SEISMIC REFRACTION ON-SITE INTERPRETATION METHOD

by

Richard D. Markiewicz

November, 1989

ARTHUR LAKES LIBRARY

COLORADO SCHOOL oi MINES
GOLDBN, COLORADO 80401
 ProQuest Number: 10782776

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 10782776

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 48106 - 1346
T-3164

A thesis submitted to the Faculty and the Board of

Trustees of the Colorado School of Mines in partial

fulfillment of the requirements for the degree of Master of

Science, Geophysical Engineering.

Golden, Colorado

Signed:

Approved:
Thomas L. Dobecki
Thesis Advisor

Golden, Colorado

Head, Geophysics Department

ii
T-3164

ABSTRACT

This thesis presents a fast method for obtaining high

quality seismic refraction interpretations, called the

Seismic Refraction On-Site Interpretation Method (SROSIM).

The method consists of a series of steps which allow the

geophysicist to interpret refraction field data practically

as fast as it is collected. Benefits of the method

include: capability to recognize site-specific problems

such as lack of velocity contrast; freedom to modify field

data acquisition parameters during the field program;

completion of preliminary interpretation results before the

field crew leaves the site; and, most important, the

ability to spend more time thinking about the interpret­

ation and less time doing routine data processing tasks.

SROSIM consists of the following steps; forward model­

ing, data acquisition parameter design, field data collec­

tion, preliminary interpretation, and final interpretation.

The process is iterative in that preliminary interpret­

ation results for a seismic line are used as forward model

input for subsequent lines. Ideally, each line is

interpreted before shooting the next, allowing the geophys­

icist to vary data acquisition parameters on-site. In the

iii
T-3164

worst case, seismic lines are interpreted on a daily basis,

resulting in at most one day lost field time.

The essence of SROSIM is a rapid refraction interpret­

ation software package called RF. This package provides

the following capabilities: 1. A communications program

designed to accept data from seismographs supporting the

proposed SEG-1 format and several other formats, 2. A

trace display and picking program which allows interactive

or automatic first break picking, 3. An interactive-

graphics-based interpretation program that enables the

interpreter to assign arrivals to layers by moving a

graphics cursor on the computer monitor, and 4. A

refraction forward-modeling program allowing checks on data

acquisition parameters and interpretation uniqueness.

iv
T-3164

TABLE OF CONTENTS
Page

A B S T R A C T ...........................................iii

LIST OF F I G U R E S ...................... ... vii

LIST OF T A B L E S ............................... ix

ACKNOWLEDGEMENTS ................................ x

INTRODUCTION ................................... 1

SEISMIC REFRACTION ON-SITE INTERPRETATION METHOD . 4

Forward Model and
Data Acquisition ParameterDesign . . . . 9
Data C o l l e c t i o n .............. .................15
Preliminary Interpretation ................. 15
Final I n t e r p r e t a t i o n ............................ 17

SURVEY OF EXISTING ENGINEERINGSEISMOGRAPHS . . 19

Introduction ................................... 19
Abem T e r r a l o c ™ .................................. 20
EG&G Geometries E S - 1 2 2 5 ™ .........................22

THEORY ............................................23
Wave P r o p a g a t i o n .................................. 23
Interpretation Methods ......................... 26

PROGRAM D E S I G N ..................................... 39
Forward Model - R F M ............................... 40
Data Transfer - R F T r a n s .........................45
Trace Display and Picking - R F P i c k ................49
Interactive Interpretation - RF ............... 57
User I n t e r f a c e .................................. 63

E X A M P L E S ............................................65
Synthetic Model ............................... 65
Field E x a m p l e ..................................... 70
Benchmark T e s t .................................. 82

C O N C L U S I O N S ........................................ 88

REFERENCES CITED ................................ 90

V
T-3164

TABLE OF CONTENTS (cont'd)

Page

APPENDIX A USER'S MANUAL ...................... 92

APPENDIX B THE PROPOSED SEG-1 STANDARD . . . 134

APPENDIX C RF SYSTEM P R O G R A M S ................. 137

vi
T-3164

FIGURES

Page

Figure 1 - General Exploration Process . . . . 6

Figure 2 - RF S y s t e m ................................ 7
Figure 3 - A Seismic Refraction
Forward Model Example ............. 11
Figure 4 - Model Travel Time Curves,
Shots at 5, 125, 245, and 445 . . 12
Figure 5 - Model Travel Time Curves,
Shots at 5, 205, 325, 445, and 645 13
Figure 6 - Model Travel Time Curves,
Shots at 205, 405, 525, and 645 . 14
Figure 7 - Wave R e f r a c t i o n .........................24
Figure 8 - Direct and Refracted Arrivals . . . 25
Figure 9 - Time-distance Graph Showing the
Intercept Time and Critical Distance 28
Figure 10 - The Reciprocal Method ................. 33
Figure 11 - Travel time curve parallelism . . 34
Figure 12 - The Generalized Reciprocal Method (GRM) 37
Figure 13 - Three layer ray path model
referenced to the horizontal, , Æij,
i r e f e r e n c e d to layer normals . . 42
Figure 14 - Automatic First Break Picking . . .53
Figure 15 - Travel Time Curves, Palmer, 1980 . . 67
Figure 16 - Arrival Assignments for Figure 15 . . 68
Figure 17 - Depth Section .................. 69
Figure 18 - Black Mountain Seismic Line 1,
Location M a p .........................72
Figure 19 - Seismic waveform data transferred by
R F T r a n s ............................ 73
Figure 20 - Results of layerassignment,RFInterp 74
Figure 21 - Seismic depth section, 0 - 320 feet,
showing drill l o g s .................. 75
Figure 22 - Seismic depth section, 320 - 660 feet,
showing drill l o g s .................. 76
Figure 23 - Black Mountain Line 1,
Forward seismic model 1 . . .79
Figure 24 - Black Mountain Line 1,
Forward seismic model 2 . . .80
Figure 25 - Black Mountain Line 1,
Forward seismic model 3 . . .81

vii
T-3164

FIGURES (cont'd)

Page

Figure 26 - Representative interpretation times

using RF ...................... 86
Figure 27 - RF Organization chart ............ 96
Figure 28 - Scheme Used to Compute the Reciprocal
Time Using an Internal Half­
intercept T i m e ................ 128
Figure 29 - Configuration and Initialization
Procedure for R F ................ 133

viii
T-3164

TABLES

Page

Table 1 - Two-layer Travel T i m e s ................. 43

Table 2 - Three-layer Travel T i m e s .............. 44
Table 3 - .TRC Data File Format . . . . . 48
Table 4 - .PIK Data File F o r m a t ..................50
Table 5 - Slope and Amplitude Test Results . . . 55
Table 6 - Forward Model Parameters,Black Mountain
Line 1 ............................... 77
Table 7 - Benchmark R e s u l t s ........................ 83
Table 8 - Benchmark Times for Interpretation of
One S p r e a d ......................... 85
Table 9 - Proposed SEG-1 S t a n d a r d ................ 135

ix
T-3164

ACKNOWLEDGEMENTS

I would like to express many thanks to the following

people for their help in completing this work. Professors

Tom Dobecki, Phil Romig, and Walt Whitman at the Colorado

School of Mines; Andy Viksne, Bureau of Reclamation;

Ulrich Schimschal, Geodyne; David Butler, John Nicholl,

and Tom Baring, MicroGeophysics Corp.; Rob Huggins, EG&G

Geometries ; Jim Hunter, Energy, Mines, and Resources

Canada; Mats Langmanson, ABEM Geophysics; and finally to

my wife Nancy.

x
T-3164 1

INTRODUCTION

Although seismic refraction surveys are conducted

routinely for several kinds of engineering site investiga­

tions, methods used to collect and interpret seismic data

have not changed significantly in the last decade (Romig,

1986). Most refraction data interpretation is performed in

the office after the field crew has left the site, often

resulting in deficiencies in the amount and quality of

field data. Final interpretations derived from such field

data are therefore not as complete as possible.

This thesis presents a Seismic Refraction On-Site

Interpretation Method (SROSIM) which overcomes limitations

imposed by off-site interpretation. The technique uses a

rapid interpretation process to form preliminary results

on-site. These results are used to revise the field

exploration program as it is being conducted.

The core of the technique is the use of a refraction

interpretation software package called RF. This package

consists of six modules which are presented in four

programs. RFTrans is a communications program designed to

transfer data from engineering seismographs to computers

over an RS-232C port. RFPick is a trace display and

picking program which allows the user to make first break

determinations either automatically or interactively. The

T-3164 2

program RF contains three interpretation modules of the

package. These modules, RFData, RFInterp, and RFCalc,

enable the user to interpret refraction data with a minimum

of manual effort. The actual refraction interpretation is

performed by using interactive-graphics routines to assign

arrivals to layers. RFM is a forward-modeling program

allowing up to four non-horizontal planar layers. It

enables the interpreter to design field acquisition

parameters and check on interpretation limitations.

Using the RF package, a geophysicist could reasonably

interpret refraction data fast enough in the field to

modify the field program on at least a day-to-day basis.

This capability is important for several reasons. 1) Fast

interpretation methods allow the geophysicist to identify

site specific problems such as lack of velocity contrast or

thin layers. 2) By having preliminary results on a daily

basis, the interpreter can modify the field program to

optimize definition of the target(s). 3) Engineering

geophysics clients typically would like and often require

results closely following the field program's conclusion.

By using fast interpretation methods, the interpreter

can spend less time doing routine data processing tasks and

more time studying subtle subsurface problems that often

confuse the final interpretation picture. These problems

often are not resolved until extremely late in the project,

T-3164 3

if at a l l , due to the large volume of data processing

required before interpretive problems become obvious.

Within the SROSIM framework, these problems are forced to

appear more quickly, resulting in their resolution at an

earlier stage in the project.

T-3164 4

Seismic Refraction On-Site Interpretation Method

The Seismic Refraction On-Site Interpretation Method

(SROSIM) is a method of obtaining high quality refraction

field data resulting in geologically-sound final inter­

pretations .

SROSIM is part of an iterative general exploration

process consisting of the following steps:

1. Create a refraction forward model using available

geologic information.

2. Design the data acquisition parameters based upon

the results from step 1.

3. Collect refraction field data using the designed

parameters and optimized field techniques.

4. Interpret refraction data with the RF package.

5. If there are more seismic lines to shoot, or if

the model from step 1 is misleading or incorrect,
modify the geologic model based upon the results
of step 4 and go to step 1.

6. Complete final interpretation of all seismic

lines from the site.

This general exploration process might be viewed as

shown in figure 1. This figure indicates the iterative

nature of the process. If seismic refraction is an

appropriate solution to the exploration problem, then

SROSIM provides tools for obtaining the solution. The RF

system, discussed in detail later and shown schematically

T-3164 5

in figure 2, is presented as one software implementation of

SROSIM.

SROSIM is well-suited to engineering site investiga­

tion. Such investigations are almost always site specific,

as they are tailored to help design or modify a specific

man-made structure at a specific location. This is in

contrast to resource exploration techniques which prospect

for resources over a wide area. Using SROSIM results in

several advantages over traditional refraction exploration,

including knowing whether the exploration program is

succeeding and giving the engineer results more quickly.

A formal definition of the interpreter's role in the

refraction exploration process results from SROSIM.

Namely, the interpreter's job is to form an initial

geologic model of the site, determine acquisition para­

meters, and interpret field data. SROSIM allows the

interpreter to follow the exploration program flow at a

site without becoming overwhelmed by a large volume of

data.
T-3164 6

Program
design

Form
geologic
model

Choose
geophysical
method

Forward
model

Acquire
fi el d
data

In te rp re t
(p re lim .)

Survey
complet

Tee

Interpret
(final/

Figure 1 - General E x p lo ra tio n Process

T-3164 7

Field survey Forward model

program

RFTrans .TRC files — » RFPick .PIK files

!
/
.RFR files RF

Other site
Forward model Interpretation exploration

Final
Interpretation

Figure 2 - RF System
T-3164 8

In essence, SROSIM is a management technique. The

method gives the interpreter automation of seismic

refraction data reduction. The time thus saved is used to

make important interpretation decisions which require the

expertise and judgement of the human interpreter. SROSIM

makes crucial interpretation decisions the job of the

interpreter, not the computer.

Many of the tasks traditionally associated with

refraction interpretation are missing from the above

description. The interpreter should not have to spend time

on first break picking (other than to verify correctness),

plotting arrivals, calculating velocities or thicknesses,

or any other task which can be automated. By leaving the

often tedious data processing chores to machines, the

interpreter can concentrate on producing a high quality

final interpretation within time and budget limits typical

in engineering geophysics.

As artificially intelligent computers become capable

of more difficult tasks, the role of human interpreters may

eventually become that of validating geologic soundness of

interpretations.
T-3164 9

Forward Model and Data Acquisition Parameter Design

The geophysicist creates an initial seismic forward

model of the site by using available information such as

previous geophysical surveys, drill and trench logs, and

geologic maps. A number of modeling routines, including

RFM, can be used to create forward models (Spence, 1983 ;

MicroGeophysics Corp., unpub.). By examining a forward

model, the geophysicist can determine the following data

acquisition parameters: 1) the geophone spacing required

to define the target, 2) the location of offset shots to

profile deeper refractors, and 3) the length of seismic

line necessary to provide overlap on the refractor of

interest. The model should also give insight into possible

problems such as existence of hidden layers. In the

extreme case, the forward model may suggest use of an

exploration method other than a refraction survey.

As an example of the modeling phase of SROSIM,

figure 3 shows the postulated subsurface configuration for

a river channel carved in rock (V = 10000 ft/s) filled with

saturated alluvium (V = 5000 ft/s), overlain by unsaturated

alluvium (V = 2500 ft/s). Model travel time curves for

shots fired into a single 64 channel seismic spread are

shown in figures 4 - 6 . These travel times were generated

by the computer program ROUGH-86 (MicroGeophysics Corp.,

T-3164 10

unpub1.) and plotted with the commercial software program

Gremix (Interpex, Ltd., 1987).

Several features of the model travel time curves

merit close attention.

1. There is only a slight apparent velocity

contrast, 5000 ft/s versus approximately 7000
ft/s, between the saturated alluvium arrivals
(segment A, figure 2) and the down-dip portion of
the refracted bedrock arrivals (segment B ) .
Therefore, it is possible to mistakenly assign
the bedrock arrivals to the saturated alluvium
and make an incorrect interpretation. This
confusion is all the more plausible if record
quality is poor.

2. Sufficient overlap to map bedrock exists from the

shots at stations 0 and 640 (figure 3), so long
as the arrival assignments are correct. However,
offset shots, for example at stations -100 and
740, would help the interpretation considerably
by identifying bedrock refracted arrivals.

3. The 10-foot geophone spacing provides good

coverage on all apparent velocity segments. A
15-foot spacing would be sufficient for this
determination, while a 20-foot spacing is too
coarse to allow identification of short segments,
e.g. segment C in figure 4.
T-3164 11

o
o
Xo

o
.o
in

Example
o
.o

Model
</>

o -M
o o c

Forward
<4-
o o o
in o o
CXI in
o
ii o

Refraction
o
o
CXI

3 - A Seismic

o
.o
Figure

o o o o o o
cxi in co o

(U) Mldaa
T-3164
12

( S Q N C m S I T m ) 3WI1
o O O o o o
05 co CO CN T- O LD

-uU IIIII11iI11 O ej-

.1 1 I
o
*
o
c
m s_•
in +j
«O
o «
o CdM
-ufO
CO
s-
" 4-
LD CU
CM S-

<U
o * E
if) <u
o C O
□ m +J <U
m
a
o sz
tO 4-)
ce +->
^ o
§ | tz> S-

□
o ’*■ I
o

o O <u
s_
o fO
rn CU
E CO
%
in I
— T3
i—
c
<T3
O S«c
O fO
CM S- to
t
— 4->
r- OI
c
05 E
O "O
O OJ
CD
O Z= 1/5

CU
S-
TT r-(- i >i i | i i T1 «i111 11 * Z5
05
O o O o o o o O
05 GO r^ CO in '
*■ rO CN

( s a N o o a s m m ) 3 mii
T-3164 13

un
T
CD

(SQNOOHSinm) 3MI1 -a
c
ra
o o O 0 o O
O) 00 CD m rO CM en
11 » Iiiii1 1 I i i T
T
r-
en
CM
en
o
o LD
o
co CM

o ra
o
O•*a m en

^Lü
+->
O
jC
EX (Z)

§ S (Z
)
eu
>
S-
3
O
«o O 0 eu

1
o
rO t—

eu
>
ra
O s-
O
CM
eu
-a
o

»a O
O
eu
s-
3
CD
T T [iiii|iiiI|I
T I I I T I T T -I I I I I I I I I I I I II I I

O O O o o
O O o
en 00 f"» co m M" rO CM
(SaN033SmiM) 3WI1
T-3164 14

( s a N O o a s m m ) 3Mii c
in •!—
T 4-
o o o o o o o o kO <U
r>* CO in rO CN -a
05 00 ■a
II:Il tLIIII1II .iiIiiii
IIII1III1III O c -a
fO o
' ' :
O o
vn
cxi $-
LD O

-
o LO CD
o o c
0o CO T - •!—
-
u
o " (T3
LO Q .
O V)
a OJ
••oo o <U
o 4-> C
(O O
m
-
a •
• a V) CL
4-> O
01 o <u
x: CD
LU 00
a♦ -
o m <u
« a o 2
" in
in O
N*
3 CJ i—
- Z > u
s_
3 ra
z 0J

o o
O in
a) cu
rO E S-
<
t—3cr
- &) 1— cu
a
>> i-
-
O ra CJ
— O s_
CN t— +->
c
- 1— cu
CU E
-o CD
o <u
O z : oo
-o I
U3

CU
s_
» rj
CD
T T rr i i i i i i i i r i i i i i i i i i t i »i i i i i i i i I I I I | I 1 T'T

O O o o o o o o O O
05 00 co m N- rO CN

(saNooasmin) 3Mii
T-3164 15

Data Collection

By using field parameters based on modeling results

the geophysicist can expect to collect high quality field

data. This expectation can only be met if 1. the geologic

model is somewhat realistic, and 2. the field crew is

careful in their data acquisition technique (Knapp et al.,

1986). Good source and receiver coupling is especially

important when using a shotgun, mechanical source, or

sledgehammer, or when the site is inherently noisy, such as

near a powerplant or highway. If the site is windy, or if

the surficial soil is loose, geophones should be buried.

Site conditions may require filtering raw data before

digitizing the input signal, though recording raw data and

filtering later is preferable.

Preliminary Interpretation

Each stopping point in the field work provides an

opportunity to reduce raw seismic data. The stopping

point is not always the end of the field day, since other

pauses in field work, such as picking up and moving seismic

T-3164 16

cables, also give the interpreter time to transfer data

from the seismograph and make first break picks. By the

end of the field day, all "data reduction" is finished and

"data interpretation" can start immediately. A battery

powered laptop computer is a convenient means of transferr­

ing and picking raw seismic data.

Preliminary interpretation results are used to modify

the original subsurface model and possibly the data

acquisition parameters. The value of a preliminary

interpretation is the ability it gives the interpreter to

change acquisition parameters so that target(s) are mapped

effectively.

Often, drilling records and other site information

will not indicate existing lateral subsurface changes.

Detection of such lateral changes is one of the primary

reasons for performing seismic refraction surveys at a

given site. Lateral changes will impact final interpreta­

tion accuracy so it is important to find these changes

before shooting the next seismic line in an area. By

keeping the interpretation and subsurface model fairly

current with the field work, the geophysicist can spot

lateral changes and modify data acquisition parameters

accordingly. At most, one field day is lost because

seismic lines need to be repeated with different acquisi­

tion parameters.
T-3164 17

Final Interpretation

By the time all seismic lines are completed at a site,

the geophysicist will know roughly what layer velocities

and thicknesses are for the refractors of interest. The

final seismic data interpretation can then be completed at

the office.

The final interpretation should incorporate additional

information such as recent drilling results, a borehole

velocity survey, or results of other geophysical surveys.

Also, final sections should not show any discrepancies such

as mis-ties between lines. The final interpretation

should conform to known geology. It can and should be

used to guide the location of any additional subsurface

exploration. Due to the nature of refraction profiling,

thin layers and velocity inversions will not be detected

directly, in spite of careful planning or attention to

detail. Therefore, it is wise to drill at least one hole

at the site to confirm or correct the seismic interpre­

tation (Romig, 1986).

Forward models provide a means of checking final

interpretations for uniqueness. By creating a suite of

reasonable forward models based on interpretation results,

the geophysicist determines whether hidden layers can exist

T-3164 18

undetected by the survey. He can also gain insight into

maximum undetected thicknesses of hidden layers. Given

borehole logs and layer thicknesses, forward models

provide velocity ranges for undetected layers, as shown in

the field example.

T-3164 19

SURVEY OF CURRENTLY AVAILABLE ENGINEERING SEISMOGRAPHS

Introduction

Several manufacturers offer portable seismographs used

for refraction and shallow reflection surveys. This

chapter reviews the characteristics of two representative

seismographs capable of communicating with a computer

running RFTrans. This review does not constitute an

endorsement of these manufacturers' products.

An engineering seismograph usually includes 12 or 24

input channels, an equal number of amplifiers and filter

modules, a monitor screen, a hardcopy printer, and a data

output port, usually RS-232C or IEEE-488. Almost all such

machines also have a built-in or auxiliary data storage

device such as a magnetic tape recorder, floppy disk, or

non-volatile random access memory.

Engineering seismographs typically incorporate some

form of signal stacking. Stacking is the algebraic

addition of digitized time series to enhance the signal to

noise ratio. Truly random noise will eventually stack to

zero, while perfectly synchronous signals will stack as the

product of individual signal amplitude and number of

stacks. In order for signals to be synchronous, source

and seismograph must be mated to start simultaneously and

T-3164 20

repeatedly. Furthermore, spread geometry must not change

between stacked shots. The theoretical increase in signal

to noise ratio (S/N) is proportional to the square root of

number of stacks, (equation 1).

SN = si * /N (1)
Equation 1 - Increase in signal to noise
ratio with the number of stacks, where

Sjj = signal to noise ratio after N stacks

Si s signal to noise ratio for one trace
N = number of stacks

At the current time, such portable engineering

seismographs are produced by ABEM, Bison, EG&G Geometries,

OYO, and Scintrex. Other larger seismographs which may be

used for engineering applications are produced by AMF-

GeoSource, Input/Output, and Texas Instruments, among

others.

ABEM Terraloc™

The ABEM Terraloc is a 12 or 24 channel portable

signal enhancement engineering seismograph. In its

standard configuration, the Terraloc has 24 input channels,

T-3164 21

a CRT display, digital magnetic tape recorder, and an

internal battery (Atlas Copco ABEM, 1986). The system is

mounted on a pack frame, and weighs approximately 50

pounds without batteries.

The Terraloc offers the user a keyboard-style inter­

face to set recording and display parameters. In addition,

the user can define and save sets of data acquisition and

playback parameters. These parameters may be recalled from

non-volatile memory to quickly set the seismograph

configuration prior to a shot.

The seismograph has an 8 bit input digitizer for a

dynamic range of approximately 48 dB. Seismic signal

memory consists of 1000 samples per trace, with a 16 bit

word for each sample (Atlas Copco ABEM, 1986).

Data output is accomplished through an RS-232C or

IEEE-488 port. All data transmission parameters are user-

selectable. The Terraloc supports the proposed SEG-1

exchange format standard (Hunter, pers. comm., 1987). When

transmitting data, the Terraloc first transmits SEG-1

headers, then seismic data in binary or hexadecimal-coded

ASCII formats (Atlas Copco ABEM, 1986).

T-3164 22

Geometries ES-1225™

The Geometries ES-1225 is a 12 channel portable signal

enhancement engineering seismograph. In its standard

configuration, this instrument has 12 input channels, a CRT

display, and an external battery. The system weight is

approximately 30 pounds without batteries.

The seismograph has a keyboard and menu system for

entering recording and playback parameters. Up to 12 user-

defined instrument settings can be stored in non-volatile

memory. The 8 bit input digitizer has a dynamic range of

approximately 48 dB. Seismic signal memory consists of

1000 samples per trace, with an 8 bit byte for each sample,

(EG&G Geometries, 1987).

The ES-1225 will output data through an RS-232C port,

and supports the proposed SEG-1 exchange format standard.

When transmitting data the ES-1225 first sends SEG-1

headers, then seismic data as ASCII-coded decimal values

(Hunter, pers. comm., 1987).

T-3164 23

THEORY OF SEISMIC REFRACTION INTERPRETATION

Seismic refraction surveys have been conducted since

the 19201s. There is extensive literature describing the

method, its field techniques, and numerous interpretation

methods. This chapter summarizes major points found in

that literature.

Wave Propagation

Refraction surveys rely upon detection of seismic

waves which have bent (refracted) at interfaces of diff­

erent seismic velocity and have subsequently arrived at the

ground surface.

Snell1s Law

The physical principle which describes wave refraction

is called Snell's Law, a corollary of Fermat's Principle

which states that wave phenomena will travel such that the

travel time between two points is a minimum. Snell's Law

can be expressed as equation 2,

T-3164 24

sin i Vi
(2 )
sin r Vr

where i is the angle of incidence with respect to the

normal to a velocity interface, r is the angle of refrac­

tion, Vi is the seismic velocity in the incident layer

(layer 1), and Vr is the velocity in the refracting layer

(layer 2), as shown in figure 7.

Layer 1

Vr
Layer 2

Figure 7 - Wave Refraction

where:
i a angle of incidence
r a angle of refraction
V;
V^ a seismic velocity in Layer 1
v = e^ismic velocity in Layer 2
T-3164 25

The Refraction Head Wave

At the so-called critical angle for the underlying

high speed layer, a seismic wave is refracted exactly

along the interface between two layers (i.e. angle

r = 90*). Seismic waves travel in layer 2 with faster

speed Vr and in layer 1 with slower speed Vi. The

associated wave fronts in layer 1 lag behind those in layer

2 creating a "bow wave" effect (Heiland, 1963). A portion

of refracted energy, therefore, returns to the ground

surface where seismic instruments detect it (see Figure 8).

1. 2.

Figure 8 - Direct and Refracted Arrivals

The direct arrival at geophone 1 will be faster, while at
geophone 2 the refracted wave arrives sooner.
T-3164 26

Interpretation Methods

All refraction interpretation schemes use offsets and

arrival times as input. Offsets are source-to-receiver

distances. Arrival times are amounts of time taken by

seismic waves to travel from the source to each receiver.

First arrivals are used in standard refraction interpreta­

tion. At small offsets the direct wave from the source is

fastest, so the first arrival is the direct wave. At

increasing offsets, waves which have been refracted are

faster and are, therefore, first arrivals (figure 8). The

smallest offset at which the refracted wave is fastest is

called the critical distance and its value depends on

subsurface layer velocities and thicknesses.

There are several different methods of seismic

refraction data interpretation. Differences between these

methods result from different assumptions made about the

subsurface wave travel path. Required assumptions for

all standard refraction interpretations are: 1) the layer

velocities increase with depth, and 2) the contrast

between layer velocities is large (Sjogren, 1984).

Assumptions which distinguish various interpretation

methods are: 1) whether refractors are planar or undulat-

T-3164 27

ing, and 2) whether individual layer velocities are

constant or varying, (Palmer, 1980).

Most interpretation methods in use today are derived

from the reciprocal method or method of differences (Edge

and Laby, 1931; Hales, 1958? Hagedoorn, 1959? Hawkins,

1961? Palmer, 1980). These methods use arrival times from

waves which have refracted along the target layer.

Required are arrival times from shot points on opposite

sides of the geophone spread. Although there are minor

differences between these interpretation methods, they are

all fundamentally the same in that they use two emerging

rays from opposite ends of the seismic line to perform

depth and velocity calculations.

Another class of refraction interpretation schemes

uses the geometry of refracted rays and efficiency of roll

along shooting to form a matrix presentation of traveltime

data (Dobecki and Romig, 1985? Kuo, 1985). These linear

inversion techniques have not seen substantial use for two

main reasons: 1. reciprocal methods are widely accepted in

the profession, and 2. some of these inversion techniques

require significant computer time on a microcomputer.

Future advances in computer technology will probably make

linear inversion methods more attractive.

A third class of refraction interpretation methods

utilizes iterative ray tracing (Scott, 1977). Such methods

T-3164 28

form an initial subsurface model using the reciprocal or

similar method. Ray tracing techniques determine a set of

forward travel times for the initial model. Successive

refinement of the subsurface model minimizes differences

between calculated and observed travel times, and ultimate­

ly produces a final interpretation.

Intercept Method

Starting with the simplest and most restrictive

assumptions, namely that the subsurface consists of planar

layers with constant velocity, the intercept method (Ewing

et al., 1939) can be used. This method uses the so-called

intercept time for each refractor, which is the time

derived by extrapolating the refractor apparent velocity

segment back to the shot location (figure 9).

Intercept
time

x
^Crossover distance

Figure 9 - Time-distance Graph Showing the Intercept Time

and Critical Distance
T-3164 29

Overburden velocities are calculated as apparent

velocities of travel time curve segments. The harmonic

mean velocity provides a representative average layer

velocity when both forward and reverse shots are used

(Greenhalgh and Whiteley, 1977). A thickness is calculated

at each shot point and plotted vertically below the shot.

The interpretation proceeds as follows. Assign

arrivals to layers, forming straight line segments on the

travel time curves. Extrapolate line segments representing

layers back to the shot point location. Define the

intercept time for the i^*1 layer, as shown in figure 9,

as t £ . Then the normal refractor depth, h^, under the shot

is given by equations 3 (after Sjogren, 1984).

1 2 cos(i12)

h2 = t3 V 2 hl
2 cos (i23) cos (i23) sin (i12)

_ ^n V (n-1)
(n-1) 2 cos(i(n_1)n)

Vfn-1) k;n “2 h k cos dkn)

cos(i(n-l)n) k=l Vk

Equations 3 - Depth equations for intercept method

Where:
h = depth to top of layer
t s intercept time from travel time curve segment
V s layer velocity
i s critical refraction angle
T-3164 30

Critical Distance Method

The critical distance method (Heiland, 1963) is

similar to the intercept method in that it is applicable to

planar-layered problems with constant layer velocities.

The basis of this method lies in the observation that

at some offset from the shot, the refracted arrival from a

given layer will arrive before the direct or refracted

arrival from the overlying layer (figure 9). By measuring

this offset, (also called the crossover distance) and

knowing layer velocities from forward and reverse shots,

the interpreter can determine layer thicknesses for

materials overlying the refractor.

The interpretation process consists of assigning

arrivals to straight travel time curve segments similar to

the intercept method. However, instead of defining the

intercept time of a segment, the interpreter defines a

distance j , at which refracted waves from layer j arrive

sooner than refracted or direct waves from layer i.

Then normal layer depths under the shot are given by

equations (4) (Heiland, 1963 ? Sjogren, 1984). For dipping

layers, vertical depths under the shot are derived by

dividing normal depths by the cosine of layer dip.

T-3164 31

X 12 ~ s^n ^i2^ )
h
1 2 cos(i12)

X23 ^ ~ S^n ^23^ ^

h
2 2 cos(i23)
(4)

h
(n-1)

k=n-2
2
k=l

Equations 4 - Depth equations for critical distance method

Where:
h s depth to top of layer
x = critical distance from travel time curve segment
V = layer velocity
i = critical refraction angle

Reciprocal Method

When the subsurface is assumed to consist of nonplanar

layers, the reciprocal method (Hawkins, 1961; Edge and

Laby, 1931) is an appropriate interpretation tool. Redpath

(1973) discusses and presents an example of reciprocal

method interpretation technique.

T-3164 32

The reciprocal method requires arrivals from both ends

of the seismic line in order to determine refractor depth

(figure 10). The advantage to this method versus the

critical distance and intercept methods is that a depth can

be computed for each geophone which receives arrivals from

both directions of shooting. Such geophones comprise the

interval known as overlap on the refractor.

Interpretation follows the steps outlined below.

Assign arrivals to travel time curve segments which

correspond to subsurface layers. The arrivals need not be

in a straight line, although the least-squares-fit line

segment through the arrivals gives a representative layer

apparent velocity.

Planar layer boundaries produce straight travel time

curve segments. Non-planar boundaries produce arrivals

which lie above or below straight segments. Shots at

several locations along and off-the-end of the seismic line

produce travel time curve segments (figure 11) which

parallel each other when in the same direction and from the

same refractor (Sjogren, 1984). This "parallelism" helps

to identify the refracting layer for a given arrival.

T-3164 33

S-
s
i-
o
-M
□ U
ro
S-
4-
O
5-
en
4-1
-a
o o
- C 4-1
4-1
O en
r- CL
ro CI
U *o
o
S- 4—
GL O

u C n.
o O ro
VI cz:
4-1 $-
a (O O
C >
csa i— O
E
S- 4-
o Cl O
4-i
O Cl
o 13 C
s- o
□ 3 V) N
en S
o Cl
Lu J=.
4-1
ro
C
*o
O
JC en
4-1 a
<y c
csa E o
en CL
o
Cl
h - en
T-3164 34

«
g
i—

AT

AT

D istance
8 0

Figure 11 - Travel time curve parallelism

after Sjogren, 1984
T-3164 35

Based on assigned arrivals, select travel time curves

to maximize overlap (figure 10) on a given refractor. In

general, different shot pairs provide overlap on different

layers at any one geophone location. Calculate time depths

for each geophone which shows overlap on a given layer.

Then calculate normal depths as in equations 5 (Hawkins,

1961; Redpath, 1973).

Zi _ t2 * v 12

Z2 = (t3 - (Zi / V 13)) *V 23 (5)

Z3 = (t4 - (Z1 / V 14) -(Z2/ V 24)) * v 34

ti = (fcA + t-B ~ tag) / 2

Equations 5 - Layer thicknesses,

Where :
= Thickness of layer i
t^ =Time depth to layer i
t& =Arrival time at a geophone from shot A,
through layer i
tB s Arrival time at a geophone from shot B,
through layer i
tAB = Arrival time at A from shot B, through
layer i
Vi sVelocity of layer i
V-i sVelocity of a layer underlying layer i
Vij scos (sin-1 (Vi / V j ))
T-3164 36

Generalized Reciprocal Method

This method, known as GRM, is explained in detail in

an SEG monograph (Palmer, 1980). The GRM extends the

reciprocal method by using refracted arrivals which reach

different geophones (figure 12). This approach does not

assume that the refractor is planar between emergence

points, as does the conventional reciprocal method and

hence may be more "general". However, the GRM does assume

the ground surface between geophones used in the analysis

is planar. In many cases, this assumption may be as

restrictive as the conventional reciprocal method assump­

tion of planarity on the refractor.

In order to fully utilize the GRM, refraction data

must include a sufficient number of arrivals from each

layer. In practice, this condition is often not met.

T-3164 37

o
CL
(U
E
03
V)
<u

1^
sz

CD S-
-
O)
■O c
o
x: E
4-> O
<U s-
z= 4-
,__ < D
03 O)
U S_
o <u
s- E
Q. <U
u x: S-
<u u o
o: 4->
XI u
*o S 03
<u 5-
SZ

N V) 4-
<U
03 C.
03 S-
S- 0)
OJ *o XI
c <D 4-i
<u -M
CD U c
03 o
CU $-
XI 4-
1— <L)
I S_
C\J (/)
O)
tZ)
<u 13
s_
13 T3
^ 7

CD O
XI
Lu 4->
QJ
E
T-3164 38

Iterative Forward Modeling

Another method of deriving a final refraction

interpretation is based upon successive refinements of a

forward model (Scott, 1977; Haeni, et al., 1987). The

first iteration in this process is a subsurface model

derived through use of a conventional interpretation

method, such as the reciprocal method. Subsequent models

result from adjustments made to the initial model.

Successive adjustments are based upon differences between

observed travel times and synthetic travel times for rays

traced through the model. When synthetic travel times show

a high degree of consistency with observed time values, the

subsurface model is judged to be an accurate representation

of the subsurface.
T-3164 39

PROGRAM DESIGN

The RF package consists of four computer programs :

RFM, RFTrans, RFPick, and RF. RFM is a refraction forward-

modeling program used to help design field surveys and to

check interpretations. RFTrans is used to transfer

digitally recorded seismic data from a seismograph or tape

recorder to a computer. RFPick is used to display seismic

data on a computer monitor screen and interactively or

automatically pick first arrivals. RF is used to interpret

seismic data in terms of a layered earth with increasing

layer velocities. This layered earth may contain lateral

velocity variations and relief or undulations on layer

boundaries.

This chapter explains some design features of these

programs. For a detailed explanation of the programs' use,

see the user's manual in Appendix A.

T-3164 40

Forward Model - RFM

RFM models up to four planar layers with downward-

increasing layer velocities. The user supplies an earth

model through simulated drill log data, i.e. vertical

interface depth, drill hole location, and elevation.

Velocities are assigned to each layer by the user.

After checking model validity (increasing velocities,

non-negative layer thickness) RFM calculates travel times

for rays through each valid layer. The program writes

data files compatible with other modules in the RF system,

so the user can examine synthetic and field data with the

same programs.

RFM begins travel time determinations by calculating

critical, incident, and emergent refracted ray angles for

all valid layers. Then the program computes travel times

and ray incidence point locations for each overburden layer

except the layer immediately overlying the refractor.

Finally, RFM determines the remaining two-layer travel time

including the critically refracted path. This critically

refracted path must include a segment of at least zero

length on the refractor surface in order to produce an

arrival at ground level.

Requiring a non-negative refractor ray segment is

equivalent to determining the critical distance for the

T-3164 41

ray. The critical distance, i.e. minimum offset required

to produce anv arrival, is not to be confused with the

crossover distance, i.e. offset required for refracted

rays to arrive sooner than rays from an overlying layer.

The travel time determination method used in RFM is

iterative in that part of a four-layer calculation is a

three-layer calculation which contains a two-layer

calculation. RFM contains a general function to calculate

two-layer travel times given layer velocities, vertical

thicknesses, and incident and emergent point locations.

Three- and four-layer calculations do nothing more than

determine overburden travel times and then call the two-

layer function.

Figure 13 shows a schematic three-layer case. Tables

1 and 2 show the necessary arrival time calculations for

two and three layers. Four-layer travel times are similar

to three-layer times, with an additional overburden time

calculation.
T-3164 42

Figure 13 - Three layer ray path model

(Djreferenced to the horizontal ,
CL. Pjj ijj referenced to layer n o r m a l s .
T-3164 43

Table 1 - Two-layer Travel Times

See figure 13. Consider the lowermost two layers, i.e.

layers 2 and 3. The travel time from A to D is the two-
layer travel time.

Given:
A , A Coordinates of incident point
x' y
D , D Coordinates of emergent point
x' y
AF, DH Vertical layer thicknesses

Overburden and refractor velocities

V V3

Then:

abs (Dx - Ax )
AD =
COS#.

AK = AD cos (#2 - #g) = AD cos (<p^ - <p^)

AE = AF cos#. DG = DH cos#.

AE DG
AB = CD =
cosi, cosi.
23 23

EB = AB sini CG = CD sini
23 23

BC = AK - EB - CG

= AB + CD + BC
V.

Where:

0 . = dip of layer i
. -1
= critical refraction angle, sin (V2 / V^)
23
T = travel time

AE, DG s Normal layer thicknesses

T-3164 44

Table 2 - Three-layer Travel Times

See figure 13. The travel time from A1 to D1 is the three-

layer travel time.

Given:
A1 , D1 :Incident and emergent ray locations
x yy x ,y
A1F1, D1H1,
F1F2, H1H2 :Vertical thicknesses
Vi# V2, V3 :Layer velocities

Then:

^23 = i23 + (*3 " ^2^ 523 = i23 ~ ^ 3 ” <i>2 )

a13 = sin”1 (V1 sin^23/ V^) t 13= sin”1 (V1 sin<S23/V2)

COS0- COS0-
A1A = A1F1 --- — ~ - DID = D1H1
cosa13 cosr13

Ax = Alx + A1A sin(a13+ <p2) D^ = Dl^- DID sin(r13- <p2)

Ay = Aly - A1A cos(a13+ <p2) Dy = Dly~ DID cos(r13- 0 2)

A1A sin (a..+ <p )

AF = F1F2 +----- — T a Ï----- (H1H2 - F1F2)

DID sin(r - <p )

DH = H 1 H 2 -----— --------- -- (H1H2 - F1F2)

1 = A1A +DID + two-layer time (A to D, AF, DH, V 2 , V 3)

Where:
0^ s dip of layer i

ifj = critical refraction angle, sin 1 (V^ / V^ )

T s travel time

a, (3, 8 , t = incident and refracted ray angles, see

figure 13.
T-3164 45

Data Transfer - RFTrans

RFTrans is used to transfer seismic data from a

seismograph or digital tape recorder to a computer.

RFTrans reads data from seismographs supporting the

proposed SEG-1 format standard (Hunter, unpubl.), and from

several other seismographs using other formats.

The proposed SEG-1 header format describes a sug­

gestion for a standard data set to accompany engineering

seismic records (J. Hunter, pers. comm., 1987). Appendix B

describes the proposed SEG-1 standard. Several manu­

facturers offer seismographs which record and transmit

SEG-1 standard data.

From SEG-1 seismographs, RFTrans reads seismic data

and creates a seismic trace data file including a standard

SEG-1 header. From non-SEG-1 seismographs, RFTrans copies

as much information as possible from the seismograph

output, and writes a data file containing a SEG-1 header,

though this header is incomplete.

Output files produced by RFTrans all have the default

extension *.TRC1 to indicate trace data. The .TRC files

are written to disk and are later read and processed by

RFPick. The SEG-1 header, individual trace headers, and

T-3164 46

waveform amplitude for each sample of each trace are all

contained in trace data files.

RFTrans uses serial (RS-232C) communications to

transfer seismic data. The program sets communications

parameters on the computer serial interface board and then

controls the seismograph through data communications

lines.

Engineering seismographs typically have a storage

device such as a tape drive or floppy disk where seismic

trace data is stored. This trace data is then transmitted

through an RS-232C or IEEE-488 port to a computer or modem.

The computer (or modem) must be able to accept character

data through the port and control the sending device

(seismograph) when the receiving data buffer is full.

RFTrans controls seismograph output by sending an

XOFF character (ASCII 19 decimal, 13 hex) when the computer

receiving buffer is almost full. After a sufficient number

of characters in the buffer have been processed, the

program sends an XON character (ASCII 17 decimal, 11 hex)

to tell the seismograph to resume transmission.

The input data format from each type of seismograph

will vary according to manufacturer's designs. RFTrans

decodes and reformats these data so that .TRC files have

the same format regardless of seismograph used. Table 3

lists the .TRC file data format. RFTrans currently

T-3164 47

supports communications with the following seismographs :

ABEM Terraloc? Bison GeoPro? EG&G Geometries ES-1210F, ES-

1225, and ES-2415F; and 0Y0 McSeis-1500.

The user can adjust several parameters while running

RFTrans, including serial communications settings and

default data file drive and path.

T-3164 48

Table 3 - .TRC File Data Format

Variable Data Tvoe Comments

System Header

file_no string[80]
job_no string[80]
date_time string[80]
opers note string[80]
manuf_code string[80]
num_channels integer 12, 24
samples jper_channel integer 0..1024
sample_rate real in milliseconds
delay_time integer in milliseconds
highjpass integer in Hertz
low__pass integer in Hertz
notch integer 0 (out), 50, 60 Hz
shotX real x-coordinate
shotY real y-coordinate
shotZ real z-coordinate

Trace Header

gp_locX real x-coordinate

gp_locY real y-coordinate
gp_locZ real z-coordinate
gP_note string[80] operator's note
num_stacks integer
gp_gain integer in dB
waveamp array
[1..1024] waveform sample
integer

Note: There is one system header and 12 or 24 trace headers

per .TRC file. The trace headers contain the waveform data
in array 1waveamp1. One .TRC file with 24 trace data
occupies roughly 50kB.
T-3164 49

Trace display and picking - RFPick

RFPick reads .TRC files generated by RFTrans and

displays trace data graphically. The user can decide to

let the program automatically pick first breaks, or he can

make picks manually by moving a graphics cursor on the

monitor screen.

The user can set several parameters which affect the

size of displayed traces and accuracy of the automatic

first break picking algorithm. These parameters are trace

amplitude and starting and ending time for the trace

display, and amplitude threshold and slope value for the

auto-picker. The automatic picking algorithm is discussed

below.

RFPick writes data files with a default extension of

.PIK which contain first break pick values for traces from

one shot. The .PIK files are read by RF. Table 4 lists

the data format for .PIK files.

T-3164 50

Table 4 - .PIK File Data Format

Variable Data Type Comments

System Header

job__no string[80]
date t i m e string[80]
opers note string[80]
num_channels integer 12, 24
delay_time integer milliseconds
shotX real x-coordinate
shotY real y-coordinate
shotZ real z-coordinate

Trace Header

gp_locX real x-coordinate

gp_locY real y-coordinate
gp_locZ real z-coordinate
gp_note string[80] oper1s note
num_stacks integer
gp_gain integer in dB
pik array [1..24]
of real milliseconds

Note: There is one system header and 12 or 24 trace

headers per .PIK file. The trace headers contain
the arrival times in variable 'pik'.
T-3164 51

Interactive Picking

RFPick allows the user to display screens of eight

traces at a time from shot records. The user can scroll

through traces using cursor movement commands. The cursor

move commands are analogous to text editor move commands,

with the exception that the only picking options are insert

and delete.

The user moves a graphics cursor to identify first

break time. The cursor can be moved 1, 10, or 25 samples

at a time, or to the first or last sample. In addition,

the cursor can be moved up or down one trace, to the top or

bottom of the display screen, or to the first or last trace

in the shot record.

There are three insert and delete options for picking

arrivals. Two of these simply toggle the arrival time at a

specific trace to be either zero or the current cursor

time. The other option sets arrival time to current

cursor time then moves the cursor to the next trace so that

picking can proceed quickly.

Automatic Picking

In addition to interactive first break picking, the

user can choose to have the program automatically pick

first breaks. RFPick uses a method to detect first breaks

T-3164 52

based upon absolute value of waveform amplitude and changes

in this amplitude between adjacent samples (EG&G

Geometries, 1986). The user can set two parameters which

affect auto-picking results. These parameters are called

amplitude and slope. Amplitude is defined as the absolute

value of the waveform at any given sample. Typical

engineering seismic data has a range of 8 bits, or -128 to

+127, so the amplitude for such data ranges from 0 to 128.

Slope is defined as the absolute value of difference

between two adjacent samples.

To illustrate how the auto-picking routine works,

consider a seismic line recorded with a sample interval of

0.25 milliseconds and a refracted first arrival with a

frequency of 100 Hz (10 ms period) as shown in figure 14.

T-3164 53

en
c

j*:
u

_v:
ro
<U
s-
CÛ

«■

u
-M

—i H aidujBS
T h r e s h o l d

M.eajg isjij —
S l o p e

6> i o e g - - - - - - J ■
a s i O N p u n o j
H (U
S-
Z3
CD

Gaidures

! i i l l " 1
S 8 e

epniijduuv uujO}8Ae/v\ oiuusiag

T-3164 54

At 0.25 ms sample interval, there will be 40 samples

through the first full cycle of the arrival, or 10 samples

through 0 to 90 degrees in phase, which corresponds to the

interval between onset and first amplitude maximum. The

auto-picker looks at a given sample and nine samples

following. Conditions for labeling a givensample as the

first break are as follows:

1) The amplitude difference between adjacent samples

must exceed the slope parameter for any of five pairs
of samples following the given sample, or

2) The amplitude of any sample betweenfive and nine

samples ahead of the given sample exceeds the
amplitude parameter.

Consider two samples in figure 14, sample 3 and

sample 14. The results of slope and amplitude tests for

these two samples are shown in table 5. Neither the slope

nor amplitude threshold is exceeded for the window starting

with sample 3. Both thresholds are exceeded at sample 14,

however. Since sample 14 is the first sample in sequence

which exceeds either threshold, sample 14 is labelled the

first break.
T-3164 55

Table 5 - Slope and Amplitude Test Results

Slope test Slope Ampl. test Ampl.

_____________ samples sample

Slope thresh. = 5
Ampl. thresh. = 2 5

Sample 3: 3 & 4 2 8 0
4 & 5 2 9 0
5 & 6 2 10 -2
6 & 7 2 11 0
7 & 8 0 12 0

No thresholds exceeded.

Sample 14: 14 & 15 0 19 14

15 & 16 2 20 20
16 & 17 2 21 26 *
17 & 18 4 22 30 *
18 & 19 6 * 23 34 *

* Threshold exceeded.

In practice, noisier data requires a larger amplitude

threshold for accurate picking. Seismic data from

explosive sources allows a large slope threshold, while

data from mechanical sources requires a smaller slope

threshold in general.

The picking method used in RFPick assumes that the

refracted wave contains at least 10 samples from onset to

first amplitude maximum. This condition will be satisfied

for a maximum frequency of 100 Hz at a 0.25 ms sample

interval, for a maximum frequency of 200 Hz at a 0.125 ms

T-3164 56

sample interval, and for a maximum frequency given by

equation 6 below:

25 / f = I (6)

Equation 6 - Sample interval I in milliseconds needed for

auto-picking a maximum frequency f in Hertz.

At this point in the processing sequence, the user has

transferred seismic data from the seismograph, picked

first arrivals, and created input files for the interpreta­

tion program. Benchmark times for these steps as well as

for interpretation are discussed in the Benchmarks section

below.
T-3164 57

Interactive Interpretation - RF

RF is the interpretive portion of the package. It

allows the user to define segments of travel time curves

and assign these segments to subsurface layers. RF

consists of three modules, RFData, RFInterp, and RFCalc,

and a main driver, RFMain.

RFData performs two primary functions. One is to

accept manually entered project data, elevations, shot and

geophone locations, and arrival times. The other is to

read this information from disk files created by RFPick.

Manual data entry is easily accomplished by typing data

into pre-defined fields. Appendix A describes the use of

RFData in detail.

RFInterp graphically shows the travel time curves for

up to nine shots into a 1 to 24 channel spread. By using

pairs of reciprocal shots, the user can assign arrivals

(first breaks) to layers. Once all the layers have been

adequately defined, RFCalc is used to assign reciprocal

shot pairs, and calculate and display time and depth

sections.
T-3164 58

RFInterp

This module is used to assign arrivals to layers. The

arrivals and all other data pertinent to a single refrac­

tion spread enter the program through RFData.

The user can assign arrivals from each of nine shots

to one of four possible layers. The resulting groups of

assigned arrivals are called "model segments". A model

segment is characterized by the following: direction

relative to the shot (forward or reverse); apparent

velocity; intercept time at the shotpoint; constituent

arrival times. A model segment may consist of 1 to 24

arrival times, but all of these arrivals must have the same

direction relative to the shot. By creating 1 point model

segments, the user can constrain layer velocities, as

explained below.

There is also one more segment used by RF which is

characterized by the above parameters. It is called the

current segment. The only difference between the current

segment and model segments is the current segment may be

considered "interactive". That is, it can be created,

edited, and deleted. Model segments can only be created as

a copy of the current segment. The interpreter forms the

current segment on the monitor screen. When the inter­

preter is satisfied that the current segment accurately

represents a portion of a travel time curve, the current

T-3164 59

segment is added to the model and it becomes one of the

model segments.

In order to correctly interpret travel time data,

RFInterp must be used to define at least one forward and

one reverse segment for each layer indicated by the seismic

data. Of course, it is preferable to define as many

indicated segments as possible.

The arrival assignment method offers the user maximum

interpretation flexibility. RF allows the user to choose a

velocity for a segment by assigning only one arrival to the

segment and then inserting that segment into the model. In

this way, hidden layers may be incorporated into the model

with their greatest possible thickness, and hence, provide

the most conservative estimate for potential error in the

final interpretation.

RFCalc

RFCalc computes and displays time, velocity, and

depth sections. To form the time section, RFCalc first

tries to use overlapping arrivals from reciprocal shot

pairs assigned by the user. If there are not at least two

overlapping geophones, the program assigns reciprocal shot

half-intercepts to geophone locations closest to those

shots, and interpolates time depths for all geophone

locations between the two shots. Finally, if RFCalc cannot

T-3164 60

find a model segment for the given layer from these two

shots, time depths are interpolated from adjacent shot

pairs. In every case, RFCalc will compute, estimate, or

interpolate time depths for each defined model layer at

each geophone location. If RFCalc does not find any model

segments for a given layer for all shot pairs, it will

write an error message to the screen and go to the next

layer.

To find values for "true" layer velocities, RFCalc

uses a scheme similar to that used in forming the time

section. The user's reciprocal shot pair assignments are

used to look for overlapping geophones on the layer of

interest. If there are at least three overlapping

geophones on a layer between two reciprocal shots, then

RFCalc performs a velocity analysis function (Hagedoorn,

1959; Redpath, 1973; Greenhalgh and Whiteley, 1977; Palmer,

1980) on arrivals at overlapping locations as in

equation 7. The value derived from this velocity analysis

function is assigned to the model segment at geophone

locations between two opposed shots.

T-3164 61

t = (Tf - Tr ) / 2 (7)

V = dt / dx

Equation 7 - Velocity analysis function,

Where:
Tf s
travel time from forward shot
Tr =
travel time from reverse shot
t s
half-difference time
dt / dx =
partial derivative of t with respect to
distance along the seismic profile line
V s velocity from velocity analysis
function

If there are less than three overlapping geophones,

then the program uses the harmonic mean velocity (Heiland,

1963) of model segments from reciprocal shots. The

harmonic mean velocity is shown in equation 8 below.

2 V if V ir
Vh i = (8)
Vif + v ir
Equation 8 - Harmonic mean velocity,

Where:
Vhi 55 harmonic mean velocity of layer i
Vff s layer i apparent velocity from forward
shot
Vir - layer i apparent velocity from reverse
shot

If there are no reciprocal shots assigned for a given

layer, then RFCalc will write an error message to the

screen and continue with the next layer. Once layer

T-3164 62

velocities have been calculated at some geophone locations,

velocities at remaining locations are interpolated.

To form the depth section, RFCalc uses layer thickness

equations 5, presented in the reciprocal method discus­

sion.

If there are no time depths for a given layer, the

layer thicknesses are set to zero at all geophone loca­

tions, and the program writes a message to the screen. To

complete depth section calculations, RFCalc uses ground

surface elevations entered by the user as the top of

layer 1. Tops of remaining layers are obtained by

subtracting layer thickness from top elevation. This

process is correct for horizontal layers and approximate

for dipping layers.

T-3164 63

User Interface

The design of a clean user interface is as important

and often more critical to the acceptance of a computer

program than the "calculation" portion of the program.

Because the range of possible user responses is so much

greater than number of types of required calculations,

making a program easy to use requires several times as much

code as making a program do some task. The user must know

what to expect from a computer program in order to use it

quickly, accurately, and easily.

The interface for the RF package is a menu-driven

system at outer levels, and a command-driven system at

deeper levels. Typically, menu-driven programs are easy to

learn while command-driven programs are fast. The

interface used in RF has the advantage of combining

strengths of both systems. The user selects a program

option through one or more menus. Commands are used once

the user reaches a keystroke-intensive point in the

program.

The data entry modules in RF are designed to allow

fast and accurate data entry. All of the data entry

screens show length of data fields in inverse colors,

T-3164 64

allow user editing, show default values, and accept only

those values within a prescribed range.

The graphics interpretation and picking routines are

constructed analogous to a text editor. Cursor movement

includes incremental moves and jumps, and all cursor

movement wraps around the screen.

Finally, all file operations show default file names

whenever possible and always ask to overwrite a pre­

existing file.
T-3164 65

EXAMPLES

Two examples shown here demonstrate the RF system.

First is a synthetic example which shows general agreement

with other published examples. Second is a field example

compared with drilling results. Finally, representative

benchmark times show the possibility of using the RF system

on-site in a production setting.

Synthetic Model

As an example of data interpretation using RF,

consider Palmer's (1980) synthetic line 3, irregular

topography with irregular refracting surface. Travel time

curves for this line are shown in figure 15. Palmer's

synthetic data have been modified to include direct

arrivals through the 1500 m/sec layer, since these arrivals

appear when using a shallow source. Distance units have

been changed to feet throughout. Every third arrival from

the synthetic data is used to form a 24 trace spread.

(Palmer's original example contained 71 traces.)

Arrival assignments are shown in figure 16. As there

is a lateral velocity change at profile distance 173 in

Palmer's example, two forward and two reverse model

segments are used for layer 2, enabling a lateral velocity

change within the RF model.

T-3164 66

Figure 17 shows the RF-derived depth section along

with the synthetic model. Several points are significant.

First, only 24 of the original 71 geophone locations were

used in the interpretation. This sample of roughly one-

third of the data necessarily smoothes the final inter­

pretation. Still, there is good agreement between

interpretation and subsurface model. The maximum error

occurs at profile distance 50 at a vertical step.

Elsewhere errors average approximately 5%.

Second, a depression in the refractor at profile

distance 223 is not well defined in the interpreted depth

section. This is partly due to the smoothing effect

discussed above. The nature of refracted ray geometry also

contributes to a lack of definition in the bottom of 'V 1-

shaped features (Palmer, 1980).

Third, the depth section shown in figure 17 does not

extend beyond the coverage provided by two shots at

profile distance 0 and 350. Offset shots and phantomed

arrivals are required to extend refractor coverage.

T-3164 67

o
CO
CTi

S-
<D
E
<n
C l.

to
GJ
>
i
_
3
(_)
QJ
E

cu
L.
3
CD
T-3164 68

a
n a
S
□

cm

15
a a

for Figure
a

a a S
CD d
a

Assignments
ir> CD
re
a

i
a a
GB

16 - Arrival
CB
LTD

CD
Figure

CD CD

GB CB

CO
 o
UTI
«XI
T-3164

r ac
CD
CD

CD
oo

VI
cu
ca

Palmer's model is shown with the int erpretation

69
T-3164 70

Field Example

The following field example illustrates results

obtainable with the RF system. These data are part of a

refraction survey completed by MicroGeophysics Corporation

under contract to Bureau of Reclamation at the proposed

Black Mountain Pumping Plant site (Bureau of Reclamation,

unpub.). The survey objectives include the top-of-bedrock

configuration and representative bedrock velocities. These

objectives aid in estimating excavation quantities and

costs, and influence the structural design.

Seismic line 1 (figure 18) consists of three overlapp­

ing spreads of 24 geophones each. Between six and seven

shots per spread provide continuous coverage on the main

refractor (bedrock).

A laptop personal computer received trace data from a

24 channel engineering seismograph by use of RFTrans.

Figure 19 shows a small part of this data along with first

break picks, as displayed by RFPick. Figure 20 presents

representative travel time curves for this survey created

by RFInterp.

Figures 21 and 22 show the final depth section for

this survey, as produced by RFCalc, along with drill hole

results. The agreement between drilling and refraction

results is quite good at DH-103-SHPP. The agreement is

T-3164 71

questionable at DH-308-SHPP and DH-304-SHPP, and is

discussed in more detail below.

Generally, the seismic data indicate two layers

overlying the main refractor, assumed to correspond with

bedrock. The overburden velocities range from 1180 to

1310 feet/sec for the uppermost layer and from 3540 to

4490 feet/sec for the lower overburden layer. Bedrock

velocities range from 9249 to 10200 feet/sec. Indicated

depths to bedrock average roughly 40 feet.

At DH-103-SHPP, drilling results indicate a fresh

bedrock contact, logged as moderately to slightly weathered

(ibid.). Seismic results in this vicinity indicate good

velocity contrast. Agreement between drill and seismic

results is within a foot, or about 2%.

At DH-3 08-SHPP and DH-304-SHPP, drilling results

indicate a gradational bedrock contact, with intense

weathering and fracturing (ibid.). The USER fracture

density index, shown in figures 21 and 22 as "FD n"

indicates more intense fracturing with greater n values.

As seen in figure 22, a zone of intensely fractured rock

overlies relatively fresh rock in drill holes -308 and

-304.
755000
T-3164

/ÿ
8
E 755500
i .755750

Figure 18 - Black Mountain Seismic Line 1, Location Map

72
T-3164 73

co

by RFTrans.
19 - Seismic waveform data transferred
Picks marked by vertical bars.
c»

6— •

CO
Figure

CD

OO oa cu

ARTHUR LAKES LIBRARY

COLORADO SCHOOL oi MINES
GOLDEN. COLORADO 80401
 T-3164

Line 1 - Sprd 1 Dldtil FlUUIIldlll lUIllUUl rip Travel tine curves

CD
oa
i
oa
5
o

tn
on

Figure 20 - Results of layer a s s i g n m e n t s , RFInterp.

74
 T-3164

«M
OK

ZUI

CD
CD
9 Qd
s «

L Qd

C9
CO
co

CO

GB
<X3

GB

CD
so

OO
o
OJ
75

Figure 21 - Seismic depth section, 0 - 320 feet, showing drill logs.

 T-3164

m
VI
eu
ça

zui

CO

55
6 Od

1 Qd
£a
*h
e5 ®

!i:
CD

CD
oo

CD

m
s
o
OJ
Figure 22 - Seismic depth s e c t i o n , 320 - 660 feet, showing drill logs.
76
T-3164 77

In order to determine if this intensely fractured zone

could manifest itself as a seismic hidden layer, RFM was

used to create a suite of forward (synthetic) travel time

curves. Seismic refraction results and logs from drill

holes -308 and -3 04 provide input to the forward model

routine.

Figures 23 - 25 show four-layer synthetic travel time

curves obtained by use of RFM. Table 6 lists parameters

used in these models.

Table 6 - Forward Model Parameters, Black Mountain Line 1

At Station 520:

Laver Depth Velocity

(feet) (feet/sec)
1 0.0 1200
2 8.0 4250
3 23.1 5000, 6000, 70001
4 47.0 9240

At Station 620:

Laver Depth Velocity

(feet) (feet/sec)
1 0.0 1200
2 10.5 4250
3 25.0 5000, 6000, 70001
4 37.0 9240

1 Models 1, 2, and 3 respectively.

T-3164 78

Examination of the models shows the fractured zone

could exist as a hidden layer, so long as its velocity is

less than roughly 6000 feet/sec. At greater velocities,

the fractured rock would appear as a distinct segment on

travel time curves. Therefore, even though the fractured

rock zone did not appear on the travel time curves, one can

still estimate the excavation characteristics of this

material, knowing that its velocity cannot be greater than

about 6000 feet/sec.

T-3164 79

Travel tine curves

U
<U
co

ir >
•O
i— <d"
OJ OJ
□ -a cr»
cs o
oo E "a
c
LT» U ro
coo
•I-o
a>o
CO I f )

-a *
S-o
III Iflfil 1

\x
□ a
■XI
£
ro tn
3
S-^r
OJ

o
.
co "o
I—I OJ

■S**
r— I
OJ
VI C II

□ —I CO
cu
ro —
+-> U
c o
3i—
0 cu
z: >
$-
u cu
a —ro (
i >i
O
CQ —1

1
CO
OJ

cu
s-
V3 LT» 3
CO 03 CD
Ljne i
T-3164 80

«M

U
<u
Oj
to
4->
QJ
QJ
a4-
-a
C3 o O
E «3-
CXI
u O'!
E
co O
O
ao
COto
■a 9s
S- o
rO tn
3 CNJ
s
-«d
-
o
O
O

\ \
C3 ■ QJ
CXI

c II

—I to
QJ
c
•M
rd
4-> U
□ c O
3
O QJ
>

L
u QJ
ro>,
□ a CQ
to
1 _

«y
CXI

OJ
s_
3
CD
CO
 T-3164

Line i 1 9 1)!?! 3 Travel tine curves

:«Mi >
81
T-3164 82

Benchmark Test

RF is designed for fast interpretation in the field.

As such, benchmark times given here are an indication of

time required to interpret one spread. However, since RF

is an interactive program, the interpretation rate-limiting

factor is the human interpreter, not the computer.

Therefore it is difficult to arrive at a typical time

required for interpretation.

Another measure of interpretation speed is number of

computer keystrokes required to produce a final section.

Other factors being equal, an interactive program which

requires less user input will be faster.

Table 7 presents average times and keystroke counts

for steps necessary to produce a final depth section from

raw field data. Figure 26 shows representative interpreta­

tion times using RF.

T-3164 83

co CO
TJ ■o T3 T3 o
4)
o O o O O CO
u o o u (_
4» 4> 4* 4) <0 <0
c
O) o
CO CO c
u O C O c
JZ J=
■o c c_
Remarks

VI "□ co
0> 4> TJ o
41 4> C 4)
U U 4) CO
c U U a
3 3 3 41 co O
oc </> OO o a c
keystrokes
(min)
Time
Module

41 4) OJ
> E E C .c E >
O 3 CO a. u O 4) o co « co
CO E -Q
TJ co co >#• CO CO CO o CO o co
< CM <Z) 4) OJ OJ
<0 o a co co
C 4) TJ co JD CO CO -Q
4» co x : CO CO E E
TJ TJ o. 4J TJ Q. <0 o •*-> O
co CO CM C CD c_ co co co
t. O C u c_ 4) u 3 3 >.
4J 4» O) co 4» Ol co Ol CO Ol
o JO xz O 4» c c
E o co E E Ol Ol co
TJ C -X 3
Task

CO o OJ co C CO CO c c
<0 <0 o c_ *-> co o o o o CJ C
O 4) 'O 4J o 41 CO CO co
co o a *-> CO a. CL 3 a a. 3 CL a. E
T-3164 84

co
4-4
o
>»
co
co X X
6) «1 c o
E — o 3 3
Q. "e™ a a. 4-1
4-> « E 4-. o c c e
o a c E
o ♦<
co o
> C L. o 41 o
o a. co CO 73
3 3 3
0) U Ol co v co c_
-X CO • 73 *4 CO 41 4) 41 41
c c 4) E 4-4
m TJ E 3 C
E a. " c_
ti
oc
o> c_
OC H-
4) <*-
a o
CO
>
CO c_
< O.

co
e>
Table 7 (cont'd) - Benchmark Results

co
tt) Ol Ol
> co >
o CO > co
u
CO o o o
> r- o XI
« m
*

CO
c 41

E O) Ol
> CO >
CO > CO in in
4) CM CM
E
(XI m CM o o m

a
4> co 4» U o o u
3 (0 c co CO CO CD
TJ Q u o u U
O
X OC oc oc oc oc OC

o
-C
co >»
<0 o c
♦J 4-1 c o
(0 co o o c
"O co o o o
o o
CD o 41 41 4-*
o > a 41 CO > o
o co 41
c_ o » CO
o « 41 41
CO E a E -C
Q. 41
c <e c 73 *4 a
o> C- a> CO 41
_x 4) c_ 73 73 44 73
co 10 >L CO C C O
CO c CO CO CO CO
>— LU < -4 < OL a. o
T-3164 85

Assuming a single, 24 channel spread shot seven times

with varying offsets, benchmark times are as follows.

Table 8 - Benchmark Times for a Single Spread

Module Task Estimated Time

(min)

RFRes Load resident

drivers.

RFTrans Transfer 7 shot 26

records.

RFPick Pick 7 shot 10

records (auto­
pick option).

RFData Enter project

data and read
arrival times.

RFInterp Assign arrivals 15

to layers.

RFCalc Assign recipro­ 10

cal shot pairs,
find time
section,
calculate depth
section, plot
sections.

TOTAL TIME 64 minutes.

T-3164 86

Field
survey
nrnm

RFTrans .. ...... RFPick

10 m i n . 1
I .PIK file
shots 3.5 m i n / s h .IRC file
25 min.
1

RF
30 min (?)

.RFR file

Interpret 'n
approx.
65 min.

Figure 26 - R e p r e s e n t a t i v e interpretation times using RF.

T-3164 87

The total interpretation time for one spread is

roughly an hour. Transferring data from seismograph to PC

requires almost half that time. Transferring data during

the seismic survey, for example while the crew prepares the

next shot, reduces actual interpretation time to about 40

minutes. A laptop computer powered by batteries can

transfer seismic data successfully. It can also help

prepare preliminary field interpretations.

Seismographs which write data to disk directly in

MS-DOS format will enable future interpretation programs to

produce results much more quickly. Using seismograph-

produced trace data files, RFTrans is not required, and a

modified RFPick would accomplish any necessary re-formatt-

ing.
T-3164 88

CONCLUSIONS

SROSIM enables the geophysicist to create high

quality seismic refraction interpretations in a timely

manner. This is done by following a series of steps in the

general exploration process. The steps are: forward

modeling, data acquisition parameter design, field data

collection, preliminary interpretation, and final inter­

pretation. The process is iterative in that the results of

the preliminary interpretation can be used to revise the

geologic model and data acquisition parameters.

The RF system, a software implementation of SROSIM, is

a rapid means of producing seismic refraction data

interpretations on-site. The features which allow the user

to accomplish on-site interpretation are the following:

1. Straightforward program interface design providing ease

of use without compromising interpretation speed.

2. Flexible and robust refraction model allowing user-

editing of travel time curve segments in order to create a

geologically-sound subsurface model. 3. Ability to

incorporate hidden layers as ”one point segments", thus

providing these layers with their maximum possible

thickness and maximum potential for causing errors in the

final interpretation. 4. Ability to allow lateral velocity

changes and thinning to zero thickness within a seismic

T-3164 89

layer. 5. Ability to forward-model a given subsurface

configuration to check on data acquisition parameters and

the possible existence of seismic hidden layers.

Use of SROSIM allows the geophysicist to spend more of

his time resolving complex interpretation problems and

requires less time to complete routine data processing

tasks. SROSIM provides computer-aided interpretation

capability, without removing the human from the interpret­

ation process. All of the crucial refraction interpreta­

tion judgments are left to the human interpreter.

The benefits of SROSIM include the following. 1. The

method fits into a powerful, iterative, general exploration

process discussed above. 2. By reducing data reduction and

interpretation time, the RF system provides a form of data

management. This management derives from the interpreter’s

ability to maintain an overview of refraction survey

progress and preliminary results. 3. The RF system

produces results in the field quickly enough that each

day’s field work is reduced the same day. At most, only

one field day is lost due to incorrect acquisition

parameters. And perhaps most important, 4. SROSIM

identifies interpretation problems in the field, while

there is still time available to resolve these problems.

T-3164 90

REFERENCES CITED

Atlas Copco ABEM AB, 1986, Sales literature describing the

ABEM Terraloc™, ABEM Printed Matter No. 90248,
Broiruna, Sweden.

Bureau of Reclamation (unpub.), Geologic Design Data Report

for Black Mountain and Snyder Hill Pumping Plants,
Central Arizona Project, Arizona.

Dobecki, T.L. and Romig, P.R., 1985, Geotechnical and

groundwater geophysics, Geophysics, v. 50, pp. 2621-
2636.

Edge, A.B.B., and Laby, T.H., 1931, The Principles and

Practise of Geophysical Prospecting. Cambridge Univ.
Press

EG&G Geometries, 1987, Operator's Manual for the ES-1225

Seismograph. Sunnyvale, CA.

_______ , 1986, Source code for computer program "SeisView",

Sunnyvale, CA.

Greenhalgh, S.A., and Whiteley, R.J., 1977, Effective

application of the seismic refraction method to
highway engineering projects, Australian Road
Research, v. 7, pp. 3-19.

Haeni, P.P., Grantham, D. Ellesen, C. 1987, SIPT-1: A

microcomputer-based version of SIPT, U.S. Geol. Survey
Open File Report 87-103.

Hagedoorn, J.G., 1959, The plus-minus method of interpret­

ing seismic refraction sections, Geophysical Prospect­
ing. v. 6, pp. 385-394.

Hales, F.H., 1958, An accurate method for interpreting

seismic refraction lines, Geophysical Prospecting, v.
29, pp. 715-729.

Hawkins, L.V., 1961, The reciprocal method of routine

shallow seismic refraction investigations, Geophysics.
v. 26, pp. 806-819.

Heiland, C.A., 1963, Geophysical Exploration. Prentice-

Hall, New York, 1012 pp.
T-3164 91

Hunter, J. 1987, personal communication regarding the

proposed SEG-1 exchange format for engineering seismic
data.

_______ (unpubl.) Draft of ”A Recommended S.E.G. Engineer­

ing Seismograph Header Format Standard, SEG-1".

Interpex, Ltd. 1987, "Gremix" version 1.0, seismic

refraction interpretation software, written by Dr. c.
Stoyer, Golden, Colorado.

Kuo, S.Y., 1985, Using refracted data and linear inversion

technique to solve the long-wavelength static problem
in reflection seismology, Colorado School of Mines,
PhD Thesis 3 038.

MicroGeophysics Corporation, unpubl. "Rough-86", seismic

refraction forward modeling software, written by Dr.
D. Butler, Wheat Ridge, Colorado.

Palmer, D., 1980, The Generalized Reciprocal Method of

Seismic Refraction Interpretation. Society of Explor­
ation Geophysicists, Tulsa, 104 pp.

Redpath, B.B., 1973, Seismic Refraction Exploration for

Engineering Site Investigations. Explosive Excavation
Research Laboratory, Livermore, California, avail,
from National Technical Information Service, US Dept.
Commerce, Springfield, Virginia, 51 pp.

Romig, P.R., 1986, Introduction to Special Issue on

Engineering and Groundwater, Geophysics. v. 51, pp.
221-222.

Scott, J.H., 1977, SIPT - A seismic refraction inverse

modeling program for timeshare terminal computer
systems, U.S. Geol. Survey Open File report 77-3 65.

Sjogren, B., 1984, Shallow Refraction Seismics. Chapman and

Hall, Ltd., New York, 268 pp.
T-3164 92

APPENDIX A - USER'S MANUAL

RF is an interactive seismic refraction interpretation

software package. Its name comes from refraction. One of

its most important capabilities is enabling you to

interactively define segments of travel time curves on the

computer monitor screen. With RF you can create and print

velocity, time, and depth sections for a seismic line.

Data Input

The RF package will read data from a seismograph or

you can enter data through a series of user friendly input

screens. These screens allow fast data entry and editing.

Whenever possible, a default value is provided.

Graphics

RF uses graphics intensively. There are graphic

display options to draw travel time curves as well as the

calculated velocity, time depth, and elevation for each

layer versus geophone position. The actual ”interpreta­

tion" phase is performed by assigning arrivals to layers

with a graphics cursor which you can move at will.

T-3164 93

Hardware

The RF package can drive almost all popular video

cards, and approximately 20 dot matrix and laser printers.

To run RF, you will need an I B M ™ -PC, -XT, or -AT

compatible computer, MS-DOS 2.0 or greater, 420 kB of free

RAM, and two floppy disk drives or a hard disk.

T-3164 94

Format of this Manual

The commands which you will use to run RF are explain­

ed in this manual. When you are required to press a key in

order to do something, that key will be printed in bold

type, for example "type A to enter the Arrival Times

option".

Sample screens for various stages in the program will

be accompanied by text explaining that particular screen.

The screens will correspond to menus that you will see as

you work through the program. The order of topics in this

manual roughly follows the flow of the program as you

interpret seismic data.

T-3164 95

Program Organization

The RF package consists of the following modules :

RFM, RFTrans, RFPick, RFData, RFInterp, RFCalc, and RFMain.

RFM, RFTrans and RFPick are self-contained programs. The

other modules are combined into one program called RF.

When you first enter RF, the RFMain module is active. From

the main module you may choose one of the other three

modules or you may Quit and go to DOS. Figure 27 shows the

organization chart for RF.

RFM

RFM is a refraction forward-modeling program allowing

up to four planar layers. The user enters an earth model

by specifying vertical depths at two locations, analogous

to drill log data. By then specifying assumed layer

velocities, the earth model is complete.

RFM will calculate travel times for each valid layer

and write its results to a data file compatible with the

other RF system programs.

T-3164 96

/
RFData
DOS

Init
RFInterp
environment

W e lcome RFCalc

menu Configure

User
choice Exit Cleanup DOS

Figure 27 - RF O r g a n i z a t i o n Chart
T-3164 97

RFTrans

RFTrans establishes communications over an RS-232C

port with engineering seismographs including the ABEM

Terraloc™, Bison GeoPro™, EG&G Geometries ES-1225™, ES-

2415F™, and ES-1210F™, and 0Y0 McSeis 1500™.

RFPick

RFPick is an interactive trace display and picking

program. You can rapidly determine first breaks with

RFPick or allow the computer to make the picks automati­

cally.

RFData

RFData allows you to enter information specific to a

single seismic line and save this information to a disk

file. You can record the field parameters such as geophone

location, elevation, and spacing. You can also record

comments such as line location or unusual conditions.

RFInterp

RFInterp is the interactive interpretation portion of

the package. By moving a graphics cursor over selected

travel time curves, you can define segments on these

T-3164 98

curves. Once you have defined enough segments to produce a

model of the subsurface, you can use RFCalc to calculate

various parameters of that model.

RFCalc

RFCalc performs calculations necessary to form the

time, velocity, and depth sections of your subsurface

model. You will also use RFCalc to match reciprocal shot

points and choose end-to-end (reciprocal) times for those

matched shots. RFCalc will produce plots of the various

sections on the computer screen, which you can then print

with a printer.

Data Files and Filenames

RF system programs write data files in binary-format.

The files storing raw seismic data created by RFTrans have

a '.TRC1 extension. RFPick reads '.TRC* files and creates

’.PIK* files. RF can import '.PIK* files, and stores the

interpretations in 1.RFR* files. RFM creates .RFR files

with synthetic seismic data.

The .RFR files can contain all entered data as well as

the interpretation and subsurface model. You can go back

at any time and reinterpret or edit your data by reading

the .RFR file for a particular seismic line.

T-3164 99

Seismic Refraction Model

RF uses a subsurface model to interpret seismic

refraction data. Although this model is fairly general, it

has some constraints. As with all refraction models, the

layer velocities are assumed to increase with depth.

Subsurface layers are assumed to have constant velocities

between reciprocal shot points, though the user can select

the shots. Layer thicknesses may vary from zero to

thousands of feet, but wildly dipping layers will be

smoothed by the interpretation process.

The numerical subsurface model used by RF consists of

the following parameters at each geophone location:

• ground surface elevation

• top-of-layer elevation
• layer thickness
• apparent and true velocity
• time depth

RF also treats the interpreted travel time curves as a

numerical model. The model consists of segments. There

are forward and reverse model segments for each shot and

for each of the four allowable layers. Each model segment

is characterized by its direction relative to the shot, its

apparent velocity, intercept time at the shotpoint, and

component arrival times. A model segment may consist of 1

T-3164 100

to 24 arrival times, but all of these arrivals must have

the same direction.

There is also one more segment used by RF, called the

current segment. This is the segment that you as the

interpreter create on the screen. When you are satisfied

that the current segment accurately represents a portion of

the travel time curves, you add the current segment to the

model and it becomes one of the model segments.

T-3164 101

Input Editing Commands

When you use one of the RF options, you will see

highlighted fields on the screen. The length of the field

tells you how many characters are allowed to fill that

field. For example, a numeric field may be 8 characters

long, while a comment field may be 40 characters long.

Whenever possible, a default number, filename, or option

appears in the field to be edited. If you like the

default, press Return and RF will use it.

If you want to edit the field you are working on, use

the editing commands described in detail below.

Left Arrow Moves the cursor one space to

the left. By typing any
character, you write over the
old character at that cursor
position.

Right Arrow Moves the cursor one space to

the right. Again, typing any
character writes over the old
character.

Up Arrow Moves the cursor to the

previous field. If the
cursor was on the first
field, it will next move to
the last field on the screen.

ARTHUR LAKES LIBRARY

COLORADO SCHOOL of MINES
GOLDBN, COLORADO 80401
T-3164 102

Down Arrow Moves the cursor to the next

field on the screen. If the
cursor was on the last field
on the screen, you will leave
the screen.

Spacebar Writes a blank space over the

character at the cursor.

Delete Removes a character.

Insert Shifts the characters to the

right of the cursor over by
one position. Any characters
on the extreme right of the
field will be lost.

Return Causes the cursor to move to

the next field.

Escape Lets you leave the screen

entirely, saving the contents
of all the fields in memory.

A summary of these editing instructions are shown at the

bottom of each input screen.

T-3164 103

RFM

RFM creates refraction forward model data from an

earth model input by the user. The details of travel time

calculation are presented in the Program Design section.

The options available in RFM include model definition,

travel time calculation, and file handling. The RFM main

menu screen is immediately below. Options for program

operation follow the screen.

RFM 1.0 Main Menu >> Choose an option below by typing the Bold letter.

drill Information

Velocities

spread Geometry

Travel times

Save .RFR file

Directory of the
data disk

X - drive\path of the
data disk

eXit and go to DOS

Choice:
T-3164 104

drill Information

Use this option to define subsurface layer

geometry. You can enter vertical depths (drill
depths) for tops of three layers beneath the
ground surface. This gives a total of four
layers. In addition, you can give the ground
surface a slope by entering different drill hole
elevations.

Velocities

Enter a velocity for each model layer.

Velocities must increase with depth, so the third
layer velocity is greater than the second layer
velocity, and so on.

spread Geometry

You can position the shot and geophone

spread anywhere in a line with the two drill
holes. Enter a location for the shot and first
geophone. For reverse spreads, locate the shot
at a greater line location than the last geophone
in the spread.

Travel times

Once you have defined the earth model, use

this option to generate synthetic travel times.
RFM will check each model layer for increasing
velocity and non-zero layer thickness, and
calculate as many travel times as possible.

X set data drive\path

Use this command to tell the program where

to write data files. The program does not check
for valid path names, though, so be sure of the
path you want the program to use.
T-3164 105

Save .RFR File

Use this option to save the seismic data in

a binary-format file. Similar to the Read option
described below. Save will prompt with any
available default. Use Esc to cancel the save.
If your save request will result in a file
being overwritten, you are asked permission to do
so. Answer Yes at this point to write over the
old file, No to change file names, or Esc to
cancel the save operation.

Directory of data disk

Use D to obtain a directory listing of the

disk specified in the data path. If no path is
specified, the default drive will be used.

eXit and go to DOS

When you are done, leave RFM with eXit.

T-3164 106

RFTrans

RFTrans is a communications program designed to

transfer seismic data through an RS-232C port to a personal

computer. The program will support communications at rates

up to 9600 bits per second with many types of engineering

seismographs, especially seismographs which transmit data

according to the SEG-1 standard format.

Using RFTrans is simple because there are only a few

items to choose before beginning the communications. The

main menu screen for RFTrans is shown below. Each choice

is discussed in detail following the screen.

RFTrans 1.0 Main Menu >> Choose an option below by typing the Bold Letter.

P - set the serial S - Save trace data

communications Parameters

1 - Transfer data from the D - Directory of the

seismograph to the computer data disk

C - Choose your seismograph X - drive\path of the

data disk

X - exit and go to DOS

Choice:
T-3164 107

P - set the serial communications Parameters

This screen is shown below. Use this option

to set the specific parameters such as data bits,
stop bits, baud rate, and parity. A code number
appears next to each parameter on the menu. If
you want to change the baud rate to 9600, for
example, type 7 in the field labeled baud rate.
Press Esc when you are done.

rans 1.0 - Set the communications parameters by typing the code number
next to the value you want. Press <Esc> to leave

Baud Rate 110 150 300 600 1200 2400 4800 9600
Code number 0 1 2 3 4 5 6 7 Baud: 7

Parity bits None Odd Even

Code number 0 1 2 Parity: o
Data bi ts 5 6 7 8
Code number 0 1 2 3 Data: 3
Stop bi ts 1 2
Code number 0 1 Stop: 1

Port C0M1 COM2

Code number 1 2 Port: 1

C - Choose your seismograph

The communications parameters have been pre­

set for a number of seismographs. If you see
your seismograph on the menu, type the number
that is next to its name. The parameters for
that particular machine will be used.

T - Transfer data from the seismograph

Once you have chosen the communications

parameters or the seismograph type, you can
T-3164 108

transfer data into the PC. The program will ask

to make sure everything is ready, then will try
to transfer data. RFTrans will write messages to
the screen to keep you informed of its progress.
After each shot record is transferred, the
program will ask you for a file name to save the
data. If you do not want to save the file, press
Esc. You can save the file later with S if you
change your mind.

S - Save .TRC file

RFTrans will also let you save files

explicitly with the S command. This command is
useful if you are experimenting with communica­
tions parameters and want to transfer files one
at a time. Type a file name at the prompt or
press Esc to cancel the save.

\ - Set the Drive and Path of the Data Disk

Use this command to tell the program where

to read and write files. The program does not
check for valid path names, though, so be sure of
the path you want the program to use.

D - Directory of the data disk

Use D to get a directory listing of the

disk specified in the data path. If no path is
specified, the default drive will be used.

X - Exit to DOS

Use this command to leave when you are done.

T-3164 109

RFPick

RFPick is the trace display and picking program. It

allows you to determine the arrival time of the first break

(or any other event) to within one sample interval. You

can move a graphics cursor on the screen and read corres­

ponding time. The cursor can be moved in increments of 1,

10, and 25 samples ? to the first or last sample? up or down

1 trace ? to the top or bottom of the screen? or to the

first or last trace. Displayed traces can be scrolled with

an adjustable trace overlap.

RFPick also offers automatic trace picking with

adjustable picking parameters. You can select the ampli­

tude threshold and the difference threshold between

adjacent samples. You can edit all picks, both on a user

input screen and on the graphics screen with the movable

cursor.
T-3164 110

The main menu screen for RFPick is shown below. Each

choice is discussed in detail following the screen.

RFPick 1.0 Main Menu >> Choose an option below by typing the Bold letter.______

Traces - plot waveforms and interactively pick

Parameters - set the plot scales

Auto-pick

Edit picks

set data drive\path

Save .PIK file

Read .TRC file

Directory of data drive

eXit to DOS
Choice: _

T - display the Traces

This option displays eight traces at a time

on the screen and allows you to move a graphics
cursor over the traces and enter or edit picks.
Within this option, there are several commands
available for interaction with the cursor. They
are described below.

Left,Right Arrow - moves the cursor by one

sample.

Ctrl Left, Ctrl Right Arrow - moves the cursor by

10 samples.

Tab - moves the cursor by 25 samples.

T-3164 111

Home, End - moves the cursor to the first, last

sample.

Up, Down Arrow - moves the cursor by one trace.

Ctrl Home, Ctrl End - moves the cursor to the

first, last trace.

PgUp, PgDn - displays the previous, next screen­

ful of traces, overlapping by the number you
selected in Parameters.

Ctrl PgUp, Ctrl PgDn - Displays the first, last

screenful of traces.

Two commands allow you to change the Y-scale

on an individual trace independent of the other
traces. (The X-scale is set for all traces with
the Parameters option.)

+ (plus) - makes the trace appear twice as large.

- (minus) - makes the trace appear twice as

small.

There are also commands to enter and edit

the first break picks. These are described
below.

Enter - makes the current cursor location the

pick and moves the cursor to the next trace down.

Ins - makes the current cursor location the pick

but does not move the cursor.

Del - deletes the current pick but leaves it on

the screen for reference. The cursor does not
move.

To leave the graphics display press Esc.

Your picks will remain in memory so you can Save
them.

P - choose the plot Parameters

Use this option to set the initial scales

for the trace display and the first break picking
options. You can window one portion of the
T-3164 112

traces by choosing the starting and ending times.

The Y-scale is set by choosing a maximum ampli­
tude for the plot. (The amplitude is also
adjustable once you enter the interactive
graphics portion of the program.) Using this
screen you also set the amplitude threshold and
slope (difference between adjacent samples).

A - Auto pick the traces

RFPick will automatically pick the traces

using the picking parameters you set. The picks
can be displayed with T (Traces). The amplitude
threshold should be roughly five times the size
of the slope threshold for optimum picking
accuracy.

E - Edit the picks

You can edit the picks on a user-input

screen if you don't need to see a graphics
display of the traces. Use the editing commands
discussed in the INPUT EDITING COMMANDS chapter.
This editing method is faster than the graphics
display method.

\ - Set the Drive and Path of the Data Disk

Use this command to tell the program where

to read and write files. The program does not
check for valid path names, though, so be sure of
the path you want the program to use.

S - Save the .PIK file

Use this option to save the picks to a file.

You will be asked for a file name and prompted
with the name of the last '.PIK1 file saved.
Press Esc to cancel the save.
T-3164 113

R - Read .TRC file

Use this option to read one file of trace

data (one shot record) as produced by RFTrans.
The trace data file contains the sample interval
and other information to accurately plot the
data.

D - Directory of the data disk

Use D to obtain a directory listing of the

disk specified in the data path. If no path is
specified, the default drive will be used.

X - Exit to DOS

Use this command to leave when you are done.

T-3164 114

RFData

RFData is the "bookkeeping” portion of RF. You will

use RFData to enter the locations and elevations for the

geophones and the shot points, the arrival times, and other

notes and comments specific to a single seismic line. Enter

the RFData module by typing D from the RF Main Menu. The

main screen for RFData is shown below. Each option is

discussed in detail following the screen.

RF 1.0 - RFData >> Choose an option by typing the Bold letter.

Data Input Files

Arrival times Read .RFR file

Locations - shot & geophone Save .RFR file
Elevations - geophone set drive\path for data disk
Project data Directory of data area

Data Adjust Import

Uphole time rfpicK files

Multiply by scale factor

Data List Exit

prinT eXit - go back to RF menu

Choice ?
T-3164 115

A - Arrival Times

Use this option to enter the times from each

shot, 1 through 9, for which you have entered or
will enter a shot location. After you type A to
select the Arrival Times option, type the number
of the shot for which you want to enter times.
Type in an arrival time for each geophone
which showed a first break for the current shot.
Press Esc when you are finished entering times
for this shot. You may edit, add, or delete
travel times for any shot in the future if you
want, so you do not need to enter all of the
times at one sitting.

L - Shot and Geophone Locations

Use this option to establish the in-line

coordinates for the shot points and the Number 1
geophone. Also use this option to type in the
geophone spacing. Use a positive (> 0) geophone
spacing if geophone 1 is at a lower in-line
coordinate than geophone 24. Use a negative (<
0) geophone spacing if geophone 1 is at a higher
in-line coordinate.
Note that RFInterp will always plot from
lower to higher in-line coordinates, so by using
a negative geophone spacing, you can choose to
plot the arrival times from right to left.
Also notice that RF uses a constant geophone
spacing. For those of you who are used to
shooting "tapered spreads" this may appear to be
inconvenient. There are two ways to handle this
situation. The first is to enter one or two
arrival times at the constant-spaced geophone
locations, by knowing the approximate arrival
times from the tapered portion of the spread.
The second and preferred, is to shoot shorter,
constant-spaced seismic lines which give far
superior overburden velocity control.
T-3164 116

E - Geophone Elevations

Use this option to enter the elevations at

each geophone in one of several ways. You can
enter each elevation manually by editing the
elevation fields. This method is not much fun.
It should only be used if the ground surface is
very irregular.
Alternatively, you may also enter one
constant elevation for the whole line if you shot
the line on level ground. The last and most
powerful method is to enter a few topographic
break points (with non-zero elevations) and let
RF interpolate between the break points. This
gives you more flexibility and speed if you have
reasonable topographic control for the site.

P - Project Data

Use this option to enter and save notes and

documentation for each seismic line. Edit the
Project Data fields just as you would edit the
numeric fields for the other options. The
available notes include the name and state of the
project site, the line and job numbers, the field
date, the operators, and two lines of general
comments. These notes are optional; they are not
used in the interpretation, except that the plot
and printout titles are taken from the project
name and the line number.

U - Uphole Time

Use this option to add a constant time to

each of the 24 arrival times for a given shot.
You will be asked for a shot number and then an
uphole time. Uphole times may be negative so
that you can undo a previous experiment. As
always, you may Escape out of the option at any
time.
T-3164 117

M - Multiply by scale factor

You may multiply the arrival times or the

geophone spacing by a constant real number. This
capability is useful if, for instance, you have
picked all of your seismograms incorrectly and
the travel times you entered are all too great or
too small by some factor. It is also useful if
you need to enter the shot and geophone locations
in feet or meters and then change scales to the
other system.

T - Print

Use this option to print the field and model

data. Make sure your printer is on line and
ready, then press Return, or press Escape if you
change your mind.

R - Read .RFR File

Use this option to read the binary-format

seismic data files which RF generates. When you
type R, RF will ask for a file name, prompting
with any available default. If you press Esc at
this point the read operation is cancelled.

S - Save .RFR File

Use this option to save the seismic data in

a binary-format file. Similar to the Read option
above, Save will prompt with any available
default. Use Esc to cancel the save.
If your save request will result in a file
being overwritten, you are asked permission to do
so. Answer Yes at this point to write over the
old file, No to change file names, or Esc to
cancel the save operation.
Set the Drive and Path of the Data Disk

Use this command to tell the program where

to read and write files. The program does not
check for valid path names, though, so be sure of
the path you want the program to use.

Directory of the data disk

Use D to obtain a directory listing of the

disk specified in the data path. If no path is
specified, the default drive will be used.

Import .PIK files

Use this option to import first break pick

files created by RFPick.

Exit (Back to RF Menu)

This option takes you to the RF main menu.

From there you may select another module
(RFInterp or RFCalc) or Quit and go to DOS.
T-3164 119

RFInterp

The RFInterp module allows you to draw travel time

curves on the screen, to compute apparent velocities of

line segments, and to assign specific arrivals to one of

four possible subsurface layers. The main screen for

RFInterp is shown below. Each option is discussed in

detail following the screen.

RF 1.0 - RFInterp >> Choose an option below by typing the Bold letter.

Travel-time curves Files

Any - Draw any number of Read .RFR file
shot records. Save .RFR file
(No layer assignment) set drive\path for data disk
Directory of data area
Two Draw one or two shot
records and a half­ Model
difference plot. reinitialize - Set veloc's,
(Assign arrivals to layers) thicknesses, depths
to zero.
Params - Choose the plot
size, x- and y- eXit - go back to RF menu
scales.

Choice ?
T-3164 120

A - Draw any number of travel time curves

Use this option to produce a quick summary

of any or all of the shots on a single seismic
line. Each shot number and location is shown
along with a Y/N option next to it. Type Y if
you want to display that particular shot. When
you are done, press Esc and a plot of the
selected shots will be produced. You are given
the option of printing this screen on your
printer. By using the scale parameters
discussed below, you can zoom in on one portion
of the graph or step back to see the larger view.
Since the shot locations are shown on the
Select Shots screen, this is a good way to check
that all of your shot locations are correct. You
can easily generate a travel time graph with all
of the shot records shown at once.

T - Draw any Two travel time curves

This option is where you will spend most of

your time in RFInterp, as it is the option which
allows interactive graphics interpretation and
layer assignment of the arrival times. As the
name of this option implies, Two uses pairs of
travel time curves. Usually these are curves
with a forward and reverse segment over some area
of interest. The curves need not overlap on a
refractor, but of course, if they do, the
interpretation is much stronger.
The pairs of travel time curves are referred
to as 'A* and 1B * . No sense of direction is
implied, and both shots may be interior split
spread shots containing forward and reverse
segments.
Two plots the 'A* and 'B* shots along with
the half-difference plot for the two shots. The
half-difference plot is formed by subtracting the
value of the 'B* arrival from that of the ’A'
arrival and dividing by 2. This value is plotted
as a small *+1 at each geophone location. Notice
that the value of the half-difference can be
negative when the 1B 1 arrival is greater than the
•A* arrival.
The commands available to you under the Two
option are shown in the screen below. These
commands are arranged to be convenient for use on
T-3164 121

a PC-, AT-, or enhanced-style keyboard. Make

sure that your NumLock is off if you are using a
combined cursor and numeric pad such as the one
on the standard PC keyboard. The specific
commands are discussed below.

Home and End keys - Change 'A* and 1B* shots

Use these commands to change the 'A*

and 1B* shots which Two is currently using
to perform the interpretation. It does not
matter whether 'A1 or 'B* is the forward or
reverse shot, RF will figure out the
direction of the arrival times relative to
the shot locations which you entered using
RFData. When you press Home or End, you
will be asked for a new shot number. You
may Escape at this point and return to the
former pair of shots. If you select a new
shot, the graph will be redrawn on the
screen and the status line will indicate the
new shot numbers. Any model segments will
be redrawn as well, but any segments which
were not Inserted will not be saved or
redrawn.

Up and Down keys - Change active shot

These commands allow you to switch the

active shot from 'A' to ’B' or vice-versa.
They act identically; both Up and Down are
included for ease of use. The current
cursor position will switch travel time
curves, but will not change geophone
positions.

PgUp and PgDn keys - Change current segment

number

Use PgUp and PgDn to increment or

decrement the current segment number. The
segment number will change by one and will
appear on the status line. The current
T-3164 122

segment number will "roll" from 4 to 1 and

from 1 to 4.
Note that Inserting a segment also
performs a PgUp, so you shouldn't have to
manually change segment numbers except when
you switch current shots.

* key (shifted 8 and unshifted PrtSc) - Redraw

If you need to Redraw the screen for

any reason use *. If you have inserted and
deleted many segment points. Redraw will
clean-up the screen.

+ key - Assign an arrival to the current segment

The + key is used to assign an arrival

to the current segment. The symbol used to
indicate the arrival will change from a
square to a diamond when the arrival has
been assigned to the current segment. The
status line will indicate the direction
(forward or reverse) of the arrivals
relative to the shot point, based upon the
shot locations you entered using RFData.
When more than one arrival has been assigned
to the current segment, a least-squares-fit
line will be drawn through the assigned
arrivals. The apparent velocity of this
segment will be displayed on the status
line. The line through the arrivals of the
current segment will be dashed, to disting­
uish the current segment from all of the
model segments, which are indicated with
solid lines.
There are two constraints in assigning
arrivals to the current segment. First, you
must choose arrivals having only one
direction relative to the shot. In other
words, you may not choose one forward and
one reverse arrival and try to assign them
both to the same segment. Second, you must
not skip more than one arrival in a row when
assigning arrivals to the current segment.
You may skip one arrival, then assign the
T-3164 123

next arrival to the current segment, and

then skip another if you choose, and so on.

- key - Delete an arrival from the current

segment

The - command performs the inverse

operation of the + command. Use this
command to delete the current arrival from
the current segment. The least-squares
apparent velocity is recalculated for the
remaining assigned arrivals and drawn as a
dashed line and the symbol for the current
arrival is shown as a square. If the last
assigned arrival is deleted, the current
segment is deleted as well and reinitial­
ized. The delete command and its associated
redraws tend to make the screen display
appear messy, so you may need to redraw with
the * command.

Ins key - Insert the current segment into the

model

Once you have established a reasonably

good current segment, and have made sure
that the segment number is correct, you will
probably want to Insert the current segment
into the model. The dashed line on the
display will be changed to a solid line, and
the status line will be reset. Note that
the segment number has been automatically
incremented by one.

Del key - Delete a segment from the model

If you want to delete a model segment,

position the cursor on any arrival which
belongs to that segment. Press the Del key,
and you will be asked if you are sure you
want to delete the segment. You may answer
T-3164 124

No and leave the model intact. To delete

many segments, it may be easier to re­
initialize the entire model from the
RFInterp menu.

P - Print the screen on the printer

You are given the opportunity to print

the screen as soon as each screen is
displayed. At any other time while a
graphics screen is shown, press P (not
PrtSc). You will then be told to Return
when the printer is ready. You may also
Escape out at this point.
The reason PrtSc is not used to perform
screen dumps is that Graphics.Com and PrtSc
work well with CGA (640x200) displays, but
not with many others. The screen dump
utility used in RF is independent of display
type and printer type, so long as you have
installed the correct drivers for your
hardware. So you can dump your EGA display
to a laser printer, for example, by using
Print.

Esc key - Leave the Two option

When you are finally done with the

layer assignments, press Escape and you will
return to the RFInterp menu. If you have
Escaped by mistake, you may return to the
interpretation again by typing T. Your 'A'
and 'B* shots, model segments, and plot
parameters will all be retained, though you
will have deleted and reset the current
segment.
T-3164 125

P - Select the plot parameters

RFInterp does not automatically scale the

screen plots. You need to choose the starting
positions and increments for the distance and
time axes. You are given the freedom to zoom in
on an interesting portion of the travel time
curves, or to pan out and see the whole picture.

I - reinitialize the model

Use this option to set all of the model

segments, layer thicknesses, velocities, and time
depths equal to zero. You are asked if you
really want to reinitialize the model. Type Y if
you do, N if you don't. If you want to delete a
single model segment, use Del described above.

Other options - R, S, D, \, and X

These are the file handling and exit

options. They are identical to those described
in the RFData section above, and are available in
RFInterp for convenience.
T-3164 126

RFCalc

RFCalc is the module which takes your layer assign­

ments and produces the time, velocity, and depth sections.

There is really no interpretation performed using this

module. All of the interpretation is completed using

RFInterp. The only judgement you will exercise while using

RFCalc is the assignment of reciprocal shot pairs and the

selection of reciprocal (end-to-end) times. The main menu

for RFCalc is shown in the screen below. The options are

described in detail following the screen.

RF 1.0 - RFCalc >> Choose an option below by typing the Bold letter.
Time Section Graphs

Assign - Choose shot pairs, choose plot Parameters

end to end times. draw Time section
draw dEpth section
Find - Compute the time section draw Velocity section
using assigned shot pairs.

Manual - Edit the time section Files

by hand
Read .RFR file
Save .RFR file
set data drive\path
Directory of data area
Depth Section
Leave
Calculate - Using the time section
and "true" layer
velocities eXit - go back to RF menu

Choice ?
T-3164 127

A - Assign reciprocal shot pairs

This option allows you to match up to six

pairs of shots for each of the four available
layers in the RF model. When you press A, you
will see an input screen with several highlighted
fields for each pair.
Enter a forward and reverse shot for each
reciprocal pair. You are next asked to choose a
method to figure the reciprocal time. There are
five choices. The N choice does nothing and you
are allowed to enter the time manually in the
field which appears just below. The F and R
choices use the forward or reverse segments by
themselves to estimate the end time by linear
extrapolation of the model segment apparent
velocity. The X choice extrapolates both of the
segments and averages them. The H choice uses an
internal shot (figure 28) which must have a
segment from the same layer as the layer that you
are trying to compute the end time for. The
half-intercept of the segment at the internal
shot is found, and it is compared to the delay
time formed by using arrivals of the reciprocal
shots at the geophone closest to the internal
shot. Then the reciprocal time is calculated for
the shot pair so that the delay time matches the
half-intercept.
With any of the choices you select, you are
next prompted at the reciprocal time field
itself. If you like the value, just press
Return, or you can edit the value to be whatever
you want.
For the first layer, the main reason to
assign reciprocal pairs is to average the
adjacent apparent velocities using a harmonic
mean, so as to make the thickness calculations
more realistic. The reciprocal time is not used
except when there is at least two geophones'
overlap. So you can choose the Nothing option
and leave a zero value, or try some extrapolation
and see what values you get.
The second through fourth layers need
reciprocal times if there are more than two
overlapping geophones on one of those layers.
The X choice works very well with well-behaved
data. Your average reciprocal time for such data

ARTHUR LAKES LIBRARY

COLORADO SCHOOL of MINES
dOLDRN. COLORADO 80401
T-3164 128

is usually within 5% of the end times for either

model segment.

it -

Figure 28 - Scheme Used to Compute the Recipro­

cal Time Using an Internal Half-intercept Time.

F - Find the time section

Once you have Assigned the shot pairs you

may choose to have RF calculate the time depths
for layers 2 through 4. The time depths are
computed using the Reciprocal Method (Hawkins,
1961, Palmer, 1980). If you have not assigned
reciprocal shot pairs, or if you have used
improper reciprocal times, then Find may complain
with an error message and halt before it is done.

C - Calculate the Depth Section

After Finding the time section, you can

Calculate the depth section for the seismic line.
This option uses the geophone elevations which
you have entered. The arrival times should be
assigned to layers, the reciprocal shot pairs
should be assigned, and the time section should
be found before this option is used.
T-3164 129

P - Choose the Plot parameters

This option is similar to the Parameters

option in RFInterp. It allows you to choose the
scales for the time and depth section plots.

T - draw the Time section

After you have calculated the time section,

you can plot it on the screen using T. RF uses
the scales for the time and distance axes that
were set with the Parameters option.

E - draw the dEpth section

This option is analogous to T and plots the

depth section using the geophone elevations you
entered in RFData and the plot scales set in
Parameters.

V - draw the Velocity section

Use this option to plot the layer velocities

at each geophone location. The shallower layers
will have lesser velocities, and the deeper
layers will have greater ones. The plot scales
set in Parameters will be used for the velocity
section.

Other options: R, S, D, \, and X

These are the file handling and exit

options. They are identical to those described
in the RFData section above, and are available in
RFCalc for convenience.
T-3164 130

CONFIGURING RF FOR A PARTICULAR SYSTEM

To run RF successfully, you must have an I B M ™ PC-,

XT-, or AT-compatible computer, MS-DOS 2.0 or greater, and

at least 420 kB of RAM above what you require for resident

utilities. Two floppy disk drives or a hard disk are

required.

RF requires several hardware drivers and auxiliary

files to function correctly. The use of these files is

described in this section.

A typical RF program disk might contain the following

files.

R F .EXE The RF program.

RFRES.BAT Loads resident files, then runs
RF.
RFSCR01.PAK Screen library.
R F .CNF Default hardware configuration.
HALOxxxx.DEV Video device driver.
HALOxxxx.PRN Printer driver.
HALORTP.EXE Graphics resident driver.
SPEEDSCR.COM Screen utility resident driver.

The *xxxx1 characters in some of the above file names will

vary depending on what kind of hardware you have.

The graphics portion of RF uses H A L O ™ graphics from

Media Cybernetics. HALO achieves its compactness and speed

by use of a resident driver program, HALORTP. You must run

this program once after you have booted your system, and

before you run RF. If you leave RF and want to go back to

T-3164 131

RF again, you do not need to run HALORTP as long as you

have not re-booted your system. Run HALORTP simply by

typing its name and pressing Return. You will see a

welcome and copyright message to let you know HALO is

resident in memory.

The screen handling facilities in RF are provided by

Speed S c r e e n ™ from Software Bottling Co. Similar to HALO,

Speed Screen uses a resident driver. Run this driver by

typing Speedscr m=20 and pressing Return. You will see

another welcome and copyright message. Speedscr will

reserve 20k bytes of memory for the screen library. Again,

you must run this program once before you run RF, but you

do not need to run it again if you want to return to RF and

have not re-booted. Speed Screen uses the file

RFSCR01.PAK, which contains all of the screens for RF.

This file must also be on your program disk.

A batch file, RFRES.BAT, will install these resident

drivers and then run RF. You can use RFRES.BAT once after

you boot-up, by typing RFRES and then pressing Return. If

you leave RF and want to use it again though, just type RF

since the resident drivers will remain in memory until you

re-boot your computer.

The next set of required files are the video and

printer drivers, HALOxxxx.DEV and HALOxxxx.PRN. The

T-3164 132

particular video and printer drivers you need must be on

your program disk when you run RF.

RF uses a configuration file, RF.CNF to store the

names of the drivers it uses. When you start RF, the

program will look for the configuration file on the default

drive\path. If RF finds a .CNF file, it will use it

without asking. If it cannot find one, it will ask you

what kind of hardware you have and attempt to write a new

.CNF file. Figure 29 shows a flow chart of RF's

configuration and initialization procedure.

T-3164 133

DOS

Halortp

ii

no

Bead config
info from
diak file

Ialt BF
eavlreameat

Figure 29 - Configuration and Initialization

Procedure for RF.
T-3164 134

APPENDIX B - THE PROPOSED SEG-1 STANDARD

The SEG-1 exchange format is an attempt to standardize

the way in which engineering seismic data is recorded and

transmitted. The standard has been proposed by the

Engineering and Groundwater Committee of the Society of

Exploration Geophysicists (J. Hunter, pers. comm., 1987).

This appendix is a summary of the proposed standard as

outlined by Hunter (unpubl).

The purpose of the SEG-1 standard is to give manufac­

turers and users of engineering seismographs access to

needed information regarding not only the recorded seismic

data, but details such as source and receiver configura­

tion, recording parameters, and operator's notes. In

addition, the proposed standard defines the format of data

transmitted from engineering seismographs, allowing easier

data transfer to computers.

To initiate a data transfer, the computer is connected

to the seismograph through an RS-232C or IEEE-488 port.

The computer sends an ASCII string <Esc>"SEG-l"<CR> to the

seismograph, whereupon the seismograph begins transmission.

All of the header information is transmitted as ASCII

strings. The seismic data itself may be transmitted in

binary form depending on the manufacturer. Table 9 shows

the transmitted data format.

T-3164 135

Table 9 - Proposed SEG-1 Standard

Seismic Line Information

File Number <CR>
Job Number* <CR>
Date and Time* <CR>
Operator's Note* <CR>
Manufacturer's Code <CR>
Number of Channels <CR>
Number of Samples per Channel <CR>
Sample Interval <CR>
Delay Time <CR>
Analog High-Pass Filter Cutoff* (Hz) <CR>
Analog Low-Pass Filter Cutoff* (Hz) <CR>
Analog Notch Filter* (Hz) <CR>
Shot Location X-coordinate* <CR>
Shot Location Y-coordinate* <CR>
Shot Location Z-coordinate* <CR><LF>

Geophone Information
Geophone #1 X-coordinate* <CR>
Geophone #1 Y-coordinate* <CR>
Geophone #1 Z-coordinate* <CR>
Geophone #1 Operator's Note* <CR>
Number of Stacks performed for Geoph #1*<CR>
Gain for Channel #1*(dB or AGC) <CR><LF>

. . . for geophones 2 through N . . .

Geophone #N X-coordinate* <CR>

Geophone #N Y-coordinate* <CR>
Geophone #N Z-coordinate* <CR>
Geophone #N Operator's Note* <CR>
Number of Stacks performed for Geoph #N~<CR>
Gain for Channel #N*(dB or AGC) <CRxLF>

. . . after all N geophone headers . . .

<ETX>

* Optional, except that <CR> must be sent

ASCII codes: <CR> decimal 13, hex OD
<LF> decimal 10, hex 0A
<ESC> decimal 27, hex IB
<ETX> decimal 03, hex 03
T-3164 136

The trace data follows the header data. Some manuf­

acturers require that the computer send additional charact­

ers to the seismograph before the seismograph transmits the

trace data. The right-handed coordinate system is used for

the shot and geophone locations.

T-3164 137

APPENDIX C - RF SYSTEM PROGRAMS

These programs are presented in partial fulfillment of

requirements for the Master of Science degree at Colorado

School of Mines, Department of Geophysics. This software

is presented "as is", without warranty of any kind, either

expressed or implied, including but not limited to the

implied warranties of merchantability and fitness for a

particular use. The Colorado School of Mines and the

author do not assume any responsibility for results

derived through use of this software.

Software may be obtained by contacting Colorado School

of Mines, Department of Geophysics, Golden, CO 80401.