NEW METHODS

IN
SHALLOW SEISMIC REFLECTION

BY

ZUHAR ZAHIR TUAN HARITH

BSc. (Malaya); MSc. (London); DIC (Imperial College)

November, 1998

This thesis is submitted for the degree o f Doctor of Philosophy

to

The Faculty of Science,

Department of Geology and Applied Geology,
University of Glasgow
 ProQuest Number: 13815565

All rights reserved

INFORMATION TO ALL USERS

The quality of this reproduction is d e p e n d e n t upon the quality of the copy subm itted.

In the unlikely e v e n t that the a u thor did not send a c o m p le te m anuscript

and there are missing pages, these will be noted. Also, if m aterial had to be rem oved,
a n o te will ind ica te the deletion.

uest
ProQuest 13815565

Published by ProQuest LLC(2018). C opyright of the Dissertation is held by the Author.

All rights reserved.

This work is protected against unauthorized copying under Title 17, United States C o d e
M icroform Edition © ProQuest LLC.

ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 4 8 1 0 6 - 1346
New Methods In Shallow Seismic Reflection
Z Z T Harith Dedications

Dedicate to

my wife, my parents and all my families....

“ Your undying love and support have assisted me in

putting me where I am to d a y ”

University Of Glasgow 1998 PhD Dissertation Page II

New Methods In Shallow Seismic Reflection
Z Z T Harith Declarations

This thesis represents work carried out between January 1996 and November 1998.

I declare that this thesis is entirely my own work except where explicitly declared in the
text. All published and unpublished materials used in this thesis have been given full
acknowledgement in the text.

This thesis has not been submitted for a degree at any other higher institution.

Zuhar Zahir Tuan Harith

University Of Glasgow 1998 PhD Dissertation Page III

ACKNOWLEDGEMENTS
New Methods In Shallow Seismic Reflection
Z Z T Harith Acknowledgements

I would like to acknowledge the University of Science Malaysia and Malaysian

Government for giving me an opportunity by granting me a three year scholarship so
that this project can be completed despite economic turbulence in the South East Asian
region. An additional acknowledgement is also expressed to John Nevett of Glasgow
Development Agency (GDA) for giving permission to experiment with the first full 3-D
shallow seismic survey on their site and who agreed to fund the cost of the field work.
Without this financial help, this project could not have been done.

This project was initiated by Prof. D. K. Smythe and who supervised the whole project.
Besides supervising the project, he also gave lots of guidance, support and advice on all
matters whether it project, career or personal. He also even postponed his retirement to
make sure that I could finish on time. For that I would like to express my thanks and
appreciation.

I would also like to express my gratitude to the two other geophysical lecturers - Dr.
Doyle Watts and Dr. Ben Doody - for their valuable insight into this work.

Appreciation is also due to the technical staff who assisted (as well as kept me
entertained) throughout the years - Bob Cumberland and George Gordon. Also thanks
to Colin Farrow and Kenny Roberts for computer support work, Mr. Douglas McLean
for photographic work and Mr. Roddy Morrison (chief technician).

Thanks to all the other postgraduates who also go through years of torture in order to
contribute to new knowledge -Bel, Andy, Carsten, A1 Hamali, Fitchy, Joan, Keijan,
Hadi, Gloria, Andy Cav., and others.

Also thanks to

1 Peter Fraenkel and Partners (the GDA consulting engineers) and J. W. H. Ross
and Co. (drilling engineers) for their help during the planning stage and kindly
providing drilling reports.

2 Roger Caldwell of OYO UK Ltd. for generously providing the recording and
minivibrator equipment at a very economic cost, and for supplying free
training in its use.

3 The Department of Archaeology for kindly loaning its semi-total station

equipment (EDM theodolite).

4 The Robertson Trust for a substantial grant to the Department o f Geology and
Applied Geology to enable purchase of the 3-D seismic processing package
ProMAX/3D.

5 Prof. Ewald Briickl of the Technical University of Vienna for his comments
and suggestions in tackling the processing problems and also for the

University Of Glasgow 1998 PhD Dissertation Page IV

New Methods In Shallow Seismic Reflection
Z Z T Harith Acknowledgements

willingness to test the author’s first break data in his refraction modelling
program.

6 Sarah Hamilton, Catherine McGee, David Sullivan and Claire Harrison - the
student field assistants - for their hard work.

Thanks also to my family who give their blessing and ‘doa’ during my three years here.

Finally I would like to thank my wife for her understanding and willingness to go
through two ‘agony’ years together even though it means sacrifice of two years of
priceless experience as an engineer.

University Of Glasgow 1998 PhD Dissertation Page V

ABSTRACT
New Methods In Shallow Seismic Reflection
Z Z T Harith Abstract

The seismic reflection exploration technique has long been established in the oil
industry as one of its most successful geophysical exploration tools. It has evolved from
a very simple technique (2-D line profiling) to become the most advanced geophysical
exploration technique of all (the 3-D reflection technique). Besides its successful use in
the oil industry, seismic reflection can also be used to investigate shallow targets. To
date, most shallow seismic investigation practitioners have only used the 2-D line
profiling technique, as it is easy and cheap to carry out.

Depending on the target, some geophysical investigation techniques can provide an

answer to archaeological needs, since they are fast, reliable and above all non­
destructive. However, this is not the case for the 2-D seismic reflection method, as it is
generally considered to be an unsuitable technique. One of the aims of this dissertation
is to test the 2-D line profiling technique with a spatially highly condensed geometry to
detect the now-buried vallum (defensive ditch) of a Roman fort at Inveresk, Edinburgh,
Scotland. The results are compared with the results of two previously obtained
resistivity profiles. An excavation programme around the site had established the
existence of only one ditch fort, although this is unusual for a Roman fort, as most of
them had at least two ditches enclosing them. The qualitative resistivity results revealed
four low apparent resistivity anomalies, with two of them being possible ditch-related
anomalies in the area where the ditch (or ditches) was predicted. The 2-D seismic
reflection profile acquired a very condensed dataset using a CDP jump-roll geometry.
The geophones with a hammer source were spaced at 0.25 m interval along a 50 m long
line, yielding a total of 2992 traces. The result of this seismic reflection profile confirms
the presence of two ditch-like structures about 3 m and 1 m deep, as inferred from the
resistivity profiles. Since one of them is too shallow to be a Roman ditch, the seismic
results appear to confirm the existence of only one ditch at this fort. The condensed 2-D
method is thus capable of imaging very shallow buried structures such as ditches only
one or two metres deep.

The use of current 3-D seismic reflection oil industry technology technique in shallow
investigation work is still in an early phase. The limiting factor of transferring this
technology is primarily the cost. There is also a problem in scaling down the survey to
suit a shallow target. The primary aim of this dissertation is to describe a case history of
an attempt to map out the underground tunnels o f an abandoned coal mine using 3-D
seismic reflection methods, suitably adapted from the 3-D seismic reflection technology
used in the oil exploration industry. The site selected for the survey is located in a
brownfield site, part of West of Scotland Science Park, Glasgow, Scotland. The 3-D
survey was designed to acquire a highly condensed dataset with a full range of source-
receiver azimuths, very high fold of coverage, and in the fastest acquisition time
possible. For that, a combination of a full and random 3-D survey geometry methods
was applied. The data were recorded using a 144-channel, 24-bit recording system. A
100-700 Hz sweep of the mini-vibroseis source was used to try to obtain high frequency
broad-band data. To randomise the common mid-points, each source was placed
pseudo-randomly within a 4 m2 area at every 4 m interval. The single element
geophones formed a rectangular swath of 8 by 18 on a 2 m grid. The data were
processed using industry standard 3-D processing software, using three different

University O f Glasgow 1998 PhD Dissertation Page VI

New Methods In Shallow Seismic Reflection
Z Z T Harith Abstract

approaches successively. The results reveal an undulating top bedrock surface 3-7 m
below ground surface. Although convincing images of underground voids or coal seams
were not successfully obtained, the results reveal a general pattern of possible
mineworks (or a thick coal seam), showing that such 3-D work is feasible and cost-
effective. In retrospect, both the geology and the abandoned workings at the chosen site
were probably too complex as a test target.

A by-product of the 3-D reflection study is the development of a new simple and robust
mapping technique. This new technique is named ‘surface tomography’ as it utilises the
tomography concept in a simplified form. The aim is to develop a semi-automated
method which may pinpoint areally small velocity anomalies on the top bedrock
velocity map (say a 2 m diameter anomaly due, for example, to a buried mineshaft).
Only the first breaks from ail the source-receiver pairs corresponding to the refracted
arrival from the top bedrock surface are used. Any propagation through a low velocity
zone within the bedrock will experience a time lag (an increase in travel time) and vice-
versa for a high-velocity zone. The first arrival times are reduced to a representative
average velocity. Any reduced time that is higher than the average reduced time is
considered to have experienced a time lag and vice-versa. To ensure that only raypaths
propagating along the bedrock surface were used, the first arrivals from a distance less
than the crossover distance of the bedrock arrival and the direct arrival (from the layer
above) are deleted. Colour-coded maps of some 80,000 reduced times from the 3-D first
breaks were produced to display the anomalies. The results show a systematic pattern of
colours striking in the same direction as the general strike of the area. The combined
information from this study, the 3-D reflection work and borehole information has been
used to revise the geological model of the area.

University Of Glasgow 1998 PhD Dissertation Page VII

TABLE OF CONTENTS
New Methods In Shallow Seismic Reflection
Z Z T Harith Contents

page

Dedications II
Declarations III
Acknowledgements IV
Abstract VI
Table of contents VIII
List of figures XIV
List of tables XXI
List of photos XXIII

CONTENTS

Chapter One Introduction

1.1 Seismic reflection - an overview 1

1.2 Shallow seismic reflection 3

1.2.1 Introduction 3
1.2.2 Some problems 4

1.3 The research 5

1.3.1 Research outline and aims 5

1.3.2 Funding 8

1.4 Thesis presentation and organisation 8

1.5 Previous work 9

Chapter Two Seismic reflectionin archaeology

2.1 Introduction 13

2.2 Aim 14

2.3 Previous work 14

University Of Glasgow 1998 PhD Dissertation Page VIII

New Methods In Shallow Seismic Reflection
Z Z T Harith Contents

2.4 The site 15

2.4.1 Introduction 15
2.4.2 Location 15
2.4.3 Geology 16
2.4.4 History 19
2.4.5 Target 20

2.5 Resistivity survey 23

2.5.1 Introduction 23
2.5.2 Interpretation theory 23
2.5.3 Results and discussion 25

2.6 Reflection survey 32

2.6.1 Can reflection work? 33

2.6.2 Data acquisition 35
2.6.3 Seismic data processing 40
2.6.4 Discussion 54

2.7 Conclusions 56

Chapter Three 3-D Shallow seismic reflection

3.1 Introduction 58

3.2 Aims 59

3.3 Previous work 59

3.4 The site 66

3.4.1 Location 66
3.4.2 Geology 66
3.4.3 The problem 70
3.4.4 Suggested solution 76

University Of Glasgow 1998 PhD Dissertation Page IX

New Methods In Shallow Seismic Reflection
Z Z T Harith Contents

3.5 Planning and basemaps 77

3.5.1 Planning 77
3.5.2 Basemaps 79

3.6 Preliminary study 79

3.6.1 Preliminary geophysical surveys 82

Resistivity 82
Seismic refraction 85
3.6.2 Borehole studies 86
The 1966 Drilling programme 86
The 1981 Drilling programme 87
The 1995 Drilling programme 88

3.7 Topographic survey 90

3.7.1 Equipment 90
3.7.2 Surveying strategy and aims 91
3.7.3 Controls 93
3.7.4 Conversion from Column-Row coordinate to National
Grid 93
3.7.5 Precision and errors 94
3.7.6 Problems 94

3.8 3-D survey design 94

3.8.1 3-D field layout 98

3.9 Acquisition 99

3.9.1 Instrumentation 99
OYO DAS-1 recording system 99
Sources 104
3.9.2 3-D data recording 106
Crew 106
Receiver swath 106
Randomised source location 106
Vibroseis sweep 108

University Of Glasgow 1998 PhD Dissertation PageX

New Methods In Shallow Seismic Reflection
Z Z T Harith Contents

Recording parameters 110

3.9.3 2-D data recording 113
Introduction 113
Geometry 113
Recording parameters 114
3.9.4 Data recording and transfer 114
3.9.5 Progress 117

3.10 Seismic data processing 121

3.10.1 Introduction 121

Vibroseis correlation 122
Processing strategies 122
3.10.2 2-D data processing 125
Introduction 125
Frequency analysis 125
Phase comparison andfirst break picking 128
Refraction model 132
Pre-processing 132
Processing - approach-I 135
Approach-I: result and comments 146
Processing - approach-II 146
Approach-II: result and comments 155
3.10.3 3-D seismic data processing 155
Introduction 155
Field correlation 156
Lab correlation 156
Comparison between field and lab correlated data 159
Processing strategies 163
Pre-processing 164
Editing and geometry assignment 164
Binning 165
Fold o f coverage 170
Processing - approach-I 175

University Of Glasgow 1998 PhD Dissertation Page XI

New Methods In Shallow Seismic Reflection
Z Z T Harith Contents

Approach-I: results and comments 178

Processing - approach-II 182
Approach-II: results and comments 185
Processing - approach-111 (re-processing o f approach-II) 192
Approach-Ill: results and comments 199

3.11 Discussion and conclusion 214

Chapter Four Surface tomography

4.1 Introduction 218

4.2 The concept 220

4.3 Previous work 223

4.4 Aims 224

4.5 First breaks 227

4.5.1 Picking 227

4.5.2 Time distance curves 229
4.5.3 Reduced time distance curves 236

4.6 Surface tomography 239

4.7 The display 241

4.7.1 Reduced time (vector) map 241

4.7.2 Reduced time (vector) anisotropy map 242
4.7.3 Reduce time (image) map 244

4.8 Interpretation 244

4.8.1 Control 247

4.9 Discussion 250

University Of Glasgow 1998 PhD Dissertation Page XII

New Methods In Shallow Seismic Reflection
Z Z T Harith Contents

Chapter Five General conclusions and recommendations

5.1 Seismic reflection as a tool for archaeological survey 267

5.1.1 Conclusions 267

5.1.2 Recommendations 269
5.2 3-D seismic reflection for engineering scale site survey 270

5.2.1 Conclusions 270

5.2.2 Recommendations and further work 272
5.3 Surface tomography mapping - a new tool 273

5.3.1 Conclusions 273

5.3.2 Recommendations 274

References 276

Appendixes

Appendix 1 Basis of surface tomography. 284

Appendix 2 Correlation of coal seams and strata encountered in
borehole 1981-1 to -29 inclusive. 285
Appendix 3 Calculation of 3-point problem to establish regional strike
and dip. 287
Appendix 4 Calculation of conjectural outcrop points of Twechar no. 1
and Twechar no. 2 coals 290

University Of Glasgow 1998 PhD Dissertation Page XIII

LIST OF FIGURES
New Methods In Shallow Seismic Reflection
Z Z T Harith Figures

page

Figure 2.1 Inveresk: Site and physical setting. 17

Figure 2.2 Richmond’s proposed plan for Inveresk Fort. 21

Figure 2.3 Schematic cross-section of the defence (Western Wall). 22

Figure 2.4 Details of the study area. 24

Figure 2.5 Resistivity profiles (CST) across buried ditch. 25

Figure 2.6a L3 resistivity profiles. 27

Figure 2.6b L4 resistivity profiles. 28

Figure 2.7 Relation between resistivity anomalies (A-D & Z) and their
near surface features. 31

Figure 2.8 The excavation cross-sections showing the relation between

archaeological remains and their surrounding. 34

Figure 2.9 CDP jump roll geometry. 38

Figure 2.10 Shot gathers. 41

Figure 2.11 CDP gathers. 42

Figure 2.12 Frequency content. 44

Figure 2.13 F-K analysis. 45

Figure 2.14 CDP’s gathers with front or top mute applied. 46

Figure 2.15 CDP’s gather after pre-stack processing applied. 49

Figure 2.16 Processing flow chart. 50

Figure 2.17a Brute stack. 51

Figure 2.17b Stack section with DSS applied to remove linear dipping
events. 52

Figure 2.17c Time migrated section. 53

Figure 2.18 Interpreted time migrated section. 55

University O f Glasgow 1998 PhD Dissertation PageXTV

New Methods In Shallow Seismic Reflection
Z Z T Harith Figures

Figure 3.1a Simple 3-D field geometry design (after Corsmit et al. 1988). 60

Figure 3.1b Simple 3-D field geometry design (after Hasbrouck 1994). 61

Figure 3.1c Simple 3-D field geometry design (after Lanz et al. 1996). 62

Figure 3.Id Simple 3-D field geometry design (after Miller et al. 1997). 63

Figure 3.2 West of Scotland Science Park, Todd Campus West: The site
location. 67

Figure 3.3 Geological map and the cross-section of the site and
neighbouring area. 69

Figure 3.4 Plan of a coal mine as abandoned in 1897. 74

Figure 3.5 Basemap as supplied by Peter Fraenkel and Partners. 80

Figure 3.6 Simplified basemap with extra information such as borehole

locations and refraction lines added. 81

Figure 3.7 Example resistivity curve (Wenner array) plotted on log-log

graph. 84

Figure 3.8 Summary of the 1981 mineral investigation programme. 89

Figure 3.9 Prearrange layout map for a grid of survey pegs at 2 m

spacing. 92

Figure 3.10 3-D three-line swath. 95

Figure 3.11 Basic swath design for 3-D. 96

Figure 3.12 Example of simple fixed receiver pattern full 3-D survey. 97

Figure 3.13 Traditional moving receiver full 3-D survey. 97

Figure 3.14 An example of three spreads (CDE) making one swath. 100

Figure 3.15 Layout of all spreads. 101

Figure 3.16 Layout of patterns CDE and V’W’Y- a detailed example. 102

Figure 3.17 Layout of all patterns. 103

Figure 3.18 Pegs, receivers and randomised source positioning. 107

University Of Glasgow 1998 PhD Dissertation Page XV

New Methods In Shallow Seismic Reflection
Z Z T Harith Figures

Figure 3.19 Seismogram concept. 123

Figure 3.20 Graphical illustration of the superposition concept of the

vibroseis data and their cross-correlating effect. 124

Figure 3.21 Shot gather of three different types of source at same

location. 126

Figure 3.22 Vibroseis shot gathers from three different shot-points. 127

Figure 3.23a Shot gather and frequency spectrum - vibroseis. 129

Figure 3.23b Shot gather and frequency spectrum - hammer. 130

Figure 3.23c Shot gather and frequency spectrum - weight drop. 131

Figure 3.24 Comparison of the polarity of the signal of three different

sources. 133

Figure 3.25 The refraction model produced by the Technical University

of Vienna from 2-D line (vibroseis source) . 134

Figure 3.26 Raw vibroseis shot record showing that the gather is
dominated mostly by unwanted noise. 136

Figure 3.27 Shot gather no. 13 from vibroseis data. 137

Figure 3.28 Shot gather no. 13 after AGC, TAR, 60-100-300-400 Hz

frequency bandpass, DSS and another AGC. 138

Figure 3.29 F-K analysis used for designing the F-K fan filter parameter. 140

Figure 3.30 Shot gather after F-K filtering. 141

Figure 3.3 la Stacked section using 600 m s'1constant velocity. 142

Figure 3.3 lb Stacked section using 1350 m s'1constant velocity. 143

Figure 3.3 lc Stacked section using 2054 m s 1constant velocity. 144

Figure 3.32 Three original CMP gathers (2 m trace spacing)showing

severe aliasing. 148

Figure 3.33 CMP gathers after summation of 4 adjacent CMPs. 149

Figure 3.34 Supergather after severe front and bottom mute applied. 150

Figure 3.35a 2-D vibroseis data - brute stack. 152

University Of Glasgow 1998 PhD Dissertation Page XVI

New Methods In Shallow Seismic Reflection
Z Z T Harith Figures

Figure 3.35b 2-D Hammer data - brute stack. 153

Figure 3.35c 2-D Weight drop data - brute stack. 154

Figure 3.36 Graphical presentation of 100-700 Hz, 200-700 Hz and 100-

700 Hz with time delay sweeps. 158

Figure 3.37a Raw shot gather for very short offsets (2-15 m) correlated in
three different ways. 160

Figure 3.37b Raw shot gather with medium offsets (15-30 m) correlated in
three different ways. 161

Figure 3.37c Raw shot gather with long offsets (>25 m) correlated in three
different ways. 162

Figure 3.38 Detail of offset vectors. 166

Figure 3.39 All corrected source vectors. 167

Figure 3.40 CMP coverage map. 168

Figure 3.41a CMPs with 2 m binning grid overlain. 169

Figure 3.41b Detail of binned CMPs at southern edge of survey. 170

Figure 3.41c Detail of binned CMPs from centre of survey. 171

Figure 3.42a Fold of coverage in plan view. 173

Figure 3.42b 3-D view of fold of coverage. 174

Figure 3.43 Raw and pre-processed data showing severe front mute
applied in approach-I. 176

Figure 3.44a Sample inline and crossline sections - approach-I 179

Figure 3.44b Sample inline and crossline sections - approach-I (interpreted). 180

Figure 3.45 Contour map of base of clay horizon based on approach-I

result. 181

Figure 3.46 Raw and pre-processed data showing only bottom mute
applied in approach-II. 183

Figure 3.47a Samples from two inline sections - approach-II. 186

Figure 3.47b Samples from two crossline sections - approach-II. 187

University Of Glasgow 1998 PhD Dissertation Page XVII

New Methods In Shallow Seismic Reflection
Z Z T Harith Figures

Figure 3.48a Samples from two inline sections - approach-II (interpreted). 188

Figure 3.48b Samples from two crossline sections - approach-II

(interpreted). 189

Figure 3.49 Bedrock or base of clay (red horizon) map - approach-II. 190

Figure 3.50 Yellow horizon (possibly underground voids) - approach-II. 191

Figure 3.5la CMP coverage with rotated 2 m binning grid - approach-III. 194

Figure 3.51b Relation between inline-crossline and National Grid (E-N)

coordinates for approach-III. 195

Figure 3.52 Raw and pre-processed data showing severe front and bottom
mute applied in approach-III. 196

Figure 3.53a Samples from two inline sections - approach-III. 200

Figure 3.53b Samples from two crossline sections - approach-III. 201

Figure 3.54a Samples from two inline sections - approach-HI

(interpreted). 202

Figure 3.54b Samples from two crossline sections - approach-III

(interpreted). 203

Figure 3.55a The view of time slice at 17.5 ms TWT. 205

Figure 3.55b The view o f time slice at 20.0 ms TWT. 206

Figure 3.55c The view of time slice at 24.0 ms TWT. 207

Figure 3.55d The view o f time slice at 25.5 ms TWT. 208

Figure 3.55e The view of time slice at 30.0 ms TWT. 209

Figure 3.55f The view of time slice at 35.0 ms TWT. 210

Figure 3.56 Plan view of the red (clay/bedrock) reflector. 211

Figure 3.57 Plan view of the yellow (? mineworkings or thick coals)

reflector. 213

Figure 4.1 Fan-shooting refraction, a basic idea of surface tomography. 221

Figure 4.2 Schematic cross-section of the study area showing the

relatively simple geology. 225

University Of Glasgow 1998 PhD Dissertation Page XVIII

New Methods In Shallow Seismic Reflection
Z Z T Harith Figures

Figure 4.3 Example of first break picking on a CMP gather. 228

Figure 4.4 Time-distance curve for all source-receiver pairs. 230

Figure 4.5 Third year refraction spread geometry. 232

Figure 4.6a Test of three different reduction velocities (400, 600, and 800
m s’1) in order to estimate velocity for second layer. 237

Figure 4.6b Test of three different reduction velocities (1400, 1600, and
1800 m s'1) in order to estimate velocity for third layer. 238

Figure 4.7 Schematic relation between surface path and the surface
projected path within the third layer. 240

Figure 4.8 Reduced time (vector) map. Back pocket 1

Figure 4.9 Colour code as used in all reduced time maps. 243

Figure 4.10a Reduced time (vector) 0-90° anisotropy map. Back pocket 2

Figure 4.10b Reduced time (vector) 90-180° anisotropy map. Back pocket 3

Figure 4.1 la An example of a series of 2 m interval points (red crosses)

calculated along two corresponding vectors (solid lines). 245

Figure 4.11b Details of the points (left) and 0.25 m pixels of different
colour (right) used to produce the image map. 246

Figure 4.12 Reduced time (image) map. Back pocket 4

Figure 4.13 Conjectural outcrop (red dashed lines) of various coal seams
in and around the study area. 248

Figure 4.14a Reduced time (vector) map - enlargement of the southern part
of the study area. Back pocket 5

Figure 4.14b Reduced time (vector) 90-180° anisotropy map - enlargement

of the southern part of the study area. Back pocket 6

Figure 4.14c Reduced time (vector) 110-155° anisotropy map -

enlargement of the southern part of the study area. Back pocket 7

Figure 4.15 Bedrock types from 1966, 1981 and 1995 boreholes. 252

Figure 4.16 Calculated outcrop points of the Twechar no. 1 and Twechar
no. 2 coals based on boreholes 1981-11, -12, and -13. 256

University Of Glasgow 1998 PhD Dissertation Page XIX

New Methods In Shallow Seismic Reflection
Z Z T Harith Figures

Figure 4.17 3-points problems showing a rather string strike and dip for
Twechar no. 2 coal. 257

Figure 4.18 Revised conjectural lines for Twechar no. 1 and 2 coals. 261

Figure 4.19 Relation between the true thickness, horizontal thickness and
maximum width of the blue. 262

Figure 4.20 Graphical representation of the variation of the overburden. 265

Figure 5.1 Points within 1 m search radius which contribute to the

average value of the central point. 275

University Of Glasgow 1998 PhD Dissertation Page XX

LIST OF TABLES
New Methods In Shallow Seismic Reflection
Z Z T Harith Tables

page

Table 1.1 Chronology of seismic instrumentation and methods. 2

Table 2.1 Summary of the geology of the Midland Valley. 18

Table 2.2 Summary of the low apparent resistivity anomalies. 29

Table 2.3 Elevation of the reflection line. 36

Table 2.4 Summary of the recording parameters. 39

Table 2.5 Stacking velocity function. 48

Table 3.1 Summary of the thickness of top two layers from resistivity
data. 83

Table 3.2 Velocity and thickness of the layers from refraction survey. 85

Table 3.3 Approximate depth of the bedrock in the site based on 1981
drilling. 87

Table 3.4 Bedrock depth from 1995 drilling programme. 88

Table 3.5 Survey control points. 93

Table 3.6 3-D recording parameters. 112

Table 3.7 2-D recording parameters. 115

Table 3.8 Summary of the seismic recording daily production. 119

Table 3.9 Spreads of the swath and the number of vibroseis shots fired
into each. 120

Table 3.10 TAR velocity function. 135

Table 3.11 Summary of approach-I processing flow. 145

Table 3.12 Processing flow for vibroseis and impulsive data - approach-
II. 151

Table 3.13 Sweep parameters. 158

Table 3.14 Synthetic sweeps parameters used for lab correlation. 158

Table 3.15 3-D vibroseis data statistics. 172

Table 3.16 Approach-I processing flows. 177

University Of Glasgow 1998 PhD Dissertation Page XXI

New Methods In Shallow Seismic Reflection
Z Z T Harith Tables

Table 3.17 1-D time velocity pairs used for NMO stack in approach-II. 182

Table 3.18 Approach-II processing flow. 184

Table 3.19 Approach-Ill processing flow. 197

Table 3.20 Approach-I, -II, and -III processing flows - similarities and
differences. 198

Table 4.1 Results of the refraction survey undertaken by the third year
students. 233

Table 4.2 The general sequence of the underlying seams together with
their average thickness and their approximate depth. 249

University Of Glasgow 1998 PhD Dissertation Page XXII

LIST OF PHOTOS
New Methods In Shallow Seismic Reflection
Z Z T Harith Photos

page

Photo 3.1 Aerial photograph of the site before the survey. 68

Photo 3.2 Exposures of rooms and pillars from an abandoned coal

mine. 71

Photo 3.3 Damage to buildings caused by underground subsidence. 72

Photo 3.4 No 7 pit, which has been exposed for treatment. 75

Photo 3.5 Mini-vibrator for vibroseis source. 105

Photo 3.6 Yellow coloured pegs mark source locations to avoid

confusion with the grid pegs (white). 109

Photo 3.7 Recording truck. 116

Photo 4.1 An abandoned mineshaft at the Todd Campus West site

during treatment for stabilisation. 226

Photo 4.2 Area of stripped topsoil, where grouting holes were to be

drilled. 234

University Of Glasgow 1998 PhD Dissertation Page XXIII

CHAPTER ONE

INTRODUCTION
New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

1.1 Seismic reflection - an overview

One of the greatest things about humankind is the ability to make use of bad
experiences as a starting point to a better life. The sinking of the ‘unsinkable’ Titanic in
1912 during its maiden journey after being struck by an iceberg, claiming thousands of
lives, was one of the greatest human disasters. In attempts to prevent the same such
disaster happening again, many inventions have been produced. Among them, Reginald
Aubrey Fessenden comes up with a great idea, in which he developed a model
submarine telegraph which had the capability to detect an iceberg. This was a turning
point in geophysical exploration. After some modification and improvement, in 1917 he
registered with United State Patents Office his method and apparatus for locating ore
bodies (Allen 1980, Sheriff and Geldart 1995). The basis of his method was “being
given the distance between two points in a mine, and having determined the time taken
by a sound wave to travel between the two points, it is possible to draw conclusions in
regard to the probable nature o f the rock between the two points, or if an echo be
observed, or refraction o f the sounds, it is possible to estimate the distance o f the
reflecting or refracting vein (Fessenden 1917)”. Later, this idea became a main concept
in the seismic reflection technique.

Beside being a tool for locating ore minerals, this technique was also capable of locating
salt domes, oil wells, etc., as recorded by his wife: “It is a great pleasure to record their
successful use in a totally different and important field - that o f locating salt domes, oil
wells, sulphur beds etc. (Fessenden 1940 in Allen 1980) ”.

Because of its capability in locating oil fields which were in demand during that era,
during the introduction of petrol engined cars and in search for a more environmentally
friendly fuel than coal, lots of capital was spent in improving this technique. The
computer revolution starting in the 1960’s accelerated the advance of this technique.
The introduction of the CDP surveying method and magnetic tape recording for data
storage were two of the milestones in seismic reflection history. Prior to 1970’s all
surveys were done based on 2-D line profiling technique. The introduction of the 3-D

University of Glasgow 1998 PhD Dissertation Page 1

New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

reflection method in the 1970’s was another milestone in seismic reflection technique.
Nowadays, 3-D reflection is common practice in petroleum exploration. Most
petroleum companies will require to shoot a 3-D reflection survey prior to deciding on
the location of wells. Thus the seismic reflection method has evolved from a very
simple technique to become the most successful geophysical technique of all. In nearly
80 years of constant improvement, it is now almost mature. The chronology of seismic
instrumentation and methods is listed in Table 1.1.

Table 1.1 Chronology of seismic instrumentation and methods (after Sheriff and
Geldart 1995).

1914 Mintrop’s mechanical seismograph 1952 Analog magnetic recording

1917 Fessenden patent on seismic method 1953 Vibroseis recording; weight drop
1921 Seismic reflection work by Geological 1954 Continuous velocity logging
Engineering Co. 1955 Moveable magnetic head
1923 Refraction exploration by Seismos in 1956 Central data processing
Mexico and Texas 1961-2 Analog deconvolution and velocity filtering
1925 Fan shooting method 1963 Digital data recording
Electrical refraction seismograph
1965 Airgun seismic source
Radio used for communication
1967 Depth controllers on marine streamer
1926 Reflection correlation method 1968 Binary gain
1927 First well velocity survey 1969 Velocity analysis
Transit satellite positioning
1929 Reflection dip shooting
1931 Reverse refraction profiling 1971 Instantaneous floating point amplifier
Use o f uphole phone 1972 Surface consistent statics
Truck mounted drill Bright spots
1932 AGC 1974 Digitisation in the field
Interchangeable filters 1975 Seismic stratigraphy
1933 Use o f multiples geophones per group 1976 3-D
1936 Rieber sonograph; first reproducible Image ray migration (depth migration)
recording
1984 Amplitude variation with offset (AVO)
1939 Use o f closed loops to check misties DMO processing
1942 Record section 1985 Workstation interpretation
Mixing 1986 Towing multiple streamers
1944 Large scale marine surveying 1988 S wave exploration
Use o f large patterns Auto-picking of 3-D volumes
1947 Marine shooting with Shoran 1989 Dip and azimuth display
1950 CDP method 1990 GPS positioning
1952 Medium range radio navigation_______ Today 4-D; Neural networking

The future development of the seismic reflection technique depends very much on the
price of oil. Today, with the help of computer technology, the research is concentrated

University of Glasgow 1998 PhD Dissertation Page 2

New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

in 4-D (time lapse 3-D) methods (Thomas 1997), and how to move into 4-C technology
(Greenberg 1997). 4-D technology is done by comparing a new set of 3-D data (shot
over an area that has already been shot with 3-D) to the existing 3-D data (Greenberg
1997). It is not strictly an exploration technique, as it is normally done in the oilfield
production stage to measure the changes in fluid flow. This can improve recovery and
make development and production of a field more efficient. 4-C technology on the other
hand is the latest technological development in petroleum exploration, where at each
location the reflected signals are recorded by 3 component geophones - attached to the
sea bed - for shear wave detection as well as by hydrophones (Greenberg 1997).

1.2 Shallow seismic reflection

1.2.1 Introduction

Beside its successful use in petroleum exploration, the application of seismic reflection
can also be found in engineering site investigations, hydrogeological studies, shallow
stratigraphy or even in archaeology, in other words for all manner of shallow subsurface
investigations. The term ‘shallow’ is here defined as all such work done using a portable
engineering seismograph with a simple source. Most of the time, the target for shallow
reflection work is less than 100 m deep. The reason for this branching away from ‘deep’
seismic methods is the advance in computer and electronic technologies, which have
reduced the size and cost, and at the same time have increased the capability of the
engineering seismograph.

Basically, shallow subsurface reflection investigators copy the technology from

petroleum exploration with some alteration to suit the shallow investigation needs. As
shallow investigation work usually proceeds with relatively small amounts of funding,
adopting the latest 3-D seismic reflection methods from the well-funded petroleum
exploration industry is almost impossible. Most of the shallow investigation
practitioners who apply the seismic reflection investigation methods have therefore

University of Glasgow 1998 PhD Dissertation Page 3

New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

limited their work to the use of the 2-D line profiling technique, some 30 years behind
the current reflection technology (3-D). Compared to petroleum exploration where
seismic reflection is now considered a mature exploration tool, in shallow investigation
work, it still has lots of room for improvement.

Shallow seismic reflection practitioners adopt the technique from petroleum industry
with the assumption it should give as good results as in petroleum exploration.
However, most of the time the results are not very promising compared to other simple
geophysical methods such as seismic refraction (bedrock mapping for engineering) and
resistivity (for hydrogeological and archaeological studies). Failure to get good results
sometimes leads to prejudice about seismic reflection: “why bother spending your
money if it is not guaranteed to work”.

1.2.2 Some problems

One of the greatest problems in adopting the technique originally designed for exploring
deep targets, covering huge areas and requiring lots of capital, to suit an investigation of
a shallow target in a limited area with a small amount of money, is the scaling down
factor. Most of the time the shallow seismic practitioners did not know how to scale
down a simple reflection survey to a specific requirement to utilise the advanced
techniques used in the petroleum industry (Bay Geophysical Inc. 1998). Beside the
scaling problem, the success of the shallow seismic reflection method also depends on
many other factors such as the source and receiver used, the acquisition design, noise
etc. But the most important of all is the ground condition.

In many cases, the shallow reflection surveys were designed based on ‘rules of thumb’
or so-called ‘experience’. This is a dangerous approach, as no two sites have the same
characteristics. Recorded signal and noise are very site dependent. Their characteristics
should be known prior to the design of the survey. Walk-away noise tests are one
technique where the signal and noise characteristics can be studied. Besides giving an

University of Glasgow 1998 PhD Dissertation Page 4

New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

idea of the signals and noise, a walk-away test also can be used to set the optimum
window (Hunter et al. 1984, 1987).

Ground conditions also control the mean frequency of the data. The uppermost earth
layer is imperfectly elastic in its response to the passage of seismic wave. Therefore
high frequency energy is lost while propagating through the earth. In some conditions,
most of the high frequency components are attenuated leaving only low frequency
components. Within a few metres of the surface, seismic wave velocities can drop as
low as 30 m s'1 (Rice et al. 1990) and the quality factor (Q) for seismic propagation
drops to values as low as 0.3 in unconsolidated, uncompacted soil which results in the
attenuation length (the distance over which amplitude decreases by 1/e) being as short
as 0.3 m for 100 Hz P-waves (Rice et al 1990). This leads to resolution problems. Since
shallow seismic reflection investigation deals with very shallow reflectors which may
appear as early as 10 ms, separation of the reflected from the direct or refracted signals
is almost impossible. This is the biggest problem in processing shallow reflection data.
In addition, the problem of static and lateral variation is also a challenge in processing
shallow reflection data.

1.3 The research

1.3.1 Research outline and aims

The nature of the uppermost few tens of meters below the earth surface needs to be
determined for several applications. For example, knowledge of this region is required
for underground water and mineral resources exploration, as well as foundation qualities
for engineering projects. Geophysics offers the tools for investigating this region. The
choice of geophysical tools used depends on the target. For example, the magnetic
method only can be used to detect a magnetic target, and the gravity method on the
other hand can only be used if there is a significant contrast in density o f the target
relative to its surroundings. Among all geophysical tools available, the seismic

University of Glasgow 1998 PhD Dissertation Page 5

New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

reflection method is the most advanced, and it is widely been used in petroleum
exploration. However, usage of seismic reflection in shallow subsurface investigation is
still limited.

For decades, seismic reflection was assumed to be useless for imaging the near surface.
The reason was that the normal seismic wavelengths used are too long to detect small
features or thin beds. Furthermore, near source shallow reflections were always masked
by noise while farther away they were hidden by refraction and surface waves.
Recently, major improvements in bandwidth, noise abatement and processing capability
have produced dramatic improvements in the capability of this method. These can be
seen by the increasing numbers of publications of the applications of 2-D seismic
reflection for imaging shallow subsurface as discussed in section 1.5.

Despite the increasing number of publications on 2-D seismic reflection for imaging the
shallow subsurface, there is still room for improvement. A list of possible research
topics relating to near-surface imaging was discussed by Steeples et al. (1997). Among
them are the application of 3-D seismic reflection methods for shallow subsurface
imaging (less than 150 m) and the application of seismic reflection (especially 2-D
seismic reflection) imaging methods to image very shallow (ultrashallow) subsurface.

This research was carried out in two stages. An initial stage involved testing the
suitability of the Oyo-1600 McSeis recording system for intensive or serious reflection
investigation. In this stage a 2-D reflection experiment with closely spaced geophone
stations was introduced to image an ultrashallow archaeological structure. The same
techniques were applied to the main stage of this research to establish the design
parameters for a 3-D survey. The experience in the early stages of the investigation lead
us to acquire a DAS-1 recording system to give a greater number of channels, dynamic
range, and sensitivity necessary for the 3-D investigation.

3-D seismic reflection surveys are widely used in the petroleum industry, but in general
have not been applied to the near surface. Publications on this topic (discussed with

University of Glasgow 1998 PhD Dissertation Page 6

New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

details in section 3.3, chapter 3) are very limited and the earliest date to only 1996.
Some of the problems with shallow 3-D reflection include the high level of cost and
effort necessitated by small geophone intervals as well as the difficulties posed by the
spatial aliasing of ground roll.

Like the 3-D shallow seismic reflection, ultrashallow 2-D seismic reflection surveys are
far from routine. Although several examples of reflection in studies the 3-15 m depth
range have appeared in some scientific literature, especially in archaeological work,
more case histories and continued research are needed to improve the capabilities in this
range, particularly for geological conditions in which GPR does not work well.

This research project is primarily about finding the best way of adopting the latest
successful 3-D seismic reflection technology, as used in petroleum exploration, to
investigating the shallow subsurface. With a small grant from the Glasgow
Development Agency (GDA) we (the Geophysics Group at Glasgow University) carried
out what is believed to be the first small-scale full 3-D seismic reflection survey over an
abandoned coal mine. The main objective of this work is to map out the rooms and
pillars of the mine. This technique is adopted from latest technology used in petroleum
exploration. Besides adopting the latest technology, this research also concerned with
developing a fast, simple and robust subsurface mapping technique that can be produced
during the 3-D fieldwork. No additional data collection is needed in this study as it only
utilises the first arrival times from the 3-D reflection data. This new technique is named
surface tomography as it uses the ideas of seismic tomography. These are two main
objectives of this research.

In archaeology, seismic reflection is the least popular geophysical investigation method

compared to others such as resistivity, ground penetrating radar or even gravity and
magnetic. Most of the time the target is not suitable for the use of seismic reflection. It
is either too small or lacks sufficient contrast in acoustic impedance. Despite those two
main objectives, the secondary objective of my studies is to image a very shallow
(ultrashallow) subsurface which is less than 7 m deep using highly condensed 2-D

University of Glasgow 1998 PhD Dissertation Page 7

New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

seismic reflection method. To prove that the seismic reflection technique can also be
used as an archaeological geophysical investigation tool, work was also done in locating
archaeological remains, a Roman Fort ditch. The aim of this work is to show that, given
the right conditions, seismic reflection can offer a better result than the other techniques.

1.3.2 Funding

These studies were sponsored jointly by the Malaysian Government (MG) and
University of Science of Malaysia (USM) as part of the USM Academic Staff Training
Scheme program for a duration of 3 years starting from January 1996 to December
1998. These sponsors, MG and USM, paid for tuition fees and a monthly allowance for
this period. The 3-D fieldwork was funded by a grant of some £13K from Glasgow
Development Agency (GDA) to the University of Glasgow. Without this fund, it would
have been impossible to continue this work.

1.4 Thesis presentation and organisation

The presentation of this thesis is based on the British Standard Institution

‘Recommendation for the presentation of theses (BS 4821: 1990)’. It is organised into
five chapters, where each chapter discusses different topics related to the shallow
seismic reflection work. After discussion (Chapter 1) of the brief history and the
advances in seismic reflection, the importance of seismic reflection for investigating a
shallow subsurface region, and the difficulties and aims of the project, Chapter 2
discusses the application of shallow seismic reflection method in archaeological work.
Although this is not the main project, this work came first chronologically, and only
involved simple fieldwork and modest amounts of processing time.

The main part of the thesis is the full small scale 3-D seismic reflection survey. This
work is discussed in Chapter 3. As we believe this is the first ever such survey, the

University of Glasgow 1998 PhD Dissertation Page 8

New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

discussion on planning, pre-fieldwork survey, data acquisition, processing, and

interpretation is discussed in great detail for future reference. Chapter 4 deals with
developing the new mapping method known as surface tomography. The last chapter
discusses the conclusions from all the work, the future of the shallow seismic reflection
method and also some recommendations for future improvement of such shallow
seismic reflection work.

1.5 Previous work

As the reflection technique used in shallow investigation work is an adaptation from the
more advanced technology used in petroleum exploration, it obviously needs lots of
modification and improvement before it can give as good results as in its original field.

First, the seismograph (recording system) needs to be as portable as possible without

compromising its capability. Advances in electronic and computer industry have helped
a lot in scaling down the huge recording systems used in petroleum exploration to a
portable machine. Today there are a number of commercial portable engineering
seismographs such as the OYO McSeis 1600, Bison, OYO DAS-1 etc. All these
commercial seismographs claim the capability of recording high-resolution seismic
reflection data and are suitable for shallow subsurface investigations. In choosing a
seismograph, the most important factor is its dynamic range - the ratio of the largest
recoverable signal to the smallest recoverable signal. The dynamic range of the
seismograph is linked closely to the number of bits used in the A/D (analog to digital)
converter. As each signal bit represents very nearly 6 dB (Knapp and Steeples 1986a),
by multiplying the number of A/D bits in the converter by 6, an estimate of the
seismograph’s dynamic range can be obtained. Besides the seismograph, the use of the
detector (geophone) is another important factor in any seismic reflection investigation.
Shallow seismic reflection needs high-resolution detectors. Most of the geophones used
in petroleum exploration are not appropriate for shallow work (Lepper 1981).

University of Glasgow 1998 PhD Dissertation Page 9

New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

Choosing appropriate acquisition parameters are also important in shallow seismic

reflection work, as the results strongly depend on it. The choice of recording parameters
is determined by the objectives of the survey, the solid geology, top layer conditions
and also the resources available (Knapp and Steeples 1986a). Pre-acquisition knowledge
of the site and/or a walk-away noise test is helpful in choosing the recording parameters.
In order to record reflected signals with minimal noise, Hunter et al. (1984, 1987)
introduced the concept of the optimum window. The optimum window is a range of
offsets (source to receiver separations) that allow the target reflectors to be observed
without interference from other events. It can be determined from analysing the walk­
away noise test record.

Almost anything can be used as a seismic generator (source) for shallow reflection
work, as long as it is capable of imparting sufficient energy to provide reflections from
the reflecting target, and is cheap, safe, portable and repeatable. In practice it ranges
from a simple hammer source to a complicated minivibrator. However, the selection is
dependent on the objectives of the survey, the target depth and also the ground
conditions. Pullan and MacAullay (1987) and Miller et al (1994) discussed the
characteristics of various sources on various sites, and found that the source
characteristics are largely site dependent. Rules o f thumb for source selection as
addressed by Miller et al (1994) are, ‘I f the mar surface is saturated and fine grained,
a downhole source is the best. I f it is hard and dry, weight drop should be a top choice.
Always try to bring several types o f sources but the first choice should be based on near
surface conditions, site restrictions and target o f interest’.

Ground coupling is another important factor in shallow reflection work, as bad coupling
can severely distort the quality of the recorded data. Distortion can cause the response
wavelet to be longer and more complex, and the changes in the wavelet character of key
reflections could be misinterpreted as a stratigraphic change (Knapp and Steeples
1986b). The coupling between receiver and ground can be represented by a damped
oscillatory system (Lamer 1970), whose properties depend upon the geophone mass, the
diameter of earth contact and local soil conditions (Hoover and O’Brien 1980). An

University of Glasgow 1998 PhD Dissertation Page 10

New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

efficient geophone ground coupling requires a lightweight geophone and a long spike to
firmly couple it into the ground (Hewitt 1980).

An understanding of the difficulties and their solution related to the shallow seismic
reflection work leads to its appropriate utilisation. The application of shallow seismic
reflection can be classified into three main categories, which are:
• Engineering site investigations (Hunter et al. 1984, Pullan and Hunter 1985, Hill
and Ali 1988, Miller et al. 1989, and Jeng 1995),
• Hydrogeology (Birkelo et al. 1987, Hunter et al. 1987, Miller et al. 1990, Johnson
and Clark 1992a, Johnson and Clark 1992b, Kidd et al. 1990, Clark et al. 1994), and
• Mining or void monitoring (Al-Rawahy and Goulty 1995, Briickl et al. 1997,
Kourkafas and Goulty 1996, Gochioco and Cotton 1989, Gendzwill 1990, Steeples
and Miller 1986).
Besides these, there is also work done in mapping mineral veins or even in locating
archaeological remains (Stright 1986, Karastathis and Papamarinopoulos 1997).

Most of such surveys are limited to the use of 2-D line profiling and using a
conventional compressional wave (P-wave) source such as hammer, weight drop or
buffalo gun. There was also some work (2-D reflection) utilising shear waves (S-waves)
as their source (Clark et al. 1994, Johnson and Clark 1992a, Jeng 1995, Pullan et al.
1990, Norminton 1990). The advantage of using S-waves as a source in shallow
reflection investigation is that the effect of water table (giving a false geological
reflector) is minimised. S-wave propagation is not influenced by the saturation level as
its propagation is controlled only by the shear modulus of the medium. The other
advantage is that the wavelength of S-waves is shorter than P-waves, which means that
they can resolve thin layers better than P-waves. One problem of employing S-wave
reflection, on the other hand, is that it may consume twice the time as the equivalent P-
wave survey, since it requires each shot-point to be shot in both directions
(perpendicular to the line) to obtain a difference record. Besides that, the data also need
to be recorded for twice as long as P-wave data for the same target depth as they have
travelled more slowly than the P-waves.

University of Glasgow 1998 PhD Dissertation Page 11

New Methods In Shallow Seismic Reflection Chapter One
Z Z T Harith Introduction

The problem of using the old-established 2-D line survey method is that it does not give
the whole picture of the area, and sometimes the image is unclear due to the lack of
resolution. Briickl et al (1997) and Kourkafas and Goulty (1996) both suggested that
their unclear image could be improved by running a 3-D survey. The suggestion from
these two projects led to the idea of conducting a 3-D reflection survey (within a limited
budget) over an abandoned coal mine. The essence of the 3-D method is areal data
collection followed by the processing and interpretation of a closely spaced data volume
(Brown 1996) which increases the resolution.

3-D is considered to be a very recent trend in shallow reflection investigation, although

the technique was introduced in the petroleum industry in early 1970’s. Corsmit et al
(1988) carried out a shallow 3-D survey on a tidal flat to investigate the shallow
stratigraphy. Their work was purely academic and pedagogic, as the prime objective
was to demonstrate to students how to conduct a 3-D survey and how to process it.
Smythe (1994, 1995) also conducted a 3-D survey to study the fault patterns in volcanic
rock to assess its suitability as a site for radioactive waste disposal. Smythe (1994) used
the same 3-D technique as normally used in petroleum exploration, however the target
is too deep (around 500 m) for this experiment to be considered as shallow reflection
work. The details of other preliminary 3-D experiments will be discussed in Chapter 3.

The latest 3-D survey conducted using portable seismograph was done by Miller et al
(1997) where they deployed a low fold 3-D survey to evaluate better lateral variability
of the interpreted faults at a sinkhole. The target (around 400 m) was also deep for an
ordinary shallow reflection survey. Bay Geophysical Inc. (1998) also claim to have
done a 3-D survey using a shear wave source in their web pages, however no published
article is available for reference.

At the time of writing (August 1998), Biiker et al (1998b) published a paper on the
same area ‘Shallow 3-D seismic reflection surveying’ discussing acquisition technique
and preliminary processing strategies in mapping a glacial sediments deposit. Details
of this paper are discussed in Chapter 3.

University of Glasgow 1998 PhD Dissertation Page 12

 CHAPTER TWO

SEISMIC REFLECTION IN ARCHAEOLOGY

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

2.1 Introduction

Excavation - coring and trenching - is a normal practice in archaeological investigation.

Their objectives are to acquire cultural and biological artifacts for subsequent laboratory
and statistical studies, and to document the three-dimensional relationships between
artifacts by using accurate, on-site recording procedure as the work progresses (Wynn
1986). The decisions on where to excavate are normally taken by the field archaeologist
based on historical documentation, surface artifacts, surface morphology or on aerial
photographs suggesting an earlier human occupation. In the absence of surface
indicators, random coring is done and based on that, trenches are dug. This routine is
time consuming. As time is often very precious in less well-funded projects like in
archaeology, fast and reliable investigation methods are needed. In addition to that,
those methods should be non-destructive. A non-destructive method is very important in
all archaeological investigation as it preserves the spatial relations between the artifacts.

Depending on the target, a geophysical investigation survey can provide an answer to

archaeological needs. It is fast, reliable and non-destructive. Because of that, most
geophysical survey methods have found some application in the archaeological context,
either as site exploration (the search for an undiscovered site) or as intra-site mapping -
used to guide excavation programs within an already discovered site (Wynn 1986).
Although many methods have been used in archaeological investigation, only a few are
frequently used. Magnetic, resistivity, electromagnetic (EM) and ground penetrating
radar (GPR) are the premier methods, while the gravity method is sometimes used.
Seismic methods - refraction and reflection - on the other hand, are usually found to be
unsuitable for archaeological investigations (Reynolds 1997).

In this chapter, I compare the results from two contrasting geophysical methods -
resistivity and seismic reflection - used in archaeological investigation. The first was
chosen to represent the most common geophysical investigation method in archaeology,
while the latter was chosen to represent an unsuitable geophysical investigation method

University Of Glasgow 1998 PhD Dissertation Page 13

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

in archaeology, even though it is one of the most successful methods in other fields such
as in petroleum exploration.

2.2 Aim

The aim of this project is to locate a Roman fort ditch (vallum), which was filled and
covered by surface material. We use a spatially highly condensed seismic reflection
method. The results from the seismic profile were then compared to the results of the
resistivity survey done earlier by Baker (1996). Besides locating the target, it is also
hoped that it will give a new perception about this method in archaeological
investigation, from being considered as an unsuitable investigation method to at least
being thought of as a secondary if not a primary investigation method.

2.3 Previous work

Not many archaeological investigators are keen on using the seismic method, as it is
expensive and involves a complex field procedure compared to other geophysical
investigation methods (Wynn 1986). Furthermore, the success of this method depends
on the presence of a contrast in elastic properties between the target and the
surroundings. The bigger the contrast the higher the chances are of success. The size of
the target also plays an important rule in the success of seismic reflection work. In most
cases the target is either buried metal, foundation stone or man-made structures, and is
thus too small for a normal shallow (hammer source) reflection survey. Only certain
targets which are big and show a significant contrast in elastic properties are suitable.
Goulty el al. (1990), and Goulty and Hudson (1994) employed the refraction method to
delineate the 8 m wide Hadrian Wall’s vallum, while Karastathis and Papamarinopoulos
(1997) applied both refraction and reflection techniques in detecting the 35 m wide
King Xerxes’ canal. Both of these works showed promising results.

University Of Glasgow 1998 PhD Dissertation Page 14

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

Baker (1996) compared the results from several geophysical surveys such as seismic
refraction and reflection, resistivity and magnetic methods in detecting the Roman fort
ditch at Inveresk. Her failure to get a good result in the seismic surveying methods led
to a conclusion that this technique was not suitable for the detection o f the vallum,
compared to the other two methods.

This work is a continuation of Baker’s project. However, instead of applying both

refraction and reflection as Karastathis and Papamarinopoulos (1997), a highly
condensed reflection survey was employed. The results are then compared with two
resistivity profiles reinterpreted from Baker’s work.

2.4 The site

2.4.1 Introduction

Discussion on this subject is based on three archaeological papers by Dennison and

Coleman (1996), Richmond (1978), and Baker (1996). As the nature of this study is an
application of geophysical techniques in archaeology, I have no intention of explaining
in detail the excavation history that the site underwent and its findings, except to give a
general perception of the site history and its relation to the target. Furthermore the
author has no basic background in archaeology.

2.4.2 Location

Inveresk is part of modem day Musselburgh. It is situated within the East Lothian
District of East Lothian Region approximately 10 km east of Edinburgh, Scotland
(Figure 2.1). The fort at Inveresk lies on the slightly higher ground facing the Firth of
Forth, surrounded by the river Esk on three sides.

University Of Glasgow 1998 PhD Dissertation Page 15

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

Today, nothing of the fort itself is visible on the ground. Most of it is buried under the
ground of St. Michael’s Kirk. The field where the surveys were undertaken is also part
of St. Michael’s Kirk. It is a reserved cemetery ground. Physically it situated on the
gentle westerly sloping flank of the hilltop.

2.4.3 Geology

The south of Scotland can be divided into three major zones, which are the Highlands,
Midland Valley and Southern Uplands. Two major faults, the Southern Upland Fault
and the Highland Boundary Fault separate the two outer zones, and the Midland Valley
is a graben bounded by those two faults. The Highland zone is made up of metamorphic
rock while the Southern Uplands are formed mainly of folded Ordovician and Silurian
strata. On the other hand, younger rocks such as an Old Red Sandstone and
Carboniferous sandstone and shale are the main rocks of the Midland Valley
(MacGregor and MacGregor 1948). Around the Edinburgh area, the youngest rock
recorded after the Carboniferous rock is Quaternary glacial deposits, which implies that
this area was subjected to weathering and erosion for a long time.

During the last two million years, most of northern Britain has been subjected to
repeated glaciation periods. The latest was 15,000 years ago. The meltdown of the ice
produced a vast amount of running water, which transported glacial debris before it lost
its energy and settled on top of the Carboniferous rocks. The meltdown also raised sea
level by a few tens of metres. Most of the lowland areas were submerged and the beach
deposits were deposited further inland, which later became raised a beach deposit after
the sea level fell again.

The geological events around Edinburgh and the Midland Valley can be summarised as
in Table 2.1. Musselburgh is situated in this Valley and is developed on raised beach
sand and glacial deposits.

University Of Glasgow 1998 PhD Dissertation Page 16

 r"
Chapter T w o
In A r c h a e o lo g y

<
u0
&
K
J
Cl
Seismic Reflection

Ss
00 T3
a <U
J—( a

■>S w
c /d ^
cd
u r
n ^*“ •
til u
f—
i >
JC
a, cd
C
cd
C C/D
cd Li
0) o
C/D
C+H "S
o
o
o CL
3
o cO
cd
dh
O 00
o
CL
O
o
C/D
£
(N o
a>
u, >
g -s
op •|*3 s00
uH
CL
cd

"O
Reflection

*y—s'
<u
Q
Seism ic
M ethods In Shallow

£
O
CDU
m
J2
o
<*-
o
Z Z T Harith

£
(/)
<U
>
c
New

D
 o

Page 18
£ 00

* ■ 3
fe g 3* XI
X)
O. P o»-
_>
t:
41
C/3
cd
00
Cd O
m Vi 00
e* 3 3
H
C
O
^
5+1
o
C/3
“
l - s ,
3
"-3
CD O
.a ^
►3 C, 41 C
Ofl 03 41
a O h

Q 41
C/i

&3
3 3
O
2
sv
o
m
3
£ 13
§ 2 +->
13
C/3
■4-*
§
fc
w ^ cd
> *§ S C/3
2 S rt cd
l 2a
■41
CO
03
J
cd
C/3
>
cd

c
o
I
*—* >
03
J
A

U
O h
C/i

co
•5
3
O
& B C/3
cd
01 §

1998 PhD Dissertation

O 13
"3 Si
cd
§ 13
+-»
.a 41
is •2
C/3 41
o
13 g 3 ^
c 3S;
.a >
c$ .a
CO o
o, 2® c3 3
4> O
<d
a 53
Q O
§ 411
u
+-H
a !i " *3
41 C
C/3
13
3 41 41 41 41 41 41
s T3 cd ” cd
I- o .2 .2 ^ 2
■i .a .a .a .a .a .a
8 1 2
> O ’>
e 3 o 'o
O rt .s
! s 2 •§ 8 .
03 41 c3 g a> 2a g b b b c3 c3 c3
EGO S n a C/i

S C/3
55 4i
<U 41
L- 1>
g S
13/1 sb
cd J2 tS M
•fj 13
41

13
31
|
CO
lg .§ C/l C/i
3 — 5 4) 03 j *2
13 TO
g c/i .5 > 41
13 W > «8 >i § S o op ! *
o
41 •S b I & is C/l
t3 13
53
£
S
a§ 3* .S
^e « H
2 § .S M
41
.3 ll O O
C*
«2 ■s lS lo § < 2
b §
C/i 1OQ2CL, u+
s S | c/il S| C§Q bc3 -cd O £h
Oh O
s CQ £ U U G j
41 <N
C/l

| I
<u ”T3

58
41
t41o O
41
a
University Of Glasgow

Ld -4—»
H s ■&
? J5 g "O o>
£ g-s
Oh
41
Oh U^
g3 c&
o c/i£ c/i 13

13
O o
X3 Ch ’ o Vh O
"S 5 N 41 N
Oh O
S
*
W
B
H
r v 41
►=T cd
Id
I §
- '- 3
4) Oh Oh
z
New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

2.4.4 History

The settlement of Scotland took place around 7000 BC (in the Mesolithic or Middle
Stone Age), however no evidence for humans’ existence in this area has been found.
The earliest evidence found, such as a stone axe for clearing woodland (SE of
Musselburgh) and the stone macehead used for ceremonial purposes (at Woodside
Garden, Musselburgh) suggest that people began to settle around this area at around
3500 BC, which is the Neolithic or New Stone Age (Dennison and Coleman 1997). The
evidence found (refer to Ritchie and Ritchie 1972, for details) shows that the
settlements around Musselburgh have continued until today. Among the archaeological
findings, the findings from an early Iron Age settlement are dominant. This settlement
continued until the Romans arrived in the first century AD.

The conquest of Britain by the Romans started in 43 AD with the invasion of Emperor
Claudius. Under the general Gnaeus Julius Agricola (Ritchie and Ritchie 1972), they
reached as far as the northeast of Scotland and circumnavigated islands during the first
century of campaigns in Scotland. These campaigns met with much resistance by the
local people, and the Romans were forced to retreat from most parts of Scotland in AD
87 - 88. In order to keep their southern region (Britain) away from the northern
rebellious factions, Hadrian’s wall was built in 120 AD. During the invasion, forts were
also built as part of their defensive and support system. After their retreat, most of the
forts were abandoned or demolished.

After the death of Emperor Pulius Aelius Hadrianus (Hadrian’s wall was named after
him), his successor - Emperor Antoninus Pius - ordered Lollius Urbicus, the governor
of Britain at that time, to recapture Scotland. A second barrier - the Antonine wall - was
also built stretching for 60 km from the Forth (Edinburgh) to the Clyde (Glasgow) for
the same purpose as Hadrian’s Wall. Associated with that, forts were also built. The fort
at Inveresk was built during this time. As it is located on the mouth of the Esk, it is
believed that this fort functioned as a port or a harbour.

University Of Glasgow 1998 PhD Dissertation Page 19

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

2.4.5 Target

No two Roman forts are the same (Breeze 1983), however most of them shared basic
common features. For example, a fort is normally rectangular in shape, and faced either
the enemy, the line of march or east. The forts are surrounded by a defence system
which usually consists of ramparts of tuff or stone with at least two ditches
(Collingwood and Richmond 1969, Breeze 1983).

The Roman fort at Inveresk was first investigated by Ian Richmond in 1946-47, when
he established the lines of ramparts on the north, south and west sides, enclosing an area
of around six acres with the main entrance facing to the west. These ramparts are made
of beaten clay standing to a height of 1 m, edged by a neat kerb of fresh hammered
dressed sandstone blocks, and is approximately 7 m wide (Richmond 1978). In one
trench on the west of St. Michael’s Kirk, the kerb rampart front was discovered at 2 m
depth.

The ditches were normally dug beyond the ramparts (Breeze 1983). The excavation on
the southern wall revealed the outer scarp and the bottom of a V-shaped ditch some
10 m wide. The ditch was dug in a yellow sand and clearly outlined by a filling of
chocolate-coloured silt and by a coating of red clay almost 10 cm thick (Richmond
1978). Only one ditch was found, which is rather unusual for a Roman fort defence.

Between the rampart and the ditch lies the intervallum or roadway. Two superimposed
intervallums were found at the site. The lower is 4 m wide, while the upper is 11 m
wide. These intervallums were found lightly gravelled, standing upon an artificial
terrace of clay 60 cm thick (Richmond 1978). Based on his excavation records, the fort
plan and a cross-section of the defences are shown in Figures 2.2 and 2.3, respectively.

Our target is to detect the V-shaped ditch, which is part of the defence system as
described above.

University Of Glasgow 1998 PhD Dissertation Page 20

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

CD
C /J

Cl

(N
L4

CL.
Ta3>

<N

X
5 £
LL to
C
l

c
oo
fN

£o
ON
On

D
a

O
GO
Q
Q
GO
<D
o

Page 22
E s
b |r a
4- -
D - _C
cd CJ

U<
c
co

oo
r-
o\
T3
a
o

$—«
CD

cd

cd
Q.

£
e
<d

t/D
-t-J
<d

1998 PhD Dissertation

<d

c
C+-I
cd
T3
CD

+-*
C+H
O
C
_o
%->
o
<d

cd
00
CD

o
GO
O
m
cd <n
CD
S-H
3
CJD
• T—I
Ph
University Of Glasgow

■£•§
_C u.
New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

2.5 Resistivity survey

2.5.1 Introduction

Resistivity surveys are one of the most common geophysical surveying methods
employed by the archaeologist, simply because they are both cheap and easy to conduct.
However the interpretation of resistivity profiles sometimes can be very difficult. To
discriminate the interpretation’s ambiguities, good geological or non-geological control
is required.

Here, two resistivity lines - L3 and L4 - were surveyed by Baker (1996) prior to the
seismic reflection work. L3 was acquired on the same line as the planned seismic line
whereas L4 was acquired 10 m to the north of L3 and parallel to it (Figure 2.4). As a
control, L3 was designed to run through two man-made features, an archaeological
trench at the eastern end and an area of stripped topsoil at the western end.

In order to obtain a 2-D model (i.e. a picture) of the subsurface, an expanding Wenner
array with a moving central point (a combined of vertical electrical sounding (VES) and
constant separation traversing (CST)) was used. Since the target is less than 4 m deep,
the readings were read at only 1, 2, 3 and 4 m electrode spacings before the central point
was moved down the line by a 1 m interval.

2.5.2 Interpretation theory

The apparent resistivity reading is controlled by the presence of conductive material

such as water or metal. Since our target is the buried Roman ditch, we would expect a
low resistivity anomaly because the ditch had subsequently been filled up, normally by
loose material, and such material traps a quantity of conductive water.

University Of Glasgow 1998 PhD Dissertation Page 23

New M ethods in Shallow Seism ic Reflection C hapter Two
Z Z T Harith S eism ic Reflection In A rchaeology

Figure 2.4 Details o f the study area.

South fence
130 -

120 -

110 -

100 - Trench
I
90 -

S3 80 H 00
cs
£ -e
• **-w
o
s P H
o 70 - i, 1991 Trench
;
/ . <
D o0
c
• J U

: cj
•
0
>+—
60 - -t—•
Co CO
o 0
....:D a
50 -

40

30 - - [mm C Area of
Se/c stripped
s'$tivi topsoil
20 -
ltyline(L3~) 1
Re ■ ■
10 -
Slsti* ty line

0- North fence

10 20 30 40 50 60 70 80
meters (from wall)

Note: A, B, C. D and the shaded area are GUARD trenches.

U niversity O f G lasgow 1998 PhD D issertation Page 24

N ew M ethods In Shallow Seism ic Reflection C hapter Two
Z Z T Harith Seism ic Reflection In A rchaeology

Figure 2.5 shows an exam ple o f how the resistivity profiles over the buried ditch
normally look like. We would also expect to see a high resistivity anomaly on the
profile due to the presence o f the dressed sandstone at the edge o f the clay ram part and
the heavy paved berm.

Figure 2.5 Resistivity profiles (CST) across a buried ditch. N ote that relatively
low resistivity value on the right (10 - 20 m) is associated with poorly com pacted infill
materials (ditch)(ffom Kearey and Brooks 1996).

36

Apparent
Resistivity 26
(Q m) 36

20 m

The effect o f the topography and the near surface resistivity variation should also be
taken into consideration during the interpretation o f resistivity profile, since these can
mask the deeper resistivity values.

2.5.3 Results and discussion

The resistivity readings were plotted on an Excel spreadsheet graph for qualitative
interpretation before transfer to the Generic M apping Tools (GMT) program for
presentation purposes. In general the profiles for L3 (Figure 2.6a) and L4 (Figure 2.6b)

U niversity O f G lasgow 1998 PhD D issertation Page 25

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

show resistivity values following the same pattern as the topography, which is
decreasing towards the western end of the lines. This is a good example of how the
topography can mask the effect of the deeper variations.

The 1 m electrode spacing profile on both lines shows the most variable profile of all.
The reason is simply that it reads the resistivity values of a highly disturbed topsoil. As
the target is deeper ( 2 - 4 m), the variation of this profile is less significant than the
variation in the deeper profiles. This shallow profile is important for checking the
interpretation validity of the resistivity profiles.

The interpretations of the resistivity data are made based on the low resistivity value on
all the profiles on lines L3 and L4. Comparing the resistivity profiles on both lines, four
resistivity low anomalies were identified, namely A, B, C and D (Figure 2.6a and 2.6b).

Anomaly A is picked up on line L3, and is characterised by a narrow and strong low
anomaly. This anomaly, however, is absent on L4. Anomaly B, on the other hand, is
clear on L4 but very subtle on L3. It is characterised by a strong but broad low anomaly.

Anomaly C appears on both lines L3 and L4. It is also a broad anomaly; however it is
not as strong as anomaly B. It only appears on the 1 m spacing profile on both lines.
Anomaly D only exists on the 1 m electrode spacing profile of L3. It is a very weak low
anomaly.

Note that the ‘strength’ of the anomaly as referred to here is based on maximum depth
at which the anomaly can be detected. A ‘strong’ anomaly means that the anomaly can
been seen on the deeper profile (2, 3 or 4 m electrode spacing profiles); on the other
hand, a low or weak anomaly only appears on the 1 or 2 m electrode spacing profiles.
The summary of the low resistivity anomalies is given in Table 2.2. All these anomalies
are difficult to explain unless they can be related to a physical control.

University Of Glasgow 1998 PhD Dissertation Page 26

 t-~
Chapter Two
Seismic Reflection In Archaeology

(N
<U
GO
03
Cl

J 1 I I L J I I L J L
<p a? s
so
CL C l
U) Q)
GO
V3
a a
UJ Ui SO
VI

00 T VI

Vl
o
Vi

so <u
't—>
0>

C/3
jo
T3
2 U
o o C/3
J-c C/3
Q<
>% O
G
O C/3
,H-v
'>
SO £
03
C/3 u
<
y> fi
—
(3 Q
tt
■pH Q
hJ -C
Q a.
(I) oo
J3
■*-ji O'
O'
<u
S3
vq n c3
<N so N
0) T3
3
GO
E Q
M
u
a PQ
G
O <
New Methods In Shallow Seismic Reflection

"St
I"*
<N

o
»"H

G
O

<o
LU £
o
“T T r t—i— r TO T 00
T T C/3
-2
8 3 i8 S 8 1 ? P !8 8 !fiS !5 ^ S ^ !S S VI a-<
L
o
Z Z T Harith

(jei©LU LULio) At!A)|as©d

c
D
 oo
Seismic Reflection In Archaeology
Chapter Two

<N
<u
00
cd
CU

m cp

a. a

cn <»

■>

ss
a>
w

J£ T3
s a>
o*-h C/5
C/5
Cl 3
oC/5
I I
C/5
a>
C/5
<JJ
s—
io Q
— 1 §
Q
<D oo
-t—> O
O'
<U
X
so3
tN N
<D T3
3
V-H

CO
E
U
CQ
Seismic Reflection
New Methods In Shallow

£
o
00
a
o
Z Z T Harith

(jejeuu-uuiio) AiwsjSBa £*

c
New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

u
Anomaly A AnomallyB Anomal Anomaly D
L3 L3 L4 L3 L4 L4

cn
h-1
X

~7-
Appearance

m
m

X i
m
m
Centre of the
■

I* <N
T—

On

On
H
anomaly (m)

i
Width narrow wide wide wide wide narrow
I i

m
'3'

o-
Max. depth of resis. 1 or 2 1 or 2

1
minima (m)

i
I

O-
Strength strong strong weak weak weak

University Of Glasgow__________________________________________ 1998 PhD Dissertation Page 29

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

Figure 2.7 shows the relationship between these anomalies and the field. Anomalies A
and D on L3 lie directly over the trench C and an area of stripped topsoil respectively.
These were the control of the interpretation. The trench was 2 or 3 m deep and it is
filled by loose excavated materials.

Anomaly A is very prominent on 1 and 2 m electrode spacing profiles, however the

anomaly is weak on the other two (3 and 4 m electrode spacing profiles). This anomaly
is believed to be a result of the trench C. The appearance of the anomaly on the deeper
profile (3 and 4 m electrode spacing profiles) is an example on how a strong near
surface resistivity reading can affect the readings of the deeper section.

Anomaly D (on L3) is also believed to be a result of the stripped topsoil area and this
was confirmed with the result of the twin probe resistivity survey (Baker 1996). Since
only the topsoil was stripped off, we would expect this anomaly to be only picked up by
the shallow profile (1 m spacing profile) and not by other profiles as shown in Figure
2.6. The reason why the anomalies A and D did not appear on the L4 profile is simply
because that line was deliberately chosen to avoid these two known man-made
structures.

Anomalies A and D can be confidently interpreted since they have good control. But the
anomalies B and C are still unexplained. These anomalies lie in the area where the
target ditch was postulated. From the excavation in the nearby area, only one ditch was
found, and the idea of a Roman fort with one ditch is always questionable, since most of
them had two or more ditches as their defence.

Can the pair of anomalies B and C be taken as a proof of the existence of two trenches
around this fort?

University Of Glasgow 1998 PhD Dissertation Page 30

N ew M ethods In Shallow Seism ic Reflection C hapter I wo
Z Z T Harith Seism ic Reflection In A rchaeology

Figure 2.7 Relation between resistivity anom alies (A-D and Z) and their near
surface features.

"■TO -**

60
_ —^ 50
30

U niversity O f G lasgow 1998 PhD Dissertation Page 31

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

Without a careful analysis, we might agree with the idea of the existence of two ditches,
since the anomalies B and C are very strong on both 1 m spacing profiles on both lines;
however this idea contradicts what the field evidence shows. On the other hand, if we
agree with the idea of a single ditch, which of these two anomalies represents the ditch,
and why are there two low resistivity anomalies?

Anomaly B is a wide and strong low anomaly. It appears on all profiles of L4, however
on the L3, only the shallow profile shows the effect. The other profiles in that region are
undulating, and the recognition of the anomaly is very difficult. By comparing the
anomaly B on L4 and anomaly A on L3, we can say that this anomaly results from the
deep trench-like structure. However, the difficulties in recognising any anomaly on the
deeper profile on L3 may indicate that the structure is not continuous. Comparing
anomaly C on both lines to anomaly B on L4 and to the anomaly D on line L3, we can
say that this anomaly also results from the trench-like structure. However the maximum
depth of this structure is less than 1 m, since the deeper profile did not show the same
anomaly.

Although a trench-like structure which gave anomaly C is continuous (based on its

appearance on both lines), it is unlikely that it can be the Roman ditch since it is too
shallow. Since anomalies A, D and C cannot correspond to the Roman ditch, then the
ditch must correspond to anomaly B. However, the problem with anomaly B is that it is
discontinuous - i.e. there is no clear indication of the anomaly on line L3.

The other interesting anomaly is a prominent high resistivity anomaly - anomaly Z

(Figure 2.6a and 2.6b). This anomaly is very narrow and its centre is at 14 m on L3 and
16 m on L4. This anomaly can be seen on all profiles on both lines L3 and L4.
Normally, a high resistivity anomaly results from the presence of less conductive
material such as a rock boulder. From the knowledge of the area, this anomaly may
correspond to the dressed sandstone in front of the clay rampart.

University Of Glasgow 1998 PhD Dissertation Page 32

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

In conclusion, although there is good excavation control, the interpretation of the

resistivity data is still subject to many ambiguities. The reasons why resistivity is very
popular amongst archaeologists are simply because it is cheap and easy to operate.

Although the interpretation of the resistivity data is ambiguous, it can however give a
rough idea of where to dig a trial pit. To overcome those ambiguities, other geophysical
surveys should be conducted in the same area. Since the seismic reflection method has
shown great success in imaging the subsurface for engineering and hydrological work
(not to mention its huge success in petroleum exploration), it also can be used in
archaeology if the conditions are right.

2.6 Reflection survey

2.6.1 Can reflection work?

The success of the seismic reflection survey method depends on the acoustic contrast
between layers. Here, our target is to locate the ditch which was dug into the yellow
sand. The target was probably filled by building stones (archaeological remains) or
other materials and covered by loamy topsoil. Excavations done in 1993 (the 1993
trench shown in Figure 2.4) by the Glasgow University Archaeological Research
Division (GUARD) team shows that the thickness of the loamy material (top soil) was
less than 0.75 m. The archaeological remains sat on loose sand, which was believed to
be a beach deposit of Quaternary age (Figure 2.8).

In another area (west of St. Michael church), a rampart front was found at
approximately 2 m depth (Richmond 1978). Depending on topography, the thickness of
the topsoil varies. Normally it is thinner on the higher ground relative to the down slope.
As the site slopes to the west, the thickness of the eastern part (near to the cemetery
wall) is thinner compared to the end of the line (down slope).

University Of Glasgow 1998 PhD Dissertation Page 33

 w

u <
9- -c
cd cj
b |C3
o >.
[E J 1

£
(-1
<
oo
C4

.SP
D
o; n>

University Of Glasgow 1998 PhD Dissertation Page 34

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

The loamy topsoil has a lower acoustic impedance (the product of density and velocity)
than yellow sand. As long as there is a contrast in acoustic impedance, the reflection
technique can be applied. However, the success rate is much higher in the area where
the contrast is bigger. Besides the necessary acoustic impedance contrast, the target size
and geometry are also important. The horizontal dimension must be bigger than the first
Fresnel zone otherwise it is undetectable, since the major contribution to the reflected
signal comes from this zone (Sheriff and Geldart 1995). The width (radius) of the first
Fresnel zone ( w ) is controlled by the dominant wavelength (A ) and the reflector depth
( z ) and their relation is

w
2 )

For the target of less than 3 m depth and if the normal mean frequency produced by a
hammer source is 80 -100 Hz, the first Fresnel zone is only about 1.8 m in radius. As
the target is wider than first Fresnel zone, horizontally it can be detected by the
reflection method. The success of the reflection method in detecting King Xerxes’
Canal (6-8 m below the surface, and 25-35 m wide) using a sledge hammer as a source
by Karastathis and Papamarinopoulos (1997), shows that our target should also
detectable.

2.6.2 Data acquisition

As the aim of this work is to detect the Roman fort ditch using the reflection technique
and to compare it to the resistivity profile, the decision was made to lay the reflection
line along the L3 resistivity line (Figure 2.4). Before any recording was done, the line
was first sighted for elevation. The sighting were done using a Sokkisha Automatic
Level (Model C3E) and only elevations relative to the observation point (assumed as 0
m) were read. The actual height of the observation point is 37.82 m above sea level.

University Of Glasgow 1998 PhD Dissertation Page 35

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

Topographically, the site is sloping gently toward the west without any variance.
Because of that, only four points of elevation (one at each end and two in the middle)
were sufficient, as shown in Table 2.3. The elevations for other points along the line can
be interpolated from these four points by assuming a linear relation.

Table 2.3 Elevation of the reflection line.

Distance ( m) Relative Elevation ( m)

0 0.82
30 -1.16
40 -2.18
50 -3.24

The data were recorded using a 24 channel OYO McSeis-1600 (version 3.1)
seismograph. One of the advantages of this seismograph is that it can facilitate high­
speed multicomponent sampling with high resolution. 48 Hz low cut and 1000 Hz high
cut pre-filters were applied during the recording. This range was chosen to allow the
seismograph to record its maximum possible range of frequency. The 1000 Hz high cut
was also chose to eliminate the frequency fold-back problem (aliasing) which can
deteriorate digital signals. Low frequency noise was further reduced by using single 100
Hz high-resolution geophones.

The signal was generated by hammering a 3 kg steel hammer on a steel plate. It is the
cheapest and commonest source in shallow seismic exploration. It produces peak
amplitudes in the frequency range 80-100 Hz. Each record is displayed on the OYO
display unit and depending on a visual estimate of quality, it can be stored in memory,
deleted or stored directly to floppy disk. The signal stored in memory can be stacked
with the later records, if necessary. To improve the signal to noise ratio, a stack of 6
good shots was generally recorded at each shot location.

Depending on the offset (how far it is from the source) and the reflector conditions
(acoustic impedance contrast), the energy of the signal detected by the geophone varies.
The near offset geophones normally detect high signal amplitudes, which most of the
time are so high that they mask other signals on the display of the seismograph, while

University Of Glasgow 1998 PhD Dissertation Page 36

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

the far offset geophones signal levels is so low that it sometimes hardly seen. To
overcome those problems, appropriate gain was applied on the all channels. The near
offset channels were assigned a lower gain in order to scale down the high energy signal
while the far offset channels had a higher gain in order to amplify the signal.

The survey was designed to acquire a very condensed dataset using a 24 channel
seismograph across a 50 m wide field. For this, a common depth point (CDP) jump-roll
geometry was designed, where 24 geophones equally spaced at a 0.25 m spacing and 13
shots at every 0.25 m starting between geophone 6 and 7 were shot. Then geophones 1
to 12 were moved down the line (to become geophones 25 to 36) (Figure 2.9). Each
spread was only 5.75 m long and moved 2.75 m down the line at each ‘jump’. 208 shots
were recorded on the line, giving a total of 4992 traces. The subsurface reflecting
points, which are at half of the receiver spacing are thus every 0.125 m, with maximum
coverage of 13 fold (1300%).

In order to get the highest resolution possible, the signal was sampled at every 0.1 ms
interval for 102.3 ms. 102.3 ms is known as the record length and is controlled by the
pre-triggering record length set (1024 samples) and the sampling rate. The pre-
triggering record length is fixed depending on the seismogram, and for the OYO
McSeis 1600 it can be selected between three values of 1024, 2048 and 4096 samples.

As the signal is sampled at a very small interval (0.1 ms), the sampling frequency is 104
Hz. The Nyquist frequency, the maximum frequency that can be recorded before
aliasing occurs is defined to be half of sampling interval, i.e. 5000 Hz. However by
applying the 1000 Hz high cut pre-recording filter, aliasing, which folds back
frequencies higher than Nyquist frequency, is unlikely to occur.

University Of Glasgow 1998 PhD Dissertation Page 37

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

so
J-t

CM
g
<

CJ
0)

(X

On
Oh
00

Uh
o00
<L>
D

Q.

Q.
Q.
co
O

Q.
™
to t
>> d
in
CM

University Of Glasgow 1998 PhD Dissertation Page 38

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

(N
<D
Spread type Jump roll
Geophone 100 Hz single element vertical component ‘carrots’
Source 3kg hammer and steel plate
Geophone spacing 0.25 m
Source spacing 0.25 m

m
Nominal fold of coverage
Record length 1024 samples
Sample rate 0.1 ms
Record length 102.3 ms
Pre- filter 1000 Hz high cut
48 Hz low cut

University Of Glasgow__________________________________________ 1998 PhD Dissertation Page 39

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

2.6.3 Seismic data processing

The data were processed using ProMAX/2D (v 6.1) industry standard software on a Sun
SPARCstation-20, with the exception of figure 2.13 which was produced using
SierraSEIS processing software.

Before any processing was done, the correct FFID (field file identification number)
from the observer’s log sheet was added to each file. Noise tests and other non­
production files were edited out. Then the geometry was loaded from ProMAX/2D
database spreadsheet and stored in the header of each trace. Once the geometry was
installed the noisy and bad traces were killed and any reversed traces were corrected.
Shot gathers show that the near offset data were dominated by refracted waves and
ground roll and it is hard to recognise any reflection signature, on the other hand, the far
end records shows an interesting reflection-like signature at around 20-30 ms TWT (two
way time) (Figure 2.10). This reflector-like signature shows as a linear event in some
records but in most of the records it is a slightly hyperbolic curve. Examples are found
in source index number (SIN) 135 and SIN 161 in Figure 2.10. The reason is that
refraction arrivals are strong and mask the reflection signal since the target is very
shallow.

The length of the spread also contributes to this problem. Since the length of the spread
was only 5.75 m, only a small part of the hyperbolic reflector event was recorded. The
longer the offset, the better the hyperbola that is recorded.

In order to get a better hyperbolic curve, the edited data were sorted into common depth
point (CDP) gathers prior to any further processing. By doing this the reflection signal
can be confidently recognised especially on the western end gathers such as CDP 330 in
Figure 2.11. The CDP gathers also show the same as shot gathers, where the near
offset traces was dominated by ground roll and refraction arrivals. Only the far offset
traces on gathers at the western end of the line showed a clear reflection signature (e.g.
CDPs 303 and 330 in Figure 2.11).

University Of Glasgow 1998 PhD Dissertation Page 40

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

£
V
o
■4—•

x:

h
C/3

<
d
xi<L)

360
<N
e60

D
University Of Glasgow__________________________________________ 1998 PhD Dissertation Page 41
New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

U
Q

00
Ph

CN
a>

<L>
oC5D

i ii111nt in huhuni iiiiii nun in

11innii hi imiii
(sui) auiji

(sui) amil
^510877831
University Of Glasgow__________________________________________ 1998 PhD Dissertation Page 42
New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

Frequency analysis of the CDP gathers shows that energy is concentrated in the 80 - 120
Hz range (Figure 2.12). F-K analysis of the shot gathers using SierraSEIS processing
software shows the data in this domain are dominated by early arrivals (Figure 2.13).
These early arrivals were a combination of refraction and reflection arrivals as the
reflector is very shallow. It was impossible to separate these arrivals in the F-K domain.
An F-K filter was therefore not applied, in order to avoid the removal of the reflected
signals.

The first arrival was only muted out on the gathers where the reflection signal was
clearly recognised (Figure 2.14). In the noisy gathers, or in gathers where the reflection
signals was questionable, the first arrival mute was applied based on the nearest clearly
recognised reflection signals. This was to ensure that no reflected arrivals were muted
and to ensure that only the refracted arrival were muted.

Reflection times on the seismic traces recorded on land have to be corrected for time
differences resulted from near surface irregularities. These irregularities have the effect
of shifting reflection events on adjacent traces out of their true time relationship. Two
major sources of irregularities are 1) elevation differences between shots and receivers,
and 2) the presence of a weathered layer of extremely low seismic velocity (weathering
effect). The thickness of weathered layer ranging from a few meters to tens of meters.
Although it is thin, its abnormally low velocity can caused large time delay to passing
rays which may lead to false identification of significant features on seismic section.

A flat horizon can sometimes appear as a dipping event in topographically uncorrected

section, with the thinner end appearing to be shallower than the thicker part. This is one
of example how the surface irregularity can effect the reflection result, thus it need to be
corrected prior any other processing. The correction for these effects is known as ‘field
static or static correction’. The static corrections remove the effect of the low velocity
layer and reduced all reflection time to a common height or datum.

University Of Glasgow 1998 PhD Dissertation Page 43

Seismic Reflection In Archaeology
Chapter Two

5
<u
C3
CL

CD
_>
o
<d
CL
C/3
<D
Lh
o
O
(N
'TD

m
a
<u
c Cl
o a
o u
o
3
(U
3 a
CT o o O n>
<D
i—i
CD
sc Q
Uh o
<d Q
>, sz
CL
o
c oo
O '
CD
CN 3 O'
cr
(N <d
cfa
CD
Lh CD
3 JC
-*->
CD
CD

c3
NO C /3

£
CL
Jc
CD

M
New Methods In Shallow Seismic Reflection

<d

I I 1 1 l l l l I I I I I l l l l I I I I I I l l l l 11 I I I I I I I I I I M i l l l l l I | l i l l l l l I I I I I I I I I I I I I I I I I I I I I I I I l l l l 11 I I I I I I I I I I

© © ©
w
O o
iyi o O
£
o
ao
CZ>
janoj^udOJdj cq
a
L—
o
Z Z T Harith

*
CD
>
c
3
Seismic Reflection In Archaeology
Chapter Two

<L>
50
03
Q_

o o o o o o o o
(n <on ys u'' ^ o

<u
£
o
wo
00
C
C /3
C /3
<
o
U
cc o
s_
t— Q.
o GO
LU 5
C Ci_ co
cs—d
c3 CO
Ui •b
aj
Lu LU CO
O 00 Q
c Q
.c
Cu
oo
O'
O'
04
< L> c3
<u
s -<
3 T3
<D
Uh
<
»-U
o
0c33
£
<
t-L>
3
00
C /3

2
E
—1
Seismic Reflection

o o o o o o o o o
O O)
<^4 03 N C0 U3 1 ' ro I'm O O O O O O O O O
(Jt CO N 'D U3 'T "3 OJ —
New Methods In Shallow

£
o
to

O
c-
O
Z Z T Harith
Seismic Reflection In Archaeology
Chapter Two

Page 46
(sui) auiji

"! !!!?" U!!!"! !!PP!!"U ""

Ta3>

§•
<L>
■4— »

1
CU
o
-4-*
Vh
O
£

1998 PhD Dissertation

o
cfcI
^3

X3
td
00
PH
a
v

<N
<U
00
•^
pH
New Methods In Shallow Seismic Reflection

iiiiiiiii iii huh nun in IllllllII llll lllllll lllllll I lllllIllllll Illllll111
9 8

(SUl) 3111IX
University Of Glasgow
Z Z T Harith
New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

Calculation of the static correction requires knowledge of the velocity and thickness of
the weathered layer and the velocity of the underlying layer under all shot and detector
locations. It is also need to assume that the rays travel in vertical path through the low
velocity layer.

The first step in static corrections is to remove the effect of low-velocity layer
(weathered layer). In this ‘weathering’ correction, we effectively change the low-
velocity layer (weathered layer) with the un-weathered velocity value. In mathematical
form, the weathering corrections time (tw) can be represent as follows.

where hw is the thickness of the weathered layer beneath the shot or receiver and vwand
v/are the velocity for weathered and fresh layers, respectively.

The second step of static correction is the correction for topography elevation
irregularities. The correction for this effect involves correcting the signal to one
arbitrary horizon, normally known as the datum. If the value chosen is lower than the
surface elevation level, the trace is shifted so that the arrivals are earlier by the
correction time (excess height divided by replacement velocity). On the other hand, if
the value chosen is higher than the surface elevation level, the correction will shift the
arrival time so it is later by the amount of the correction time. If the processing system
is such that events above zero time - datum, are lost or muted, then shallow data can
disappear. Ideally the datum is chosen so that the corrections are as small as possible.

In shallow reflection work where the early arrivals are significant, the choice of a
processing datum is important (Rogers 1981). For a small-scale survey and where the
topography does not vary much, the lowest point is normally taken as the datum (Burger
1992). In the study area the topography changes by approximately 4 m (Table 2.3)
along the 50 m line. Thus taking the lowest point as the datum will mean truncating up
to 16 ms of valuable records from the higher region traces. To overcome this over-
truncation of the traces, the datum is chosen to be the middle of the field. The reason for

University Of Glasgow 1998 PhD Dissertation Page 47

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

this is that the thickness of the weathering layer was unknown, and if the datum is too
deep, the elevation correction will ‘remove’ the surface layers by bringing the traces up.
Since the target is shallow, an over-truncating of the traces should be avoided to
preserve the near surface information. An elevation correction with a 250 m s'1
replacement velocity was applied to the data. This value was an estimation of the first
(glacial till) layer velocity taken from the direct wave velocity in the shot gathers.

Figure 2.15 shows the CDP gathers after all the corrections and data enhancements were
applied. The relatively small dataset enabled a constant velocity stacking analysis to be
done to find the suitable stacking velocity. A normal semblance velocity analysis was
very hard to do since the record length was short (only 100 ms) and the velocity
changes were not very clear. Using this technique, an approximation of a stacking
velocity function was obtained (Table 2.5) and used in normal moveout correction and
stacking process. Only the reflectors between 10 - 40 ms were important in this survey
because the target is less than 5 m deep. A summary of the processing flow applied to
the data is given in Figure 2.16.

A brute stack section (Figure 2.17a) shows a strong reflector starting from 20 m (CDP
160) to the western end of the line. The first 20 m of the section were contaminated by
steep dipping events. The autostatic horizon for static correction was picked based on
this reflector. However, after this static correction, the data shows only a slight
improvement. To remove the steeply dipping linear events, a dip scan stack was
applied (Figure 2.17b). The section was further improved by applying post-stack
Kirchhoff time migration using the same stacking velocity (Figure 2.17c).

Table 2.5 Stacking velocity function

Time Velocity m/s

0 200
30 400
60 1000
100 1600

University Of Glasgow 1998 PhD Dissertation Page 48

o

Page 49
£ 60
§ 8
C l -C
CIS O
J=
U<
j=
c
o
ts
<u
5=
<L> (sui) auiji
o'

T3
♦U
0<0u ' f t
fcdt MlllUMUW lllllltlllUIUliil
CO
•
cd -4 -*

T<3L>
Lh
*8
£
<D

d>
00
_c

University Of Glasgow__________________________________________ 1998 PhD Dissertation

*co
55
<D
O
o
ft
c3
<
*-iu
ft
S-H
0)
c3
f<tU
d
00
co
ft
Q
O

<<
5D
V= <N
i i 2 i ii 16123089
0£ <
»L
-i>
u 3
00
ft
o
CO (suu) auiu
£
”JO
c3
CO

o.-S
* c3
§ K
t5 H
N
Z N
 N ew M ethods In Shallow Seism ic Reflection C hapter Two
Z Z T Harith Seism ic Reflection In A rchaeology

Figure 2.16 Processing flow chart.

Field data
SEG-Y Format
^ Geometry

Editing
Remove bad traces
Reverse Traces

CDP sort
Max. pow er auto
static
Further editing

Elev. static

Band-pass filter
2-D spatial filter

NMO
CDP/stack
AGC R eflector picking
Top mute

Stack

Time migration
DSS

i
Stack with Time migrated
Brute stack
DSS stack
Figure 2.17a
Figure 2.17b Figure 2.17c

U niversity O f G lasgow 1998 PhD D issertation Page 50

 Chapter Two
Seismic Reflection In Archaeology

Page 53
1998 PhD Dissertation
New Methods In Shallow Seismic Reflection

University Of Glasgow

ii n i n ii |i ii n ii ii pi n ii ii i| ii ii hi n p ii n ii n |ii ii ii ii i| ii ii i n n |i n ii n ii pi n n n ij ii ii i n i

O O O Q O O O O O
i - c s j c o ^ r i / ' 5 C D r - - . c c c 3
Z Z T Harith

<g (SU l)3U I!l

PJ
New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

2.6.4 Discussion

Figure 2.18 shows the interpreted final result of the reflection survey. The reflector (red)
is found dipping to the west, following the topography. At the eastern, leftmost end of
the profile (low CDPs), the reflector is only 8 ms TWT deep, while at the western end
the depth increases to approximately 30 ms TWT. The most interesting feature in the
profile is the ‘synclinal’ pattern in the middle of the section, between CDPs 150 and
200. This feature is approximately 10 m wide and about 10 to 15 ms TWT deep.

From the archaeological excavation evidence, the Roman fort was built on the sandy
beach deposit and was subsequently covered by loamy topsoil. From the exposure in the
nearby area, the sandy beach deposit was deposited on top of Carboniferous Coal
Measures group (sandstone and shale). The velocity of the topsoil was measured from
the direct arrival analysis and it is in the range of 180 m s'1to 250 m s'1, with an average
of 200 m s'1. This value (lower than the velocity of sound (air blast, 332 m s'1)) is a
result of the wave travelled through a loamy and grassy wet top surface.

Stacking using the NMO velocity function (Table 2.5) of which was estimated from the
CVSs analysis shows the appearance of maximum reflection amplitude at
approximately 10 ms TWT to 40 ms TWT. This maximum reflection is indicated by
the red marker on Figure 2.18. Calculating the dip of this horizon shows it is a flat
horizon. The Dix equation shows that for uniform horizontal layers and small offsets,
NMO velocity is equivalent to the root-mean-square velocity (VRMS). The conversion
from time to depth section requires interval velocity information, which can be
calculated from the relation below;

1/2
T2x VRMS22 - T xx VRMS\
int vel =
T2 - Ti \

where T 1&2 are TWT to the shallow and deeper layers, while VRMS1&2 are the root-
mean-square velocities to the respective levels.

University Of Glasgow 1998 PhD Dissertation Page 54

 Archaeology
Chapter T w o

a> oil
£ ca
0-
iiiiiii Iiii iiii n Inn mi ilti iitnii Iiii ii ii ii lii n ii ml n ii in n Im ii milii ii ii ii iln ii mi
Seismic Reflection In

a>
43
<L> O
fH
£ o
i 's
s -G
30 —
C £
co
to
<D O
Cu
c
G to
C <u
*— • 3 ■£
a T3 to
<
to
D e
G (U
"G
t)
ca ac l, sa>
OiJ
2
s
c
E
-t-j * £
"G i! Vi
<D <u Q
aj r* 33 Q
Q- 3 3
i-T X>
-t—> C
3 3 O'
o
5
S
O
00 ca .Stof2
43
to u
01 <u G
*-
<o -C
~ 43
"3
.2P C3
£ WT-l
to
<U C
o. toa
<U
1-1(i>
<U t
Seismic Reflection

.S O
•— • u
"O
< e</3
UU
. C U
O
<u
33 =3
(—1 <u
<u
3
New Methods In Shallow

<L>
43 O
H ou
IIII III II |l IIII till |ll IIIIII l| IIII III II |! IIIIIIII III IIIIIIIIIIII III II llll IIIIIII IIIII III I
a
Z Z T Harith

>
c
(SUJ) 3UJJ1 D
New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

Based on this relation, the interval velocity for the first interval is approximately 400 m
s'1. This value is bigger than the velocity of the direct arrival (200 m s"1). This value is
also used to calculate the depth of the red reflector with respect to the datum. Based on
the velocity information, the depth of the red reflector is approximately 1.8 m at the
eastern end of the line and progressively increases to 7.2 m down the line. The depth of
the syncline or pit is approximately 2 m.

This reflector (red) is believed to be a boundary between loamy topsoil and a sandy
beach deposit. The synclinal shape of the reflector is presumably the indication of the
V-shaped Roman ditch, since it is located at the place where the ditch is expected.
Furthermore, the width and the depth of this reflection feature is also about the same as
estimated from archaeology.

The brute stack (Figure 2.17a) shows the good reflector from about the 20 m position
(CDP 160) to the western (right-hand) end of the line and it is missing (hardly
recognisable) from the eastern end to that point (CDP 160). This phenomenon might be
expected to appear due to up-slope thinning of the topsoil. Because the topsoil is thin in
the higher level (eastern end), the problem of interference between direct wave
(travelling along the surface) and reflection signals (from the top of the sandy beach
deposit) might be occurred at this end which degrade the quality of the section.

2.7 Conclusions

The result of the resistivity survey suggested there were two possible anomalies related
to the ditch - anomaly B and C. However, anomaly C was too shallow, while anomaly B
was discontinuous. Anomaly B is more likely to be due to the ditch than anomaly C
since it is much deeper, although the presence of the C anomaly on L3 was unclear. On
the other hand, on the reflection survey only one ditch-like feature was detected.
Compared to the resistivity profile L3, this feature appears to be at the same location as
anomaly B.

University Of Glasgow 1998 PhD Dissertation Page 56

New Methods In Shallow Seismic Reflection Chapter Two
Z Z T Harith Seismic Reflection In Archaeology

The seismic reflection survey employed here has produced a better image of the Roman
ditch than the semi-qualitative resistivity survey. It confirms that there is only one ditch
in this area, and that the centre of this ditch is 21 m from the wall on line L3. The width
of the ditch is approximately 9 m wide and has a depth of approximately 2 m.

In conclusion, although the reflection survey is technically more complex than the
resistivity method, it did nevertheless produce a better image, and a more definitive
detection of the target.

University Of Glasgow 1998 PhD Dissertation Page 57

 CHAPTER THREE

3-D
SHALLOW SEISMIC REFLECTION
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.1 Introduction

The 3-D seismic reflection exploration method is the latest exploration technique and
has been widely used in search for petroleum. From the time when it was introduced in
1972 until now, it has undergone a great deal of improvement and modification. The
usage of this technique is widespread all over the world in the search for petroleum. In
1996, approximately 42.5% of world seismic exploration (for petroleum) was done
using the 3-D seismic technique, compared to only 35% using the 2-D technique, while
the method used for the other 22% is unknown (Stefanic 1997). The deployment of this
3-D exploration method is still increasing. Many major oil companies will not drill oil
exploration wells without consulting 3-D seismic data first. After constant improvement
throughout the years, it was considered by 1991 to be almost a mature technique
(Brown 1991).

Although 3-D is considered a mature exploration tool, only a few attempts have been
made at employing this technique for small-scale near surface investigation work.
Application of the 3-D petroleum technique to near-surface investigation has been
proven not only challenging but also expensive (Lanz et al. 1996, Miller et al. 1997). As
the 3-D technique matures within the petroleum exploration field, involving lots of
capital, adapting the technique affordably to shallow investigation work (mainly for
engineering purposes) which traditionally involves less capital expenditure is the main
restriction.

In this chapter we are going to discuss the application of full 3-D seismic reflection in
investigating a very shallow structure for engineering purposes. This work is believed to
be the first of the kind ever done. Because it is considered a novel trend in shallow
seismic investigation, details from the planning through to presentation of results will be
discussed.

University Of Glasgow 1998 PhD Dissertation Page 58

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.2 Aims

The main aim of this work was to map out underground tunnels (rooms and pillars) of
an abandoned coal mine using 3-D seismic reflection methods adapted from the latest
and most advanced 3-D seismic reflection used in petroleum exploration. The fieldwork
costs of this project were sponsored by the Glasgow Development Agency (GDA).
Designing the 3-D survey or finding the best way to scale down the survey from
petroleum exploration to suit engineering work is the most important aspect, as it is
believed to be the first of its kind has ever been done. Besides that how the data are
processed is also important, as shallow records cannot be treated like deep records.

3.3 Previous work

The application of the 3-D surveying technique in shallow investigation is still in an

early phase of development. Corsmit et al. (1988) conducted a 3-D survey to study the
shallow stratigraphy over a tidal plain. The data were acquired using a 12-channel, 10
bit engineering seismograph with a 30 kg steel ball source dropped from 3 m height on
to a 30 kg steel plate. Their 3-D survey was designed so that the data points of different
source - receiver pairs exactly coincide, so that the CMPs lie on a regular grid. For that,
the geophone stations were located on two parallel lines while the shot locations were
placed on four traverses perpendicular to the receiver line (Figure 3.1a). His work was
purely academic and pedagogic, as its main purpose was to demonstrate to students how
a 3-D survey is conducted, and how the data are handled and processed.

A simple 3-D survey can be designed by placing the line of receivers perpendicular to
the line of sources. Using this simple concept (Figure 3.1b) Hasbrouck Geophysical Inc.
(1994) conducted a 3-D seismic survey to test the use of 3-D data in defining the near
surface structures and properties that control groundwater flow and contaminant
transport.

University Of Glasgow 1998 PhD Dissertation Page 59

 Page 60
f—
£ <u
q. aL
E
—
‘5
c/3
£
o

C/0
G

oo
oo
On

T raverse 4
'I
on
O Traverse 3
U
a/
cC
a
cop 0)
O)
•5a>3

1998 PhD Dissertation

a a
o
TD a I
• o
>,
U. oo . o
+-» a
<L> p o
e o
o
o o
aao
> 5
2
<u
▲ T raverse 2

m
2 T raverse 1
a.
S Q) 0)
CN
5 =
2 2 a>> os
© a)
cS <v; cc

rn
<u
s_
3
op
2

oo
£
_o
University Of Glasgow

-C
C/3

o .-s
■§ <“1k-1
2 x
t H
fc N
Z N
 c

Page 61
ar - 1 -i<u
irj C
S. 5
•S.H
U fl
4>
CO
£
_o
~c3
-C C/3

ON

LO
co co

CM

CN

1998 PhD Dissertation

<3

m
a.
C/3

C
_o
o<U
*-4->

(U
oc;
o
.5?
C/3
£
*
_o
"rt
University Of Glasgow

-C
c/3

■8-s
JZ ,
tr:
^ x
N
N
 Page 62
(J
d>
fW
ra C<O
0
CL O i
Jo .H
U E
V)
'5
GO
£
_o
13
-d
GO

NO
On
On

cd
i—l
Li
CD

w
C
Op
co
a>

1998 PhD Dissertation

"O
£
a>
B
C
o
<u
00
2
a>
VS
Q
m
—
Cl
t—1
C
c/5

rn
<U
Li
P
_0p
co
0L)
&

CO
£
_o
"5
University Of Glasgow

GO

TO
3
-C
flj

£
&

Page 63
<1) £o
5 "8
e
M (U
"s
CL o2

J= .52
U c

l>
ON
O n

C3

CD

CD
c£
o3
C
_op
c/5

1998 PhD Dissertation

CD
tp

+ ->
<d
£
o
CL)
tJD
<D

m
Cl
B
(Z3 CO

t uD
m
c CD
_o Vh

o P
00
c:a<>
d
s
02

c/i
University Of Glasgow

J=
1/3

TO j
.c
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

A simple 3-D survey also can also be designed by placing the receiver lines parallel to
the source lines. This design (Figure 3.1c) was deployed by Lanz et al. (1996) in order
to overcome the problems in the 2-D section where the critical parts of the 2-D seismic
section were dominated by out-of-plane reflections and diffractions. For each receiver
deployment, a total source configuration consisting of four rolling lines of 24 sources
(shotgun) points at a 5 m interval was used.

Miller et al. (1997) also deployed a 3-D survey to evaluate the lateral variability of the
interpreted fault around a collapse feature or sinkhole. The 3-D dataset was recorded
using a single fixed square array centred on the sinkhole (Figure 3.Id). This is another
simple method of acquiring 3-D data.

All these simple methods were designed so that the midpoints from different source
receiver pairs exactly coincide so that the CMPs lie on a regular grid. Because of the
regular CMP grid, these simple methods can be considered as a subset of a more general
3-D survey, in which the source receiver midpoints and the azimuths should be as
random as practicable, so that there is complete freedom in binning the CMP data. A
discussion on 3-D survey design is presented in Section 3.8.

Smythe (1994) pioneering the use of the vibroseis source in imaging the fault patterns
for hydrogeological work. The aim of his work was to image the relationship of faults
within the target volcanic rocks in 3-D, in order to investigate the suitability of the area
as a radioactive waste disposal site. This survey covered 1 km of CMP area with the
prime target in a depth range of 400-1000 m. Although the aim was to provide
information on fault patterns for an engineering site investigation, the depth of the target
excludes this work from being categorised as a shallow investigation work. However,
the success of this survey has led to the present shallow 3-D experiment, the author’s
project.

Based on the success of Smythe (1994) in mapping a complex fault system at a

relatively shallow depth (compared to petroleum exploration) and the suggestion from

University Of Glasgow 1998 PhD Dissertation Page 64

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Briickl et al. (1996) and Kourkafas and Goulty (1996) that the 2-D reflection images
over a mining area (coal and gypsum) can be improved by conducting a 3-D reflection
survey, in early 1996 the decision was made to experiment with full 3-D seismic
reflection over an abandoned coal mine.

The year 1996 marked a new era in shallow seismic industry where the transformation
to a full 3-D seismic reflection technology (from the petroleum industry) has been
brought in. At the time the present project was initiated, it was believed that it was the
first ever of its kind, as there were no publications (except for the previously mentioned
simple 3-D surveys). However by late 1996, several publications on the same subject
had been published (Barnes and Mereu 1996, House et al. 1996, Siahkoohi and West
1996).

Barnes and Mereu (1996) studied the limitations of high resolution 3-D seismic
reflection techniques for mapping near surface stratigraphy. They also compared a 2-D
with a 3-D data set to evaluate the effectiveness of the 3-D method in imaging structures
as seen on the 2-D data set. Siahkoohi and West (1996) on the other hand used the 3-D
seismic reflection techniques to image complex stratigraphic structures within glacial
deposits. Several problems were encountered such as: (1) maintaining the consistency
of shot amplitude and spectral content, (2) strong surface wave generated especially at
single points shot, and (3) in obtaining a uniform S/N ratio as well as the offsets and
azimuth within each bin. The uniform offset and azimuth within each bin can be
achieved by using a less linear shot and receiver distribution (Siahkoohi and West
1996).

House et al (1996) also used the 3-D survey method to determine the feasibility of
using this methodology to constrain bedrock topography. Based on their work, they
also suggested that the targets should have a depth of more than 15 m with a sharp
velocity contrast in order to have good results in 3-D shallow seismic reflection work.
The latest publication on this subject has been by Buker et al (1998b) who outlined a
comprehensive strategy of 3-D seismic reflection data acquisition and processing.

University Of Glasgow 1998 PhD Dissertation Page 65

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.4 The site

3.4.1 Location

The site, at national grid556705, is part of the West of ScotlandSciencePark. It is

located to the NW of and 7.5 km distant from Glasgow City centre. It lies north of the
junction of Acre Road and Maryhill Road, and is elongated 320 m along the Maryhill
Road with 110 m width. The total area of the Science Park is 35.2 km2 (Figure 3.2,
Photo 3.1). At the time of initial survey, the site was a green field, however later during
the course of fieldwork, part of the topsoil was stripped off from the site.

3.4.2 Geology

Glasgow is situated in theMidland Valley. The valley is an ancient riftvalleyor graben

bounded by two parallel faults orientated NE-SW, the Highland Boundary Fault to the
north and the Southern Upland Fault to the south. The rocks in the valley are relatively
young, with two main systems, the Old Red Sandstone of Devonian age and
Carboniferous sediments. Broadly speaking, the Carboniferous rocks are arranged in a
wide compound syncline underlain and flanked by strata of Old Red Sandstone age
(MacGregor and MacGregor 1948).

In the Glasgow area, most of the solid rocks are covered by glacial moraines and
outwash sand (George 1973). The Drift edition of the geological map 1” Sheet 30
clearly shows that the site is covered by boulder clay from recent glacial deposits. This
glacial deposit is deposited on the “d2b rocks, which are the Limestone Coal Group of
the Carboniferous Limestone Series (Geological map, Sheet 30, Solid Edition, part of
which is reproduced in Figure 3.3). These groups were deposited during Lower
Carboniferous time. The cross section drawn through the nearby area (Figure 3.3)
shows that the Carboniferous rocks are dipping to the east at a low angle.

University Of Glasgow 1998 PhD Dissertation Page 66

Figure 3.3 Geological map and the cross section o f the site and neighbouring area
(Geological Survey o f Great Britain, sheet 30)

Horizontal sale: 1 inch : 1 mile

Vertical scale: 1 inch : 0.5 mile.

(:wi;e the horizontal scale)

Stratigraphic column o f the sedimentary rocks in the site and nearby area

A'jr.'.Yt .V jtS I j i u W l 'J .C'nff.1

, ,
!A/fmnwr*
J (ftCi
i C’uWs'A'ii/rroiLT
)•LiwiVtw
j Sufnu >7J
&
GafctfersriU
, .V ittfrA C h w i i p
Series
one1
— ......... .... i
r^sa x E o serm ; l . i '& S f f & K r f>M

3<
:
-J - t t I .■ n r / . i - i v . . : , ' r«
tiOf*vr$9ntu)kii • n t", -ft*

< 'unn* .'..rrti*! 1V

[ J3KS ] . H ur;kr Scrii-3

'Atentfl

Cross-section o f A-B showing northeasterly dips

Site
c* r r c ij'.- v jf V * T t
 % «y>r

M-SSi

E ^ § M r? r^
£ h ^H h £E38i
' ■f •^^KSsSSS^’-■- ■
?-V-.it ’rV -
SggBS^y ft .:SjJ1■<•:-> !-,1,r-‘A
vyKy*,.'.

^S rw SmH
I g S S M

,.n V ;
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.4.3 The problem

Many cities in the United Kingdom, especially in the northern England and in the
Midland Valley of Scotland, such as Sheffield, Newcastle, Edinburgh and Glasgow are
underlain by rocks of Carboniferous age which have been substantially mined not only
for coal but also for fireclay, ironstone, clays or even sandstone during the past. These
materials were extracted in several fashions depending on the era. In medieval times
(15th century), a shaft of about 2 m square was dug down to the coal layer and the coal
was extracted radially from the bottom of the shaft until it was difficult to support the
hanging roof (Hyde 1987, Waltham 1997). Only then was it abandoned and a new shaft
dug nearby. The excavated material from the new working was used to fill old shafts,
but normally this fill was not properly compacted. This kind of working was known as a
bell pit.

Later, the ‘stoop and room’ or room and pillar excavating method become popular
because more coal can be extracted, and it is safer than the bell pit method. In this
method, coal is extracted from the rooms, and pillars are left to support the roof. The
extraction of these rocks creates an underground void. Once the area is found
uneconomic, the mining stops. After some time has elapsed, the pillars might no longer
be able to support the roof, so that the roof collapses into the rooms (Photo 3.2). This
may cause a depression or a hole at the surface. Adding additional pressure or weight by
building on these areas accelerates the collapse process, which can damage the
buildings (Photo 3.3).

For that reason, land over an old mining area, which has old uncollapsed voids at
shallow depth is considered unsuitable for redevelopment. However, due to the
increasing demand in recent years for city-centre land, either for housing or industrial
development, these areas, which were formally regarded as unsuitable, should be
reassessed.

University Of Glasgow 1998 PhD Dissertation Page 70

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

The site used in the present dissertation is situated in a green field with pasture, over an
area of an old coalmine within the Garscube Mineral Field. The mine was abandoned in
1897. Figure 3.4 shows a map of the known workings on the site at the time it was
abandoned. Associated with the coalmine, three mine shafts - Pit no 7 and Pit no 8, are
within the site, while Pit no 5 lies just outside the field. These pits were completely
covered by topsoil. The outlines of these pits were only rediscovered in 1981 after
extensive trial-and-error drilling. The size of No 7 pit is approximately 3 x 2.5 m
(Photo 3.4) and was found after 44 closed spacing boreholes had been drilled (J. W. H.
Ross and Co., unpublished report 1981). Pit No 8 (100 m deep) is bigger, at
approximately 6.1 x 3 m2 and was found after 16 boreholes had been drilled (J. W. H.
Ross and Co. unpublished report 1981). These two workings were for a deep coal
horizon which creates no problems for the foundations of any new buildings. The
outlines of No 5 pit were undefined despite the fact that 40 boreholes were sunk at the
suspected location (J. W. H. Ross and Co. unpublished report 1981). The boreholes
drilled, however, found filled material at approximately 6 and 10 m depth. Compared to
the depth of No 8 pit, this working is obviously near to the surface and could give rise to
problems to the foundations if any buildings were going to be built.

In this green field, single and double story buildings were planned as an expansion or
enlargement of the West of Scotland Science Park. As the subsurface has been mined,
the strength of the rocks has also being disturbed. The ability o f the rocks beneath the
proposed buildings to hold the weight of the proposed buildings needs to be reassessed.
The extraction of the coal also creates zones of weakness along the existing sewage
tunnels. Assessing the suitability of the site for construction is done by geotechnical
engineers as part of the site investigation.

In an area of an abandoned coal mine, the primary aims of the geotechnical

investigation are to determine the extraction ratio, the pattern of the layout and the
condition of old pillars and rooms. The methods can be either by exploratory drilling,
trenching, closed-circuit television or where possible entering old workings, depending
on the conditions and depth of the workings (Bell 1979).

University Of Glasgow 1998 PhD Dissertation Page 73

N o rth
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

All these methods are already established among geotechnical engineers. For important
structures, it requires thorough exploration and sampling of all strata likely to be
significantly affected by the structure load. Once the coal working pattern has been
recognised, the building can be either moved to a more stable area, or when there is no
alternative an appropriate treatment to the subsurface void should be carried out.
Normally, the voids are grouted. The amount of grout material used depends on the void
volume.

At the present site, the decision had already been made by the consulting engineers to
grout the areas which were likely to affect the proposed buildings, prior to this 3-D
seismic project being approved.

3.4.4 Suggested solution

As the objective of the geotechnical site investigation is to recognise the pattern of old
underground tunnels and other voids, novel methods of accurate subsurface imaging for
shallow structures and underground tunnels need to be developed. Ideally the
investigating method should be able to picture or image the target with great detail and
give a continuous image at minimum cost. The current geotechnical investigating
methods such as drilling do not give the continuity needed, and grid drilling is
expensive. Entering an old working, on the other hand, is considered to be too
dangerous.

One answer to such an engineering site investigation problem is the 3-D seismic
reflection method. Although the 2-D reflection method already has been widely used in
engineering investigations, it still cannot meet the continuous imaging requirement of
the site investigation. Furthermore, the resolution is sometimes poor. The 3-D seismic
reflection method on the other hand has long been established in the petroleum industry
and proven capable to continuously image the rocks and their structures with great
resolution. Few studies have been done to date on how to scale down the advanced oil

University Of Glasgow 1998 PhD Dissertation Page 76

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

industry seismic technologies to suit shallow investigations. The main delay in

progressing to 3-D at the site engineering scale has probably been the fact that the
processing software and cost are still beyond the capability of engineering companies.
3-D for engineering investigation purposes is considered a new challenge for the
shallow subsurface investigator.

3.5 Planning and basemaps

3.5.1 Planning

The success of any work depends on careful planning. The most important factor is
financial. Unlike petroleum exploration, shallow investigations (for engineering
purposes) are normally provided with very limited funding. This is the main constraint
in adopting expensive 3-D survey technologies. Furthermore, the knowledge of the site
such as surface material, surface conditions, accessibility, possible noise and problems
are also should be taken into account in designing the survey.

After a series of discussions with the Glasgow Development Agency (GDA), the GDA
permitted the University of Glasgow to conduct a small-scale shallow 3-D seismic
reflection survey over one of their sites. The site, within the West of Scotland Science
Park, is part of an area previously classified as unsuitable for development because of
the presence of underground voids left over from the old coal-mining work.

The aim of this survey was to find the best methods of adopting the latest full 3-D
technologies used widely in petroleum exploration industry to image the shallow
underground tunnels of an old coalmine working. The methods developed as a result of
this work hopefully can be used to give useful information on the pattern of
underground tunnels and their depth at other sites. As part of the agreement, GDA
funded the field costs at the site. However, it should be stressed that the results from the

University Of Glasgow 1998 PhD Dissertation Page 77

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

seismic work were not intended to influence the development of thisparticular site,
since a programme of orthodox drilling and grouting was to takeplaceregardless.
Rather, the aim was to use the drilling results as a calibration, if possible, for any
positive seismic images.

The survey was designed to make it as economical as possible. The costings were
designed to allow:
1. A week of on-site topographic survey (refer to Section 4.6.2)
2. 1-2 days of experimental acquisition methods
3. 4-6 days of intensive production mode data acquisition
4. Hiring high resolution recording equipment
5. Paying the field assistants.

Among all these factors, hiring the high-resolution seismic recording equipment was the
most expensive.

The fieldwork took place during July 1996. To minimise the cost, only a maximum of
eight people were involved, compared to dozens or even hundreds in normal petroleum
exploration. They comprised three geophysics academic staff, of whom two were
present at anyone time, the author as a postgraduate student, and up to four student field
assistants. Besides hiring the recording equipment and paying for the students, other
expenses were minimal. Transportation and software for processing (ProMAX/2D &
3D) and interpretation (GMAplus 3D) were already available in the department.

Prior to the actual 3-D survey, preliminary surveys in the form of undergraduate
geophysical field training were done to improve our understanding of the site. At the
end of the 3-D survey, a conventional 2-D reflection line was also planned for
comparison purposes.

University Of Glasgow 1998 PhD Dissertation Page 78

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.5.2 Basemaps

Basemaps in the form of paper copies and digital AutoCAD DXF (data exchange
format) ASCII files of the site were provided for us by Peter Fraenkel and Partners,
Geotechnical Consultants for GDA, using their AutoCAD software. Since it had been
decided to make all maps and figures (where possible) using the public domain GMT
(Generic Mapping Tools) software, the AutoCAD DXF ASCII generated files were then
converted to a suitable format for input to GMT using acad2gmt, a specially written
Geophysics Group Fortran program.

The basemaps supplied contained all information relating to the site such as the field
and other boundary outlines, main road, neighbouring buildings, and topographic
contours (Figure 3.5). However, only the field and main road outlines are relevant to the
survey, so for clarity the others were edited out. On the other hand, information on the
borehole locations, preliminary refraction survey lines and the outlines of the common
depth point subsurface coverage were added into the base map (Figure 3.6).

3.6 Preliminary studies

The objective of the preliminary study is to obtain an idea of the near surface geology
and conditions as well as to examine possible noise sources and physical limitations, so
that the acquisition of 3-D could be designed to suit all the conditions and limitations.

The preliminary studies consisted of an afternoon site visit to discuss the best time to do
the survey as well as to observe and check the accessibility of the site. This was
followed by the Honours third year students’ geophysical field training week, of which
seismic refraction and resistivity surveys were conducted at this site. Besides those
surveys, borehole information from mineral and site investigation reports was also
studied.

University Of Glasgow 1998 PhD Dissertation Page 79

 New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Figure 3.5 Basemap as supplied by Peter Fraenkel and Partners, Geotechnical

Consultants for GDA. Contour line is every 0.5 m.

55500 55600 55700

70600 70600

70500 -
70500

ipfc

70400 70400

70300 -
- 70300

MME

55500 55600 55700

University Of Glasgow 1998 PhD Dissertation Page 80

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Figure 3.6 Simplified basemap with extra information such as borehole locations
and refraction lines added.

/0 1 9 9 1 BH 12 \
70550 2.3m (Sandy clay)

'NMH

01981 BH 11
2.7m (Sandy day)

70500 -

3.3m
T 996 BH 4 4A 48
4.4ni\(Drift)

70450 -

4.0.1

BH 5

70400 -

(Drift)
01981 B \ l 3
2.5m (Sandy day)
BH 1 3.6m

1995 BH 6 o
l.6m

01972 BH1
70350 -
01995 BH,
3.9m (Dril

1981 B| 14o
5.6m (J clay)
55500 55550 55600

Shaded area is the seismic reflection CMP coverage.

The values (in m) beside boreholes represent the thickness of the overburden.

University Of Glasgow 1998 PhD Dissertation Page 81

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.6.1 Preliminary geophysical surveys

Prior to the actual 3-D survey, some of the undergraduate geophysical field training for
the 1996-97 Honours Geology class was done at the site. It was supervised by two
geophysics lecturers and two postgraduate students including the author. This field
training for the students’ benefit was also designed so that the author could use the data
from this fieldwork to understand the site characteristics and physical limitations in
order to design the 3-D survey. The students were divided into four groups and each
group spent half of a day on each of two geophysical methods.

Two types of geophysical investigations were conducted - electrical resistivity and

seismic refraction. These two investigation methods were chosen as they represent the
most common geophysical investigation methods applied in shallow engineering and
hydrogeological investigations.

R e s is tiv ity

Resistivity is the most common geophysical methods used in shallow investigation

works, particularly in engineering and hydrogeological investigations. It is used to study
the horizontal and vertical variations in the electrical properties of the earth’s
subsurface. It works by measuring the potential differences at points at the earth surface
that are produced by introducing or directing an artificial current into the ground. The
changes in potential difference give information on the subsurface electrical properties.

The movement of the electrical current through the rock is mainly controlled by the
content of ions (salinity) in pore waters and pore connectivity. Unlike water, most rock-
forming minerals such as silicate are insulators, which means that they do not let current
flow. In general terms, resistivity decreases with increasing salinity and connected water
filled pore space.

University Of Glasgow 1998 PhD Dissertation Page 82

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

In electrical resistivity investigation practice, two resistivity surveys procedures are

routinely done - Vertical Electrical Sounding (VES) or 1-D electrical drilling, and
Constant Separation Traversing (CST) or 2-D electrical profiling. Each student group
had to conduct at least two VES and one CST. The VES’s were conducted using the
Wenner electrode configuration where four electrodes, aligned at equal spacing,
progressively expanded about a fixed location.

The resistivity values were plotted on log-log graph paper and qualitatively analysed
before computer modelling. Figure 3.7 is an example of the resistivity values plotted on
the log-log graph, which give an idea of the layers present. The plotted resistivity curve
shows a common S or C shape. In the example shown there are at least three different
layers. The first layer has a considerably higher resistivity and corresponds to the near
surface material, which was dry at the time of the survey. The second layer has a lower
resistivity, which implies that this layer is porous and contains lots of connected water.
This layer may correspond to the saturated overburden or glacial till. The deeper high
resistivity layers on the other hand can be assumed to correspond to the Carboniferous
rocks, which are more compacted and therefore with less pore space. Based on the
rough idea of the layers present and their resistivities, the estimated depth or thickness
of each layer was calculated. The summary of the bedrock depth from the undergraduate
studies is given in Table 3.1.

Table 3.1 Summary of the thickness of top two layers from resistivity data.

Thickness (m) Group A Group B Group C Group D

Top layer 0.5 1 0.9 0.9
Second layer (Glacial till?) 4.5 2.3 2.3 2
Third layer 16 - 4.9 -

Bedrock depth (m) 5 3 3.2 2.9

University Of Glasgow 1998 PhD Dissertation Page 83

R£S\£ti V l'rf
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

S e i s m i c r e fr a c ti o n

As with electrical resistivity, the seismic refraction surveying method is another

common geophysical tool used in investigating shallow subsurface targets, especially
for estimating the depth of the bedrock for engineering works. Besides that, the
information gathered from this survey can also be used to assess the rock strength and
potential fluid content (Reynolds 1997).

The refraction data were recorded using a 24 channel OYO McSeis 1600 seismograph.
Most of the lines surveyed were oriented in the dip direction except line A-A’ (Group
A) where it was oriented in the strike direction (Figure 3.6). The geophones were
arranged on a 46 m line at a 2 m interval while the shots were placed at 2 and 5 m offset
from each end and also at the midpoint. Time-distance plots were produced to get the
velocity and thickness of the layers at every shot point. A summary of the thicknesses
obtained from the refraction survey is listed in Table 3.2.

Table 3.2 Velocity and thickness of the layers from refraction survey.

Group A Group B Group C Group D

Layer 1 - Velocity (m s'1) 455 425 385 370
- Thickness (m) 4.5 2.7-3.7 1.87 2.36
Layer 2 - Velocity (m s'1) 2235 2370 1571 2000
- Thickness (m) - - 9.38 -

Layer 3 - Velocity (m s'1) - 2647 -

- Thickness (m)

Three of the groups found a two-layer earth except for Group C, where a three-layer
earth was recognised. If the two-layer earth is assumed (glacial till overlain by
Carboniferous bedrock), the velocity for layer 1 is too low for glacial till, but on the
other hand the velocity for the second layer could correspond to shallow weathered
Carboniferous rock. Group C’s model is somewhat more sophisticated, where the first,
second and third layers correlate to the topsoil, glacial till and Carboniferous rock
respectively. These results have not been re-interpreted by the author.

University Of Glasgow 1998 PhD Dissertation Page 85

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.6.2 Borehole studies

Borehole exploration has long been established as an integral part of any shallow
subsurface investigation. Depending on the investigation purposes and cost, the drilling
methods can range from cheap and simple auger boring to expensive rock core drilling.
It is normally done prior to the engineering construction to examine the subsurface
geology and other related issues. The information from the drilling programme is
recorded on borehole logs, which record (as tabulated text) the depth, rock types and
any special remarks. Samples from rock core drilling can be used to examine the rock
quality, minerals present and also the shallow structure such as faults. The advantage of
borehole study is that it is a direct investigation method, however the investigation is
limited to the immediate vicinity of the hole. Furthermore, subsurface correlation
usually cannot reliably be done by borehole since it needs continuous sampling of the
borehole, which is expensive.

Drilling on the site had been undertaken since 1966, when three boreholes were sunk in
order to investigate the unrecorded mine workings. In 1981 another drilling programme
was carried out. The aim was to investigate the mineral (i.e. coal seam) location under
the site and adjacent area. The latest drilling programme was conducted in 1995 with
the main purpose of providing information on the ground conditions at the site relative
to the proposed development. All three drilling programmes are discussed individually
as their objectives were different.

T h e 1 9 6 6 d r illin g p r o g r a m m e

The 1966 drilling programme was conducted by the Institute of Geological Sciences,
Edinburgh and supplied to Glasgow University by Peter Fraenkel and Partners. Three
rotary drilling logs - 1966-1, 1966-2 and 1966-3 - were studied. The quality of the
borehole logs were so poor that it was difficult to correlate one with another; however

University Of Glasgow 1998 PhD Dissertation Page 86

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

they clearly indicate the bedrock at approximately 3 m depth, and the presence of coal
seams at various depths.

T h e 1 9 8 1 d r illin g p r o g r a m m e

The 1981 drilling programme was conducted by J. W. H. Ross and Co. as instructed by
the GDA. The main purpose of this drilling was to investigate mineral locations under
the site and adjacent area. Twenty-nine boreholes were drilled, of which six were drilled
within the site. These six boreholes were drilled using the rotary core drilling method
where continuous samples were collected and the minerals were analysed for
nomenclature and correlation (explained below in Chapter 4). However, due to the lack
of borehole information, the nomenclature and correlation of the minerals in all the
boreholes drilled in the site might not be correct (J. W. H. Ross and Co. unpublished
report 1981).

The borehole information shows that the bedrock belongs to the Carboniferous
Limestone Series of Lower Carboniferous age, dipping 9° to 18° toward the NE. The
depth of the bedrock, however, varies from approximately 2.3 m to more than 8 m as
tabulated in table 3.3.

Table 3.3 Approximate depth of the bedrock in the site based on 1981 drilling.

Borehole Number Approximate Depth (m)

1981-11 2.7
1981-12 2.3
1981-13 2.5
1981-14 5.6
1981-15 5.1
1981-16 8.5

A more complete analysis of the borehole information is carried out in Chapter 4 in the
context of the surface tomography results. The summarised result of the 1981 drilling

University Of Glasgow 1998 PhD Dissertation Page 87

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

programme is shown in Figure 3.8, where the meroon (red) hatched area is unsuitable
for development, while the green hatched area is deemed suitable for single storey
buildings.

T h e 1 9 9 5 d r il l in g p r o g r a m m e

The 1995 drilling programme was carried out by Nicholson (Site Investigation) Limited
on behalf of Peter Fraenkel and Partners. The aim of this drilling was to provide
information on the ground conditions in relation to development of the site. Like the
1981 drilling programme, this programme also evaluated the nearby area. Six out of
nineteen borehole sunk were drilled within the site, however only first five were
available. Borehole 1995-6 was not supplied. The boring was conducted using the cable
percussion method, which is a poor method of sampling rocks. The percussion boring
often destroys the parent structure, resulting in a mix-up in identification of the bedrock
materials (parent rock), and boulder material is often mixed up (Nicholson Limited
unpublished report 1995).

The five borehole logs studied show that drilling was stopped after encountering
bedrock or obstruction. Because of that, the use of this borehole information is limited
to the estimation of the depth of the bedrock only (Table 3.4).

Table 3.4 Bedrock depth from 1995 drilling programme.

Borehole Number Approximate Depth (m)

1995-1 2.3
1995-2 5.8
1995-3 3.9
1995-4 Not reached
1995-5 4.1

University Of Glasgow 1998 PhD Dissertation Page 88

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.7 Topographic survey

Knowledge of the precise location of sources and receivers is very important in all
seismic exploration. For a land seismic investigation, positioning is often carried out
with the Global Positioning System (GPS) or by a theodolite survey, or by a
combination of both GPS and theodolite depending on the area of coverage. GPS
readings give the coordinates with respect to the satellite coordinate system, and have to
be transformed into local coordinates (Sheriff and Geldart 1995). The accuracy achieved
depends on the type of codes (P or C/A-code) observed and the level of SA (selective
ability), and it ranges from several tens of meters to tens of centimetres (Leick 1992,
Dana 1998).

For a small scale high resolution survey, the accuracy need to be much better than this,
which can be achieved by using theodolite surveying. A theodolite system can be used
to calculate the distance and bearing as well as the relative height between the station
and specific locations. The theodolite positioning technique offers a precision up to 0.06
m with an additional uncertainty of 0.01 m per kilometre of line length (Heliker et el
1986).

3.7.1 Equipment

An Electronic Distance Measurement (EDM) theodolite was borrowed from the

Glasgow University Archaeological Research Division (GUARD). It was a Sokkia 5-G-
D20816 set (serial no 116057). It consists of three parts, the tripod, EDM theodolite and
reflection prism.

EDM theodolite is widely used today as it is easy to operate, is frilly automatic and
gives a very accurate reading. The precision error depends on the model used. Sokkia
set5 (the model used) has an angular accuracy (horizontal / vertical) of 5 second. For
measuring distance, it provides very accurate readings with ± (5 + 5ppm x D) mm

University Of Glasgow 1998 PhD Dissertation Page 90

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

accuracy, where D is the measuring distance. In EDM theodolite surveying, a laser

beam is aimed at the reflection prism and the time taken for the beam or pulse to travel
from theodolite to the prism and be reflected back to the theodolite is calculated to give
the distance as well as the horizontal and vertical angles. The vertical angle can be used
to calculate relative height between the station (theodolite position) and the prism
location.

3.7.2 Surveying strategy and aims

The Autocad DXF file (supplied by Peter Fraenkel and Partners) containing the outline
of the site was used to guide the theodolite surveying work. As the site is elongated
parallel to the Maryhill Road, a 2 by 2 m grid was plotted to cover the site with the Y
axis or column parallel to the baseline which also parallel to the main road. The site was
covered by 34 columns and 140 rows (Figure 3.9). The location of the grid points were
denoted by a five digit notation, where the first three digits represent the row number
and the other two represent the column number. This map is known as a prearranged
layout, and forms the basis of the acquisition pattern design. Using a specially written
Fortran program, the range and bearing (azimuth) of all the grid positions with respect
to certain known points are calculated.

The aim of this theodolite survey was to mark the actual location on field with a peg, as
defined in the prearranged layout, so that the geophones and sources can be positioned,
and at the same time to calculate the elevation differences between the grid point
location and the fixed (control) point. To save time, only 8 m by 8 m grids were sighted
using the theodolite and then interpolated to a 4 m by 4 m grid using measuring tape.
All the positions, whether sighted using theodolite or by tape, were marked by sticking
specially constructed peg into the ground. The pegs were constructed of softwood,
250x5x10 mm3 in size and numbered uniquely on a white background using the five­
digit notation.

University Of Glasgow 1998 PhD Dissertation Page 91

 New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Figure 3.9 Prearranged layout map for a grid of survey pegs at 2 m spacing.

eg 14034
70550 -

70500 -

70450 -

70400 -

70350 -

70300 -

55500 55550 55600 55650

University Of Glasgow 1998 PhD Dissertation Page 92

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.7.3 Controls

The digital base maps supplied by Peter Fraenkel and Partners define the position of
two manhole covers in the site, one in the middle and the other one at the north end of
the field. The central manhole is noted by CMH while the north end one is denoted by
NMH. These two manholes were the control points and all the distances and azimuths
for theodolite survey were calculated with respect to either these two points. The line
connecting these two points is taken as the baseline, used to align the azimuth of the
theodolite. The elevation and the national grid (easting and northing) of these two points
is recorded in Table 3.5.

Table 3.5 Survey control points.

Station Easting (m) Northing (m) Height (m abpve OD)

CMH 55568.08 70421.31 37.80 *
NMH 55505.12 70534.07 35.51

3.7.4 Conversion from column, row to National Grid

As rows and columns were aligned with respect to the baseline (an arbitrary line
connecting CMH and NMH) the column and row coordinates need to be converted into
National Grid coordinates (easting and northing). The conversion process was done in
two phases where firstly they are rotated into an arbitrary grid and then secondly shifted
to true National Grid coordinates.

The rotation was being done by taking the 00000 position as the rotational axis and
rotating the column/row coordinates clockwise by 29°.645. The rotated coordinates
were then shifted to their true coordinates by adding the actual Easting and Northing
coordinate of the 00000 position.

University Of Glasgow 1998 PhD Dissertation Page 93

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.7.5 Precision and errors

Setting out the pegs, based on the pre-arrangement layout by walking the prism to the
correct range and bearing, following instructions from the surveyor at the theodolite,
resulted in field accuracy of better than 0.1 m. This is within the precision required of
the high-resolution seismic reflection method, in which errors of the order of 0.1 m in
horizontal and vertical coordinate are acceptable. The accuracy of the peg positioning
can be observed by checking the line-up of pegs along rows and columns. Vertical
errors were checked at the processing stage by re-gridding and contouring the data, to
see any anomalous peg positions stand out, as well as comparing the height with the
gridded topographic elevation supplied by Peter Fraenkel and Partners.

Comparing the calculated CMH and NMH eastings and northings with the values
supplied give zero error, which means the calculated coordinates are correct. Eastings
and northings were calculated to a precision of 0.01 m.

3.7.6 Problems

The main problem in the theodolite survey was that sometimes the pegs disappeared.
The reasons were either that they had been removed especially in the area where the
topsoil was stripped for drilling injection boreholes, or they were run over by vehicles.
The missing pegs were normally replaced by using tape measuring from the
neighbouring pegs, however in case of pegs removed for topsoil stripping, new pegs
were replaced by re-sighting the position using theodolite.

3.8 3-D survey design

A good survey design is a design of which achieves the investigation goal and at the
same time minimises the cost and time of acquisition and processing. Designing the 3-D

University Of Glasgow 1998 PhD Dissertation Page 94

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

survey varies depending on survey objective, cost and time. In the late 1970’s, 3-D
surveys were shot using a three-line swath (Figure 3.10). A 3-D three-line swath is
actually an expansion o f the ordinary 2-D line survey, where the centre line is an in-line
spread while the other lines are simply off-line spreads. The data were processed using
the same concepts as for 2-D data. All the traces with the same CM P num ber were
grouped and N M O -corrected before stack. The N M O for off-centre lines needed an
azimuthal factor and proved a little more difficult than anticipated (Stone 1994).

Figure 3.10 3-D three-line swath (after Stone 1994).

Source
Receiver

Today, 3-D surveys are often designed so that the source and receiver lines are arranged
perpendicular to each other. The CMP concepts were replaced by the binning concept,
as developed for crooked-line 2-D CMP gathering. The bin is a small area, normally
rectangular, into which the 3-D data are sorted. For each bin, the fold, offset and
azimuthal distribution is im portant although the exact position o f each CMP within it is
(by definition) unimportant.

Some o f current 3-D land design methods are parallel, cross-spread and perim eter

techniques. Dickinson et al. (1990) compared these techniques and concludes that the
results from the first two techniques differ little. The perim eter survey method is not as
good as other two due to low fold, on the other hand it is cheaper. Other methods such
as swath, full 3-D and loop were explained in detail by Stone (1994).

U niversity O f G lasgow 1998 PhD Dissertation Page 95

N ew M ethods In Shallow Seism ic Reflection C hapter T hree
Z Z T Harith 3-D Shallow Seism ic Reflection

In a swath survey the receiver array is much greater in length in one direction than in
the other and the source moves through them orthogonally (Figure 3.11). The source-
receiver azimuths are almost unidirectional. The limited range o f azim uths make this
design unsuitable for exploring com plex targets. This design is recom mended for survey
that are large but have relatively simple structure in one direction (Stone 1994).

Figure 3.11 Basic swath design for 3-D.

Soui*ce line (vertical)

i
<>

<>
W
<>
w
< Receiver lines (horizontal)
►

To investigate a more complex target, the survey must be designed so that a full-range
o f azim uths can be obtained in each bin. This can be achieved by a full 3-D survey
design. A full 3-D survey can be designed by arranging the receivers in a rectangular
grid and for each receiver position the shot is placed adjacently. This kind o f design
will produce a full 360° range o f azimuths within most o f the bins.

A simple full 3-D design survey (Figure 3.12) in which a rectangular fixed receiver
pattern covers the survey area is relatively easy and cheap to conduct. However if the
survey requires moving the receiver pattern to cover the whole area, it becomes rather
expensive, as lots o f time is needed to move the cables and geophones. Traditionally,
moving the receiver pattern in a full 3-D seismic survey is done by moving one or two
grid units to either the right, left, up or down after a line o f shots has been observed
(Figure 3.13). This kind o f survey requires lots o f equipm ent to be laid beforehand. In
addition, telemetry type receivers must be used to enable efficient switching.

University O f G lasgow 1998 PhD Dissertation Page 96

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.12 Example o f simple fixed receiver pattern lull 3-D survey.

A i k
1 o' k o 'k o 'k o k o 'k o 'k o '

0
o o o o o o o o
o o o o o o o o
o o o o o o o o
o o o o o o o o
o o o o o o o o

Receivers

^ Source lines

Figure 3.13 Traditional moving receiver full 3-D survey.

Source line 1
Spread
Movement

.Spread _L

U niversity O f G lasgow 1998 PhD D issertation Page 97

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

In highly cultivated and populated areas where a normal 3-D survey such as a swath or
full surveying technique is difficult to design (due to obstructions), Bertelli et al (1993)
introduced a new kind of survey which could overcome the coverage restriction while at
the same time increasing the spread efficiency as well as improving the field operations
flexibility. This new technique is known as the ‘random’ technique.

This technique is based on the assumption that in order to achieve the desired 3-D
coverage of an area it is sufficient to subdivide the surface into basic elements, and put
into each of them the average number of sources and receivers positions that provide the
desired coverage (Bertelli et al 1993). In the random survey, the basic recording block,
according to the number of seismic channels available, includes a number of the
elementary areas.

3.8.1 3-D field layout

In order to achieve the desired 3-D coverage of the area as well as achieving the full
source-receiver azimuth range, a combination of full and random 3-D survey methods
was deployed. Basically, the receivers and sources were arranged and moved in or
along a grid pattern similar to the full 3-D survey, however instead of moving the whole
spread after each shot line is completed, the recording block is divided into three
elementary areas (groups) as in the random 3-D survey design so that only parts of it are
moved. The full 3-D survey will ensure a full range of source-receiver azimuths, while
the random 3-D survey will ensure that most of the survey area will be covered with a
desirable fold, which is important in the binning process.

Using 144 channel recording equipment, the receivers were divided into three groups of
48 receivers each. This division is constrained by the input cabling to the recording
instrument. The receivers in each group were arranged in 6 rows and 8 columns or 8
rows and 6 columns. In geophysical terms, this group is known as a spread. Each spread
was connected to the recording instrument by a separate set of cabling. Three spreads

University Of Glasgow 1998 PhD Dissertation Page 98

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

were placed next to each other to form a rectangular layout of 18 rows and 8 columns or
8 rows and 18 columns. This rectangular receiver layout is known as the swath (Figure
3.14). Each row and column has a 2 m interval. Figure 3.15 shows the layout of all the
spreads employed in the 3-D survey.

The sources were arranged in a variable pattern over each swath. Figure 3.16 shows an
example of the area of shot coverage for two of the swaths. After all shots for the
pattern were fired, one or two spreads were moved instead of the entire swath. This
strategy was employed as it maximises acquisition efficiency.

Spreads and swaths were defined over the whole area of interest (Figure 3.15). Ideally
the swath geometry (consisting of three adjacent spreads) should have a rectangular
shape, however in some cases due to unavoidable obstructions such as drilling rigs, the
geometry of the swath was set in a L or even a T shape (Figure 3.16). The spreads are
labelled in alphabetical order. Although there is no overlap of receiver spreads, many
shot positions were observed or shot several times (Figure 3.17). This is a more efficient
method than a geometry in which each shot position is observed only once, but in which
many more movements of the receiver spreads would than have to be made.

3.9 Acquisition

3.9.1 Instrumentation

O Y O D A S -1 r e c o r d in g s y s te m

The data were recorded using an OYO 144 channel DAS-1 system. It is a multi-purpose
high performance seismic acquisition system designed for engineering, environmental
and hydrocarbon exploration. It can be used for all seismic investigation such as P- and
S-wave reflection and refraction, VSP, and crosshole tomography.

University Of Glasgow 1998 PhD Dissertation Page 99

 CJ c o
£ .2 o
r“ (u
£ D
<d aj 03
CL
a.
cd
x: .
CJ

C/0
£ r-

aC
o-
c/o J I I L CO
ai CO
CO
M-
CM
CO
13 ^
G c/o
m- G
♦■■ C P O CO .3 g?
43 - o ° acd
cd
od CO Oh
C/5
£ o> H-»
<D M
C/5 - 0)
CM o _>
<u <D
c Lj ‘ S
o ~ CO G o
CM o <d
bX) o P4
.g K <D
CM H C/5
cd
6 - CO ?C/5 ' M«
CM CD g
w
Q ~
If) 3 .3
o cd
CM
U
C/5 ~' *t 1CD) TTj"f
CM
T3
cd c *h—> zCm
(D - CO
CM E s X°3
43
o ^ td
<d
CM „ M £
<u CM ° CD crt Q
O Q
£h—• r-
<m CM S J 8M—» £
O C/5 oo
O On
JD CM £ " On
"a ,
O) > ”3
B T3 £
cd
X
<d CO
JZ cO/3h
£2
G
< ^-S *8
o
C/5 “

CD « r2
T“ O
in •S ?
CO T“ o o

<D
M Tt g Js>
G C/5 H
•SP CO .52
<D _ S
G
1) T—
Zp
<D
O h £P

c2 "T T CM 1 1
i— r o
CD
CD
<D
co CM ® CO N tO ID CO CM T3 ^
r*
-j— oT*~ t—
o r*
o oT” T
o*• T
o” O O Cd 2
a> -G
0)
c/o
£ Moy C/5
o hG
£
"rt c3 O
X U)
c/o w
a
Cm
o .-S o
■Q§J I--Hc3

S
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Figure 3.15 Layout of all spreads.

112
110 -

108 -
D
106 -

104 -

102 -

100 -
Q R
98
96 -

94 - a
92 U
90 -

88 -

o5 86
a:
84 -
V’ W
82 -

80
78 -

76 -

74
72 -

70 -
N
68
66 -

64 -

62
60 "i i i i i i i r

0 2 8 10 12 14 16 18 20 22 24 26 28 30 32
Column

Spreads of 48 channels are labelled A, C-W, Y, Z. Spreads V and W were moved right
from their original positions to V’ and W’ respectively due to an unavoidable
obstruction

University Of Glasgow 1998 PhD Dissertation Page 101

N ew M ethods In Shallow Seism ic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.16 Layout o f patterns CDE and V ' W ' Y - a detailed example.

-J C
iII I
! !
I II Ii II I| II !! !i ‘!I-----!! —~—
!! II—-—------
II iI II Ir
l_

114 - "

112 -

110 -

108 -
106 -
104 -
102 -

100 -

98 - -
96 -
94 -

92 -

90 -
5 88 -
o
cc 86-
84 -
82 -
80 -
78 -
76 -
74 -
72 -
70 -
68 -

66 -

64 - -
62 -
60 -
i TI I I I i I I T T I ^ I T T I
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34
Column

Dark shading indicates swath receiver area w h ile lighter shading indicates
area o f the shots into the swath

University O f G lasgow 1998 PhD Dissertation Page 102

New M ethods In Shallow Seism ic Reflection C hapter T hree
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.17 Layout o

off all patterns.

■j -j 0 ! i
' i' i
1 i1 i 1 1 1 I1 I l1 I' l1 !
1 I
1 l1
114
112
110
108
106
104
102
100
98
96
94
92
90
5 88
o
<r 86
84
82
80
78
76
74
72
70
68
66
64
62
60
58
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34
Column

Dark shading with alphabet represent the receiver swaths while the outlines represent
the area o f the shots into the swaths. Note that the area o f shots overlap each other to
obtain a good fold o f coverage even at the edge o f the swath.

U niversity O f G lasgow 1998 PhD D issertation Page 103

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

The DAS-1 uses 24 bit Sigma Delta Technology, which increases the system dynamic
range by 18 dB and the instantaneous dynamic range by 24 dB, to give superior digital
recording resolution according to its manufacturers OYO.

Sources

Three types of sources were tested in order to find the best source to be deployed on the
site. These were a portable mini vibroseis, a 7 kg sledgehammer and 7.26 kg weight
drop.

The vibroseis was supplied by OYO as part of the recording system equipment. The
Oyo minivibrator is a compact portable electromagnetic device weighing about 75 kg. It
has a cylindrical shape, and is approximately 40 cm diameter and 60 cm in height
(Photo 3.5). Two people are needed to move it around and for that it is fitted with
handles. The vibrator is connected to its power source and the DAS-1 by a 100 m long
armoured umbilical cable.

Beside the vibroseis source, more common shallow seismic sources (7 kg

sledgehammer and 7.26 kg weight drop) were also tested. Both of these strike a steel
plate of 40 cm in diameter, 2.5 cm thick and weighing about 20 kg. The weight drop
was dropped manually by field assistance from 1.75 m height measured by sighting rod.
The plate has approximately the same diameter as the base plate of the vibrator. The
plate has a tangentially drilled cylindrical hole into which is put the trigger. For better
coupling with the earth, the plate and the vibrator are placed in a pre-prepared hole.

University Of Glasgow 1998 PhD Dissertation Page 104

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.9.2 3-D data recording

C rew

Recording the 3-D data was done by the Glasgow University Geophysical Team
comprising 3 academic staff and one postgraduate student (the author). Additional to
that, 4 students were paid on a daily basis as field assistants. Although 8 people were
involved during the recording phase, not all of them were present on the site at one
time. Academic staff took charge, but only one or two were involved each day. The
permanent members in the field were the author and two student labourers. On average,
there were 5 personnel present at any time during the seismic survey period.

R e c e iv e r s w a th

The receiver swaths were based on the columns and rows of the 4 m grid. However, for
part of the survey area the receivers were laid out at a 2 m interval in order to acquire
the densest data possible. The 2 m ranging rod was used to mark a 2 m position from the
peg. As the pegs are very important for location, to avoid removing them the receivers
were offset to the right and upwards by 0.1 m from the pegs (Figure 3.18).

R a n d o m i s e d s o u r c e lo c a t io n

The source was moved along columns or rows of the survey grid. However, it was not
placed exactly on the grid, to avoid a regular common shot point (CSP) pattern. The
combination of randomised CSPs in conjunction with the regular grid of receivers leads
to a randomised coverage of CMPs. This is desirable in all 3-D so that CMP binning can
be carried out in any desired fashion.

University Of Glasgow 1998 PhD Dissertation Page 106

 Chapter Three
3-D Shallow Seismic Reflection

<L»
00
CO
0_

00

co
O, O.
o
Cl
°+
<
od
-
sh
3
O
C /3

-a
a) CO
c /3

£
o
"O
c
0J-,3 o, o
T3 C\i
C
03 Q
Q

O'
O'

00
<u
a.

CO
a> ~T~
3
00 CO CNJ O
Uh
New Methods In Shallow Seismic Reflection

UJ Uj 0 O U B IS IQ

£
o
00
O
L—
o
Z Z T Harith
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

The shot position was placed randomly within a 1.8 x 1.8 m2 area from the pegs, as
shown by the shaded square in Figure 3.18. In helping to determine the shot location so
that the CSP is as random as possible, a random number generator was used to generate
a table of offsets in the row and column direction for every possible shot position, in
increments of 0.2 m. With 10 permissible such offsets in each coordinate, there is a
total of 100 possible positions within the randomised area.

In advance of each shot, a field assistant consulted the table and measured the offset
required in the row and column direction using a ranging rod. On the turf-covered sites,
the position was then marked by digging a 40-cm radius hole as a preparation for
placing the source. On the stripped topsoil site, the position was marked by an un­
numbered peg. To avoid confusion with the grid pegs, these pegs were coloured yellow
(Photo 3.6).

V ib r o s e is s w e e p

In all high-resolution work, the ideal source is the one that can generate lots of high
frequency energy. However, depending on the near surface materials, most of the high
frequency components are attenuated. Unlike impulsive energy sources, which generate
the energy during a very short time, the vibroseis source transmits energy at a low rate
(i.e. low power) into the earth for several seconds. This long-duration encoded quasi-
sinusoidal signal contains a range of frequencies at a constant amplitude, and is known
as a sweep.

The sweep can be designed to start at the high frequency and end at the low frequency
(downsweep), or vice-versa (upsweep). Usually the change of frequency over time is
linear, however in some cases a non-linear sweep may also be deployed. The non-linear
sweep is usually designed so that the sweep spends relatively more time at the upper
end of the sweep spectrum, to compensate for losses due to the attenuation from the top
layers.

University Of Glasgow 1998 PhD Dissertation Page 108

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

In an attempt to record the broadest bandwidth of frequencies possible, a 100-700 Hz

linear upsweep was chosen. The reason behind this was if the higher frequency signals
were badly attenuated, the lower end of the band can still be used since low frequencies
are less affected by the attenuation. Although a non-linear sweep can be used to
compensate for the loss of high frequency components, the decision was made not to
use it, as a non-linear sweep increases the recording time. Furthermore, a non-linear
sweep might have been too risky if the high frequencies did prove to be elusive, unless
the total duration of the sweep were also lengthened to maintain sufficient time at the
lower frequencies.

The value 3850 ms of sweep duration is derived from the DAS-1 set up parameters of
4000 ms total recording time. The total recording time is a combination of the sweep
duration and the listening time (the time after the sweep is finished). 150 ms of listening
time gives a correlated length of 150 ms, the same as the record length chosen for an
impulsive record.

R e c o r d in g p a r a m e te r s

The recording parameters were chosen during the planning phase with the aim that the
acquisition would acquire the data in the densest and broadest band mode as possible.
Most of them are limited by the equipment capability. Although the recording
parameters were chosen during the planning phase, some of them were changed during
the actual fieldwork due to unavoidable problems.

To get the spatially densest data possible, the receiver station interval was set to 2 m.
The shots were at a 4 m interval, i.e. at every second row or column. In the southern
comer of the area, where the topsoil had been cleared, the shot interval was changed to
2 m.

University Of Glasgow 1998 PhD Dissertation Page 110

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

The 144 channel DAS-1 system is made up of a 48 channel main unit and two 48
channel analogue input extension units. Each unit handles each spread, which means
that the swath is handled by 3 units of 48 channels.

The primary seismic source was the OYO minivibrator. The impulsive sources
(sledgehammer and weight drop) were only done for 3-D preliminary tests. The signals
were detected using SM4 30 Hz single channel geophones already available in the
department. Although 100 Hz geophones have in the past been considered more
appropriate for high resolution shallow work, their primary purpose was to filter out
ground roll. Modem high dynamic range recording systems now make this analogue
filtering unnecessary.

As the aim o f this work is to map the abandoned mineworkings at less than 60 m depth,
only short records were required. For impulsive sources, the record length was 150 ms,
however for the vibroseis data, 4 s of uncorrelated data were recorded. After correlation,
the record length is only 150 ms. The data were sampled every 0.5 ms. The sweep was
also tapered using a 50 ms cosine taper. A list of recording parameters is summarised as
in Table 3.6.

Both correlated and uncorrelated vibroseis data were recorded in the field. The field-
correlated data is done using a convolution function available in the DAS-1. To increase
the signal to noise ratio, uncorrelated vibroseis data were stacked 4 times and the
impulsive data 8 times before storage. This vertical stacking method will increase
coherent signals such as reflection by ‘n’ times and reduce the random noises by the
1 /2 »
power of ‘n ’. This hopefully can increase the signal-to-noise ratio by the factor of
‘n ’ , where ‘n’ is the number of stacking times.

University Of Glasgow 1998 PhD Dissertation Page 111

 Chapter Three
3-D Shallow Seismic Reflection

Page 112
o.
a oL i
o X
1-»
JO
JO .00
•
00
C
D
C/3
£
*s
oLi £
C
D £
3 OO S3
cd
cd ">< 3
cC3, l t-l OO 3
<
L> cd
00 3

University Of Glasgow________________________________________ 1998 PhD Dissertation

.s 3 OO c/3
>—i X 3
o OO
Q 00 <
d o
<
d LT)
(-<
c/5 x)< d
CN
X
< C/3
cd <
d
00
Q N
<
d
t-i
13 0 cd >
cd X > X
cd
Li o
JO m 8 x)
C
L)
<L)
3 o
13 o
JO ■> o +->
vq o 1 ■d" <D cd
■'d- ’3 : S 13 <N fc: m
co ■d- C/5 "5 fc m o X
o VO
j/ 3 Li *s
o O <N
CD <D i
JD
cd
o O OcS3/3 cd CD Li
S3 d - a N CQ
x
H > a> 3 o d- w IT) ffi oo o
o O c/5 CN ■d- £ C/5 m d-
New Methods In Shallow Seismic Reflection

c
c/3
Li
Li
§ 1-* <D
3
Li
1-> S I 8-I
-L
1/3 C
D
cd
> >
cd
o
cd
JS3 X
S3 Li <0 t
•^ <
d
l*
<> <D &
L a> oo 1—*
00 D S3
C
O c S3 a>
•S3 .3 .3
.If Li
< .3 Sa>
jd

33 33 cd
JD ft
i->
O O OO &
d
X > S3 <D S3 <L> > XL i X & Oh
Li
a> • O o O O a> Li

£ JO
O o o o >> CD <D <D
o o cd o cd O o oo c3 (D (D <D
<
d
Pi! X
+->
C/5
s "cd s
o i-»
C/5 C
O a> i->
/5 C/5 X C /5 P4
I
C/5
CD o £
X L-l ffi W C/5
£ £
C/5 C/5 c/5
Z Z T Harith
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.9.3 2-D data recording

In tr o d u c tio n

On completion of the last of the 3-D acquisition, it was decided to run a single 2-D
reflection using all available sources in turn - mini-vibrator, sledgehammer and weight
drop. Only one line of 2-D was recorded as there was only one afternoon (and evening)
left before the termination of the recording equipment hire. The 2-D profile was shot
along row 072, where the topography was nearly flat. Furthermore, half of the line was
stripped of topsoil, while the other half was still covered. This would provide a further
comparison of the effect of topsoil.

G e o m e tr y

The total length of row 072 is 70.5 m. To acquire the densest data possible, all 144
channels were used, which means the receiver spacing was 0.5 m. The geophones were
placed directly at the pegs without the offset as in the 3-D survey. The pegs at every 2 m
were shifted to the left of the line to make room for placing the geophones. The first
was placed on column 00 while the channel 140 was on column 34. As there was no
more room at the NE edge of the field, channels 141 to 144 were not connected.

The source was moved towards the higher column number at every 2 m or at every
column starting on column 01 and ending on column 34. As the peg position was
occupied by the receivers, instead of placing it directly on the location, the source was
offset by 0.5 m towards the higher column number. All sources were placed in a 0.5 m
wide, 10 cm deep pre-prepared hole for better coupling.

University Of Glasgow 1998 PhD Dissertation Page 113

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

R ecording param eters

Most of the recording parameters used for 2-D profiling were the same as those used for
3-D surveying. The only differences were the receiver and source geometry and their
interval. A summary of the 2-D recording parameters is tabulated in Table 3.7.

3.9.4 Data recording and transfer

The recording took place inside the recording truck (Photo 3.7). The recording truck is
equipped with three independent 12 VDC power supplies. These power supplies were
used to supply the power to the DAS-1 seismograph. The power for the vibrator was
supplied by an external power generator. This external generator was placed quite a
distance from the swath in order to reduce the noise generated from it.

Summing the data to improve signal to noise ratio and correlation of the uncorrelated
vibroseis data were done automatically by the DAS-1. The data were written to an
external SCSI 500 Mb hard disk. Both correlated and uncorrelated vibroseis data were
recorded. Each time the data were written on to the external hard disk, information such
as shot number, source position and source type were recorded on the observer’s
logsheet. At the end of the day, the recorded data were transferred into a processing Sun
workstation back in the Department via a networked personal computer into which the
hard disk was plugged. Visual checks of the data were made before the contents of the
disk were erased for the next day’s surveying.

In order to save time re-laying the spread every morning, the spreads were left on the
ground overnight. The risk of theft or vandalism was considered to be minimal, as the
site is fenced. Furthermore, there was also a guard on duty in the yard at the south end
of the field. The DAS-1 and other hired equipment were locked up in the recording
truck. The portable generator was left outside the security guard’s hut.

University Of Glasgow 1998 PhD Dissertation Page 114

 Chapter Three
3-D Shallow Seismic Reflection

Page 115
cd

<o
2
c>d-i <o

University Of Glasgow ________________________________________ 1998 PhD Dissertation

cd co
Pi CO
00
'■a /3
C
*-( <
o
a Q N
< o
S-H
cd X
o o
<N cd m
o
o CO
.2CO
<N CQ
T3
■"d- C/3 co O
o
t-4 CN VO
o O
CO 2
i
N PQ c3 0?
'Td CO
e'­
en o
c
CO
C3
o o
fit
o
w
/3 o ’
C
ffi oo .2 o
u C/3 o £ £ CO l-J IT)
I
H
New Methods In Shallow Seismic Reflection

c
CO
Vh
I3 CO
&
to <o ?Vh .*3 .P
.2 CO CO -4->
& & I1 CO "§
00 co
0J3| & C
. s -4-» CO S r2 .2
3o3 aoo -2 &
TJ a> a CO a CO p
T3
<-4 o 04
Pi Pi
o o ‘<o • Ho O .2 o ip CO CO
co
o S O ed s td g
o +-» o +-> o
Icd4oo £
o
oo c3 £ (O
CO CO
£ £
P4 C/3 p^! C/3 C/3 C/3 C/3 C
/3 C
/3 pi! hJ S /3 C/3 C
W C /3 C/3
Z Z T Harith
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

After the last day of production (Day 11) the morning of the last field day (Day 12) was
spent in removing all ground equipment and the remaining pegs.

3.9.5 Progress

The actual 3-D field work started on Monday 15 July 1996 with the arrival of the hired
DAS-1 and portable mini-vibroseis. Day 1 and most of day 2 were devoted to laying out
the cable - swath CDE, and to test shots. The 3-D data production started on the
evening of the second day using the vibroseis source. After all 70 shots fired into this
swath the production acquisition was repeated using the hammer and weight drop
sources. The first swath - CDE - was completed in four days. At the end of day 4,
based on the time consumed to complete just one swath, it was decided not to proceed
recording the data from both hammer and weight drop, but to concentrate only on the
vibroseis source. It appeared to produce the broadest-band data.

The surveys were also done during the weekend when the drilling engineers were on
weekend break. This meant that there was no drilling noise and we were also able to
work on the stripped top soil area especially at the southern end of the site. Most of
swaths MNO and PQR were shot with a 2 m source spacing, so that that area has an
extra dense coverage. Starting on the following Monday, the survey had to continue in
the presence of five active rigs drilling grouting holes. The noise from these rigs has
undoubtedly reduced the quality of the data to some extent.

In general, moving a whole swath (three spreads) took around two hours. The
geophones were left connected to the cable and each spread was loaded hand over hand
into the Landrover. This method of speeding up the survey progress also had the
additional advantage that the two or three bad channels remained at the same relative
place within the swath. This made the equipment checking easier at the start of
recording of each new swath.

University Of Glasgow 1998 PhD Dissertation Page 117

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Table 3.8 summarises the progress rate of the 3-D acquisition. Table 3.9 summarises the
total number of vibroseis shots collected for the whole area.

Time lost to production recording because of unforeseen events, breakdown, etc. is

known as down time. It includes the necessary daily travel time and equipment
checking. Over the 12 survey days, the amount of down time was approximately as
follows:
1) Spread faults 5 hours
2) Recording truck problems (flat batteries) 4 hours
3) DAS-1 problems (setting up configuration files) 1 hour
4) Severe rain (Day 8) 3 hours
Total 13 hours

The working day average about 10 hours, so there was a loss of about 1 hour per day.
Additional to this time was the period spent each evening downloading the data from
the hard disk, which took another 2-3 hours.

University Of Glasgow 1998 PhD Dissertation Page 118

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3Cj
co
oo
Date (July 96) Day No. Work Swaths Shots Weather Comments

o
Mon. 15 Field testing CDE Very good First swath laid; test shots
Tues. 16 Testing; CDE 42 Very good Vib. only

(N
production

CO
Wed. 17 Production CDE Very good Vib. and WD

CO
CO
Thu. 18 Production CDE Very good WD

IT)
Fri. 19 Production I FGH, 132 Very good Vib. only; Impulsive source
HIJ terminated

r-
On
Sat. 20 Production | JKL Very good
San. 21 Production MNO 194 Hot, cloudy

vo r- oo
Mon. 22 Production | PQR Heavy rain Drilling started

On
C
Tues. 23 Production 154 Good Drilling noise

O H
UC/3
0$ [D

o
Wed. 24 Production 173 Good Drilling noise

o
Thu. 25 Production YZA Good End of 3-D: 2-D line shot

I
O
Fri. 26 1 Clear up Good All spreads and pegs removed

<N
University Of Glasgow 1998 PhD Dissertation Page 119
 |
c

-C
o
_o

C/3
(U
C/3
<13
03
od
JZ
c

b «
£r - mo

C/3
£ .2

C/3
'53
a . 03

O
U an
E

3
M

H
M
M

co
j§
o

Os
T3
M

C/3
>

c/3

.£3
.C

a/
C/5

ctf

C/3
.O

O
O

cd
<13
O
O
C3

Vi
cd

Cm
T033
<u

^03
Cm
V3

tB

o3
.1o3

Qi
<l>
Pattern no. Swath No of No of pairs FFIDs Remarks
Stations of files

-
CDE 7099 70 101-240
FGH 68 241-376

<N cn
09
HIJ 64 377-504

■'3'
JKL 197 197 505-898 2m shot spacing

cn
Os
MNO 194 899-1286 2m shot spacing

On
CO

NO
PQR 139 1287-1564

I 1

r-
06 06
STU 1565-1744

oo
Os
oo
V’W’Y 98 1745-1940

OS
YZA 70 70 1941-2080

3
3
2
3
3

'm

w
o
o
01

«M

T3

Xi
C/3
03

03
03
C/3

cd
.fel
O
T3

O
O
C3
fc

Jg

c3
c3

53
cd

Cm
fc

33
T133
03

Qi
^03

C/3

Cu
University Of G lasgow 1998 PhD Dissertation Page 120
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.10 Seismic data processing

3.10.1 Introduction

All the reflection data - 2-D and 3-D - were processed with ProMAX/2D and
ProMAX/3D, industry standard software installed on a Sun SPARCstation20, with a
capacity of 132 Mb RAM and 27 Gb hard disk storage. As ProMAX does not have the
capability of data picking and displaying time slices, the processed stacked data were
transferred to a 486 PC running GMAplus 3-D software, a software for viewing,
manipulating, and interpreting 3-D data sets.

The seismogram (seismic record) is basically a record of the amplitude of a wave

disturbance (together with background noise) that is detected by a detector at different
times. The amplitude of the wave is controlled by the reflection coefficient (R) which
depends on changes in acoustic impedance (I). Acoustic impedance (I) is defined as the
product of velocity (V) and density (p) of the material:

Acoustic Impedance I = V.p

Reflection Coefficient R =
V2p2 + V\P\

In seismology, the layered earth can be represented by a series of reflection coefficient

values at different times (depths). If the earth’s surface is disturbed, the wave generated
at the disturbance centre will travel through the earth as well as along the surface.
Normally the disturbed wave is generated by a sudden release of energy, which leads to
the generation of a very short pulse known as a wavelet. An initial waveform (wavelet)
travelling through the earth will travel at various speeds, depending on the properties of
the medium. The changes in physical properties that the wave encounters are like a
filter. When the initial waveform (wavelet) passes through the boundaries of layers with
different properties, the shape of the wave will change. Figure 3.19 shows graphically

University Of Glasgow 1998 PhD Dissertation Page 121

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

the changes of the wavelet when it meets the boundaries represented by the reflection
coefficients of different value (left panel), thus producing the seismogram.

The changes in shape can also be understood mathematically as being a convolution

operation between the disturbing wave and the reflection coefficient response series.

V ib r o s e is c o r r e l a ti o n

Since reflections occur at intervals smaller than the sweep length, the returned signal
detected by a receiver is the superposition of the reflected wave trains (Figure 3.20)
together with the noise. It is a very long and complex seismic record, and it is
impossible to interpret. It must therefore be converted into a more conventional seismic
record form (as would be gathered from an impulsive type source). To do that, the field
data must be cross-correlated with the pilot sweep.

Cross correlation measures the resemblance between two time series. When both time
series are identical, the cross-correlation between them is known as the autocorrelation.
The autocorrelation of the sweep is known as the Klauder wavelet. Note that the
autocorrelation value is a maximum where two functions are superimposed without a
relative time delay (at zero lag), and the resulting function is symmetrical.

P r o c e s s i n g s tr a t e g i e s

Processing of the recorded data started with the 2-D data as it involved a relatively
small amount of data. Furthermore, the results from these three different data types
(vibroseis, hammer and weight drop) would hopefully give an idea of the layers present
and their velocity. The phase relationships between the three sources and in particular
the vibroseis polarity could also be checked by comparing the signatures of the first
arrivals.

University Of Glasgow 1998 PhD Dissertation Page 122

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Figure 3.19 Seismogram concept (after Yilmaz 1987).

Principle of Superposition
"i r

L
i :
t-
(a)
~! r
■jc>b
i i
1 r (b)
+
~i r

+
(c)
i r

Oh
(d)

oh
J L (e)
~i 1- ----- 1- - - - - - - r 1 T n r

4 o t-
j L i i i J L
(f)
n 1- ----- 1- - - - - - - r

— |JIa-
J 1-----------1----------- L J. L 1i----1----u J ! J I L L
(9)
0.2 0.4 0.6 0.8 s 0 0.2 0.4 0.6 0.8 s

A wavelet (a) travelling in the earth repeats itself when it encounters a reflector along its
path (b, c, d, e, f). The resulting seismogram (g, right) represents the composite
response of earth reflectivity (g, left) to the wavelet (a, left).

University Of Glasgow 1998 PhD Dissertation Page 123

 Page 124
-C
f-
Q . DC
.J »
'<-> i/)
E
'5
<n
£
_o

"5
x:
in
Q
m "O
c
CCS
c<J
4—<
CCS CC CC
TS
in .5 CO

’<5 2
on <
O £§■
X> Time
I s
’> 5w r-J l^
<4-1 Og-
O ^ o
O- ON
o VWIAA/Www
O 2
c
O TD
O C
co 8B yw w w w v^o
•■5 £«* ^AAA/IAAAMa^ " ,
S *

1998 PhD Dissertation

Q
J-,* <
s-iu VlMAAA/vww\,
cL<£
3* w
D o k/\A/\jayvwww- >■0
V> 3 £
8)

33
4—> <2
<+-, <£
O 0)G VxjAAAAyjywjw^
c _o
.2 4— >
03
as
in aj -V X A a A/www^ !
fc
o
o
o+
— m

33 u
Cl, <u
© TD
©©
2 £ -KAAAAAA/wvw- o aH >
O
-V W A A A M M m ^

c
_o o
<N ■WVA/IAAAMw-
<D
<G
<D o <u
a: 3
:=
L (5U “VW/lAAA/Ww^
•SP
<u
<n
_o
University Of Glasgow

15
x:
C/3

"© _C
©s
H
<Uh«M
3

t6 H
N
Z N
N ew M ethods In Shallow Seism ic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Although all the sources were tested for the first swath (CDE), only the vibroseis source
was continued to the end of the 3-D survey. The other two were terminated due to
shortage of time. As both correlated and uncorrelated vibroseis data were recorded, it
was decided to correlate the uncorrelated data again in the laboratory using a synthetic
sweep, and to compare the quality to the field correlated data. Processing of the 3-D
data was done in three different stages, as discussed in section 3.10.3.

3.10.2 2-D data processing

In tr o d u c tio n

All the data were processed using ProMAX/2D. The data used for 2-D profiling were
hammer, weight drop and vibroseis lab correlated data. Although vibroseis field
correlated data were recorded, they were not used, for reasons discussed below.

F r e q u e n c y a n a ly s is

All the data were stacked in the field vertically 4 times for the vibroseis data and 8 times
for the hammer and weight drop data. Figure 3.21 and Figure 3.22 are examples of shot
gathers from the three methods at the same location. The quality of the data ranges from
poor to moderate. Most of the shot gathers were dominated by direct waves, ground roll
and air blast. The far-offset traces were badly contaminated by noise. To make
processing even more difficult, there was no clear sign of reflectors present on any shot
gathers.

University Of Glasgow 1998 PhD Dissertation Page 125

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

University Of Glasgow 1998 PhD Dissertation Page 127

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

The frequency content of the shot gathers shows that the amplitude spectrum of the
vibroseis data is confined essentially to the correlation range of 100-700 Hz, with the
peak amplitudes at the low end of this range, around 150 Hz (Figure 3.23a). The
spectrum shows a severe loss of the higher frequencies, at a rate of about 10 dB/oct. In
contrast, the impulsive source records have a lower frequency spectrum, in the range of
20 to 100 Hz with the peak at 30 to 80 Hz.

Figures 3.23b and 3.23c show a shot record with its corresponding amplitude spectrum
for hammer and weight drop respectively. The figures show a great similarity between
the hammer and weight drop records. There appear to be no useful source frequencies
higher than 100 Hz.

The 30 to 80 Hz effective bandwidth for the impact source records implies they are not
suitable for high resolution work, as the minimum thickness that can be resolved is
approximately 6 m (assuming a velocity of 1200 m s'1 and mean frequency of 50 Hz).
On the other hand the peak value of the returned frequency for the vibroseis data is
around 150 Hz, which means the minimum thickness it can resolve is approximately 2
m, using the same criteria.

P h a s e c o m p a r is o n a n d f i r s t b r e a k p ic k in g

The SEG signal polarity convention is that for an upward ground motion, the signal is
represented by a negative number on the tape which in turn means a trough on the
displayed trace. The polarities for all records were checked by comparing the first
arrival signatures from all the records.

The aim of this was mainly to ensure that the polarity of the vibroseis data matched that
of the impulsive source data, because the zero-phase nature of the former was not clear
on the shot records.

University Of Glasgow 1998 PhD Dissertation Page 128

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Jattojgp

University Of Glasgow 1998 PhD Dissertation Page 129

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

£
£

d
,d
d
d
on

J-l

o
i

00
ds_

-d
s

ro
a)
B
B

oC
-4-11

ce

(N
c3
a.

<o

at-*>
a"

&o

.2P
o<D

<D
<D

(S U )
Jonojgn

OU T I
University Of Glasgow 1998 PhD Dissertation Page 130
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

d
d
£
o

-d
03

<N
o
od
’5

6d

m
C/5

<
>

S-H

43
cr

-t—>

*
-*—
cdtf
TD

GO
00

O
<u
<o
u
h

oo
op

dJ-l>
(X

o<D

00 Q.
• |
0J
©
o
TT
• | i |
o
tr

CO

<su)
janojap

|
o
CO

a u jx
O
▼
H
©
|

rH
0J
o
^
,
o
I
t—
University Of Glasgow 1998 PhD Dissertation Page 131
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Figure 3.24 shows that the first breaks are to the left (trough) on both impulsive sources.
The first ‘kick’ represents the first arrival signature. Comparing short offset data (where
the signal is strong) between impulsive and vibroseis data, it is seen that the first arrival
in the impulsive records corresponds to the first trough on the vibroseis record. This
confirms that the polarity for all three datasets is same. First break times were picked
using the ProMAX first break picker. On impulsive records, the onsets (the first ‘kicks’)
were picked, while on the vibroseis records, the corresponding first troughs were
picked. Only clear first breaks were picked, and noise-contaminated traces were not
picked.

R e fr a c tio n m o d e l

To obtain the near surface velocity structure, the first break times picked on the hammer
data for the 2-D line were sent to Professor Ewald Briickl of the Technical University of
Vienna for refraction inversion. His results show that the average velocity of the first
layer is about 350 m s'1 (Figure 3.25). This value has been used in elevation and static
corrections applied subsequently to all the data, both 2-D and 3-D.

P r e -p r o c e s s in g

The field recorded data were assigned with their appropriate field file identification
number (FFID). The FFID number is unique and is based on the Observer’s logsheets.
Any non-production files such as noise tests were deleted prior to applying geometry.
All 34 shot records of vibroseis data and 31 shots each of hammer and weight drop were
carefully inspected to look for any signs o f anomaly or inconsistency. Visual
inspections of the data were also done to look for any signs of reflector; unfortunately
the data were dominated by low frequency noise which would mask any reflectors.
After visual inspection of the files on the screen, any dead and noisy traces were zeroed
and reverse traces were corrected.

University Of Glasgow 1998 PhD Dissertation Page 132

 <U c
V o
r-
t—
S_ C
<U <u
DO
« <E> CO
a.
Q . ft£
CO
4=
U E
<u
c/>
(su i)a u iL L

CO
-C
on
Q
<D

' 4-*
O
<u
&
C/5
<u
O Vi
C/5
3
-t->
$3 03
<L>
Vi
T3
O h
<£ O

43
00
"S
< 4-1 £
o T3

3
GO Vh
CL>

(L) s
43 B
43
O

Q
’C
C ti o Q
Vi
o 43 £
C3 00
<u ON
43
■*—>
ON

C+H
O
3
3 1)
0 C/5
C/5 <u
Vi
c3 D h
<L>
Vi

1
o
a>
3
U C/5
'U
S
p i

c (N £
_o 00
o
<D
<3
1-1
<u
3
*9 .SP
T
T33
v
on ' I 1I ' I 1I ' I 1I 1I ' I ' I ' I • I ' I ' I • I '
5
o
93
<u
£
”43c0 v -I o
GO
C/3 ( s u i)a iu u .

O
<4-
o
•§ &
£ X
<u
tI ^N
’£
X N D
 <UE

Page 134
« o
£r - (o
o
j- cp
<u
Cl Oi
U E
5—3
'co
£
zet
_o

_C c«ou
GO OS—i
Qi
ro :S
1)
.s
0
1
<N
so 601.
* sot
cd

<u
>
tM : £6
o
: 68
Vi
<
d
_>

1998 PhD Dissertation

*s
o
: LL
"3
o : £L

o : 69
<u
H
<D
Jl
-LI

X)
-o
<u : es
o
3
TD
O

<D
T3
O
s : Z£
.2 : ee
-L
-*
C
_o
o : 63
cd
'-i-»
o(D <£
<
L>
q= Vi
<D
.C
H

C<O
U : ei
o£
cr>
(N
University Of Glasgow

"5
-C
CO <
L>
Vi
c 3 o m co
.EP ■*r CO
03 CO
CO
CO
co
CM
co
o

o.-S £ (ui) uouBAaig

■s
<D H&
r*
t% *-
N
;z n
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

P r o c e ss in g - a p p r o a c h -I

As the vibroseis source has a higher amplitude frequency spectrum, the vibroseis data
were processed first. The aim of this strategy was to process the data following common
practice in petroleum industry or in other words, the data were treated the same as for
deep seismic records. Shot gathers (e.g. Figure 3.26) are very noisy. The most
interesting noise appears on the far end of the shot record, where it shows a hyperbolic
pattern. This noise also appears on the impulsive source shots, which means it is not a
correlation artefact of vibroseis data, nor is it from the external generator for the
vibroseis source. It is probably reflected air blast energy from a building off to one side,
in line with receiver channel 125.

The raw data (Figure 3.27) were first automatically scaled by a factor of 35 and
corrected for spherical divergence based on a simple spherical spreading equation

(—— ) correction with true amplitude recovery (TAR) velocity function as listed in
dist
Table 3.10. Air blast signal was also attenuated (Figure 3.28). An elevation static with
350 m s'1 replacement velocity was applied to the data Datum for this elevation static
was set at 37.0 m.

Since only high frequency signals were important in this work, the data were filtered
using a frequency band pass filter (60-100-300-400 Hz). A dip scan stack (DSS) was
also applied to the shot gather, with the main aim to scan and filter out the linear
dipping noise events such as air blast and ground roll.

Table 3.10 TAR velocity function.

Time (ms) Velocity (m s*1)

0 350
10 800
20 1200
50 1500

University Of Glasgow 1998 PhD Dissertation Page 135

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

o
J-i
on

(-1

■(->

S
GO
ai
<

1)
*>

(N
L)

O
G

op
O
O

JO
o3

00
<D

(s u i) a u jji
•

>
c

O
>

■I—>
a

<u
<u
'5b
<L>

a<D

University Of Glasgow 1998 PhD Dissertation Page 137

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Although the deconvolution process can increase the temporal resolution (by
compressing the wavelet into a spike), it was not applied to the vibroseis data since it
only works best with a minimum or maximum phase wavelet. The vibroseis data has a
zero phase wavelet. In order to enhance the signal and remove the noise, an F-K filter
was applied. This was designed using F-K analysis (Figure 3.29). The F-K filtering
removed most of the noise (Figure 3.30). The data were not muted for the first arrivals,
as reflectors of interest were suspected to appear at very early times, less than 20 ms.
After the F-K filtering, the data were then sorted into CDP (common depth point)
gathers.

As a fixed spread geometry is used, the maximum fold is the same as the number of
shot records, i.e. 34 fold. Because the record length is short (150 ms) and dominated by
the unwanted signals, the velocity analysis for CMP gathers was difficult. Furthermore
there was no clear or sharp contrast between the glacial till and the bedrock as the top of
bedrock was weathered (thus lowering the velocity). The difficulties of velocity
analysis were compounded by lateral velocity variations that distort the expected
hyperbolae. So instead of using a velocity function, the data were stacked using several
constant velocities or CVS’s (constant velocity stacks).

Figure 3.31a, 3.31b, and 3.31c show the stacked section using 600 m s'1, 1350 m s'1 and
2054 m s'1 constant velocities. It seem that by using stacking velocity of 1350 m s'1 the
maximum reflection amplitude appear continuously uninterrupted around 20 ms. No
residual static were applied to improve the quality of the reflector. This stacking
velocity value is assumed to be an appropriate stacking value for the interface of the
bedrock. A summary of the approach-I processing flow is tabulated in Table 3.11.

University Of Glasgow 1998 PhD Dissertation Page 139

 oo oo oo oo oo og oo oo oo
cn co r-^ co co ^ cn cm t-
1) c

Page 140
P o
f—
r~ uu
h *+=
a>
a.
_c
U
<D
C/3
£
_o
"5
J=
C/3
If '!# *£ f.«
*.v;> j./**;i
Ch
-*(L>
->
a>
a
Cd
03
a,
«:V* 0 " ; . y
M* It •ITJ# "
%*%\ Af
_<•. i.
, : , : '' Xa ,y *Ti*
■
(-Li
<D
,C
+->
6JD
.s

1998 PhD Dissertation

’2 ■ . : .1 VVj
.S
"wf*
Ta>
3
Ci
<8
T<3U
(Z3
3
C/3

cd
C
cd

Puii

ON
(N
c <
o CU
l
o
<L> .2P
<u ptH

a>
C/3
£
O
”3
University Of Glasgow

4c/3d

"§
£ X&
*5 H
N r+-
Z N (suu) auuLL
ru
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

!•

l l l |I S S i l l ; S

University Of Glasgow 1998 PhD Dissertation Page 141

 Chapter Three

Page 145
3-D Shallow Seismic Reflection

bo
’o
cd
Cl
C /3
d>
O
Li
3
o<73

(N
bo
o
cd
CL
<73
Li C/3
<u CD
*
> ocd
<o -5*-1->
o
< 0 T3
5-1 cd
X, C3
cd

University Of Glasgow________________________________________ 1998 PhD Dissertation

cd
■'d' CL
cd "d- m
T3
T<3L> T3 o
cd -'d-
td <Lui
3 C M rr
fc V} r - ir>
o T3 O
0 <u ^d- CN
XI r- CL>
X T3 C/3
X
o
o
c-1
Li
o
a
§ o XO o
X
o
o
^ri m
o
<d <L> o C
bo o o
CD
T3 cd
m■ Li
CD
-«->
(7 3
CO X o
c/3
O
Li o o a
x X Q ^r a o
cd IT) m cd o
H > (N m m Ph X
New Methods In Shallow Seismic Reflection

O'
<D
a
cd
aj V h
+-» C
cd
* Vi
/5
'4—
cd CD
ts
o
c3
Q cu C/5
"3 o
§ ."tl
.a 5 *§ C/3
T3 x I Q >
&O <d
w <
cd
« PH U U
Z Z T Harith
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

A p p r o a c h -I: r e s u lt a n d c o m m e n ts

The section shows a discontinuous reflector at approximately 20 ms TWT on the 1350

m s'1 CVS section. Using a simple equation (velocity = distance over time, or distance
= 1350 m s'1 * 10'2 s), the approximate depth to this reflector is around 13 m. The
actual depth (the depth calculated from an appropriate velocity function) is less than that
as the section is stacked using single value of velocity 1350 m s'1. This reflector may
represent the top of the bedrock (Carboniferous sandstone and shale) since it is flat-
lying.

If it represents the bedrock reflector, it should be continuous over the whole section.
Since the line was oriented parallel to the dip direction, the deeper dipping reflectors are
also to be expected under this horizontal reflector (according to the refraction model).
Because of these unsatisfactory results there was little point in proceeding to process the
impulsive data using the same processing flow.

P r o c e s s in g - a p p r o a c h -II

In approach-II, the data were processed using a bare minimum of processing. The
primary goal was to stack the data without having to worry about adverse processing
effects such as the risk of spoiling the wavelet (by deconvolution) or stacking the
refracted arrivals (only short offset traces were considered).

In this strategy, the data from all three sources were processed. Since there was no clear
indication of any reflectors in shot gathers, and the far offsets were severely
contaminated by unwanted noise, all the data were sorted into CMP gathers prior to any
processing. The data were sorted into 276 and 263 CMPs for vibroseis and impulsive
data respectively. The spacing between each CMP is equal to half of the receiver
spacing, i.e. 0.25 m. As the data were collected using a fixed spread with 0.5 m receiver
and 2 m source spacing, the fold of coverage increases by 1 every 4 CMPs. The

University Of Glasgow 1998 PhD Dissertation Page 146

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

maximum fold of coverage is same as the number of shots, which is 34 for vibroseis and
31 for both impulsive datasets. Since the source interval was 2 m, the average offset
increment between traces in CMP gather is 2 m. Figure 3.32 shows example CMP
gathers for each of the sources. The data are severely aliased spatially. In order to cure
the spatial aliasing problem, the CMP offset was rounded down to the nearest integer
using a simple mathematical expression as follows:
i j ( c D p +v

Then all the CMPs with the same integer number were combined. Thus four CMPs at
0.25 m interval were combined to form supergathers at a 1 m interval. This reduces the
number of CMPs by a factor of 4. The total numbers of supergathers is 71. Figure 3.33
is an example of a supergather for each of the three sources. They all show quite an
improvement of the raw data appearance, compared to the original gathers (Figure
3.32). However, air blast and ground roll are still dominant. In order to avoid
misinterpreting the refraction events as reflections, extra care was taken in muting the
first arrivals. However, as the reflection signal was expected to be very shallow (right
below or even masked by the refracted signal), the muting process was only applied to
positively identified refractions (Figure 3.34). All the signals below the air blast were
also muted out (Figure 3.34). The muting was done on the supergather CMPs. A
bandpass filter was also applied to the supergathers. An F-K filter was not applied as
most of the noise was successfully muted out.

At this point the supergathers were re-sorted back to their original CMPs. To improve
the quality of the data, a minimum phase predictive deconvolution with a very small (1
ms) prediction distance was applied for the impulsive data. Because the deconvolution
would only be successful if the input signal is of minimum phase, such deconvolution
was not applied to vibroseis data. For the vibroseis data, frequencies in the 100 to 400
Hz range were boosted using spectral shaping. The CMPs were stacked using a 1-D
velocity function with a 60% NMO stretch mute. Results are shown in Figures 3.35a,
3.35b and 3.35c.

University Of Glasgow 1998 PhD Dissertation Page 147

(siu) auiii
University Of Glasgow 1998 PhD Dissertation Page 148
 c

Page 150
0)
£ o
£r™ S<u (SUl) 3Uljl
£>- ^<u
CL
£ .2
O E
crt
'S
00
£
o
-C
CO

T3
.(U
'Hh
cCd, l

<D
-4—>
CS
s
so
•*->

o
X
T3 ^
§
■4— *
e

1998 PhD Dissertation

o
<
VD
i
<U
>
<L>
CO
Vi
CL)
<K
cd
Vi
u
-CS
-4 -1
cd
00
Vi
<D
&4
3
C/D

m
0)
Vi
co 3
op
ou
4—»
<

<u
oo
£
J2
University Of Glasgow

-C
00 (sui) aui;i

”o ■S
- s
f6 NH
IZ N
 Chapter Three
3-D Shallow Seismic Reflection

Page 151
xi
o
cd <N <0
o
o* Ml
.C #o
§■ 3cd 3
<o
cd Pi
C/3
Vh
Pi
od J-i C/3
Td <o
<o aj
• ^"4
3<o
C/D CO
3 £ CO

l IT) SJ
(3
0)

1998 PhD Dissertation

O
C/3
CO td
J-h
oVh CO
x> -a
CO
Pi
O
x
53
O
£ 4->
<o
O
53 ”2
bO o
<o <0 o
M|
"cC/3o TS33
CO

CO *C/3
o cd
3
2 Pi Q
Pk
> <N

(N
m
New Methods In Shallow Seismic Reflection

<o

H to
.B
&
'w
b£) 3
8 .s -*->
o
o
University Of Glasgow

CO
CO
o m
Pi
Z Z T Harith
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

University Of Glasgow 1998 PhD Dissertation Page 152

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

(SUI)
3UIT|
University Of Glasgow 1998 PhD Dissertation Page 153
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

University Of Glasgow 1998 PhD Dissertation Page 154

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

A p p r o a c h -II: r e s u lt a n d c o m m e n ts

The results (Figures 3.35a, 3.35b and 3.35c) show much resemblance to each other. As
with approach-I, all three sources show a reflector at 15 -20 ms TWT. This is probably
the reflection from the top of the bedrock. However, the results of approach-II show
another reflector below the bedrock reflection. This reflector dips to the right (NE), in
the approximately down-dip direction for layers within Carboniferous bedrock. The
refraction model (Figure 3.25) also shows this feature.

3.10.3 3-D seismic data processing

I n tr o d u c tio n

Although the actual idea of this project was to conduct the first combination of frill and
random 3-D survey for shallow investigation using three possiblessources - vibroseis,
sledgehammer and weight drop, the decision had been made toacquire the majority of
the 3-D data using only the vibroseis source. The reasons were that to use all three
sources would have required triple the field work time, which was not available. After
analysing the frequency content of the three different sources from the first swath
(CDE), it was evident that both sledgehammer and weight drop have lower dominant
frequencies (30-80 Hz) than the vibrator.

Both correlated and uncorrelated vibroseis data were recorded. The uncorrelated data
were stacked four times to increased the signal to noise ratio before they were cross
correlated with their pilot sweep to produced the correlated data. Like the 2-D where the
data were processed by ProMAX/2D, the 3-D data was also processed using ProMAX
software, however instead of ProMAX/2D, the 3D version was used. Only the vibroseis
data were processed. The impulsive datasets (from the first swath CDE) would only be
processed if the results from the vibroseis data were promising.

University Of Glasgow 1998 PhD Dissertation Page 155

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

F i e l d c o r r e l a ti o n

Correlation is the first process applied to the uncorrelated vibroseis data. Normally, the
correlation process is done in the time domain by a purpose-built ‘correlator’, and it is
common practice to record the correlated data only, as the long uncorrelated data use
lots of recording space or tape. With the DAS-1 and minivibrator recording system, the
field correlation is done slightly differently from the conventional method. Here the
baseplate and reaction mass accelerations are measured and recorded on two auxiliary
traces. The weighted sum of the two accelerations is used in a convolution filter
process, which replaces conventional correlation. This is done in order to get a better
estimate of force applied to the ground (Smythe personal communication). This
recorded correlated data is defined as the field correlated data. Because the DAS-1
system correlates the uncorrelated signals using an essentially ‘blind’ convolution
process, both correlated and uncorrelated records were recorded. Although the
uncorrelated data consume lots of recording space, by recording the uncorrelated data
(which were later correlated using conventional methods), comparison between the
correlation methods can be done.

L a b c o r r e l a ti o n

In conventional correlation, a synthetic pilot sweep was generated using ProMAX

software. This synthetic sweep matches the sweep used in the field, as listed in Table
3.13. Using the vibroseis correlation processor in ProMAX, all the uncorrelated data
were then correlated and stored as ‘100-700 Hz Lab. Corr.’ datafiles.

The correlation process can also be used to filter the data; for example if the
uncorrelated data are correlated with a matching sweep as before, but only having finite
amplitude over 200-500 Hz, the correlated data will only contain frequencies in that
range.

University Of Glasgow 1998 PhD Dissertation Page 156

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

With that idea, a secondary dataset was constructed by correlating the uncorrelated data
with a 200-700 Hz synthetic sweep. Thereby the low frequencies (lower than 200 Hz)
are filtered out. The original 100-700 Hz sweep was a linear sweep, meaning that the
rate of increase of frequency is linear. So for the frequency range of 200-700 Hz, the
sweep must be 3208.33 ms in length to keep the same linearity relation:

700Hz -2 0 0 Hz _ IQOHz-lOOHz
x 3850ms
x = 3208.33 ms

But if we generate a synthetic sweep of 200-700 Hz over 3208.33 ms in the same way
as generating the 100-700 Hz sweep, we create a sweep which has different phase from
the 100-700 Hz sweep. So this sweep cannot be used to correlate the uncorrelated
vibroseis data.

To generate the 200-700 Hz synthetic sweep with same characteristics as the 100-700
Hz sweep and maintain the linearity relation, we generate a 100-700 Hz sweep with the
front end muted up to 200 Hz. Since the sweep has a linear relation, the amount of
muting can be calculated precisely using a linear equation as follows:

200Hz -100Hz 100Hz -100Hz

x 3850ms
x = 641.67ms

Because the sweep (100-700 Hz with 641.67 ms time delay) was generated using 100-
700 Hz sweep, the relation between these sweeps were same, however this sweep only
has frequencies in a range of 200-700 Hz (Figure 3.36).

University Of Glasgow 1998 PhD Dissertation Page 157

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Table 3.13 Sweep parameters.

Sweep Types Linear

Start Frequency 100 Hz
End Frequency 700 Hz
Sample Interval 0.5 ms
Sweep Length 3850 ms
Taper Type Cosine on both ends
Filter 1000 Hz high frequency;
96 dB/octave slop

Table 3.14 Synthetic sweeps parameters used for lab correlation.

Parameter 100-700 200-700 200-700

Delay time (ms) 0 0 642
Start time (ms) 0 0 642
End time (ms) 3850 3208 3850
End tapers (ms) 100 100 100
Result 100-700 Lab Corr. — 200-700 Lab Corr.

Figure 3.36 Graphical presentation of 100-700 Hz, 200-700 Hz and 100-700 Hz

with time delay sweeps.

100-700Hz, 3850ms sweep

200-700Hz, 3208.33ms sweep

100-700Hz with 642 ms delay

0ms 3850ms
100Hz 400Hz 700Hz

University Of Glasgow 1998 PhD Dissertation Page 158

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

C o m p a r i s o n b e tw e e n f i e l d a n d la b c o r r e l a t e d d a t a

Figure 3.37a, 3.37b and 3.37c shows the quality of the field correlated data compared to
the 100-700 Hz lab correlated and 200-700 Hz lab correlated data. Each figure has three
panels of the same shot gather representing the results of three different correlation
methods. The left panel represents the result from the OYO field deconvolution process,
while the middle and right panels show the output from the 100-700 Hz and 200-700 Hz
lab correlated processes, respectively.

The field correlated dataset shows a different polarity compared to the lab correlated
datasets. From the polarity comparison in 2-D of the 100-700 Hz lab correlated
vibroseis data and the impulsive datasets, the polarity of the lab correlated vibroseis
data was same as the polarity of the impulsive datasets. For that reason the field
correlated dataset has been reversed in polarity in the displays of Figures 3.37a, b and c.

Figure 3.37a shows the first 48 channels for shot 401, which represents the typical
characteristics of data from small offset (2-15 m) gathers. Figures 3.37b and 3.37c on
the other hand show the qualities of the first 48 channels for the shot 101 and shot 51
respectively, which represent typical traces at intermediate (15-30 m) and far (>25 m)
offsets. The chevron effect that appears on all shot gathers is the result of the receiver
swath geometry where the source-receiver offsets varying up and down the six-column
spread (see Figure 3.14 above).

The effectiveness of all the correlation methods can be discerned partly by analysing the
relative quality of the air blast on the three panels. As it is a coherent signal, the
convolution process in theory should enhance this signal as it enhances the reflection
signals; however the field correlation datasets show the air blast in the left hand panel of
Figure 3.37c is very poor compared to that on the other two (lab correlated) panels. This
effect is less seen in near offset gathers (Figure 3.37a) as it is masked by early arrival
signals.

University Of Glasgow 1998 PhD Dissertation Page 159

0> C

Page 160
u o

si u
C3. 0, £,
Q
U
eo
C/D
£
_o
15
x:
c/ d 75

Q cd ? 8 '• g • 8 2 8 .8 •?.
iiiiiiiiiiiiiiiini i n ii ii in ii n ii in iiiiiin ln n iiili mmiiiliiiiinii i ill IIII ll nm ii nln n n n iln i m i itl n i
£
+-*
e
<L>
S-l
,<o

<u
-»->

"a3
fc!
O
0
If
IT)

1998 PhD Dissertation

1
<N

O
rC
'JPjr
!>

W
Vh
1)
X2
cd
00
p f jl
■4—> w fii
Xo753
£ l\
p*43 -I:*:>.s's&
U/JS.Sr* *
r'T
r v /’. r mT
cd
r-
cn
m III II l l l l J ll III I III | l I I I I I I l l | l i l t II III | l III I III t | I I I I I t i l l | | | | III | | t | ft | | 11| | | | | | | III | | | | | | | | 11| It | | HI | | | | | | HI 111| | | |H I I I I I I l | | l | I I I I j l l l

<D
*-l •2 .. 8 8 - ?■. 8/. 8;.;: 8,'. 8 8 8 /2 -'8 • B'W?
3
00
(cui)diu(i
£
o
University Of Glasgow

X
C/5

o5
5 DC

EN
N
(sui)aiuii
 <uc

Page 162
CD o

i— CD<u
1—GZ
£cl aC
CS o
LD
U E
—
‘ <U
co
£
_ o

0. c0
on 2 8 8 ? 8 8':>8^A§ ’>• g ■ S- -g.- 9
IU1UU1 miiiiii Liiiimi miniii
imiiiiiliniilrii miuui
03
£
+ -»
C
CD
Lh
<£

T3
CD
■4— 1
J g
*0)
Lh
5 -!
O
o

1998 PhD Dissertation

IO
CM
A

<D
O
oo
c
o

Lh
t ?&y. $_
0)
_C
oj
00
+ ->
o
x:
CO

£
03
Pd
c
_o * H fS
o<u o
r-
rn
<D
rA
<u ,".,,-J,l,|,"""474 > - , i|,,ii|,|," ,,," ,| '," ,iii,|"n ," n ||iiniiii|iiiiMiii|iMiiiiH|Miiiiiiiiiiiiiiiiiiiiiiiiiii|iiimni m)
«-
3
00
c<L>/3 ■: ■■.••-■-. (tiujauiti v ; >' ' i*
o£
University Of Glasgow

JZ
CO

"o "B
•§
<D*Tc3
*

tS H
N
Z, N
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

From the air blast effect, we can conclude that the weighted sum of the two
accelerations as input to a convolution operator for correlating the vibroseis data is
possibly unstable. Because of this, it was decided not to use the field correlated data in
any subsequent processing.

Field correlated data are very ringy and dominated by low frequency signals. The
figures also show that most of the low frequencies are filtered out from the record when
the data are correlated using 200-700 Hz synthetic sweep (Figure 3.37c). The interesting
effect from this 200-700 Hz correlation is that the first arrivals (from near offset traces)
are very sharp (Figure 3.37a). These results suggested that for near offset processing the
highest resolution might be obtained by using 200-700 Hz correlated dataset.

P r o c e s s i n g s tr a t e g ie s

In order to obtain the best result as possible, both 100-700 Hz and 200-700 Hz lab
correlated data were used. The 200-700 Hz lab correlated dataset was processed first.
For this data, only the short offset traces were used as the 200-700 Hz. correlated data
show a very sharp first arrival (from near offset traces, Figures 3.37a, b, and c).
Furthermore, for the long offset traces, the returned signals were very weak and
contaminated by random noise. The reflection signal is curving with the offset. The
longer the offset, the better the hyperbola that can be observed. The problem with using
only near offset traces (as in first approach) is that it is very difficult to differentiate the
reflection from the refraction signals, as both appear straight. This may create a problem
during the stacking as the refraction signal is assumed to be a reflection signal.

To overcome these uncertainties, longer offset traces were considered for the second
approach. However, because the returned signal was very weak and hardly seen in 200-
700 Hz shot gathers, the broader range of frequency dataset (100-700 Hz) was used.

University Of Glasgow 1998 PhD Dissertation Page 163

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

P r e -p r o c e s sin g

Processing the 3-D data is not fundamentally different from processing the 2-D data.
For example, the data need to be labelled with their appropriate FFIDs prior to any
processing, then editing bad traces or removing unnecessary or problematic files based
on the Observer’s logsheets and visual inspections is a task that has to be done. Other
processes, such as geometric spreading correction, elevation statics, and data
enhancement need to be applied to all seismic reflection data. Basically, almost all the
concepts of 2-D seismic data processing apply to 3-D data processing.

Complications in 3-D data processing arise in geometry quality control, statics, velocity
analysis and migration (Yilmaz 1987). Unlike in 2-D, where the data are acquired using
a linear spread, the 3-D data is acquired in an areal mode spread, in this instance with
randomised source locations. This leads to differences in CMP sorting.

The 3-D CMP binning concept is essentially the same as in 2-D processing. However
the traces in a 3-D CMP bin have different azimuths. This leads to difficulties in
analysing the velocity for NMO stacking, especially in areas where lateral velocity
changes are present.

E d itin g a n d g e o m e tr y a s s ig n m e n t

The first task before any processing was carried out was to add the correct FFID (field
file identification number) to each file. This FFID number is unique, based on the
Observer’s logbook. Besides referring to notes in this logbook, like the 2-D data, the 3-
D data were also visually checked to identify any previously un-noticed bad or reversed
traces. These bad traces and any other irrelevant data (such as noise tests) were edited
out prior to any processing. The reversed traces were also corrected to their correct
polarity.

University Of Glasgow 1998 PhD Dissertation Page 164

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Before the geometry was assigned to the traces, any discrepancies in shot location must
be corrected. This is a very important step in this 3-D survey, since the sources were
placed in pseudo-random positions. For that, the shot positions (determined from the
randomised table) were sighted using the theodolite after the shots had been taken.

Checking the source locations was done by making detailed GMT maps of the proposed
location (from the randomised table) and the actual location (surveyed after the shot had
been taken) as shown in Figure 3.38. In most cases the discrepancies between the
proposed (from table) and actual (surveyed) position are very small, however in some
cases there were also big discrepancies such as in column 12 of Figure 3.38. For such
big discrepancies, checks on shot location (to find which one is correct) were done by
examining the air blast travel times (and hence shot-receiver distance at the constant
velocity of 330 m s'1) on the correlated data. Figure 3.39 shows the entire corrected
survey source positions as vectors from the peg to the actual shot.

After the survey information had all been corrected, this geometry information was then
loaded into the ProMAX database and then applied to the seismic data. From this stage
onwards, every seismic trace carries the geometry information in its trace header.

B inning

Figure 3.40 shows all the source-receiver midpoints. They are so abundant in most of
the survey area that they turn the diagram solid black. The one gap (white area) in the
middle of the survey area was a result of the spread shifting due to an unavoidable
obstruction (drilling rigs). A grid of 2 m square bins was defined and fitted as a
uniform grid to the dataset (Figure 3.41a). The binning grid azimuth was defined to be
the same as the surface geometry azimuth, although the subsurface grid could have been
aligned in any direction. Details of the binned CMPs are shown in Figures 3.41b and
3.41c where the former shows bins at the southern edge with few CMPs in each, while
the latter shows central bins with a very high fold of coverage.

University Of Glasgow 1998 PhD Dissertation Page 165

 New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Figure 3.38 Detail o f offset vectors.

80

78

76

74

72
6 8 10 12 14

The black arrow vectors are the randomised positions from the table while the white
arrow are actual source positions as surveyed after the shot

University Of Glasgow 1998 PhD Dissertation Page 166

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.39 All corrected source vectors.

i i ' * 1 ■ * ■ i
116 : 1 : ' 1 1 ' ' ' : : - : ! ■
1 1 4 ------------------------ t ----------
/ *---- / y — y ---- ---------- — '-
-J-J2 ___________y ^ »i /t t J/ ~- \---------------
.L. y t —yZ—y—s ' 1-------------
y
■J-J Q __________ S ______ t tt___ d£y______
^ 1____ t 1_t.____ tt___ —. ..y^.-1.1. ..._^r
------- f .SJ.
«*---- f------- s.... y
---—----- ----------

1QQ 4. / ' t4
/ K —■ y • / * -y^ ; I . /?—
j ... / / ' y . *-y£-L.. i 1. I.i —
4 \

-jQg t/ ^ tt // !/ y* y/ y
S 1 // ‘ 4 . —d I
S^J /— / ------
Z
-j Q 4 / ^ SS' '-*ri_d
__ / y—. Sy L/ __ L/ ----- / / •. tt / / - // // ij_

102 // tt * t1 ^* _ >t Z/ sy ''1 / // // *t— . // ^ —i— _

1 0 0
/ 1*— *4— ----------- ;—*— *— ^
^— y*' / / / ^ -
gg // /' y t /y /y t4 / / - ___ //_ —Z/ _— tc----- /-/------i 4 ----- --------

gg It /I t *t___/ / y* f y —*■_—t— //.

/ /y 1 xZZ3— —s . —
g4 t .__ / / t/----- /y ~, /
/____ 1t , / /___ —«— Lt /
Z_— t .-r^.
-«u---«------ {-------- 4
— t--------

g2 // t > S ^ /y 1 J/. . /./ (/ _/ / /.4 ,_ , t {

/----- y --
Si
g0 / 'J4 SS t\ // // - j ^ T Z/ —<_
1 . _ It ]7--------
/
gg *, //'' ft // \f // 1t / / ^ / t t y. f--- // j / _
gg - —-JZZ i __ _ y —Z
i-----Z / ■—- y .-*• /J'
Z' Z
/ ---------
•
g 4 4< ^/ < ^ __// . .. I
d______ L y
dZ! /
Z £t -rJ *------ -------------- fi *1----------------
Q2 - ■■ // ■'. / / ’ — ■ y, _ /t ■ / .._ii Z/ .' dr. . ✓
s —: / t ■■ Xy. ■ , -

gg / - _— ,< //' ' ii : t/ S

^ y*1' I / // : y tf /..._ /~ J/
78 - V/ t/ . // . /y ' /i f/ X /' T/ '7-/; y t /
s\ ■/ > / / [ , / ■/
7g » y V '- - y y y> S' / __ Z
t s>•' - i/ t____________________/ / ! -*__ //__/ ; If Z_______
/
/ /!/ /'/ >> L>gi^y /..
/ >>• 'i________ ______ *- ;__;---- :________ _______
7 4 sS '// '!/>
f f . l1 t t_ y/ < ^ 1 .«• /_^ \ S *y.' / ^---- 1—: 4 ____________________
f — «— t
/ Lr Sy ’y f / // t/ .": t1. 1.N y/ >^ / y/ - Vy ________ _______________________
72 —^ ^y >I >i 'y t / .\ t#- f/ 1t /t -// /1 - /Z : / / y *4 sy '! Sy / /' '
/s S s / L_ / ^. ; t1. // /4 4/ /f t/-i--- jj__f T_— ---- ----------------------
7 0 _______ . y/ /y - !/ *-r * . / - // /S< '. N X i1— L/ L L_ ( Z - — Z_ / -tj___________ ______
jf _ ^ / Z /
> \m. . / S /. / i Z_______________ _______ ____ :____ _ ; ;

ea ^- / ^* /f j„, /f li /t ,.,, - yy L ^ y . /L / t...s

/ L L c _Z/ Z' y ,__ :____ ; .... _
/ t sS'S
S't 'S'^'- v/ jl
S It t y >y / /J £l£i / / ■t \.: • : :__ ;__ _____
eg _ . S. iL yS'/
/'< '/ S y '// /y /f 4 ( sy / 1 ^ // i f «— /— / . . . , ___ _
bb > y* y L y/ -^ y/ 'y, . yiLl y w.. / y :__ ;__ ________ :____ ... ..
g4 1 / S -/> / /'d i_^
J y / y / ' ^ ____
S’ t t___Z/ -!_______ -cj_________________
/ - r / i d ./— / yf y/ Zy il. ._ /4 / yL,«. / // __ Z __ i____ i____ ;_____ '___
M * / / > < / z Ir S / t / / \ , > ■ - I— f Z__. ____ !___ db___' ;________ _
y"\
y .// / /Zd y ’- // I» ■_ 1 ^ / / . / /r / - ■ _____
go /L-t/ ,. f/ 4, ys?//-'-~ - / /'■— y y ~- 11____ _____ ______________ _____________
_ // // y0 - JL/

58 . | ■ |I i1 |I ■
1 |-
I"■1 v
T 1 iI 1 ,, ■
1 iI" ■
11 iI " '■ iI ■
1 iI ■
' iI ■
i ■i ■
— 1 iI ■'■ iI 1 iI 1
-2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34

The arrowheads represent the position of the source.

U niversity O f G lasgow 1998 PhD Dissertation Page 167

 N ew M ethods In Shallow Seism ic Reflection C hapter T hree
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.40 CMP coverage map.

70505.000

70500.000

70495.000

70490.000

70485.000

70480.000

70475.000

70470.000

704B5.000

704G0.000
CDP Y Coordinate

70455.000

70450.000

70445.000

70440.000

70435.000

70430.000

70425.000

70420.000

70415.000

70410.000

70405.000

70400.000

70395.000

70390.000

55520.000 55540.000 55560.000 55580.000

COP X Coordinate

University O f G lasgow 1998 PhD Dissertation Page 168

 N ew M ethods In Shallow Seism ic Reflection C hapter T hree
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.41a CMPs with 2 m binning grid overlain.

70505*000^

70500*000-=

7 0 4 9 5 .0 0 0 -5

70490*000-5

7 0 4 8 5 .0 0 0 -5

70480*000-5

70475*000-=

70470*000-5

7 0 4 6 5 .0 0 0 -5

7 0 4 6 0 .0 0 0 -5
CDP Y C oordin ate

70455*000-5

7 0 4 5 0 .0 0 0 -5

70445*000-5

70440*000-5

70435*000-5

70430*000-5

70425*000-5

70420*000-5

70415*000-5

7 0 4 1 0 .0 0 0 -5

70405*000-5

7 0 4 0 0 .0 0 0 -5

7 0 3 9 5 .0 0 0 -5

70390*000 —

55520.000 55540.000 55560.000 55580.000

CDP X C oordin ate

The grid is laid parallel to the surface column and row directions.
In this way it minimises the occurrence of empty bins.

U niversity O f G lasgow 1998 PhD Dissertation Page 169

 N ew M ethods In Shallow Seism ic Reflection C hapter T hree
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.41b Detail o f binned CMPs at southern edge o f survey.

7 0 3 9 8 .0 0 0 -

7 0 3 9 7 .5 0 0 -

7 0 3 9 7 .0 0 0 -

7 0 3 9 6 .5 0 0 -

7 0 3 9 6 .0 0 0 -

7 0 3 9 5 .5 0 0 -

70 3 9 5 .0 0 0 -
■■ . ( V / V v y y ; .V
7 0 3 9 4 .5 0 0 -

70 3 9 4 .0 0 0 -

7 0 3 9 3 .5 0 0

' \ '/' 'v \ ■ V

CDP Y C oordinate

70 3 9 3 .0 0 0

70 3 9 2 .5 0 0

703 9 2 .0 0 0

703 9 1 .5 0 0

70 3 9 1.000

703 9 0 .5 0 0

70 3 9 0 .0 0 0

70 3 89.500

7 0 3 8 9 .0 0 0 -3

7 0 3 8 8 .50 0 —

55542.000 5 5 5 4 4 .0 0 0 55546.000 55548,

5 5540.000
CDP X C oord in ate

University O f G lasgow 1998 PhD Dissertation Page 170

 N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.41c Detail of binned CMPs from centre of survey.

7 0 4 1 4 .5 0 0 -

7 0 4 1 4 .0 0 0 -

7 0 4 1 3 .5 0 0 -

7 0 4 1 3 .0 0 0 -

7 0 4 1 2 .5 0 0 -

7 0 4 1 2 .0 0 0 - m

7 0 4 1 1 .5 0 0 -

7 0 4 1 1 .0 0 0 -

7 0 4 1 0 .5 0 0 -

7 0 4 1 0 .0 0 0 -

7 0 4 0 9 .5 0 0 - 8
CDP Y C oordin ate

7 0 4 0 9 .0 0 0 -

7 0 4 0 8 .5 0 0

70 4 0 8 .0 0 0

7 0 4 0 7 .5 0 0

7 0 4 0 7 .0 0 0

70 4 0 6 .5 0 0

7 0 4 0 6 .0 0 0

7 0 4 0 5 .5 0 0

7 0 4 0 5 .0 0 0
S
70 4 0 4 .5 0 0
m
7 0 4 0 4 .0 0 0
I
70 4 0 3 .5 0 0
U
7 0 4 0 3 .0 0 0

70 4 0 2 .5 0 0
m

55538.000 5 5 540.000 5554 2 .0 0 0 5 5 5 4 4 .0 0 0 5 5 5 4 6 .0 0

CDP X C oord in ate

University O f G lasgow 1998 PhD D issertation Page 171

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

F o ld o f c o v e ra g e

Figure 3.42a summarizes the fold of coverage (in plan view), whereas an isometric view
is shown in Figure 3.42b. The fold of coverage ranges from a minimum of 1 to
extremely high, as high as 570 fold. The exceptionally high value of 570 fold is
achieved at two peaks in the lower left comer where the 2 m source spacing was used
instead of 4 m spacing. The total number of bins is 53x33 or 1749, with the mean fold
coverage being about 80 fold. This is a very high figure compared to 3-D surveys even
in petroleum exploration. A summary of the statistical figures for the 3-D data is given
in Table 3.15.

Table 3.15 3-D vibroseis data statistics.

Total no. of shots 979

No of receivers per swath 144
Total no. of traces 140976
No. of 2x2 m grid bins 53x33 =1749
Average no. of traces per bin (fold) 80.6

University Of Glasgow 1998 PhD Dissertation Page 172

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T I larith 3-D Shallow Seism ic Reflection

Figure 3.4 2 a Fold o f coverage in plan view .

70500

400

■380

■300

-1 4 0

100

80

55 5 4 0
CDP X_C00RD GEOHETRY

University O f G lasgow 1998 PhD D issertation Page 173

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z f I larith 3-1) Shallow Seism ic Reflection

f igure 3.42b 3 -D view o f fold o f coverage

pD P FOLD GEOHETRY

.CDP Y.COORD GEOHETRY

50 100 150 200 250 300 350 400 450 500 550
CLP POLL GEOHETRY

Peak corresponds to the area where dense data were collected.

The average fold is 1-150 although some bins reached 550 fold.

U niversity O f G lasgow 1998 PhD D issertation Page 174

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

P r o c e s s in g - a p p r o a c h -I

As the aim of this work was to image the shallow underground tunnels as well as the
shallow geological conditions, high-resolution data are really important. Therefore the
200-700 Hz lab correlated dataset was used instead of 100-700 Hz lab correlated
dataset. By using the higher frequency range dataset, it was hoped that the smallest
possible reflection times could be resolved, which would be very useful in mapping the
base of the boulder clay (top bedrock surface) at 3-6 m depth. In this approach, the data
were reversed in polarity to match the polarity of the field correlated data. It was only
discovered subsequently that the polarity of the lab correlated dataset has the correct
polarity, based on the analysis of the impulsive 2-D datasets discussed in section 3.10.2
above.

Instead of using all the traces, only small offset (<5 m) traces were used. This was due
to the fact that in far-offset traces, the returned high frequency energy was very weak
and was hardly recognised in the gather. To ensure only reflected arrivals were stacked,
and to avoid any possibility of stacking refracted arrivals, a severe front-end trace mute
was applied to the data (Figure 3.43).

This approach also had the aim of not degrading the source wavelet, hence no
deconvolution was applied to the data (since the deconvolution only works optimally if
the data are minimum phase). However in order to enhance the frequencies of likely
use, 200-500 Hz spectral shaping was applied. Spectral shaping keeps the phase of the
wavelet unchanged while increasing the amplitude of the frequencies specified.

Because only near offset traces were used and no reflectors were visible in the deeper
section (later times), the data were stacked using a single stacking velocity o f 350 m s'1.
This value is the velocity of first layer (overburden) as given in the refraction analysis
(Figure 3.25). A summary of the processing flow used in this approach is given in Table
3.16.

University Of Glasgow 1998 PhD Dissertation Page 175

 Page 176
Si <u (sui) auu;i
u aP
d> © © © ©
00 W CO ^
£
_o
Xo3 iiiu Jiuii uuiudmi JiiJiiliujmnLjiuuill
-c ccS
on O
&.
&

T3
.a
*0.
&
<L>
1
C
O
cfa
<L)
>
<L>
C/2
00
a

1998 PhD Dissertation

M
C/2

x:
•2P
’C
cj
■4— >

TccS3
T3
(L
C/>2
C/2
<D
O
0
C1o
<L>
&

| V
ccS
04
qs
u
Od
m
^t
<u m’
on <o
£o II III 111IIIII11II IIIII III IMil II III111II III 11 III I
op
University Of Glasgow

© © ©
X £ CM CO ^
c/2

TO
S (sui) auiM
X
id
S
£
Z
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

jd
ot-i

<

'S
H
uo

-a

m
ot-H
&
C/3
C/3

p,
P<

VO
Ph
Data 200-700 Hz Lab Correlated
0-50 ms time
<5 m offset only
reversed polarity (matches the field correlated polarity)
Elevation Static Datum 37 m; replacement velocity 350 m s'1
AGC 20 ms centred windows
Spectral Shaping 200-500 Hz
NMO / Stack Constant velocity stack at 350 m s'1
AGC 10 ms centred windows
Pad stack volume
Output

University Of Glasgow________________________________________ 1998 PhD Dissertation Page 177

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

A p p r o a c h -I: r e s u lts a n d c o m m e n ts

Since the polarity of the data was reversed, a positive impedance contrast is represented
in displays by a black peak. The results of this approach are shown in Figure 3.44a,
where it represents a pair of typical sections - inline 133 and crossline X I6 from the
data volume. There are no clear indications of any reflector, however there is one fairly
continuous signal at a very early time (<20 ms TWT). This is interpreted as the base of
clay or the top of the bedrock. This reflection wavelet is clearly made up of peak -
trough - peak - trough - peak, spanning 10 ms. Thus the dominant frequency is about
200 Hz.

The geology of the site is very simple, a glacial deposit un-conformably overlying
gently dipping Carboniferous rocks. The velocity of the overburden is less than the
velocity of the bedrock, which means the reflecting interface has a positive impedance
contrast, marked by the black peak. The central peak was picked as the top bedrock
reflector and marked by a red circle (Figure 3.44b). The picks are very jittery, indicating
a low signal to noise ratio.

The contour map of the bedrock horizon is shown in Figure 3.45. The time on the map
is represented as un-migrated TWT (two way time). The values range from 6 ms to
about 18 ms and a division by 3 gives an approximate depth in meters below the datum
of 37.0 m. Thus the depths vary from about 2 m to about 6 m below the datum.

University Of Glasgow 1998 PhD Dissertation Page 178

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

o
o

o
o
T-

T—
o
o
CM

CM
(s rn )iM i
(SUJ)1M±

o
o

00
CO
o

o
o

o
Lf>

LO
University Of Glasgow 1998 PhD Dissertation Page 179
C o
i— <u
in
Si
^ (sui) iM i
a. oi
& o
jd
U
<u
GO
£
_o
73
x:
cn

* i
a>
d>
s- T<U3
a,
>-(
<u _o
c ’S h

-O T“ cd
u
cd
O
X
J-H
Oh

cOd. _. \ 1 /1 ^ > > i

O
.2
-t—'
d>
o
CO ^ S
Q Ji
rr\
d)
.s
O
TO
3
cd
<L>

Cd GO

cd
00
T3
<u

hO §
T* - I
© o
c
c
€co0 esc O
<u
h

0
<u o H
hO

zp
<D .2P
01 Ph

<u

£
_o

_c
GO I
o o o o
CM CO Tf
o x:-
<U hhH (sui)iAU
6N
ZN
N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.45 Contour map o f base oI clay horizon based on approach-I results.
50
48

i t . '. ' •• . ■ . i—
!
-
,: ■■ ■* :
• --:u‘: . r!

I : If

2 4 6 8 10 12 14 16 18 20 22 24 23 28 30 32
Contour interval 1 ms TW T. Dark blue - 6 ms, Dark orange - 18 ms. grey -n o pick.

U niversity O f G lasgow 1998 PhD D issertation Page 181

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

P rocessing - approach-II

Unlike the first approach where the aim was to image the very shallow clay/bedrock
interface, this second approach aimed to image the deeper section. For that the broader
bandwidth dataset (100-700 Hz Lab Correlated) was used. It also decided to use the
whole 150 ms long trace, as hopefully some deeper structure would be imaged.

In this approach, the polarity of the data is kept as recorded, so the positive impedance
contrast is now represented as a white trough instead of black peak as in approach-I.
The long offset traces are also used instead of short offset traces as in approach-I. In
order to correct the topography errors, an elevation static correction with 37.0 m datum
and 350 m s 1replacement velocity was applied to the data.

Like the first approach, spectral shaping was applied to whiten the zero phase vibroseis
wavelet. This process retains the zero phase character while at the same time scales up
all frequency components within the desired window to the same amplitude. The air
blast and later arrivals were muted from the data. The first arrivals were not muted in
order to keep any early reflections arriving soon after the refracted first arrivals (Figure
3.46). To avoid stacking the refracted signals as if they were reflections, a low stacking
velocity at early times was used. For that, a 1-D velocity function (listed in Table 3.17)
with 60 % NMO stretch was used.

Table 3.17 1-D time velocity pairs used for NMO stack in approach-II.

Time (ms) Velocity (m s'1)

0 350
10 800
20 1200
50 1500

The stacked section was padded with zero traces to produce 52 inlines and 33 crosslines
before they were migrated and transferred to GMA-plus for interpretation. Table 3.18
lists the processing flow for this approach.

University Of Glasgow 1998 PhD Dissertation Page 182

 Page 183
i_ gh
(U
« <u (S U l) 3UIIJ.
Q. OL
.2
u P
.a
O O O O
C(U
/3 t- co m ^
£ JJ.LLLLLLLjjJJ.LilLLlimLlUllluiJJJ.iuiiLJJJJJ.iJ.
XoI
-C cd
c/o O
>-.
CL.
D h
cd

T3 v>_
.a C
M
■a
§■
3 r^
6 §
fi
o
ts irt_ ^1 ^ /“ -w■
o “ '
X
"5
o
00
c

1998 PhD Dissertation

'%
-8
c/3

£
00

cd
-*->
cd
T3
T4)3
C/3
C/3
<X>
o
o
<L>
t-i
a.
T3
G
cd

«
a>
c
o £
’■oC 3
<u
co<u
f*
VO
<D cn
C/3
* <L>
i-i
JO G III lllll I 11 III I I I I
University Of Glasgow

75 00 II II III II I I I II I I I I II II II I II

£ E Cd ?
c/i
’B s (sui) 3IUIJL
c3
§ DC
t% N H
12 N
 Chapter Three
3-D Shallow Seismic Reflection

Page 184
C/D
a X
o
o cd

£
o
T<3D O ' £
00
O
.a
C/3
cd
3
VO C
o —
\£
G
x>1
CO
<L> fc £ £ cd <u
To3 o To
g
O 0 3 M
OX) 3
C /3
o
l-l U g 0 G
ex,
•8 ^cd 01 a s

University Of Glasgow ________________________________________ 1998 PhD Dissertation

i—i TC3D T3 £ T3 b .&
X N N l-C
<D
N
o a
a a *o * g -§ a
Cd a
o o o <D o CD o
l-l
o o O C/3 3 O
ex r -1 I 3 > 3 a> o ■'d-
o o 6 oI
o o .G o cC/Dd o
< •'d-

oo
m
<L>

IH 00
g
<L>
00
G
Cd
H—> & o &
C/3 J >
G C /3 g C /3
<L» 3l-H 00 cd
.2 3I-,
C/3 cd H
a > O
-> .a u «-i -*—
o
>
New Methods In Shallow Seismic Reflection

c3 JD <L) o
00 cd
O h
Q 3 C /3 c s C
/3
Z Z T Harith
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

A p p r o a c h -II: re su lts a n d c o m m e n ts

Four examples of the vertical sections resulting from this approach are presented as two
separate figures. Figure 3.47a shows examples of two inline sections - inline 17 and 29,
while Figure 3.47b shows examples of two crossline sections - crossline 6 and 20.
These examples reveal two reflectors between 5 ms and 30 ms TWT. A hint of deeper
reflectors also appears in some inline sections.

Interpreted versions of the same four vertical sections are shown in Figure 3.48a and
3.48b. The first reflector is a very shallow strong trough and is picked as red circles.
This reflector is picked as the shallowest identifiable consistent event. Because the data
have a polarity corresponding to the SEG convention, the trough represents an increase
in impedance contrast, or the boundary between the upper low velocity layer and the
higher velocity layer below. This reflectors may represent the clay/bedrock interface.
Figure 3.49 shows the migrated time structure map of this red horizon. Not surprisingly,
it shows similarities to the map produced from the first approach (Figure 3.45).

The second identifiable consistent event is the strong peak just below the red picked
trough. This reflector is marked by yellow circles (Figure 3.48a, Figure 3.48b). The
black peak represents a negative impedance contrast, or in other words the upper
medium has a higher acoustic impedance than the lower. A condition where such a
reversed polarity event would occur is at the top of a water- or air-filled underground
void. Thus the yellow reflector might represent in some form the old mine workings.

The overall structure of the mapped yellow horizon (Figure 3.50) shows a general
deepening to the top right hand comer of the 3-D area, cut across by possible fault zones
(or trough-like features) trending WNW to ESE. This strike is same as the strike of the
Carboniferous bedrock below the overburden boulder clay. The yellow event reaches a
maximum reflection time of 25 ms at the top right hand comer of the area, which
corresponds to about 16 -17 m depth.

University Of Glasgow 1998 PhD Dissertation Page 185

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

University Of Glasgow 1998 PhD Dissertation Page 186

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

-Aaa

University Of Glasgow 1998 PhD Dissertation Page 187

(stu) i M l
University Of Glasgow 1998 PhD Dissertation Page 189
N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.49 Bedrock or base clay (red horizon) map - approach-II.

si
50

M ia

{TiiJB M i

2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32
Contour interval l ms TW T. Dark blue - 8 ms, dark orange - 20 ms, grey - no picks.

University ( )f Glasgow 1998 PhD Dissertation Page 190

N ew M ethods In Shallow S eism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.50 Yellow horizon (possibly underground voids) - approach-IT.

hi i' -

■■.i

Contour interval 1 ms TW T. Dark blue - 9 ms, dark orange - 22 ms, grey - no picks.

University O f Glasgow 1998 PhD D issertation Page 191

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Although the overall trend of the yellow horizon follows the general direction of the
strike in the Carboniferous, there remain questions to be answered. For example, there
is no independent evidence for fault-like structures. Furthermore, there is still much of
area where picks cannot be made. These problems lead to reprocessing the data using a
third approach.

P r o c e s s in g - a p p r o a c h -III (r e p r o c e s s in g o f a p p r o a c h -II)

In a 2-D survey, the spread is normally laid parallel to the perceived general dip of the
layers to obtain a true dip cross-section. The same principle can be applied to designing
a 3-D survey. To get the best image of the subsurface, the inlines and crosslines should
be chosen to be parallel to either the strike or dip direction of the layers.

In the first and second approaches, the data were binned such that the inlines are parallel
to the rows and crosslines are parallel to the columns of the surface peg geometry.
However, the directions of these rows and columns are somewhat oblique to the general
dip and strike direction of the layers, but this binning strategy was chosen to avoid lots
of empty bins. Results from the second approach (yellow horizon), show an interesting
pattern, where the contour lines are elongated in the NW-SE direction, parallel to the
general strike of the layers within the Carboniferous. The results might be improved if
the data are re-binned with respect to the geological dip and strike. Such re-binning is a
freedom available from the randomised CMP distribution shown in Figure 3.40, but
would not have been possible if the sources had been arranged on a regular grid like the
receivers.

This third approach is a modification of the second approach. Using the same data (100-
700 Hz Lab Correlated), the data were re-binned, based on the results of the second
approach. The bin grid was rotated clockwise until the crosslines were parallel to the
strike.

University Of Glasgow 1998 PhD Dissertation Page 192

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Compared to the second approach in which we have 53 inlines and 33 crosslines (or
1749 bins), the rotated binning grid changes the number of inlines and crosslines to 50
and 57 (or 2850 bins), respectively (Figure 3.51a). Obviously this rotation also increases
the number of empty bins. Figure 3.51b shows the relation of the inline and crossline
system with respect to the National Grid coordinate system.

The data (100-700 Hz rotated bin) were processed in the same style as the second
approach. The elevation static of 350 m s'1 replacement velocity with 37.0 m datum was
also applied to the data to correct for elevation differences. No change of trace polarity
was made. The main difference between this (third) and the previous (second) approach
was in the way that the data were stacked. Unlike in previous two approaches, in this
approach we have decided to apply both top and bottom mute (Figure 3.52). This
hopefully will ensure that only the reflection signals will be stacked. Instead of stacking
using a 1-D velocity function, as in the second approach, these data were stacked using
a 3-D velocity function.

The reason of using the 3-D velocity function instead of 1-D velocity function is
because of the presence of lateral velocity variations. By using this velocity function, it
is hoped that the data would be stacked with the appropriate stacking velocity, thus
giving a better image of the subsurface. Finally, the data were not migrated, as in the
second approach. A summary of the processing flow for this approach is given in Table
3.19, while Table 3.20 is a summary of all processing flows for the three different
approaches.

University Of Glasgow 1998 PhD Dissertation Page 193

 New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Figure 3.51a CMP coverage with rotated 2 m binning grid - approach-III.

70520.000-E

70510.0004

70500.0004

70490.0004

70480.0004

70470.0004

70460.0004

70450.0004

70440.0004

70430.0004

70420.0004

70410.0004

70400.0004

70390.0004

70380.0004

r^ - - T r r~ I — i r ~r ~r T
55480.000 55500.000 55520.000 55540.000 55560.000 55580.000 55600.000
X

University Of Glasgow 1998 PhD Dissertation Page 194

N ew M ethods In Shallow Seism ic R eflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.M b Relation between inline-erossline and National Grid (E-N)

coordinates for approach-111.

554 4 5 5 5 4 96 55546 55597 55648

■70592

70529
5 0 , 57

1,57

■70467

70404

■70341
Scale 1 2000.00
0 100
i i Meters

University O f G lasgow 1998 PhD D issertation Page 195

 o c

Page 196
£ .1
£r - t>
<u (sui) auiix
<u
C l Q£
Jc .y
u E 1 /3
"5 X
C/3 ocd
£
o
oLh
a.
C l,
x cd
C/3

73
• Si
"a,
a.
cd
-LJ
l)
3
£
S
o
-L->
-<-»
o
X

a
o
uS-
1)
>
a>

1998 PhD Dissertation

C/3
00
c
o
X

X
bO
cd
-L-*
cd
73
73
<L>
C/3
C/3
<D
o
0Lh
p,
1
o
73
§
eo
£T
<L> 03
«s
u
oe^ cd
£
04
v/3
C
* (N
_o cn
"«3
University Of Glasgow

xC/3 a>
Vh
3 in mi ii|ii mi ni|i mu ui| inn mi |im iiiiiini mu ii
• £P
9823^8
o.-5 to
4* Xc3
«
, N (sui) aui|i
£ N
 on

£r - ou
< 43
cG
dO
Q. 0443 a.
x .2
<-> C/1
E
‘5
C /3
£
0
"rt
X
C /3
Q
1
m

C/3

£
o
*/"3
m
H
£ .td
o
o o
to 13
bO >
.S
C/3 "S
C/3 43
0O3 603
o
O
cd

X p4
o
03
O Q
Vh
CU Q
£
£ oo
03
O n

<
03
i H
rn
03

I C
b0
H .s
*& &
.3
00
3
C/3
Id
t-i
lj»-l
-(-* -t-*
o C3
-*-» 03 03
O to
43
E43C C/D C
/3
04
o

£
o
00
O
C+-1
o

43
>

Z N s
 Chapter Three
3-D Shallow Seismic Reflection

Page 198
fc
<D
o c<Do
U o
„ cd
<D
*h ^ to J_,
p <D -*-»
53 J 3 O +->
cd <d
N Vh ,co N

TP3 43
o
X ^ fc a o ffi
cd o O
o o
cd O
t-i r- Icn oa
CO PJ o■ I
o ^ CN • P Q ox
O o
o o
< (N O O < m 'O £

£
o
O
£
o •S
cd
CO
53 CD CD
►—] O

1998 PhD Dissertation

bD cd
a N Vh N
• i-H ffi +->
CO 43 K
CO o O o
<u cd O p
o
o O
O Vh r-■ o "d"■
P, CM a t n CN
P o cd o> o
o
< »—i CN oi < cd o

-a
43
o O 43
cd m O
-«-»
oS-4 m a>
a, t-i
cd <D .« 53 N
43
o 53 O ffi
o #o O
cd ’p •
O > o
cd £ oun
»-i i <D
cM c/3
o Pi
aT o p
New Methods In Shallow Seismic Reflection

(N P o o
cn CN CN P4 & oS CN Z

3 bO
cd _o P
H
cd 44
O ’I
c/3 cd 43
J-4
<D a +->
(Z) C /3 a
’pi-t o
-*->
.2 bfi
od P -*-> od
O
University Of Glasgow

cd cd > o bb
od Pi
CD
c3
P Q w z C/3
Z Z T Harith
New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

A p p r o a c h -I ll: r e su lts a n d c o m m e n ts

As in Approach-II, examples of vertical sections resulting from this approach are also
presented as two separate figures. Figure 3.53a shows examples of two inline sections -
inlines 28 and 37, while Figure 3.53b shows examples of two crosslines - 19 and 31.
All these examples also show the presence of two reflectors, but they are slightly deeper
(between 12 and 35 ms TWT). The first reflector is a horizontal strong trough around
16 - 20 ms TWT and is marked by red circles. The second reflector is a strong peak
dipping towards a higher value of crossline number (i.e. in a NE direction). This
reflector is marked by yellow diamond marker. Interpreted versions of the same four
vertical sections are shown in Figure 3.54a and Figure 3.54b.

The data were interpreted in a same way as in second approach. Because the polarity of
these data is not reversed, according to the SEG convention the trough is represents an
increase in impedance contrast. This means that the upper layer has a lower velocity
than the deeper layer, assuming that density changes are less important than velocity
changes. As the area has a relatively simple geology, this condition occurs at the
boundary between overburden and bedrock. So it is assumed that this red reflector
represents the clay/bedrock interface. From the geological map cross section (Figure
3.3), this interface is drawn as a flat glacial erosion surface.

The second reflector, marked by the yellow marker, is a peak. This means that this
reflector has a negative impedance contrast. In most common conditions the velocity
increases with increasing depth, but here the situation is reversed. The negative
impedance contrast is a result of a higher velocity layer sitting on top of a relatively low
velocity layer. Since this site is over an abandoned coal mine, a condition where such a
reversed polarity event would occur is at the top of a water- or air-filled underground
void. Besides that, it also can occur at the top of a thick coal seam as coal has lower
velocity compared to sandstone or shale (Carboniferous rock). Thus, it can be inferred
that the yellow reflector might represents, in some form, the top of the old minework or
a thick coal seam.

University Of Glasgow 1998 PhD Dissertation Page 199

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

University Of Glasgow 1998 PhD Dissertation Page 200

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

University Of Glasgow __________________________________________ 1998 PhD Dissertation Page 201

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

Only those two reflectors are picked as our aim is to image the underground workings.
No attempts were made to pick the topsoil/clay boundary, as this reflector is too
shallow. Furthermore the reflected signal from this boundary is masked by the first
arrivals. The other reason of limiting the picking is that no direct well tie could be
made. The picking was based only on the continuity of the signals.

The picking of the red reflector is relatively easy, as it can be picked almost on all
vertical sections. In contrast, the picking of yellow reflector is not as easy. To see the
behaviour of these reflectors and guide the picking, the 3-D data volume was sliced at
different times. Figure 3.55a to Figure 3.55f are examples of six time slices at 17.5,
20.0, 24.0, 25.5, 30 and 35.0 ms TWT of the data processed under the third approach.
These time slices show the distribution of positive and negative amplitudes at a constant
time. The positive amplitudes or peaks are coloured as red, while the negative
amplitudes or troughs are coloured as blue. Picks are also displayed on these time
slices.

The time slice at 17.5 ms (Figure 3.55a) shows an irregular area of blue (as marked by
the red pick marker) over most of the section. This feature becomes clearer in on the 20
ms slice (Figure 3.55b), however, it is absent from the deeper slices (Figure 3.55c, d, e
and f). This feature corresponds to the trough picks (red marker) on the vertical section.
The distribution of these picks indicates that this reflector is almost flat/horizontal.
Figure 3.56 shows the plan view (structure map) of the red horizon. The red horizon is
picked with reference to the trough, thus it is represents the boundary of glacial clay and
Carboniferous rock (bedrock). The time (TWT) variation ranges from approximately 10
ms to 24 ms TWT, which corresponds to approximately 3 m to 7.2 m deep (assuming
the velocity for the layer above this reflector is 600 ms"1- glacial till). The map of this
reflector shows an alternating curved region of blue (shallow) and red (deep), with the
overall trend dipping towards the N-NW. This shows that the bedrock reflector is not
totally flat but it is rather an undulating surface. This curved feature however is not
observed in approach-II processing result (Figure 3.49).

University Of Glasgow 1998 PhD Dissertation Page 204

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow S eism ic Reflection

Figure 3.55a Time slice at 17.5 ms TWT.

T-l
S cale - 1 : 979.68
0 10 20 30 40 50 60 70 80 % e te rs

U niversity O f G lasgow 1998 PhD D issertation Page 205

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.55b Time slice at 20.0 ms TWT.

4.0

2 . 6

-1.3

—0.1

•1.5

S cale - 1 : 979.68
0 10 20 30 40 50 60 70 80 9 ^ e te rs

University O f G lasgow 1998 PhD D issertation Page 206

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-L) Shallow Seism ic Reflection

Figure 3.55e Time slice at 24.0 ms TWT.

^ M e te r s

University O f G lasgow 1998 PhD D issertation Page 207

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.55d Time slice at 25.5 ms TWT.

University O f Cdasgow 1998 PhD D issertation Page 208

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.55c Time slice at 30 ms TWT.

U niversity O f G lasgow 1998 PhD D issertation Page 209

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.55f Time slice at 35.0 ms TWT.

r c ift

S cale - 1 : 97 9 . 6 8
0 10 20 30 40 50 60 70 80 9 ^ e te rs

University O f Glasgow 1998 PhD Dissertation Page 210

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.56 Plan view o f the red (clay/bedrock) reflector.

University O f G lasgow 1998 PhD Dissertation Page 211

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

The main target of this work is to map out the reflector from a shallow mineworks such
as tunnels or thick coal seams. This target is deeper than the bedrock reflector and can
be recognised by a strong continuous peak, representing the expected negative
impedance contrast. Vertical sections of the data (Figures 3.53 and 3.54) show that it
appears at around 22 to 35 ms TWT, thus analysing the time slices of the 3-D data
volume around these time should show interesting features.

The slice at 24.0 ms TWT (Figure 3.55c) shows a strong red coloured line parallel to
inline No 23. This feature corresponds to a peak, and is marked with the yellow pick
marker. As the bins have been rotated so that the crosslines are aligned parallel to the
strike, this line is parallel to strike. This is an indication of either air- or water-filled
mineworking or a thick coal seam, as both of these can produce a strong peak parallel to
the strike direction. The next deeper slices (25.5 ms (Figure 3.55d) and 30 ms (Figure
3.55e)) show the band moving towards the NE of the area, which means that it is
dipping towards that direction. Tracing this reflector back to shallower slices (e.g. the
time slice at 17.5 ms TWT) shows that this reflector is absent, thus it can be assumed
that this reflector originates somewhere within the range of 17.5 ms to 20.0 ms TWT.

Unlike the red horizon, the yellow reflector shows a NE dipping structure (Figure 3.57),
and it is much deeper (approximately 22 - 40 ms TWT). This is in concordance to what
is being shown in the time slices, i.e. it starts to appear clearly in or around inline 23 and
dips to the NE. As this reflector is truncated by the red reflector around inlines 20 to 23
(time slices 17.5 and 20.0 ms TWT), the picking of this reflector from inline 1 to 20
cannot be done. This is why the structure map (Figure 3.57) is blank SW of inline 22.

Time slices at 35.0 ms TWT (Figure 3.55f) and deeper show a mixed of blue (trough)
and red (peak) without any structural pattern. Nothing can be deduced from these
slices except that it is all noise. The vertical sections also confirm that no reflectors can
be picked at these deeper times.

University Of Glasgow 1998 PhD Dissertation Page 212

N ew M ethods In Shallow Seism ic Reflection C hapter Three
Z Z T Harith 3-D Shallow Seism ic Reflection

Figure 3.57 Plan view o f the yellow (? mineworkings or thick coals) reflector.

U niversity O f G lasgow 1998 PhD D issertation Page 213

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

3.11 Discussion and conclusion

The use of seismic investigation in shallow engineering surveys normally involves

finding or calculating the depth of the bedrock. For this purpose, the refraction
technique is the primary choice, as it is simple and easy to be conducted. In the area
where the geological conditions are simple such as an overburden overlying flat
bedrock, this technique (with some borehole information) may be enough in answering
the depth of the bedrock problem. However today, the aims of shallow seismic
investigation diverge into wider areas such as investigating movement of shallow
contaminants (aquifer study), imaging dangerous underground tunnels or voids, or even
in searching for underground water reservoirs. The seismic refraction method is thus
becoming more and more inadequate as it has several shortcomings, such as the hidden
layer problem.

As the refraction technique becomes unsuitable for these new tasks, shallow seismic
investigators are now using the reflection technique in much of their work. Seismic
reflection can be used to image the subsurface more accurately if the acoustic
impedance contrast is big enough to be detected. To date, seismic reflection, especially
the 2-D technique, has widely been used for shallow seismic investigation either to
accurately map the bedrock or to study shallow water reservoirs. The reasons are
mainly that it is simple to conduct, and involves little computational time. As the
technology becomes increasingly more advanced, the previously exclusive techniques
(used in petroleum exploration) are now also available for shallow seismic investigation
usage. These include the 3-D seismic reflection method.

This study is involved with transferring one previously exclusive petroleum exploration
technique - 3-D seismic reflection - to shallow seismic engineering investigation work.
This work is one of the pioneering works in this field, thus descriptions on the
acquisition processing techniques have been discussed in detail.

University Of Glasgow 1998 PhD Dissertation Page 214

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

The aim specific of this work is a case study or trial to image shallow underground
mining tunnels, which have been abandoned for decades. These areas are considered
dangerous and not suitable for development. However, because of the increasing
demand for land for either housing or industrial estates, these previously dangerous
areas are now considered to have potential for development. Before they can be
developed, the unstable ground needs to be treated, normally by grouting. For this
remediation work to be economical, an estimate of how much grouting material is
needed to stabilise the area is important, as this is the most expensive cost involved. To
estimate the volume of grouting material needed, a reliable method for imaging the
unstable ground is important. The 3-D seismic reflection method may be one of the
solutions.

With a grant of some £13K, the 3-D survey was planned over a 6000 m2 area that is
known to have been extensively mined. This amount is small compared to the cost of
closely spaced boreholes (a more traditional engineering method of continuously
imaging the subsurface) survey, since the average drilling cost of each borehole - e.g.
rotary coring in rock - is £40 / m with an additional £50 per hole and £400 to supply the
rig to the site (Waltham 1994). For example, the total cost to conduct a grid of 50 m
deep boreholes (using a rotary coring method) at a regular 10 m spacing to cover an area
of 6000 m2 is about £120K.

The 3-D data were processed using three different approaches. The results of the
second approach indicate the presence of two possible reflectors that have been marked
as red and yellow reflectors. The red reflector is interpreted to be the clay/bedrock
interface, while the yellow reflector is within the bedrock. The yellow reflector shows a
strong reverse polarity characteristic which is interpreted as the reflection from a
mineworking or a thick coal seam. The reason for this interpretation is the reversed
polarity, which could only happen in the presence of a negative impedance contrast.
This condition (in the site) could occur at the upper boundary between solid rock and an
underground void, or at the upper boundary between the solid rock such as sandstone
and a thick coal seam. No boreholes were drilled to confirm these pickings.

University Of Glasgow 1998 PhD Dissertation Page 215

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

An attempt to improve the result leads to the third processing approach. In this
approach the data were re-binned so that the inline/crossline direction are parallel to the
dip/strike directions respectively. The randomness of the CMPs and their azimuths
allows this re-binning to be done. This is the advantage of the data collected using this
kind of survey. The disadvantage, on the other hand, is that it produced lots of zero
trace bin, which increases the number of dead traces in the vertical sections. The
processing sequences were kept same as the second processing approach with a
difference in the way the data were stacked. In this approach the data were stacked
using 3-D velocity function. The result of this approach shows an improvement in
quality. As in the second approach two reflectors are recognised and picked. The
consistency in picking the reflectors become easier as the reflectors are clearly
recognised either on the vertical section or the time slices.

The structure maps of these two reflectors show interesting information. The red
reflector (clay/bedrock boundary) is undulating with two deep areas. The overall
pattern is dipping in a N or NW direction. More interesting is the structure map of the
yellow horizon (? tunnel or thick coal seams). The contours strike in the NW-SE
direction, which is the same direction of the solid geology. The horizon dips NE.

Although this 3-D seismic reflection survey failed to produce a detailed picture of the
target, it did however manage to image the presence of the general pattern of possible
underground voids or thick coal seams. Information such as the depth of the
underground minework is important for remediation work, because it can limit the area
to be grouted, as only a shallow unstable area needs to be grouted. Thus the volume of
grouting material can be estimated by assuming the average height of the minework.

The results of this 3-D seismic reflection survey show that this method is also feasible
for a shallow engineering site survey. The cost and time consumed to conduct this
survey also proved that it is very cost-effective.

University Of Glasgow 1998 PhD Dissertation Page 216

New Methods In Shallow Seismic Reflection Chapter Three
Z Z T Harith 3-D Shallow Seismic Reflection

The delay in producing any result from this trial survey was due to the fact that lots of
time was spent to find the best way to process the very shallow seismic data. Clearly
the processing time would have to be reduced to a month or less for the method to
become acceptable and useful to engineering contractors. Time was also spent on
introducing a new mapping method as a by-product of the 3-D seismic reflection,
named as ‘surface tomography’. This new mapping technique is discussed in Chapter
Four.

University Of Glasgow 1998 PhD Dissertation Page 217

 CHAPTER FOUR

SURFACE TOMOGRAPHY
New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

4.1 Introduction

Tomography by definition means a picture of a cross section of an object. Like a loaf of

slice bread, stacks of neighbouring cross sections can give a 3-D picture of an object. A
computerised axial tomography or CAT scan is one of the famous examples of this
concept. A CAT scan is used to define normal or abnormal structure in the body as well
as assisting in procedure by helping to guide accurately the placement of instruments or
treatments. In a CAT scan, a large doughnut-shape X-ray machine takes X-ray images
(closely spaced X-ray slices or cross-sections) at many different angles around the body
over a full range of penetration angles in a controlled manner. The rays pass through
different materials of the body such as bones and soft tissue. Different body materials
show different X-ray attenuation characteristics. The loss of intensity due to the unequal
absorption by different parts of the body is measured and reconstructed to give the
picture of the part that has been scanned or examined. In order to reconstruct the
intensity of the X-ray recorded, four major reconstruction algorithms have been
introduced which are: (1) summation methods, (2) convolution methods, (3) Fourier
methods and (4) summation-expansion methods (Peterson et al. 1985).

Deep-earth geophysicists, in studying the internal structure of the earth also use this
technique. It was first introduced into this field by Aki et al. (1977) and known as
seismic tomography. Seismic tomography uses the same principles as in a CAT scan but
rather than using the attenuation characteristic, the travel-times of the signals are
observed (Lowrie 1997). In most cases, the recorded arrival time (travelled from a
source location - normally an earthquake epicentre to an observation centre around the
globe) is different from the theoretical times calculated based on a homogeneous
mathematical earth model. Seismic tomography is concerned with studying this
difference. The main reason for this difference is that the waves travel through
inhomogeneous earth materials. Using a modelling technique - adjusting the velocity of
the model until it give an observed travel time - a velocity model of the earth’s interior
can be obtained. Therefore seismic tomography may be described as the three

University Of Glasgow 1998 PhD Dissertation Page 218

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

dimensional modelling of the velocity distribution of seismic waves in the Earth

(Lowrie 1997).

The success of seismic tomography in mapping the velocity structure of the Earth leads
to smaller scale work which is known as refraction tomography (Lanz et al 1998). Like
seismic tomography, refraction tomography also involves analysing the arrival times of
the rays travelling from sources to receivers. However instead of utilising the seismic
energy from natural sources such as an earthquake, refraction tomography utilises
controlled seismic sources, and the receivers can be arranged over a small area of
interest.

Crosshole tomography is another method where the tomography concept is used. It is

used primarily in the petroleum industry to describe the lateral velocity variation in the
reservoir description phase of exploration and production. Unlike both a CAT scan and
seismic tomography - where rays are detected by exterior detectors (receivers),
crosshole tomography studies need two neighbouring boreholes, one for the sources and
the other for the receivers. The sources and receivers are placed at various depths and as
in seismic tomography, the arrival times are observed. The arrival times from each
source-receiver pair are then reconstructed to give a picture in terms of velocity
distribution of the reservoir. Three forms of reconstruction methods are normally used
which are: (1) algebraic reconstruction technique (ART), which was developed from the
summation-expansion technique used in medicine, (2) simultaneous reconstruction
technique (SIRT) and (3) Gauss-Seidel method methods (Sheriff 1990). Details of the
mathematics involved in seismic tomography can be found in many tomography
textbooks.

Although CAT-scanning, seismic tomography and cross-hole tomography each study a

different characteristic of the rays, they share one thing in common, which is that they
need to have lots of rays travelling through the medium. The rays radiating from the
source in a fan-like path are similar to a simple fan-shooting survey used in detecting or

University Of Glasgow 1998 PhD Dissertation Page 219

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

delineating localised anomalous zone such as salt domes and mine shafts (Kearey and
Vine 1990).

In this chapter, we are going to introduce a new and simple surveying method
employing the tomography technique, named herein as surface tomography. Surface
tomography involves analysis of the arrival times from lots of ray paths (using the
tomography concept) that travel by refraction along a very shallow interface, but
without applying any reconstruction technique. The data used in this study are the same
as the data used in the 3-D seismic reflection studies (Chapter 3), however instead of
whole records, only the first arrivals are used.

4.2 The concept

All the tomography techniques, regardless their field of application, share the same
basic idea, where the properties of rays (travelling through different media) are studied
to give a picture of the target object. Surface tomography is, in effect, a development of
the old concept of the fan-shooting refraction method.

In fan-shooting refraction methods, receivers are located such that raypaths form fan­
like arrays radiating from the shot location (Figure 4.1). Normally an approximate
location of the target is known prior to the survey, and the survey is designed so that the
target can be placed in the middle of the fan. The source-receiver offsets are
approximately constant, except for a normal or control linear profile. The arrival time
for each receiver is then compared to the control profile (a profile taken from the area
where no anomalies are present). Whenever the ray travels through an anomalous
velocity zone, it will be subjected to a time lag or time lead depending upon whether the
zone has a lower or higher velocity anomaly respectively, compared to the surrounding
velocity.

University Of Glasgow 1998 PhD Dissertation Page 220

Uh ct
bfc; 0g>
■
« IO
ci
6 H
8
<3
!3
C/3

kb
Io
-t->
<
0u
0.Cd-H
1 M—
( Vh ,
<+ 2^ J7,
Oh
oH
043>
T3
C
0/33
cd

Od
• i*<
+->
ocd
o3
*
CD
V-h
00
.S 8<B
o
JB
C/3
c
cd
|Jh tM
OS w'?
ri W

^3"
<U
op
£
New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

Surface tomography also shares the same idea - whenever the ray path encounters an
anomalous zone, it will be subject to a time lead or time lag. However, the approximate
location of the anomalous zone is not known. Furthermore, source-receiver offsets are
not constant, and no normal (control) profile is needed. The main characteristic of
surface tomography is that it needs to have lots of random raypaths. Details of the travel
time differences due to the anomalous velocity which form the basic idea of this
technique are explained in Appendix 1.

Unlike the fan shooting method, in surface tomography the arrival time for each source-
receiver pairs are not compared to their corresponding control arrival time (to determine
the time lag or lead); instead all the arrival times for each source-receiver pairs are
‘reduced’ to be relative to one velocity value which represent the datum or reference
velocity. The difference between the observed arrival time and the arrival time as if
only one material had been present gives the total delay time (Sheriff 1973), which can
be higher or lower than the datum reduced time.

The reduced travel time (graphical detail will be explained in section 4.5.3) is a function
of velocity and of the structure (velocity, thickness) of the overlying layers. Consider
these three two-layer models situations:
1) The control situation where the velocity of the first and second layers is
600 m s'1 and 2400 m s'1 respectively, and the thickness of the first layer is 2
m. The arrival time at 10 m distance is 10 ms. The reduced time for this point
at a 2400 m s'1 reduction velocity is 6 ms.
2) The first and second layer velocities are 600 m s'1 and 1200 m s'1respectively,
with the thickness of the first layer being 2 m. The arrival time at 10 m
distance is 14 ms. The reduced time for this point at 2400 m s'1 reduction
velocity is 9 ms, or 3 ms slower than the control time.
3) The first and second layer velocities are 600 m s'1 and 2400 m s'1 respectively,
with the thickness of the first layer being 4 m. The arrival time at 10 m
distance is 17 ms. The reduced time for this point at 2400 m s 1 reduction
velocity is 12 ms, or 6 ms slower than control.

University Of Glasgow 1998 PhD Dissertation Page 222

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

Thus, instead of being due to low velocity anomalies, a time lag could also be due to
thicker overlaying layers. This is a drawback of the technique.

This drawback is not crucial to the success of the method, as the bedrock interface (the
layer of interest) is assumed to be roughly horizontal. The assumption is based on the
general geological cross-section (Figure 3.3) as well as the information from boreholes
and the preliminary study (Section 3.6, Figure 3.6). The physical properties of the
overlying layer are consistent throughout the area, thus the velocity is also consistent.
The effect of the topography variation has been corrected for (by elevation and static
corrections) during the seismic data processing. So the time lag or lead is not
influenced too much by the structure (the thickness and velocity) of the overlying layer.

4.3 Previous work

As mentioned in Section 4.1, most tomography applications are in medical imaging

(CAT-scanning), reservoir characterisation (crosshole tomography) and in
understanding the internal structure of the earth (seismic tomography). Impressive
results from crosshole tomography studies in describing the petroleum reservoirs have
influenced many geophysicists to adopt the technique for shallow geophysical
investigations. Peterson et al (1985) and Cottin et al (1986) used the crosshole
tomography technique to image the area for a radioactive waste repository and
foundation quality, respectively. Lytle and Dines (1980) also used the crosshole
tomography technique in mapping buried voids, shaft and tunnels.

Kilty and Lange (1990) used the tomography technique to detect shallow ore mineral
zones. In contrast to conventional crosshole tomography, they used a surface to surface
technique where the sources and receivers were positioned along two parallel profiles
with the target approximately in the middle. First arrivals from the waves that travelled
through the bedrock (vertically bedded high velocity material covered by alluvium)

University Of Glasgow 1998 PhD Dissertation Page 223

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

were reconstructed using transform reconstruction methods to give an image of the ore
mineralization zone.

Lanz et al. (1998) used the tomography concept in mapping the depth and geometry of
landfill. They studied the first arrival times from several neighbouring 2-D refraction
profiles. The first arrival times were then reconstructed using a tomographic algorithm
that employs a fast fmite-difference eikonal solver for propagating wavefronts through
2-D heterogeneous media, and an inversion method that incorporates appropriate
damping and smoothing constraints (Lanz et al. 1998) to produce a tomogram for each
refraction line.

Although the surface tomography introduced here also employs a surface to surface
technique, it is different from the techniques of Kilty and Lange (1990), and Lanz et al.
(1998). Instead of placing the sources and receivers along parallel lines like them, the
sources and receivers are placed randomly. This allows the rays to travel a full range of
azimuths. This technique also examines the first breaks from the wave that travels
through the bedrock, however, instead of applying an algebraic or transform
reconstruction technique, the author uses a simple mathematical manipulation
performed using a spreadsheet. We compare the time taken to travel through different
materials with the time taken as if it travels through only one type of material. The time
lead and time lag are mapped instead of the velocity.

4.4 Aims

The site (refer to Chapter 3 for a detailed description) is situated on an abandoned

coalmine. It has a relatively simple geology comprising NE-dipping coal bearing
Carboniferous sandstones and shales overlain by Quaternary glacial deposits and
covered by a loamy top soil (Figure 4.2). In the study area, two recorded and one
previously unrecorded mineshaft have been found and treated (Photo 4.1). The size of
the shaft is approximately two meters square.

University Of Glasgow 1998 PhD Dissertation Page 224

 Page 225
£
.
8H
*

s
» §P
43
UH

Qfi ID
O
o
<u
on

V
JC
■ _Q '
op
a
o
4=
C/5
cd

1998 PhD Dissertation

a/

-a
3
<D
43
+->
C+-t
O
c
0
■4-*
CJ
1>
C /3
1
C/3
C/3
O
u
CJ
CJ
*-«
cd

v
JC
CJ
CO
_o
c
o<D
qC
a! (N
<u
t-.
<L> c
GO sij C/D
£
_o
University Of Glasgow

"rt
43
00

c/1
■8«
■§ k
£ N
Z N
 id ^

Page 2 2 6
3
o coa*
tx
x :

b
S- s

JO
cd

1>

b£
g
3
"O

D isser ta tio n
Cl
§
cd
CJ
TJ
T 3
O
H

1998 PhD
(D

<cd35
XC/5
<L>
C

TD
0/
eo
-a
c
cd
_D
cd
C
<

'd-
O
■H1
O
-C
Of G la s g o w

Cm
University
New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

The aim of this technique is to detect anomalous velocity zones. Mineshafts, which have
a very low velocity comparedto thesurrounding material, would be anideal target,
although it is also hoped thatthetechniquewillpickup geological effects such as the
variations in velocity of sandstone, shale and coal.

4.5 First breaks

4.5.1 Picking

Since surface tomography deals primarily with analysing first arrival times, accurate
picking of the first arrival times is very important. Inconsistency in picking will result in
huge errors. The dataset used for this study is same as the one used in the 3-D shallow
seismic reflection, and was gathered using a portable OYO minivibrator as a source and
30 Hz single geophones as receivers. Detailed descriptions of the acquisition methods
have been presented in Chapter 3.

At large offsets unwanted signals (noise) degrade the data quality. For the near offset
traces where the energy (of the signal) is stronger than the noise level, the data quality is
good. In order to minimise errors, picking of the first-arrivals has only been done on the
good data, where they are clearly identified. Since the long offset traces were
contaminated by unwanted noise, the first break picking was limited to offsets of less
than 10 m. As the aim of this technique is to pinpoint any anomalies of a size of
approximately 2x2 m2, such as a buried mine shaft, there are enough such near offset
traces to serve this purpose.

Prior to first break picking, the polarity of the signal needed to be confirmed. This was
done by comparing the vibroseis data to the data gathered using hammer and weight
drop (impulsive source) from the 2-D survey carried out on the same area. The first
trough was picked on all shot gathers and CDP gathers (Figure 4.3) when it is clearly
recognised. All the first arrivals were picked manually using ProMAX/3D and then
exported to a Microsoft Excel 97 spreadsheet for simple manipulation.

University Of Glasgow 1998 PhD Dissertation Page 227

New Methods In Shallow Seismic Reflection Chapter Four
ZZTHarith Surface Tomography

(s iu ) a u i;i
New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

4.5.2 Time distance curves

Picked first arrival times were then plotted against their offset (distance) to produce a
time-distance curve. All the time-distance pairs were initially plotted onto the same
graph (Figure 4.4). This is known as the all-points time-distance plot. The advantage of
this plot is it gives an idea (first impression) of the geological information such as the
layers present, their average velocity and their estimated thickness or depth. This
information is sufficient for surface tomography studies since there is no intention of
mapping the structure of the refracting interface.

The all-points time-distance plot shows that the variation of the first arrivals (Figure
4.4) increases with offset. At very near offsets it is small (±2 to 4 ms from the mean),
however the variation for the far offsets is big, between ±6 to 8 ms around the mean. As
the energy of the signal decreases with offset and the noise level increases relatively,
this increase in variation with offset is to be expected. One of the reasons for the
variation is the difficulty in recognising the first breaks on the far offset traces. This is
also the reason why the first break picking is restricted to the near offsets (less than 10
m) only.

Inhomogeneity of the medium (lateral variations in velocity) is, of course, another

explanation of this variation. Because the rays have to travel through different velocity
layers, the time-distance curve will show a scattered plot rather than a classical straight
line section.

What information can we gather from the all-points time-distance plot? Based on best-
fit lines drawn onto the travel-time plot, there are three different segments or layers
which can be recognised. The first and third segments are clearly seen on the curve. The
first is determined by drawing a least-squares best-fit line through the origin. It has a
slope of approximately 5.45x1 O'3 s/m. The third segment dominates most of the graph;
it has a least-squares best-fit slope of approximately 6.06x1 O'3 s/m, and crosses the time
axis at 10 ms.

University Of Glasgow 1998 PhD Dissertation Page 229

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface T o m o g ra p h y

(sup «wulL
University Of Glasgow 1998 PhD Dissertation Page 2 3 0
New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

The second segment is more subjective, however, based on the knowledge of the local
geology, a best-fit line with a slope of approximately 1.74x1 O'3 s/m can be drawn in
between the first and third segments. It intercepts the time axis at 4.9 ms. The crossover
distances between the three line segments thus obtained are at approximately 1.5 and 4
m, respectively.

In a normal refraction work, data from forward and reverse shots are observed for the
same line. The differences in intercept time indicate the structure of the interface. The
interface is flat if the intercept times are the same, otherwise it is dipping. The direction
of dip is towards the higher intercept time. However, in the case where only one
direction is shot, the interface must be assumed to be horizontal. A plot of all-pairs
travel-time graph can be considered as the case where only one direction is shot in a
normal refraction survey. Therefore, it has to be assumed that the interface is horizontal.

As the plot indicates a three-layer model, the thickness of the first and second layer can
be obtained. Equation 4.1 below needs information on the intercept time and the
velocities for all the layers in order to calculate the layer thicknesses. Assuming no dip,
the thickness of the first layer is 0.47 m and that of the second layer is 1.51 m.

Equation 4.1 Thickness of the Zi & Z2 for a horizontal interface.

tV V
Z, = ----- -' l ■ 2 , ■ where Z ,, Z 0 are the thicknesses of layers 1 and 2
W l - v ? ) 'n
VliV2, V3 are the velocities of layers 1, 2, and 3
t VV Z V (V 2 - V 2)112
Z 2 =-----2 2 3 1 2■ 3----1... t,, t2 are the intercept times for layers 2 and 3
2(F, - F 2 ) ^(K Z -K ,2)

University Of Glasgow 1998 PhD Dissertation Page 231

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

The geological map (Geological Survey of Great Britain, sheet 30, Figure 3.3) shows
this area is situated over low angle NE-dipping inter-laminated Carboniferous
sandstones and shales with seams of coal of various thickness. This Carboniferous rock
is overlain by Quaternary glacial deposits. The contact between these two groups is
drawn as a flat erosion surface. On top of the glacial deposits are thin black and loamy
materials known as topsoil.

Prior to the 3-D survey, a week of third year geophysical fieldwork (refraction and
resistivity survey) was conducted on the site under the supervision of the geophysics
lecturers and the author. The students were divided into four groups. All groups
gathered their data using the same inline spread geometry pattern - a line of 24
geophones with a 2 m interval. At least five shots were recorded for each group. The
shots were placed 2 and 5 m away from both ends and in the middle of the spread
(Figure 4.5). Group A and B aligned their spread parallel to the strike while the other
two aligned theirs perpendicular to the strike. The result of these refraction surveys is
shown in Table 4.1.

Figure 4.5 Third year refraction spread geometry.

1 12 13 24
1____ 1 1____ 1 1— 1 1------- 1
2m 2m 1m 2m
i--------------1
5m 3m

• Receiver at 2 m spacing
■ Source

University Of Glasgow 1998 PhD Dissertation Page 232

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

Table 4.1 Results of the refraction survey undertaken by the third year students.

Group Layers Velocities Thicknesses

A 2 440 m s'1; 2000 m s'1 4.5 m
B 2 425 m s'1; 2370 m s'1 2.8 m
C 3 441 m s'1; 1349 m s'1; 2778 m s'1 1.8; 8.9 m
D 2 370 m s'1; 2300 m s'1 2.3 m

A week before the 3-D reflection survey was done, the loamy topsoil was stripped from
the area where the injection grouting boreholes were to be drilled (Photo 4.2). Between
0.3 and 0.7 m of the topsoil were stripped off. The daily drilling logs from the grouting
operation were also available for reference. They show that the rockhead - the rock
underneath the glacial deposit - is approximately 3 m beneath the surface.

Based on direct evidence from the geological and drilling records, the site clearly
comprises three layers as graphically shown in Figure 4.2. If so, why do three of the
four student refraction surveys only detect two layers? To explain this contradiction, the
author re-analysed each of the student groups’ datasets and their results. One of the
reasons that can be deduced lies in the spread geometry.

The travel-time curve (Figure 4.4) indicates that the crossover distance between
segments one and two is approximately 1.5 to 2 m offset. The crossover distance is the
distance where the refracted wave has travelled through a higher velocity medium and
has arrived at the same time as the direct arrival, or the refracted arrival which has
travelled through a lower velocity interface. The crossover distance is different from the
critical distance (the distance at which the reflection time equals the refraction time).
The critical distance is less than crossover distance; however in designing a seismic
survey, the crossover distance is often taken as an approximation of the critical distance
(Milson 1989). The critical time is obtained by dividing the crossover distance by the
direct wave velocity (Milson 1989). This will avoid the detection of unimportant signals
such as the direct wave.

University Of Glasgow 1998 PhD Dissertation Page 233

 Page 2 3 4
£ g-
£ I
* 1°
-C I
UH <
OD
.ca
t+-i
i_
3
00

T<3D

a>
jo

o
JO
tic
.S
3
O
t-,
GO

D isserta tio n
a>
JO

ot/i

1998 PhD
a,
o
03
<D
Ou
ex

Kj
(U
—
<

(N
of-
O
+-*
o
JO
ex
Of G la s g o w
University
New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

Since the nearest offset receiver is at 1 m, and the receiver spacing is 2 m while the
approximate critical distance is 1.5 m, only one receiver can detect the direct arrival, an
arrival from the first layer (top soil). In an area with a lateral variation in velocity like
this area, drawing a best-fit line was not an easy job. The problem is, to which line
segment does this single point belong?

Without any prior knowledge of the layers and their velocity, there is a high possibility
that this single point was considered as a part of the second segment which means that
only two segments were recognised. Drawing the best-fit line through the origin and
honouring all the points will decrease the slope, which implies an increase in the
velocity. Depending on the slope of the best-fit line drawn in this way, the velocity
should give the value of approximately 183 m s'1 (first layer velocity), 580 m s'1
(second layer velocity) and 1680 m s'1 (third layer velocity) based on the time distance
curve of Figure 4.4.

An average velocity value for the sandstone-shale (typical rock type for the third layer)
is 1500 to 4000 m s'1 (Waltham 1994). Pennant Sandstone of Carboniferous age
however has a higher velocity value in the range of 4000 - 4500 m s'1 (Kearey and
Brooks 1991). However, because rockhead or bedrock is very shallow (approximately
3 m), it is subjected to weathering, which will reduce the velocity. The weathering
effect may be the reason why the velocity for the third layer (Carboniferous rock of
sandstone-shale type) is as low as 1600 m s'1.

University Of Glasgow 1998 PhD Dissertation Page 235

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

4.5.3 Reduced time distance curves

A reduced time distance curve is another way to display time-distance pairs. Reduced
time where the travel time corresponding to a single constant velocity is subtracted from
the recorded time (Equation 4.2). The constant velocity is known as the reduction
velocity. The reduced time plot is normally used to display large-scale refraction data,
which are concerned with late arriving phases such as wide-angle reflections or S-wave
arrivals (Kearey and Brooks 1991).

Equation 4.2 Reduced time equation.

^ X
T = t ----- T = Reduced time
Vr
t - Observed time
x = Offset
Vr = Reduction velocity

Six different reduction velocities - 400, 600, 800, 1400, 1600 and 1800 m s'1 were
chosen in order to estimate velocities for the second and third layers. The first three
were chosen to estimate the velocity for the second layer, and the other three were used
for estimating of the velocity of the third layer.

Figure 4.6a and 4.6b show all-points plots reduced by the six different reduction
velocities. The middle part of the plot in is flattened at a 600 m s'1 reduction velocity
(Figure 4.6a), while the end section (third segment) is flattened when 1600 m s'1 of
reduction velocity is applied (Figure 4.6b).

Inspection of the 600 m s’1reduced time plot also reveals evidence of the existence of
three layers, since there appear to be three different segments with positive, flat and
negative slopes, respectively. These estimates of velocity for the second and third layers,
of 600 and 1600 m s'1, respectively, are clearly very approximate, in view of the scatter
of points.

University Of Glasgow 1998 PhD Dissertation Page 236

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

Figure 4.6a Test of three different reduction velocities (400, 600, and 800 m s'1) in
order to estimate velocity of second layer.

|«flTVn»(v-*OQ>(

OtaiM (w»m|

|*ATiim<vw400) )■ *
--------- 1
♦

J • ♦ • •• • -r v
r

M
▼ • a 10
OtMmtiMl

University Of Glasgow 1998 PhD Dissertation Page 237

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

Figure 4.6b Test of three different reduction velocities (1400, 1600, and 1800 m s'1)
in order to estimate velocity of third layer.

! o t 5 T im * ( v » i4 o 0 ) I

| • & Tim* <v» 1600) j

♦

♦
♦1 iknikSS&2L
• •«

■ ■ ■ j
i

-----------------------*1

V A 2
•1 EMM

0 1 2
r ' | '• •
3
•

4
•

5 8 7 * 9 10

Olilimi

[ # A Tim* <v-l600) )
•

*
Sh H
k h |M
*

- k :“ i *
0 1 2 3 4 5 8 7 8 9 10
CW
w— 1

University Of Glasgow 1998 PhD Dissertation Page 238

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

4.6 Surface tomography

The surface tomography method is a simple and robust way of mapping the velocity of
a refractor, using the first arrivals from a 3-D survey. The arrival times from many ray
paths are displayed along a projection (map) of the respective ray paths along the
refractor. Each ray will show either a time lag or lead relative to a reference time,
depending on the medium within which it is travelling. These time lags or leads are
assigned an appropriate color scheme to make them easy to recognize. In calculating a
time lag or time lead, only three basic pieces of information are needed, which are: (1)
source and receiver location, (2) azimuth (which can also be calculated from source and
receiver location) and (3) first break time. Using these, the surface length of the raypath,
an average velocity and the direction of travel can be obtained.

As the dataset used in this study is the same as that used in the 3-D reflection survey, all
these data are already available in the database (the first breaks having already being
picked during the data processing phase). The surface length (the distance between
source and receiver) represents a projection of the propagation path from source to
receiver through all the media (topsoil, glacial till and Carboniferous). However, as the
main target (e.g. a hidden mineshaft) extends only up to the second interface (between
glacial till and Carboniferous bedrock), only the portion of the raypath travelling along
that interface is useful in detecting the anomaly related to the shaft. The projection of
the raypath through the first and the second layers should be edited out.

The length of the path along the second interface (refracting along the top of the third
layer) can be obtained by subtracting the critical distance from the surface length. A
crossover distance from the all-points time-distance curve gives a good approximation
to the critical distance. Re-calculating the travel path also means shortening the surface
length, however the direction is still the same. In order to draw an accurate third layer
path, the subsurface projected coordinate (vertically projected from the entry point to
the third layer) is calculated using Equation 4.3. Figure 4.7 shows a schematic relation
between the surface path and the surface projected third layer path.

University Of Glasgow 1998 PhD Dissertation Page 239

N ew M ethods In Shallow Seism ic Reflection C hapter Four
Z Z T Harith S urface T om ography

Equation 4.3 X & Y coordinate for the projected third layer entry point.

. x sin (a )
o „= S„ H------------------------------------ w h e r e X c = C ritical d ista n ce
2
a - A zim u th

x cosia)
Sy =S y H— -— ------ S X’ S y = ^ & Y c o o rd in a te for orig in a l poin t

S x , S y = X & Y c o o rd in a te for p ro jected p oin t

Figure 4.7 Schematic relation between surface path and the surface projected
path within the third layer.

N orth
S u rfa ce len g th

C ritical d ista n ce

A z im u th
S s’ Sr’

S u b su rface len g th

W h ere S = so u r c e
R = r ec eiv e r
S s = su b su r fa ce e q u iv a le n t o f S
Sr = su b su r fa ce e q u iv a le n t o f R
S s ’ = su rfa c e p rojected S s
S r ’ = su rfa c e p r o je cte d Sr

University O f G lasgow 1998 PhD Dissertation Page 240

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

4.7 The display

The best way to look at the surface tomography had to be found by producing different
kinds of maps such as vector and raster image maps. All these maps were produced on a
UNIX Sun workstation using the GMT (Generic Mapping Tools) programs. Because
the target is small (approximately two meters square) a useful map must be big enough
to see those anomalies. The decision has been made to produce maps with 1:50 scale,
*
of 24 cm
*
i.e. 1 m on the ground is equal to 2 cm on the map. Therefore, any anomalies
(2 by 2) on the map (1 m2 on the ground) should be recognised without any difficulties.

Three types of maps were produced. Each of them aims to tell a different story. All
these maps were plotted using a Hewlett Packard Design Jet 350C plotter. Because the
original maps were too big to be plotted as a single file, the size had to be reduced (for
presentation purposes) to an appropriate size prior to sending to the plotter. The scale
used for three main back-pocket maps is approximately 1:192, which means the original
size of the map is reduced by a factor of four (approximately).

4.7.1 Reduced time (vector) map

Figure 4.8 (Back pocket 1) shows a reduced time map for the whole dataset. All the rays
travelling through the second interface are represented by vectors. There is a small
arrowhead showing the travel direction, but this is only visible on larger-scale plots or
screen displays.

The aim of this map is to show the zones of time lead and time lag with respect to their
path so that anomalous zones can be identified. The data, which comprise the third
layer’s (X, Y) coordinates of the point Ss’ (Figure 4.7), the azimuth from source to
receiver, the path length and the reduced time, were sorted by their reduced time and
binned into several files.

University Of Glasgow 1998 PhD Dissertation Page 241

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

Each file was assigned a different colour. The datum or average time was assigned as
green. The time lag, or time bigger than the datum, was assigned blue increasing in
intensity towards higher times, whereas the time leads were assigned red increasing in
intensity towards lower values (Figure 4.9). This colour code was used for all surface
tomography maps.

4.7.2 Reduced time (vector) anisotropy map

The solid geology of this area is north-easterly dipping coal bearing sandstones and
shales of Carboniferous age, striking NW-SE. Because of the heterogeneous nature of
the stratigraphy, rays travelling along the strike direction are likely to have a different
velocity from those travelling across the strike. This is a form of anisotropy.

The reduced time (vector) anisotropy map displays this anisotropy. This reduced time
(vector) anisotropy map is produced in the same way as a reduced time (vector) map.
The individual files which had been previously sorted by reduced time were sorted once
again, this time by their azimuth. All raypaths with azimuths between 0° to 90° and 180°
to 270° (i.e. down- or up-dip) were stored into different files from the ones with
azimuths of 90° to 180° and 270° to 360° (i.e. along strike).

Figure 4.10a (Back pocket 2) is the resulting map where the azimuths lie between 0° to
90° and 180° to 270°, representing all the rays that travelled roughly perpendicular to
the strike. Figure 4.10b (Back pocket 3) represents all the rays travelling more or less
parallel to the strike.

University Of Glasgow 1998 PhD Dissertation Page 242

N ew M ethods In Shallow Seism ic Reflection C hapter Four
Z Z T Harith Surface Tom ography

Figure 4.9 Colour code as used in all reduced time maps.

T im e L ead
(5 m s , red )

D atu m
R e d u c tio n v e lo c ity = 1600 m s '1
( 10 m s, g r ee n )

T im e L a g
( 1 7 m s , b lu e )

U niversity O f G lasgow 1998 PhD D issertation Page 243

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

4.7.3 Reduced time (image) map

The reduced time (vector) map is produced by plotting lots of thin lines with an
appropriate colour representing each source-receiver vectors. Where there is lots of
vector covering the area (e.g. southern parts of the area, Figure 4.8), finding a small
clustering of different colour (small anomaly) is not a problem. On contrary, where the
number of vectors is not sufficient to cover the area, it create gaps (e.g. middle part of
the map, Figure 4.8), thus it is hard to recognise any small anomalies. This is the
drawback of the reduced time (vector) map.

The reduced time (image) map is produced to overcome this problem. Instead of using
lines, the area is coloured by different colours corresponding to the local reduced times.
Instead of lines, points were used. For each vector, coordinates for a series of points at 2
m intervals were calculated. The points were calculated using Equation 4.3, but instead
of using the critical distance, a 2 m crossover distance was used. Figure 4.1 la shows an
example of two lines (vectors) and the series of 2 m interval points along the lines. The
points were then gridded onto a 0.25 m equal grid with a 1.0 m search radius (Figure
4.1 lb), thus giving fine pixels for producing an image map. The reduced time values of
all the points inside the circle were averaged. Using the same colour code as before,
these new points were then used to produce a raster image map (Figure 4.12, Back
pocket 4).

4.8 Interpretation

How do we interpret these maps? As the map is produced using different colours to
show the time lag and the time lead, interpretation is qualitative. The differences in the
general pattern (of colour) and the concentration of blue (time lag) are the main
observations. The pattern difference will hopefully indicate the bedrock geology while a
small concentration of time lag colour (blue) may reveal the location of a low velocity
anomalous zone, such as a mineshaft.

University Of Glasgow 1998 PhD Dissertation Page 244

 Page 245
it
H h ctf
b g>
-4-J O
& a
ri O
□H
8
43
00

T3
a>
c3
3
jy
3
o
r~N
C/3
1>
C/5
VI
O
»-c
X<L!> ^
*“* C/5
w 4)

|S
'a s
« s,
a>

1998 PhD Dissertation

-*c-■ -*
o a3
■ S u-* O
es
E >f cr
O)
CN &P
'tJ
O t3 CN
(75 C cn
■g a
hs o
cd fc!
•s y
*I
Oh O

1 9
8 -3

cd

1<» T^
<L>
c* 00
o
E
V
C/3
*
_o
"c3
University Of Glasgow

JZ
C/3

"8«
f |
tC N<-
Z N
3!=( sc
^

Page 246
UOh CL
Id
eCcL gC
U ^
u Cl
o cd
,ctj

<D
CD
ccj
J
<D
JC
4—*
cu
o
3 I
"O
oc
a
o>
-i—
~o
<u

43
CD

1998 PhD Dissertation

T3
L^
O
C /2

><
'a ,
B jt—!---L L! "I- 4 i—
Uk
04 if; r 4 t > + 4 it4V \ '
T3
!
■w, k * + i ++ | 4 t ++|4: I i ! 4 -4>
V ' +! i+14 +i + 7+ V+ + 44 U*+-«. i+l L2S
C r—i t ♦*- 4 f- • *• "'V V 4?+
3 -V * r'± +I i +i + **i + <i + 4
C3 i- -t 4 k V
!v +1+" - ..14 yn o ...v + *
f. r/+ T
, t . * i* . 4
;i
144 - +
, 4 - i++
- i ..+ , I- tfk 4 - i. 4
& % ! +4 1+ i * £ . j - t - 1 +♦ Lin 4 - iI44 ++ _ k! ♦ i i !I _ _ ■ + t ' .1I i J+ ! iI
£ - k J k 4 Vk 4 ■+'%, 4 ^ ++^ I " 41
* * f c -+ * : ''4 *_ 4
; -V o 4! . .+ : V ±*J o ^ k t^ V f4 -
M i (**■! , I I !_ U+ " 4 + !-i i+ f !+ \ + ! t + ! "1
O Lk V r v r t-~ 4 4 i! k - + : ■**■■ v t
Cl
<u
? .t •,-•
• *. ! 1 !* * ; ,*< »■u*tff-<
A /r f j,i !k
*■*'k
- «'* k. w.C
,*
43 . +- +..4. k -r -p^ 'V 4i. 4 V V f~ + ......... 4 * ■; : it ^
*—•
§
oa>
<-t—.
O
i++ t+
k M* k. - 4 4,- ' ■..
44 1 +* l£ * +t+H- !' tVf - 4.1 V
4W
* -t -iV 4f 1+1 ;k
' +’44
v 4+k7 rl 4^V4
l i *t. ..v
+T <*. u'w:
i- • ■i + L 1 i4 !+
,i +f -' i yt + “'; { 4 ? :v 1
C/2
03
<L>
aS '53 .-4/4 '.i k ± *+; 4 t.,...*4 V . . ................ L 4*- • l. •

Q
w
. ^
i t+: 4 -4 .+rv +v +
+ ‘ '4 i -f4 ,’ +4 -+
'.4.v V4. ^ + j4 + I +4 f• + i ' 4 k it 4 + i|.+
„ 4 4 'kr V .f .i* ,4* A14 . + +>-,4
•4 j; +4i;t + ^ 1 I . ! 5
-L l!’ 4 : 14 ^ 4j 1:1: j + +j j ++ +-i-+ + 1 + -40
S ; !; ?, 4 4k t ; V hA +4I«t..-+ 7 4 +4 4+ 4..+ ' k+4‘• > '
i * l± •h T■ > j ' + ■++■
h*± ;V*f
3- ^ , k k r : ri V4k5 : f K k
cd * 1 ' t ’- k - V ’k k k k '* ;'.* . ,*v
University Of Glasgow

^r
£ cu v - ^ / - 4- ‘ o 4 ■ -7 4+; +'+4 v r - 4 -4
c 1—
3
op
! «
- «
|x
£
i
(-
C N
Z N
New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

4.8.1 Control

Because the interpretation of these maps is qualitative, the controls on the interpretation
are very important to make it meaningful. Our controls are based on all information
available such as: (1) the geological map (Figure 3.3), (2) the 1966 and 1981 boreholes
and mineral report, (3) the 1995 boreholes and geotechnical report, and (4) the plan
view of working coal seams as abandoned in 1897 (Figure 3.4). An attempt is also made
to correlate these maps with the result of the 3-D seismic reflection study (Figures 3.52
to 3.54).

The bedrock geology is known to be a low angle NE-dipping Carboniferous sandstone

and shale sequence with seams of coal of various thicknesses. The general strike is NW-
SE. Extensive drilling in 1981 identified several coals, as listed in Table 4.2. Based on
this, a map with conjectural coal outcrop was drawn (Figure 4.13). Among these, only
two coals were conjectured to crop out at the site - Twechar no. 1 and Twechar no. 2,
which have an average thickness of 0.38 m and 0.16 m, respectively (Table 4.2).

Borehole information from the 1995 drilling programme was not useful because most of
the drilling was terminated once bedrock was encountered. The only useful information
that can be gathered from these boreholes is the depth to bedrock and its type.

Figure 3.4 shows two superimposed plans of the worked coals as abandoned in 1897.
The blue outline is the plan of the Jubilee Coal working, while the green outline shows
the plan of the combined Meiklehill Main and Meiklehill Wee Coals (also known as
Possil Main Coal). The Jubilee coal was deeper (113 m) than the green outline
workings (41-46 m). The working of the green coals was more extensive, where the
working pattern started in the middle of the study area and extended in the NW-SE
(parallel to the general strike direction) and moved in the NE direction (down-dip
direction). The working pattern for blue (deeper) outline is more subtle where it
recorded two main working areas connected by a tunnel. Both blue and green workings
were terminated at the NW part with the same pattern (abruptly at a NE-SW line).

University Of Glasgow 1998 PhD Dissertation Page 247

N ew M ethods In Shallow Seism ic Reflection C hapter Four
Z Z T Harith Surface T om ography

Figure 4.13 Conjectural outcrop (red dashed lines) o f various coal seams in and
around the study area.

• 1981 \
B orehole \ 2
\
/T 9 8 1
• Borehole 11

1966
• 198\ Borehole 3
Borehole 7

1972
Borehole 1

1981
Borehole 14

1981
B orehole 5

1*881
/O ,4981 Borehole 16
B orehole 15

B orehole 2

University O f G lasgow 1998 PhD D issertation Page 248

N ew M ethods In Shallow Seism ic Reflection C hapter Four
Z Z T Harith Surface T om ography

The time slices o f the 3-D seismic reflection survey show a continuous peak (red)
parallel to the strike direction (see for example Figures 3.55c and 3.55d). This peak is
interpreted as an indication o f the tunnel or low velocity layer such as the coal layer
(refer to Chapter 3 for detailed discussion).

Table 4.2 The general sequence o f the underlying seams together with their
average thickness and their approxim ate depth (in m). Green and blue colours are the
worked coals whose plan views are shown in Figure 3.4.

Seam Average Approximate Note

Thickness (m) depth (m)
Twechar no. 1 Coal 0.38 0 D atum
Twechar no. 2 Coal 0.15 8
Kilsyth no 1 Blackband Coal 0.69 15 W orked
Kilsyth no 1 Blackband U nder Coal 0.46 21
Possil Wee Coal 0.61 38
Upper Possil or Meiklehill Main 0.92 41 W orked
Coal
Upper Possil Ironstone or Meiklehill 0.46 46 W orked
Wee Coal
Lower Wee Coal 0.69 53
Fourteen Inch Ironstone and Coal 0.15 59
Fourteen Inch Under Coal 0.46 66
Ashfield Cooking Coal 0.33 73
King Coal 0.46 79
Possil Main Coal 0.99 84 W orked
Cloven Coal 0.46 95
Shale Coal 0.92 108 W orked
Jubilee Coal 0.61 113 W orked
Knightswood Coal 0.76 122 W orked

University O f G lasgow 1998 PhD D issertation Page 249

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

4.9 Discussion

The discussion is based on (1) the reduced time (vector) maps, (2) the reduced time
(image) map and (3) the reduced time (vector) anisotropy maps with an assumption of a
horizontal interface and constant thickness. The interpretations of these maps were
controlled by all information available as mentioned in Section 4.8.1.

The reduced time maps, whether vector (Figure 4.8, Figure 4.10a, Figure 4.10b) or
image (Figure 4.12), show an interesting pattern of colours - alternating strips of blue
and green striking NW-SE. They also show two areas - one of orange and the other of
blue in the very south of the area. These are shown at a larger scale (1:128) in Figure
4.14a, back pocket 5. All these features are enhanced in the reduced time (vector)
anisotropy maps for vectors along strike. For example, Figure 4.14b, (Back pocket 6)
and Figure 4.14c (Back pocket 7) show enlargements of reduced time anisotropy maps
associated with the zones of orange and blue at the south end of the area, as shown in
Figure 4.14a. Figure 4.14b shows a map of the vectors propagating in the 90°-180°
direction while Figure 4.14c shows the more limited propagation direction. It is found
that, by limiting the vector to a certain propagation angle (Figure 4.14b, Figure 4.14c)
the separation between bands of colours becomes clearer. These features (alternating
strips) also show up on the map of bedrock resulting from the 3-D seismic reflection
survey, Figure 3.56). What do these patterns mean?

Assuming the bedrock is flat (as discussed in section 4.2), the time lag or lead is a
function of velocity alone, which means that a time lead corresponds to a higher
velocity layer, and vice versa. This assumption is based on the depth to the bedrock
information obtained from the boreholes (Figure 3.6) where it is consistently recorded
around 2.5 m to 5.0 m. Without any information on the geology of the bedrock, it can
be concluded that these patterns reflect the main bedrock geology (inter-laminated
sandstone and shale). This is based on the fact that P-waves to travel faster in sandstone
than in shale.

University Of Glasgow 1998 PhD Dissertation Page 250

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

The result of 3-D survey for bedrock (Figure 3.56), however, does not entirely support
the assumption (flat bedrock) where the presence of the undulating bedrock surface is
mapped. One of the possible explanations is that the bedrock detected by this survey is
most likely to be the boundary between the weathered and unweathered rock and not the
boundary between the glacial deposit and the weathered rock. This is based on the fact
that the impedance contrast between weathered and unweathered rock is larger than that
between the glacial deposit and weathered rock (bedrock surface). The boundary
between the weathered and unweathered rock is deeper than the boundary between the
glacial deposit and the weathered rock. Furthermore, the weathering effect on shale is
more pronounced than the effect on sandstone because clay (main component in shale)
is less resistant to weathering compared to silica (main mineral composite in sandstone),
thus producing an uneven weathered-unweathered boundary.

The red colour on the bedrock map (Figure 3.56) represents the deeper area while the
blue is shallower. In geological terms based on the argument above, the red is related to
shale, as shale shows less resistance to weathering compared to sandstone, and thus may
have a deeper weathered-unweathered boundary. On the other hand, the blue is related
to the sandstone. The depth to the weathered-unweathered boundary variation also could
be in a more resistant rock such as sandstone due to the presence of different amounts of
water, as water is a good weathering agent.

Comparing this map (Figure 3.56) with the reduced time (image) map (Figure 4.12),
there is a good correlation between the deeper area (red on Figure 3.56) with the blue or
time lag area (Figure 4.12). The shallow area (blue on Figure 3.56) in contrast
corresponds to the red (time lead) on Figure 4.12.

To test this idea, information on the type of bedrock (the rock type underneath the
glacial till) from the boreholes is analysed and is presented as Figure 4.15. Most of the
boreholes in the blue area showed that weathered sandstone is the bedrock except 1995-
3 and 1981-14 where shale is recorded.

University Of Glasgow 1998 PhD Dissertation Page 251

 N ew M ethods In Shallow Seism ic Reflection C hapter l our
Z Z T Harith S urface T om ographs

Figure 4.15 Bedrock types from 1966, 1981 and 1995 boreholes.

55450 55500 55550 55600 55650 55700

l I_____________i L _
70600 I

- 70550 1

- 70500

- 70450

1995 BH 5

- 70400

70350

70300

—I-- r 70250
55450 55500 55550 55600 55650 55700

Red and blue circles represent sandstone and shale respectively.

U niversity O f G lasgow 1998 PhD D issertation Page 252

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

This physical evidence contradicts the idea of blue corresponding to shale, thus an
alternative explanation is that those bands are related to more weathered sandstone
possibly caused by to the presence of more water. This idea agrees with the fact that the
northern part of the area is nearer to the River Kelvin (Figure 3.4). Although we know
that the P-wave velocity increases in a water saturated rock, it is still true that the
velocity of P-wave velocity is less in the saturated weathered rock compare to the
saturated unweathered rock (if the rock is same).

Can we correlate the surface tomography results with the 3-D seismic reflection results?
Time slices around 24 and 25 ms (Figure 3.55c, Figure 3.55d) of 3-D seismic reflection
data show a continuous peak (indicated by red marker) striking NW-SE and dipping
NE. Based on the geological map (Figure 3.3) the general strike is NW-SE and dipping
approximately 18° toward the NE. This general strike and dip was confirmed by
analysing the three point problem for several coal seams identified during the 1981
drilling programme (Appendix 2). The calculations of strike and dip are attached as
Appendix 3.

The 3-D seismic results show that the strong continuous red marker corresponds to the
general dip and strike of the area. As discussed in Chapter 3, this strong continuous
peak (reverse polarity) occurs at the boundary with a negative acoustic impedance
contrast, which means that the upper layer has a higher acoustic impedance value than
the lower layer. In the study area, a high negative impedance contrast can occur at the
top of a thick coal seam or at the top of water- or air-filled tunnels/mineworkings. The
general strike of the coal seams is parallel to the general strike of the bedrock. As the
red stripe (Figure 4.12) is parallel to the general geology strike direction, it also reflects
the geology/structure within the bedrock, thus this red stripe may indicate of the
presence of thick coals or tunnels.

The idea that the strike is parallel to the red stripe (seen on the reduced time maps) or
continuous peak (red) in the 3-D time slices contradicts the drawn outcrop strike of
Twechar no. 1 in which it was drawn oblique to this direction (Figure 4.13). The

University Of Glasgow 1998 PhD Dissertation Page 253

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

conjectural strike of Twechar no. 1 has been drawn in an open S-shape. This
conjectural outcrop is based on the 1981 drilling report in which it was only reported in
two boreholes, 1981-12 and -13.

As the regional dip is to the NE, any borehole to the NE of that conjectural outcrop
should therefore intersect the Twechar no. 1 coal. Borehole 1981-11 was drilled
between the conjectural outcrop and borehole 1981-12. In theory, it should record the
Twechar no. 1 coal at shallower depth compared to the depth in borehole 1981-12, but it
did not record any. The position of this conjectural outcrop is rather peculiar in being
placed SW of borehole 1981-11 (Figure 4.13).

As a check on this conjectural outcrop, we can use the depth at which it was recorded in
boreholes 1981-12 and -13, together with the regional dip (approximately 18° to the
NE). The calculation of these points is attached as Appendix 4. Twechar no. 1 should
then crop out 29 m and 16 m up-dip from boreholes 1981-12 and -13, respectively,
assuming a horizontal surface (Figure 4.16). The calculated outcrop location for
borehole 1981-13 lies on the conjectural outcrop. However for the borehole 1981-12, it
lies to the NNE of borehole 1981-11 (Figure 4.16). This explains why there was no
Twechar no. 1 recorded in borehole 1981-11.

The conjectural outcrop of Twechar no. 2 also was drawn to the SW of boreholes 1981-
11, -12, -13, -14 and -16 (Figure 4.16). Using the same arguments as above, these
boreholes should have recorded this coal unless the drilling depth was too shallow.
From the list (Appendix 2) Twechar no. 2 coal was indeed recorded in boreholes 1981-
11, -12 and -13, but was absent from boreholes 1981-14 and -16. As boreholes 1981-14
and -16 are nearer to the drawn conjectural outcrop compared to borehole 1981-13,
these boreholes should have recorded Twechar no. 2, but they did not. So what is wrong
with that drawn outcrop?

Calculating the outcrop location from their depths in boreholes 1981-11, -12 and -13,
and using the regional dip, Twechar no. 2 coal should outcrop at 53, 51 and 43 m up-dip

University Of Glasgow 1998 PhD Dissertation Page 254

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

from those boreholes, respectively (Figure 4.16). The calculated distance for borehole
1981-13 matches the location of the drawn outcrop, but the distances for the other two
boreholes do not fit at all. The calculated outcrop of Twechar no. 2 for borehole 1981-
12 lies to the NE of the conjectural outcrop while for borehole 1981-11 it lies to the SW
(Figure 4.16).

To check this mis-match between the re-calculated outcrop of Twechar no. 2 points
based on boreholes 1981-11 and 1981-12, and the drawn conjectural outcrop supplied
by the contractor, the dip and strike of Twechar no. 2 coal is re-examined. Calculating
the dip of the Twechar no 2 using three points (boreholes 1981-11, -12 and -13) shows
an anomalous strike of NE-SW and a dip of 1.52° towards the NW (Figure 4.17). This
strike and dip is totally different from the regional dip and strike of the area, and
suggests that there may be an error during the logging or mineral identification. If it is
true, which of those three sources of information is the most likely to cause the
problem?

In order to solve the problem, drilling logs for boreholes 1981-11 and 1981-12 were re­
examined. Borehole 1981-11 drilling log recorded the presence of coal at the depth of
16.99 m below ground level. From the GDA list (Appendix 2), this coal has been
identified as Twechar no. 2 coal. Examination of drilling log 1981-12 showed that only
one coal was encountered. However in the list given in the GDA report (Appendix 2),
two coals were identified named as Twechar no. 1 and Twechar no. 2 coal at the depth
of 9.49 m and 16.50 m, respectively. The coal found during the drilling (at 9.49 m
deep) is related to the Twechar no. 1, whereas at the depth where they identified
Twechar no. 2, the drilling log recorded ‘sandy mudstone’. No trace of coal was
recorded at this depth.

University Of Glasgow 1998 PhD Dissertation Page 255

New M ethods In Shallow Seism ic Reflection C'hapter Four
Z Z T Harith Surface Tom ography

Figure 4.16 Calculated outcrop points for the Twechar no. 1 and Twechar no. 2
coals based on boreholes 1981-11, -12, and -13.
55450 55500 55550 55600 55650 55700
70600

® 1981 BH 12
70550

T 2

70500

V .

70450

\<a
BH 5

•1 9 8 1

BH 3 70350

1 9 8 1 B l-I 1 4

70300
•1 9 8 1 BH 15 \

70250
55450 55500 55550 55600 55650 55700

T1 and T2 are the calculated outcrop points for Twechar no. 1 and 2 Coals.

University O f G lasgow 1998 PhD D issertation Page 256

New M ethods In Shallow Seism ic Reflection C hapter Four
Z Z T I larith Surface Tom ography

F igure 4 17 The 3-point problem calculation showing anom alous strike and dip
for Tw echar no. 2 Coal.

55450 55500 55550 55600 55650 55700

70600

Tw echar no 2. Coal
B orehole 1981-11 17 86 m
B orehole 1981-12 18.53 m
B orehole 1981-13 22.93 m

(1 9 8 1 B H 1 2
sca le (a p p ro xim a te ly) 1:1820 70550

(1 9 8 1 B H 11

\ l .52
70500

70450

70400
> 1981

70350

70300

70250
55450 55500 55550 55600 55650 55700

University O f Glasgow 1998 PhD D issertation Page 257

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

Although the Twechar no. 2 is not identified in borehole 1981-14 and 1981-16 (refer to
Appendix 2), the drilling logs from these boreholes were also re-examined as the
conjectural outcrop of Twechar no. 2 was drawn to the SW of these boreholes. This is
because they should record the presence of Twechar no. 2. Extrapolating down dip
from the drawn conjectural outcrop of Twechar no. 2, the coal should be logged at
around 4 m and 2 m depth in boreholes 1981-14 and 1981-16, respectively. However at
those depths, both drilling logs showed the presence of sandy mudstone.

In all three drilling log examined (1981-12, 1981-14 and 1981-16), where there should
be Twechar no. 2 coal found, all of them recorded the presence of sandy mudstone
without any evidence of coal. This raises the question whether the Twechar no. 2 coal
identified in 1981-11 is actually the Twechar no. 2 or a deeper coal. To check this,
based on the drawn conjectural line of Twechar no. 2, a down dip extrapolation was
done for 1981-11 and shows that the Twechar no. 2 should be found at a more shallower
depth (around 4 m deep below ground level). The drilling log for 1981-11 borehole at
this depth logged the presence of the sandy mudstone, which is consist with other
boreholes (1981-12, 1981-14 and 1981-16).

The other question about the Twechar no. 2 identified in 1981-11 is that it was found
deeper than in borehole 1981-13 or in borehole 1981-12. Yet, the location of 1981-12
borehole is to the west of those boreholes. Thus the Twechar no. 2 in 1981-11 should
be found at a shallower depth.

This simple analysis shows that although the coal found in borehole 1981-11 was
identified as the Twechar no. 2 coal, this should be treated with care as it may be
confused with the deeper coal. To avoid any confusion, the Twechar no. 2 information
from boreholes 1981-11 and 1981-12 was did not considered for modelling or
discussion.

As the drawn conjectural outcrop for Twechar no. 1 and Twechar no. 2 coal (from the
GDA report) are questionable, can we revise the conjectural outcrop to satisfy all the

University Of Glasgow 1998 PhD Dissertation Page 258

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

conditions? To re-draw the new outcrop line for Twechar no. 1 and 2 we have to
consider all strike/dip direction information as well as the calculated outcrop points for
both Twechar no. 1 and 2 coals for 1981-13 and Twechar no. 1 in 1981-12 borehole
(Figure 4.16). The information from the 3-D time slices at 20 -25 ms (Figures 3.55b,
3.55c, and 3.55d) and the strike of the tunnels/mineworks (or thick coal layers) across
the survey area in the NW-SE direction should also be taken into account.

In boreholes 1981-13,-14,-15, and -16 there is record of the presence of Kilsyth no. 1
Blackband (0.69m thick) or Blackband Under (0.46m thick) Coal at an approximate
depth of 10-20 m (Appendix 2). It is also recorded that the Kilsyth no. 1 Blackband
Coal was being mined (Table 4.2).

If the peak signal detected by the 3-D survey is related to these two coals or related
mineworks, then the strike of the upper coal such as Twechar no. 1 and 2 should also
be drawn parallel to this strike direction. Drawing of the outcrop line for Twechar no. 1
by connecting two calculated points (Borehole 1981-12 and -13) produces a NNW-SSE
direction line, which is oblique to the strike direction. This is unlikely as the strike of
Twechar no. 1 should parallel the regional strike direction.

The second alternative is to introduce a fault or faults running across the field.
Although there was no evidence of faults recorded in any of the reports available,
judging from the extensive fault system shown on the geological map (Figure 3.3) the
possibility of faults being present in the area cannot be ruled out.

The existence for the east fault across the field is the NE-SW (across strike) continuous
red signature on some of the 3-D time slices (e.g. Figure 3.55c). The second supporting
evidence of the east fault is more subtle and was deduced by analysing the plan view of
the abandoned Jubilee coal mine (Figure 3.4), where it shows a strange pattern, where
instead of one continuous working field, there are two fields separated by a tunnel
system in the middle of the field - as if they were the same seam separated by fault.

University Of Glasgow 1998 PhD Dissertation Page 259

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

The existence of the west fault was also deduced from analysing the pattern of the
workings. The plan view for both mineworks (Figure 3.4) also shows an abrupt
termination parallel to the NE-SW direction, as if they also had been faulted. The
location of the west fault was placed so that it coincides with the edge of the mine
workings.

Based on the calculated outcrop points for both Twechar no. 1 and 2 coals, and the
possible existence of two faults, the new revised Twechar no. 1 conjectural outcrop is
drawn as shown in Figure 4.18. If we accept this fault model, then an explanation is
needed to explain why the fault did not show on any reduced time map. One possible
explanation is, since the reduced time map is based on analysing the refractions from
the top of bedrock, which in most of the survey area is sandstone, the differences in
velocity are negligible, thus the differences in velocity did not show up on the reduced
time maps.

On the image map (Figure 4.12) the conjectural outcrop of Twechar no. 1 is drawn on
the blue stripe (time lag) rather than on the red (time lead). The reason for this is, if the
conjectural outcrop shows the presence of the coal (Twechar no. 1, 0.38 m thick) and
coal has relatively lower velocity than sandstone, then the blue zone is an appropriate
place to draw it. The effect of this coal also will slow down the travel time of any
raypath crossing this coal, thus creating a zone of very slow (dark blue) travel times.
The width of this blue stripe (approximately 6-10 m) is also appropriate to the time lag
zone created by this coal although it is only 0.38 m thick. Figure 4.19 shows the relation
between the thickness of the coals and the time lag zones created by them.

University Of Glasgow 1998 PhD Dissertation Page 260

 N ew M ethods In Shallow Seism ic Reflection C hapter I our
Z Z T Harith Surface Tom ography

f igure 4.18 Revised outcrop lines for Twechar no. 1 and 2 Coals.

55450 55500 55550 55600 55650 55700

i l_ _
70600

Note

T1 C alculated point fo r T w echar no 1

T2 C alculated point for T w e ch a r no 2

Blue line is revised outcrop o f T w echar no 1

- 70550
Green line is revised outcrop o f Tw echar no 2

Y ellow line is a N W -SE stru ctu re based

on C h a p te r 3 results

- 70500

- 70450

• _
1995 BH 5

70400

70350

- 70300

70250
55450 55500 55550 55600 55650 55700

University O f G lasgow 1998 PhD Dissertation Page 261

N ew M ethods In Shallow Seism ic Reflection C hapter Four
Z Z T H arith Surface Tom ography

Figure 4.19 Relation between the true thickness, horizontal thickness and maximum
width o f the blue zone. Red lines represent the dipping layer. Wm represents the
m axim um w idth o f the vector band on the surface.

Wm

6m
6m
6m

W= tan (a)

Wm = 2 ( 6 + j W )
Wm(aniso) = 2(6sin(45)-1-yfT)

The width (horizontal thickness) o f the layer is controlled by its true thickness and dip.

Assuming the topography is flat, the relation is W = where W is the width; H is

the true thickness and a is the dip.

Based on the geological map and investigation reports, the average thickness o f

Twechar no. 1 is 0.38 m and it dips at 18° to the NE. This means the horizontal
thickness is approximately 1.2 m; however this is still too narrow com pared to the width
o f the blue strip (6-10 m). To explain this, instead o f the horizontal thickness o f the
coal, the total vector (raypath) lengths should be considered. Any raypath originating in
or passing through this zone will experience a delay. Because we are dealing with the
raypaths, the width o f the blue does not necessarily represent the width o f the low

University O f G lasgow 1998 PhD Dissertation Page 262

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

velocity layer, in this case the Twechar no. 1 coal. Instead, it represents the maximum
length of the path which is travelled perpendicular to the strike in either direction. As
the maximum length of the path is 6 m (maximum offset 10 m, less the critical distance
4 m), the maximum width will be approximately 13 m.

The discussion so far relates the blue stripe to a thick coal layer sandwich by a more
weathered sandstone. This particular coal is may be related to the Twechar no. 1 and
Twechar no. 2 coals. If we accept this idea than we have to explain with what velocity
does the P-wave travelled through the coal? As has been shown above, the horizontal
thickness (h) of the Twechar no. 1 is approximately 1.2 m. If we take the maximum
reduced (RT) delay time for the P-wave to travel through 1.2 m of coal is 5 ms (refer to
Figure 4.9), then from a calculation shown below the P-wave velocity through coal is
approximately 200 m s'1.

RT = — --------—
Vcoal Vref

_ 1.2m 1.2m
5ms =
Vcoal 1600m / s

Vcoal ~ 200m / s

This value is far too small for the velocity of P-waves in coal. Thus, it is impossible to
relate the blue strip to the presence of a coal seam alone.

Since the coal alone cannot be exclusively responsible for the blue strip* another
possible explanation for it is it represents the area where the weatberecj rock is thicker.
In section 4.2 the author discussed how this could occurs. The first model represents a
control condition where the overburden with 600 m s"1 velocity over laying
Carboniferous sandstone-shale bedrock (of 2400 m s'1 velocity). The second model
represents the condition where the bedrock velocity is lower than normal, thus reflecting
a low velocity anomaly. The third model on the other hand was discussed in order to

University Of Glasgow 1998 PhD Dissertation Page 263

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

explain the effects of a thicker overburden over the bedrock. The discussion of these
models was based on two assumptions which are 1) the bedrock interface is flat and, 2)
no variation in topography - thus the thickness of overburden is constant. The
conclusion drawn from this discussion shows that the delay time (blue zone) could be
caused by either a geological anomaly (model one) or the variation in thickness of
overburden (model three).

Picking of the first arrival was done on all shot gathers after the data were corrected for
field statics. The second assumption is based on the fact that the field static correction
corrects all the errors resulting from the differences in elevation and weathering effects.
All the picked first arrivals are therefore referred to a common datum.

The first assumption on the other hand was based on the geological cross-section of the
area (Figure 3.3). However, this assumption is not entirely consistent with the
information from boreholes, the preliminary geophysical survey (Figure 3.6) and the
result of refraction modelling from the Technical University of Vienna (Figure 3.25)
that indicate the bedrock is not flat. Figure 3.25 clearly shows that the top layer
(overburden) varies in thickness by approximately 2 m from West to East along the 60
m long line with a velocity of only 370 m s ' 1. As the thickness variation is
approximately 2 m over the 60 m line, the dip of this reflector is approximately 2°. If
the dip is constant the maximum variation over the 10 m length (the maximum offset of
the data used in this studies) is 0.35 m. The conditions discussed above are presented
graphically as shown in Figure 4.20.

University Of Glasgow 1998 PhD Dissertation Page 264

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

Figure 4.20 Graphical representation of the variation of the overburden.

X=10

2-35

What is the effect of a variation in thickness of overburden on this study? Consider

model one, discussed in section 4.2 where the velocity of the overburden is 600 m s'1
and the velocity of underlying layer (bedrock) is 2400 m s'1. Depending on if the ray
propagates up-slope or down-slope, the arrival time will differ. The up-slope travel
time ( tu) and the down-slope travel time (td) are given by the expression below
(Reynolds 1997).

x[sin(flc - a ) ] | 2zt (cosflJ

vi v.

s[sin(<9c +a)\ | 2z„(cos6>J

where 0C= sin 1

\ v2 J

«14.5°

University Of Glasgow 1998 PhD Dissertation Page 265

New Methods In Shallow Seismic Reflection Chapter Four
Z Z T Harith Surface Tomography

If the minimum thickness is 2 m, then the maximum thickness at the deeper end of the
10 m line will be 2.35 m. Therefore the za and zbthickness will be approximately 2 m

and 2.35 m, respectively (referring to Figure 4.20).

From the expression above, the total up-dip travelled time is 11 ms which gives a
reduced time of 7 ms. This means the time taken to travelled up-dip is delayed 1 ms
compared to the average time. The total down-dip travelled time on the other hand is
12.3 ms which gives the reduced time of 8.3 ms, and a delay of 2.3 ms.

From the discussion above, both up-dip and down-dip reduced time show a delay time,
which could cause the blue zones. Thus, it is fair to say that the observed variations
(blue and red zones) could also be substantially due to variations in drift thickness.

University Of Glasgow 1998 PhD Dissertation Page 266

 CHAPTER FIVE

CONCLUSIONS AND RECOMMENDATIONS

New Methods In Shallow Seismic Reflection Chapter Five
Z Z T Harith Conclusions and Recommendations

5.1 Seismic reflection as a tool for archaeological survey

5.1.1 Conclusions

A number of geophysical techniques are very useful in archaeological studies either in

looking for valuable buried artifacts/objects or in locating archaeological sites. These
include ground penetrating radar (GPR), resistivity, and magnetics or even gravity.
GPR and resistivity are normally used in locating buried ditches while the magnetic
method is used to detect any iron-related objects, which are often associated with the
ancient site. The gravity method is not as popular as those three methods, but has
sometimes been used in locating buried building foundations such as Roman buildings
or roads.

Unlike the above-mentioned four techniques, the seismic method has found itself least
favourite among the usage of geophysical technique in archaeological investigation.
This perception, however, is now changing. To date, some work has been done in this
field utilizing the seismic method. The main target is normally to detect buried ditch­
like structures. The success of the seismic method in archaeological work depends on
the size of the target and the impedance contrast between the host material and the
foreign material filling in the structure. The main seismic technique used in
archaeological work is seismic refraction, as it is easier than the reflection technique.

The work described here was done to prove that the seismic reflection technique also
can be used as an archaeological investigations tool despite its bad reputation among the
‘archaeo-geophysicists’. The aim of this work as to a Roman fort ditch which has been
buried a couple of meters deep. The target itself has a V-shape with a size of
approximately 10 m wide and couple of meters deep. Because the target is very small
and shallow (for the normal reflection survey), the author introduced the highly
condensed acquisition technique where a total of 208 shots were taken along a 50 m -
long 2-D profile. The results of this very shallow seismic reflection survey were then

University Of Glasgow 1998 PhD Dissertation Page 267

New Methods In Shallow Seismic Reflection Chapter Five
Z Z T Harith Conclusions and Recommendations

compared to the results (re-interpreted) of the resistivity survey (the more common
archaeo-geophysical survey method) which had been done by Baker (1996).

The resistivity survey shows the presence of four low resistivity anomalies, with two of
them most likely being ditch-related anomalies. However from excavation records, the
fort is surrounded by only one ditch, which is unusual for a Roman fort as most of them
were typically surrounded by at least two ditches. The two ditch-related anomalies seen
on the resistivity profiles show that the resistivity profiles alone cannot solve the
problem, although it narrowed down the options on where to lay the excavation
trenches.

The results from the highly condensed 2-D seismic reflection profile also shows the
presence of two structure-like ditches. These two ditch-like structures on the seismic
reflection data are found to be correspond well to the two possible ditch-related
resistivity anomalies. The size of one of the ditch-like structure on the seismic
reflection data is approximately same as the excavation records and more interestingly it
lies a few meters off from the expected location (based on excavation records).

Unlike the resistivity result, where the anomalies were interpreted qualitatively and
ambiguously, the seismic reflection result shows an image of the ditch-like structures
itself. Furthermore a quantitative interpretation can also be made based on the image
produced. This means that a more confident determination of the target location can be
archieved using the seismic reflection survey method.

Comparing the results of these two methods, it can be concluded that the seismic
reflection method gives a better result, despite the pre-perception that this method is
very difficult to conduct and the chances of failure are high for a very shallow
investigation.

University Of Glasgow 1998 PhD Dissertation Page 268

New Methods In Shallow Seismic Reflection Chapter Five
Z Z T Harith Conclusions and Recommendations

5.1.2 Recommendations

To increase the chances of success, the physical properties of the target and the
surrounding area need to be understood, as the seismic reflection method only works
best in an area where the contrast in acoustic impedance is high.

Although the result of this reflection survey (together with the resistivity results) gave
the possible location of the buried Roman fort ditch, as well as confirming the existence
of only one ditch, the seismic image can still be improved to show the existence of the
target beyond any doubt.

To produce a clear image, instead of using hammer to generate the seismic source,
higher frequency sources should be considered. This includes the buffalo-gun or even
detonators. The source should also be placed a few feet under the surface (inhole
sources) to avoid the noise contamination. The inhole sources also reduce the
attenuation of the high frequency energy component by the top layer.

Besides the choices of the source used, any future similar survey also should consider a
‘walk-away test’ to determine the optimum window and design the acquisition
parameters based on this window. The combination of a highly condensed CDP and
optimum window concept will concentrate recording of the less contaminated reflection
arrival zone, thus reducing the failure rate.

This work was done using an OYO McSeis 1600 recording instrument (seismograph).
This particular seismograph only has 16 bits of dynamic range (the ratio of the highest
and smallest recoverable signals can be recovered by the system). The smallest
recoverable signal is often taken to be the noise level of the system. However, signals
often can be extracted even though they are buried in the noise. So in future, a
recording system with a higher dynamic range such as OYO DAS-1 which has 24 bit
dynamic range should be considered as it can increases the recoverable ratio, thus
increases the chances of extracting useful signals from high-amplitude noise.

University Of Glasgow 1998 PhD Dissertation Page 269

New Methods In Shallow Seismic Reflection Chapter Five
Z Z T Harith Conclusions and Recommendations

5.2 3-D seismic reflection for engineering-scale site survey

5.2.1 Conclusions

Seismic reflection exploration is deployed mainly by deep seismic explorationists to

map possible petroleum accumulations, and also to investigate the deep interior of the
Earth. The applications of this technique to shallow work are relatively new, but it
became more familiar to shallow geophysical investigators in the 1980’s as seismic
recordings system become smaller and more portable, without reducing their capability.
Today’s high-resolution seismic recording systems are very important in shallow
seismic investigations, as high frequencies (up to 1 KHz or more) can readily be
recorded. Beside the development of high-resolution recording systems, advances in
computer technology have enabled more data to be processed, and in less time.

Today seismic reflection is widely used to investigate targets of less than 200 m depth.
Most of such work is aimed at mapping the depth of the bedrock interface in
engineering foundation studies, or in mapping the aquifer zone in groundwater studies.
Current applications of this technique, together with case studies, are regularly
documented in the Proceedings volumes of the annual symposia or in house journals of
learned societies in the fields of geophysics, engineering geology, groundwater or even
archaeometry. Mostly, a simple 2-D CDP technique is used; however, with the
increasing availability of high-resolution portable seismographs, in the last three years
or so there has developed a trend amongst the geophysicists to try more advanced
techniques such as the 3-D seismic method. This technique is now considered essential
in identifying and mapping hydrocarbon accumulation targets in petroleum exploration.

The major part of thesis is a discussion of a new imaging method in shallow seismic
investigation - a full 3-D shallow seismic survey. At the time this project was initiated
(early 1996), it was believed to be the first frill 3-D seismic reflection survey that had
ever been done for engineering purposes, despite the fact that other shallow 3-D projects
had started at around the same time (referred to in Chapter 3). Most of these, however,

University Of Glasgow 1998 PhD Dissertation Page 270

New Methods In Shallow Seismic Reflection Chapter Five
Z Z T Harith Conclusions and Recommendations

have turned out to be not much more than an extension of a 2-D line reflection survey,
or in other words, 2-D lines with offset sources. In contrast, the high fold of coverage
and range of azimuths and offsets used in the present trial survey are even greater than
in most full-scale 3-D deep seismic reflection surveys.

The aim of the present 3-D work was to image underground mine workings at depths of
approximately 20-35 m. The roofs of the tunnels and rooms are probably collapsed, as
the workings were abandoned on hundred years ago. Besides imaging the tunnels, it was
also hoped to image the bedrock geology.

The results of this work show promising images. The top bedrock, made of inter-layered
Carboniferous sandstone and shale, at very shallow depths (less than 10 m) was partly
mapped over much of the 2 m grid (Figure 3.53). However, the positive identification of
the old mineworkings is much less certain. Reliance has been made on the reversed
polarity of the mapped reflector (Figure 3.54) as an indication of the upper surface of a
void or a thick coal layer as coal, has lower velocity than the sandstone.

The main problem of this work occurred during the acquisition phase, as time was very
limited. The data were collected intensively within a week. There was no time for
extended source suitability comparisons, as only three sources were available, the 7 kg
sledgehammer, weight drop and mini-vibroseis. The condition of the top layer (covered
by thick grass and containing lots of moisture) also contributed in degrading the data
quality. The hammer and weight drop sources did not yield a broad or high enough
frequency range, as was expected. On the other hand the surface materials absorbed
most of the high frequency signal produced by the mini-vibrator despite it being placed
in a 5 to 10 cm deep pre-prepared hole. Besides the absorption problems, the data were
also contaminated heavily by unwanted noise, in particular, the several drilling rigs
operating simultaneously during the acquisition.

The data were processed using industry standard seismic data processing software. The
processing largely followed the standard practice in the petroleum industry, with a few

University Of Glasgow 1998 PhD Dissertation Page 271

New Methods In Shallow Seismic Reflection Chapter Five
Z Z T Harith Conclusions and Recommendations

minor differences. Only 150 ms o f data were recorded, as only the top 50 m was the
region of interest. The shallow bedrock interface meant that the corresponding reflection
arrival was very early. This early reflection arrival tended to overlap with earlier
arriving signals such as the direct wave and refraction first arrivals. Spiking
deconvolution (or spectral whitening for vibroseis data) is ineffective at separating these
closely-spaced signals unless there is a wider-frequency bandwidth than was recorded.

The combination o f short trace length (150 ms), lateral changes and small impedance
contrast with depth created problems in analyzing the velocity variation. The choice of
stacking velocity functions was very difficult. A constant velocity stacking technique
was found to be the best way to stack the data.

5.2.2 Recommendations and further work

Despite the fact that good images of the primary target were not produced, this full 3-D
seismic reflection experiment can be considered a success. However, in retrospect, the
site conditions and target o f the Todd Campus West site may have been too complex for
a preliminary trial 3-D survey. However the experiment shows that it is logistically
feasible, even in the presence o f other site contractors. A better target for future
experimentation would be a somewhat deeper target set of voids in a flat-lying
geological structure, and without the presence o f contractors on site.

The results presented here may be improved by further processing, especially in

separating the refraction first arrivals from the early reflection signals. Preliminary work
in this respect has been started, using wavelet packet transform filtering applied to the
present 3-D dataset (Watts et al. 1998).

In a future trial survey, the acquisition phase is very important, and there should be
more effort in choosing a suitable seismic source. Perhaps a buffalo gun placed down a
hole 0.3-0.5 m deep can be investigated as a source. This may eliminate (to some

University Of Glasgow 1998 PhD Dissertation Page 272

New Methods In Shallow Seismic Reflection Chapter Five
Z Z T Harith Conclusions and Recommendations

extent) the loss of high frequency signal due to the absorption by the loose surface
material, which is inevitable with a surface source such as the minivibrator. Besides the
choice o f source used, it is also vital to study what (minimum) fold is required to obtain
good results from a full 3-D survey. This will minimize the cost of conducting the
survey, or alternatively, more area can be surveyed in the same time and at similar costs.

In view of cost o f grouting suspected or known mine workings, and the fact that the 3-D
seismic method might become useful as a pre- and post-grouting imaging tool, it is
believed to be the most promising geophysical survey method available today.

5.3 Surface tomographic mapping - a new tool

5.3.1 Conclusions

This dissertation has also introduced a new robust and simple tomographic method for
mapping the velocity of the top bedrock. The basic concept of this technique is same as
the fan shooting refraction technique, which requires only the first arrival time from
each source and receiver pair. However, the data are not acquired especially for the
purpose, but are a by-product of the 3-D reflection survey. The aim was to develop a
semi-automated quick process, with minimum user input, to reveal small-scale velocity
anomalies at the level o f (say) the top bedrock.

The surface tomography maps, which use colour-coding to reveal the velocity
anomalies, show some interesting results. Bands of different colours were observed
striking parallel to the strike o f the Carboniferous rock. More interestingly, one of the
low-velocity bands closely follows the re-interpreted outcrop o f the Twechar no. 1 coal.
There was some disagreement on the location o f the conjectural outcrop lines for
Twechar no. 1 and 2 coals (as supplied by the geotechnical investigators in their
reports), thus revised outcrop lines for these coals are presented (Figure 4.18).

University Of Glasgow 1998 PhD Dissertation Page 273

New Methods In Shallow Seismic Reflection Chapter Five
Z Z T Harith Conclusions and Recommendations

This new method is a robust surface mapping technique. The results (in the form of
reduced time maps) show that it can successfully image the bedrock surface geology.
Besides the success in imaging the bedrock, the technique is also shows potential for
detecting any localized anomaly such as a mine shaft or even surface mineralization
related anomalies. This is based on the presence of a few ‘blobs’ of different colours
surrounded by different colours. Unfortunately the author could not prove any shaft-
related anomaly on the map. But after the survey was completed, the geotechnical
investigation found one previously un-recorded shaft, which unfortunately lay just
outside the survey area.

In conclusion, this technique shows potential as a by-product of a 3-D reflection survey

especially in investigating an area which has a complex surface geology such as the
study area. The technique is also can be used to image or study the surface anisotropy.

5.3.2 Recommendations

The reduced time (image) map is produced based on an average reduced times at every
0.25 m grid within 1 m search radius, thus all the point within the seatch radius
contributes the same weight, regardless their location with respect to the central point.
For example, consider three theoretical points A, B and C (Figure 5.1). Point A, B and
C are one out o f a series of 2 m interval points (from 3 corresponding vector lines) with
a reduced time of 5, 8 and 9 ms. An average time out of these three value that
contribute to the central point is 7.3 ms. This value is higher than the nearest reduced
time value which is 5 ms (point A). This value is not appropriate to represent the
reduced time for the central point as it should have a value nearer to 5 ms. This problem
becomes more serious if one o f the point in the area has an extreme value.

University Of Glasgow 1998 PhD Dissertation Page 274

N ew M ethods In Shallow Seism ic Reflection C hapter Five
Z Z T Harith C onclusions and R ecom m endations

Figure 5.1 Points within 1 m search radius which contribute to the average value
o f the central point (red dot).

To overcom e this problem , a different kind o f averaging technique, w h ich takes into

account the effect o f distance from the centre o f the averaging area should be

considered. Such a m ethod should m axim ize the near centre contribution and at the

sam e tim e m inim ize the contribution o f the far (from centre) points. This m eans that

the value carried out by the centre point w ill reflected the con d itions o f the sm all

centered area. One o f the technique can be used to carry out this task is by using a

G aussian (norm al) distribution function and cou ld be d evelop ed in future.

University O f G lasgow 1998 PhD D issertation Page 275

LIST OF REFERENCES
New Methods In Shallow Seismic Reflection
Z Z T Harith References

Aki K., Christofferson A., Husebye E., 1977. Determination of three dimension seismic
structure of the lithosphere. Journal o f Geophysical Research 82, 277-296.

Allen S. J., 1980. Seismic method. Geophysics 45, 1619-1633.

Al-Rawahy S. Y. S., and Goulty N. R., 1995. Effect of mining subsidence on seismic
velocity monitored by a repeated reflection profile. Geophysical Prospecting 43,
191-201.

Baker G. S., Steeples D. W., and Drake M., 1998. Muting the noise cone in near-surface
reflection data: An example from southeastern Kansas. Geophysics 63, 1332-1338.

Baker L. M., 1996. The application o f geophysical techniques to locate the defences of
the Roman Fort at Inveresk, Midlothian. Unpublished BSc dissertation, University of
Glasgow.

Barnes M. K., and Mereu R. F., 1996. An Application of the 3-D Seismic Technique
for Mapping Near-Surface Stratigraphy near London, Ontario. Journal o f
environmental engineering geophysics (JEEE) 1, issue 3, 171-177.

Bay Geophysical Inc., 1998. Company web-page, http://www.bavgeo.com/

Bee M. F., Bearden J. M., Herkenhoff E. F., Supiyanto H., and Koestoer B., 1994.
Efficient 3D seismic surveys in a jungle environment. First Break 12, 253-259.

Bell F. G., 1979. Methods o f treatment o f unstable ground. The Butterworth Group Ltd.

Betrelli C., Mascarin B. and Salvador L., 1993. Planning and field techniques for 3D
land acquisition in highly tilled and populated areas - today’s results and future
trends. First Break 10, 23-32.

Birkelo B. A., Steeples D.W., Miller R. D., and Sophocleous M., 1987. Seismic
Reflection Study o f a Shallow Aquifer During a Pumping Test. Groundwater 25,
703-709.

Breeze D. J., 1983. Roman Forts in Britain. Aylesbury, Shire Publication.

Brown A. R., 1991. Interpretation o f three-dimensional seismic data (third edition).

AAPG Memoir 42, AAPG Publishing.

Brown A. R., 1996. Interpretation o f three-dimensional seismic data (fourth edition).

AAPG Memoir 42, AAPG Publishing.

Briickl E., Castillo E., Madhusudan, and Salzmann H., 1997. Detection o f abandoned
underground coal mine workings by seismic methods. Unpublished article.

Buhnemann J., and Holliger K., 1998. Comparison of high-frequency seismic sources
at the Grimsel test site, central Alps, Switzerland. Geophysics 63, 1363-1370.

University Of Glasgow 1998 PhD Dissertation Page 276

New Methods In Shallow Seismic Reflection
Z Z T Harith References

Biiker F., Green A. G., and Horstmeyer H., 1998a. Shallow seismic reflection study o f
a glaciated valley. Geophysics 63, 1395-1407.

Biiker F., Green A. G., and Horstmeyer H., 1998b. Shallow 3-D seismic reflection
surveying: Data acquisition and preliminary processing strategies. Geophysics 63,
1434-1450.

Burger H. R., 1992. Exploration geophysics o f the shallow subsurface. Prentice Hall
Ltd.

Chiu S. K. L., and Stewart R. R., 1987. Tomographic determination o f three-

dimensional seismic velocity structure using well logs, vertical seismic profiles and
surface seismic data. Geophysics 52, 1085-1098.

Chiu S. K. L., Kanasewich E. R., and Phadke S., 1986. Three-dimensional

determination o f structure and velocity by seismic tomography. Geophysics 51,
1559-1571.

Clark J. C., Johnson W. J., and Miller W. A., 1994. The application of high resolution
shear wave seismic reflection surveying to hydrogeological and geotechnical
investigations. In Proceeding o f the Symposium on the Application o f Geophysics to
Engineering and Environmental Problems, vol. 1, March 27-31.

Collingwood R.G and Richmond I., 1969. The Archaeology of Roman Britain 3rd
edition. Methuen and Co Ltd.

Cottin J. F., Deletie P., Jacquet-Francillon H., Laksmanan, J., Lemoine Y., and Sanchez
M., 1986. Curved ray seismic tomography: application to the Grand Etang Dam
(Reunion Island). First Break 4, 25-30.

Corsmit J., Versteeg W. H., Brouwer J. H., Helbig K., 1988. High-resolution 3D
reflection seismics on a tidal flat: acquisition, processing, and interpretation. First
Break 6, 9-23.

Dana P. H., 1994, (revised 1998). Global Positioning System Overview. The
Geographer’s Craft Project, Dept, o f Geography, The University of Texas at Austin.

Dennison E. P. and Coleman R., 1996. The Scottish Burgh Survey, Historic
Musselburgh: the archaeological implications o f development. Historic Scotland and
the Scottish Culture Press.

DevaneyA. J., 1983. A Computer Simulation Study o f Diffraction Tomography. IEEE

Transactions on Biomedical Engineering 30, 377-386.

Dickinson J. A., Fagin S. W. and Weisser G. H., 1990. Comparison of 3-D Seismic
Acquisition Techniques on Land. SEG Expanded Abstract 1990 technical program
vol. 1, 60th annual international SEG meeting Sept 23-27 1990.

University Of Glasgow 1998 PhD Dissertation Page 277

New Methods In Shallow Seismic Reflection
Z Z T Harith References

Dines K. A., and Lytle R. J., 1979. Computerised Geophysical Tomography.

Proceedings o f the IEEE 67, 1065-1073.

Dolan F., and Pratt T. L., 1997. High-resolution seismic profiling o f the Santa Monica
Fault Zone, West Los Angeles, California. Geophysical Research Letters 24, 2051-
2054.

Doll W. E., Miller R. D., and Xia J., 1998. A noninvasive shallow seismic source
comparison on the Oak Ridge Reservation, Tennessee. Geophysics 63, 1318-1331.

Fessenden H. M., 1940. Fessenden, Builder o f tomorrow. New York, Coward-McCann

Inc.

Fessenden R. A., 1917. Patent no 1240328, Methods and apparatus for locating ore
bodies. US Patent Office

Francese R. G., and Hajnal Z., 1997. Reflection seismic imaging of complex
stratigraphic features in glacial deposits. EAGE extended abstract, 59th Conference
and Technical Exhibition - Geneva, Switzerland, 26-30 May.

Gendzwill D, 1990. High-Resolution Seismic Reflections in an Underground Mine.

SEG Expanded Abstract 1990 technical program vol. 1, 60th annual international
SEG meeting Sept 23-27 1990.

George T. N., 1973. Geology Around Glasgow. In Excursion guide to the geology of the
Glasgow district, edited by Bluck, B.J., 1973. Geological Society of Glasgow.

Ghose R., Nijhof V., Brouwer J., Matsubara Y., Kaida Y., and Takahashi T., 1998.
Shallow to very shallow, high-resolution reflection seismic using a portable vibrator
system. Geophysics 63, 295-1309.

Gochioco L. M., and Cotton S. A., 1989. Locating faults in underground coal mines
using high-resolution seismic reflection techniques. Geophysics 54, 1521-1572.

Goulty N. R., and Hudson A. L., 1994. Completion o f the seismic refraction survey to
locate the vallum at Vindobala, Hadrian’s wall. Archaeometry 36, 327-335.

Goulty N. R., Gibson J. P. C., More J. G., and Welfare H., 1990. Delineation of the
vallum at Vindobala, Hadrian’s Wall, by a shear-wave seismic refraction survey.
Archaeometry 32, 71-82.

Green A., Pugin A., Beres M., Lanz E., Btiker F., Huggenberger P., Horstmeyer H.,
Grasmiick M., Delaco R., Holliger K., and Maurer H. R., 1995. 3-D high-resolution
seismic and georadar reflection mapping o f glacial, glaciolacustrine and glaciofluvial
sediments in Switzerland. Annual symposium environmental engineering geophysics
soc. (SAGEEP) Expanded abstracts, 419-434.

Greenberg, J., 1997. Activity buoys new technology. AAPG Explorer Nov 97, 8-9&18.

University Of Glasgow 1998 PhD Dissertation Page 278

New Methods In Shallow Seismic Reflection
Z Z T Harith References

Hasbrouck Geophysical Inc., 1994. Shallow 3-D reflection seismic example. Hasbrouck
Geophysical Inc. homepage http://www.hasgeo.com

Heliker, Griggs, Takahashi, and Wright, 1986. Volcano monitoring at the US

Geological Survey’s Hawaiian Volcano Observatory. Earthquake Information
Bulletin 18, 35.

Hewitt, 1980. Seismic Data Acquisition. SEG Continuing Educational Program, Soc.
Exploration Geophysics, Tulsa.

Hill, I. A. and Ali, J. W., 1988. Application of shallow seismic reflection methods to
site investigation. Trans. Inst. Min. Metall. (Sect. B: Appl. Earth sci) 97, B29-B35.

Hoover, G. M. and O’Brien, J. T., 1980. The influence of the planted geophone on
seismic land data. Geophysics 45, 1239-1253.

House J. R., Boyd T. M., and Haeni F. P., 1996. Haddam Meadows, Connecticut: A
case study for the acquisition, processing, and relevance of 3-D seismic data as
applied to the remediation of DNAPL contamination. In AAPG studies in Geology
no 42 and SEG Geophysical development series no 5, edited by Weimer P and Davis
L. AAPG/SEG publication, Tulsa, 257-266.

Hunter J. A., Pullan S. E., Bums R. A., Good R. L., Harris J. B., Pugin A., Skvortsov
A., and Gorianiov N. N., 1998. Downhole seismic logging for high-resolution
reflection surveying in unconsolidated overburden. Geophysics 63, 1371-1384.

Hunter J. A., Pullan, S. E., Bums, R. A., Gagne, R. M. and Good, R. L., 1984. Shallow
seismic reflection mapping for the overburden-bedrock interface with an engineering
seismograph - Some simple techniques. Geophysics 49, 1381-1385.

Hunter J. A., Pullan, S. E., Bums, R. A., Gagne, R. M. and Good, R. L., 1987.
Applications o f a Shallow Seismic Reflection Method to Groundwater and
Engineering Studies. Proc. o f exploration 87, Ontario Geological Survey Special
Volume: an application o f geophysics and geochemistry, 704-715.

Hyde E. D., 1987. Coal Mining in Scotland. Scottish Mining Museum, Edinburgh.

Jefferson R. D., Steeples D. W., Black R. A., and Carr T., 1998. Effects of soil-
moisture content on shallow seismic data. Geophysics 63, 1357- 1362.

Jeng Y, 1995. Shallow seismic investigation of a site with poor reflection quality.
Geophysics 60, 1715-1726.

Johnson W. J., and Clark J. C., 1992a. High resolution S-wave Reflection Surveying for
Groundwater Characterisation. In Proceedings o f RD 92, National research and
development conference on the control o f hazardous materials, Feb. 2-6, 1996 San
Francisco.

University Of Glasgow 1998 PhD Dissertation Page 279

New Methods In Shallow Seismic Reflection
Z Z T Harith References

Johnson W. J., and Clark J. C., 1992b. Improving subsurface resolution with the seismic
reflection methods: Use S-wave. In The Proceedings o f the 6th national outdoor
action conference national ground water assoc., May 11-13 1992, Las Vegas.

Karastathis V. K., and Papamarinopoulos St. P., 1997. The detection of King Xerxes’
Canal by the use o f shallow refraction and reflection seismics - Preliminary results.
Geophysics Prospecting 45, 389-401.

Kearey P., and Brooks M., 1991. An Introduction to Geophysical Exploration (second
edition). Blackwell Science Publications.

Kearey P., and Vine F. J., 1990. Global Tectonics (first edition). Blackwell Scientific
Publications.

Kidd D. F., Kruppenbach J. A., and Ricketts B., 1990. Multidiscipline Approach to
Very Shallow Exploration Targets. SEG expanded abstract 1990, tech. program 1,
60th annual international SEG meetings, Sept 23-27 1990, San Francisco.

Kilty K. T., and Lange A. L., 1990. Acoustic Tomography in Shallow Geophysical
Exploration Using a Transform Reconstruction. In Geotechnical and Environmental
Geophysics, edited by Stanley H. Ward. Society of Exploration Geophysicists.

Knapp R. W., and Steeples, D. W., 1986a. High-resolution common-depth point seismic
reflection profiling. Instrumentation. Geophysics 51, 276-282.

Knapp R. W., and Steeples D. W., 1986b. High-resolution common-depth-point

reflection profiling: Field acquisition parameter design. Geophysics 51, 283-294.

Kourkafas P., and Goulty N. R., 1996. Seismic reflection imaging of gypsum mine
workings at Sherbum-in-Elmet, Yorkshire, England. European Journal of
Environmental And Engineering Geophysics 1, 53-63.

Lamer A., 1970. Couplage sol-geophone. Geophysical Prospecting 18, 300-319.

Lankston R W., 1989. The seismic refraction method: A viable tool for mapping
shallow targets into the 1990s. Geophysics 54, 1535-1542.

Lanz E., Maurer H., and Green A. G., 1998. Refraction tomography over a buried
waste disposal site. Geophysics 63, 1414-1433.

Lanz E., Pugin A., Green H., and Hurstmayer H., 1996. Results o f 2- and 3-D high-
resolution seismic reflection surveying o f surficial sediments. Geophysical Research
Letters 23, 491-494.

Leick A., 1992. Delineating Theory for GPS Surveying. Journal o f Surveying
Engineering 118, 33-42.

University Of Glasgow 1998 PhD Dissertation Page 280

New Methods In Shallow Seismic Reflection
Z Z T Harith References

Lepper C. M., 1981. Guideline for Selecting Seismic Detectors for High-Resolution
Applications. Bureau of Mines Ref. Of Investigation 8500, US Dept of the Interior.

Lowrie W., 1997. Fundamentals of Geophysics. Cambridge University Press.

Lytle R. J., and Dines K. A., 1980. Iterative Ray Tracing Between Boreholes for
Underground Image Reconstruction. I.E.E.E Trans. Geosci. Remote Sensing. GE-18,
234-240.

MacGregor M. and MacGregor A.G., 1948. British Regional Geology: The Midland
valley o f Scotland, 2nd edition. Her Majesty’s Stationary Office.

Miller R. D., and Xia J., 1998. Large near-surface velocity gradients on shallow
reflection data. Geophysics 63, 1348-1356.

Miller R. D., Pullan S. E., Steeples D. W., and Hunter J. A., 1994. Field comparison of
shallow P-wave seismic sources near Houston, Texas. Geophysics 59, 1713-1728.

Miller R. D., Steeples D. W., and Brannan M, 1989. Mapping a bedrock surface under
dry alluvium with shallow seismic reflections. Geophysics 54, 1528-1534.

Miller R. D., Steeples D. W., and Mazzella A., 1990. Reflection from geologic interface
shallower than 30 m at the Pittman Lateral, Henderson, Nevada. SEG expanded
abstract, 1990 technical program vol. 1, 60th Annual International SEG meeting Sept
23-27 1990 San Francisco.

Miller R.D., Villela A., Xia J., 1997. Shallow High-Resolution Seismic Reflection to
Delineate Upper 400 m around a Collapse Feature in Central Kansas. Environmental
Geoscience 4, 119-126.

Milson J., 1989. Field Geophysics. Open University Press.

Narbutovskih S. M., Michelsen F. B., Clark J. C., and Christensen E. W., 1995. High
resolution seismic reflection test at the DOE Hanford site. Proc. Symp. Application
Geophysics to Engineering and Environmental Problems, 947-951.

Norminton E. J., 1990. Seismic Model Studies o f the Overburden Bedrock Reflection:
P-wave and S-wave. SEG expanded abstract, 1990 technical program 1, 60th annual
international SEG meeting Sept 23-27 1990.

Ongkiehong L., and Askin H. J., 1988. Towards the universal seismic acquisition
technique. First Break 6,46-63.

Owen T. R. E., and Sinha M. C., 1990. A PC-based seismic acquisition system. First
Break 8, 57-62.

Peterson J. E, Paulsson B. N. P., and Mc.Evilly T. V, 1985. Applications o f algebraic

reconstruction techniques to crosshole seismic data. Geophysics 50, 1566-1580.

University Of Glasgow 1998 PhD Dissertation Page 281

New Methods In Shallow Seismic Reflection
Z Z T Harith References

Pullan S. E., and Hunter J. A., 1985. Seismic model studies of the overburden-bedrock
reflection. Geophysics 50, 1684-1688.

Pullan S. E., and MacAullay H. A., 1987. An in-hole shotgun source for engineering
seismic surveys. Geophysics 52, 985-996.

Pullan S. E., Hunter J. A., and Neave K. G., 1990. Shallow Shear-Wave Reflection
Tests. SEG expanded abstract, 1990 technical program 1, 60th annual international
SEG meeting Sept 23-27, 380-382.

Reynolds J. M., 1997. An Introduction to Applied and Environmental Geophysics. John

Wiley and Son Publication.

Rice J. A, Krohn C. E., and Houston L. M., 1990. Shallow Near-Surface Effects on
Seismic Waves. SEG expanded abstract, 1990 technical program 1, 60th annual
international SEG meeting Sept 23-27, 747-749.

Richmond I. A. (the late), (1978). A Roman Fort at Inveresk, Midlothian, edited and
prepared by W. S. Hanson. Proc. Soc. Antiq. Scot. 110, 286-304.

Ritchie J. N. G. and Ritchie A., 1972. Regional Archaeologies: Edinburgh and

Southeast Scotland. Heinemann Educational Books Ltd.

Rogers A. W., 1981. Determination o f static correction. In Development in Geophysical

Exploration Method-2, edited by Fitch A. A. Applied Science Publication.

Sheriff R. E., 1973. Encyclopaedic Dictionary of Exploration Geophysics, 1st edition.

Society o f Exploration Geophysics, Tulsa.

Sheriff R. E., 1990. Encyclopaedic Dictionary of Exploration Geophysics, 3rd edition.

Society o f Exploration Geophysics, Tulsa.

Sheriff R. E., and Geldart L. P., 1995. Exploration seismology, 2nd edition. Cambridge
University Press.

Siahkoohi H. R., and West G. F., 1996. 3-D seismic imaging o f complex structures in
glacial deposits. SEG Expanded Abstract, 60thAnnual International Mtg., 873-876.

Smythe D. K., 1994. 3-D seismic reflection trial- data acquisition report. Unpublished
NIREX report no 622.

Smythe D. K., 1995. The 3D structural geology o f the PRZ in Radioactive waste
disposal at Sellafield, UK, edited by R S Haszeldine and D K Smythe. Department of
Geology and Applied Geology, University o f Glasgow.

Steeples D. W., 1990. Spectral Shaping during Acquisition of Seismic Reflection Data.
SEG expanded abstract, 1990 technical program 1, 60th annual international SEG
meeting Sept 23-27, 917-921.

University Of Glasgow 1998 PhD Dissertation Page 282

New Methods In Shallow Seismic Reflection
Z Z T Harith References

Steeples D. W., and Miller R. D., 1986. Some shallow seismic-reflection pitfalls. SEG
Expanded Abstracts with Biography, 56th Technical Program, 101-104.

Steeples D. W., Green A. G., McEvilly T. C., Miller R. D., Doll W. E., and Rector J.
W., 1997. A workshop examination of shallow seismic reflection surveying. The
Leading Edge 16, 1641-1647.

Steeples D. W., Knapp R. W., and McElwee C. D., 1996. Seismic reflection
investigations o f sinkholes beneath Interstate Highway 70 in Kansas. Geophysics 51,
295-301.

Stefanic V., 1997. Seismic industry rides the waves, plenty of work and success but
where are the new hires. AAPG Explorer June 1997, 1,6,10.

Stone G. D., 1994. Designing Seismic Surveys in Two and Three Dimensions, edited by
Meeder C. A. SEG publishing.

Stright M. J., 1986. Evaluation of archaeology site potential on the Gulf of Mexico
continental shelf using high-resolution seismic data. Geophysics 51, 605-622.

Thomas K., 1997. Faster, better, data is industry goal. AAPG Explorer June 97, 8-9.

Veen, M. van der, and Green A. G., 1998. Land streamer for shallow data acquisition:
Evaluation o f gimbal-mounted geophones. Geophysics 63, 1408-1413.

Waltham A. C., 1997. Foundation o f Engineering Geology. Blackie Academic and

Professional.

Watts D. R., Smythe D. K., and Deighan A. J., 1998. Removal of airblast from a 3-D
engineering seismic survey with a wavelet packet filter. SEG expanded abstract with
biographies, 1998 technical program 1, 821-824.

Whiteley R. J., Hunter J. A., Pullan S. E., and Nutalaya P., 1998. “Optimum offset”
seismic reflection mapping o f shallow aquifers near Bangkok, Thailand. Geophysics
63, 1385-1394.

Wright C., 1996. A note on the computation of the offset distance XY in the
generalized reciprocal method o f seismic refraction interpretation. Geophysical
Prospecting 44, 41-53.

Wu J., 1996. Potential pitfalls o f crooked-line seismic reflection surveys. Geophysics

61, 277-281.

Wynn J. C., 1986. Archaeological Prospection: An introduction to the special issue.

Geophysics 51, 533-537.

Yilmaz O., 1987. Investigations in Geophysics. Vol. 2: Seismic data processing. SEG
publication.

University Of Glasgow 1998 PhD Dissertation Page 283

APPENDICES
New Methods In Shallow Seismic Reflection
Z Z T Harith A p p e n d ix 1

Appendix 1

The Basis o f surface tom ography (after Lowrie 1997).

(A) (B) 1 .9 5 s

N 0% +5 v /
/ 1 .8 5 s
6 .6 6 .3 + 10% + 5%
/
c 1% 2 .1 s
6 .0 5 .4 w
0% -1 0 %

f ^r 2 .0 s
P -w a v e v e lo c itie s (K m /s ) V e lo c it y a n o m a lie s (% ) l. 9 s 2 . 05 s

T r a v el tim e s (s e c o n d s )

(D) + 0 .0 5 s + 2 .5 %

3% 3%
+ 0 .1 5 s + 7 .5 %

0%
-5 %

0 .0 s 0%
+ 0.1 s - 0 .0 5 s + 5% - 2 .5 %

T r a v el tim e a n o m a lie s ( s e c o n d s fa st) T ra v el tim e r e s id u a ls (p e r c e n t fa st)

Consider the passage o f rays in six directions through a square region containing four
equal areas w ith P-wave velocities 5,4 km s '1, 6.0 km s '1, 6.3 km s '1 and 6.6 km s-i
(App. la). Let the expected velocity in the region be 6.0 km s '1. The velocity anom aly
in each area is found by subtracting the reference value and can be expressed as a
percentage o f the expected value. This gives zones that are 10% and 5% fast, a zone
that is 10% slow and a zone w ith no anom aly (App. lb ). Suppose that six seismic rays
traverse the square region (App. lc ), and let the expected travel tim e be 2.0 s for each
ray; some o f the observed travel tim es are shorter and some are longer than the expected
value. The travel time anom aly is com puted by subtracting the observed tim es from the
expected travel tim e (App. Id); it is positive or negative depending on w hether the ray
travels faster or slow er than average. The travel tim e residual is obtained by expressing
the anom aly as a percentage o f the expected travel tim e (App. 1e). A positive residual
result for an early arrival that has travelled through a fast zone w ith positive velocity
anomaly; a negative residual corresponds to passage through a slow zone w ith negative
velocity anom aly. The set o f travelled tim e residual form s the database for topographic
inversion, w hich is designed to yield the original velocity structure (App. la), or, more
com monly, the distribution o f fast and slow regions (App. lb).

University Of Glasgow 1998 PhD Dissertation Page 284

New Methods In Shallow Seismic Reflection
Z Z T Harith A p p e n d ix 2

Appendix 2

Correlation o f coal seams and strata encountered in boreholes 1981-1 to -29 inclusive
(after Scottish D evelopm ent Agency).

Bore no D epth to pavem ent o f seams

1 K in g B la ck b a n d C o a l 8 .2 0 m
P o ssil M ain C o a l (c a v ity ) 1 0 .4 0 m
2 U p p er P o ssil Iro n sto n e (c o a l) 8 .0 0 m
L o w e r w e e c o a l (c o a l) 1 0 .9 1 m
F ou rteen In ch iro n sto n e (c o a l) 1 6 .9 7 m
F ou rteen In ch U n d er C o a l 2 5 .0 9 m
3 F ou rteen In ch U n d o - co a l 1 2 .3 5 m
A s h fie ld C o k in g C oal 1 9 .6 8 m
K in g C oal 2 6 .2 5 m
P o ssil m a in C o a l (c a v ity ) 2 8 .2 0 m
4 A s h fie ld C o k in g C o a l 8 .4 3 m
K in g C oal 1 5 .0 3 m
P o ssil m a in C o a l 1 8.5m
5 P o ssil w e e c o a l (w a s te ) 1 7 .0 0 m
U p p er p o s s il or M e ik le h ill m ain c o a l (L o o s e p a c k e d w a s te ) 2 0 .4 0 m
6 U p p er P o ssil Iro n sto n e (p o s s ib le w a s te ) 1 8 .0 0 m
L ow er w e e coal 2 2 .5 0 m
F ou rteen inch Iron ston e (c o a l) 2 7 .5 5 m
F ou rteen In ch U n d er co a l 3 3 .8 5 m
7 K ilsy th n o 1 B la ck b a n d U n d e r co a l 1 0 .4 5 m
8 P o ssil W e e C o a l 2 1 .2 0 m
U p p er P o ssil or M e ik le h ill m a in coal 2 2 .9 0 m
9 L ow er w e e coal 9 .4 0 m
F ou rteen In ch iro n sto n e (c o a l) 1 5 .5 5 m
F ou rteen in ch U n d er c o a l 2 2 .4 0 m
10 K in g c o a l 9 .7 5 m
(or K in g c o a l (w a s te )) ( 1 3 .6 0 m )
P o ssil M ain C o a l (w a s te ) 1 3 .6 0 m or 1 7 .8 5 m
11 Twechar No 2 Coal 1 7 .3 7 m
12 Twechar No 1 Coal 9 .4 9 m
Twechar No 2 Coal (p o s itio n ) 1 6 .5 0 m
13 Twechar No 1 Coal 5 .2 0 m
Twechar No 2 Coal 1 3 .9 2 m
K ilsy th N o 1 B la ck b a n d C o a l (p o s s ib le c lo s e d w a s te ) 1 9 .7 0 m
14 K ilsy th N o 1 B la ck b a n d (w a s te ) 1 5 .3 0 m
K ilsy th N o 1 B la ck b a n d U n d er C o a l (w a s te ) 2 2 .3 0 m
15 K ilsy th N o 1 B la ck b a n d U n d er C oal 9 .9 6 m
16 K ilsy th N o 1 B la ck b a n d (w a s te ) 1 4 .5 0 m
K ilsy th N o 1 B la ck b a n d U n d er C o a l (p o s itio n ) 2 0 .9 0 m
17 P o ssil M ain 3 .5 m
C lo v e n C oal 1 0 .5 6 m
S h a le C o a l (c o lla p s e d w a ste ) 2 2 .4 2 m
J u b ilee C oal 2 8 .6 7 m

University Of Glasgow 1998 PhD Dissertation Page 285

New Methods In Shallow Seismic Reflection
Z Z T Harith Appendix 2

18 Shale Coal (void) 13.65m

Jubilee Coal 20.08m
Knightswood Gas Coal (waste) 27.68m
19 Cloven Coal 5.47m
Shale Coal (waste) 18.3m
Jubilee Coal (waste) 24.20m
20 Possil Main Coal 5.49m
Cloven Coal 12.48m
Shale Coal (waste) 26.56m
21 Shale Coal 3.91m
Jubilee Coal 9.66m
Knightswood Gas Coal (void) 18.71m
22 Jubilee Coal 3.66m
Knightswood Gas Coal (void) 13.06m
23 Shale Coal (waste) 6.91m
Jubilee Coal 11.69m
Knightswood Gas Coal 20.59m
24 Knightswood Gas Coal 7.55m
Knott Coal 26.92m
25 Knott Coal 21.17m
26 Knott Coal 13.98m
27,28,29 Penetrated strata referred to as the Black Metals known to underlie
the Knott Coal

University Of Glasgow 1998 PhD Dissertation Page 286

 N ew Methods In Shallow Seismic Reflection
Z Z T Harith Appendix 3

Appendix 3

Calculation of 3-point problem to establish regional strike and dip.

a) Strike and dip o f Blackband Coal based on boreholes 1981-13,-14, and -16.

Borehole 13

Surface level 36.85 m

Blackband Coal
at 19.70 m d e e p 1 7.15m

Borehole 14

Surface level 39.45

Blackband Coal
at 15.30 m depth 24.15

Borehole 16

Surface level 40.82 m

Blackband Coal
at 14.50 m depth 26.32

Dip of the layer is

24.15 m 3.5 cm @ 3.20 m Diff. Between Bh. 13 and 16 is 9.15 m

Dist. Between Bh. 13 and 16 is 10 cm

7.0 m which m eans

17.15m 1 cm = 0.915 m or
1 m = 1.090 cm

TanO = 7.01/3.20 Diff between Bh 14 and 16 is 2.17 m

= 2.18 On m ap 2.17 m is equal to 2.4 cm from Bh 16
O = 65.4 degrees.

University Of Glasgow 1998 PhD Dissertation Page 287

N ew Methods In Shallow Seismic Reflection
Z Z T Harith Appendix 3

b) Strike and dip o f Blackband Under Coal based on boreholes 1981-14, -15, and
-16.

Diff. Between Bh. 14 and 15 is 12.97 m

Dist. Between Bh. 14 and 15 is 3.8 cm

which m eans

1 cm = 3.4 m or

Diff. Between Bh. 14 and 16 is 2.17 m

on map, 2.17 m is equal to 0.65 cm from Bh. 14

Borehole 14

Surface level 39.45 m

Blackband Under Coal
at 22.30 m d e e p 1 7. 1 5m

Borehole 16

Surfadejevel 40.82 m
Blackbara^Uqder Coal
at 20.90 m cteeo 19.92 m

Dip of the layer is Borehdt©J5

Surface level 40.08 m

3 cm @ 10.2 m Blackband Under Coal
.12m at 9.96 m d e e p 30.12 m

10.2 m

19.92 m

TanO = 10.2/10.2
= 1
O = 45 degrees.

University Of Glasgow 1998 PhD Dissertation Page 288

 N ew Methods In Shallow Seismic Reflection
Z Z T Harith Appendix 3

c) Strike and dip o f Blackband Under Coal based on boreholes 1981-7, -14, and
-15.

Borehole 7

Surface level 33.56 m

Blackband Under Coal Diff. Between Bh. 14 and 15 is 12.97 m
at 10.45 m deep 23.11 m Dist. Between Bh. 14 and 15 is 3.7 cm

which m eans

1 cm = 3.5 m or
1 m = 0.28 cm

Diff between Bh 7 and 14 is 5.96 m

On m ap 5.96 m is equal to 1.67 cm from Bh 14

Borehole 14
Surface level 39.45 m
Blackband Under Coal
at 22.30 m d e e p 17.15m

Dip of the layer is

30.12 m 1.9 cm @ 6

7.01 m
Borehole 15

23.11 m Surface Level 40.08 m

Blackband Under Coal
at 9.96 m d e e p 30.12 m
TanO = 7.01/6.65
= 1.05
O = 46.50 degrees.

University Of Glasgow 1998 PhD Dissertation Page 289

N ew Methods In Shallow Seismic Reflection
Z Z T Harith Appendix 4

Appendix 4

Calculations of conjectural outcrop points o f Twechar no. 1 and Twechar no. 2 Coals.

Calculation of the conjectural outcrop o f Twechar no. 1 and no. 2 based on the presence
o f these coals in boreholes 1981-11, -12, and -13. Note that the dip o f these layers
(18°) is taken from the geological map and documentation reports as the readings
calculated from the three point problem using the deeper coal horizon shows a rather
steep dip (>42°, appendix 3).

Conjectural width (CW)

Depth

CW = Depth/Tan (18°)

Scale (Figure 4.16)

5.5 cm (on map) = 100m (on surface) or
1 cm (on map) = 18.2 m (on surface) or
0.05 cm (on map) = 1 m (on surface)

Twechar No 1 Coal

Borehole 11 Borehole 12 Borehole 13

Surface level (m) 35.23 35.03 36.85
Twechar No 1 at (depth, m) 9.49 5.20
Twechar No 1 25.54 31.65
Conjuctural Outcrop (m) 29.20 16.00
(depth/tan 18)
Conjuctural on map (cm) 1.5 0.8

Twechar No. 2 Coal

Borehole 11 Borehole 12 Borehole 13

Surface level (m) 35.23 35.03 36.85
Twechar No 2 at (depth, m) 17.37 16.50 13.92
Twechar No 2 17.86 18.53 22.93
Conjuctural Outcrop (m) 53.45 50.78 42.84
(depth/tan 18)
Conjuctural on map (cm) 2.7 2.5 2.1

U niversity O f Glasgow 1998 PhD Dissertation Page 290

WEST of SCOTLAND SCIENCE PARK
(TODD CAMPUS WEST)

REDUCED TIME (VECTOR) MAP

M sm

w m m

m m m

Figure: 4.8

Back p ock et: 1

Grid : 2 m x 2 m

Reduction velocity : 1600 m/s

Datum CD
UvO

<n ®
000
M C -*-1
—<6 O
L lQ L L O C
 WEST of SCOTLAND SCIENCE PARK
(TODD CAMPUS WEST)

REDUCED TIME (VECTOR) MAP

(0-90 anisotropy map)
55500 55510

Figure : 4 .10a

Back p ock et: 2

Grid : 2 m x 2 m

Reduction velocity : 1600 m/s

Datum
Ro t a t e =
WEST of SCOTLAND SCIENCE PARK
(TODD CAMPUS WEST)

REDUCED TIME (VECTOR) MAP

(90-180 anisotropy map)

20 55530 55540 55550 55560

Figure : 4.10 b

Back p ock et: 3

Grid : 2 m x 2 m

Reduction velocity : 1600 m/s

5 _ocr
Datum
□CM
a*—
Ll.0"
tSM-
®co
Ro t a t e =

EO<
c — -
null
—ix3
Li_QQ_
 WEST of SCOTLAND SCIENCE PARK
(TODD CAMPUS WEST)

REDUCED TIME (IMAGE) MAP

55500 55510 55520 55530 55540 55550 55560 55570 55580

70510 70510

70500 - 70500

70490 70490
TWECHAR NOI

70480 70480

70470 70470

70460 70460

TWECHAR N02.

70450 H 70450

70440 - 70440

70430 - Figure : 4. 12
70430
Back p o ck et: 4

Grid : 2 m x 2 m
70420 - 70420
Reduction velocity : 1600 m/s

70410 - 70410

70400 - 70400

70390 - 70390 DATUM

\
70380 t | i i i r | i i i— i— |— i— r 70380
55500 55510 55520 55530 55540 55550 55560 55570 55580
 p a w
w on
D H
O o
n
w
O oo
o o
D
oo
H
np
> >
f f l

< 3c/3 °C/3

nH
w
n
L_J W
O m iy

% n

3 m
> 0 0

>
hd pa

O w h-* •
0> 3. P
0- o 00
C
c Ou p*r i-i
O • •
T3 O)
o O'* to O
Q)
c
o O
3
<
3 ?r
CD
x
hL
o to O i
p
o
»—►
1— * • 3
v:

O n
O
O
§.
C/5
 3
p a f f l

tn 00
O 3
H
o
- v Cj H |- b
£ n O oo
9 m on
h - a
oo D O
o H
p
n^
3
K* f f l
>>
o ^
h -4

o < 2c O C/5
f f l
n c/5

5 q m
w ^m
-g s 00 n
w
13
m M m m
>
13
>
pa

o 00 ■fl
CD
CL
3. 8FT CTQ*
CL C
co
ah>*
o
£D
r-t
H-*•
o
3
tO
3
O
o 4L
3
r+
c <
2-
oo *
X
to
3
T?
a
rl-

ON
cr

ON.
o
o
C/3
 tn
m go
H
O
c 3H o> -b
n O go
o tn
L A
a
j
O Q
Do
LA □
/ B
£3 S L_J
»-*• U J
>>
° o 3 o
PT ^ c |
o tn 00 00
^ n n
^ H
3 O m w

£ * H n
w

> >