, 346–349, 358 181, 190, 310, 325, 339–340, 371, 386, 392
monitoring, 267–272, 350 Travel times, 230, 248, 261, 271, 279, 296, 301, 303, 305, 306
monument design, 347, 348 Trenching/trenches
reference datum and benchmarks, 268–270 very large, 223–224
regional, 58, 60, 268 Triggering mechanisms
risk assessment, 292, 336, 357–359 dewatering, 67, 68, 72, 83
site specific, 268–272 drilling operations, 70–71
Subsidence and sinkhole activity, 3, 19, 35 guideline to minimize, 70
Surface geophysical methods leaky water pipes and sewers, 69
echo sounder, 177, 189, 190, 194 pumping, 67, 68
electromagnetics, 174 surface water run-off, 67–70
ground penetrating radar (radar), 162, 164, 170, 173, 174, 176, vibrations, blasting and earthquakes, 70, 71
178, 196, 197 water levels, 67, 68
magnetometer, 172 water related, 67–69
microgravity Type of data, 109, 120, 127–131, 133, 202, 280, 310
analysis of, 406
low gravity values, 406–408
multi-frequency analysis of surface waves (MASW), 164, 170 U
multi-frequency electromagnetics (MFEM), 174 Underground mines, 4, 62–63, 71, 197, 281
radiometric, 161, 165 Unmanned aerial systems (UAS), 137–139, 158
resistivity, 164, 165, 167–170, 173, 174, 181, 182, 185, 189–191 Unsaturated bedrock, 280–282
seismic reflection, 163, 164, 173, 174, 181, 194 Unsaturated zone, 150–151, 168, 229, 232, 277, 280–282
seismic refraction, 162, 164, 167, 170, 173, 174, 181, 184, 194, 196
spontaneous potential (SP), 164, 173, 188
time domain electromagnetics (TDEM), 164, 185, 186 V
Surface geophysics Visual inspection, 56, 131, 178, 205, 251–256, 258, 350
brief overview, 162–163 Void space, 32, 60, 68, 71, 73, 75–76, 78, 81, 82, 87–95, 169, 176,
direct and indirect detection, 168 197, 240, 257–258, 260, 261, 322, 324, 326–327,
ease of use, 173 353, 361
geophysical anomalies, 164–165, 167–168, 173 Vulnerability of cave habitat, 301
guidelines for the selection, 162, 173–175
parameter measured, 164–165
penetration of measurements, 168 W
processing of data, 170–171 Wakulla Spring system, 26, 57, 58, 187, 188, 253, 301–304
resolution (lateral/vertical), 169–170 Wastewater sprayfield, 303
survey grid, 171–172 Water and sewer lines, 4, 58, 59, 69, 397
use over water, 164–165, 173 Water level data, 279, 284, 403
utilities, buried drums, tanks and trash, 172 Water quality
Surface water, 8, 9, 23, 26, 49, 50, 55, 58, 67–70, 80, 83, 126–128, specific conductance, 126, 284–286, 289–290, 356
135, 154, 275, 278–280, 298, 300, 323, 324, 340, 357–359, Winter Park Sinkhole, 19, 22, 49, 73, 91–94
382, 397, 413, 414 Woodville Karst Plain (WKP), 57–58, 185, 187–189, 234, 236, 238,
Surface water management, 50 239, 253, 301–304
Swelling soils, 58 Work plan, 103, 119–120, 122, 131, 132, 158–159, 388