Petroleum Geology & Geophysics
Actual Trace Blocked Impedance
Wavelet
Blocked
Velocity
*
-1
Blocked
Density
FWSchroeder 1
Objectives
• Comparison between zero phase and impedance data
• Our ultimate goal
• Relative impedance
• Absolute impedance
• Low frequency component
• Basic inversion process
FWSchroeder 2
A Quick Review
Zero Phase Relative Impedance
Trace Trace
Marine Shale
A
Shoreface Sand
Marine Shale
B
Fluvial Sand
Shoreface Sand
Marine Shale
Zero Phase – used to identify and map major changes in impedance
Relative Impedance – used to quantify properties above/below major
impedance boundaries
FWSchroeder 3
Forward Seismic Modeling
The Convolution Model
Blocked Logs Refl. Coef. Wavelet Modeled Trace
• Input is a sonic log and a density log
• It is performed at well locations
• It allows us to tie well markers to the seismic trace
• Well data can be modified to understand the seismic
signature of lateral variations (e.g., what if the
reservoir was twice as thick)
FWSchroeder 4
Inverse Seismic Modeling
Seismic Inversion
Actual Trace Wavelet Blocked Impedance
Blocked
Velocity
*
-1
Blocked
Density
• Input is any seismic trace
• It is performed away from well locations
• Well data can be used to calibrate the results
• It can help us understand what the seismic is telling us
about rock properties (velocity and density)
• Software can generate multiple realizations of the
velocity and density, range of possibilities
FWSchroeder 5
The Ultimate Goal
• Ideally, we would like to turn each seismic trace into
an impedance trace or, going even further, into a
velocity and density pseudo-log
• That would allow us to predict rock properties in an
interval of interest rather than acoustic properties at
a major rock boundary
• Variations in velocity and density at each seismic trace
would be better indicators of lateral changes in
rock properties, such as thickness, net:gross or
porosity
FWSchroeder 6
Definition
Seismic inversion is the process that starts with a
reflection amplitude trace, generates a model
for the impedance structure that produced that
trace and then transforms the impedance into
estimates of Vp, Vs and density
It is an inverse of the convolution method that
we use to model a seismic trace from a sonic
log and a density log
FWSchroeder 7
Relative Impedance
• Since the input zero phase trace has a limited frequency
range, the derived relative impedance is also band-limited
• It is relatively easy to derive from a zero phase trace
by two methods:
– Integrating the zero-phase trace (a running sum)
– Rotating the phase of the zero-phase trace by -90°
Zero Phase Relative Impedance
Trace Trace
Trace Integration
Phase Rotation
FWSchroeder 8
The Major Difference
Impedance Log Relative Impedance Units
-150 -100 -50 0 50 100 150 200 250
650
700
Two-Way Time (ms)
750
Relative Impedance
Two-Way Time (ms)
800 lacks the low frequency
trends and the high
850
frequency oscillations
900
950
1000
1050
Based on Chopra and
Sharma, 2012
FWSchroeder 9
Limitation of Relative Impedance
• Due to bandwidth limitations, relative acoustic
impedance from seismic provides only an
approximation of acoustic impedance as seen on logs
• If we go to the extra trouble to restore the missing low
seismic frequencies (< 10 Hz) we can calculate
absolute impedance, which provides a better fit to log
data
FWSchroeder 10
Frequency Content
Impedance Log Zero Phase Relative Impedance
Trace Trace
A seismic trace has Relative impedance will
a limited range of have the same limited
Sonic and density logs
frequencies, about range of frequencies as
have a wide range of
10 to 50 Hz the trace it was
frequencies, from 0 to
derived from
200+ Hz
Two-Way Time (ms)
FWSchroeder 11
Two Types of Impedance Data
1. Relative Acoustic Impedance
• A simple mathematical integration of zero phase seismic
reflectivity data or a -90° phase rotation
• Indicates relative changes (+/-) in seismic impedance
• In some cases can be calibrated locally to an average
physical property of a reservoir unit
• Not useful for accurately measuring vertical impedance
changes over large intervals
FWSchroeder 12
Two Types of Impedance Data
2. Absolute Acoustic Impedance
• Begins with relative impedance, which was described on
the previous slide
• Restores low frequency component of seismic data using:
– Well data from the project area, and/or
– Seismic derived stacking velocities (from processing)
• Precise calibration of impedance measurements is possible
for multiple reservoirs at multiple well locations
• Frequently it is used to build input for a detailed
reservoir model (for reservoir simulation)
FWSchroeder 13
Low Frequencies from Stacking Velocities
Velocity
• The hotter colors indicate the
velocities that result in a good
flattening of gathers
• The blue “Xs” are the values
that the velocity interpreter
selected
Two-Way Time
Vint • The red block curve show the
computer interval velocities
The gradual increase in
velocity with depth is what we
use to get the low frequency
trends in impedance
FWSchroeder 14
Frequency Spectra
Amplitude LOW FREQUENCY COMPONENT
100
Seismic
Bandwidth
50
0
0 10 20 30 40 50 60 70 80
Frequency
FWSchroeder 15
Acoustic Impedance Data
• Estimates impedance/sonic/density log responses from
migrated and stacked seismic traces
• Is an interval attribute, as opposed to an interface
property
• Can better resolve beds that are thinner than tuning
thickness than with zero phase data
• Useful as a lithology or porosity indicator
• In some cases reveal HC/water contacts better than with
zero phase data
• Commonly is used for pore pressure prediction
FWSchroeder 16
Impedance Data: Value
• It represents the subsurface better, especially when low
frequency information is included:
– Low frequency geologic variations
– Increased vertical resolution
• It may yield better estimates of reservoir thickness
• It can be transformed to rock properties, such as net
pay, porosity, hydrocarbon pore thickness
• It allows us to perform enhanced seismic attribute
analysis
• Increasingly it is being used by interpreters to delineate
reservoir rocks and surrounding seals
FWSchroeder 17
Assumptions for Post Stack Inversion
• The input seismic data has been imaged properly and is
zero phase
• Seismic amplitudes are proportional to reflection
coefficients (controlled amplitude and phase seismic
processing)
• The input seismic data represents normal incidence
(zero-offset) traces (AVO effects have not influenced
the migrated/stacked data)
• Input data is free of multiples and coherent noise
• An appropriate wavelet has been extracted and it does
not change with time
We use a near offset
stack as input if the S:N
ratio is adequate
FWSchroeder 18
Inversion Process
• Zero Phase • Sonic, Density Logs
Seismic • Full, Partial Stacks Well • Markers
• •
Data •
Stacking Velocities
Horizons Data •
Directional Survey
Check Shots
Horizons Filtered
Logs
Stacking
Velocities
Low
Frequency
Wavelet Zero
Component
Phase
P-Impedance
Lithology
INVERSION
Density
Porosity
S-Impedance
FWSchroeder 19
Example 1
A zero phase seismic line that shows a
gentle structure with an amplitude Horizon 1
anomaly that is caused by a gas
field.
Horizon 2
Horizon 3
Modified from Russell and
Hampson, 2006
FWSchroeder 20
Example 1
Horizon 1
A zoomed portion of the zero
phase seismic line over a
known gas field.
Horizon 2
Horizon 3
An inversion of the seismic
Horizon 1
line based on a recursive
type of inversion. Note
how the gas field stands
Horizon 2
out.
Horizon 3
Modified from Russell and
Hampson, 2006
FWSchroeder 21
Example 2
Well 1 Well 2
50 ms
A zero phase seismic line that intersects two wells. The P-impedance
curves are shown in black.
50 ms
Modified from Misra
and Chopra, 2011
1 km
A P-impedance section. This was obtained with a probabilistic neural
network.
FWSchroeder 22
Example 3
50 ms
An estimated P-wave velocity section obtained by a PNN analysis.
The black curves are P-wave sonic logs.
50 ms
Modified from Misra
and Chopra, 2011
1 km
An estimated density section obtained by a PNN analysis. The black
curves are density logs.
FWSchroeder 23
Summary
• Provides a log estimate at every single seismic trace
• Intervals are easier to interpret geologically than the
interface properties that zero phase data measures
• Estimated acoustic impedance can be compared to well-
derived acoustic impedance to calibrate the results
• Acoustic impedance can help us estimate lithology,
porosity, fluid content, and pore pressure anomalies
• The inversion process attempts to remove the band-
limiting effect of the seismic wavelet, hopefully
providing increased geologic detail
• Geo-statistical data derived from well control can be
combined with seismic impedance values to produce
acoustic impedance volumes with improved vertical
resolution
FWSchroeder 24
References
Slide 9: Chopra, S. and Sharma, R., 2012, A Solid Step towards Accurate Interpretations, AAPG Explorer,
December, p. 36-37.
Slide 19-20: Russell, B. and Hampson, D, 2006, The Old and the New in Seismic Inversion, CSEG Recorder,
V. 31, N. 10, December.
Slide 21-22: Misra, S. and Chopra, S., 2011l, B. and Hampson, D, 2006, Neural Network Analysis and
Impedance Inversion – Case Study, CSEG Recorder, V. 36, April, p. 36-39.
FWSchroeder 25
FWSchroeder 26
