Llli1 Ill Lilt Il: Assessment and Selection of An Interpretation Procedure

AD-A255 749
lllI1 Illtililt Hr
11111 il TECHNICAL REPORT GL-92-12
ASSESSMENT AND SELECTION OF AN

*
-AUTOMATED ELECTRICAL RESISTIVITY
INTERPRETATION PROCEDURE
by
Janet E. Simms, Dwain K. Butler
Geotechnical Laboratory
Z;, DEPARTMENT OF THE ARMY
Waterways Experiment Station, Corps of Engineers
3909 Halls Ferry Road, Vicksburg, Mississippi 39180-6199
DTIC
S ELECTE
SEP 181992
A
Electrode Spacing (ft) L'-0_ '$
a
___ August 1992
___ NFinal Report
Approved For Public Release; Distribution Is Unlimited
Prepared for DEPARTMENT OF THE ARMY

US Army Corps of Engineers
LABORATORY Washington, DC 20314-1000
, :err neire t r n
,r
rn feee Do not return
t :o !he orginxaor
The firdings
in this report are not to be construed as an official
Department of the Army positinn jnless so cesiqnated
by other authorized documents
The contents of this report are not to be used for

advertiswq. Publication or promotional purposes
Citation of trade names does not constitute an
,fficial erdorsement or approval of the use of
such commercial products
Form Approved
REPORT DOCUMENTATION PAGE OMB No. 0704-0188
data sources,
Public reporting burden for this collection of information isestimated to average I hour per response,including the time for reviewing istructions. searching existingq
gathering and maintaining the data nedd n opeigadreviewing the collection of information Send comments regarding this burden estimate or any other aspect0f this
collection Itinformation, includig sugge=t.om br reducing this burden, to Washington H~eadquarters Services.Directorate for information Operations and Reports, 12IS Jefferson
DavisHighway. Suite 1204. Arlington. VA 22202-4302. and to the Office of Management and Budget. Paperwork Reduction Project (0704.0188). Washngton. DC20503
1. AGENCY USE ONLY (Leave blank) 2. REPORT DATE 3. REPORT TYPE AND DATES COVERED
August 1992 Final report
4.TITLE AND SUBTITLE A5. FUNDING NUMBERS
Assessment and Selection of an Automated Electrical
Resistivity Interpretation Procedure Project AT40,
Task WS,
6.AUTHOR(S) Work Unit 001
Janet E. Simms
Dwain K. Butler
7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) B. PERFORMING ORGANIZATION

REPORT NUMBER
Station
USAE Waterways Experiment
Geotechnical Laboratory Technical Report
3909 Halls Ferry Road GL-92-12
Virksburg, MS 39180-6199
2. SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESS(ES) 10. SPONSORING/MONITORING
AGENCY REPORT NUMBER
US Army Corps of Engineers
Washington, DC 20314-1000
11. SUPPLEMENrARY NOTES
12a. DISTRIBUTION /AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE
Approved for public release; distribution is

unlimited.
13. ABSTRACT (Maximum200words) The Defense Science Board and Joint Chiefs of Staff have
idendified ground-water supply development capability as a major technological
shortfall. The goal of research and development under the Corps of Engineers
Water Supply Program is to provide the military with the capability to detect new
ground-water sources to support operations, aid in humanitarian relief programs,
and as a part of nation assistance. The detection of ground-water sources will be
achieved through an integrated, automatic data acquisition and interpretation
capability for electrical resistivity and seismic refraction data. This report
addresses the requirements for the resistivity data acquisition and interpretation
capability. These requirements include a) computer controlled, automatic data
acquisition, b) direct data input into the resistivity interpretation program, c)
automatic interpretation option, and d) equivalence analysis capabilities. The
data acquisition system will include a multi-conductor sounding cable and electrod
switch box which provides a fast method for data collection. Several resistivity
interpretation programs are evaluated based on the requirements stated above plus
(continued)
14. SUBJECT TERMS 15. NUMBER OF PAGES
Geophysics 83
Fes stivity,Geoelectrical 16. PRICE CODE
Resistivity Inversion
17. SECURITY CLASSIFICATION 18. SECURITY CLASSIFICATION 19. SECURITY CLASSIFICATION 20. LIMITATION OF ABSTRACT
OF REPORT OF THIS PAGE OF ABSTRACT
Unclassified Unclassified I
NSN 7540-01-280-5500 Standard Form 298 (Rev 2-89)
Prescribed by ANSI Stid 39-18
98*-102
13. (continued)
other desirable features. Both theoretical and actual field data are used to
evaluate the performance of each interpretation program. The program RESIX
PLUS, written by INTERPLEX Ltd., Golden, Colorado, performed well and satisfies
the majority of requirements.
PREFACE
The work documented in this report was performed during the period October
1991 through May 1992, and was sponsored by the US Army Corps of Engineers
(USACE), Project AT40, Task WS, Work Unit 001, "Subsurface Water Location." The
USACE Technical Monitor was Dr. Donald Leverenz.
The work was performed by Drs. Janet E. Simms and Dwain K. Butler,
Earthquake Engineering and Geosciences Division (EEGD), Geotechnical Laboratory
(GL), US Army Engineer Waterways Experiment Station (WES). This report was
prepared by Dr. Simms. General supervision was provided by Mr. Joseph R. Curro,
Chief, Engineering Geophysics Branch, EEGD, Dr. Arley G. Franklin, Chief, EEGD,
and Dr. William F. Marcuson III, Director, GL.
At the time of publication of this report, Director of WES was Dr. Robert
W. Whalin. Commander and Deputy Director was COL Leonard G. Hassell, EN.
ccesion Fo:
NTIS CRA&I
DTIC TAB
Unannouncej ,
Justification
By
Disth.Di to .. Iy..................
.........
ib'otion .....
A.vailc"A; I:ty o"_-
Dist
A-1
DTIC QUALry fNSPECTD 3
1
CONTENTS
Page
PREFACE ............................................................. 1
LIST OF TABLES .. ................................................... 4
LIST OF FIGURES .. .................................................. 4
CONVERSION FACTORS, NON-SI TO SI (METRIC) UNITS
OF MEASUREMENT .. ................................................. 6
PART I: INTRODUCTION .. ............................................ 7
Background .. ................................................. 7
Detection versus Exploration ................................ 7
Scope ......................................................... 9
PART II: CONCEPT DESIGN CONSIDERATIONS FOR AUTOMATED

GEOPHYSICAL SURVEYING CAPABILITY ...................... 10
Seismic Data Acquisition Capability ........................ 13

Resistivity Data Acquisition Capability .................... 16
PART III: DESIRABLE ATTRIBUTES OF RESISTIVITY

INTERPRETATION . ....................................... 19
Inversion Algorithm . ........................................ 19

Direct Data Input . .......................................... 19
Automatic Interpretation ................................... 20
Equivalence Analysis . ....................................... 20
Graphical Output . ........................................... 21
User-Friendly Program ....................................... 21
Incorporation Into Overall System .......................... 21
PART IV: SURVEY OF AVAILABLE RESISTIVITY INTERPRETATION

PROGRAMS . ............................................. 22
2
Page
DCRESI ....................................................... 22
RESIX........................................................ 22
SVES... .......................................................24
RESINV......................................................... 24
ATO.......................................................... 25
PART V: EVALUATION OF RESISTIVITY INTERPRETATION PROGRAMS ....... 26
Theoretical Examples . ....................................... 27

Field Examples . ............................................. 36
PART VI: SUMMARY AND CONCLUSIONS ................................. 55
REFERENCES .........................................................57
APPENDIX A: MILITARY REQUIREMENTS FOR GEOPHYSICAL GROUND WATER

DETECTION AND EXPLORATION ........................... Al
Abstract . ................................................... A2
Background . ................................................. A2
Geohydrological Models ...................................... A3
Detection Versus Exploration ............................... A4
Detection Principles . ....................................... A7
Emerging Technology ......................................... A O
Ground Water Detection Field Trials ........................ All
Conclusions . ................................................ A18
Military Deployment of Geophysical Ground Water Detection
Capability . ............................................... A20
References . ................................................. A22
3
LIST OF TABLES
No. fAe
I Features of resistivity interpretation programs ............ 23
2 Summary of best resistivity interpretation program
for modeling each data set ............................... 56
LIST OF FIGURES
No. Lne
1 Conceptual diagram of automated geophysical
surveying capability ..................................... 11
2 Flow chart showing steps involved in the implementation
of the automated geophysical surveying capability ........ 12
3 Desired features of the automated geophysical
surveying capability ..................................... 14
4 Seismic refraction data acquisition system ................. 15
5 Resistivity data acquisition system ........................ 17
6 Example 1, (a) three-layer model and (b) corresponding
theoretical Schlumberger sounding curve with 5% Gaussian
noise added .............................................. 28
7 Example 1, inversion results using the (a) user-aided and
(b) automatic interpretation methods ..................... 29
8 Example 2, (a) three-layer model and (b) corresponding
noise added .............................................. 30
10 Example 3, (a) four-layer model and (b) corresponding
noise added ............................................. 33
12 Example 4, (a) four-layer model and (b) corresponding
noise added ............................................. 35
14 Example 5, (a) Schlumberger field data, (b) borehole data,
and (c) seismic interpretation ........................... 38
16 Example 6, Schlumberger field data collected at a landfill
site ..................................................... 41
18 Example 7, (a) Schlumberger resistivity data set SW-19
and (b) seismic interpretation ........................... 43
19 Example 7, inversion results of resistivity data set SW-19
using the (a) user-aided and (b) automatic
interpretation methods ................................... 44
4
Page
20 Example 8, (a) Schlumberger sounding curve T-14 and
(b) seismic interpretation ............................... 46
21 Example 8, inversion results of resistivity data set T-14
interpretation methods ................................... 47
22 Example 9, (a) Schlumberger sounding curve VES-2 and
(b) borehole data ........................................ 49
23 Example 9, inversion results of resistivity data set VES-2
interpretation methods . ................................... 50
24 Example 10, Schlumberger sounding curve VES-3 .............. 51
26 Example 11, Schlumberger sounding curve VES-4 .............. 53
5
CONVERSION FACTORS, NON SI-TO SI (METRIC)
UNITS OF MEASUREMENT
Non-SI vnits of measurement used in this report can be converted to SI

(metric) units as follows:
Multiply By To Obtain
feet 0.3048 metres
6
Assessment and Selection of an Automated
Electrical Resistivity Interpretation Procedure
PART I: INTRODUCTION
Background
1. The Defense Science Board and Joint Chiefs of Staff have identified
ground-water supply development capability (which includes detection) as a
major technological shortfall. In many areas of the world, particularly the
hot, dry desert regions, existing developed water sources are nonexistent or
insufficient to meet military requirements. The goal of research and
development under the Corps of Engineers Water Supply Program is to provide
the military with the capability to detect new ground-water sources to support
operations, aid in humanitarian relief programs, and as a part of nation
assistance.
Detection versus Exploration
2. Geophysical methods are routinely used throughout the world in

civilian exploration programs for the assessment and development of ground-
water resources. These exploration programs are far different from exploration
in the military environment because there is a nearly unlimited amount of time
available, there is no competition for available logistic support, and the
scientific skill and experience levels of the explorers are much higher than
can reasonably be expected to be available in a combat theater. The surface
geophysical methods that are predominantly used in these ground-water
exploration programs are gravity, electrical resistivity, electromagnetic, and
seismic refraction methods. Although occasionally only one of these methods
will be used in an exploration program, generally at least two of the methods
are used in a complementary approach. A geophysical ground-water exploration
program will normally use all available geological and borehole data in order
to produce the best possible assessment of the ground-water potential and
conditions in an area (Butler and Llopis 1984).
3. The primary objective of geophysical ground-water exploration is the
7
mapping of subsurface structural and stratigraphic indicators of the possible
occurrence of ground water, such as buried river channels, fracture zones in
bedrock, confining layers (aquacludes), etc. Actual detection of the ground-
water table with any of the geophysical surveys may be noted but may not be of
primary importance in the overall ground-water exploration assessment.
4. The expression "ground-water detection", in contrast to ground-water
exploration, applies to the concec of actually detecting the presence (or
absence) of ground water and the depth to the water table beneath a given
"point" on the surface by conducting one or more types of geophysical tests at
that point. Ideally, aquifer thickness and water quality would also be
determined. For some cases, information regarding ground-water occurrence and
other geological factors might be available but, in general, the assessment of
the presence of ground water must rely solely on the geophysical results at
the given surface location in the detection scenario. However, it is
envisioned that many geophysical ground-water surveys will be conducted to aid
in choosing between alternate sites in an area already identified as having
good ground-water potential by other methods. Of the geophysical methods most
commonly used in ground-water exploration programs, only electrical
resistivity, electromagnetic, and seismic refraction methods are applicable to
the ground-water detection problem. These methods are complimentary for
locating possible ground-water sources, with the electrical techniques
identifying different layers and variations in water salinity based on the
resistivity (or conductivity) of the subsurface materials, whereas seismic
data yields the subsurface structure, layer thickness, and layer velocities.
Water saturated earth material has a characteristic velocity (-1500 m/s) and
range of resistivities (10-300 0-m), therefore by combining the results
obtained from the electrical and seismic methods, it is possible to determine
if a ground-water source is present. Detection principles for these methods
are described in Appendix A. The capabilities provided will be general enough
to perform in a "detection" mode for a tactical application, or in an
"exploration" mode for either fixed base water siupply or humanitarian
relief/nation assistance applications.
8
ScoiDe
5. Initially, the detection of ground-water sources in a military

environment will be achieved through an integrated, automatic data acquisition
and interpretation capability for electrical resistivity and seismic
refraction data. This will relieve the need for extreme levels of civilian
scientific skill and experience, and also accommodate the military time
constraints. The final capability is also likely to incorporate
electromagnetic methods also. The second part of this report describes the
design considerations for an integrated data acquisition system.
6. Prior to integrating the geophysical data sets, it is first necessary
to evaluate each geophysical method independently to determine the optimum
method for data acquisition and interpretation. The third part of this report
addresses the electrical resistivity interpretation process. It discusses the
features desirable in a resistivity interpretation program and evaluates
several inversion programs based on these requirements.
9
0.
C) 0
Q. CE CL
oc Q-
0 0
c04-) 4)
Cl* 0~ qq
LA~i 0
4
U
to u wt U'
(LU
0L 0
.
a 0
C 4
WQ
0
........0
CL 44i
So 0
0 0
0 4.)
I4J 0
wE
0 .tO
CO I--O4J
zD -0~
~C
0 0 4
(ID 3rs < 0
oo 00.
.....
0 0
% 0)
LU.. CC. 0 .)
0-0t
4 c
F34
--climate
e) drilling rig capability
--drilling method
--depth limitations
f) logistical considerations
--location/operational requirements
--accessibility/mobility
--security
--estimated drilling time

8. Desired features of the resistivity and seismic refraction data
acquisition and interpretation capability are summarized in Figure 3. The data
acquisition systems for both the resistivity and seismic methods are described
below, with the remaining sections of this report concentrating on the
resistivity interpretation aspects.
Seismic Data Acquisition Capability
9. The seismic refraction data will be acquired using conventional

methods with the addition of computer controlled acquisition (Figure 4). The
system consists of a 24 channel seismograph with roll-along data acquisition
capability. A typical geophone cable consists of twelve geophones, with
geophone spacings available from 10 feet to 50 feet (ft) (total cable length
120-600 ft). The length of the seismic line is dependent on the desired depth
of investigation, generally 4 to 5 times the depth of investigation. For
drilling rig capabilities of 600 ft and 1500 ft, the minimum length of a
seismic line would be 2400 ft and 6000 ft, respectively. The knowledge
assisted system will guide the user through the setup of acquisition
parameters and the data collection procedure. The data will be exported
directly to an external computer for interpretation using the seismic
refraction data processing software SEISMO (Yule and Sharp, 1990). SEISMO is
an interactive program which determines the velocity structure based on a
given set of travel times. At the present time, the program requires the input
of an apparent velocity model in order to obtain the true velocity profile.
The apparent velocity model can be entered in one of two ways: 1) the model
can be entered via keyboard, or 2) the model can be obtained interactively by
13
CC
00 0
_0 N c
co2 Lo .5 -
CL E0 N Ia-
V 0% CD 0 *
R 10. W 2 M~I CD
t.j 0 -9 0U 0 C.
%0 > 00~ 0
<> 0 z 4
0 'E iz -9 L-01; ~
0- 0 01 .0O0C
oi 0ot
0 -f 0 o 0. :- 0
-M T E m E 0 0 'a c
a -o 0. o2V >
0- EL 0 0
0 w00 az0
C4 sU - (D'
V c '0C ' 5 La)P

CD u 0
0 so
a0 2. '9 002
ou .1 !. 0
E E 9
0~ 0 D
Q 414
CL
C-
EE
oc
oc
CL 0
4
0*
ca
.0
oD4
0)0
-,4
0~r
identifying points where the slope in the travel time versus distance curve
changes. There are plans to make this step optional (where the user aids in
determining the initial model) and allow for the program to be completely
automatic.
Resistivity Data Acquisition Capability
10. The conventional method for collecting d.c. resistivity data is to

use four electrodes, two current electrodes (C) and two potential electrodes
(P), placed in a straight line (Figure 5a). The spacing between the electrodes
is dependent on the type of array used and the desired depth of investigation.
A general rule of thumb is that the maximum current electrode spacing should
be 3 to 5 times the desired depth of investigation. Current is injected into
the ground through one current electrode and returns through the other. The
potential difference measured between the two potential electrodes can be
related to the resistivity of the subsurface material (refer to Engineer
Manual 1979). Several measurements are required to construct one sounding
curve. Each measurement is taken at a different electrode spacing, requiring
the electrodes to be moved after every measurement. This requires extra
manpower and is a time consuming process.
11. In the a" omated data collection system, the electrodes are placed
at each desired spacing and connected to a multi-conductor sounding cable,
using an electrode clip lead, prior to taking any measurements (Figure 5b).
The multi-conductor cable is connected to an electrode switching box which
allows a quick method for either manual or automatic (via computer) switching
of electrodes. This reduces the manpower requirements and data acquisition
time for performing a resistivity survey. The electrode switch box is
connected to the resistivity instrument, which is powered by a 12 volt
battery. The system will allow data collection to be computer controlled or
manually controlled through the resistivity instrument. If computer
controlled, the program will prompt the user for array type, maximum electrode
spacing, and other information necessary to perform the survey. The data are
input directly into the computer as it is collected. The program will prompt
the user to either accept the measurements or remeasure to ensure good quality
data. Upon completion of the survey, the resistivity interpretation program is
16
C P, C2 ground
surface
(a)
ground
7 7 J 7 surface
electrode o uI-ouco
switch box cable
00 - resistivity meter
6 %
-d4- computer
(b)
Figure 5. Resistivity data acquisition system. (a) conventional method

and (b) automated system.
17
activated and, either automatically or user assisted, the data inverted to
obtain the number of (electrical) layers present and the resistivity and
thickness of each layer. This information will be combined with the results
from the seismic refraction survey to determine if the local site is a
potential ground-water source.
18
PART III: DESIRABLE ATTRIBUTES OF RESISTIVITY
INTERPRETATION
12. The following discussion addresses the features desirable in a

resistivity interpretation program. Parts IV and V evaluate several inversion
programs based on these requirements. The resistivity inversion programs are
evaluated based on the following features: 1) how robust is the inversion
algorithm, 2) ability for direct data input, 3) automatic interpretation
capabilities, 4) equivalence analysis, 5) graphical output, 6) how user-
friendly is the program, and 7) ease of incorporating the program into the
overall multi-interpretation capability.
Inversion Algorithm
13. The inversion algorithm is the core of any resistivity

interpretation program and it is necessary that it be robust. Noise will
always be present in any resistivity data set, therefore the algorithm must
adequately handle the noise and any problems encountered due to the inherent
nonuniqueness of resistivity data. The need for a robust algorithm is evident
when one considers the possibility that a nonrobust algorithm can invert a set
of data which has a unique solution and converge to the wrong solution. The
primary criteria for evaluating a resistivity interpretation program should be
the robustness of the inversion algorithm.
14. The inversion algorithm should also have the capability to handle
Schlumberger, Wenner, and dipole-dipole array data. These three array types
allow the versatility to collect sounding or profiling data with various
depths of investigation.
Direct Data Input
15. Thb data should be directly imported into the resistivity inversion
program from the data acquisition system. This would eliminate errors which
may arise from manually inputting the data and reduce the time required for
interpreting the data. If a commercially available inversion program is used,
then the software company must be willing to either supply the program source
19
code or adapt the program to satisfy this requirement.
Automatic Interpretation
16. The ultimate user of the data acquisition and interpretation system
will be military personnel with no formal training in or knowledge of
geophysics. Since the system is geared toward the layman, it is highly
desirable to have the option of automatic data interpretation. This option
will automatically generate the required initial model for input into the
inversion program. Since the automatic interpretation option may generate an
initial model having similar adjacent layer resistivities or layers that are
too thin to be resolved at depth, it is recommended that a model integrity
routine be run following the formation of the initial model. Suggested
criteria are: (i) the thickness of a layer should be a minimum of 20% of the
total thickness of the layers above it; (ii) the resistivity of a layer should
differ from the layer immediately above it by more than 35%; and (iii) the
fitting error of the smoothed model should be within 10% of the fitting error
obtained with the unsmoothed model. These criteria will ensure that the
initial model consists only of layers that are significant for fitting the
data. None of the resistivity interpretation programs evaluated have this
feature, though it would be easy to incorporate.
17. Caution must be exercised when using the automatic interpretation
option to avoid pitfalls encountered when the human factor is removed from the
decision process. It is recommended that the user have at least minimal
training in the interpretation process. The traditional method of the user
supplying the initial model will also be available as an option.
Equivalence Analysis
18. Resistivity data is inherently nonunique, therefore an inversion

program should include a routine which performs equivalence analysis.
Equivalence analysis gives the range of variation in the parameters which will
fit the data within a specified error bound. The range of variation in a
parameter is a measure of the resolution of the parameter, with a smaller
range of variation indicating better resolution. Also, knowing the range of
20
parameter variation could aid in identifying the soil or rock type.
Graphical Output
19. The resistivity inversion program should provide on-screen and hard
copy plots of the data, best-fit sounding curve, inversion model, and range of
equivalent solutions. The program should also provide a tabular listing of the
data and inversion results.
User-Friendly Program
20. A user-friendly inversion program is a necessity since it will be

used by individuals who may not have the time to become familiar with the
intricacies of the program. It should guide the user through the
interpretation process, preferably through menus, with the possibility of
being fully automatic. If a commercially available program is used, aid from
the software company in meeting these requirements may be required.
Incorporation Into Overall System
21. The resistivity inversion program must be compatible with the entire
data acquisition and interpretation capability (resistivity, seismic, etc.). A
UNIX operating system is desired which will give compatibility (interfacing
capability) with the TERRACAMMS data base and geographical information system,
and other logistics planning tools (Falls et al. 1991). Willingness of the
software company to aid in meeting the compatibility requirements would be
advantageous.
21
PART IV: SURVEY OF AVAILABLE RESISTIVITY
INTERPRETATION PROGRAMS
22. Five resistivity interpretation programs are compared. Two of the

programs, RESINV and ATO, are available through government publications and
the other three, DCRESI, RESIX, and SVES, are commercially available. Each
program is evaluated based on the features described in Part III, with the
inversion algorithms evaluated in Part V. The programs and their features are
listed in Table 1 for reference.
DCRESI
23. DCRESI is no longer commercially available but has been replaced by

RESIX, which is written and distributed by INTERPEX Ltd., located in Golden,
Colorado. The U.S. Army Engineer Waterways Experiment Station (WES) possesses
a copy of DCRESI. The program is not user-friendly. It is a command driven
program, as opposed to menu driven, and the user must be familiar with the
program in order to execute the commands in the proper sequence. DCRESI
supports several electrode array types but it does not have an option for
automatic interpretation or equivalence analysis. The program does meet the
minimum graphical output requirements. It cannot run on a UNIX platform. Data
cannot be directly imported into the program and, since DCRESI is no longer in
distribution, the software company is not favorable to amending it to meet our
requirements. Therefore, the program DCRESI will not fit well into the overall
capability.
RESIX
24. The program RESIX PLUS is the revise" version of DCRESI. It is menu
driven which makes the program more user friendly. The program can model data
collected from several electrode array configurations, and has options for
automatic interpretation and equivalence analysis. At the present time, the
data cannot be automatically read into the program via the resistivity meter,
but the software company, INTERPEX Ltd., is willing to modify the program to
meet our needs. Personnel at the 'WES can also provide this service. INTERPEX
22
Table I
Features of Resistivity Interpretation Programs
PROGRAMS
Feature DCRESI RESINV SVES RESIXPLUS
user-friendly yes yes
graphic output yes yes yes yes
direct data input yes
various yes yes yes

electrode
arrays
automatic yes yes

interpretation
equivalence yes yes

analysis
unix platform yes yes
easily yes
incorporated
DCRESI and RESIX PLUS: INTERPEX, Ltd., Golden, Colorado.

SVES: Atlas Copco ABEM, Bromma, Sweden
RESIN': Butler et al. (1982), Davis (1979a, b), Mooney (1979)
23
is currently adapting the program to run on a UNIX platform. RESIX PLUS would
fulfill our resistivity interpretation requirements. The WES possesses a copy
of RESIX PLUS.
SVES
25. SVES was written by the Swedish company Atlas Copco ABEM for use
with their resistivity instrument, the Terrameter. The data (Schlumberger
array only) can be imported directly into the resistivity inversion program
SVES when used in conjunction with the GEOMAC (ABEM trademark) hand-held
computer and Terrameter. The program has an automatic interpretation option
but does not offer an equivalence analysis routine. It is menu driven,
relatively user-friendly, and satisfies the graphical output requirements.
SVES runs on a DOS platform and there are no plans to modify it to run on a
UNIX platform. SVES has problems running under DOS 5.0. ABEM recently
announced that they will support the INTERPEX resistivity interpretation
software, therefore support or future upgrades to SVES will be limited or non-
existent. Because of the limited support for SVES, it is not recommended that
SVES be used in the Water Supply Program. WES does possess a copy of the
program SVES.
RESINV
26. The program RESINV is available through government publication

(Butler et al. 1982, Davis 1979a, b, Mooney 1979). It is written in Fortran
and will run on a UNIX system. The program is neither menu or command driven,
but runs through a complete interpretation of the data without pausing at
intermediate steps waiting for user input. It could be modified to run in a
menu driven mode. In its present state, it does not offer automatic
interpretation or equivalence analysis options but WES personnel could write
these routines. The program will accept either Schlumberger, Wenner, or
dipole-dipole data. RESINV could be modified to accept data input directly
from the resistivity meter. With the above changes, program RESINV could
satisfy our requirements. These changes, however, would require a considerable
amount of time to incorporate.
24
ATO
27. Resistivity inversion program ATO is also available through a

government publication (Zohdy and Bisdorf 1989). It can only be used to model
Schlumberger data. The program is a type of direct interpretation program in
that it does not require an initial model, but fits a model consisting of as
many layers (N) as there are data points (M, i.e., N-M). This violates the
general principles of nonlinear inversion, where typically N<M or even N<<M.
The resulting model shows more of a continuous variation in resistivity with
depth rather than a discrete layer structure. This type of resistivity
interpretation is not readily amenable to joint interpretation with other
methods which result in a layered structure. Therefore the program ATO will
not be given further consideration.
25
PART V: EVALUATION OF RESISTIVITY INTERPRETATION
PROGRAMS
28. Resistivity interpretation involves determining the number of

subsurface layers represented by the sounding data and the resistivity and
thickness of each layer. Prior to the advent of computers, all resistivity
interpretation was done using a set of master curves (Orellana and Mooney
1966). These curves represent simple two and three layer models, and by
matching the field sounding curve with the appropriate master curve it is
possible to determine the number of layers present and the resistivity and
thickness of each layer. As the popularity, capabilities, and availability of
computers grew, methods were developed which computerized the curve matching
process.
29. All interpretation programs evaluated here utilize a nonlinear
least-squares inversion scheme (Marquardt 1963, Inman 1975). Generally, a
resistivity interpretation program requires input of field data (actual or
theoretical), the number of model layers, and an initial estimate or guess of
the resistivity and thickness of each layer. The layer parameters are usually
estimated based on the shape of the sounding curve. Using the initial guess of
the layer parameters, a set of data is generated which is compared to the
field data. Through an iterative process, the inversion algorithm minimizes
the root mean square error between the field data and generated data, updating
the parameters until a given error criterion is satisfied. The automatic
interpretation programs do not require an initial guess of the layer
parameters, but instead use a computer algorithm (Koefoed 1976) to generate
the initial model.
30. Both theoretical and actual field data are used to evaluate the
resistivity inversion algorithms. Theoretical data are used because the
solution is known, whereas with field data the results can only be
corroborated with other types of field data, if available. Five percent
Gaussian noise is added to the theoretical data to simulate field conditions.
The theoretical examples range from simple three layer models which can be
well resolved through the inversion process, to models which exhibit both
nonuniqueness and poor resolution. Five of the real field data examples are
from an arid or semi-arid region. These examples were chosen because it is
26
important to see how the various algorithms respond to different types of
data. The four theoretical examples are discussed first, followed by the seven
field examples. It should be noted that, for the inversion programs which
require an initial model, the same initial model was used in each program for
a given example.
Theoretical Examples
Example 1
31. The first theoretical example consists of three layers, with the
second layer being conductive. The true model and corresponding sounding curve
are given in Figure 6. This example can be considered relatively unique and
therefore the inversion algorithms should have no problem in resolving the
true parameter values and the range of equivalent solutions should be small.
32. The programs DCRESI, RESINV, and the user-aided (initial model
provided by the user) SVES and RESIX programs do well in fitting the true
parameter values, all obtaining very similar solutions (Figure 7a). The
automatic interpretation of SVES fails to achieve a suitable inversion model
(unable to plot at the scale of Figure 7b). The RESIX automatic interpretation
yields a five-layer model when smoothing is not applied. If the initial model
estimate is smoothed prior to inverting the data using the criteria discussed
in Part III, then the resulting model is identical to the user-aided RESIX
inversion model (Figure 7b).
Example 2
33. Example 2 also consists of three layers but now the second layer
resistivity has a value of p2-100-m, making the model highly nonunique
(Figure 8). The algorithms are not expected to perform well in resolving the
second layer parameters, so there should be a large range of parameter values
which can fit the data within a small specified error.
34. The inversion results are given in Figure 9. Program DCRESI does
surprisingly well in estimating the true layer parameters for this nonunique
model. Unfortunately it does not have an equivalence analysis routine. The
27
- -4
*CF 0
-4
V) 0)
-4
0
0
0 -4
* J
-od
0 ::
s 4
x 0
0.0
4)
.0
E E
EE
C I
C C;
Ec
II
I
E
cr-'
Itd
0 a)
-I-
W 0 0
CL
U->w-
LLJ EC0:ina
L - 0
.2m
(wU) 44dea
'-4
00
* C) 0o
0 40
-00
.9q-
() 4
0L -4
00
00
~4)
'Jo
E E
co C
E
.r v
I I
E Z'
z 4)
E.E
0 .0 lJ
LLI 4
(wj) a.1 dea

user-aided and automatic SVES inversion results are similar, both
overestimating the second layer resistivity and underestimating the thickness.
Programs RESINV and user-aided RESIX also obtain similar results, but they
underestimate the resistivity and overestimate the thickness of the second
layer. Nonuniqueness is affecting these parameters. The automatic model
estimate of RESIX initially obtains a four-layer model, but by applying
smoothing criteria (i) the third layer can be eliminated (Figure 9b). This
smoothed RESIX model is the best of the automatic fitting programs.
Example 3
35. The third example consists of four layers with moderate contrasts in
layer resistivity (Figure 10). The parameters should be fairly well resolved
through the inversion process with a bounded range of equivalent solutions.
However, as the number of layers increases, the parameter resolution decreases
which corresponds to an increase in equivalent solutions.
36. Again, DCRESI achieves the best inversion model (Figure lla). RESINV
and the user-aided program RESIX also result in good inversion models. Both
automatic interpretation programs overestimate the number of layers in the
true model (Figure lib) (the arrow indicates that the curve extends beyond the
bounds of the graph). Of the two, the six-layer model of RESIX is a better
estimate. The RESIX initial model does not meet any of the smoothing criteria.
However, since the additional layers have resistivity values between those of
the surrounding layers, it is an acceptable model.
Example 4
37. The fourth data set is an example of poor resolution, where the data
cannot resolve the presence of a thin layer or a layer having an intermediate
resistivity. The four layer model consists of three layers with gradually
increasing resistivities overlying a conductive basement (Figure 12). The
sounding curve, however, suggests a three layer model, where the second layer
with a resistivity of 100-m appears invisible. The inversion algorithms
probably cannot identify the presence of the hidden layer. This example will
appear like Example 2, thus the second layer parameters should show a large
32
*c
4$
00
0 .4
0. 0
*4 o
0 .,
* L
0 "
*L 0
- (DP
x 0
E E E
o 0o
E EE -A
Eana
. . . . . . . .. .
I 4I
0 -4
U)xn: )4
LLJJ
'4) eno
41)
rL i
(W 4de
00
-L0
00
* 00
0 ~co
*l 0
< c
LLi
E E.
0.
oc
-4
range of parameter variation which can give equivalent solutions.
38. The automatic interpretation SVES model is the only one that
contains four layers, thus identifying the presence of the hidden layer
(Figure 13b). The model is a good fit to the true solution. The automatic
RESIX program initially obtains a six-layer model but applying pre-inversion
smoothing reduces it to five layers. This inverted model is a good
approximation to the known solution. The user-aided programs only identify
three layers present, mainly because the initial model was inputted by the
user and that initial model was based on the shape of the sounding curve,
which suggests three layers. This example points out an important failure of
resistivity inversion programs - they cannot increase the number of layers in
an initial model, but only reduce them by indicating a very small layer
thickness or a layer resistivity similar to that of an adjacent layer. Three
of the user-aided inversion models are quite similar, although poor, while the
SVES model is very bad (Figure 13a).
Field Examples
Example 5
39. The first field example is a set of data collected for the purpose
of detecting possible ground-water contamination. Both seismic and resistivity
data were collected at this site. The data suggest a subsurface structure
consisting of three layers but a borehole indicates four layers present
(Figure 14). All of the user-aided programs and one automatic program (SVES)
obtain three layer models, most having similar results (Figure 15). The
automatic RESIX inversion model consists of four layers. Without any other
geologic information available, the automatic RESIX model is the favored
interpretation. The resistivity inversion results indicate an overburden layer
as does the seismic data (Figure 14c), however the seismic data identifies an
intermediate interface at 52 feet whereas the resistivity models indicate a
deeper bedrock interface. Using the borehole information and seismic data, the
original interpreter fit a five-layer model and was able to identify the
36
,Ii i I I I I I
obo
* 0°
J"J
L&,! (w 4;do~l
#A -1 -
°I 2
--
_(w) ,i 0
I I
_ _ IbU,
cx10
EA I;.)
> . .
> LJO'.-
0
> to
C - 4
- I II II
I.1
'4
4J
4.4)
0"4
'U C o
< 0 T a E
X 000
o0 0 0 0 0
V4)
01-w~) ju'-dd
AIIIISI08
"6
° ~ 'w
ac
- ------------------------------
In >5
.,., mnc oon

Lii
&0
~0
InI 0
La>LjC
UN ., '
x ao (/o o a
.I- 0
i1i2
14
o' 0 0 0 O
w 3
Li (11)
Li>
C4.Jm
intermediate layer in the resistivity data2 (Figure 15). (For this example
and the ones that follow, the curve designated as "original" is the
interpretation obtained by the original interpreter). However, based on the
resistivity data alone, a five-layer model is not justified.
Example 6
40. This set of data was collected at a landfill area to identify the
boundaries of the landfill. The Schlumberger sounding curve suggests four or
five layers present (Figure 16). For the user-aided programs, a five-layer
model was used to fit the data. The resulting inversion models are fairly
similar, with the major difference being the thickness of the fourth layer
(Figure 17a). The model obtained using DCRESI reduces to four layerE since the
resistivities of the third and fourth layers do not differ by more Chan 10-m.
The automatic SVES program obtains a two-layer model which is a very poor fit
to the data, whereas a good fit to the data is achieved by a five-layer model
using the automatic RESIX program (Figure 17b). The original interpreter also
found that a five-layer model adequately fits the data.
Examples 7.8
41. These two data sets are from the White Sands Missile Range in New
Mexico (Butler and Llopis 1984). The region is semi-arid and the data were
collected over an unconfined alluvial aquifer. Seismic and resistivity data
were collected at both survey sites.
42. The first Schlumberger resistivity data set (SW-19) suggests a four
layer model with the seismic data also indicating four layers (Figure 18).
With the user-aided programs, a three-layer model had the smallest fitting
error using programs DCRESI and SVES, while RESINV and RESIX achieved a best
fit with a four-layer model (Figure 19a). The automatic fitting programs
(RESIX and SVES) obtained an optimum fit with four layers (Figure 19b). When
1Personal
Communication, Keith J. Sjostrom, Civil Engineer, US Army
Engineer Waterways Experiment Station, Vicksburg, Mississippi.
40
EXAMPLE 6
4--
>a
a, S
< o00
1 0 •. .
1 10 100 iee
Electrode Spacing (fi)
Figure 16. Example 6,Schlumberger field data collected at

a landfill site.
41
0
(n ?5c)(
L.i (fl 0)
> Lit0,.
(/.
V I
~~,I
1. 4 - - -- - - - -- - - - -
II-o
0 I I "
0 01 I -
..- '
0 an o 0 in
0 0 t 41
c
(11) q dac]
. 0
CD(u)) qldec]
.............. .......................................... -
LaJ III
cI I ----- I on
(;I 0 e u0
0'
-4
4-3-
0~~~$ (n4l(0(
0~0
L O
0) 0C0D
C)
-lCJ4t
0.)0
a0 0C0
3 0 0 03 0 0
0 0 in 40 in U 0 in
(11) qjdaa]
La
-r
4 4
(0
-4
x4
4
00 W-4
*D
*5
* -4
*0
.4-4
CCu
,4
* 4)
(11-wqo) AI!A!IS!Se8 4u9,ioddv

Nj a n"0 cnm 0
,-4
I In
w
-$44
I @3I
0-
D,- E -C
0 -f -: -- - - - - - 4
4J
Aj
-4 43
I , -I
oi >% 4)
0 a
(11) ti daIi
L U)W
r-4
fn s
-4-4-
C4 on 0d w 0
L) L 0j
-4
@)
-4
o 0 o 0 0E
- -~ o.,
(~) fl
q~dex
6
using the automatic RESIX program, the initial model estimate met smoothing
criteria (ii), however, the fitting error of the three-layer model exceeded
the 10% limit (criteria (iii)), thus the three-layer model was rejected. All
inversion results are similar, although the DCRESI first layer resistivity is
lower than the other models and the user-aided SVES is a rough approximation
of the four-layer models. The resistivity models identify the overburden and
two intermediate interfaces, as does the seismic interpretation. The first
intermediate interface determined from the resistivity inversion is
approximately 80 feet shallower than that determined from the seismic data,
and the lower resistivity interface is about 85 feet deeper than the seismic
interface. The two methods measure different properties so it is not
surprising that the interface depths do not coincide. The four-layer models
(RESINV, automatic SVES and RESIX) are comparable to the original
interpretation (Butler and Llopis 1984).
43. Resistivity data set T-14 indicates four or five layers while the
seismic data identifies three layers (Figure 20). The user-aided and automatic
inversion results are given in Figure 21a and 21b, respectively. Programs
DCRESI and SVES (user-aided and automatic) found a four-layer model to best
fit the data whereas a five-layer model was optimum using the other programs
(RESINV and both RESIX). The original interpretation consists of five layers.
The second interface in the five-layer resistivity models correlates well with
the first seismic interface at a depth of approximately 15 feet. The two
methods detect the effects of the water table at much different depths;
seismic 96 ft, resistivity 40-50 ft. Again, this is due to the different
subsurface properties each technique measures. The deepest interface the
resistivity models identify is due to a change in ground-water salinity
(Butler and Llopis 1984). Since there is not a significant density contrast,
the seismic method is unable to detect this interface.
Examples 9. 10, 11
44. The following three data sets were collected in a desert region of
Egypt (Butler et al. 1990). The purpose of collecting this data was to
identify possible ground-water sources. Seismic data are not available to
correlate with the resistivity data but borehole information is available for
45
>L.
ina: 0
4.)
-4-- .- 4
In I-
-4-4o
4)
-. .. .. .. .... ....
00 oW on on -r4
(11) L41dea -r4

Li 0~
-J ri*
x oxtnc/ 0
II 0
in 0 inAn 0 i
Nt
correlation with sounding VES-2.
45. A three or four layer structure is suggested by the first data set
VES-2. A borehole distinguishes five layers of alternating sand and clay
(Figure 22). The user-aided programs DCRESI, RESINV, and RESIX achieve a best
fit to the data with a three-layer model, while a four-layer model is best
using SVES (Figure 23a). Both automatic modeling programs (SVES and RESIX) fit
a four-layer model (Figure 23b). The first and second layer resistivities of
the SVES model are nearly the same plus the second layer is very thin, so this
model should reduce to three layers. The first two layer resistivities of the
RESIX model are also similar and meet smoothing criteria (ii), however, the
fitting error of a three-layer model increases more than 10% (criteria (ii)),
so the four-layer model is kept.
46. All models have basically a high, low, high resistivity structure.
The low resistivity layer corresponds to the zone of saturation. The original
interpretation fit five layers to the data (Butler et al. 1990). The six
inversion results are a rough approximation to this original interpretation,
with the automatic RESIX model fitting quite well.
47. Sounding VES-3 also infers a three or four layer earth (Figure 24).
Both the user-aided (Figure 25a) and RESIX automatic programs (Figure 25b)
find that a four-layer model adequately fits the data. The SVES automatic
inversion fits a three-layer model. The user-aided programs exhibit an
alternating high and low resistivity structure as does the original
interpretation (Butler et al. 1990), which consists of five layers. However,
the third layer is a smoothed representation of the third and fourth layers in
the original interpretation, and is much thinner. The automatic inversion
models (RESIX and SVES) do not fit the original interpretation as well, having
a low-high resistivity sequence.
48. A subsurface structure consisting of three layers is indicated by
the sounding curve VES-4 (Figure 26). All user-aided inversion programs were
used to fit a three-layer model. The automatic SVES program also fit three
layers but RESIX fit a four-layer model (Figure 27). The layer resistivities
of all models are similar to the original four-layer interpretation (Butler et
al. 1990), but the total layer thickness is about 35 feet greater.
48
-4
UlU
030
00 -4
0)D 0)(
(40w
C4 0
o03
-1 C
o~~~b 03
-. ,4. cov-
E
0 -4
4 54
c) bfl
-4
EUD
41
U 4)
XV
I~LI ~a'o
(Lu-w~o)
AAIJSI~q~4 "
4UG.od0
o01
(n cr
-4
-I.-
E .En
4))c
04.
0~~U 0 0
> 0 4)4.
0)~~~~ V)- ( 3 o q .
44
0~0
Lii ---------
OC, C'
(W) t(Ne
EXAMPLE 10
E
E"
-
0
Q)
U¢)
0)
rw
C0
I
L
a
0.
0<
<1............I ........ ,''....I
10 100 100
Electrode Spacing (m)
Figure 24. Example 10, Schlumberger sounding curve VES-3. Data

collected in a desert region of Egypt (Butler et al. 1990).
51
> Li
E r. ,
-0 C3 I 4)
04 40 Go G 4-0
41)
1) -4
~K V
(w) 6j~V).2)
LJ
~co
,4
~~:j
I(---t9:
on 40 0
qld~a-4
EXAMPLE 11
E
I "
E
-C 10 ' •
C1)
<.
110 1 00 '0O0
Electrode Spacing (m)
Figure 26. Example 11, Schlumberger sounding curve VES-4. Data

collected in a desert region of Egypt (Butler et al. 1990).
53
LV ) -C
> Li .Z
Wm) 0
'Io
E0
(A0
:>0
"4
L. - 4
II
_o II
o0 -4-A
go I, C
--4
CL G)) n n04J
Eno
--4
0)
r -4
> Cd
0
C -V
L~ '-4
0 0
C4 ID coe
(w) tN-a
PART VI: SUMMARY AND CONCLUSIONS
49. It is evident from the examples presented that there are differences
in the various algorithms available for interpreting d.c. resistivity data.
Overall, both the user-aided and automatic interpretation schemes were able to
find a model that had some similarity to the true or originally interpreted
solution. Table 2 summarizes which inversion algorithm was best at modeling
the various data sets. Of the user-aided inversion programs, DCRESI and RESINV
performed equally well, with RESIX close behind. RESIX out performed SVES when
using the automatic interpretation option. DCRESI is no longer commercially
available and does not contain the required automatic interpretation or
equivalence analysis routines. The program RESINV also does not contain these
routines and it is not in a menu driven format. It would be possible for the
WES personnel to revise RESINV to meet our needs, but this would require many
man-hours. The only inversion program that performed well and contains the
majority of the desired attributes is RESIX PLUS (refer to Table 1). At the
present time RESIX does not have direct data input capabilities. However, this
in not a major obstacle and two solutions are possible. The company which
wrote RESIX, INTERPEX Ltd., is considering writing an interface to allow
direct data input, or it is possible that WES personnel could perform this
task. Based on the performance of the various inversion algorithms evaluated
and the number of desirable attributes each program contains, it is
recommended that the program RESIX PLUS (INTERPEX, Ltd., Golden, Colorado) be
used as the resistivity interpretation program in the Water Supply Program.
55
Table 2
Summary of Best Resistivity Interpretation
Program For Modeling Each Data Set
EXAMPLE USER-AIDED AUTOMATIC
1 all RESIX
2 DCRESI RESIX
3 DCRESI/RESIX RESIX/SVES
4 none RESIX
5 RESIX RESIX
6 none RESIX
7 RESINV RESIX/SVES
8 RESINV/RESIX RESIX
9 DCRESI RESIX
10 RESINV none
11 SVES SVES/RESIX
56
References
Butler, Dwain K., Gangi, Anthony F., Wahl, Ronald E., Yule, Donald E., and
Barnes, Donald E. 1982. "Analytical and Data Processing Techniques for
Interpretation of Geophysical Survey Data with Special Application to Cavity
Detection." Miscellaneous Paper GL-82-16, U.S. Army Engineer Waterways
Experiment Station, Vicksburg, Mississippi.
Butler, Dwain K. and Llopis, Jose L. 1984. "Assessment of Two Currently

Fieldable Geophysical Methods for Military Ground Water Detection," Military
Hydrology, Report 6. Miscellaneous Paper EL-79-6, U.S. Army Engineer Waterways
Butler, Dwain K., Llopis, Josd L., Yule, Donald E., Sharp, Michael K., and
Dardeau, Elba A. Jr. 1990. "Water Detection Response Team Geophysics Element
Case Histories." Technical Report GL-90-23, U.S. Army Engineer Waterways
Davis, P.A. 1979a. "Development and application of resistivity sounding

inversion for several field arrays." M.S. thesis, University of Minnesota,
Minneapolis, Minnesota.
1979b. "Interpretation of resistivity sounding data-Computer

programs for solutions to the forward and inverse problems." Information
Circular 17, Minnesota Geological Survey, St. Paul, Minnesota.
Department of the Army 1979. Geophysical Exploration. Engineer Manual EM

1110-1-1802, Office of the Chief of Engineers, Washington D.C.
Falls, T.C., Butler, C.D., Webb, B.T., Gates, B.Q., Williamson, J.L.,
Turnage, G.W., and Chase, A. 1991. "ALBE Geographic information system/user
interface/graphics, Volume 1 programmer's guide (version 1.0)." Miscellaneous
Paper GL-91-, U.S. Army Engineer Waterways Experiment Station, Vicksburg,
Mississippi.
57
Inman, J.R. 1975. "Resistivity inversion with ridge regression." Geophysics,
Vol 30, pp 798-817.
Koefoed, 0. 1976. "Progress in the direct interpretation of resistivity

sounding: an algorithm." Geophys. Prospect., Vol 24, pp 233-240.
Marquardt, D.W. 1963. "An algorithm for least-squares estimation of nonlinear

parameters." J. Soc. of Industrial and Applied Mathematics, Vol 11, pp 431-
441.
Mooney, H.M. 1979. "Handbook of engineering geophysics, volume 2--Electrical

resistivity." Bison Instruments, Inc., Minneapolis, Minnesota.
Orellana, Ernesto and Mooney, H.M. 1966. "Master tables and curves for
vertical electrical sounding over layered structures." Madrid Interciecia.
Yule, Donald E. and Sharp, Michael K. 1990. "Seismic refraction data

processing software: SEISMO version 2.7." Instruction Report GL-90-2, U.S.
Army Engineer Waterways Experiment Station, Vicksburg, Mississippi.
Zohdy, Adel A.R., and Bisdorf, Robert J. 1989. "Programs for the Automatic
Processing and Interpretation of Schlumberger Sounding Curves in Quickbasic
4.0." U.S.G.S. Open-File Report 89-137 A.
58
APPENDIX A: MILITARY REQUIREMENTS FOR GEOPHYSICAL
GROUND WATER DETECTION AND EXPLORATION*
"Reprinted from Proceedings, Symposium on Surface and

Borehole Geophysical Methods in Ground Water Investiga-
tions, Fort Worth, TX, 1985 (Feb.) National Water Well
Association, pp 228-248.
MILITARY REQUIREMENTS FOR GEOPHYSICAL GROUND WATER
DETECTION AND EXPLORATION
Dwain K. Butler
U. S. Army Engineer Waterways Experiment Station

Vicksburg, Mississippi
Abstract
Adequate water supply is a critical requirement for support of

military operations in arid and semi-arid regions and for fixed military
bases. Ground water exploration typically will utilize all available
information to aid the interpretation of geophysical survey data and
produce an integrated assessment for an area. Situations are envisioned,
however, in which little or no supplementary information will be available
to aid or constrain the interpretation of geophysical survey data.
For this latter case, information about ground water table depth, aquifer
thickness, and water quality is required expeditiously at selected,
perhaps widely separated, locations. Ground water detection is a term-in-
ology properly applied to rapid ground water assessments at selected,
widely-spaced locations. Case histories are presented illustrating
both ground water exploration and detection. A ground water detection
study at five locations on White Sands Missile Range, New Mexico, illu-
strates the application of seismic refraction, electrical resistivity,
loop-loop low induction number electromagnetic (El), and transient EM
methods. Results of the geophysical methods are compared to known geo-
hydrological conditions.
Backgrou d
Ground water detection methodology is the subject of several research
projects at the U. S. Army Engineer Waterways Experiment Station (WES).
The methodology comes under the field of military hydrology, which is
a specialized field of study dealing with the effects of surface and
subsurface water on the planning and conduct of military operations.
Responsibility for management of a Military Hydrology Research Program
was assigned to WES by the Office, Chief of Engineers. Ground water
detection is part of the water supply thrust area; other thrust areas
are weather-hydrology interactions, state of the ground, and streamflow.
A2
There is no device or black box that can be set on the ground at
a given location and, with just the press of a button, determine with
a 95-percent probability that potable ground water is present at a depth
of X feet. Even in the foreseeable future, there is little likelihood
that such a device will be available either in this country or elsewhere.
In the majority of cases, ground water is usually detected as a matter
of course in field investigations not specifically intended for ground
water exploration. A Ground Water Detection Workshop was held at WES
in January 1982. It was attended by Department of Defense representatives
interested in improving military capability to develop and exploit local
water sources to support military operations in arid regions. The conclu-
sions of the Geophysics Working Group at the Ground Water Detection
Workshop were: (a) there are two currently "fieldable" geophysical
methods, electrical resistivity and seismic refraction, that are applicable
to the ground water detection problem and may offer a near-term solution
to the need for ground water detection capability, and (b) there are
several state-of-the-art and emerging geophysical techniques that may
have potential in the far-term for application to the ground water detec-
tion problem. The near-term solution, i.e., the use of currently fieldable
methods, has the potential of significantly reducing the risk of dry
holes during water well drilling operations, but the field operations
are somewhat cumbersome and time-consuming for possible deployment in
support of forward area operations. Development of one or more of the
emerging geophysical techniques offers the possibility of delivering
something closer to the desired capability than the near-term methodology.
Geohydrological Models
Geophysical exploration for ground water refers to surface remote

sensing techniques as shown in Figure 1. The objective of the geophysical
surveys in ground water exploration is the determination of subsurface
structural or stratigraphic indicators of the presence of ground water
I. Direct Methods
A. Drilling
B. Surface Reconnaissance
II. Indirect
A. Aerial/Satellite Remote Sensing Methods

Objectives: Structural, Geomorphic, and Vegetative
Surface Indicators of Ground Water
Occurrence.
B. Surface Remote Sensing (Geophysical) Methods
Objectives: Structural, Stratigraphic, and Aquifer
Property Subsurface Indicators of Ground
Water Occurrence.
Figure 1. Methods for ground water exploration
A3
or the measurement of a parameter that is an actual physical property
of the aquifer itself. The indicators are indirect clues to the presence
of ground water. A physical property of the aquifer itself could be
a more direct clue of the presence of ground water. It is important
to be aware of the various ways in which usable quantities of ground
water may occur in the subsurface. Ground water occurrence can be illus-
trated by models which illustrate unconfined aquifers (Figures 2 and
3), confined aquifers (Figure 2), perched water (Figure 3), and water
which is concentrated along fracture zones in otherwise nearly impervious
rock (Figure 4). As suggested by Figures 2, 3, and 5, more than one
of the above models or conditions will more than likely occur at a given
site.
Detection Versus Exploration
Geophysical methods are routinely used throughout the world in

exploration programs for the assessment and development of ground water
resources. The geophysical methods that are predominantly used in these
ground water exploration programs are gravity, electrical resistivity,
and seismic refraction methods. Although occasionally only one of these
methods will be used in an exploration program, generally at least two
of the methods are used in a complementary approach. A geophysical
ground water exploration program will normally use all available borehole
and other geological data in order to produce the best possible assessment
of the ground water potential and conditions in an area.
The primary objective of geophysical ground water exploration is

the mapping of subsurface structural and stratigraphic indicators of
the possible occurrence of ground water, such as buried river channels,
fracture zones in bedrock, confining layers (aquacludes), etc. Actual
detection of the ground water table with any of the geophysical surveys
may be noted but may not be of primary importance in the overall ground
water exploration assessment. Figure 6 is an example of the use of
the seismic refraction method to delineate a buried channel in an arid
region in western Kansas; identification of material type was made by
correlation with exploratory borings near each end of the profile.
In this example, the water table was actually detected by the occurrence
of the characteristic seismic velocity (to be discussed later in this
paper) in the central part of the survey profile. However, even if
the ground water table had not been detected in this example, the strati-
graphic indicators would dictate the greatest ground water potential
for a well placed in the center of the subsurface channel.
The expression "ground water detection," in contrast to ground

water exploration, applies to the concept of actually detecting the
presence (or absence) of ground water and the depth to the water table
beneath a given "point" on the surface by conducting one or more types
of geophysical tests at that point. In the ideal case, the aquifer
thickness and water quality would also be determined. For some cases,
information regarding ground water occurrence and other geological factors
might be available but, in general, the assessment of the presence of
ground water must rely solely on the geophysical results at the given
surface location in the detection scenario. It is envisioned, however,
that many times the geophysical ground water surveys would be conducted
A4
RECHARGE
IAREA 8ESRT
GROUND URFACE
WATRER&
PERCHEDW
WATERBTABLE
AQUIFER
UUNCONFINED
Figure . Hydrogeological model of erch nedantbl
unonindaqufr
FRA CTURE
• = PLANE
%FAULT
ZONE O ""JOIN/o.
CONCENTRA TION j
Figure 4. Hydrogeological model of ground water

concentrated on fracture zones
0 PERCHED
] ALLUVIUM
CLAY1wEATHERED TRANSITION ZONE
SANDWLAYER IMPERMEABLE ROCK
O WELL LOCATION TO TAP GROUND WATER IN FRACTURED ROCK

G WELL LOCATION TO TAP PERCHED GROUND WATER
WELL LOCATION TO TAP UNCONFINED AQUIFER
L L IL LI L L LL L
5. Hydrgeological model illustrating multiple
Figure
of ground water occurrence

modes
A6
3300
U_ -O
- SN
, ,IO0SF
Figure 6. Example of water table detection and of delineation of a

buried channel in western Kansas by the seismic refraction method
to aid in choosing between alternate sites in an area already identified

as having good ground water potential by other methods. Of the three
geophysical methods most comonly used in ground water exploration programs,
only two, electrical resistivity and seismic refraction, are applicable
to the ground water detection problem. Figure 7 sumarizes geophysical
methods and their present or projected applicability to ground water
exploration and/or detection programs. Detection principles for the -
electrical resistivity and seismic refraction methods are discussed below.
Detection Principles
Electrical resistivity method
The electrical resistivity mthod applicable to the ground water

detection problem is vertical resistivity sounding, where the objective
is to make electrical measurements at the surface from which the vertical
variation of electrical resistivity with depth can be interpreted.
The resistivity of a material is a fundamental geophysical property
of the material. Although the range of resistivities of geological
materials is that of the order of 1020 ohm-n, the range commonly enc oun-
tered in ground water exploration and detection is typically 10 ' ohm-rn.
Most soils and rocks conduct current primarily electrolytically,

i.e., through interstitial pare fluid. Thus, porosity, water content,
and dissolved electrolytes in the water are the controlling factors
in determining resistivity rather than the soil or rock type. A major
exception to this generalization are clays, which can conduct current
both electrolytically and
bulk resistivity of aelectronically.
soil or rock and The
the general relation between
porosity (volume
fraction), pore fluid saturation S (volume fraction of b), and pore
fluid resistivity 0 can be expressed by the following empirical equation:
0 b ao -ms-n
A7
Ground Water Ground Water
Geophysical Method Detection Exploration/Assessment
Seismic Refraction X X
Seismic Reflection
(Profiling) X
Seismic Reflection
(V /V Sounding) X
Electrical Resistivity X X
Gravity X
CW Electromagnetic (EM) ? X
Transient EM X X
Pulse "Radar" EM ?
Magnetic X
Airborne (Gravity,
Magnetic, EM) X
Figure 7. Summary of applicability of geophysical methods

to ground water exploration and detection
where a , m , and n are constants which depend on the soil or rock

type. Below the water table Sw w 1 (100 percent saturation). Qualita-
tively, equation 1 indicates: (a) as porosity increases, bulk resistivity
decreases; (b) as pore fluid saturation increases, bulk resistivity
decreases; and (c) as pore fluid resistivity increases, bulk resistivity
increases.
A common and successful use of resistivity sounding is for detecting

the fresh water/salt water interface, which will always be indicated
by the occurrence of a prominent resistivity decrease. Detection of
the water table itself is a more difficult problem. Under favorable
conditions, the water table will be detected as the top of a conductive
or less resistive layer; since, except for unusual conditions, even
fresh potable ground water is much lower in resistivity than the dry
aquifer material. The most favorable conditions will be when the water
table occurs in unconsolidated sediments with little clay content.
Dry silts, sands, and gravels will have resistivities of 300 ohm-m and
greater; for fresh water, the resistivity at the water table will typically
decrease to a range of 20 to 100 ohm-m in areas like the southwestern
United States. In sediments with considerable clay content, the resistiv-
ity contrast will be mach smaller and may be undetectable. At the fresh
water/salt water interface, the resistivity of the aquifer will decrease
considerably, perhaps to < 1 ohm-m. Zohdy et al. (1969, 1974) adopted
a qualitative criterion of o -10 ohm-m to differentiate fresh from
saline ground water conditionsb. in a large ground water assessment program
A8
at White Sands, New Mexico. Clays can have resistivities intermediate
to the resistivities of highly saline and fresh aquifer conditions.
Seismic refraction method
The seismic method applicable to the ground water detection problem

(in the near-term) is the refraction method. From a seismic refraction
sux',ey at a given location, it is possible in principle to determine
depths to interfaces between materials with contrasting bulk density
and seismic velocity and to determine the seismic velocities of the
different materials. Generally, only compression-wave (P-wave) velocities
are easily determined from seismic refraction surveys.
The physical principle involved in the detection of the water table

by seismic methods is that the P-wave velocity of saturated sediments
is considerably greater than the same sediments in dry or only partially
saturated conditions. Typically, the P-wave velocity will increase
from 300 - 700 m/sec to 1375 - 1675 m/sec at the water table, where
the water table occurs at shallow depths (< 30 m) in unconsolidated
sediments (silts, sands, and gravels). The occurrence of a characteristic
1,500 m/sec velocity at shallow depths at a site is generally strongly
indicative of a ground water table, although some weathered rocks and
massive clay deposits can have this velocity also.
If the water table occurs at greater depths (> 30 m, for example),

the sei-mic velocity of the saturated sediments can be as high as 2,300
m/sec; but in these cases, the velocity of the unsaturated sediments
just above the water table can be as high as 1,200 m/sec. The smallest
velocity contrast at the water table will occur in very fine-grained
sediments, where the velocity contrast can be as small as 150 m/sec.
When the water table occurs as an unconfined surface in rock, there
will always be a velocity increase at the water table, but it may be
small. Where the ground water occurs in a confined rock aquifer, there
may be little in the seismic data to suggest the presence of ground
water without independent or complementary information. Whether the
water table in an unconfined aquifer will be detected or not depends
on the thickness of the saturated zone above high-velocity rock. In
some cases, where the contrast in seismic velocity between rock and
saturated sediments is large and the saturated zone is thin relative
to its depth, the water table refraction will not be detected in an
"ordinary" seismic refraction interpretation (blind zone problem).
Complementary methods
The resistivity and refraction methods are complementary in the

sense that they respond to or detect different physical properties of
geologic materials. Both methods can detect the water table, hence,
the presence of ground water under certain conditions. In cases where
both methods detect the water table, one method serves to confirm the
results of the other method or to resolve ambiguities. Certain conditions,
however, such as the presence of a fresh water/salt water interface,
can be detected by one method but not the other.
A9
When depths to interfaces determined by geophysical methods are
compared to "ground truth data" from nearby boreholes, typically the
agreement is within +107 for the seismic refraction method and +20%
for the electrical resistivity method. Of course, the difference between
the actual interface depth and geophysical interface depth can occasionally
be greater due to the effects of blind zones and velocity inversions
(departures from the normally assumed case where seismic velocity increases
with depth) in seismic refraction interpretation and highly equivalent
solutions in electrical resistivity interpretation. The problem of
geophysical determination of the water table depth is complicated by
the physical nature of the "interface." The "geophysical interface"
commonly may be somewhere within the capillary zone, the velocity and
resistivity interfaces may be different, and neither may agree with
the standing water depth in a borehole (and the standing water depth
itself may be different from the actual water table). The difference
in geophysical and borehole water table depth determinations will be
greatest in fine-grained sediments and least in coarse-grained sediments.
Emerging Technology
An advancing technology is the use of seismic reflection methods

to determine both compression (Vp ) and shear-wave (Vs ) velocities from
primary reflection records (collections of all geophones receiving signals
from a single source location). Thus, both compression- and shear-wave
interval velocities can conceivably be determined from a single 'split-dip"
spread setup, although different sources might be required to generate
separate compression- and shear-wave reflection records. In this procedure,
V /V ratios would be determined as a function of depth and, due to
t~e fact that shear-wave velocities are generally much less affected
by water saturation than compression-wave velocities, the Vp /Vs profile
should be highly indicative of the occurrences of ground water. Because
only a single reflection spread setup is required, the logistical complexi-
ties associated with the continuous reflection profiling procedure are
avoided.
Electromagnetic (EM) methods
If there is ever a device that even comes close to the "black box"
water detector ideal, it will likely be an EM device. There are numerous
EM techniques ranging from near-DC induction techniques to GHz wave
propagation techniques. Hopefully, some innate property of the aquifer
system will ultimately be amenable to interrogation or probing by an
EM technique and allow direct ground water detection. Direct ground
water detection, however, must be viewed as a long-term goal, and the
immediate application of the EM methods is as a replacement or supplement
to electrical resistivity in a comniementary exploration or detection
program.
There are several EM techniques such as magnetotellurics and various

types of low frequency, continuous wave inductiun (CWEM) methods that
can be used to determine resistivity or conductivity as a function of
depth. Compared to the electrical resistivity techniques discussed
A1O
previously, these EM techniques can be more rapid and less logistically
cumbersome, and they do not require surface contact.
One of the most promising of the emerging technologies is the transi-

ent electromagnetic (TEM) method. In the TEM method, a very broad band-
width EM signal is input to the ground and, because the signal is transient
(i.e., not a continuous wave source), very high power levels are possible
and measurements can be made during the off-time of the transmitter.
The return signal is interpreted to give resistivity as a function of
depth. The exciting aspect of the TEM method is that as many as 20
soundings per day can be conducted under favorable conditions. The
TEM method still has the same non-uniqueness as any other method used
to determine resistivity as a function of depth; however, the TEm method
has superior vertical and lateral resolution and is less effected by
lateral variations than electrical resistivity and other EM methods.
Ground Water Detection Field Trials
Two field sites were selected as representative of two common aqui-

fers: an unconfined alluvial aquifer and a confined (artesian) rock
aquifer. White Sands Missile Range, New Mexico, was selected as the
alluvial aquifer site, Rnd Fort Carson, Colorado, as the confined rock
aquifer site. Geophysical investigations at the field sites were conducted
in two phases. In the first phase, electrical resistivity and seismic
refraction surveys were conducted at five widely separated locations
at White Sands and at one location at Fort Carson. During the second
phase, CWEM surveys were conducted at the five locations at White Sands
and at Fort Carson, and TEM surveys were conducted at four of the White
Sands locations. This paper will specifically address selected results
from the White Sands locations where all four geophysical techniques
were applied. Complete details about the field test sites and the results
of the first phase of field investigations are given by Butler and Llopis
(1984), and results of the CWEM surveys of the second phase are given
by Butler (1984).
Figures 8 and 9 illustrate the results of seismic refraction and

electrical resistivity surveys at the SW-19 location at White Sands.
The geophysical models resulting from the data in Figures 8 and 9 are
shown graphically in Figure 10. A ground water assessment or geohydro-
logical model is deduced from the geophysical models using the detection
principles discussed earlier. The interpreted geohydrological model
for SW-19 is shown in Figure 10.
Geophysical ground water assessments for all five locations at

White Sands are summarized in Table 1. The known geological and ground
water information about the five locations are summarized in Table 2.
Comparison of Tables I and 2 indicates general qualitative agreement
between the geophysical ground water assessments and the known ground
water data for all the locations except HTA-I. The predicted water
table depths are consistently too shallow, however, compared to borehole
water depth measurements, by amounts ranging from 12 percent at SW-19
to 28 percent at B-30 and T-14. Direct application of the detection
principles resulted in misidentification of the water table in the case
All
see 5eO f N 5'W
488
' 2ee
TRUE VELOCITIES. fps 0
V 1: 1375
!/2-2800 c
Iee 6 :7 V3=3700
V4-7400
C 4,
d- WELL 10 d=12"
e soc. . 'XYz
388 688 98 1288 1588 1888 218
DISTANCE, ft
Figure 8. Example of seismic refraction results, SW-19 site,

White Sands, New Mexico
19' Initial 4adel
final '4adel
0 Field Data
I-
.
a
0 -
is , t ! I I I toIt I i t l ll I I i i i s'I
L. FT
Figure 9. Example of resistivity interpretation procedures for

SW-19 site, White Sands, New Mexico
A12
RESISTIVITY VELOCITY INTERPRETATION
0 ohm-It ft/sec ________
600
1900 MOIST. LOOSE
- - - -- - -- -- SEDIMENTS
100
200- MOIST. DENSE

753700 SEDIMENTS
40-
30070 SDMET
400-
Figure 10. Geophysical models and interpretation for the SW-19 site,
White Sands, New Mexico
Table 1
Summary of White Sands Geophysical Ground Water Assessments
Predicted
Water Predicted Confidence in
Table Water Quality Aquifer Ground Water
Location Depth, Ft Statement Thickness Assessment
HTA-1 8 Fresh .100 ft Poor

B-30 65 Fresh from 65- ?Fair to Good
125 ft, becom-
ing very saline
below 125 ft
T-14 95 Fresh from 95- ?Poor to fair
150 ft, becom-
ing saline
below 150 ft
MAR 160 Fresh from 160- Base of Fair
300 ft; very aquifer,
saline from 1000 ft
300-1000 ft
SW-19 400 Fresh ?Very Good
A13
c0 0 9000 0
w ~ u-0 ,
-,4
CO 00 tor ~ 3 ,0
00 10 0 0 .
-4 WS.= -4 S0"=-
M 41 0.a ~0)
40 C)
WE 0 W 0 -4 W0 04C
Air-60 A 60 0 0
%D A
ev 0 cc ) 0 0 00
.4
C A; C.;
0 4.5 4
-4 MJ
>- -4 8
Ai W. I a 00 -4 W 444 E
to00 .C co U4 0 (A -
aE 6 - A Ao'j C)C 00
5- C4 00J
U4 = %&4C 00~
&0
Li 4 -
00iGo4
0k
-o3
0
I'
-4 .tn
4C4 -1 Co 0
0
0) cn.4 c 0j 4.003tomo 0-
0o - bg-4 to cca a -4 CCVC-40C-
0C. " 1
$.4 % c l -4 13
AJ C 0U cc1 CC 00 DOM r0r0 A U'
- w 4 D -t0A
;-
01 0)4 JW 0044 0 Z. 0 W
-4.C . 4 0.4
0 -0 0d r. J 0 - -4 A-4 ca i -4
0 07 Ai 0a Ai 0
cI 0 0) * 0 .
C > -4-S C- 4- 40
0 i 0 >% "4 a- ca cc
0X
C >.1J A£ 0JA
3 U4 .4 4 04 44 04U0
0J :> > a iAj~C~0
X 004
-
cc, -. A0 I- a Go 0 -40
4. =0 0 c' a 0 r
0 w- 004
C4-
0 00
.E 0 AJ
EA
.0- cc . %0
44
0 U -
to' co -e
W1~ C.. 00 U-1A-

-
S 0 -- Cc0
" r- - 0I
ca. IN '.
'0 -4 00
- 0
-E
0 0
4u <-0
o -
-3 ~49
A14
A4
4 4 C)h
so o1O ('
-Li 4* ca
0 00c 1
41 ed :) ), --- I
w1 cc0-4-cn fn
0
go ca 0 -I a0 AiA
r= - _ 0 - 0*g 0.
w4 4) .0 u- 0a
0 4J :P -4~ $40 -
0 o 410 d>
-0 - V4 W. 2-0, -. 0 co-4
C 0 -4 toC4 -4 &1 $4
u* a 4- AJ 50
w 00
-4 hi1 .1 041ci 410
00 41 -. 4 CD44 C14 Ai
0 an.('4 anC
00 cc w 4 0 C1so J
-d -4 w1-4 1-4wci 4 U4 0U0 c
(,... D > I > AIj 4 S- 0
414 ca X4o k. ' 4; 2. 0 4- 0 0
W-L. 0 - 1f
00
ot
A-S 04i 0%4
Ihi WJ AJ. 4 hCic

.4 w -4. 0 ' fn4'404ri
X Ai e I 'D44
4
cc0 A. 0 . 1% C4 " 40
0 0 ,0 *'O : 0 \4
-4
Ai
-0cJ CI
V0''
O j G
ca 0000
54,4C4 C'J04
41 t
A 15
Rho late stage (aim-.)
10. 100. 1k 10k
I
CC
a.
* VC
0.0
CL
- 0
CV
0.0
30o.
100.
Figue 1. moel
Rsistvit or S-19sit corespndig t th
inFigue cure
soli 1
oA17
Table 3
Comparison of Measured and Predicted
Depths at Three White Sands Locations
B-30 T-14 SW-19

Measured Water Table
t
Depth (m) 27 40 138
Predicted Water Table
20** 29** 122
Depth (m)
WES Electrical Resistivity
Interface* (m) 38(30-46)t 46 122 tt
USGS Electrical Resistivity

Interface* (m) 46(39-57)t 49
TEM Resistivity Interface (m) 30 50-62 t 122 t t f
* Selected White Sands data were also interpreted using a USGS

inversion program.
** Based on seismic refraction model.
t Range of model predictions for equivalent solutions.
tt At production well.
tft 150 m from production well.
between the predicted water table and resistivity interface depths. For
T-14, the TEM interface agrees with the electrical resistivity interface;
while for B-30, the TEM interface depth is within 10 percent of the
measured water table depth. The TEN interface for SW-19 agrees exactly
with the seismic refraction and electrical resistivity interfaces.
The TEM method fulfilled all expectations regarding ease and rapidity
of field use and depth of investigation capability. Although the TEM
method is not a stand-alone ground water detection device, it is a possible
replacement for electrical resistivity in a complementary geophysical
ground water detection methodology. The primaty problem with the TEM
method currently is the lack of comonly available interpretation tools.
There are only limited numbers of master curve solutions available.
Also, even the direct TEM multi-layer response problem requires a minicom-
puter, and the USGS multi-layer inverse program currently operates on
a VAX 11/780. Hopefully, inverse programs can be configured to operate
on the emerging "super-microcomputers."
Conclusions
Based on the results of this work and other work reported in the
literature, the following conclusions are made regarding the applicability
of a complementary geophysical methodology for ground water detection:
A18
a. For cases in which the water table occurs in coarse-grained
sediments (sands and gravels), the geophysical methods can
be used very successfully for ground water detection.
b. For cases in which the water table occurs in fine-grained sedi-

ments (clayey sands, silts, silty clays, sandy clays, etc.),
the geophysical methods can be used for ground water detection;
however, the interpretation will sometimes not be as straightfor-
ward as for case a, and the difference between predicted and
actual water table depth can sometimes be much greater than
for case a.
c. A fresh water/salt water interface is easily detected by the

electrical resistivity method or TEM method, but will not show
as an interface in seismic refraction results; detection of
this interface is useful in that any fresh water present will
be shallower than the interface depth.
d. Rock aquifers can be detected by the geophysical methods, but

there may be nothing in the survey results to differentiate
a rock aquifer from an unsaturated rock unit (except for the
case where the rock unit has high resistivity, in which case
the unit is not an aquifer).
e. For some field situations, such as at the Fort Carson site,

topographic variations and complex, lateral geologic changes
make a straightforward data interpretation impossible.
f. In some cases, such as the HTA-I location at White Sands, the

straightforward interpretation method can lead to false identifi-
cation of the water table.
g. In order to be conservative when specifying drilling depths,

geophysical water table depth estimates should be increased
by 30 to 40 percent.
h. It is envisioned that the desired depth of investigatiorn will

probably be dictated by considerations such as maximum desired
drilling depth or maximum probable depth to water in an area;
geophysical ground water assessment productivity is strongly
dependent on depth of investigation.
The conclusions of the study can be summarized as follows: Comple-

mentary seismic refraction and electrical resistivity surveys (a) can
generally be used successfully for ground water detection when the water
table occurs in unconsolidated sediments, and (b) can generally not
be used successfully for detection of ground water in confined rock
aquifers. For the case of rock aquifers, a ground water exploration
program is required. The complementary geophysical methodology currently
fieldable consists of seismic refraction and electrical resistivity
methods. In the near future, the TEN method may advantageously replace
the electrical resistivity method.
A19
Military Deployment of Geophysical
Ground Water Detection Capability
Development of ground water detection and assessment capability

in the military is developing in conjunction with water well drilling
and production capability. Geophysical methodology will never be applied
in a stand-alone mode but always as part of an integrated system approach.
Figure 13 illustrates a possible flow sequence for field deployment.
The key problems which must be addressed are the skill levels required
for the geophysical survey teams and the organizational structure.
If the decision is made to develop a geophysical ground water detection/
exploration capability in or for the field military forces, the following
options are considered feasible:
a. Recruit or assign junior officers with degrees in geology,

geophysics, or other science/engineering fields with strong
geoscience backgrounds to teams which receive intensive special-
ized training.
b. Utilize teams with special training to conduct surveys and

then relay data to a rear area interpretation unit or data
analysis contractor that could handle data from several survey
units and be better able to incorporate information from ground
water maps and data bases into the ground water assessments.
c. Develop geophysical survey expertise in National Guard or Reserve

units which already have identified professional geoscience
expertise.
c. Establish arrangements with Government agencies and/or geophysical

firms for on-call geophysical testing and interpretation services
for areas that are reasonably secure; these personnel should
have full access to ground water maps and data bases. A quick-
reaction team is a possible approach.
It is important that the military track and contribute to research

and development on stage-of-the-art and emerging geophysical techniques
for ground water detection, such as frequency-domain and time-domain
electromagnetic methods and the concept of determining the ratio of
compression wave to shear wave seismic velocities as a function of depth
as a ground water indicator. Another important area is the development
of training manuals and programs for geophysical survey operators and
for geophysical ground water interpretation procedures. The ultimate
goal is the development of an automated system for assessing ground
water potentials as part of a totally integrated system that would incorpo-
rate (1) existing water resources-related information, (2) remote imagery
analysis and interpretation capabilities, and (3) geophysical expertise.
A20
DRLLY~
WELL UNITSCA
DEEMIE GUPORSCLSRVYTA4
SELECTIOO WLLLSIITS
POSSOSILEE
WISTRVEYAREAROF INTEREST
MOBI=IZAWELL DEVELOPMENTT O
~ ~
Figure~ ~ ~ ~ ~~~~~E
tlzaino
F~darmilsrtn epyiaME3. uvyta
forseLec
L ING
of IT sitsRreulin inrdue GEellSCA
sJoryhls
A 2- 1IC ,P
E IT
References
Butler, Dwain K. and Llopis, Jose' L. 1984. "Assessment of Two Currently

Fieldable Geophysical Methods for Military Ground Water Detection,"
Military Hydrology, Report 6. Miscellaneous Paper EL-79-6, U. S. Army
Engineer Waterways Experiment Station, Vicksburg, Mississippi.
Butler, Dwain K. 1984. "Assessment and Field Examples of Continuous

Wave Electromagnetic Surveying for Ground Water, Military Hydrology,
Report 8. Miscellaneous Paper EL-79-6, U. S. Army Engineer Waterways
Experiment Station, Vicksburg, Mississippi (In Preparation).
Fitterman, David V. 1984. "Interpretation of Transient Electromagnetic

Soundings from White Sands Missile Range, New Mexico." Administrative
Report, U. S. Geological Survey, Prepared for U. S. Army Engineer Waterways
Kaufman, Alexander, A. and Keller, George V. 1983. "Frequency and

Transient Soundings." Elsevier Science Publishing Company, New York.
McNeill, J. D. 1980a. "EM 34-3 Survey Interpretation Techniques."

Technical Note TN-8, Geonics Limited, Ontario, Canada.
McNeill, J. D. 1980b. "Electromagnetic Terrain Conductivity Measurements

at Low Induction Numbers." Technical Note TN-6, Geonics Limited, Ontario,
Canada.
McNeill, J. D. 1980c. "EM 37 Ground Transient Electromagnetic System:

Calculated Depth of Exploration." Technical Note TN-10, Geonics Limited,
Ontario, Canada.
Zohdy, A. A. R., Jackson, D. B., Mattick, R. E., and Peterson, D. L.

1969. "Geophysical Surveys for Ground Water at White Sands Missile
Range, New Mexico," Open File Report 69-326, U. S. Geological Survey,
Denver, Colorado.
Zohdy, A. A. R., Eaton, G. P., and Mabey, D. R. 1974. "Application

of Surface Geophysics to Ground Water Investigations," Techniques of
Water Resources Investigations, Book 2, Chapter DI, U. S. Geological
Survey, Denver, Colorado.
A22
Waterways Experiment Station Cataloging-in-Publication Data
Simms, Janet E.
Assessment and selection of an automated electrical resistivity interpre-
tation procedure / by Janet E. Simms, Dwain K. Butler ; prepared for
U.S. Army Corps of Engineers.
83 p. : ill. ; 28 cm. - (Technical report ; GL-92-12)
Includes bibliographic references.
1. Water, Underground. 2. Geophysics. 3. Hydrogeology. I. Butler,
Dwain K. I1.United States. Army. Corps of Engineers. Ill. U.S. Army
Engineer Waterways Experiment Station. IV. Title. V. Series: Tech-
nical report (U.S. Army Engineer Waterways Experiment Station);
GL-92-12.
TA7 W34 no.GL-92-12

Llli1 Ill Lilt Il: Assessment and Selection of An Interpretation Procedure

Uploaded by

Copyright:

Available Formats

Llli1 Ill Lilt Il: Assessment and Selection of An Interpretation Procedure

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Llli1 Ill Lilt Il: Assessment and Selection of An Interpretation Procedure

Uploaded by

Copyright:

Available Formats

AD-A255 749

ASSESSMENT AND SELECTION OF AN

Electrode Spacing (ft) L'-0_ '$

Approved For Public Release; Distribution Is Unlimited

Prepared for DEPARTMENT OF THE ARMY

The contents of this report are not to be used for

7. PERFORMING ORGANIZATION NAME(S) AND ADDRESS(ES) B. PERFORMING ORGANIZATION

11. SUPPLEMENrARY NOTES

12a. DISTRIBUTION /AVAILABILITY STATEMENT 12b. DISTRIBUTION CODE

Approved for public release; distribution is

A.vailc"A; I:ty o"_-

DTIC QUALry fNSPECTD 3

PART I: INTRODUCTION .. ............................................ 7

PART II: CONCEPT DESIGN CONSIDERATIONS FOR AUTOMATED

Seismic Data Acquisition Capability ........................ 13

PART III: DESIRABLE ATTRIBUTES OF RESISTIVITY

Inversion Algorithm . ........................................ 19

PART IV: SURVEY OF AVAILABLE RESISTIVITY INTERPRETATION

PART V: EVALUATION OF RESISTIVITY INTERPRETATION PROGRAMS ....... 26

Theoretical Examples . ....................................... 27

PART VI: SUMMARY AND CONCLUSIONS ................................. 55

APPENDIX A: MILITARY REQUIREMENTS FOR GEOPHYSICAL GROUND WATER

Non-SI vnits of measurement used in this report can be converted to SI

Detection versus Exploration

2. Geophysical methods are routinely used throughout the world in

relief/nation assistance applications.

5. Initially, the detection of ground-water sources in a military

(ID 3rs < 0

--estimated drilling time

Seismic Data Acquisition Capability

9. The seismic refraction data will be acquired using conventional

V c '0C ' 5 La)P

Resistivity Data Acquisition Capability

10. The conventional method for collecting d.c. resistivity data is to

Figure 5. Resistivity data acquisition system. (a) conventional method

12. The following discussion addresses the features desirable in a

13. The inversion algorithm is the core of any resistivity

Direct Data Input

18. Resistivity data is inherently nonunique, therefore an inversion

20. A user-friendly inversion program is a necessity since it will be

Incorporation Into Overall System

22. Five resistivity interpretation programs are compared. Two of the

23. DCRESI is no longer commercially available but has been replaced by

Feature DCRESI RESINV SVES RESIXPLUS

user-friendly yes yes

graphic output yes yes yes yes

direct data input yes

various yes yes yes

automatic yes yes

equivalence yes yes

unix platform yes yes

DCRESI and RESIX PLUS: INTERPEX, Ltd., Golden, Colorado.

26. The program RESINV is available through government publication

27. Resistivity inversion program ATO is also available through a

28. Resistivity interpretation involves determining the number of

(wj) a.1 dea

.,., mnc oon

Figure 16. Example 6,Schlumberger field data collected at

(11-wqo) AI!A!IS!Se8 4u9,ioddv