Report No.

FAA-RD-77-36

WIND SHEAR MODELING FOR

AIRCRAFT HAZARD DEFINITION

Walter Frost
Dennis W.Camp

1977
Interim Report

~~~~March

>- "

i1

i
4.

-..

DDO

'"-'--

U'I-

~Document

is available to the U.S. public through

( _
C:)

the National Technical Information Service,

Springfield, Virginia 22161.

Prepared for

.2

C ..

U.S. DEPARTMENT OF TRANSPORTATION

FEDERAL AVIATION ADMINISTRATION
Systems Research & Development Service
Washington, D.C. 20590
q4

NOTICE
This document is disseminated under the sponsorship of the
Department of Transportation in the interest of information
exchange. The United States Government assumes no liability
for its contents or use thereof.

Tochnicul Report Documentetgie

WiM

No.

4.Title end

WPrformin

2. ovenme

nt Accesionr Me.

Pe"

3. Recip.ernt's Catalog No.

wtie

Orgotetsen Mote

Adres

-".

Work Uit N.. (TRAIS)

National Aeronautics & Space A~n

Atmospheric Sciences Division
Spab
Sciences
."P p s O
gan,
~ , lonmeLaboratory
and dr-s s

1)

Assoc.,
Tu.1ah1a,
3-

Inc.

154-451-014A
tic CN
d

f
10.' Wrk tMe.

Marshall Space Flight Center, AL 35812

for

__ _7620

U. S. Department of Transportation

Apr t

managed by the Aviation Weather Systems Brach, ARD-50.

(
RASs-32217)
to FG Assoc., TuSlahomia,
Apr
37388

Contracted by MSFC

t-

Mardi 377

6. AbotDcc

Thsdocu.ment presents a discussion of the various types of wind shear

which cause problens to aircraft operations. The presentation is
limited to lw-level wind shear sch as culd be enountered in the
vicinity of airports by aircraft on approach to landing and on takeoff.
The types of shear discussed are primarily limited to frontal,
thundersto m and shears associated with stable and neutral Conundary
layers. The discussions ehasize otodeling of the wind shear for
simulation purposes.
Windi shear is discussed not only as a change in absolute wind speed
but also as a result of wind direction change. Sxme insight is
given relative to a crtbinatfion of wind speed and d ion
change.
It is to be noted that the crndition affecting aircraft operations a~s
not one shear parameter but is a orbination of several, e.g.
hrizontal s hear,vertical shear, wind direction change and
height of shear above ground leel.o

17.

Key W

p,

p ose s,.tr~huron

Storemeut

Wind shear
Thunderstorm Gust Fronts
Frontal Wind Shshe

Doisment is available to the U.S. public

through the National Technical
Infortion Service, Springield,
Virginia

19. Security Clsl.

(of this report)

Unclassified
Form DOT F 1700.7

(8-72)

22151.

20.Security Clessif. (of this peg

21. No. of Paes

Unclassified
Reproduction of completed page authorzed

"r

22.

Price

PREFACE
The authors are grateful for the many people who were
helpful in the preparation of this report. A complete
acknowledgment list would be too lengthy; however, there are
some individuals we feel we must recognize. These are Messrs.
Frank Coons, Guice Tinsley, and Frank Melewicz of the FAA
for many helpful discussions and especially for their patience
in critiquing the report; Messrs. R. Craig Goff* and Jean
Lee of NSSL for discussions and providing thunderstorm data;
Mr. Hubert McCaleb and other members of NTSB for providing
needed information and reports; Mr. Paul Jernigan of Douglas
Aircraft for discussion on how data are used in aircraft
simulators; Messrs. Mike Kennan and Dean Babock of SRI for
many helpful discussions and comments; Drs. John McCarthy and
Edward Blick of the University of Oklahoma for discussions
and providing data; and we especially wish to recognize
Dr. George Fichtl without whose help and assistance the report
would not have been possible. The authors are indebted to
Mrs. Judy Wright for typing the report.
*Mr. R. Craig Goff's present affiliation is FAA/NAFEC.

A/

iv

5 ;El
--

0 _
-I-

I-

, 0..

Am

LU

CC

3.0

E E

ot

:E...

I-c

TABLE OF CONTENTS
SECTION

1.0

PAGE

INFLUENCE OF WIND SHEAR ON AIRCRAFT OPERATIONS. . .

1.1

1.2
1.3

Introduction. .
. .1
Problems of Wind Shear in Aircraft Operations
Effects of Wind Shear on Aircraft Flight. . .

2.0

TYPES OF WIND SHEAR . .

3.0

FRONTAL WIND SHEAR. .

4.0

THUNDERSTORM GUST FRONTS. .

4.1
4.2
4.3

4.4

5.0

5.4

6.0

. ..

..

. 19

. 19
. 19
. 24
. . .
. 24
. . . . . 28
. . . .. . 28
. . . . . 30
. . . . . 35
. . . . . 36
40
. . . ..
41
. . . . . 41
.
.
.

.
.
.

.
.
.

.
.

. 53

Introduction . . . .. . .
.
Fronts . . . . . . . . . . .

.
.

.
.

.
.

.
.

.
.

.
. .

vi

13

Introduction ...............
.
Mean Flow in the Lower Surface Layers . . . .
Turning Layer . . . . . . . .
....
.
5.3.1
Introduction .
. . .
..
...
.
5.3.2
Barotropic Turning Layer ......
5.3.3
Baroclinic Turning Layer. . . . . . .
Boundary Layer Over Non-Homogeneous Terrain .
5.4.1
Surface Roughness Transition. . . . .
5.4.2
Other Surface Variations. . . . . . .

SUMMARY . .
6.1
6.2

0 .

11

Introduction . . . . . . . * * . * .
Storm Cells . . . . . . . . . .. . .
Gust Fronts . . . . . . . . . . . .
4.3.1
Gust Front Types.
4.3.2
Gust FiDnt Speed. . . . . .
4.3.3
Gust Front Characteristics.
4.3.4
Gust Front Shape. . . . . .
4.3.5
Gust Characteristics . ..
.
4.3.6
Wind Speed Fields . . . . .
4.3.7
Wind Shear..
..... . . .
Wind Shear Prediction Techniques
4.4.1
Raw Data ........
. . .
4.4.2
Mathemat4cal Models . . . .

BOUNDARY LAYER OVER FLAT TERRAIN. .

5.1
5.2
5.3

43

53
53
64
64
65
73
81
81
95

. 96
. 96
. 96

6.3
6.4
6.5
6.6
6.7

Thunderstorm Gust Front. . . . . . . . .

98
Atmospheric Boundary Layer . . . . . ..
103
Turbulence Models
...
........
107
Flight Simulator Applications . . . . . . 109
Conclusion . . . . . . . . . . . . . . . 111

APPENDIX A . . . . . . . . . . . . . . . . . . . . .

112

REFERENCES . . . . . . . . . . . . . . . . . . . . . . 133

vii

LIST OF ILLUSTRATIONS
PAGE

FIGURE
1 1. Flight Recorder Traces of Two Different Aircraft

Passing Through Unexpected Wind Shear ......

1-2

Approach in Tailwind with Shear .

1-3

Approach in Headwind with Shear ...........

3-1

Cases of Directional Wind Shear Associated with a

Front . . . . . . . . . . . . . . . . . . . . . .

3-2

14

Estimated Position of Cold Front Relative to the

Distance of the Front from the Airport. .

16

3-3

Approximate Conditions of Wind Speed and Temperature

Difference which will Induce Turbulence along with
. . .
.. .. .. 17
............
Wind Shear ....

4-1

Typical Features of a Thunderstorm . .

4-2

Three-Dimensional Motions of Air Parcels in a Thunderstorm Cell. .

...

.....

. .

20

22

4-3

Contours of Vertical Velocity Relative to the Storm

Motion . . . . . . . . . . . . . . . . . . . . . .

4-4

Hodograph of Horizontal Velocity in the Thunderstorm

Cell . . . . . . . . . . . . . . . . . . . . . . . 25

4-5

Common Features of a Gust Front. . . . . .

26

4-6

Streamline Pattern Associated with a Gust Front.

31

4-7

Nonlinear Frontal Slope as Described by Colmer . .

33

4-8

Altitude at which an Aircraft Approaching an Airport

Might Expect to Encounter a Gust Front which
Passed the Airport Earlier. .

4-9

. .

...

23

34

Height.-Time Plots of Gust Front Velocity Fields.

38

4-10 Height-Distance Plots of Gust Front Speed Fields

39

45

4-11 Nondimensional

Wind Speed Contours Displaying

Similarity Proposed by Sinclair et al . .

viii

4-12

Vertical Wind Speed Along a 3* Glide Slope Through

Gust Fronts as Correlated by Fichtl and Camp..

4-13

51

Wind Speed from Combined Tower and Rawinsonde Data

Near the Time of Maximum Wind Speed During the

Passage of a Gust Front ..

52

55

5-1

Surface Roughness Length (Z.), cm .....

5-2

Nondimensi'onal Wind Shear as a Function of Height,

Z, over Monin-Obukhov Stability Length, L'.

5-3

Height Scale for the Planetary Boundary Layer...

5-4

Hodograph of Dimensionless Velocity for the Neutral

Barotropic Turning Layer .........

5-5

68

Departure of the Nondimensional Wind Component

Parallel to the Surface Wind from the Wind at

z=0.15 . . . . . . . . . . . . . . . . . . . . .

5-6

........

..

5-7

Experimental Values of Al(p)

5-8

*Experimental Values of B()

5-9

Calculated Wind Spiral for. a Stable Atmosphere,

5-10

Comparison of the Log-Linear Velocity Profile with

the Wind Component Paralled to the Surface Wiad

. .

........

5-13

74

75
76

L=llm . . . . . . . . . . . . . .

as Determined from Figure 5-9 . .

5-12

71

Departure of the Nondimensional Wind Component

Perpendicular to the Surface Wind from the Wind
at z=0.15 ..

5-1

59

..

77

78

Numerically Computed 3aroclinic Wind Hodographs for

Four Different Directions of the Thermal Wind .

79

80

Schematic Illustration of the Developing Internal

Boundary

Layer .

. .

84

85

Computed Growth of the Internal Boundary Layer

Relative to a 30 Glide Slope For Some Typical
Roughness Changes Near Airports. . .

ix

5-14

Typical Horizontal Velocity Profiles at Given

X Stations Downwind of a Change in Surface
Roughness. . . . . . .

5-15

. ...

Velocity Variation Over Patches of Trees. .

6-1

Vertical Velocity Contour Given By Goff

Compared With Tabulated Values For Computer
Look-Up Grid System . .

. .

94

99

Computerized Flight Path of Aircraft Character.

. 101

Wind "Seen" by DC 8 Landing With Fixed Controls

In Case 9 Thunderstorm .

3o

87

.. *

..

istic of DC 8 Through Case 9 Thunderstorm. .

-3

. .93

5-17

6-3

86

Computed Wind Profiles At Various Stations

Downstream of A Discontinuity In Surface
Roughness. .

6-2

Characteristic Vertical Velocity Field Over A

Change in Surface Roughness. .

5-16

..

. . . .

...

... .

. 105

. . .

. 106

74ith Time of Day . .

. 108

6-4

Longitudinal Wind Component . .

6-5

Lateral W -1 Component. . .

6-6

Variation of

Ix

II

II

. 102

LIST OF TABLES
TABLE
1-i

Page
Representative Winds Along Flight Path of DC
10-30 Airliner Which Crashed at Logan Airport, December 19, 1973......
. . . .

1-2

Data From Tower Log of A Major Airport. . ..

1-3

Winds Over JFK on January 4, 1971 . . . . . . .

4-1

Gust Front Speeds of 17 Different Storms.

4-2a

Maximum Absolute Shear (m s -1 100 m -

4-2b

Average Maximum Absolute Shear (m s "

4-3

Typical Values of Cold Aij Outflow Parameters

(Based on NOAA/NSSL Data). . . . . . . . . . 50

4-4

Worst Case Values of Cold Air Outflow Parameters (Based on Eastern 66 and NOAA/NSSL

. 29

42

100m -1 ) . 42

Data . . . . . . . . . . . . . . ... . . . . 50

5-1

Values of A(h/L) and B(h/L) Under Neutral

Conditions h/L = 0 . . . . . . . . . . . . . 67

5-2

Comparison of Modified Logan and Fichtl

Solution With Bradley's Data . . .

. 92

. 97

6-1

Summary Of Literature Survey . . . . . .

6-2

Longitudinal and Lateral Wind Components As A

Function of Dimensionless Height Z And
Stability p. . . . .*. . . . .. . . . . . .104

xi

LIST OF ABBREVIATIONS AND SYMBOLS

A(h/L)

Empirically determined stability functions

B(h/L)

Empirically determined stability functions

Gust front speed

Cp

Specific heat

Coriolis parameter

Heat flux

Height scale

Von Karmans constant

Monin-Obukhov stability length

L'

Alternate scaling length (see Equation 5.6)

Average slope of gust front

Ri

Richardson's number

Ro

Rossby number

Time

Temperature

uc

Cold air velocity component

uwWarm

air velocity component

U,

Friction velocity

Wg

Geostrophic wind

Wx

Wind speed in horizontal direction

Wy

Wind speed in lateral direction

Wind speed in vertical direction

xii

Horizontal distance

Lateral distance

Vertical distance

zo

Surface roughness

Atmospheric boundary layer thickness

/ P

Density

xiii

1.0

1.1

INFLUENCE OF WIND SHEAR ON AIRCRAFT OPERATIONS

Introduction
Wind is an important consideration in the analysis

of airplane flight in the atmospheric boundary layer, both

because of short scale gusts or turbulence and because of large
scale variations of the mean wind.

In the planetary

boundary layer, the mean wind decays toward the ground

and has considerable horizontal variations due to
irregularities in terrain.

Thus, both spatial and

temporal variations occur in mean surface winds encountered

along ascending and descending flight paths.
Many analyses of airplane motion have been carried
out (1-1, 2, 3) which, in general, consider only constant
winds and thus neglect effect of wind shear.

Recent

reports of commercial airline accidents, however, have

identified wind shear in the planetary boundary layer
as being a significant factor in the cause of these
accidents, illustrating that variable winds are an
import&

design consideration.

This report begins with a review of the reported

influ' nce of wind shear on aircraft accidents.

Recent

major accidents where wind shear was a primary factor are

described.

State of the art mathematical models, tabulated data of mean

wind profiles and fields which can be generally used in
computer simulation of aircraft flight through the wind
shears associated with these meteorological conditions are
presented.
1.2

Problems of Wind Shear in Aircraft Operations

Accidents caused by encountering strong wind shear,

turbulence and gusts during terminal flight operations are

becoming increasingly more evident as improved documentation
Qf these conditions is achieved through advanced instrument
technology, such as digital flight data recorders.

Reference

[1-41 reports that weather conditions were significant

factors in air carrier accidents within the last two years
and Wyatt 11-5] reports weather to be a factor in approximately 39% of all fatal accidents which occurred in general

15

aviation between 1964 to 1973.

To what degree these were due to wind effects is uncertain.

However, many missed approaches which have simply

been classified as pilot error may, in the light of today's

knowledge, be traceable to wind shear.
Recently, a number of well-defined incidents where wind
shear has been a major factor contributing to serious
accidents have been documented.

Laynor [1-61 reports that

an Iberian Airlines' DC 10-30 crashed on December 17, 1973,

about 500 feet short of the runway while making an Instrument
Landing System (ILS) approach to Logan International

Airport, Boston, Massachusetts.

The meteorological data indicated that the winds aloft
at altitudes above 1000 feet in the Boston area were generally
from the south at speeds approximately 40 kts.

The surface

wind measured at Logan Airport was from a westerly direction

at 9 kts.

These data suggest that wind shear existed at the

time of the accident.

Examination of the data recorded on the aircraft digital
flight data recorder generally verified that the aircraft
penetrated an altitude band where a sudden change in wind
direction and speed occurred.

The results of reconstructing

the approach in a Douglas DC-10 simulator disclosed that a

wind shear, characterized by a diminishing tailwind component,

encountered at low altitudes during the period of approach

where the pilot is transitioning from automatic to manual
flight, produced a situation in which the aircraft descended
below the glide path.

Winds representative of those

experienced during the accident are shown in Table 1-1.

On January 4, 1971, a FAA Douglas DC-*3C, N7 crashed at

LaGuardia Airport, New York, approximately 2000 feet short
of the approach threshold.

The probable cause of the

accident was reported [1-7] as the fai.lure of the pilot to

recognize the wind shear conditions and compensate for them.
The meteorological conditions which prevailed the day of the
crash were a warm front between the Kennedy and LaGuardia
airports and the winds aloft were southwesterly and quite
strong while the surface winds were from the northcast and
relatively gentle.
3

'1
TABLE 1-1
REPRESENTATIVE WINDS ALONG FLIGHT PATH OF DC 10-30 AIRLINER
WHICH CRASHED AT LOGAN AIRPORT, DECEMBER 19, 1973 (1-6]
Altitude
ft.

Longitudinal
Component
kt

Lateral
Component
kt

Direction/Speed
North/kt

1600

25 tail

26 left

169/36

1000

23 tail

26 left

176/35

650

21.5 tail

25 left

176/33

500

18.0 tail

23 left

179/29

420

12.5 tail

18.5 left

183/23

350

9 tail

14.5 left

190/17

255

3 tail

10 left

229/9

180

6 head

2 left

270/5

100

6 head

2 left

296/6

4 head

2 left

296/5

Sowa (1-81 reports four cases of wind shear ranging

from a catastrophic accident to a frightening experience
for which the primary cause can be attributed to wind shear.
Figure 1-1 from [1-8] illustrates the flight recorder traces
of two dissimilar aircrafts at widely separated regions of
the earth.

The figures suggest that the aircraft most likely

encountered a strong headwind near 3000 feet resulting in a

pitched-up attitude arid a high rate of ascent.

At 8000 feet

with the throttle and altitude probably adju. -ed to trim out
4

~~
o ., ~. ~~~
~

~~~~~

....

."
..... t

the high climb rate, a tailwind was presumably encountered,

resulting in a very low pitch attitude and a rapid rate of
descent over a distance on the order of 7000 feet.

Another

1000 feet and the incident may well have been fatal.
Nine missed approaches (see Table 1-2 from [1-8]) made
at J. F. Kennedy Airport, New York, between 2152Z and 2354Z
on January 4, 1971, highlight a problem also associated with
wind shear.

Chambers F1-9] reports a warm front approached

JFK from the south arriving approximately 2300Z.

At 2330Z,

winds were as indicated in Table 1-3.

The wind speeds shown demonstrate that a wind shear of
2 kts/100 ft was present.

Prior to the passage of the front,

the surface wind was 0400/7 kts.

During approach under these

shear conditions, a decreasing tailwind with decreasing

altitude was encountered and heavy (slow response) aircraft
were unable to prevent a speed overrun.
on the glide slope.

They were also high

This conceivably caused many of the

missed approaches shown in Table 1-2.

Difficulties occurred

during takeoff as well since departing airplanes experienced

loss of air speed due to the wind shear and also a thrust
loss due to flying into warmer air aloft.
Chambers [1-9] points out that British Overseas Airline
Company pilots have reported other cases of marked wind shear
at New York.

Additionally, he notes that BOAC lost an airplane

shortly after takeoff from Kano, Nigeria in 1965 due to wind

shear from a cold air down draft accompanying a thunderstorm.
5

~AIRSPEED__.

450

ALTITUDE

T-

C)-

c)

300
Lu

150

C).
CD

CD/

MINUTES FROM START OF TAKEOFF

9

450

AIRSPEED
ALTITUDE----

u-

CD3

I--

6 -300
Lu

.I
-JL
LLU

""
0.

~MINUTES

1.

36u
41
5
FRO, 1 START OF TAKEOFF

FIGURE 1-1

FLIGH4T RECORDER TRACES OF. TWO DIFFERENT AIRCRAFT

__6

PASSING THROUGH UJNEXPECTED WIND SHEAR [1-8]

*TABLE 1-2
DATA FROM TOWER LOG OF A MAJOR AIRPORT [1-8]
Time (Z)

Aircraft

2152
2200
2237
2300
2302

Twin Turbo
Wide-body
4-Engine Jet
4-Engine Jet
Wide-body

Missed approach & diverted

Landed
Missed approach
Landed second approach
Missed approach & diverted

2304
2329
2333

4-Engine Jet
Tri-Jet
4-Engine Jet

Missed approach
Missed approach
Missed approach

2341

4-Engine Jet

Landed second approach

2346
2349
2353

Tri-Jet
Wide-body
4-Engine Jet

Landed second approach

Missed approach
Landed second approach

2354
0013

Wide-body

0020
0023

Wide-body
Wide-body

Missed approach
Changed from runway 04R
to 22
Landed second approach
Landed second approach

Comment

TABLE 1-3
WINDS OVER JFK ON JANUARY 4, 1971

Direction/Speed

(kts).. .. .. ...............

ll-9]
t It
t ud-

220/05

Surface

215/26

1000 ft.

220/46

2600 ft.

Wind shear is also considered a major factor in two

recent airline crashes.

A Continental Airlines 727 crashed

on takeoff at Stapleton Airport in Denver on August 8, 1975,

with no fatalities, but a more ill-fated Boeing 727, Eastern
Airlines Flight 66, crashed on landing at J. F. Kennedy
Airport in New York June 24,

1975, resulting in 112 deaths.

Wind shear is now being recognized as a hazard to

terminl flight operations.

Section 1.3 describes qualita-

tively how wind shear influences the aircraft during approach

and departure.

1.3

Effects of Wind Shear on Aircraft Flight

Basically, there are two potentially hazardous shear

situations [1-10].

First, a tailwind may shear to either

a calm or headwind component.

In this instance, initially

the airspeed increases, the aircraft pitches up and the

altitude increases

(Figure 1-2).

Second, a headwind may

shear to a calm or tailwind component.

Initially, the

airspeed decreases, the aircraft pitches down and the altitude

decreases (Figure 1-3).

Aircraft speed, aerodynamic char-

acteristics, power/weight ratio, powerplant response time and

pilot reactions along with other factors have a bearing on
the severity of wind shear effects.
The potential hazards of wind shear therefore suggest
that manned flight simulators should be programmed to train
flight crews to cope with shear conditions and for fast time
computer analysis to relate the potential hazards posed by

TAILWIND

-4

......

HEADW IND

AYX
OUTER MARKER

FIGURE 1-2

APPROACH INTAILWIND WITH SHEAR

HEADWIND

TAILWIND

RUNWAY

OUTER MARKER

FIGURE 1-3

APPROACH IN HEADWIND WITH SHEAR

9

(1-101.

the wind shear phenomenon to different types of aircraft and

control systems.

Section 2.0, which follows, identifies the

various wind shear regimes while Sections 3.0, 4.0 and 5.0
provide a concise review of the state-of-the-art knowledge of
frontal wind shears, thunderstorm wind shears and wind shears
found in stable and neutral atmospheric boundary layeis, respectively.

Formulation of existing models and/or wind data

into mathematical expressions is given in these sections for

later applications to the engineering models to be developed
in the final report.

I
,I

i1

2.0
2.1

TYPES OF WIND SHEAR

Introduction
Wind shear is caused by several different motions of

atmospheric air masses.

The atmospheric boundary layer in

its natural state always contains some degree of wind

shear.

The geostrophic wind* aloft which flows parallel to

the isobars is turned and retarded by frictional forces as

the earth surface is approached.

This creates the atmospheric

turning layer which has both directional and speed variation

with height resulting in vertical wind shear.

Near the earth,

horizontal variation in winds occur due to terrain irregularities, creating horizontal wind shear.

Thus, during most

approach and takeoff operations, some degree of wind shear

is encountered.

The strength of the shear and the degree to

which it becomes hazardous is dependent upon the existing

combination of meteorological conditions.

Under conditions

of strong nighttime temperature inversions, which may have

in addition a low level jet near the top of the inversion,
severe shears can be created, whereas, under conditions
of a super adiabatic lapse rate, shear may be destroyed by
vigorous turbulent mixing.

Wind shear can occur in land/sea

*The qeostrophic wind results from a balance of the horizontal pressure gradient and the Coriolis force, and it blows along
straight parallel isobars above the boundary layer.
11

breezes and in anabatic/katabatic winds associated with

local topography.
Large wind shear occurs in association with cold air
downdrafts spreading over the ground outwards from a thunderstorm (gust front) which may precede the storm by as much as
10 miles.

Slowly moving warm and cold fronts create wind

shear primarily of a directional variation but also at times

of a wind speed variation.

Cases of low level jets with

accompanying strong shear have been noted in warm sectors

near cold fronts.
Sections 3.0, 4.0 and 5.0 of this report provide insight into the current understanding and prediction schemes
available for the three major sources of wind 'shear, i.e.,
frontal wind shear, thunderstorm wind shear and wind shear
associated with stable and neutral boundary layers.

112

3.0 FRONTAL WIND SHEAR

Fronts, as described in this section, refer to warm and
cold frontal surfaces in motion and not to the thunderstorm
gust front discussed later.

Not all fronts produce signifi-

cant shear and, in fact, many have gradual transitions in

wind speed over broad transition regions.

Those cold and

warm fronts which do have sharp, narrow transition zones

generate significant wind shear that is particularly disturbing due to associated bad weather.

It was wind shear

created by a slowly moving warm front that caused the

numerous missed approaches reported at J. F. Kennedy Airpoit
January 4, 1971.
Mathematical expressions for the structure of wind fields
accompanying fronts are not presently available, but criteria
for those fronts which contain significant wind shear is given
by Sowa [1-8] as:

1) fronts which have a 10OF (50C) or more

temperature difference immediately across the front at surface level; and/or

2) fronts which are moving at 30 kts or

faster.
Wind shear associated with fronts can be either directional shear or speed shear.
cases of directional shear.

Figure 3-1 illustrates two

The large solid line in Figure

3-la indicates the cold front.

The wind below the front at

A (surface wind) is from 3200 and the wind above the front

13

WARM SECTOR

COLD FRONT

WITH~WPYA ROIJjaTi

(which is paralled to the isobars) is from 2200.

Similarly,

point D in Figure 3-lb which has a warm front sloping over

it, has a surface wind from 0400.

Above the front, the wind

is paralled to the isobars from 2300.

The height above the

landing surface at which the directional change occurs is

difficult to determine.

A cold front moving at 30 kts or

more might reasonably be expected to have a frontal slope

such that the front is 5000 feet above the airport three
hours after the passage of the front.

Figure 3-2 shows a

linear extrapolation of the above observation.

The shear

associated with a warm front usually exists below 5000 feet

and for approximately six hours prior to the warm front
passing the airport.

The shear becomes negligible once the

warm front has gone by.

Wind speed changes as opposed to wind directional
changes occur most frequently with slow moving warm fronts
having large temperature differences.

Very little infor-

mation on predicting these speed changes is available.

The presence of turbulence with the frontal wind shdar
may or may not occur.

Figure 3-3 indicates the approximate

conditions of speed and temperature difference which will

induce turbulence along with the wind shear.
Recent data available from the Atmospheric Variability
Experiment Number IV (AVE IV) provides three hour soundings
of wind speed and direction at 25 m increments in elevation.

15

C)

0
0-

w ci

o
-LH

u.
LLZ

0:
--

HW
(f) z'

C:)

5c

C)

-'H
U

I-l
C/)

F-

uLJ

LL-

SC)
C)
C:)

IL

C:)
C:)
LPI
CD
CD
CD

CD)
c-.
C)
CD
C)

0-4

C.)

)
C:)
CD
CD
CD

39NVHD N011NI UNIM JO IH91311 UJIVWIIS3

16

II

0-0

C4

~j

___C

__

-o4

LL =z

CU0

U.

c
V
F-

LL.

ZL.

FE

..

<

cn

-oLo
Int

U>

W'

UOJid 10

17

X UW
-)

<_

3LLJ

~LC

wwLL

co
-)

-.

-0

W.

padS

U-

These data are available for 23 stations throughout the

eastern half of the United States.

Analysis of these data

to extract detailed structure of fronts and to develop a

predictive model of frontal winds should be carried out.

-I

11

18

4.0 THUNDERSTORM GUST FRONTS

4.1

Introduction
One of the most common causes of significant wind shear

is the gust front associated with thunderstorms.

The thunder-

storm gust front is believed responsible for the Eastern 66

crash in New York in 1975, and for the BOAC crash in Kano,
The severe wind shear accom-

Nigeria in 1965, among others.

panying thunderstorms is generated by a vigorous rain-cooled

downdraft, which spreads out horizontally from the storm cell
as it approaches the ground.

The cold outflow is led by a

strong gusty wind which often occurs as much as 20 km ahead

of the storm, called the gust front.

During the passing of

the front, winds may increase from a relative calm to significant values in a very short time and then decrease as
suddenly.

This sudden increase in wind is called the gust size

and the time over which it occurs is called the gust length.
4.2

Storm Cells
The salient kinematic features of a thunderstorm are best

described by considering a particular storm reported by Kropfli

and Miller [4-1].

The storm extends to a height of approxi-

mately 13 km and covers an area of approximately 324 km2 .

Figure 4-1 schematically illustrates four typical features
common to most thunderstorms which are:

mid-level (6 km in

this case), entrainment of cool dry air from behind the

storm, a precipitation-filled downdraft, a gust front
19

0.-.

0
In-

E:

I-r
0ILL

1 ~~

(1SW)
0

C0

consisting of air originating from the middle level and a warm

air updraft ahead of the storm.

Figure 4-2 from [4-1] shows

the approximate three-dimensional motion of air parcels in

the storm.

This figure illus~.rates that the gust front at

the surface on the right forward flank has' its source of air
from the lof~t rear flank and exits at the surface under the
updraft.

The source of air for the northern portion of the

gust front comes mainly from air entrained at the left for4,,

ward flank and partially from the left rearward flank.

Air

entering at mid-level on the right rear flank descends and

exits at the rear about 0.5 km above the ground.
.4

This

downdraft out the backside is also supplemented with inf lowing

air which starts atC the lower edge of the updraft.

Air

entering from the right side and the right forward flank moves
upward with cyclonic curvature and exits near the 11 km level.
Most of the upward moving warm air inflow ~from tihe forward
central edge of the storm enters at a 250 backward tilt (the
direction of this sheet of air is often called the gust
enteingat
te
21 riht
lan desend
rotation,
exiting near
and ascendsn with
little ear
front slope) md-lvel
the 11 km level.

Part of this air splits off, however, and

exits from the left side near the 9 km level.

Contours of vertical velocity relative to the storm motion
in horizontal planes through the storm at three elevations
are shown in Figure 4-3.
4

Three kilometers above the surface

0L
1CL

wa

:
00

-L

-z

,10

uiCl

NC

0'L-I

U)/

22

N4

~
'~c*w-

t*

'

*,-i.~

w
a-

II

In

>-I

I.--

CO

Uc

.4

Cw w

r4

coto
A.,

t~

23*

A-

the inflow air is moving upward at approximately 3 m/s

whereas, the downdraft descends in the center at 4 m/s.
At 7 km the inflow has penetrated the storm with vertical
velocities reaching 11 m/s.

The highest updraft reported

during the life of the storm occured near the top and was
observed to have attained a speed of 19 m/s.

The rature of

the relative velocity in the horizontal planes is indicated

in the hodograph shown in Figure 4-4.

These velocity contours

correspond to a particular instant of time and, although retaining similar features, will show strengthening or weakening
as the storm matures and decays.
The foregoing comments were intended to describe the
large scale charactcristics of a thundcrstorm.

It is the air motion in the gust front of the thunderstorm below the 2 km level which is of most significance to

aeronautical operation in terminal areas.

Some relatively

detailed wind tower measurements cf gust fronts are reported

[4-2, 3, 4, 51 although a complete understanding of the
physical mechanism and a predictive mathematical model of the
three-dimensional motion is still needed.
4.3

Gust Fronts

4.3.1.

Gust Front Types

Figure 4-5 illustrates the common features of a gust

front.
The figure shows that the pre-storm warm moist air is
2i1

6 m/s

4 m/s
9
.3
2 m/s

10

12

HEIGHT IN KI.OMETERS---ABOVE SEA LEVEL

FIGURE 4-4

15

HODOGRAPH OF HORIZONTAL VELOCITY IN THE THUNDERSTORM

CELL., [4-].

25

Ito

zz
w
C)i

-JCz

<z

LL

LI.

CC

cn

<

Ll-

26-

K
displaced upward by the more dense outflow from the thunderstorm.

The mechanical forcing occurs in a 1 km wide band

in advance of the gust front.

After rising about 800 m

over the head, gravitational instability forces the uplifted

air down in the wake of the gravity current and horizontal
divergence assists entrainment of warm air across the inversion boundary into the cold outflow.

This wake, then,

is a highly turbulent zone frequently characterized by large

shears in the horizontal wind and large oscillations in
vertical motion.

Similar characteristics are observed in

secondary outflow surges.

Horizontal divergence has been

observed to be so strong in the wake of these secondary

surges that large volumes of air may descend almost to the
surface.
The extent and intensity to which the features illustrated
in Figure 4-5 may be present depend upon the evolution c,. the
thunderstorm.
frontal cases.
1.

Goff [4-2] has grouped his data into four

These are:

Gust fronts associated with intensifying storms or

accelerating outflow.

2.

Gust fronts associated with mature intense storms

or strong outflow.

3.

Gust fronts associated with dissiFating storms or

outflow decelerating with respect to the storm.

4.

Gust fronts in the final stage of life cycle.

27

4.3.2

Gust Front Speed

Table 4-1 lists the gust front speed of the 17 thunder-

storms reported in [4-2,.

Gust fronts associated with

intensifying and accelerating outflow tend to move more

slowly relative to the ground than those associated with
storms in the state of final dissipation which move more
rapidly.

The gust front speed has been related to the

velocity component in the cold air by Clarke [4-61 as:

c =0.67 uc
Frank [4-5] gives c = 0.62 u

4.1

and Colmer's [4-3] data from

a visual inspection seem to support a value on the order of

0.67.

Goff [4-2] finds Equation 4.1 suitable for case 1

type storms, but that

c = 0.7 uc + 0.3 uw

4.2

where uw is the velocity component in the warm air (normally

negative) correlates types 2 and 3 storms somewhat better.
Type 4 storms appear to have a different propagation
mechanism and are not well predicted by either Equation 4.1
or 4.2.
4.3.3

Gust Front Characteristics

Most all gust front passages are either dry, or almost

coincident with the onset of rain.

The dry gust front may

move as far as 10 to 12 km ahead of the precipitation and

28

TABLE 4-1
GUST FRONT SPEEDS OF 17 DIFFERENT STORMS

Type

No. of

Gust Front

Storms

Speed (r/s

5.0, 8.6,

11.6

6.1,

16.7, 11.0,
11.5, 13.1,

9.6, 8.0, 11.4, 6.9

17.5

11.8,
12.4,

8.5
5.5

can travel this distance from the cell center in 15 to 20 min.

[4-7].

Goff [4-2] reports that the separation distance

between precipitation and the gust front is largest for

mature storms, case 2, whereas Colmer [4-3] reports that for
the 11 thunderstorms he studied, the gust fronts remain beteen 0 to 7 km from the storm center while the storm is in
the mature and developing phases and then move ahead of the
storm by 8 to 20 km during the decay phase.

Goff [4-2] re-

ports that the storm overruns the gust front in the decaying
stages and precipitation tends to coincide with the gust
front passage.

It appears that the definitions of the storm

life by these two authors may be inconsistent or that the

storms are different.

Note the average separation distance

for case 1 storms, from Goff, is 7.6 km; case 2 storms is

10 km which correspond roughly to Colmer's 0-7 km and 8-20 km
range.
29

The important observation from the above is that gust

fronts with sudden wind changes may pass over an airport
several miles ahead of any indicated precipitation or storm
activity.
4.3.4

This creates unexpected wind hazards.

Gust Front Shape

The gust front boundary is defined as the boundary

separating the flow originating from a cold source and the

flow of air originating from a warm air source.

The boundary

is generally shaped with a nose protruding ahead of the front.

Surface drag tends to retard the outward rushing cold air
near the ground causing it to overrun itself above.

The

point of the nose is on the order of 100 to 300 m above the

surface (both Colmer [4-3] and Frank [4-5] report the peak
at 250 m),

and can extend horizontally as far as 200 m into

the warm air.

The lag of ground level air behind the foremost part
of the front is expected to be intermittent with the cold
air falling through the warm air entrained beneath the nose
and again reforming due to surface stress.

This over-

turning implies high turbulence in the nose region.

The

bulging gust front is most prevalent in mature storms.

Above the peak of the nose the gust front boundary
slopes backward.
and 450

to 750

Goff gives angles of 450 for case 1 storms

for case 2 storms.

A time-height plot of

streamlines shown in Figure 4-6 clearly illustrates this

30

_______________________________________'.~"~.--.-.C-

U.s

F-

I,

0
C/)
Ul)

z
F-

wL <

z- W
U-

LO
L=
-

31

L.L

slope which coincides with the zero streamline.

As previously

mentioned, Kropfli and Miller [4-1] report a 250 slope which

corresponds to the decaying stages of their storm.

Colmer

[4-3 ] notes that the frontal slope is not linear and parts
of the gust front will have much larger slopes than the
average (see Figure 4-7).

He correlates the ratio of maximum

slope over a 30 sec. period to the average slope, R, with

the gust length, H, by the relationship
log R = 0.667 log H + const.

4.3

Depending on the relative position of the storm to the airport, a landing aircraft will encounter the most severe wind
shear at different heights above the ground.

Also, the rain

typically follows some distance behind the gust front and

the pilot may or may not land in rain.

Since for simulation

purposes it is useful to relate the location of the shear

and the probability of rain to the position of the airport,
Figure 4-8 has been prepared from limited data.
Figure 4-8 illustrates the altitude at which an aircraft approaching an airport on a 30 glide slope might
expect to encounter a gust front which has passed over the
airport at the time indicated on the horizontal scale.

The

probability of rain at the airport relative to the position

of encountering shear is also shown.

As an example, the

curve shows that if the gust front had passed the airport
10 min. prior to the airplane's apprcach, the pilri
32

-uid

cc-

-~-

C)
LUJ
0- LL
LU

SC)

LO

w
LU

U)

C:)

0-

0
-J

Lij

LUJ
LD
UL-

33

CD-

Lt~L
=s.
S

S.

C=

.*5.0**M

>. <

LUI

ci-.C
L

LU~

CD

i-iLU
I-

CD

C)

C)

(W)INOU ISnO 111M

34

03l10fl30 JO IH913H

00

m)

cu

expect shear at 210 m and a 33% chance of landing in rain.

This curve is interpolated from the data of the preceeding
discussion and should be used only as a rough estimate.
4.3.5

Gust Characteristics
The gust size (magnitude of wind increase) and the gust

length (length of time over which the wind increases) were

plotted by Colmer for 11 case studies.

On an average, the

gust size increases by 50% between the surface and 500 m

with the most increase occurring in the first 50 m.

This is

a significant observation since a ground based anemometer

will read winds considerably less than those a hundred or
so meters above.

Note, however, that the gust size profile exhibits

considerable variability.

The gust length on the other

hand shows less variability and an almost linear decrease

with height.

Additional gust size and length data are

given by Sinclair et al [4-4] for three Florida storms.

They report at 18 m an average gust size of 4.8 m/s and a
gust length of 12.8 s, and at the height where the peak wind
occurs, they report an average gust size of 6.6 m/s and a
gust length of 14.8 s.

The length scale based on a constant

gust front speed (which was not reported) does not show the
same decrease with altitude that Colmer reports.

Goff [4-2]

finds the change in the horizontal wind component normal to

the gust front to have a value between 4 and 20 m/s with an
35

average of 12.8 m/s which shows good agreement with Colmer's

average value above 50 m.
The track of the maximum gust tends to coincide with the
track of the storm cell.

The size of the gust decreases

almost linearly from this maximum value havinc an approximate

15% decrease in gust size 5 km from the center and a 32%
decrease 10 km from the center.
Frequently more than one gust is experienced during the
passage of a thunderstorm.

Ward and Arnett [4-8] report

multiple cold air surges for a single thunderstorm which

they attribute to storm pulsation.
shows secondary surges.

Goff's [2-21 data also

There appears to be no reduction

of the data which allows a quantitative statement about

these multiple surges to be made.

Goff (4-9] in a recent

paper discusses secondary surges of high momentum air which

appeared in the outflow air mass from the thunderstorm on
June 7, 1971.

Two surges observed were both characterized

by a strong shear zone in the Wy-component and large variations in the vertical wind speed.

In one surge, a downdraft

in excess of 11 m/s was observed at the 377 m level.

4.3.6

Wind Speed Fields

The most useful data for simulation of aircraft flights

through gust fronts is the three-dimensional velocity field
in the vicinity of the gust front as a function of position

36

and time.

A step toward providing these data is the height-

time plots given by [4-2, 4, 5].

Often these plots are

converted to height-distance plots by assuming Taylor's

hypothesis.

Taylor's hypothesis allows wind speeds measured

in time at a point in space to be converted to wind speed

variation with position along a line in the direction of the
mean wind by assuming the turbulent flow is carried past the
point of measurement sufficiently rapid that the turbulence
pattern is "frozen" in the air flow and does not change over
the time the measurement is taken (ie position = mean velocity
X time).

Typical height-time plots are shown in Figure 4-9

and height-distance plots in Figure 4-10.

Genirally, these

data are highly smoothed by filtering techniques which remove

the high frequency content.

pwind

Whereas, Figure 4-9 gives the

speed and the wind direction contours in the vertical

plane, Figure 4-10 expresses the data in terms of the three

velocity components:

W , vertical velocity, Wx, horizo. ,l

velocity normal to the front, and Wy, lateral wind speed

parallel to the front.
The data corresponds with the schematic of the gust
front shown in Figure 4-3.

Inspecting the Wx component of

wind speed in Figure 4-10 one observes a negative wind speed

corresponding to warm air inflow toward the front which is
delineated by the zero wind speed.

The positive values in

the height-time plots correspond to cold air from the thunder37

Io DEGREES

Ito

120-

100

w 0-

14:03

260

140
12

I4o

w0

100

s-10

140
120

0120

25

20
15
TIME(MINUTES)

120

30

150
6

10

I'

rn/s

12b-

01

wo90w

15

10

20

25

30

TIME (MINUTES)

FIGURE 4-9

HEIGHT-TIME PLOTS OF GUST rRONT

38

SPEED AND DIRECTION, [4-4)

2 3

-0.4

0.3
0.2

00.
3

VERTICAL WIND SPEED (M/S)

(kmn)

-1'4

-2-12

-12

01

0.4

-1

22

0.2

-22
3

12

WIND SPEED PARALLEL TO FRONT (M/s)

2 0

24

-2

(kmn)

-4

-6

-8

0.4

0.2

0.1

68
4

RELATIVE WIND SPEED, COMPONENT NORKAL TO FRONT (ml/.)(m

FIGURE 4-10
4

HEIGHT-DISTANCE
FIELDS, [4-2]

39

PLOTS OF GUST FRONT SPEED

storm flowing toward the front.

Negative wind speed near the

surface in the cold air zone indicates an undercurrent probably resulting from the cold air riding over a low level
temperature inversion.

Although filtered out of the data

of Figure 4-10, a highly turbulent wake region expanding in

depth with increasing distance behind the front occurs behind
the elevated "head" of dense air which is characterized by
strong turbulence and mixing.
Inspection of the vertical wind speed in Figure 4-10
shows a vertical updraft attributed to lifting of the less
dense, warm air over the more dense forward-moving air mass.
Maximum updrafts generally occur 1 km or less ahead of the
gust front.

The magnitude of these updrafts generally in-

creases wi4h frontal strength (increasing frontal slope).

The value of the maximum updraft reported by Goff [4-2] is
6.7 m/s, and that by Browning and Harold [4-101 is in excess of
8 m/s and 10 m/s for two different fronts.

The record from

the digital flight data recorder of Eastern Air Lines Flight

66 and Eastern Air Lines 902 for the thunderstorm over J.F.
Kennedy Airport gave updrafts of approximately 4.3 m/s, and
5.2 m/s, respectively.

However, these values probably do not

correspond to the maximum.

4.3.7

Wind Shear
Strong vertical and horizontal shears occur with all

gust fronts,

Vertical shear is a variation in the wind

speed components with height, z, whereas horizontal shear

40

refers to variation with horizontal distance (in this case,

wilt

the longitudinal mean wind direction, x).

Large

vertical shear (a()/az) occurs near the surface and at the

upper boundary and large horizontal shear (()/ax)
at the outflows leading edge.

occurs

Secondary horizontal shear

may occur also due to multiple outflow surges.

The raximum wind shear averaged from the 20 thunderstorm
cases studied by Goff f4-2] are tabulated in Table 4-2a and
Table 4-2b.
relationships

The spatial derivatives were formed from the

a/ax

= c-1

a/3t.

From

the table one notes that

shears of horizontal wind are greatest near the ground

reaching a minimum at 200 m and then increasing upward.

This

tends to conflict with Colmer's [4-3] conclusion that due to

the increase of gust size and decrease of gust length with
altitude, horizontal wind shear in the gust front between
the surface and 500 m above ground doubles, i.e., the average
(of 11 cases) surface shear is 0.0063 s-

while the 500 m

level shear is 0.016 s- . It should be noted that reported

wind shear values are based on a very sparse grid system and
on highly smoothed data and are, therefore, at best gross
approximations.
4.4
4.4.1

Wind Shear Prediction Techniques

Raw Data
Despite the filtering and course grid space used to

obtain the spatial distribution of wind speeds and wind

shears given in the previous section, the results are ex41

<,

4,

--

TABLE 4-2a

1 00 m - 1 ) [4-2]

Maximum Absolute Shear (m s

*Grid level or layer of gradient:

ax

11.1

5.0

4.9

5.4

4.7

4.1

4.8

5.3

4.5

4.6

ax

7.0

3.2

3.4

3.8

3.4

3.7

4.0

4.8

4.8

7.8

aWz/ ax

0.0

1.2

2.2

3.4

3.8

3.4

2.6

2.8

3.3

3.8

DWx/ az

16.2

8.4

8.7

9.6

9.4

8.6

10.8

6.5

10.6

aWyl az

26.1

iO.2

6.9

5.6

8.2

6.4

6.4

6.3

9.7

awz/ az

3.8

2.1

2.3

2.0

2.6

3.1

3.1

3.1

3.1

aW
Wy

10

TABLE 4-2b

Average Maximum Absolute Shear (m s-1 100 m - 1) [4-2]

*Grid level or layer of gradient:

aWx /ax

4.3

2.0

2.3

2.3

2.3

2.4

2.6

2.9

2.9

3.2

aW !lx

4.0

1.5

1.5

1.7

1.9

1.9

1.9

2.2

2.2

2.6

DWz/3X

0.0

0.6

0.9

1.2

1.3

1.2

1.1

1.5

1.9

2.4

DWx/az

11.6

5.8

4.7

4.6

4.1

4.7

5.0

4.1

5.1

---

BWy/aZ

10.6

5.3

4.1

3.6

3.9

3.6

3.4

3.2

4.3

aWz/3z

2.1

1.6

1.7

1.3

1.7

1.9

1.9

1.9

1.9

Levels 1 through 10 are 50 meters apart.

42

10

-- - ----

?I-

tremely useful for estimating the influence of gust fronts

on aircraft flights.

Two ways of using the data are to

store them in the computer on an appropriate grid system

and use linear interpolation with a table look-up scheme
to give the wind input values to the computerized airplane
equations of ntion; or to superimpose geometrically scaled
flight paths on to the contour plots and pick off the winds
"seen" by an aircraft traversing that flight path.

The

latter method is more restrictive in that the aerodynamic

forcing functions computed from the input winds assume the
airplane remains on or in close proximity to the selected
flight path.
4.4.2

Mathematical Models
Mathematical schemes for computing wind fields associ-

ated with thunderstorm gust fronts are still -n the formative

stages.

The most direct mathematical approach is probably

that of Mitchell [4-11] who solved the finite difference

equation for the two-dimensional, time dependent primitive
equation of fluid motion.

Surface friction is modeled with

a bulk aerodynamic drag function applied at the lower levels

of the computation grid.

Vertical diffusion of momentum is

nmoelled with linear diffusion terms, and velocity-slip is

permitted at the surface.
This model appears to capture much of the essential
physics of the flow and provides a computational tecnnique
43

which might be incorporated into a computer code for simultaneous solution of the fluid mechanics with the airplane
dynamics.

Unfortunately, however, the computer solution

is costly and this prohibits any extensive flight

simulation.
A less sophisticated approach to modelling two-dimensional
flow fields about a gust front is proposed by Sinclair et al
[4-4].

They scaled the three Florida thunderstorms with the

scaling parameters; time, At, which characterizes gust length;

velocity AV, which characterizes gust size and; height, Az,
which is the height of maximum velocity.

Figure 4-11 shows

scaled velocity contours which Sinclair et al [4-4] conclude

arc quite similar to all other scaled gust profiles and,
hence, if AV, At, and Az are known, at least part of the
structure of the wind front can be determined.

Because of

the difficulty in evaluating Az, this approach was abandoned

and a simple hypothesis to find the wind structure was
adopted.
We will. not pursue this approach further since the
effort of reference [4-4] was toward forecasting gust fronts,
whereas our goal is to provide realistic wind models for use
in parametric simulation and design studies.
fronts of Figure

4 -11provides

The scaled gust

such a model in that it allows

us to conduct design analysis and simulation studies wifh

44

1.25
1.00

Z/t4tStt

AZ

Attou 1113t13F16

*75~

tr

1.1 10
*

ke 13

A1 y

k\

.50
.25,
4

I4

o-0

.4

A .8

Lb

1.4

Lz

1.6

1.8 2.0

~sl

175

1.501.25.'
1.00

lAz
.75
.50-i

.25

FIGURE

4-11

.2

.4

.6

.8

10

1.2

1.4

L6

Lb8

ao

NONDIMENSIONAL WIND SPEED CONTOURS DISPLAYING

SIMILARITY PROPOSED BY SINCL.AIR

45

ET

AL

[441

parametric variation of gust length,

9.ist size and height

Since we wish to establish design

of maximum velocity.

envelopes or limits rather than forecast gust fronts, we are

at liberty to select any values of Az, AV, and At within the
confines of good engineering judgement.
of this approach in

More will be described

the final report.

Updrafts and downdrafts associated with gust fronts along

a given approach path have been correlated by Fichtl and
Camp

They scaled vertical wind speeds along a 30

[4-10].

glide slIpe from the gust front data of Goff [4-2] and also
incorporated the vertical wind speeds reconstructed from the
digital flight data record of Eastern 66.

Their mathematical

representation of the sequence of vertical wind speeds encountered by an aircraft during landing is given by the
following:
Major Downdraft:
X-x ]
-- -plA sin[-f-q1

z D >z >z
r
--

4.4

Major Updraft:
3+
3X)
(l-2qo) (x
0)(-xr 0
W = A4.5

~46

- L

> z > z

2
(l_3qo) (xXr) 2 + (2qo_3q )(x-xr)
0o
r o2
(q o
-q o2

4.5

Minor Downdraft:
w

-P2 A sin[R

q2

Minor Updraft:
A

->
zr

+ q2 )L

4.6

r-l-x

sin[]

(1 + q2 ) L> z> z

-(1 + 2q)L
4.7

where

Xr
The various

Lr

z
E

4.8

quantities in the above equations are defined

as following:
w = thunderstorm cold air outflow vertical velocity
z = altitude of airplane above surface
ZD= altitude of the top of the major downdraft
Zr= altitude of the top of the major updraft
ZB= altitude of the bottom of the minor downdraft
zm= altitude of maximum updraft velocity in the major
updraft
A = amplitude of major vertical velocity updraft
L = vertical extent of major vertical velocity updraft
relative to the flight path
pl

ratio of major downdraft to major updraft velocities

P2 = ratio of minor downdraft or minor updraft to

major updraft velocity

q0 = (Zr-Zm)/L
*This parameter is not required in the calculation

but serves to physically define q0 .

47

ql

(Zr'ZD)//L

q2= (Zr-ZB-L)/L
Values for the cold air outflow parameters are given in
Tables 4-3 and 4-4.

A schematic is given in Figure 4-12.

Table 4-3 provides cold air outflow parameters for typical

vertical wind speeds as derived from data provided by the
NOAA/NSSL.

Table 4-4 provides cold air outflow parameters

which encompass the worst case (Eastern 66).

To apply the model, one enters the model with the
parameters:
Zr , ZD, ZB,

L, A, Pi' P2

consistent with the range of parameters in Tables 4-3 and

4-4.

It is recommended that a family of profiles which

encompass the full range of variation of the model as implied

in Tables 4-3 and 4-4 be used.

This will insure that certain

conditions will not be overlooked.

Wind speed as a function of height from combined tower
and rawinsonde data near the time of the wind speed maximum
was correlated by Frank [4-5] with the expression
Wx= Wo+ AW tan ( (z-z0 )/ LH)(l-e-Z'/Zm)

48

4.9

where

ZmI

94.72m

AH= 889 m
z = 402 m
W = 25.0 m/s
AW= -23.5 m/s
Figure 4-13 shows the results.

Also Frank [4-5) notes that

in the surface boundary layer regions the wind profile obeys

a power law
WX= Wxr (z/z r)n

4.10

where Wx is the wind speed normal to the front at height z,

Zr is a reference level, and Wxr = Wx(Zr).

Just behind the

front n was found equal to 0.39 from tower data and equal
to 0.35 from Doppler radar data.

The power law was reported

to give good agreemcnt with gust front data up to 350 m.

Sinclair

et

al

[4-4) demonstrated that a logarithmic

wind law is valid from the surface to approximately 100 m.

Therefore, the ratio of wind speed at two different levels
becomes simply a function of surface roughness, i.e.,

V2

ln(z
49

49

/z)

4.11

TABLE 4-3
Typical Values of Cold Air Outflow Parameters (Based on NOAA/NSSL
Data) [4-10]
Units

Parameters
<L <183 + zr

64 + 0.223z

0< A <1 + 0.011zr

0.3< pl

<

1. 2

ft

-1

ft sec

Nondimensional

0<P2_ 0.35

Nondimensional

0.36

Nondimensional

0.25 q1 < 1

Nondimensional

0< q 2 < 2.3

Nondimensional

q0 =

L< z < 1500

ft

zD = zr - qlL

f:

zB = zr -

ft

(1 + qq)L
TABLE 4-4

Worst Case Values of Cold Air Outflow Parameters

Eastern 66 and NOAA/NSSL Data)

r-

[4-11]
Units

Parameters
64 + 0.223 z

(Based on

ft

<L< 183 + z

0 <A < 10 + 0.011z

-1
ft sec

0.3< Pl < 1.2

Nondimensional

<

Nondimensional

0 <P

q0 =
0

ql

0.35

0.36

Nondimensional

Nondimensional
Nondirnensional

0<q2 < 2.3

L< z r< 1 5 0 0
Z Z
zD = zr - ql L

z B = Zr-(l + q 2 ) L

ft

ft
ft

50

2
'ZD--ZR

QI L

ZR

ZM=ZR-QOL

ZR-L

ZB: ZR-( 1 + Q

II

A
UPDRAFT

DOWNDRAFT
FIGURE 4-12

VERTICAL WIND SPEED ALONG A 3" GLIDE SLOPE

THROUGH GUST FRONTS AS CORRELATED BY
FICHTL AND CAMP, [4-]01.

51

2000-

1600.

1200

0
I-

800

5
*44O

20

10

30

14IND SPEED (m/S)

*1

FIGURE 4-13

WIND SPEED FROM COMBINED TOWER AND RAWINSONDE

DATA NEAR THE TIME OF MAXIMUM WIND SPEED
I
I
,I
.
|
i
i
DURING THE PASSAGE OF A GUST FRONT, [4-5).
52

*
,

5.1

BOUNDARY LAYER OVER FLAT TERRAIN

Introduction
Considerable information on the behavior of the atmo-

spheric boundary layer at specific locations has been obtained

with individual meteorological towers up to altitudes of
typically 150 m with some data up to 500 m.

These data are

normally assu.ted valid over flat, horizontally homogeneous

terrain and are transferred to spatial coordinates with
Taylor's hypothesis which requires that the boundary layer
be satistically stationary.

Above the altitude which can be

reached with towers; aircraft, rawinsonde and jimsphere data

provide insight into the turning layers of the. atmosphere
but these data are not as precise as tower data.
5.2

Mean Flow in the Lower Surface Layers

Below 150 m the mean wind direction has, in general,

little variation with height, and over flat terrain is considered to be steady plane-parallel flow of a horizontally
homogeneous and statistically stationary surface layer.
Numerous reviews of the atmospheric boundary layer are
available [5-1, 2, 3, 4, 5] and it is only necessary to
summarize these here.
In the surface boundary layer, the mean wind is considered to be influenced by surface conditions, atmospheric
stability, and elevations.

Mean wind models receiving the

53

greatest acceptance are expressed in terms of nondimensional

wind shear as:

zk

5.1

dW
U, a "" " (zlL)

The wind speed is then given by

z+z
u,
Wx(z) =--

~zL
z

5.2

dz

z
0

WY (z)=0
y (z)5.3

w (z) --0
5.4

where W (z), Wy(z) and Wz(z) are the wind speed in the x,
y, and z directions respectively; the x axis is aligned
with the mean wind direction, the y axis is perpendicular
to x and the z axis is vertical to the horizontal plane.
The constant k is the von Karman constant having a numerical
value of approximately 0.4 and z. is a length scale that
characterizes the surface roughness.

Typical values of zo

are given in Figure 5-1.

The surface friction velocity u, is normally assumed
constant which is experimentally verified only in the lower
30 meters, however, its variations to 150 m is generally
negligible.

54

0.5.
"

-550
FLAT 1
04ON

WOODLAND1
FOREST
.

ICOUNTRY

-500
0.,3 --.
C)
,,<

- 450

10'23
URBAN
kI AREA

c/-.)"

4/

3c-

0.2 00

-350

300

OPEN SEA
0I0-

510-1

2 102

0 2
102

SURFACE ROUGHNESS LENGTH (Zo),CM

FIGURE 5-1

SURFACE ROUGHNESS LENGTH,Z 0 ,VERSUS THE POWER-LAW

EXPONENT 1/NJWHERE WX/WX REF (Z/ZREF )1/N 15-231.

55

103

T'1

II

The quantity L is the Monin-Obukhov stability parameter

u*3 c pT

P5.5

L=-

kg H

Since th- heat flux, H, appearing in this expression is

dif"icult to measure, an alternate scaling length L' is
frequently used:

U*T

_W]

kg[ -T +

jL

The relationship between L and L' is given by

KM

5.7
H4

viherea Kt and XM are the eddy conductivity and the eddy.

viscosity, respectively.

L' in turn is related to a more

conventional stability parameter, Ri, called the gradient

Richardson's number:

Ri

5.8

herce

+
, 30

21= Ri 4(z/L')

56

5.9

Richardson's number is a nondimensional ratio between the

mechanical wind shear that tends to displace air and the
buoyancy force, which may damp or amplify this tendency.
The atmosphere is said to be stable if Ri, z/L'>O corresponding to aT/az<-g/cp indicating a stable lapse rate or
inversion; neutral if Ri, z/L'=0 giving aT/Dz =-g/cp
indicating an adiabatic lapse rate; and unstable if Ri,
z/L'<0 giving aT/az>-g/cp indicating an unstable lapse rate.
The empirically derived expressions for O(z/L') for
different conditions of stability are given by a number of
investigators. Barr et al [5-5] recommends the following
forms for 0 (z/L').

Neutral case:
o(z/L')

= 1;

5.10

z/L' = 0

Stable case;
O(z/L')

= 1 +a 'z

1> z/L'> 0

5.11

Very stable case:

=

*(z/L')

1 + a';

z/L'>l

where

=I

1 + a' Ri

57

5.12

Unstaole case:
10

(z/L')

~(1

y' Ri)

5.13

>z/L'

where
Ri
(1+

z,

Ri)k

The nondimensional wind shear has the form shown in

Figure 5-2 where a = 4.5 and y = 4a- 18.

Details of the

development of Equations 5.10 to 5.13 are given in [5-5,6].

The integrated form of Equations 5.10 to 5.13 is

u
Wx =-

z +z 0
[In

--

Zo

(z/L')]

5.14

-1 d(z/L')

5.15

where
V(z/L)

z/L'

f
0

Before considering the result of the integration, it is

noted that the above expressions for nondimensional shear
become increasingly less accurate at altitudes above z > 90 m.
This is due to the fact that their development assumes a
constant shear .tress which does not hold at high altitudes.
Barr et al [5-5] propose

an extension of the model to

altitudes on the order of 350 m by allowing the friction

velocity to vary linearly as

5.16

(l-z/6)
5*0
58

NONDIMENSIONAL SHEAR

kz
U az

kz

U*o az

awx

5.5

f~

*0

kz

awx

___

u*o az

NONDIMENSIONAL ALTITUDE, z/L'

FIGURE 5-2

NONDIMENSIONAL WIND SHEAR AS A FUNCTION OF HEIGHT,

ZOVER MONIN--OBUKHOV STABILITY LENGTH,L',1-]

59

_________

~F

~~

where 6 is the atmospheric boundary layer thickness given by

5.17

10.7f

=*O

The parameter f is the Coriolis parameter.

The nondimensional

wind shear is then given by[

kz

u*o ~z

1-Z/6) OWzL'

+z0

[n

5.18

5.19

Z +_
+

The wind profile for the given stability cases then becomes;

Neutral case:

uW[i-

z + Zo

--

z/L'

60

5.20

r~~-

-- - -

- ------

r"

Stable case:

Wxn
u*o

fl

2+z z0

z+

/6

5.21

O<z/L'< 1
Unstable case:
u*o

-z

+ z0
2 0n

2
-

5.22

where

Z/L'
-

(()

1-K ) + z/6,'

(C)dt

5.23

0
where
-

z/L' and z/L' <0

Numerical integration of Equation

5.23 is carried out in
f5-5].
Very stable case:

UW
--

z +*-I
z + a'(i+ In E,)

[in

7i

+o

+
6

z/L' >1

61

a,-

,
2+z/(LJ

5.24

The very stable wind profile is applied above the altitude zI where zl/L'= 1.
Equation 5.24 must be used with discretion since it is
based primarily on conjecture.

Little is actually known

about the very stable wind profile, the shape of which is

largely determined by the absence of turbulence above z/L'>l.
Very near the ground where z/L' is less than unity a turbulent
As height increases beyond z/L'=l a non-

layer can exist.

turbulent layer can occur which is decoupled from the lower

layer.

Change

in wind direction in excess of 450 is not

uncommon between these uncoupled layers.

Leurs [5-4] suggests

a discontinuous wind profile for very stable conditions.

Below a given level z he assumes either a calm Wx (z) =0
or a logarithmic profile.

Above z, he considers various

constant values of Wx (z).

The wind speed change across

z I is taken as discrete which is probably not physically

meaningful but represents a limiting case.
The application of the mean wind profiles is most
frequently based on a reference wind speed at a specified
height.

Present efforts to standardize tower heights at

airports to 20 ft suggests 20 ft might serve as a meaningful

reference height.

Given a reference wind and height, u~o

can be determined from

u_

Wx (Z ref )

+ zo
In Zref
____-+4(--r--,~
zo

Zref

62

Zref
-f)-

Zref

5.25
.2

In view of the fact that 6 is a function of u*o an iteration

procedure is required if the z/6 function proposed by
Barr et al [5-5] is retained.

The effects of altitude

variation of shear stress, is u,- U*o (l-z/6), is negligible

for unstable conditions but can be appreciable for stable
conditions.

The Richardson number or L' must also be

known at the reference height; or a parametric study of

variation with Ri may be carried out depending on the
design problem.
The roughness lengtb,zo,should be specified for each
particular airport and may be a function of wind direction
(see [5-7] for an example of a surface roughness wind
direction rose), wind speed (for instance tall grass and
water have changing roughness characteristics with wind speed),
season, etc.
A comparison of the various equations given above
(see [5-51 ) indicates that stable conditions result in more
severe and higher winds above a certain level for the same
reference wind speed.

However, since this wind is propagated

from the geostrophic wind, it might be expected that greater

stability at the reference height will be associated with
lower mean wind speeds at the same height, and consequently
the differences in the shears between stable and unstable
conditions may not be so great.

63

Turning Layer

5.3
5.3.1

Introduction

Above the constant stress layer the presence of mean

flow synoptic scale horizontal pressure gradients and the
action of the Coriolis force tend to produce vertical variation in the vertical transport of horizontal momentum and
clockwise turning of the mean flow, looking toward the
earth, in the northern hemisphere and counterclockwise
As indicated earlier,

turning in the southern hemisphere.

however, experimental evidence [5-8,9] appears to show that

in the height interval 30 m < z < 150 m the vertical variation
of the wind direction is approximately 20 - 40 which is

negligible for most design analyses.

Additionally, by allow-

ing u, to vary linearly with altitude, Barr et al [5-5] argues

that Equations 5.20 to 5.24 reliably predict Wx

(z)

to an

This may be true under

elevation of approximately 300 m.

barotropic conditions, b,-t if baroclinicity exists, the

turning of the wind may become important as low as 150 m and
should be considered in analyses of aircraft flight characteristics and control system design.
Most solutions for the wind profile in the turning layer
treat the barotropic case which assuames that the large-scale
synoptic horizontal pressure gradients which are related to
the geostrophic wind WG by
W
1 ap
-

xG

_f

p. ;

yC

ap

pf

D-x

64

5.26

do not vary in the vertical direction,

Surveys of such

solutions are given in [5-1, 3, 10, 11 and 121

Reference

(5-12] summarizes as well, analytical solutions which

discard the barotropic assumption and consider specified
height variation of the geostrophic wind, i.e., the baroclinic
case.

The results of these solutions will be summarized

below, foi further details the reader should see the stated
references.
5.3.2

Barotropc Turning Layer

According to similarity theory, if a variable wind is

appropriately scaled then its profile is given by an empirically

determined universal function.

For flow far from the surface,

similarity gives:

F (, )
h

- ref
u*o

5.27

where Wref is the appropriate wind scale to be discussed later

and F is the outer profile function.
For flow in the surface layer

0 - L5.28

65

where i is the unit vector ip the x direction.

It is

corrai.only assumed that there exists a layer where Equations 5.27

and 5.28 can be asymptotically matched [5-13, 14] to give

Wref

5.29

Equating components of the vector:

A(h/L) =

B(h/L)

kWye
k

5.30

re

and

U*

f
f

5.31

The functions A(h/L) and B(h/L) are determined from empirical

data.
Under neutral conditions A(0) and B(0) have a range of
values as shown in Table 5-1.
The geostrophic drag u*/IWrefl is found from Equation 5.30
as

CD =

/IWrefI

= k

I[[n h/z

+[B(h/L)]

66

A(h/L)]

5.32

TABLE 5-1
Values of A(h/L) and B(h/L) Under Neutral Conditions h/L
Reference

A (0)

B(0)

Yamada [5-15]

1.855

3.020

Fichtl [5-1]

1.6

4.7

Tannekes [5-10]

Monin [5-3]

1 to 2

4.2 to 4.7

Clarke and Hess [5-16]

1.1 0.5

4.3 0.7

and the angle between the direction of the surface wind and
that of the reference wind vector Wref is obtained from

u,

(h

f)

5.33

sin a =kWrefB(h/L)

For the barotropic atmosphere the appropriate scale for

wind is
A

--

Wref= WG

5.34

where the geostrophic wind, WG, is constant.

The appropriate height scale for neutral stability is
h = clu,/ ff where cI

0.3.

Departures from neutral

stability cause some controversy over the appropriate form

of h.

Yamada [5-15] found the relationship between h/L and

u*/IfIL shown in Figure 5-3.

The expression h/L = 0.3 u,/IfJL

67

600

400

200

0
U*

_0.3

-200

O'311 L

-400

-600
-1200

FIGURE 5-3

-400

-800

HEIGHT SCAI.E FOR THE PLANETARY BOUNDARY LAYER

68

400

[5-15].

holds reasonably well for u*/IfIL<200 but it underestimates

h/L considerably for u*/IfIL> 200.

Other workers [5-17]

find the height of the inversion layer zi during nighttime

and the mixing layer during daytime to be the correct scaling
height.
The foregoing analysis does not permit computation of
the wind profile variation with altitude.

It does express

the geostrophic drag coefficient u*/IWGI and the cross isobaric wind angle as a function of the surface Rossby number
Ro = "W

I/fz

for given values of h/L.

Note, if h can be

written u*o/f, Equation 5.33 can be rewritten

In Ro

B(h/L) + inWGI
U*o

k2

U*o

A(h/L) 2

A meaningful wind profile variation with elevation can

not be expressed in a simple mathematical form.

Most all

solutions to the wind spiral with height have been carried

out nu enrical:

The equations solved are

- f(WG-WxZ,,))
f(W

_.

xGW

XW)5.36

and

f(W

+W (z))

5.37

69

where
~

dW
"x

pK

y
y=

dW
pK dz

5.38

The eddy momentum K must be modeled to close the set of

equations.
K

i,

The expression
k 2 [(

1+

l)

5.39

where
k(z + zo )

54
5.40

5/4

1 + 4 (z/zm)
was proposed by Lettau as reported by Hanna [5-11].

Other authors

[5-18,

19) have used similar forms for K

ar.d k and obtain essentially che same solutions for

W (z) and WY
y(z).

Figure 5-4 shows Lettau's solution in

dimensionless form.

This solution is valid for all neutral

wind spirals,the only variation being the length of the

straight line extension along the left asymptote.
To construct the neutral oarotropic wind profiles
from Figure 5-4, first locate the required value of
fzo/(0.185 uo).

The geostrophic wind is the vector

from tbis point to the central point of the wind spiral.

Any other vector drawn from this point to any point on the
curve represents the wind velocity at the heigit indicated
by the end point.
70

'

CNz

'-I
00

00
0-

JH
>

V,)

0
D~

I~00
CICI

00

9Z3
711

-4

As an example, suppose zO

0.01m, f = 0.84 x 10- 4 s- 1

and u*0 = 0.22 m/s, the geostrophic wind is given by

W= VW2 +W
xg

2
yg

where W xg and WYg are determined from Equations 5.30 and

1.63 m/s
5.31 with h = 0.3u*/Ifl. Thus, Wxg = 5.11 m/s, Wyg
and Wg

5.36 m/s.

The wind speed and direction at z = 162 m

is obtained by scaling from Figure 5-4 which gives

W = Scaled Distance (0.32-0.00002) W /Scaled Distance
(center - 0.00002)
= (4.75)(5.36 m/s)/5.16 = 4.93 m/s
and a = 70

Solutions for the diabatic barotropic wind' profiles

have been numerically computed [5-20, 21] but are not
expressible in a non-dimensional form such as Figure 5-4.
The experimental results of Clarke and Hess [5-16]
provide a method of estimating the wind profile for
conditions departing from those of neutral stability.
Figures 5-5

and 5-6

give the departure of the nondimensional

wind components parallel,W x(z)/u*o, and perpendicular,

to the surface wind respectively from the
W (z)/u.,,
1
y
reference wind defined at z = zf/u*o = 0.15 as a function
of nondimensional height, z.

Using the values of AI(j)

and BI(w).where j,= ku*/IfIL, given in Figure 5-7 and 5-8

which are based on z

0.15 and Equations 5.30 and 5.31

with h =!fl/u*o the values of Wxref = Wx(^

Wyref = Wy(z

0.15) can be determined.

72

0.15) and

For a given

stability conditior v the values of AW

(W (z

0.15)

Wx(z) )/u*o and AWy

(Wy(z - 0.15)

- Wy (z))/U,

can be

scaled from Figures 5-5 and 5-6 at any height z up to a

value of 0.15.
Figure 5-9 shows a plot of the wind spiral for a
lm determined in the above manner.

stable atmosphere of L-

One observes that the turning of the wind begins at

approximately 20m indicating that the turning layer can
become very significant under stable nighttime inversion.
Figure 5-10 compares the log-linear wind profile Equation
and Equation 5.11 with the horizontal component of the
turning layer.

The log-linear law begins to depart

appreciably at 20.m.

Baroclinic Turning Layer

5.3.3

The baroclinic turning layer has been solved

numerically by Blackadar r5-22]

for the assumption of

a linearly varying geostrophic wind.

5.41

G= Go+
4.

where WGo is the surface value of the geostrophic wind

and A is the vectorial rate of increase with height.

The

results of calculations for four different directions of

4

A are shown in Figure 5-11. The curves are calculated

forIAI= 42 x 10 5 f,

1 Wg[

84 x 105 f, Zo= 0.01m, and the

indicated heights are expressed in meters.

73

The baroclinic

5.14

C:)

C:)C)".J

:-):2 )

IL

<

CD

zuz
CD

FLL

C)LL
01L

3CZ)

-j

C)

III

C))
C)

74

C)

-C)

-4

__

____

___

____

___

____

____

___

____

.uj

0*

zU
3r.1=
9z
9X
l~z
99

0
9gu

aL

u
SF

LL

IX*f
<O

9:
a

a. I

C:,J

3c

759

CD

C)

CDJ

FOI-

C:)

Uf)

wi

z
C:)

0::

C)

LCI
LUJ

zw

I<

CD) u

a) > CF)
,.- i t-i
V.-:

Z
W
se

C:)
CN

Z
0
w

U-

L)I

I:

CD
C:C)

(Ti)

76

C
C:)

0
Ul)

-j

00

>

C
UL

LAJ

-~

r-. >

0
*j

CD

CC

C)
+g
+

77

fI

LII

I.-

-J
f-a-

C/,)

C)

-I

'-I

CD

00

C)

C)

C)

i
1-1fr-

(. I
*0
SN

U-

.4

C:)<

C)

C)

C=)

o*n/AM

C)

U-L

78

120
100

ESTIMATED FROHt
FIGURES 5-4 9
AND EQUATIONS -10 9

5 -31

80

IN

60
=i(],

(IN 2. +5.2 Z)

UWX

K<

20

40

20

0L
20

FIGURE 5-10

40

60

80

COMPARISON OF THE LOG-LINEAR VELOCITY

PROFILE WITH
THE Wli., COMPONENT PARALLEL TO THE
SURFACE WIND AS
DETERMINED FROM FIGURE 5-9,
79

0
(N

N,

0)

m.

0'0
00

=r--

<

U-

L(J
L-

'-Il

CI

uJj

r-I-

80E

4]]

wind profile can be scaled from these curves only for

the specific conditions stated.
Application of Equations 5.30, 5.31, 5.36 and 5.37
requires a different scaling velocity than WG.

Yamanda

[5-15] uses a vertically averaged geostrophic wind

h
'ref

fGz)d
I

5.42

0
Scaling velocities used by other investigators are also
reported in [5-15].
5.4
5.4.1

Boundary Layer Over Non-Homogeneous Terrain

Surface Rouqhness Transition
An internal boundary layer occurs within the atmos-

pheric boundary layer when the moving air mass passes over
a change in surface roughness.

Typically the internal

boundary layer will develop as shown in Figure 5-12.

When a change from surface roughness zol to surface
roughness zo2 occurs at location x = 0, a transition
zone grows upward and spreads outward from the discontinuity
in surface roughness.

Above region XI

shown in Figure 5-12

the vin6 profile is undisturbed from the upstream profile

which, dcpending on stabili-y conditions, is described by
Equations 5.20 through 5.24.

Below region II, the wind

speed profile may be taleri as logarithmic

81

Uo- kZ
o=
2 (x)
k

in

z +o

5.43
543

Further from the discontinuity, where z/L

within region I
becomes large, the wind speed profile may more adequately
be described for the existing diabatic conditions by the
appropriate equations given in Section 5.2.

Within

region II the wind blends gradually from the velocity

profile in region I to that in region III.

At 1-arge

distances from either side of the discontinuity in

roughness, the shear stress at the surface adjusts to
values consistent with those expected of the flow had it
developed over uniform terrain.
Numerous mathematical model and numerical solutions
of the internal boundary layer have been proposed.
Surveys of these are given in f5-23, 24, 25].
experimental data to confir

The

thuse models are somewhat

Reported data from measurements in

limited, however.

the laboratory and in the atmosphere are given in

15-26, 27, 28, 29].

These data tend to corroborate

the mathematics, but are generally not measured over

sufficient distances to verify the character of the
boundary layer at elevations where encounters with
airplanes occur.

In the absence of such experimental

data, mathematical models are used to extend the data to

heights of interest for this report.

82

It is generally agreed both experimentally and

analytically that the height of the internal boundary
layer, essentially 6m in Figure 5-12, grows as:
=M

()=

cl F7

Zo2

5.44)

o2

5.44

For neutral conditions, values of n range from 0.70 to 0.80

and values for cI are given as: 0.75 - 0.03 ln (zo2 /zol)
by Elliott [5-30] and unity by Munro and Oke [5-271.
Rao [5-313 nwerically computes for unstable flow that 6
grows as x 0

88

for L =-20m and as x

for L =-2m.

For

neutral conditions he reports n = 0.77.

Figure 5-13 illustrates the approximate growth of
the internal boundary layer, based on Equation 5.44
relative to a 3

glide slope for some typical roughness

changes near airports.

Wind shear may be expected as

aircraft pass through the internal boundary layer and

knowledge of the shape of the wind profile is required to
assess the magnitude of this shear.
Solutions of the two-dimensional flow field over a
change in surface roughness by numerical solution of either
the boundary layer or Navier-Stokes equations using
various turbulence models are given in [5-12, 32, and 34].
Typical horizontal velocity profiles at various stations
in the windward direction from the surface roughness change
are shown in Figure 5-14, and a characteristic vertical

83

REGION III

ROUGHNESS

REGION I
0

FIGURE

5-12

SCHEMATIC ILLUSTRATION OF THE DEVELOPING INTERNAL

BOUNDARY LAYER [5-351.

84

URBAN AREAFLTAD
zoo 3.5
(MILES) 0

3.0

2.0

1.0

4.0

1000-

FLAT LAND

URA__:50
UAREA50Z~.O
-n

5.0

'4.0

2.0

N 3.0

1.0

0. 05

(MILCS)

WIND 1000Z -

2
ot
2'

s6.O8

WOOD
LAND FOREST

LAT LAND
1000-

a 05

0.30

0Z

1.0 (5.'Ds U. (MIES

3,0 4.'2.0

1000WID
.22WN

6.06.0,84-842

LAT LAN
os
.

500-FLA
NS
O PEL

E T

5.0

3:

- -

(.0L5.)
1.0
WI1000WN

2.0

LES)

=50 470o

2.0

LAND

000

Z..0

0:

FLAT

S-

OPEN SEA

REATVEToA

lMIL

0L
L

SOMETYPCAL10UGN06s0.63XZ
0 2.85

-2

FOR5.
3.0ES
(D.0E
CHNGED ERAVRS

IDE SLO0

04

103

100

FIGURE 5-111

TYPICAL HORIZONTAL VELOCITY PROFILES AT GIVEN

X STATIONS DOWNWIND OF A CHANGE IN SURFACE
ROUGHNESS [5-241
86

CDN

C)%

uU.,
0

>- csj

u0

C:J

r-I

L LJ
U')
<-

IC\JC

C87

u-i

f0N,1t

velocity field is shown in Figure 5-15.

The velocity

information given in the two figures are for a rough to

Similar results are
smooth transition of In z /Z =5
obtained for smooth to rough transitions and some results
for the unstable case have also been reported.

One observes

a downward motion of the air as it passes over the change

of surface roughness.

This suggests an unexpected downdraft

may be encountered by aircraft approaching, say over a

forested area breaking into a cleared airfield.
The application of the above numerical solutions to
computer simulation of airplane flight over transitions
in surface roughness requires either on-line solution of
the complete set of equations, which is extremely costly
in computer time, or a table look-up scheme with the wind
field data stored on tapes or cards.
is readily carried out.

The latter approach

However, for quick preliminary

analyses, a simple mathematical expression for the wind

profile through the internal boundary layer would be
valuable.
Logan and Fichtl [5-351 have recently proposed such
a model for neutrally stable conditions consisting of two
bounding logarithmic layers and an intermediate velocity
defect layer.

Velocities and stress distribution functions

within region II, Figure 5-12, which meet all boundary

and matching conditions of regions I and III are given by:

88

u(z)

2
u.* (z X)
where

5.45

u*(z,x) F(n) + u

(u

n- z/6 o and 6

2-

2.4

u* 0 1 ) G(r) +

5.461

is the outer boundary of region II.

The velocity ui is given by:

ui - -

ln

6
Zo
--

where

5.47

zol

6m <z <60

defines a region in which the flow has

decelerated relative to the original equilibrium

logarithmic profile:
Wx

U~o I
and

1 ln

5.48

Zol

6i <z< 6m defines & region in which the flow has

accelerated relative to the logarithmic velocity profile:

Wx
U
o2

z
kln Zo2

5.49

By appropriate matching techniques the relationships given

below can be found:
6o- 6
-

+ )5.50

72r(ln r-(l+M)) + em ]

5.51

kx

M = In(z 0 1 /Zo 2 ) ;

89

6m/Zo2

5.52

U*o2 =U*ol

++551

5.53

in (66m/Zol )
- 01

in

5.54

--

Tese relationships are based on the functions

4

G(n)

(6o+6L- 2 z)

erf[i-L)]55

LT-5.55

(60-

where
erf(r)

f C e-C2dl
5.56

and
F(n) =

5.57

The expression relating 6i , 6m and 6

6m

is:

=(6o + 6 i) /2

5.58

The model is thus complete except that 6i must be

determined from experiment.
of Bradley [5--26]

In Reference [5-351

was used to determine

6 m*

the data

The authors

of 15-351 point out that 6m must be determined separately

90

for each analysis from experimental data having the required

value of Zol14o2.
To circumvent the requirement of individual experimental
6
results for each analysis, it is proposed that m simply
be related to x by
Om - X0 . 8 z0.2

(see Equation 5.44)

A linear expansion around 6m is assumed which gives

(1

5.60

c2) 6

5.61

6i- c2 a

The solution for the velocity profile within the internal

boundary layer is then straight forward.

With c 2 = 0.5 the

computed velocity profiles using Equations 5.52, 5.54,

5.55 and 5.57 were evaluated and are compared with Bradley's
data in Table 5-2.

91

Table 5-2
Comparison of Modified Logan and Fichtl[ 5-35] solution with
Bradle

Data

5-26:.

x - 1220 cm

MODEL
z

(Cm)

BRADLEY

u (z)/U re

u (z)/U re

100

0.96

0.97

60

0.90

0.895

42

0.87

0.86

x - 210 cm
MODEL

BRADLEY

(cm)

u(z)/Uref

31.2

0.79

0.79

20
10

0.72
0.67

0.72
0.64

The agreement is seen to be very good.

u(z)/Ure f

Thus, the modified

approach tc. the solution technique of Logan and Fichtl [5-35)

provides an easily programmable mathematical model for
computing internal boundary layer wind profiles.

Figure 5-16

illustrates computed wind profiles at various stations

downstream of a discontinuity in surface roughness.

It

appears from inspection that the wind shear is probably

negligible, however, this will be confirmed by introducing
the model into the flight dynamic computer code in a
AW

later section.
92

Wn

CD

ID

C)

S0
Ow
F-

0
U-N

ID
LLJ
LnJ
:).-

CD2:

J,0

CD

.C

ci

C
00

93o

\0
\0

0%v

U,

w
w

0N

0
0-

Nc

L0

ww

01
I

3HINI

3ds ONIM AO IN33)U3d

94

--

5.4.2

zWW7 .---

fl--

717

Other Surface Variations

The preceeding discussion considers only winds over

changes in surface roughness where the wind approaches

over a lacge uniform plane and re-establishes itself over
an equally large uniform plane.

More complex boundary

layers may develop over irregularly spaced patches of

roughness such as illustrated in Figure 5-17, see [5-36].
These types of wind fields may best be analyzed as wind
barriers for which an extensive review is given in (5-37].
When the surfaces of different roughness are also at
different temperature, recirculation can occur as in sea
breezes [5-38, 39, 40].

Sea breezes have been reported

to cause difficult landing conditions [1-9].

Since

many airfields are located near the sea or near other

large bodies of water, a model of sea breezes for use in
flight simulation studies should be developed.
The atmospheric boundary layer over "heat islands",
that is, surfaces of different heat flux, temperature and
roughness such as a large metropolitan area surrounded by
flat grass land, are currently being studied [5-24, 40, 41].
In view of the fact that many approach and departure flight
paths are over expansive cities, wind shear over heat
islands is another area which requires investigation.

95

6.0
6.1

SUMMARY

Introduction
Physical models of wind shear environments potentially

encounterable by aircraft in the terminal area have been

reviewed.

Numerous mathematical models and data sources

describing wind shear conditions as reported in the literature are surveyed.

survey.

Table 6-1 summarizes the result of the

Sections 6.2, 6.3 and 6.4 discuss the tentative

conclusions and recommendation for the three major areas of

wind shear; fronts, thunderstorms and neutral and stable
boundary layers, respectively.

Section 6.5 describes the

turbulence models to be used in establishing the primitive

turbulent wind shear profiles and Section 6.6 discusses
application of the models to simulator operations.
6.2

Fronts
The search for data sources from which mathem.rtical

models of wind shear can be established for synoptic scale

warm and cold fronts is still being carried out.

Data from

tho 500 m tower at NSSL/NOAA I,aboratories in Norman,

Oklahoma have been measured and these measurements are
currently being reduced to a format similar to that given by
Goff [4-2] for thunderstorm gust fronts, see Figure 4-10.
It is expected that these data will. be available by April,
1977 and they will be tabulated and used with a computer
lookup routine as described in Section 6.3 for thunderstorm
gust front data.
96

BEST AVAiLABLE COPY

A
414

V, I

0I

,C6,-4
VV
E.0M
0 0

41.)
:.

2)
D

E)
.-

*0

0x

>1
W4J0M
V. ri
-iMV
fo 8. 0
0) 0Ha
))
U )
MV4)

Q DO,
)V
4J 0
M (L :r U 4J
)'
nW0-0r
00
'r m l
H0 c c O0 a
,1 '.4
M>
a
M
c
(0
14J0)-4,0)
a
0)
r4JI11 4 014
00)
.0 0
10)
c a)0>-4)) c41M
000)E
-4 144)1U) 0
:0
)
it &444
4.1
14
N c
4. 14w c:
-.
4
U Ul
8I
Pt
.
4
4
>%
W)
,Vf
0
) 3t ,4
--HCna V)P
0 44 ( 4) 0
, 0
.,
D W'
0
X
)04
0
01.1
.4Il
c
0 01It 94 cr0 ) 0 - c :> j 0r 0 ) 0
ow
, EO
C)

.-

iv ) M

- V

0)10) c1 X) %iO 0 r_ 4
0t
0 N
( V) 4 M)'V0
W r -44 .
M-4 " 0
10
0 0 r0
W14 I0 M 0
a Mr-0)
>
00

.ri

a
,-

V) ")

Q ,-q3 HU)

DO0
0
(D u V)
VW 401-4

_ w
*0M.0

M 1A I0 V)

'D )4J
o0
41'.rn E

41

0'0
1041
0d

14-M)-

J1

'
L
14 '0
4
n1
.,l4
UC)"

4 -

*.4

1-

v.iv v,.
4o Ma, .,,

14

4
) 0

to0.
)0)

0
00) 3
u9
14
01U0
a) 0.11
0
:1
0
(a
4) )
H.
0 4).014
-1 .,8
w m +j V
1.a)m4MV
0)

.0 r: ro.,I
M'-IHu

., 14F14
o r
)4W04'
W CO
0
c
0 4.1
(v)
ty -4
01
0)
0.0il)
0 41
1,01
ivtMMJc004

m
40)()
,(P
0U

r-O
0
4JC)
0

IU 4J CU0
1.4)d)M

00

97

CU t:3

)
Oc
,
0400
.,1
0141
W
0- 4)
0
c.
V 'l >
"
44 0 r.4-4
.,0-.4
0..'
O 0

4u

.
,
>,a)
o (v :r
, , m g0.4
T
-4 a
c ) vc.o4.
.
)4
.40 t
-4 4) Ln j
H4X
C~ c .. H k
0
3)
M(
0))
ON
40
0
41 v 4 z 0 M r : I
41M
.40,0
A
Oc)
M k0
0U
C)
, ,
r
)9
010
m4 ) c
0 a) 4.)4.100"-4
r,. 0 Mc ) ;C-4 N
N.1
N
14
*414
H
_ 0w 0)
.0 0.0' 0ON . . *hi
4)0
V
V M r,044
M
.14 cu )
.4.
0/ c:411
,
A
4
1
0:
J
(Ul
4
V
l
iji
c
0n
)
r:
.
ww141
a
:
fa
W .- 0 44I
c
-4" r. 4 1
a). 0
0--' 0 M
H0 ),c 40 0
) M
U0 H -1
*>1
14 rq0
4 0)
) .0
r
44 0 ) 0 U 0 . 4 ". V) )
0

:3

900U E M

Wc
01..V0

.1:
4'
00)V)0)
U .,q
0,44101M
041 ) 0) w

eu Z V

w 4

a> ) u41 V '

)
0 ) r )1
CI.U

In
:)"4 .14
r-)4 -4 U t -4
10
.))>140W
::34JVC) r r )
4
CM) C>
044
. 40J
Vd40
UE
H
)
-4
:9-4
3. 1 0 1 04
a.000.)
(0 VC0).C'0
04.14.-1
0 C.
5; 4-4 -- 1 0)0

~4

L_ W
>1
.~

*'.0 10
,
0)
j
t4n0)c
T r0
4 4J

V44'0
4 44010
0
0197 M
4)r)
4J '-4V
440
-4
ri X0
O) V0w )
4 4)00M . 0
40V)0.4 w
0 -4 J
4J rj) 3)
.40 U1)
- 0'.- :0 U
-4 F r,-4
m4
VE11) U ) a-4 t) :3
0Ou
0)
V0) 30-4)CA
N 0 ) 14 4) (1)
0 r.11
0
r_ 1 r H-a oV0 M
>1 0 .11V 0))r_
E 0
- 4J V0-4 4,0
M400
M
.- 0) 4 O>
0
M
M
PU0EW-I
Q) r- )rM
" uM >)0 -0
e0).
a0M
1 A
E
U
FzC
W LAW14
w
0) Q)1V)
0)..'414.
W0.10 113Q0)
40)1 )
0)-1
a1*4
0)1-10'U14P X0..
)11
Il 0)4 toZ )
'000Q.-I
41
M W00
14I O0 M0
j 0)C
C0)4
r
0) -0 r. MM M 0C:
4011
0
A4)0U.00
140
0)
(
V0, r tC.0 ) 0 000Ln
(13IP )
1
.4cM nd".71
0)
0.4 ) H10) r.4
) .' -u
-10
-0 0140
)0c-11.,C
r:c 4 4)
f. 14 r - E
r . V.4E 4 0
M 1u'-M)EI

) 1. 144 1
X0) t" ) r)

H M 44 V M0 ON 4' 4J WI

c c10

15IAZI-4A
O.
.- 44
0-

O W 11-4) 4- 0 UO r m
40
*.
a g0
A
WM 0 X X
O 1 4)
0
*'i>
1)0.0) H
0)0)
M11~
Ol
)
164H0U04)
E0. Q. M
)C
0I'1 H
CL W -H

10 w UwwF10>
000) 0 *.4'4 41- n
'dM.1i
0 U
-,
0) 1_
00c)
4
'0 1 0 to >
) *'
0 -1-140)0
-)1(04
0)4.4 r 0UOr .
En1',0
-4 On

0 1

c. N'oooo
p4
0-

1 4 .w A E 4)0
f >,u I .d,. o
0o

0
14-4EC.
m
0) 1
C
W40

0M 0 -4
4
0- 4,4 a
4
41 , w .
,4
.,v
A4) C.
W.0q))o"4004
W0
Old $40a)
$4 H

~i

1
I"'i)
( 10 1 44
1 ))0 -4

.V
)

0n
CL0
0-4--0
4(0C)
V)0-I
,0
1
0 )
4)
Q)'
.-M 0)0
>1 u U) D
104(a
(a14 L00
OUV0F_ M
04I)
V)IC1
(1H rt .40In..0)0
W rV1
ITV. 4j -4a0)0(,1
lo V 0)4) k
.0)H. $4
0))(A4V0C
0
fi- )
0
)1UL U
:9 4) 0 :l 0)M)0V
2'.
0))
c:
0~ :3) OH
M.10w E M c10
:)4L"
0 >
40
)
.) 4r A0'L0. ) 4
V)
lo 44.0
0 0
"
m MJ
l
4 r [U) 1 .0
.14 1).4t.0.c)a)
. M 4)
0Pcw >>140C)
1j
c: 0I.
W440(DC
(V)
0
0 41
c
r~
0
0
.Ha
0
0 *.'4
0
.,1
c0, 1t
0.) 0 ()40-4
00)40
10V4
L0)a0V0.'4
M MM
1 fa u
dr4)V
Mr
1
, 0-000rM4
f 4)) c t-0))1
0 O
4 r1
t
4 C0M M
1
).41 01.:)
A)0 M 0
01
0040)1V)
A
)0140
C4 -04C
C ( 0
0 P
-4_0) (0
40 .
0
1
W0) 14 0
)0.aF
Hr
)'0.4lC
w0.~ I)H 103
0)0)14M
)4)an
00)
)0l00)
r-4
) 4 )F
(L '00
-)
M a
fu
fl)'E 0. w 41
E.
d0
E
4J ( E
c ).-4 4)

r.l14

o0

) , 10U" w
M0)
k k

.,4 N"0)v.4

1"

.t
- ?C)

, .'i
C
I
I: A Ito' 41 t
0cl u
* 4.1m,-

40
.(A
0
W0
0i ) $4
u c
4- 0
-4/ O :(a04
M
$- 1.4
W r. M .4wM 0I
V)
n
)
.,-.
:*.
94 -4
4)
CO
,-4)
-3 t7
a- 14 0 0) 1.
ar
e0X
,
-.H4
.
0)a0)0 4)
0C.
..
'
)1 C-.0 C
14-)
0
0
.E
. 'a0
0
4.
0).,-

:3 -40)

c0'0.

(D 0)
0) c

(a0 u
a)4
H

0)4>5
U14
(a
01 M
P
1)
0 0.11)0.,

U)

0 0)

4) -10

H)~E
W4'-4

4 11 s 0-,
44 :.,-I
e).
$4 c

014
ra 0J V
k
r.1 a.. 0
VM w 0i a.. 0
0
44
0 H
0 4MX:
01
P
0'd'c
>
0
1 0 -4 W
E) 0 0)
-40) .10,0100) w r
0141W"-4
>,104 ) ) C).

v fO r
,0
,414
0.
*)
A0
(a
)

U)

0 4)1u,,
m
. a0"4.4
0 4)44 D
t7%9
fa :)a
0
E

G)

ami
it v
:I M
Pi
V0)w
u:,
. 19
4J n
>4 -4)

U) l

MVm V

Ur. .t
a M-4 0
.11
c '1A .,
H144
,+ 4J
J
41 a' (a.
0
0 . M
4"2,0
k 0
3 0
-. 4 .0 0 E 0
>
..00
0 Ot
X
. 0
V0 ,
a "
0 4 ,"4)V
1 )
0
1
,40)
0.--4
J.00'-4,,-0
44, 0.- )0

-_iv" >
00
40)
M.
.c
.. U4

M)0'A)
> -40

4I1
0)

E) 44
,

: -4

141M

H &. Q.4)10
) 4) .0
w

.0

o4)0

a44"
00
'

I.
4)

4) C)
:3
A:l
n0I.

M4)

w10 -4 0.

>4

...

NU04"'4 wnU

ld
j

4)'UU

44

1P=
', 01 0 0
1u14j'.'tr L,,-0
E 0
1.444 c

,
0

0I

uA

l 4

Old

4-ne

77

6.3

Thunderstorm Gust Front

The most extensive thunderstorm gust front data for

formulating detailed wind speed profiles are those measured

by Goff 14-2] in the form of cross sections of lines of
constant vertical, Wz, longitudinal, WX, and lateral, Wy
wind speed components through the thunderstorm parallel to
its mean motion.

Tabulating these data on a grid and

utilizing a computer lookup routine allows the wind speed to

be predicted at any horizontal and vertical position called
for by the simulator control program.
The range of data is 500 m in vertical and 3,321 m11,685 m in horizontal extent, depending on the data setconsidered.

Twenty data sets are availablc and have been

punched on computer cards for the table-lookup routine.

Figure 6-1 illustrates the grid system superimposed on the
wind field.

Tabulated values of wind speed for the 20

thunderstorm gust front cases are given in the appendix in

Tables A-1 to A-20.
The vertical grid spacing is the same for all tables,
however, the horizontal grid spacing varies according to the
length of record available.

The vertical grid spacing DZ

and horizontal grid spacing DX are listed in meters at the

upper right=hand corner of the table.

Each data set has 11

nodes in the vertical and 41 nodes in the horizontal; not all

nodes are shown on Figure 6-1.

Thus for example, the horizontal

spatial extent of the record in Table A-1 is 40 x 100 m = 4000 m

whereas in Table A-7 it is 40 x 196 m = 7840 m.

98

____________________________________

'*L

rq

tir

"4

caL

ri

.0
fn -43
00
-I

0%

*v

LD

S~~~u
NNN~r
-44'-l

00 0

000

0!

9S4-

0:

<

0il1%
f
3~3--.3*-

'-4

:,I a 11D

~.,99

9-

I.

Figure 6-2 shows the flight path of an airplane

characteristic of a DC-8 landing with fixed controls through
the thunderstorm.

Figure 6-3 gives the wind speed encountered

by the airplane during descent.

The wind speeds encountered

are seen to have a very complex profile indicating that

mathematical formulations of gust fronts which express the
wind speed only as a function of elevation, such as Equations
4.9, 4.10, and 4.4 through 4.7 cannot provide realistic
simulations.
In principle, three-dimensional wind fields are
required to give meaningful simulations, but these data are
not available.

Thus, simulation of flight across the

direction of travel of the storm is not at this time

possible.

Note that all cross section given by Goff [4-2]

are in planes parallel to the direction of travel of the

storm and any simulation employing these data assumes the
aircraft is flying either toward or away from the storm
along the direction of its mean motion.

Work is continuing

to provide estimates of three-dimensional gust fronts from

Goff's data [4-2].
Developing the statistics to provide risk of exceedance
criteria promises considerable difficulty.

The approach

envisioned at this time will first attempt tc establish a

common point on all 20 storm cases such as the location of
the gust front.

Second, the mean wind speed for all 20

cases will be computed at corresponding grid points relative

to the common point.

Third, the standard devi-..ion about

100

m
Z

LLO
00-

LL 0-

00%
CD
ILO
WC-C

101

4.0

STARTING POSITION

3.0

2.0

2>

Va

70mt/s

ha=91 m
-1.0

-0.2

-0.1
HEAD WIND! UPDRAFT

FIGURE 6-3

WIND 'SEEN" BY DC-8 LANDING WITH

FIXED CONTROLS IN CASE

102

9 THUNDERSTORM

TAILWIND/ DOWNDRAFT

the mean with the assumption of normality will be used to

estimate the probability of exceeding A prescribed wind

speed or wind shear at each grid point.
6.4

Atmospheric Boundary Layer

The data of Clarke and Hess (5-163 from Figures 5-5 and

5-6 have been tabulated in Table 6-2 as a function of

dimensionless height 2 - fz/u, and stability parameter
V- ku,/fL.

Coupled with a computer lookup scheme the

longitudinal, Wx , and lateral, Wy , components of wind speed

can be determined for dimensionless heights of 0 < 2 < 0.15
and for given atmospheric stability conditions of -300 <
v< 200.

The range in physical height, z, of the data is

dependent on tLe latitude through the Coriolis parameter, f,

and the friction velocity, u,.
10-4s

- I,

For typical values of u,,

z ranges from 0 to 750 m.

Wind speed profiles generated by the computer lookup

scheme are shown in Figures 6-4 and 6-5.

Note that the

strongest speed shear occurs under very stable conditions,

U = 200, but the strongest directional shear occurs under
intermediately stable conditions, 1i= 50.
Actually, the strongest directional shear occurs under
baroclinic conditions.

A computer code for computing wind

speed profiles under baroclinic conditions where the

synoptic pressure gradients vary linearly with height (5-22]

is programmed.

Realistic magnitudes for the rates of varia-

tion, however, are being justified before meaningful wind

speed profiles can be presented for simulation work.
103

I.

0"

o.

NO.

0
Pb

p
~.

..o

a.
NO-

oN

.0
.

"

.2

-" o"0 ,.,00 .

00

Z)

t. - 4

Po

1)

N C4

r4

Pb
A Pb

0
'I
N

00
.1~.

.0

0y

*"

C!

T
N
10
N

NQ

. C?~ .
N
N
4
Pb
04P
N

a.
0q
Pb
~k, ~~
N"N

s.~

Ny

N"o

Pb

Nv

Nv

M4

n
Pb
N

Pb
N

Pb
N

0
.

IN
0

r4
0.

0.-

N n

, N o

No

0,
4

p.
o
P.

rk

NA

N2

N
:

Nb

a.,.

n
N

0.

0
.

.
0

..
Pb

zb

NNN

o.

m0

In
0
N

N4

0
.

0
.
b

P.

4.

. S P
.Pb ..

.o
.

Pb

..

.
*4o

0
fft

0"
in

a
N

0
el

0p
P%.
.0

,
Pb
1:o
o 1!
0C000.

In
0

4b

b
N

.7

P:b

Pb

4n

b
7
m
0..

N04

.
:

0a
..

.
P

NN

..
P P.

0p
rl.
4

0
A
N4

.4

.4

N
-a

P-

.N.

5.

4.

4.

a.

o
4.

.
N

m0

o.

4- 4

Pb

.4

Pb
4

o
op.
4

a
Pb

0
N4

.
0

.0

-.y

t-

0
P

N:

N4,, ,. 0

4o

t-0

N, ,,. .

*,

N .4

o
N
4

r,p

P -

T ..
-! .
.
N
N
t
r
N
0. 04
0
0
0

H
S0

b ..

-a

NC,#

0
.

o o o o o o o o :: :a

.0.H

r%
Pb
. 0%
.

E-Q

fn

N%
0.

N4

, .. ,

N4

.
.4

13 Z

4v

0o
P

.1 o. o. o. .4'! CP
'1 No

"

00

. *
.

NC
'

.
0
U

S.
.

N P.

Pa

0)W,

aP

r. 0.

. a
,,
1
N
o
UP. .0 P R

*4

*b PS 0
*

N
.

Eo

U.

Pb 0

oHbN

Nt

in

0
.
N

.4a
.. 4..
N

N
C;
N

w
.
N

.
N

fmP
4
.
b

S
P

-~~~T

7M--

400

0
0

so-

100

50

100

~-

200 -

so-

5-

0,

345

Wy (m/sec)
W

FIGURE

6.5

LATERAL_ WIND COMPONENT

106

Horizontal wind $hear for the atmospheric boundary

layer models is not being considered in this program.
Figure 5-16 suggests that horizontal shear resulting from
terrain features diffuse rather rapidly and should not cause
significant effect above elevations of approximately 25 m.
Under high wind conditions, however, these low level wind
shears due to vortices or recirculation zones near buildings
and other obstructions to the surface wind may create
hazardous flight conditions.

This is an area where a follow

on effort to establish mathematical models of hazardous

wind conditions for simulation of flight near building is
needed.
Risk of exceedance predictions for the-atmospneric
boundary layer require establishing the probability of a
given value of p.

Figure 6-6 from [5-161 provides some

insight in this regard,

The figure shows a 40 day average

of the daily variation of p with two standard deviation

error bands.

Assuming a Gaussian distribution, the

probability of exceeding a given value of v can be estimated.

This curve is, of course, highly sensitive to location and
further research is required to establish a more general
daily variation in atmospheric stability.
6.5

Turbulence Models
The primitive turbulent wind speed profiles will be

generated by superimposing a randomly generated turbulence

signal on the steady-state wind speed.
107

The review of the

100-

50

0-

-50-

-100SUNP 1ISE

SLJ1 SET

10

12

14

LOCAL APPARENT TIME OF DAY

FIGURE 6-6

VARIATION OF

PWITH

108

TIME OF DAY

16
(HR)

[5-16).

18

20

22

24

literature pertainin.j to turbulence model is not included in

this interim report, hawever, the reader is referred to [6-1
through 7]

for information on current models.

The development of a realistic turbulence model depends

primarily on knowledge of the appropriate turbuJence energy
spectra.

For atmospheric boundary layers, well established

turbulence spectra are given in [5-2, 5, 6 and 6-8] and

turbulence simulation in these cases is straightforward.
For turbulence simulation in thunderstorms, spectra are
still in the development stage.

In fact, thunderstorm

turbulence is neither statistically stationary nor homogeneous and superimposing a turbulent component of wind on
the quasi-steady state values given in Tables A-1 through
A-20 may not be realistic.

However, in lieu of a more

reliable method, this approach will be employed until the

state of the art in thunderstorm turbulence modelling is
improved.
6.6

Fliqht Simulator Applications

Mathematical models for programming simulators to

respond to fronts are not ready at this time.

Models for

steady-state thunderstorm gust fronts, as described in

Section 6.3, and for atmospheric boundary layers, as
described in Section 6.4, are available as computer lookup
routines which are ready for immediate application.

Addi-

tional work is required to increase the sophistication of

the models to include turbulence and a statement relative to
the risk of encountering a specific wind shear.
109

The thunderstorm gust front model consists of 20

separate data sets.

The suggested application of these data

to programs for evaluating visual displays, for testing

avionics, for verifying operational procedures or for
designing control systems where a pilot poll is taken as the
evaluation criteria, is as follows.

Each pilot should "fly"

as many of the wind shear cases as desired, say, 10 to 20

cases, in a prescribed order.

The pilots would be unable to

learn the wind shear encountered as each one would be

different, however, the experimental sampling would be
consistent in that every member of the team would experience
the exact same sequence of wind shears.
A similar procedure can be carried out'with the atinospheric boundary layer model.

In this case, a matrix of p

values would be established prior to the experiment and the

simulator would be programmed to accept these in a specified
sequence.

The selection of values for p can be guided by

the data in Figure 6-6.

It should be cautioned that the present data decks are
not likely to include the extreme wind shear that might be
encountered once in a lifetime.

The ability to simulate

extremes is contingent on developing the risk of exceedance

statistics which are currently under investigation.
Real time simulations are easily achieved with the
computer lookup routines since interpolation between tabulated values requires only milliseconds of machine time.
Moreover, data card decks are easily incorporated into
110

simulator program logic [6-9, 10] and thus facilitate

application to existing simulator programs.
6.7

Conclusion
Efforts are continuing to develop simulation models of

detailed wind shear profiles in large-scale frontal motions,

to establish meaningful turbulence spectra for thunderstorms
and to establish risk of exceedance criteria based on valid
statistics of extreme wind shears.

Models of quasi steady-

state wind shear for thunderstorm gust fronts and atmospheric

boundary layers are completed and ready for immediate
application in wind shear hazard/flight simulation studies.

i1

APPENDIX A

112

.4
N
9

.0

*
a~~~~~~
-

0.
N

.44

4
14

-4.
4

.4

0N1;

414

0
4

10

'41
1*.4

1 '4

ff

.4

f-N4 0

0
.

11

00

0-

m0

t-N.

lot
4

o46
1

.40

.40

00

o~I
I'

4
4.
18
.

.444

*4

&4

C.

44

.
a

.4
4

14

.44

N
a

:
A
N,

Ch

0.

1:

1
0

4@

N1.
o

0.

N4

0
N

1., 4
0
-04

140

Is

No

N61

f-'4 .0
Nt

inU
6

qa
6
*

ON

4-

6N

14 N

14

0N

1N

0N

40

".,q

"4

6N N

aC;

- -

o ano

.4
.0

e. N 9)

C4

g0
0

0.
~z4

H
o

'

0.
4

01

14

N
0

0
0

14

O
.

'4

I
14

1
T
.N

N
4

'

a
4

0N

0
.

.4
-

I.U

I4

.4

-4N
4

0%
.

.4

r4

.4
.4

1q
4

U 66666

.4

44

a
0

a, to
0
0

a- D
1
4

t~

af
0
0

fn
N

ft
1

'0

'0

'O

.4

.4-113

Ift

t4ttv 0%

It
pt
0%

40

0%

0
0

0%

0%
%

0%
0%

04

. t
%

N 0

0.
On
%

*4 0.
M%
%N

W%
44 0%1C

0%*
*

0*

*:
t-4

t-

0%

Q,

f-

It

a
U.

on
N

4;*
-

0 m

0
.

ID to

A 0%

0a a a a
1)-04%

40
a
.40U.

N
a

*
N

f
%

m
.

N
.

4%

r-

aa

0%

N
4

N
4

04

4a 0

ol
*J
N

Ft
4

ft

N4
04

0%

N0

.4n

Pd

N 0% N

C!
0

00

00

P.

0%

.0

so

1.0

Ci

0
0

0%

C;

0 0 0

0%
4

P 0

4%

0a

0'

a a

tU
N

.40

4~

ft

0
1*

in

lf

0
-t

a
.

a,0 0%NS

fm
0

In N ONN0%
N

t-

.0

ft

10
4v

PN

a~ a
P.: 0%

0 Pt

ey

44
%

N 0

%0

oy
00

P.

11

N0

0.

NO

0
a

000

CP~

4 C
.4

..

.;

o~
0

i C)

4;

r-~

U4

*V

.*

1;
a 0

on

r4

4
9A

C) 0

a0

C)

04

C.

C.

0 c;00

0
.

(A

0
N

0N

go

I-

t0
-

C).0

.4

.04
)

t;

09
C-

a,N

1 0l

a4)

69
.

N;4

C
a%0

N .0

a
*n

oa

90

0 C.0

4
.4

*1F

t0)

4
1.

0
4

.4

0:

TF

1)
Q

C .4

C.

*............

.4.11)5

0)

14

4N

N
I

44

*4

0'~

C.

"

.4

.1

0 o

. *.
. .
0

N.

d%

.*

0.01.
0

.4

.'

0'

.o

W*

E-

hI

Nam0
n

0'n

'

'

'

'

40

o4 on'

00

.0

0'.4

04

a)

'

.4

.
a 0'.

440

0
0

~4

.z

1191

-44

00
0

.4

.4

.401160

o o
04

El.4

0
*.
..

0'

. 4

I4n
0

17

'
00

a 0

n n

0
N

a
)

C0

a
v

0
.

t
1

0
N

E
n N

o .4

0-N

t
N

"0.
P,
F,

Pt

.0

.0

0-4

a
P

.0

.0

)CA0

.o

'
0

%aa

.4
St

6-,
t

Pt

NN

.4*

6'

.N

Pt
.

04
"4

tP-

o.

tn

m
,N ,t

a1
.

InI

.)

P.

,-P

VII
N4

",4
0*"

N
P4

('4
.4

Pt

-t

....

N
IV

N-

:N

a)
N4

In
N

Pt

Pt

'

In

..

t
4

.
C)4
P

t.

.4

t4.4

54

#4

0'

4N

..

a.4

.4

F-

.. C

0 C

I4

I4

P.4

4 u

..

In

Nt

PI

r0

00

.4

.0

)0

C!
B
NI

C;

CA
.4

a.,

"1

11,

0
C

C. 0

n4

....
.

0aC

o)

.4

..
a

C!
C

. .
Pt

.,

.
0

o) o
0

C
0)

04

04

o4

.CA

04
4

oN

CA

04
.

C)

.
0~

.0

.
0~

.4

II

*4
04

C)
,

0.

N. . .

4 C

C;

C)

C)

..

4
09 .,1
0

N
4

C)

4 c) C

0
.-.

I4

V4
C

4
4

C)

V4Pt
N.4

0C)

E)
.
.

c;

. .
0

.4*...

-t ,, .4

m4.

1.

Pt

P-

..

4 N

....

004
.

111
4

~4 N

-0
.

C)

.tN0 U%

a)

54

,. .o ,. .N)

Pt
4

44)

40

-3

N;

PN

I .

mC
,

N )

in
Pt

..

. C4 .

C,

..

'd

ID0
.0

'

N0
N

0
.

Pt

..................
~ ~

NV

9'..

PCYP

"

i0nt

%
t
.4

Pt

.0.

14
Pt

fm.. .

N
4

In

.
,6

I!'

Pt

4.t
Pt0

0 t.

N,
...4

.t

In

Pt
.

Nv

.0

,- ,;O N ,,- .

-Or,

..03

4
N
* F-

N-

N.

.4

r-

C,

6.

%6

Q*Pt

........

..

o ,.,

Nl

.4

.......

.tCo.P
0C).
.4.t.P

.......

E.4

*
N

C
0
O

Pt

o"

.a
0
4

.4

.6
D
C

t)

4-

In
)

4.4

)117

In

.0

0~~~p
oAn t

tA
N

.4
in

A.

af

N4

N4.
-4

4.

0.

0.

44.4

00

w*

~~ ~~~~~
0.0
A

-A

.0

04
A

Up0

a' 0

C)

P.

An

4-

in

000

*f

I-.

.1"

a Na

00

t-

~0

0F-

ft0~
44
'.40

00t

1t

44

PA

4n

*~
0

~~~~~~~

400

01

of

A~NIn
-

40A

..

n4

1441

0000

SC A
NN
A

0.

.4

A
.q
N

A
N
.40
4n

If4

0
o%
.;

*:
N

01
*4

a4

04

&

0 0.

.4

44
#

0,

00

C;

A*

.
C
0

0.*
*

0N
N1

s-

o4

t-0

-a

n
0v

1;

0
0

.
C;

on

.4

SH

d%
0%

C) '

14

NOa
0

1;*

0.C4

0q

An

4 0 C;
.

t-

. 4

At

N
1- 0.

*
N

t4

0.A-

.0.4

a0
N

o
4

00

.4

%D
A4f

~~~~~~
0.

Ao

0
.4

w 0

.4

mU

In

.4
-

00
00

a0a

0A
N

04

0
In*;

C.
0
7

-0

n.1

001

.4

A~
#4

An
in

NA

61

0
0

a4..4.

U,
oft

41

.44

'

01018C

.44

Ut.N

0I

04

0a

o0

0
.

C)

0 04

.
a

in4

N U
. 0

0U

0.U

ow
4

*n

V%

44%
8-U

In
a

0
A-

O
U% M.40

--

--- o4
-

NI

04

0,

0*

)
4

0
04

44

N
0

o4~

4
9U

4.

444

4A

tv

0l

0y

C;)

%-

CY

.0N
eq0.
0
0.

~
fn4

m
.r

N
U

.)

.o

(v0
4

U1

*)

ly

ry04
4

o.Y

0l

4t

m4

004

4
0

-4

f:

t0

.
.

0..

r-4

N
.

0
0

0.

0
.

.44

r4.

N
go-

0
0

.0

0
01

44

.4

04
.4

.4

0I

n4

UI

A .N
0

N4

A4
4

U
.4
C:
0

0.

CD
0

4)

8D8-

'a.

C.0

Nl

0
U
Q

0
0.

4a

0
0

%
N

1.,0U

0
C:

01

4-*
44

)0

*0.

0
0 C.

. C-00

*.

t.

*
4n0

S
0

N-

V%
N

41194

00.

fn

.4

t*

44

N
a
N4

0%
4

od

4%

0%

.0

0%

on

.4

A4

ow

.04
4

U.

CD
b4
9~~

0
0

C;4.

0%

4%
5 4% 5
0
%

0
*4

.4
ft

0%
A

N,

.
0

r,

..

00

0 0 0

00

U:

0AN
4

0;4

-D

.4

40%

4 0.

04

r*

N
0

0
0

04

0.

44

0.

0
.

..

48
4%

00

A 0

4a

U.i

00

L%

soN

A,

so

4u

.40 in%

t-

do

t44

.0

in

In
U.0

UM

U.

b4%
U.

OD

U.

U.
%

41 4

U.U.

.4

in

N0

120

NC)0

.
0

C,

0N

N
.4

U
U.

.4

6U

N4

N
U

.0

0
0

0
0

4% 4 14

00a
a a

00

.4

N%

a4

ND

00

t4

55

aN

0%
N

fq

a4
N

0.11

0
4
4

.0

0
*

9%
.

.
0

0. a. 0a a

a a

NO0%

0;4

on4

'~~~
4

0%

.
U

0)

W%
00

0.

0
;N.0

4 N

TI

71

471

*4

f
4-

4o

.4

0o

71

-4

9 Q
00

99

9
N

No

19191*
M 4 N

.4

0*-

a4
N

ta

NIN

~~~~
4
411
*0
~~~~

8~9
El~
1

9N

.4

9w

00

Nq

No

.4

lNO
0

4 0

.0

.N

N
o
.4a0

19

7N
00

01

..
0
.

0
4

*4

9f9E~

a19a9

a0
000

*
.4.

.4 !0.O

7 -4

0
9

1N

ot

1*
-

7
N

o
4

N4

00000

9fm
1

91

r-

0.4.
91

d%

.4.4

*~~

..

011

.4

a
9

40

.4

m
.4 ..

04

.4.4o
.4
11

OP

91l

w
.14

49

P.

0.4
P4

4
N

o
4

91~~
4N)

km

.1

.4

v%44.

n9*

.
AN

4144
0

91 1 0

.91

4.4

P.

121

Nn

M0

o 0

00

c
4

a
0

p0o

0
044.

oo

c;

o1

ol0o

*4
o40

.4

44.

4
.
. . N

.4

in

Nn

A03

H4

a.

19

0.

D4

09

A4

4% 49

0%
0

00

.f0 % A

1)

9%

9%
n%

0 9

.0%

Vt

0
4

9.

4n4
.%
%

C34%
1

44

9.

.%

%
.

9%

4.

4ID

.4

9.

9.

.4

.
4.

9.

44

.40
.

~~~~~

4-

N1
I

44

*p
4
04.4
0

4r
9

440

N .4

t4

.4

m
N

.4o

9 0

Nl

0,
4

4
.

.4
0

.
4

04
4

Nf

03

-4

0%

.4

a4

aN
N

4a

.4

p
4

9.

0
0

o
0

6
9

9.
C;

o
4

0a

0 0

NO
0

0N

0 la

o44

46

0-

.9

4%

4%

9%

4N

'

.4

.0

I-

09

4%
%

.9

44

4D t4

N
4.

9w

P4

n*

.4

%
-.

.4

49

00

9
4r!

00

0 .4
0

*Q

* .
9

el9
.

.4%

0
.

.40

)
4

9.
in

00

.~

46

40

a%

Il

A1

0-4 In.
N
.

N~ 9%

.440

a%

.%
-44N

00

3a

04

00

40

*122

)0

4f

n4

09

N40 N0

94

m%

mE

*4

#490
94

9p

'a

on9

in4

94

o9% 4

9%
4

9%

N.
.

t-

P.

9%0

C%
NI

fn

P.

.444

%4

.4
n

VP.
.

in
0n

44

In

N.4

m04

0
)

0%

a%

'I

.f

40

.4

NC

9 0

C:

.%4
4

L)

C,

S
00
4

f 0

N0

.40

s09

N
I

.4

V,4

-.

4-

~~
.4
.0

.4
04

0
(P.%
9

-.

ry
0.N

n
4

P.

04

0
0

W%
N 49%

m
9

4a

.4009 44

4
-4

.
N a)

9%
9. 4

V.0

044 ID

N0

0 @000

N-

V.,

~~

4 yr

149

00

0
%

49

aol~

4.~

N9

%
4

9% 9

469 0.
9

in
4.

4.

...

F%

6%
P

4%

. 0%

all

ePC

.40
in

P.

9.% 0.44
.4i

9%9%4

4
% 94 N%

9%.4.40100 4N

9%#

94

94

9%

m9f

'?
*~~r

61123

'

00U

vi

%D

In

In

vi

vi

4.

.0

10.

ON

0i

i N

4
N

*
N

.
N

14 N

.44

.4

;.

.4

4
30

-.

1**
*4

.4

.4vi

0n
Ni

Nq

14

.40.

0v

..

i0

.
0

in-

~~~
i

.4

.
"0

14

04

iv

ei
vi~
vi

.%
4

1:
4

4v

.4

In

viv
0

*N

fn

vi

-ti

~~ IZ0

vi
0

C.
N

0
0.

tv

.1

vi

Li

.0
n

Pt

~~

'0

.1)

P;

1:.

N
v

0i

Nt

Nw40
t4

.4

t
.4

64

*4
.0

40

do

.)

0.0
N

0
4
44
o4

4
-o~

..

t-.... .......

m)
t

4N)

N'

Ai

C;

C;

N;

9
o4

C!.
o)

0 C;
.

o4

o)

a a
.

C4

Y 4

0*
0

.0

4 0
4!

.
0

C).0

0)

0~

.4

.4

*
v

0
nt

Ca
.4

1Nv

P.
4

vi.

04o r.
In.
.

on

44

f.

Ut
1 .

P4

Pt

.
. 1
N0
fN
0

~~~~~~~~~~
0
N0
4

on
4

N~

vi

4 00

.a

C)

a12a4

C.

ilk

a.

0.

*4

0.
A

a-

A AA

*
A0

4
A

A.

*9r4

0 a
Vo

N 4N

.4

A0
C

.4

ON N

44 14

0.4

-0

0'

04A

444

A 4

.4

a.

*0

, 044

N*

D 00
*
*

03

"190

94

'A 9

44
q

o9

ANN

a
-

c;9 0

A4.4

*~~~~ 0:
9
9
a4,

~a.9..

0. 0in
0
9q

97

IL

so

.4~r

f-

00

o
A00

IV

.9

.4

t-

0
0

A
0

00

.
00

N
0 0

aNO

00C

34

30

In

na

0A.

AN

b-

0.
.0
4

"1 0.
in

0-.
l

A.

.
A

P4

..

.4 1:
4.
V44,4.

9-

1- 44

m44 0

33
9

1!

.4
'-

AA

9-

1o

0
0

0o

,4

9WNN9N
:A

0.

4U

0
A4

4.

00.

-A
0
0m 9
.4.4 .4
000

.0

p0

44

V4
NO a

a..

00N,

No

A404

*
10,

N04

N .9

AA

~~00
.0
W

A.

A0

044. 04

09

of

.4

A0

fm
0

N
0
or*

.4

044
z
0

0
V,

0.0

40.0
0
0:

oI

0' 0

C
0

0
4

14
0

*125

0
0.

00

C)

0N

1
on

0a

.0 4 N

44

0% 0

.4

0a

.4

1-

*
0 .4

a
c;
4
%

.%

.4

1;
.

*
N

T*

04

.4

0N

41

f4

4%l

o%

C.

C4

C3

.
0:
.

.4

0
w

94 C)

~0

N:

C0
0

AN

0 On

U-

40
1

N1

.0 0

0.

.0

410

*
0D

On

9.
ONO

I~.

.0 Ol

c;

09

'4

V
I4.

.4

1
p

0.

0
0

0 C

4%

to
0
0

0;
4

On
0
N

N
0

C
C)0

q
04 0
%
N

.4

C34

C%

.4

U
0

4
0

c.

0)

0
0

04

C*

CI
*

)
I

0
I

0
I

0
I

a0

%N

r 0% N
0%

00
1

-.

o)

40

C%

0
I

4.4

On

4%.
-0

0
)

0
0

0%

0
.

N.
.

t0

r:

43

f4
00

a.

.
4
4

o.

0a

*y

0.wo

U
0,

-0

0,

4.

0
a0

*
N*
4

4%

C4
4

U0.

*
0

a.

.0

*0

4T

10.

4%

0o'
0

a
4

4)

D4
4
0

N0

0)4

r
0

oo

c:

C!
4

o 0

-4

ON
0

0%

i-i

N)

o4
4

0
U

-.

ilia

.4

.4

.4
o

% 0

0.4

Nl 0;

0000000

C)

.4~~~
44%

4 N

4 00

4
.4
0
.4

t
0

.
0
.94

IV

N%

04
.

0N

'4

0
.

o4

4
0

%.

. 0

4
N

.4

On o

4N

4 N
0%
Na%
N

10
4
4

p.

0n
0
o

ON

-4

9126

77

11

7 ~

4D

4 4 4

44ft

.40

C;

ob

N4

n in

0 0

00

oN
0 U

on

in

.0

4 4

V)

go0
0

do
4

0.

ft

lo4

N
.0

.4

4
~ ~~

40

gq

an

.4

.4

.4

04

4
.

.
In

4*.

so4
a

4-N

04

C;

04

Pb

Pb

4a

0;

.4

n ..

4.

.4

Pbo44a

4)

fl4
.

Z4

01

.4

4
C,

41

4)

.4

.4

127f

4Olt
40

4)

4,U

4
U

.4

N4
b

CIO

0;

a
01

4 C- P1b N 4 N N

P-

fn

N4
N

a%

-4

40

0a

fn
*x

0 a. 4

NPt

-.

4 o4.

.0

C.

'a

"I

40

10*%

N%
0.

04

.A

0%N

4%

C.44

%
N%

4..

4
m4

0-

t-

f-N

.N

A%

z%

0%

0 0%0

0.

0%

0%

01

E-

V%.4

0o

.4

.
to.4

on0%

.
ID
0

b
.

N
N

0 a%
0.

.4

N%
.4

do

Nf
P_4

v.4
4

N*

t. N

0;

- 0%* 0.

IL-0% t0N

.4

0%

t- 0

0N

-1
.

P
0

AP

.
%N

P.

40-

P.

.0-

-0'

P.

0%

*.

0012008

04N

P.

C;
.

0%

.
4

.4

c;0

NN

0
a

.
0

.0C%

0
%

0
0

94

N.-4

C4

N4

%
40

0N

4 01

C.

4. 0

0 4

40

W.
4

0
*

m.

t- .4

N C

n0
*

U)
.4

a-

%4

.0

30

.4.4..4

%%

.4 4

44

9%

44

a.

0.V
0

in
*

44A.4

4%

9%

.4-

.4

%
44

9%

%0
99,

I N a9 N 0%...
n4'

on

%
.4
9%

9
44-~~Q

0
4

in6.9 I-0

in

04.

T.
*
-4

v*
1:
"4%.4-

9
N

a%
N

C)

4
4

.4

44

.4

,.

0f-

T
00

00

%nO
0
0

4
N
.4

.4

N
N

9%
q
.

C..0
.4
-

N
.4

ft

.4

.8

44

0I

.4 c4
%

9%
*

9%
-

-4

-4

-4

.4

C-4

o.
9
-4

w9
%

t-9
"4
.4

1%
N

1.

T
4
N

.0
.4

1.

N)

C.0

r-0 .04

*0

.0

1.
.

~.- tr-9.4.40

44

.49-

9t~

6n

0
14

z9

9%

.4

9:

04

0l

00

04

4.0

0
00

m% UN
9

N
%

in
0

en
NV.

0
.

ol4.

m1

.4

04

.4

. C)0

A4.4

. 0

qq
0.

.4
.

904

9%

o0)
I

o
I

.4

:4

0A4

17
0

0C)0

., C

T.
0

.I

el
.

44-

04

1
4

04

IA
9

129

1.4-

94

14

An

.4

z%
9
0

t-

04

.M

w%9
%
.4
.4

0
1%9
4

.4

1%
%.

.0

00

.4.

0
00

.4

44

.0

-4

9
.
- .4
.4
.4

T
-

Go
0

w0

09%

4f%
4n

a,

C)

9%4

.4

09

444

0 0

.4

44

0.%a

o44.
.9

a0

a4

-9

E-%

i4n%
00

.*

6
0

4)00

a%

4:
o%
0

ft

4.

0%

In

~~~
4

IN
9%~~ 9

ow9

444~

1-

.4 N

9%

C)

'4

f
I

04

'

99

44

4.

4
.4

.4
-

.
44 4
.4
.4
N
4

o
0

o.4
*

p.
'

0Q P.
9.

'

fn

.4

.4

on

0.

0~

- 4

o4

04
.4

.4

p.

N.

0.

0
0.

4'

fn

0.p.
.6

N'N
0

6)

P.

0%

0.

0l

r.

'

ND

.4
.4

0o

0m

.4
.4

00

PN

~~~~~
5
N
6

0.

P;

4 0

N.

130

r-

4
0

0t00

C)

'

34

"

0i

P.

0'

C)

0'
0

a'

04'

'

0;

C4
4

.0

4'

'

o.

P.N

0.

(Y4
.

f.

en

0.
4t

.4

N'N
'

.
I'P

a)

.4a

0
0

00P
4

a%
4

0.h

~~~~~
'
-

.4

pt'

4040

-4

44

'

0
0

14

40

4'~~~
.

00
.

Ch
0g.
0. 4

'

4
4

N%
4'

El
4
o

toN

U) Ny

'a r.

'4

.4

a.

lo

.4

i4p

4'~f

4'

'4

-4

CI0D.

SC0

-4
0.6%

4
.4

.4

0
0

.4

NC:
.4

'

.4

0
4

-40.

m'4
N

4 16

0l

cl444
.
0

0l

0,

0p

0m

a0

.4

M4,

Mn

0N

4%

0O

r 000

0.

*0

in
0

0p
.

04

o 40-.
n

1
4 0

.4

.0

4
.4

61 4
04~

.4.

I*

a0

0f -

.4

.4

..

0.

6)

Ch

044
0

C)

(A
.

C%
4
.4

in

.4

.-

An

N)

V%

m4

.4

.4

.4

g,

.4
4

t-No

r4

.4

0,
to

Z4

'

C 0

0.

0,
4

41.U

W4-

.0'

Ua

o4

.4

4n

Z0

0'

Ol

.m

14
-..

in0

O
0

90

04
-

04

04

'0

N
o

C)

04

D0

Ua
m

44N

04

C)

0l4

.4

o,

0..4a.

4N0

0.

In

Cl

C;

..

o4
I

o0'

11

f-

14

0.

on
4

bM

0
.4

U
A

a.4

V4
C4

.4.

nN

~~~
0 U
.4

B-a~F;

4 f4

-4.4

44

on

44

44

U0**

4 N

144

a.
44

.4

.4

.0

WI~
.4

t.0-

W,
0

4,4

n.4

in

4l

0g~C

04

It

840
)0

IV

.4

*)

9.
4 04

9.

w4

*'

4%44
a%
4.

4%
4

4%
4

.a

4~

ip%
9.

.4

.4

.4

.4

(P.

~~
m4
.4a

.4

.4

.4

Np
Z

o4
9.

Q9.

4%4

'4

.
c.

4)

~9

~9

~9

4~~~~U

(N

~~ Nl

o
44
0

4.

.4

0o

.4

'.4

0
.4

0
*

C;

04o f-

N
.

*
%

a,.

.c:

.4

o4

.4

4.

fn.f
%

N4
.o
4

*1

804

00

o4

o4

N.

'4.
4*

%0

c;
N

c;
.
0

1
0

o40

o.C4
0

o0

.
4

~u

.4

~~~ ~~~~

!4N

%
4

0
0

.4

0o

w%

w)

%4
4

00

N4A

9.

0
i

.4

00

,#m

m4
.

4m

.4~ .

-4

ad~~~~~~~
*
4
4
N ~

o,
N

*.
N .

.D.4a.
Z%

NN

9. 0l
4%~Nl

09

4
.

0~~~~~c
%4
c.4

N44*
4%

4%
4

a.~
~~
N

f.
N

4%4%noA

U 4%

46

f-N

'I
04

0.

u.
.

a '

44

C) .

C,.
o'.)

.1342

. C)4.
0

-NOW

REFERENCES
REFERENCES FOR CHAPTER 1

1-1.

Etkin, Bernard. Dynamics of Atmospheric Flight,

John Wiley and Sons, New York, (1973).

1-2.

McRuer, Duane, Irving Ashkenas, and Dunstan Graham.

Aircraft Dynamics and Automatic Control, Princeton
University Press, Princeton, New York, (1973).

1-3.

Dobrolenskey, Y. P. "Flight Dynamics in Moving Air,"

NASA TT F-600 (1971).

1-4.

Bromley, Edmund, Jr., "Aeronautical Meteorology Progress and Challenges - Today and Tomorrow,"
57th Annual Meeting of the AMS, Jan. 1977,
Tucson, Arizona.

1-5.

Wyatt, S. V. "Weather Involvement in General Aviation

Accidents, "Fifth Annual International Seminar,
The Society of Air Safety Investigators,
Washington, D. C., October 1-3, (1974).

1-6.

Laynor, W. G., Performance Study DCA-74-A-14, NTS3,

Bureau of Aviation Safety, Washington, D. C., (1974).

1-7.

Aircraft Accident Report, FAA, Report No. NTSB-AAR71-11, (1971).

1-8.

Sowa, Daniel.
"Low-Level Wind Shear," D. C. Flight
Approach, No. 20, Douglas Aircraft Co., Lon. Beach,
California, (1974).

1-9.

Chambers, E. "BOAC Experience with Turbulence," Flight

in Turbulence, AGARD-CP-140 (1973), 6. 1-6.13.

1-10.

Low Level Wind Shear, Department of Transportation,

Federal Aviation Administration, Advisory Circular AC No: 00-50, (1976).

133

REFERENCES FOR CHAPTER 4

4-1

Kropfli, R. A. and L. J. Miller, "Kinematic Structure

and Flux Quantities in a Convective Storm from
Dual-Doppler Radar Observations," J. Atmos. Sci. 3
(1976) 520-529.

4-2

Goff, R. C., "Thunderstorm Outflow Kinematics and

Dynamics," NOAA Tech. Memo. ERL NSSL-75 (1975).

4-3

Colmer, M. "On the Character of Thunderstorm Gust

Fronts" Proceeding International Conference on
Atmospheric Turbulence Roy. Aeronautical Soc., (1971).

4-4

Sinclair, R. W., R. A. Anthes and H. A. Panofsky,

"Variation of the Low Level Winds During the Passage
of A Thunderstorm Gust Front," NASA CR-2289 (1973).

4-5

Frank, H.W., "Notes on the 8 August 1972 Haswell Gravity

Flow Event," personal communication.

4-6

Clarke, R.H., Mesostructure of Dry Cold -Fronts over

Featureless Terrain," J. Meteor., 18 (1961) 715-735.

4-7

Byers, H.R. and R.R. Braham, Jr., The Thunderstorm, U.S.

Government Printing Office, Washington, D.C., (1949)
287.

4-8

Ward, N.B., and A.B. Arnett, "Some Relationships

Between Surface Wind Fiel-Is and Radar Echoes,"
Proc. of the third Conf. on Severe Storms, Champaign,
Ill. (1962).

4-9

Browning, K.A. and T.W. harold, "Air Motion and Precipitation Growth of a Cold Front," Quart. J. Roy.
Meteor. Soc., 96 (1970) 369-398.

4-10

Fichtl, G.H., and D. Camp, Personal Communications (1976).

4-11

Mitchell, K.E., "A Numerical Investigation of Severe

NASA CR-2635, (1975).
Thunderstorm Gust Fronts,"

134

REFERENCES FOR CHAPTER 5

5-1

Fichtl, GH., "Problems in the Simulation of Atmospheric Simulation of Atmospheric Boundary Layer
Flows, Flight in Turbulence," AGARD-CP-140 (1973)
2-1.

5-2

Busch, Niels E., "Thp Surface Boundary Layer,"

Boundary Layer Meteorology, 4 (1973) 213-240.

5-3

Monin, A.S., "The Atmospheric Boundary Layer," Annuai

Review of Fluid Mech., Annual Reviews, Palo Alto,
Calif. (1970) 225-150.

5-4

Leurs, JoK., "A Model of Wind Shear and Turbulence in

the Surface Boundary Layer," NASA CR-2288 (1973).

5-5

Barr, N.M., Dagfinn Gangers and D.R. Schaeffer,

"Wind Models for Flight Simulator Certification of
Landing and Approach Guidance and Control Systems,"
Report No. FAA-RD-74-206 (1974).

5-6

Businger, J.A. "Turbulent Transfer in the Atmospheric

Surface Layer," Workshop on Micrometeorology, Ed.
Duane A. Haugen, Published by American Meteorology
Society, (1973) 67-100.

5-7

Blackadar, A.K., J.A. Dutton, H.A. Panofsky and A.

Chaplin, "Investigation of Turbulent Wind Fields
Below 150 m Altitude at the Eastern Test Range,"
NASA-CR-1410 (1972).

5-8

Fiedler, F., and H.A. Panofsky, "The Geostnophic Drag

Coefficient and the Effective Roughness Length,"
Quart. J. Roy, Meteor. Soc., 98 (1972), 213-306.

5-9

Thuiller, R.H. and V.0. Lappe, "Wind and Temperature

Profile Characteristics from Observations on a
1400 ft. Tower," J. Appl. Meteor., 3 (1964) 299-306.

5-10

Tennekes, H. "Similarity Laws and Scales Relations in

Planetary Boundary Layers," Workshop on
Micromoteoroloqy Ed. Duane A. Haugen, Published by
American Meteorology Society, (1973) 177-216.

5-11

Hana, S.R., "Characteristics of Wind and'Turbulence in

the Planetary Boundary Layer," ERLTM-ARL Tech. Memo.
8, Oak Ridge, Tenn. (1969).

135

5-12

Estoque, M.A., "Numerical Modeling of the Planetary

Layer," Work.s;hop on Micrometeorology, Ed., Duane
A. Haugen, Published by American Meteorology
Society, (1973) 217-270.

5-13

Blackadar, A.K., and H. Tennekes, "Asymptotic

Similarity in Neutral Barotropic Atmospheric
Boundary Layers," J. Amos. Sci. 25 (1968), 1015-1020.

5-14

Hess, G.D. "On Rossby-Number Similarity Theory for a

Baroclinic Boundary Layer," J.A ms
.,
(1973), 1722-1723.

5-15

Yamada, Tetsuji, "On the Similarity Functions A, B and

Cof the Planetary Boundary Layer,"
.Ams
c.
5 (1976) 781-793.

5-16

Clarke, R.H., and G.D. Hess, "Geostrophic Departure and

the Function A and B of the Rossby-Number Similarity
Theory," Boundary Layer Meteorology 3 (1974) 267-288.

5-17

Deardorff, J.W., "Numerical Investigation of Neutral

and Unstable Planetary Boundary Layer," J. Atmos.
Sci., 30, (1972) 91-115.

5-18

Blackadar, A.K., "The Vertical Distribution of Wind

and Turbulent Exchange in a Neutral Atmosphere,"
J. Geophys. Res. 67 (1962) 3095-3102.

5-19

Ohmstede, W.D. and J.F. Appleby, Numerical Solutdon of

the Distribution of Wind and Turbulence in the
Planetary Boundary Layer," Meteor. Res. Nlote No. 8
DA Task I-A-O-1l01-B-021-08 USAERDAA-MET-5-64,
(1964) 43.

5-20

Blackadar, A.K. and J. Ching, "Wind Distribution in a

Steady State Planetary Boundary Layer of the Atmosphere
With Upward Heat Flux," Final Rept. Contract AF
(604)-6641, Dept. of Meteor., Penn. State Univ.
(1965) 23-48.

5-21

Yamamoto, G., N. Yasuda and A. Shimanuke, "Effects of

Therwa] Statification on the Ekman Layer," J. Meteor.
Soc., Japan, 46, (1969) 442-455.

136

5-22

Blackadar, A.K., "A Single Theory of the Vertical

Difntribution of the Wind in a Baroclinic Neutral
Atnospheric Boundary Layer," Final Rept., Contract
AF (604)-6641, Dept. of Meteor., Penn. State Univ.
(1965) 1-22.

5-23

Plate, E.J., Aerodynamic Characteristics of Atmospheric

Bounday Layers. AEC Critical Review Series, TID25465, (1971) 137-160.

5-24

Huang, Chin, Hua and E.C. Nickerson, "Stratified Flow

Over Non-Uniform Surface Conditions: Mixing Length
Model," Boundary Layer Meteorology 4 (1974) 395-418.

5-25

Panchev. S., E. Donev and N. Godev, "Wind Profile and

Vel-tical Motions above an Abrupt Change in Surface
Roughness and Temperature," Boundary Layer
Meteorology 1 (1971) 52-63.

5-26

Bradley, E.F., "A Micrometeoroloqical Study of Velocity

Profiles and Surface Drag in the Region Modified by
a Change in Surface Roughness," Quart. J. Roy.
Meteor. Soc. 94 (1968) 361-379.

5-27

Munro, D.S. and T.R. Oke, "Aerodynamic Boundary Layer

Adjustment Over a Crop in Neutral Stability," Boundary
Layer Meteorology, 9 (1975) 53-61.

5-28

Antonia, R.A. and R.E. Luxton, "The Response of a

Turbulent Boundary Layer to a Step Change in
Surface Roughness. Part 2 Rough-to-Smooth," J.
Fluid Mech. 53 (1972) 737-757.

5-29

Merony, R.N., "Characteristics of Wind and Turbulence

in and Above Model Forests," J. Appi. Meteor., 7
(1968) 780-788.

5-30

Elliott, W.P., "The Growth of the Atmospheric Internal

Boundary Layer," Trans. Amer. Geophys. Union 39,
(1958) 1048-1054.

5-31

Rao, K.S., "Effect of Thermal Stratification of the

Internal Boundary Layer," Boundary Layer
Meteorology, 2 (1975) 227-234.

13

137

-4
--------------------------------------------------------------------------

-m-

5-32

Huang, Chin-Hua and E.C. Nickerson, "Stratified Flow

Over Non-Uniform Surfaces: Turbulent Energy Model,"
Boundary Layer Meteorology, 1, (1974) 107-124.

5-33

Rao, K.S., J.C. Wyngaard and O.R. Cote, "The Structure

of the Two-Dimensional Internal Boundary Layer Over
a Sudden Change in Surface Roughness," J. Atmos. Sci.
31 (1974) 738-746.

5-34

Bitte, Juergen, and Walter Frost, "Numerical Solution

of Turbulent Flow over Surface Roughness Transition:
Two Equation Turbulence Model," paper in preparation,
Univ. of Tenn. Space Institute.

5-35

Logan, Earl, Jr., and G.H. Fichtl, "Rough to Smooth

Transition of an Equilibrium Neutral Constant
Stress Layer," NASA TM X-3322, (1975).

5-36

Geiger, Rudolf, The Climate Near the Ground, Harvard

University Press, Cambridge, Mass. (1950).

5-37

Windbreaks and Shelterbelts WMO-No. 147.TP.70, (1964).

5-38

Munn, R.E., Descriptive Micrometeorology, Academic

Press, (1966).
Ogawa, Yasushi, Richard Griffil-hs and W.G. Hoydysh,

5-39

"A Wind Tunnel Study of Sea Breeze Effects,"

Boundary Layer Meteorology 2 (1975) 141-162.
5-40

Vukovich, F.M., J.W. Dunn III and B.W. Cressman, "A

Theoretical Study of the St. Louis Heat Island:
The Wind and Temperature Distribution," J. Appl.
Meteor. 5 (1976) 417-440.

5-41

Taylor, P.A., "A Model of Airflow above Changes in

Surface Heat Flux, Temperature and Roughness for
Neutral and Unstable Conditions." Boundary Layer
Meteorology, 1 (1970) 18-39.

138

REFERENCES FOR CHAPTER 6

6-1.

lHoubolt, J. C., Roy Steiner, and K. G. Pratt.

"Dynamic Response of Airplanes to Atmospheric
Turbulence Including Flight Data on Input and
Response," NASA TR R-199, National Aeronautics and
Space Administration, Washington, D.C., June 1964.

6-2.

Houb-it, J. C. "Gust Design Procedures Based on Power

Spectral Techniques," Technical Report AFFDL-TR6*/-74, Wright-Patterson Air Force Base, Ohio,
August 1967.

6-3.

Reeves, P. M., G. S. Campbell, V. M. Ganzer, and

R. G. Joppa. "Development and Application of a
Non-Gaussian Atmospheric Turbulence Model for Use
in Flight Simulators," NASA CR-2451, National
Aeronautics and Space Administration, Washington,
D.C., September 1974.

6-4.

Reeves, P. M., R. G. Joppa, and V. M. Ganzer. "A NonGaussian Model of Continuous Atmospheric Turbulence
for Use in Aircraft Design," NASA CR-2639,
National Aeronautics and Space Administration,

Washington, D.C., January 1976.

6-5.

Jones, J. G. "A Unified Discrete Gust and Power

Spectrum Treatment of Atmospheric Turbulence,"
Paper presented at the Royal Aernnautical Society
International Conference on Atmospheric Turbulence,
May 18-21, 1971.

6-6.

Tomlinson, B. N.

"Developments in the Simulation of

Atmospheric Turbulence,"

Tech.

Memo FS 46,

Royal

Aircraft Establishment, September 1975.

6-7.

Frost, Walter, and M. Perlmutter. "Three Velocity

Component Atmospheric Boundary Layer Turbulence,"
Contract No. NAS8-29548 Report, University of
Tennessee Space Institute, Tullahoma, Tennessee
37388, 1976.

6-8.

Munn, R. E., editor. "Turbulence Spectra, Length

Scales and Structure Parameters in the Stab]e
Surface Layer."

Vol.
6-9.
6-10.

4,

Nos.

3,

Boundary- .aerNeteoroioq,

2,

April 1973.

Jernigan, P. L. Douglas Aircraft Company, personal

communications.
Bray,

R. S.

NASA Ames Research Center,

comnu nica 0 ons.

-4

3 and 4,

139

peisonal