99/
DELINEATION OF WELLHEAD PROTECTION AREAS
IN FRACTURED ROCKS
Wisconsin Geological and Natural History Survey
Ground-Water Protection Division
Office of Ground Water and Drinking Water
U.S. Environmental Protection Agency
Washington, DC 204*0
1991
Z Printed on Recycled Paper
-------
-------
DISCLAIMER
This report is the result of an investigation .supported by the U.S. Environmental Protection
Agency's Office of Ground-Water Protection, as part of its efforts to provide technical assistance
to state, tribal, and local governments on the implementation of the Wellhead Protection Program.
The specific methods and approaches contained in this document have been peer reviewed but
do not constitute official Agency endorsement or policy recommendations. The Office of
Ground-Water Protection provides mis information to help solve complex technical problems
related to the delineation of wellhead protection areas in fractured-aquifer settings; Further
assistance is available from the Office of Ground-Water Protection in Washington, D.C. and from
the ground-water offices in the ten EPA regions.
a
-------
ACKNOWLEDGEMENTS
by
is document was authored, under Cooperative Agreement # OC^lM- y
^^M^^
adSn, Wisconsin for the U. S. Environmental ^otectionAgency
»dS^
Marilyn Ginsberg of GWPD served fis Project Manager.
-------
DELINEATION OF WELLHEAD PROTECTION AREAS IN FRACTURED ROCKS
ABSTRACT
In 1987 the U.S. Environmental Protection Agency (EPA) published guidelines for
delineation of wellhead protection areas (WHPA j to meet the requirements of the 1986
Amendments of the Safe prinking Water Act of 1974. In the document, EPA concentrated on
WHPA delineation in common types of aquifers-granular, porous aquifers under unconfined
conditions. In 1989 the Wisconsin Geological and Natural History Survey prepared this report,
under an agreement with EPA to evaluate methods for WHPA delineation in unconfined
fractured-rock aquifers.
\ ... * ,'
Two fractured-rock settings were selected for the study: Precambrian crystalline rocks in
central Wisconsin and Silurian dolomite in northeastern Wisconsin. In both situations, densely
fractured rocks behaved as uniform, porous media at the scale of field tests. Potential methods
for WHPA delineation were tested using a full range of hydrogeologic investigations, including
water-table mapping, collection and analysis of water samples, geophysical logging, monitoring,
and numerical modeling.
The methods tested ranged from simple "cookie-cutter" approaches to complex computer
models, including the fixed radius methods, flow-system and vulnerability mapping methods,
residence-time methods, and numerical flow/transport models. The evaluation of these methods
indicated that flow-system mapping (combined either with time of travel criterion or calculations
using the uniform flow equation) and numerical modeling are the two most viable approaches
to wellhead protection in fractured rocks that act as porous media at the WHPA scale.
Flow-system mapping is especially useful when a ground-water divide is relatively close to the
protected well, and it offers reasonable accuracy at the least cost. Numerical modeling (the most
expensive method) provides increased precision and a better three-dimensional picture, which
may justify thejncreased costs, especially in complicated settings.
Four WHPA delineation approaches are suggested for unconfined fractured-rock aquifers that
do not behave as porous media. Vulnerability mapping combined with the arbitrary fixed radius
method or the simplified variable shape method produces a WHPA that includes most of the
areas near the well that are susceptible to ground-water contamination. Hydrogeologic mapping
can be used to determine ground-water basin boundaries and in some'cases, the ground-water
basin may function as the ZOC for a given well. The geochemical approach may be used to
provide information on relative ground-water ages and source areas. Numerical ground-water
flow/transport models, used carefully, may be able to simulate flow in discrete fractures or
fracture zones. Even though'the models are based upon porous-media assumptions, careful
discretization can allow one to incorporate a few high-permeability fractures or fracture zones
into the model design.
in
-------
-------
CONTENTS
> : , ' '.' "'''.' '" ' ' ' ' , ' - '''."-'/. Page
DISCLAIMER U
' - .
ABSTRACT . m
CONTENTS , V
EXECUTIVE SUMMARY n
'
, ..
Chapter I. INTRODUCTION
Project Background _
Purpose and Scope
Description of Study Areas *
Junction City Study Area '''' 6
Sevastopol Study Area °
Previous Studies
Acknowledgments
Chapter H. HYDROGEOLOGY OF FRACTURED ROCKS JJ
, Basic Characteristics of Fractured Rocks ||
Definitions and Terminology n
Origin and Examples of Fractures **
General Characteristics of Fractured-Rock Aquifers »*
Hydrogeologic Characterization of Fractured-Rock Aquifers 13
- Discrete versus Continuum Approaches *^
Tlie Problem of Scale and the Porous-Media Assumption 15
Determining Whether Fractured-Rock Aquifers Behave as Porous Media 16
Evaluation of Fractured-Rock Aquifers for Wellhead Protection Studies ^
Characterization of Fracture Patterns and Locations 20
Determination of Hydraulic Head Distribution 20
Determination of Aquifer Characteristics 21
Chapter HI. WELLHEAD PROTECTION AREA DELINEATION 25
Zones Used for WHPA Delineation ^
Common Delineation Criteria ' TT
Distance ±L
Drawdown , ^_
Time of Travel (TOT) , *?
Flow-System Boundaries ^.
Assimilative Capacity .
Evaluation of Criteria
-------
WHPA Delineation Methods for Fractured Rocks
Vulnerability Mapping
Flow-System Mapping i
Flow-System Mapping with Time of Travel Calculations
Flow-System Mapping widi Uniform Flow Equation
Residence Time Approach |
Numerical Flow/Transport Models ,
WHPA Delineation Methods for Fractured Rocks that Do Not Behave as Porous
Media
WHPA Comparative Analysis
Cost Analysis
Assessment/Comparison^ and Selection of Methods
Conclusions
Chapter IV. IMPLEMENTATION POSSIBILITIES
Applicability and Usefulness of Selected Methods
Management Strategies for WHPAs
Transferability of Results (by WJ. McCabe, U.S. EPA)
Introduction
Acknowledgments
Mapping Criteria
Areas Where Techniques of This Document May Be Applicable
REFERENCES CITED
29
30
31
33
39
,44
50
52
55
55
56
66
67
67
63
71
71
71
71
75
79
APPENDIX A JUNCTION CITY SITE/PORTAGE COUNTY, WISCONSIN
SITE SELECTION
INVESTIGATION METHODS
HYDROGEOLOGIC SETTING
RESULTS OF INVESTIGATIONS
Water-Table Mapping
Water-Level Measurements
Vertical Distribution of Hydraulic Head
Water-Level Fluctuations
Aquifer Tests
Specific Capacity Data
Slug Tests
Pumping Test
Water Chemistry and Isotope Analyses
Water Chemistry
Isotopes
85
85
85
87
90
90
96
96
96
98
98
98
100
102
102
104
VI
-------
' "- . '. ; ' <*' ' " ' 105
NUMERICAL MODELING 105
Model Selection 105
. Conceptual Model
M^M^^ ;
Model Layer 2 - Bedrock With Open Fractures .-
Model Layer 3- Bedrock With Few Fractures ill
Model Calibration - ni
WHPA Simulations 111
Steady-State How Simulation . Ill
/ Transient Flow Simulation ,113
' Particle-Tracking Simulation
APPENDIX B - SEVASTOPOL SITE, DOOR COUNTY, WISCONSIN
' ' ', '-.' '.'-...'". ' - ' ; ' "' ' 119
119
iV AS'lAJi'^-'irf aii£«, i/v/viv v^v»«-n-«» », -i** ^- --- -
SITE SELECTION
INVESTIGATION METHODS -
'""'.''-' _ ' ' ...-:'.. .'.'.' 123
HYDROGEOLOGIC SETTING ;
RESULTS OF INVESTIGATIONS
Water-Table Mapping 126
Water-Level Measurements 126
Vertical Distribution of Hydraulic Head
Water-Level Fluctuations 129
Aquifer Tests 131
Specific Capacity Data 131
Slug Tests 131
Pumping' Tests'
Water Chemistry and Isotope Analyses
Water Chemistry .
Isotopes ,
' " ' ''-'.."-'. ' . .' . 155-
NUMERICAL MODELING 135
Justification and Code Selection 136
Conceptual Model 137
Model Grid Design 137
Model Calibration . 140
WHPA Simulations 140
Steady-State Flow Simulation 140
Particle-Tracking Simulation
vn
-------
FIGURES
1. Junction City study area, Portage County, Wisconsin
2. Generalized bedrock geology of the Junction City area, Wisconsin
3. Sevastopol study area, Door County, Wisconsin !
4. Generalized geologic cross section of the Door Peninsula, Wisconsin
5. Effect of fractures on ground-water movement
6. Pumping test responses in settings where the porous-media assumption does
and does not apply
7. Examples of water chemistry and temperature variations with time in
diffuse-flowr.and conduit-flow aquifers------ : v
8. Terminology for WHPA delineation in fractured rocks
9. WHPAs based on vulnerability mapping at the Sevastopol site
10. ZOC delineation in crystalline rocks using a water-table map
11. ZOC delineation in a shallow ground-water system in dolomite using a water-
table map
12. ZOC delineation in a deep ground-water system in dolomite using
a potentiometric-surface map
13. ZOC delineation in dolomite using water-table and potentiometric-suirface maps
14. ZOC delineation using the uniform flow equation
15. ZOC delineation in crystalline rocks using the uniform flow equation
16. ZOC delineation in a deep ground-water system in dolomite using the uniform
flow equation
17. ZOC verification using tritium data at Junction City, Wisconsin
18. ZOC predicted by numerical modeling for a well in crystalline rocks
19. ZOC predicted by numerical modeling for a well in dolomite
20. ZOC comparative analysis for a weU in crystalline rocks
21. ZOC comparative analysis for a well in dolomite
22. Areas of unconfined fractured-rock aquifers in the contiguous United States
23. Areas of unconfined fractured-rock aquifers in Alaska, Hawaii, Puerto Rico,
U.S. Virgin Islands, and Guam
24. Major physiographic regions of the continental United States
4
5
7
8
14
17
19
26
32
35
37
38
40
41
43
45
48
53
54
61
62
72
73
76
Al. Location of wells in a portion of the Junction City study area, Portage
County, Wisconsin
A2. Location of wells and piezometers in the well field at Junction City
A3. Depth to bedrock in the Junction City area ;
A4. Geologic log for core hole JC23 at Junction City
A5. Correlation of natural gamma logs for wells in the Junction Ci'y area
A6. Temperature logs for the village well (JC9) and the nearby, deep cone
hole (JC23)
A7. Geologic, cross section for the Junction City area
A8. Portion of the water-table map of the Junction City area
86
88
91
92
93
94
95
97
Vlll
-------
A9. Computer-generated map of log hydraulic conductivity values for the crystalline ^
rocks at Junction City 101
AID Drawdown for the Junction City pumping test
ill. SSvity distribution in the Junction City wellfield based on the pumping ^
test data . . tn^
A12 Hydrbgeologic boundaries for the Junction City site ... '
All Are* view of grid and spatial discretization used in the numencal simulation of
the Junction City ground-water flow system
A14. Configuration of model layers used in numerical model of ground-water flow at ^
-- the Junction City site n?
A15 Simulated steady-state water-table elevations in me Junction Uty area n-t
1 16 TrTvef Shs aM TOTs for particles reaching the Junction City village well at a
depth of 53 ft below the land surface or U06A above msK ^ , . . ."=>
A17 Travel paths and TOT s for particles reaching the Junction City village well at a
depth of 123 ft below the land surface or 1036 ft above msl »">
A18. Profile view of travel paths and TOTs for hypothetical particles reaching the ^
A19 verpaATOTfor particles reaching the Junction City. vUlage well at a
depth of 162 ft below the land surface or 997 ft above msl
BL Research sites in Door County, Wisconsin ','-.-.,
B2. Expression of bedrock fractures in alfalfi field at the Sevastopol site, Door ^
County, Wisconsin
B3 Geophysical logs for well MW1 at the Sevastopol site
B4. Distribution of hydraulic head in the shaUow and deep ground-water systems ^
at the Sevastopol site, August 1989 . _
B5. Profile of hydraulic head at the Sevastopol site in March 1V8V |^>
B6 Hydrographs of water levels at the Sevastopol site - :'
B7. Finite-difference grid used in numerical model of ground-water flow at the
Sevastopol site . '"'.". \
B8. Configuration of model layers used in numerical model of ground-water, ^
flow at the Sevastopol site 141
B9 Simulated hydraulic head distribution at the Sevastopol site .
BIO Simulated particle paths using the PATH3D code, Sevastopol site »«
Bill Cross section of the Sevastopol site showing vertical ground-water movement
from the surface to various depths along a well casing
IX
-------
TABLES
'-..'" ' : " Page
1. Differences between porous-media and fractured-media aquifers 12
2. Range of values of porosity 22
3. Qualitative interpretation of tritium concentrations in ground-water 46
4. Data requirements for WHPA delineation methods for fractured crystalline and
dolomitic rocks * 56
5. Estimated work, time, skill, and cost requirements for selected WHPA
delineation methods 57
6. Potential management tools for wellhead protection 70
7. Common depth ranges of. unconfinedsfractured-rocka^uifere in the United
States and its possessions 74
;
Al. Well and piezometer data, Junction City study area, Portage County,
Wisconsin ' 89
A2. Results of slug tests at the Junction City site | 98
A3. Results of single-well tests on core hole piezometers at Junction City 100
A4. Chemical analyses of the Junction City samples , 103
A5. Isotope results for the Junction City area 104
A6. Layer characteristics for the Junction City model "107
Bl. Well and piezometer data, Sevastopol study area, Door County, Wisconsin 122
B2. Results of slug tests at the Sevastopol test site and barnyard research site 132
B3. Results of pumping tests at the Sevastopol site H2
B4. Chemical analyses of Door County samples ' 133
B5. Isotope results for Door County . 134
B6. Layer characteristics for the Door County model i 137
-------
EXECUTIVE SUMMARY
Background
the 1986 Amendments to the Safe Drinking Water Act (SDWA) of 1974 established^
nationwide program to prevent contamination of ground-water sources tappedby public water
supply^elirth^ellh^i Protection Program. The U.S. Environmental Protect Agency
CEP A VwS required to provide technical guidance to help state, tribal, and local agencies
SnptemTthT&ologic aspects of the program. In 1987 EPA P^«hed "«
Delineation of Wellhead Protection Areas" to meet this requirement e S
rsum^arV of criteria and methods for delineating wellhead protect areas
document concentrated on WHPA delineation in one common type of aquifer--
£a^aqXexisting under unconfined conditions. Given the lack of attend
at that point to delineating specific WHPAs in confined aquifers and less common fractured-rock
aquifers, the guidance provided limited information on such settings.
This report "Delineation of Wellhead Protection Areas in Fractured Rocks," was prepared
by the Wisconsin Geological and Natural History Survey (WGNHS), under an agreement with
the EPA Office of Ground-Water Protection, to evaluate approaches to delineating WHPAs m
unconfined fractured-rock aquifers that would be applicable to fractured-rOck settings in
Wisconsin and in similar settings elsewhere in the United States.
Hydrogeologic Setting
Two fractured rock settings were selected for the study: Precambrian crystalline rocks in
central Wisconsin and Silurian dolomite in northeastern Wisconsin. The first test site is typicid
of terrains where crystalline rocks are at or near the land surface and covered only by residual
soil or a thin layer of unlithified deposits,.and where fractured crystalline rocks at; the sole
source of community water supplies. The second test site is typical of fractured sedim"^
carbonate-rock aquifers either exposed at the land surface or^covered by thin soils^ where
potential for ground-water contamination is very high. Due to abundant ^*>^*c?,
rocks at both sites behaved as porous media at the scale of field tests. The P°*-med*
assumption implies'that the hydraulic properties^ the individual fractures are not important and
that the aquifer can be treated as a continuum when the problem scale is large enough.
Approaches Used ,
Methods to delineate WHPAs weretested using a-fid! range of hydrogeologic investigations.
including water-table mapping, collection and analysis of watersimples, geophysical logging,
aquifer testing, and numerical modeling. Four approaches to WHPA delineation were tested at
XI
-------
the two test sites. The overall goal of each approach was to delineate the zone of influence
(ZOI) or the zone of contribution (ZOC) of a well in an unconfined fractured-rock aquifer.
Four main approaches were used for testing WHPA delineation methods. The first approach
included calculating the ZOI of a well by standard well and ground-water hydraulics equations.
These equations are based on the assumption that a large body of fractured rock with closely
spaced fractures acts hydraulically similar to a porous, granular medium. The second approach
involved mapping hydrologic and geologic boundaries, as well as identifying areas of the
landscape particularly vulnerable to ground-water contamination. The third approach utilized
geochemical and isotopic indicators to estimate the age and source of water produced by a well.
The isotopic results can be used to verify the validity of the ZOC estimates based on hydraulic
considerations alone. The fourth, and most sophisticated approach, was the construction and
calibration of a detailed, three-dimensional numerical ground-water flow/crafnsport model of each
study area.
WHPA Delineation Criteria
WHPA delineation is based upon the analysis of criteria, thresholds, and delineation
methods. The criteria and thresholds incorporate the general technical basis of the WHPA. The
WHPA delineation methods are techniques by which criteria and thresholds are translated into
on-the>ground or on-the-map WHPA boundaries.
Criteria used for the WHPA delineation at the two test sites in Wisconsin were adopted from
the 1987 EPA guidelines and included
1) distance from the well,
2) drawdown around the well, - -
3) time of travel (TOT) to the well, and
4) ground-water flow system boundaries.
Some combinations of these criteria will work better than others in establishing a wellhead
protection program in fractured rocks. For example, the distance criterion would not work well
by itself because it does not account for any site-specific hydrogeologic characteristics. The
drawdown criterion leads to the delineation of a ZOI only, which is not a good basis for a
WHPA in fractured rocks for two reasons. First, the dimensions of the drawdown cone may be
too small for the porous-media assumption to be valid. Second, unless thes water table is nearly
horizontal, the zone of influence will seriously underestimate the true ZQC. A more protective
WHPA is probably based on the flow-system boundary criterion. Delineation of ground-water
divides .and flow lines is done at a large enough scale that the assumption that fractured rock
behaves as a uniform porous medium is likely to be valid. Even if the porous-media assumption
is not completely valid, flow-system mapping may delineate zones of conduit fracture flow,
which can often be simulated numerically. :
xii
-------
WHPA Delineation Methods
Following selection of WHPA delineation criteria, six methods were tested for WHPA
delineation at the two test sites in Wisconsin. The methods, in older of increasing complexity
include
1) arbitrary fixed radius,
2) calculated fixed radius,
3) vulnerability mapping,
4) flow-system mapping, .
- with TOT calculations,
- with analytical equations,
5) residence-time approach, and
6) numerical flow/transport modeling.
The first two methods are not recommended for WHPA delineation in fractured rocks. The
arbitrary fixed radius method does not incorporate any hydrogeologic considerations, and the
estimates of the radius are more difficult to make in complex fractured-rock conditions than in
porous media. The calculated fixed radius method resulted in unrealistic radii around die
protected well, probably because the errors associated with aquifer parameter estimates and with
estimates of the time required for the aquifer to reach steady state are larger in fractured rocks
than in porous media.
Vulnerability mapping can be applied to any type of hydrogeologic setting. This method
does not delineate a ZOC for a well; rather, it identifies areas particularly vulnerable to
ground-water contamination. A WHPA that includes most of the areas near the well that are
susceptible to contamination can be delineated by combining vulnerability mapping with the
arbitrary fixed radius method or the simplified variable shapes method. ,
The flow-system mapping method utilizes flow-system boundaries, and it is a very efficient
method for delineating ZOCs in fractured-rock settings, especially where the flow-system
boundaries are close to the well. A water-table map can be compiled with a minimum of field
work if enough water-level data are available from ag. ncies' water-well files and if water-level
fluctuations in the area arc relatively small. In general, flow-system mapping requires the
assumption that the fractured rock behaves as a porous medium at'the scale of the problem.
However, in cases where the aquifer does not behave as a porous medium, the flow-system
mapping technique may be useful for delineating fracture conduits or zones of high hydraulic
conductivity. The WHPA for the well should then take such zones into account
Combining the flow-system mapping method with calculations of the time of travel or with
calculations using the uniform flow equation may lead to a more accurate ZOC for relatively
little additional cost These combination methods offer reasonably accurate and efficient
protection for wells in many fractured-rock settings, but they may not be adequate if the fractured
rock does not act as a porous medium.
Xl.ll
-------
The residence-time approach utilizes water chemistry and isotopes to identify ground-water
flow paths and to provide information on relative ground-water ages and source areas. This
approach does not assume that the fractured rock behaves as a uniform porous medium. The
residence-time approach does not delineate a ZOC and is therefore only useful when used in
combination with other WHPA delineation methods, especially with the flow-system mapping
and numerical models. The method can identify those tiydrogeologic settings where recent
recharge indicates vulnerability to potential contamination.
Numerical models use mathematical approximations of ground-water flow and/or
contaminant transport equations that can take into account a variety of hydrogeologic and
contamination conditions and do not necessarily require that aquifers behave as-porous-media.
The flexibility of computer models allows for handling features such as conductive fracture zones
in some relatively simple settings. These models possibly offer the most accurate ZOC
delineation, but at considerable cost, which may be justified in complex settings or when great
accuracy is required.
The study of two fractured-rock settings in Wisconsin demonstrated that standard
ground-water flow equations and field techniques developed for porous media can be used for
the delineation of WHPAs protection areas in fractured rocks. Even so, both sites were complex
enough that methods less sophisticated than flow-system mapping did not encompass enough of
the site characteristics to result in an adequate WHPA. The aquifers at both sites were
heterogeneous and anisotropic, and numerical modeling was the only method able to take the
heterogeneity and anisotropy into account Both sites were located near ground-water divides,
which limited the size of the ZOCs for each site. Due to the presence of these ground-water
divides, the numerical models produced ZOCs that were not much different from ZOCs produced'
by flow-system mapping. However, at both sites the TOT estimates produced by modeling were
more accurate than TOT estimates from other methods. At sites far from hydrogeologic
boundaries, numerical modeling will probably be the only method that can delineate an accurate
WHPA of reasonable size.
For a fractured-rock aquifer that does not act as a porous medium, WHPA, delineation am
be accomplished by a combination of vulnerability mapping, hydrogeologic mapping, tide
residence-time approach, tracer tests, and numerical modeling. Row in such aquifers can occur
mainly through discrete fracture conduits. It is particularly important to identify possible areas
where such conduits intersect either the land surface or the well to be protected because such
conduits can offer direct and rapid pathways for contamination to travel from the land surface
to the well. ,
WHPA Implementation
The WHPA delineation methods tested in Wisconsin can be applied to similar settings
elsewhere in the United States. Unconfined crystalline-rock and carbonate-nock aquifers acre
common throughout much of the eastern United States and western mountain ranges and in many
xiv
-------
«io^c in the midcontinental section. Even though similar aquifers are quite common, there are
assumption is not completely justified.
XV
-------
-------
Chapter I
INTRODUCTION
Project Background
Th. 1986 Amend^nu «o *. Safe D**ng Water AaCSDWA) of K*4 estaWishefl a
^^
actual resource-ground water.
The Wellhead Protection Program seeks to accomplish this goal by the establishment of
wellhead protection areas (WHPAT), within which potential contamination sources are
^ffemntiaUv^ia^ed A WHPA is defined in the 1986 Amendments as "the surface or
X" SJSL& te-d approach. TTms, WHPAs «e «ea«d as a mean, to reduce
the potential for contamination of public water supplies.
The Wellhead Protection Program requires the participation of all levels of government.
Individual states must develop and implement wellhead protection ^grams^Kr mee^the
requirements of the SDWA Amendments. The federal government is responsible for ^
a^v*g state wellhead protection programs and for P^f^^^^^
-local governments. Because of this, the U.S. Environmental Protection Agenpy^PA) was
required to provide technical guidance to help state and locd ag encies ^P^*e
hydrogeologie aspects of the weUhead protection program. In 198 7 EP ^A puW is^ hed
Sdllinesfor Delineation of Wellhead Protection Areas' as ?f ^^Xeatin^
EPA 1987) The guidelines provided a summary of criteria and methods for delineating^
WlJpAs The ^tocSnent, however, concentrated on WHPA delineation in one common type
er-^ ^us, g^ula, aquifer existing un^
limked attentiorto delineating WHPAs in confined aquifers and less common fractured-rock
aquifers.
Staff of the Wisconsin Geological and Natural History Survey
report under Federal Assistance Agreement No. CX-81538Ml^h ?teU.S^ Ogte <
G\6und-Water Protection, Washington, D.C., with Manlyn Ginsberg as E?A Project ^Officer
and Alexander Zaporozec as Project Manager. Pans of the report were contributed by Bruce
-------
Brown and Ronald Hennings of WGNHS, and by William McCabe of EPA. The project
period was October 1, 1988 - December 31, 1989.
Purpose and Scope
.1 , ,
The purpose of the project was to investigate methods for delineating WHPAs that would
be applicable to fracturcd-rock settings in Wisconsin and to similar settings elsewhere in the
United States. This document was prepared to assist planners, managers, and administrators
of state, county, or municipal governments in the task of protecting their water supplies
against contamination.
The main objectives of the prqjttCt were
1) to study the patterns of ground-water flow in fractured crystalline rock in
central Wisconsin and in fractured dolomite in northeastern Wisconsin;
2) to develop methods for estimating.hydraulic properties and time of travel (TOT) in
these two rock types;
3) to identify, on the basis of the results of the studies, criteria and the range of
methods that can be used for delineating WHPA in fractured-rock aquifers;
4) to examine implementation limitations that may influence the applicability and
usefulness of the recommended methods;
5) to evaluate the applicability and transferability of the recommended methods to
similar hydrogeoldgic settings elsewhere in the United States.
The scope of the project included investigation of several methods thought by the
investigators to be applicable to solving the technical problems of delineating WHPAs in
fractured rocks. These methods range from simple "cookie-cutter" approaches to complex
computer models; they include the fixed radius methods, flow-system and. vulnerability
mapping methods, analytical methods, residence-time methods, and numerical flow/transport
modeling.
The investigation of the study areas was carried out in two steps. Tilie first step involved
the collection of background data about the geology and hydrogeology of each site. The
second step included testing WHPA methods using different approaches. The background
investigations led to the assumption that the fractured rock at both sites contained enough
closely spaced fractures to function hydraulically similar to a porous, granular medium. Our
field investigations confirmed this assumption.
Four main approaches were used for testing WHPA delineation methods. The first.
approach included calculating the zone of influence (ZOI) of a well by stuidard well and
ground-water hydraulics equations. The second approach involved mapping hydrologic and
geologic boundaries, as well as identifying areas of the landscape particularly vulnerable to
ground-water contamination. The third approach utilized geochemical and isotopic indicators
to estimate the age and source of water produced by a well. This approach can be used to
-------
verify the validity of the zone of contribution (ZOC) estimates made on the basis of hydraulic
conslo^tiotalle Tne fourth, most sophisticated approach, was *e cons^cuon^
calibration of a detailed, three-dimensional numerical ground-water flow/transport model of
each study area.
The study also included the evaluation of criteria ^ t
and a comparison of each method in terms of advantages, disadvantages, skill ^u
and costs. The last chapter of the report analyzes limitations of methods "J*« .
applicability to other paVts of the United States and suggests potential tools for managing
WHPAs. The implementation possibilities were discussed at meetings with local elected
officials in the study areas.
Description of Study Areas
Two different fractured-rock settings were selected ^^f0^ ^SSrtck and
Wisconsin (at the VUlage of Junction City) in a complex of Precambnan crystalhne rock and
oneTmt ^Peninsula of northeastern Wisconsin (in the Town of Sevastopol), where the
aquifer is composed of fractured Silurian dolomite.
Junction City Study Area
on
This site was selected to represent areas where small communities e
produced by one or more wells in fractured crystalline rocks. The site is located "the
weVtem edge of the Village of Junction City (fig. 1), a small coirnnumtyj^pulanon 525)12
mS^orthlest of Steven's Point, Portage County. The vUlage is -supplied -* w^from a
drilled weU, 321 ft deep, developed in Precambrian metavolcamc rock. Average daily
pumpage is about 40,000 gallons.
Junction City is located at the southern, exposed edge of the Precambrian shield-a
geologically stable region, consisting of metamorphic and igneous rocks more _than 600
million yeis old (fig 2). The crystalline rocks are covered bya ^"*"
-------
Village well (JC-9)
3 Approximate study area
She location
in Wisconsin
Figure 1. Junction City study area, Portage County, Wisconsin.
-------
KILOMETRES
MJLES 1
SCALE 1:100 000
n 1 2
^ Village well
[ ] Cambrian sandstone
Precambrian rocks: _
Metasedimentary rocks
Granitic rocks
Mafic intrusive rocks
Felsic to intermediate metavolcanic rocks
Mafic metavolcanic rocks
Figure 2. GeneraHzed bedrock geology of the Junction City area, Wisconsin (adapted from
Greenberg and Brown, 1986).
Area of near-surface
crystalline rocks in
Wisconsin
-------
decrease in size and number with depth. Recent investigations indicate that the surficial
materials are hydraulically connected to the bedrock system (Hutasoit and Davidson, 1989).
Junction City is located in the part of Wisconsin that was glaciated prior to the late
Wisconsin glacial period. Surficial materials in the area consist of highly weathered, older
glacial deposits mixed with younger clayey, silty sand derived from weathering of the bedrock ft
and reworked by hillslope processes. Occasional erratic boulders mixed with the bedrock
residuum are the only evidence of pre-Wisconsin glaciation (Clayton, 1986). The" surficial
materials are thin and discontinuous and are generally inadequate for domestic and
community water supplies (Zaporozec and Cotter, 1985).
The general area of the test site is a relatively smooth upland plain with gently sloping
topography^ The land surface,:ranges between 4=140-and 1180 ft above mean sea level. The
area receives about 31.5 in. of precipitation annually and is drained by sunall intermittent
tributaries of Mill Creek (fig. 1), a tributary of the Wisconsin River.
The village well is situated 100 ft from U.S. Highway 10, a majorvtfrioroughfare, and 250
ft from a railroad, which pose potential contamination threats to the well,. The well was used
as a pumping well for tests conducted during the study. Ten piezometers were installed at
seven sites upgradient and one 6-in. observation well was installed downgradient of the
village well (for locations, see Appendix A).
Sevastopol Study Area
The study focus in this area was on wellhead protection in rural settings underlain by
fractured dolomite either exposed at the land surface or covered by thin sjoils, where potential
for ground-water contamination is great. The test site is located about 9 miles north of
Sturgeon Bay, Door County (fig, 3). The site contains a cluster of 17 piezometers installed in
seven wells within a 100-ft radius of each other on farm land near a small patch of woods.
Seven additional piezometers are located at a barnyard about 1 mile northeast of the test site
(for locations, see Appendix B).
The Door Peninsula is a narrow (7 to 20 miles wide) peninsula that rises more than 200
ft above Green Bay and Lake Michigan (fig. 4). Bedrock is dolomite of Silurian age
(approximately 400 million years old), a thick (400 to 600 ft), resistant formation, which
forms the prominent Niagara escarpment in eastern Wisconsin (Shenill, 1975). The
west-facing escarpment descends abruptly to Green Bay. The site lies on the back slope of
the escarpment (fig. 4), which dips southeastward under Lake Michigan into the Michigan
basin. The dolomite is a major aquifer in eastern Wisconsin and extends from the northern
tip of Door County south into Illinois. Well yields range from 10 to 500 gpm (Zaporozec and
Cotter, 1985). Pleistocene deposits are thin or absent in tl.l. area, and do not form a
productive aquifer. The Silurian dolomite is exposed at the land surface or covered by a thin
layer of soil. It is densely fractured; horizontal fractures occur on the order of every 1 to 10
-------
SCALE 1:100 000
KILOMETRES 1 _0 ',' 2 3 *
MILES 1 , _" ',,9 - ' 1 ' ' ' 2
* Sevastopol test site
Barnyard research site
D Approximate study area
Site location
in Wisconsin
Figure 3. Sevastopol, study area, Door County. Wisconsin.
-------
60 000 tt
5000 0 5000 10000ft
gHI Unlrthified material (Pleistocene)
[gHHl Dolomite (Silurian)
Shale (Ordovician)
I Well used to construct section
Figure 4. Generalized geologic cross section of the Door Peninsula, Wisconsin.
-------
ft and vertical fractures form a more or less regular pattern (fig. B2, Appendix B) Nearly
visible in cropped fields and pastures during the growing season. A detailed description of
the geological setting is included in Appendix B.
The test site is located on a local topographic divide between Etonians and Lilly Bay
Creeks, which drain southeast to Lake Michigan. Topography is of very moderate ^ ^
slopes very gently in all directions from the site. The land surface is between 790 and 810 ft
above mean sea level (msl). Average annual precipitation is 30 in. Land use at the sue,
typical of central Door County, consists of pasture, com and hay crops, and maple syrup
production.
Previous Studies, ...
The concept of wellhead protection zones is relatively new in the United States.
Attempts at establishing WHPAs in the U.S. have been scattered and not as advanced as m
Europe especially in West Germany, where the concept originated (Headworth, 1986).
Generally, WHPAs have been delineated only for porous, granular media; the authors
literature search did not locate any publications on delineating WHPAs in fractured rocks.
Acknowledgments
Appreciation is given to land owners in the two study areas who allowed access to their
land to measure water levels in wells and to take water samples. Special appreciation is
given to the Village of Junction City Board for its support of the study and permission to use
the village well for pumping tests. David Haupt, of Haupt Well and Pump Co., Inc.,
registered well driller from Aubumdale, provided information on the test holes for the Village
of Junction City well and other wells in the area.
) i .>'".'. ' '
John Leatherman, Community Resources Agent, Portage County; Dennis Skahen,
Agriculture and Resources Agent, Door County; and William Schuster, Door County
Conservationist, were instrumental in arranging public meetings in Junction City and Door
County, respectively. Without their help it would have been difficult to obtain local input on
method selection and limitations.
The editorial, cartographic, and administrative staff of the Wisconsin Geological and
Natural History Survey have all contributed to the successful completion of this project, their
patience and assistance are greatly appreciated.
-------
-------
Chapter n
HYDROQEOLOGY OF FRACTURED ROCKS
Basic Characteristics of Fractured Rocks
Definitions and Terminology
Fractures can occur in almost all geologic materials, from unlithified surficial materials
to deeply buried rocks. The term "fracture" can have many definitions, fo this report, a
fracture is defined as follows: .
Fracture: Ageneral term for any break in a rock, whether or not it causes displacement,
due to mechanical failure by stress. Fractures include cracks, joints, and faults (Bates
and Jackson, 1980).
A crack is a partial or incomplete fracture; a joint is a parting in a rock along which no
movement occurred; a fault is, a fracture or a fracture zone along which movement occurred.
Although the last two categories of fractures form in very different ways, their effect on
ground-water flow is similar.
Fracture zones consist of closely spaced and highly connected discrete fractures (Gale,
1982). Related terms often used in discussions of fractures include bedding
planes-discontinuities in the depositional form of sedimentary rocks; fracture
traces-small-scale (hundreds of feet) linear features visible on the land surface above buned
fractures; and lineaments-large-scale (miles) linear features related to major fractures
extending to great depths and visible on aerial photographs. No discussion of fractures is
complete without mentioning karst, a carbonate-rock terrain where fractures in soluble rocks
have been enlarged by chemical solution or physical erosion.
Origin and Examples of Fractures
Although the origin of fractures is not always clear, fracturing is usually attributed to one
or more of the following geologic processes:
- tectonic forces, geologic uplift, mechanical folding;
- magma movement;
- thermal processes, stress relief associated with heating or cooling;
- glacial or erosional loading or unloading;
- earth tides.
Rock types most susceptible to fracturing include
- crystalline plutonic "rocks (granite, gabbro, diorite, etc.);
11
-------
- other igneous and metavolcanic rocks (basalt, rhyolite);
- metamorphic rocks (schist, gneiss, quartzite); ' i
- carbonate sedimentary rocks (limestone, dolomite).
Such rocks generally are not ductile, and thus are subject to failure and fracture formation
under applied stress. However, these are not the only rock types in which fractures occur.
Some sandstones, especially when well cemented, are also highly fractured. Stephenson and
others (1988) discuss the formation and importance of fractures in unlithified clayey till units
in central North America. Fractures are also important in shales and other low-permeability
argillaceous rocks, but little research has been done on fractures in these materials (Gale,
1982).
General Characteristics of Fracttired-Rock Aquifers
Although fractures can be important to ground-water movement through almost any
geologic material, this report is concerned with the hydrogeologic properties of aquifers in
which water moves primarily through fractures. Fractured-rock aquifers differ from porous,
granular aquifers in several important ways, summarized in table 1.
Table 1. Differences between porous-media and fractured-mcdia aquifers.
Aquifer Porous Media Fractured Media
Characteristics
Porosity . Mostly Mostly
primary secondary
Flow Slow, laminar Possibly fast
and turbulent
Isotropy More Less ;
isotropic isotropic
i
Homogeneity More Less
homogeneous homogeneous
Flow Darcy's law Darcy's law
predictions applies may not apuply
12
-------
Porosity refers to the ratio between the volume of voids and volume of solids in a given
volume of rock mass, and it is related to the storage and transmission characteristics of an
aquifer. Type of porpsity is probably the single most important difference between
fractured-rock and granular aquifers. Primary porosity is the volume of original pores
between rock or mineral grains and is usually a function of the packing of the grains.
Secondary porosity refers to the pore volume related to spaces created within a rock after its
deposition or emplacement, through fracturing or diagenetic processes such as dolomitization
and dissolution of primary cements. Primary porosity is relatively predictable and easy to
measure; secondary porosity can be unpredictable and difficult to measure.
Anisotropy refers to the variability of hydraulic conductivity or transmissivity with
direction. Fractured-rock aquifers can be highly anisotropic inthree dimensions (Snow,
1969); hydraulic conductivity and ground-water velocity can vary by several orders of
magnitude depending on the direction of ground-water movement Figure 5A, shows the
effects of fracture anisotropy on regional ground-water movement. Ground water in
anisotropic media may not move perpendicularly to the hydraulic gradient, but at some angle
to'it,'. '' ' ' : ' ' . '.-'.' ..''
Fractured rocks tend to be less homogeneous (more heterogeneous) than porous, granular
rocks. Heterogeneity, the variation of hydrogeologic properties (hydraulic conductivity,
porosity) from place to place within the aquifer, is frequently large in fractured-rock settings
due to the discrete and linear nature of many fracture systems. Figure 5B shows how
near-vertical fracture zones can short-circuit a regional ground-water flow system in
crystalline-rock. -
The equations of flow in porous media, generally based on Dairy's law, usually require
that ground-water flow be laminar and of low velocity. In some fractured systems,
particularly in areas where fractures become enlarged by solution, ground-water flow may
become turbulent and fast, and the classic ground-water flow equations may no longer apply
(Freeze and Cherry, 1979, p. 73-74). Although turbulent flow can be modeled, the complex
equations and enormous amounts of data required make analysis of turbulent ground-water
flow impractical for most wellhead protection studies.
Hydrogeologic Characterization of Fractured-Rock Aquifers
Discrete versus Continuum Approaches
Hydrogeologic investigations in fractured-rock terrenes usually fall somewhere between
two very different approaches to fracture characterization. The discrete approach considers
each fracture individually. Careful measurement of such fracture properties as length,
orientation, width, aperture, wall roughness, and connectivity provide the parameters
necessary for detailed study of ground-water movement in each fracture. The discrete
13
-------
ISOTROPIC AQUIFER
ANISOTROPIC AQUIFER
Z0t\& Of
. contribution
water-table
contours
of
contribution
KV
A. Effect of fracture anisotropy on the orientation of the zone of contribution.
B. Numerical simulation of regional flow paths in fractured crystalline rock.
showing "short-circuiting" along vertical fracture features (from: Gale, 1982)
Figure 5. Effect of fractures on ground-water movemen*.
14
-------
approach canpotentially provide great detail on ground-water movement. However, the data
and computational requirements of this approach are so great as to make the approach
impractical for all but relatively small volumes of rock in the most simple field situations.
At the opposite end of the spectrum, the continuum approach assumes that the fractured
medium approximates a porous medium at some working scale. In this approach, the
properties of individual fractures are not as important as the properties of large regions or
volumes of fractured material. The continuum approach offers the advantage of requiring
manageable amounts of data, but cannot resolve as much detail about the flow system as is
possible with the discrete approach.
Most field-scale hydrogeologic investigations of fractured-rock aquifers, and therefore,
most wellhead protection studies in fractured rock, fall somewhere between these two
approaches. Although detailed examination of all fractures present in the field is usually
impossible, some details, such as the orientation and length of major fractures or fracture
zones, can be incorporated into analyses based on the continuum approach and porous-media
assumption.
The Problem of Scale and the Porous-Media Assumption
The use of standard ground-water flow equations, such as Darcy's law, in fracturcd-rock
settings requires the assumption that the fractured rocks behave as porous media at the scale
of the problem. Knowledge of the problem scale therefore becomes critical in field studies of
wellhead protection because "major" fractures at the scale of a few inches or feet can become
"minor" fractures in the area at the scale of thousands effect or miles. r'Major" fractures in
an otherwise uniform flow system can short-circuit the system, causing ground water to move
in unexpected directions (fig. 5B). The size of observable fractures usually depends on the
observation technique. For example, aerial photographs might identify large bedrock
fractures, observations in vertical boreholes might identify smaller fractures, and microscope
studies of laboratory samples might identify even smaller fractures.
The porous-media assumption implies that the classical equations of ground-water
movement hold at the problem scale, that knowledge of the hydraulic properties of individual
fractures is not important, and that the fractured-rock mass can be characterized by field and
laboratory techniques developed for porous media. Long and others (1982) provide
theoretical criteria for determining when fractured systems behave as porous media. They
suggest that "fracture systems behave more like porous media when (1) fracture density is
increased, (2) apertures arc constant rather than distributed, (3) orientations are distributed
rather than constant, and (4) larger sample sizes are tested" (Long and others, 1982, p. 645).
However, rigorous testing of these criteria for field situations is difficult and expensive. For
most wellhead protection studies investigators could use a combination of subjective criteria
to determine whether the fractured-rock aquifer can be treated as a porous medium. The
criteria listed below are subjective because fractured-rock aquifers fall on a continuum
15
-------
between true porous-media and conduit systems, and the decision whether ito use the
porous-media model will always require professional judgment and experience. Treating the
aquifer as a porous medium at the wellhead protection scale should never imply that the
porous-media assumption is valid at smaller scales or for site-specific problems.
Although the porous-media assumption may often be justified, critical errors can result if
the assumption is incorrectly applied in situations where it does not hold. The porous-media
assumption is usually not justified in conduit karst environments, where ground water can
move under turbulent flow through networks of cave systems. Ground-water investigations in
such areas can require detailed analyses of cave systems, including cave mapping and tracer
studies (Quinlan and Ewers, 1985).
Determining Whether Fractured-Rock Aquifers Behave as Ponms Media
Subjective-criteria for determining whether fractured rock can be treated as a porous
medium for the purposes of wellhead protection include pumping test responses, configuration
of the water-table surface, the ratio of fracture scale to problem scale, distribution of ,
hydraulic conductivity, and variations in water chemistry and water quality.
Pumping test responses.-Then are three main criteria for determining whether a
fractured-rock aquifer approximates a porous medium using an aquifer putinping test.
1) The drawdown in observation wells should increase linearly with increases in the
discharge rate of the pumped well. Rickey (1984) suggests conducting a series of
1-hour pumping tests with incrementally increased discharge ratest. A plot of
drawdown at 1 hour versus pumping rate for the series of tests should approximate a
straight line in porous-media^-equivalent settings (fig. 6A).
2) Time-drawdown curves for observation wells located in two or more different
directions from the pumped well should be similar in shape and should not show
sharp inflections, which could indicate hydraulic boundaries (fig. 6B).
3) A plotted drawdown cone from a pumping test using multiple observation wells
should be either circular or elliptical. Linear or very irregular cones can indicate a
failure of the porous-media assumption (fig. 6C).
Water-table surface configuration.--^ porous-media-equivalent fractured aquifers a
water-table map should show a smooth and continuous water-table surface without areas of
rapidly changing or anomalous water levels. In particular, the water table should not have the
"stair-step" appearance that-can occur in sparsely fractured rocks with large contrasts in
hydraulic conductivity between blocks and fractures. Although a "stair-step" water table
clearly indicates a failure of the porous-media assumption, a smooth water-table map dees not
prove a porous-media-equivalent setting. Detection of irregularities in the; water table may
require more closely spaced monitoring wells than are available for most wellhead protection
studies. ,
16
-------
POROUS MEDIA EQUIVALENT
*1
x-
/
00
Discharge. gpm
NON-POROUS MEDIA EQUIVALENT
' .' . X
x
°. MO aoo abo <»o
Discharg*. gpcn
B.
POROUS MEDIA EQUIVALENT
10 100 1000
Time
NON-POROUS MEDIA EQUIVALENT
10 - 100
Time
C,
POROUS MEDIA EQUIVALENT
NON-POROUS MEDIA EQUIVALENT
S Drawdown, ft
X
X
\ .
II
-------
Ratio of fracture scale to problem scale.-far porous-media-equivalent aquifers, the
observed vertical and horizontal fractures should be numerous and the scale of fracturing
should be much smaller than the scale of the wellhead protection problem. As a rule ot
thumb the minimum dimension of the WHPA should be at least 100 times the average
fracture spacing. For example, at the Junction City and the Sevastopol test sites, bedrock
fractures are numerous and occur on a scale of a few feet or tens of feet, but the potential
WHPAs at both sites cover several square miles.
Hydraulic conductivity distribution.-^ porous-media-equivalent settings the distribution
of hydraulic conductivity, as estimated from piezometer slug tests or from specific capacity
analyses (Bradbury and Rothschild, 1985), should be approximately log-normal. In aquifers
where the hydraulic conductivitydistribution's strongly bimodal, the porous-media
assumption is probably hot valid. j
Variations in water chemistry.-Overall variations in ground-water chemistry have been
shown to be useful in determining whether fractured aquifers behave as diffuse-flow
(porous-media) systems or as conduit (discrete-fracture) systems. In conduit systems, ^
variations in ground-water chemistry can be significant from place to plac* and through time.
Water moving through conduits usually has less contact time with mineral surfaces and so can
be lower in total dissolved solids and have lower mineral saturation indices than water
moving through a diffuse-flow aquifer having similar mineralogy. Ground water in
diffuse-flow fractured aquifers will usually have relatively uniform chemical composition
through time and from place to place within the aquifer. For example, Sliuster and White
(1971) examined water chemistry and temperature of several springs in Pennsylvania and
showed that water chemistry in conduit-flow springs varied greatly with time, .but the water
chemistry of diffuse-flow springs did not (fig. 7). Municipal wells are sampled on a regular
basis for such parameters as temperature, pH, hardness, turbidity, and bacteria. The variation
in these parameters can help determine whether a fractured-rock aquifer tehaves as a porous
medium. For example, ground water high in turbidity and bacteria is mere often a result of
conduit flow rather than diffuse flow. If a production well never has a tinrbidity or bacteria
problem, it is probably not connected to significant conduits intersecting the land surface.
Evaluation of Fractured-Rock Aquifers for Wellhead Protection Studies
Wellhead protection studies in fractured rocks usually require an evaluation of the
hydrogeologic setting and properties of the potential WHPA. Essential ctaa for wellhead
protection studies in fractured rocks can include !
1) characterization of fracture patterns and locations, i
2) determination of hydraulic head distribution, and
3) determination of aquifer characteristics.
The amount of effort required to characterize fractured-rock aquifers varies with the
complexity of the aquifer and the scope of the wellhead protection study. For example, use
18
-------
Jan' F ' M " A " M " J
67
A S O N ' D "Jan" F M A
68
Rock Spring (conduit flow)
. spruce Creek Spring (diffuse flow)
12
10
1
o>
ex
I 6
'Jan' F ' M ' A
67
J ' J A S O N D 'Jan' F M A
68
Rock Spring (conduit flow)
Spruce Creek Spring (dffuM flow)
Figure 7. Examples of water chemistry and temperature variations with time in diffuse-flow
arid conduit-flow aquifers (after Shuster and White, 1971).
19
-------
of the arbitrary fixed radius method requires little or no aquifer characterization; numerical
models require extensive data sets. The following discussion is intended as a guide to some
of the techniques that could be used in wellhead protection studies in fractured rocks.
Characterization of Fracture Patterns and Locations ,
The first step in characterizing fractured-rock aquifers usually involves detennining the
spacing, orientation,.depth, and length of fractures present in the field, which is important for,
determining whether the porous-media approximation is applicable or not The methods
employed to gain this information depend on the surface or outcrop expression of fractures,
type of material, depth of unlithified material over the aquifer, and fracture orientation
(vertical and horizontal).
In areas where unlithified surficial materials are thin or absent, bedircck fracture patterns
can be mapped from aerial photographs or from detailed measurements on the ground. Figure
B2 (Appendix B) shows surface expression of fractures in dolomite beneath thin soil in Door
County, Wisconsin. Although such mapping methods provide good detail on fracture strike
and density, they give little information about the distribution of fractures at depth and,
provide only a two-dimensional measure of a three-dimensional system. Description of
fractures and measurement of fracture dip and strike in vertical outcrops can help provide an
understanding of the three-dimensional fracture system. Surface and borehole geophysical
methods can help characterize subsurface fractures in areas where there is little visual fracture
expression. Although most surface geophysical methods have the advantage of relatively
rapid data acquisition over large areas, the acquired data generally provide a depth-averaged
measure of some geophysical property over a large volume of subsurface material. Surface
methods can provide good delineation of zones of fractured rock, but detailed data on
individual fractures are difficult to obtain. Borehole geophysical methods provide detailed
information about rock properties along borehole walls and are particularly valuable for
investigations of fractured-rock aquifers.
Drilling is usually the most accurate way to locate fractures in the subsurface, and is also
required prior to the use of borehole geophysical methods. Core drilling and recovery is the
preferred method for obtaining a good record of subsurface fractures. Figure A4 (Appendix
A) shows the log, with interpretations of fractures and fracture zones, of a core obtained in
fractured crystalline rock at Junction City. Because vertical boreholes provide little
information about vertical fractures, angle drilling is sometimes used to characterize vertical
and near-vertical fracture systems.
I ' ' '
Determination of Hydraulic Head Distribution
In any wellhead protection study, determination of the three-dimensional distribution of
hydraulic head in the subsurface provides critical information for the prediction of
20
-------
ground-water flow paths and rates. In fractured systems, where hydraulic conductivity can
vary significantly in space, a detailed three-dimensional analysis of the head distribution is
particularly important. >
Frequently, in fractured-rock aquifers, total hydraulic head varies significantly with
depth Such variations are particularly important in wellhead protection studies .because
knowledge of vertical hydraulic gradients is required for vertical flow calculations, ^ _
Measurement of the vertical distribution of hydraulic head generally requires "nests of .
closely spaced piezometers screened at several different depths. Existing domestic wells with
long open intervals are usually inappropriate for measurement of vertical hydraulic gradients.
Figure B5 (Appendix B) shows how hydraulic head varies with depth at the Sevastopol test
site. -.--' . -..'.. " ' , i ''
Hydraulic heads in fractured-rock settings can vary significantly in response to
precipitation, snpwmelt, nearby pumping, surface-water fluctuations, and other phenomena.
An understanding of such variations is important to some wellhead protection studies because
extreme hydraulic head variations can lead to variations in ground-water flow rates and
directions at various times of the year. Determination of water-level fluctuations requires
periodic measurement of water levels in monitoring wells. Modern measuring and recording
equipment, such as pressure transducers and digital data loggers, has greatly simplified the
acquisition and processing of such data. Such elaborate data collection would be necessary
Only for more complete WHPA delineation methods.
1 " , ' - * I ' . ;
Determination of Aquifer Characteristics
Knowledge of aquifer parameters, including aquifer thickness, porosity, hydraulic
conductivity, and specific yield, is essential for the analyses of ground-water flow to wells
required in more complex WHPA delineation methods. In most cases, the standard testing
methods developed for porous, granular aquifers can be applied to fractured-rbck aquifers
with little modification. However, use of these methods in fractured rock requires some
special considerations, outlined below.
Aquifer thickness is an important parameter in most standard equations of ground-water
flow to wells. In porous-media settings, where aquifers are bounded by confining beds, the
aquifer thickness is usually obvious and easy to measure. In fractured-rock settings, however.
aquifer thickness can be problematic. An example is the crystalline rock aquifer at Junction
City, Wisconsin (Appendix A). This aquifer is formed in the upper part of the Prccambnan
basement rocks of central Wisconsin; the lower stratigraphic boundary of which is unknown.
In fact, as in many non-plutonic basement rocks, the original deposititfnal (bedding) planes
are now vertical or steeply dipping. For the purpose of this report, the aquifer boundary was
defined as the depth at which significant fractures, as measured by borehole logging and
observed in a core sample, disappear.
21
-------
Porosity is an important value in wellhead protection studies because it has a large
influence on time of travel calculations. Porosities of fractured rocks are generally less than
porosities of non-fractured rocks, as shown in table 2.
Table 2. Range of values of porosity
(from Freeze and Cherry, 1979).
Material
Unconsolidated deposits
Gravel 25-40
. Sand 25-50
Silt 35-50
Clay 40-70
Rocks >
Fractured basalt 5-50
Karst limestone 5-50
Sandstone 5-30
Limestone, dolomite 0-20
Shale 0-10
Fractured crystalline rock 0-10
Dense crystalline rock 0-5
In general, porosities of unconfmed fractured aquifers can be estimated from the specific
yield values obtained during aquifer pumping tests. More accurate porosity measurements are
sometimes obtained using field tracer experiments, but such field tests art; often complex and
beyond the scope of most wellhead protection studies. ,
Methods for determining basic aquifer parameters for wellhead protection studies include
- use of "textbook" values,
- use of specific capacity data,
- piezometer slug tests,
- single-well pumping tests,
- multiple-well pumping tests, and
- packer tests.
"' t
Many widely available hydrogeology texts (Davis and DeWiest, 1966; Freeze and
Cherry, 1979) list typical values of hydraulic conductivity and storage coefficient for a variety
22
-------
of fractured-rock settings, and such values can be used as a starting pomt in wellhead
protection studies. However, location-specific data would be required for more complex
wellhead protection studies.
The specific capacity of a well refers to the drawdown occurring for a given pumping
rate at a pseudo-steady state. Wellconstructors commonly report specif*: capacity data to
government agencies upon the completion of many domestic and industrial wells. Appendices
A and B provide examples of the use of specific-capacity data to estimate transmissmty and
hydraulic conductivity in fractured-rock settings. For site-specific studies, the piezorne er
rate-of-rise or "slug" test (Freeze and Cherry, 1979) provides an estimate of the hydraulic
conductivity around the piezometer tip. Tables A2 and B2 (Appendices A and B) give
examples of slug-test data at the test sites in Wisconsin.
Full-scale pumping tests involving several monitoring wells probably provide the most -
reliable measure of aquifer parameters for wellhead protection studies. Such tests can provide
Li averaged over a large fractured-rock mass and can be used to measure the vertical and
horizontal anisotropy Of the fractured-rock mass (Hsieh and others, 1985; PaPa<^^;.,
1965) Hickey (1984) used apumping test to test the .Darcian flow assumption for fractured
limestone in Florida. Although often a preferred method of data acquisition, pumping tests in
fractured-rock systems can be expensive to conduct and complex to analyze. Appendices A
and B discuss the results of pumping tests in fractured rock at two test sites in Wisconsin.
' " . . , '.' !
Packer tests, in which inflatable packers are used to isolate specific parts of a borehole,
are a variation of the standard pumping test. Packer tests have particular application to
fractured-rock aquifers. By isolating particular fractures or fracture zones, the spwific
properties of these zones can be tested. For example, Shapiro and Nicholas (1989) used-
packer testing and analyses in fractured dolomite in northern Illinois to estimate fracture
properties. . ,
23
-------
-------
Chapter m
WELLHEAD PROTECTION AREA DELINEATION
Zones Used for WHPA Delineation
The purpose of wellhead protection area (WHPA) delineation is to approximate the area
through which water flows to a well so that management decisions regarding the control of
contamination sources in that area can be implemented. The area through which water
recharges a well is the zone of contribution, or ZOC (fig. 8). In contrast, the zone of
influence, or ZOI, is the area affected by a pumping well, and coincides with the area! extent
of the cone of depression.- The ZOI extends outward from the pumping well to the point of
negligible drawdown (fig. 8). Field measurements of drawdown resulting from well pumpage
give the most accurate determination of the ZOI.
In a homogeneous, isotropic aquifer with a horizontal water table, the ZOI is a circle and
only water within that circle flows to the pumping well. In such an aquifer, the ZOI is
equivalent to a ZOC. In the more common case of a sloping water table, however, the ZOI
and the ZOC can be quite different (fig. 8). Water upgradient of a well will flow toward the
well even if it is outside the well's cone of depression. Likewise, some areas might be within
the ZOI and yet not be drawn down enough to reverse the natural hydraulic gradient. Water
within such areas will flow away from the well.
This chapter describes the criteria commonly used in the WHPA delineation process and
examines various methods for determination of the ZOC in fractured rocks. Example
applications of the various methods are given for two test sites in different geologic settings.
The aquifer at the Junction City site is developed in highly fractured metamorphic rock. The
aquifer at the Sevastopol site is developed in fractured dolomite.
Although fractures are abundant at both sites, the characteristics of the fractures and the
sites are quite different, as described in Appendices A and B. At the Junction City site, the
aquifer is covered by up to 55 ft of silt and clay. Fractures in the igneous and metamorphie
rocks are most abundant near the bedrock surface, and decrease in size and frequency with
depth. They have narrow apertures (typically less than 0.01 in.), frequently contain fillings of
calcite or other minerals, and probably are short (typically less than 30 ft). Fracture spacings
are irregular, ranging from a few tenths of an inch to several feet. Fractures are frequently
vertical, especially between 130 and 160 ft below the land surface.
In contrast, at the Sevastopol site, the dolomite is either exposed at the land surface or is
covered by soil less than 5 ft thick. Fracture apertures range from a fraction of an inch to
several inches. Some fractures have been widened by solution, and minor karst features occur
in the area. However, these karst features are infrequent and isolated. Fractures in the
dolomite tend to be regularly spaced and can be traced across the landscape for hundreds or
25
-------
20C-
I^GROUNDWATER
I DIVIDE
A'
LAND SURFACE
BEDROCK
SURFACE
PRE-PUMPING
WATER TABLE
VERTICAL PROFILE
(B) PLAN VIEW
LEGEND:
V Water table
Ground-water Flow Direction
Pumping Well
ZOi Zone of Influence .
ZOC Zone of Contribution
NOT TO SCALE
Figure 8. Terminology for WHPA delineation in fractured rocks (adapted from U.S. EPA, 1987).
26
-------
thousands of feet. Although fractures occur throughout the dolomite aquifer, horizontal zones
of interconnected fractures with high hydraulic conductivity occur at several different depths.
Common Delineation Criteria _
Five common criteria considered in delineating a WHPA are: 1) distance from the.well,
, 2) drawdown around the well, 3) time of travel (TOT) to the well, 4) physical boundaries to
the ground-water system, and 5) capacity of the subsurface environment to assimilate
contamination (U.S. EPA, 1987). This report assesses various combinations of these criteria
in delineating WHPAs in fractured rocks.
Distance
The most basic criterion for WHPA delineation is distance from the well without regard
for the ground-water flow direction. The distance criterion delineates a WHPA by placing a
circle of fixed radius around the well. The radius can be arbitrarily chosen or calculated
.using an analytical equation usually involving a specific navel-time threshold.
Drawdown
- ' ' ' , '-""'-..' '-..'. . V
The drawdown criterion defines the areal extent of the cone of depression (ZOI) of a
well. In areas where the water table is essentially flat, the ZOI coincides with the well's zone
of contribution; however, such cases are rare in practice. Delineation of a WHPA based on
drawdown requires estimations of transmissivity and storativity, and selection of a drawdown
limit (for example, 0.05 ft), beyond which the drawdown is considered negligible.
Time of Travel (TOT)
The TOT criterion bases WHPA delineation on the amount of time it takes ground water
to travel from a recharge ppint at the land surface to a well, and theoretically incorporates all
the processes involved in contaminant transport. Practically, however, the TOT criterion
usually incorporates advection only. Incorporation of contaminant dispersion, diffusion, or
retardation is difficult given the available information on these processes for a given
compound. In addition, it would render the TOT delineations too compound-specific. The
use of TOT/distance calculations based on the average linear velocity of the ground water is a
particularly feasible approach for a generalized wellhead protection program. Limiting the
contaminant transport processes to advection only is most valid where ground-water flow
velocities are relatively high, as often occurs in fractured rocks.
27
-------
A TOT limit (for example, 10 or 20 years) can be chosen in various ways. One method,
based on chemical degradation rates, sets the limit at the point where contaminant residues
are no longer considered a threat to water quality. Alternatively, the TOT limit can be based
on the time required to contain contamination of an aquifer. The TOT limit chosen can vary
considerably depending on the nature of the main contamination risk. Delineating a WHPA
with a TOT criterion implies that the TOT is translated into a distance from the well through
determination of ground-water velocity. Delineating WHPAs based on (the TOT criterion
therefore requires more technical input than either the distance or the drawdown criterion.
Required information includes an estimate of aquifer hydraulic conductivity, an estimate: of
aquifer porosity, and determination of the magnitude and direction of the hydraulic gradient.
Flow-System Boundaries
The flow-boundary criterion uses ground-water divides, surface-water bodies, or other
hydrologic/physical features to delineate a WHPA. For a localized flow system, the WHPA
often corresponds to a ZOC, and water entering the ground-water system inside the ZOC is
assumed to eventually reach the well. The flow-system-boundary criterion therefore provides
the maximum protection for a given well, but implementation of a WHPA based on
hydrogeologic boundaries may be impractical if the defined area is too large. The
flow-boundary criterion is most useful for aquifer systems where the boundaries are easily
defined and the distances from the boundaries to the well are relatively short
Assimilative Capacity
The assimilative-capacity criterion is used in combination with the TOT criterion with
the assumption that contaminant attenuation will take place within the un saturated and/or
saturated flow system. Non-toxic concentration levels and the time it takes to reach those
levels must be determined to delineate a WHPA on the basis of this criterion. Such a
designation is compound-specific and the processes involved in attenuation are often not well
understood and not easily quantified; and this is particularly true in fractured rocks. This
criterion was not used in WHPA delineation at the two test sites in Wisconsin.
^Evaluation of Criteria
Some combinations of the above criteria will work better than others in establishing a
wellhead protection program in fractured rocks. The choice of criteria depends on the
availability of existing hydrogeologic information, the personnel and financial resources
available, and the complexity of the fractured-rock setting. Some generalizations regarding
these criteria can be made. The distance criterion alone gives a very poor measure of the
ZOC because it does not account for any hydrogeologic site characteristics. Drawdown is a
useful criterion if the pre-pumping water table is; relatively flat and if there is information
28
-------
regarding the well pumpage and accurate aquifer parameter estimates. The TOT criterion is
useful for establishing practical WHPA boundaries in situations where ground-water flow is
mostly horizontal and existing hydrogeologic boundaries are far from the protected well.
However, the TOT criterion can be misleading and difficult to establish in sparsely fractured
rock where highly conductive fracture-conduits occur and where significant vertical flow
occurs. The assimilative-capacity criterion is useful, but it is difficult to implement due to the
lack of information regarding the fate of contaminants in the saturated zone. This criterion
cannot stand alone because TOT considerations are necessary to translate a time of concern
into a distance on a map. The best criterion probably is the flow-system-boundary criterion.
A WHPA based on this criterion might also be less sensitive to the porous-media assumption
in the fractured-rock environment than WHPAs basedon other criteria. Delineation of
ground-water divides and general flow directions is carried out over a large enough area that
the assumption' that-fractured"lock behaves as a uniform porous medium is likely to be valid.
WHPA Delineation Methods for Fractured Rocks
Many methods have been recommended for the delineation of WHPAs. This study
tested six methods for delineating WHPAs using various combinations of the criteria listed
above. The methods tested, in order of increasing complexity, were
1) arbitrary fixed radius,
2) calculated fixed radius,
3) vulnerability mapping,
4) flow-system mapping,
- with TOT criterion,
- with analytical equations,
5) residence-time approach, and ,
6) numerical flow/transport models.
The first two methods are not particularly suitable for the accurate delineation of
WHPAs in fractured rocks. The arbitrary fixed radius method does hot incorporate any
hydrogeologic or contaminant transport considerations, and can best be used as a first-step
approach. When the radius is large enough, the true ZOC will be included within the WHPA
delineated by methods 1 and 2 and will be protected. However, large areas outside the ZOC
will also be protected. The application of analytical flow equations to calculate a fixed radius
brings an improvement over die arbitrary fixed radius method, but may not give acceptable
results in unconfined fractured-rock settings because it fails to account for heterogeneity,
anisotropy, ground-water recharge, and vertical components of flow, all of which can occur in
fractured-rock settings. For example, at the two Wisconsin test sites, the Theis
nonequilibrium equation (Theis, 1935) gives unrealistic radii around the protected well. The
circles calculated by this method are top large and encompass so much area that
implementation of an appropriate WHPA would be difficult. On- problem with the Theis
nonequilibrium equation approach stems from the estimation of the time to reach steady state.
The final radius is highly sensitive to this parameter that is difficult to estimate. Unless'the
29
-------
time can be somewhat accurately determined, this approach will have significant limitations
for fractured-rock aquifer analyses.
Vulnerability mapping uses geologic maps, soils maps, water-table maps, aerial
photographs, and mapping of surficial features to identify areas of the landscape particularly
vulnerable to ground-water contamination. Vulnerability mapping dries not produce a ZOC
for a given well; however, it does identify significant fractures near the well that may
contribute to ground-water contamination. Vulnerability mapping combined with the arbitrary
fixed radius method or the simplified variable shapes method (U.S. EPA, 1987) can be used
to delineate a WHPA in fractured rocks that do not approximate porous media.
The remaining three methods-flow-system mapping and flow-system mapping combined
with TOT criterion or with, analytical equations, .the,residence-time approach, and numerical
modeling-were found suitable for ZOC delineation in unconfined fractured rocks that behave
as porous media at .the WHPA scale. The following sections describe each of these methods
in detail.
Four WHPA delineation approaches are suggested for unconfined fraetured-rock aquifers
that do not behave as porous media. These methods include vulnerability mapping combined
with the arbitrary fixed radius method or the simplified variable shapes method;
hydrogeologic mapping; the residence-time approach; and numerical ground-water
flow/transport modeling. These methods are discussed in more detail in the following
sections. .,'-,
Vulnerability Mapping
£>escnpri
-------
areas particularly susceptible to ground-water contamination could be conducted as part of the
data collection far other methods.
Advantages of vulnerability, mapping are
1) the assumption of a uniform porous medium is not necessary;
2) the method does not require detailed measurements of aquifer parameters;,
3) the method uses a variety of data, ranging from office-available maps to
field-measured surface features.
Disadvantages of vulnerability mapping are
1) the method does not delineate a ZOC for the well;
2) the results are somewhat subjective.
'.''' - i ' -" -.,
: Example from the Sevastopol jife^rSchuster and others (1989) constructed a map of
exposed bedrock and shallow soils for northern Door County, Wisconsin, including the area
around the Sevastopol test site. This map considered fracture traces, areas of exposed
bedrock, solution features, closed topographic depressions, and soil attenuation potential.
Areas of the landscape containing these features and having low soil attenuation potential
have the highest potential for ground-water contamination. As shown on figure 9, the areas
susceptible to contamination are irregular and not centered on the well to be protected. A
WHPA for the well might be circular, using an arbitrary fixed radius of 4000 ft, or could be
delineated using a simplified variable shape (U.S. EPA, 1987) oriented around the well
according to the regional ground-water flow direction.
Flow-System Mapping
DMcriprion.»Hydrogeologic mapping (U,S. EfcA, 1987) identifies the physical and
hydrologic features that control ground-water flow. Physical boundaries to ground-water flow
can include the geologic contacts that form the limits of the aquifer, structural features such
as fault-block walls or zones of fracturing, and topographic features that may function as
ground-water divides. Hydrologic features, including rivers, canals, and lakes, can function as
flow-system boundaries. The flow-system mapping method, a subset of the hydrogeologic
mapping method, uses ground-water divides and flow-system boundaries derived from a
water-table map to delineate the ZOC for a given well.
How-system mapping assumes that hydrogeologic boundaries, particularly potentiometric
boundaries, are stationary through time. In aquifers where water levels fluctuate seasonally or
where well drawdowns approach potentiometric divides, caution must be used when
delineating boundaries for ZCC analysis.
Row-system mapping requires detailed mapping of the configuration of the water table.
Ideally, investigators should use field measurements in properly constructed monitoring wells
and nested piezometers for construction of such maps. In practice, funding and time
31
-------
Direction of regional ground-water flow
Test well (MW-1)
\ \ Contamination susceptible area based
on field mapping of exposed fractures,
thin soils, and other surface features
(after Schuster and others, 1989)
WHPA based en arbitrary radius
WHPA based on simplified variable
shape
FEET 600 0 SCO 1000 2000
Figure 9. WHPAs based on vulnerability mapping at the Sevastopol site.
32
-------
considerations can rule out such detailed field work. In some situations, available office data,
in the form of water levels on well constructors' reports, previous hydrogeologic studies, and
surface-water features on topographic maps, can produce acceptable water-table maps
(Blanchard and Bradbury, 1987). Field measurements of water levels in existing domestic
and industrial wells can supplement these data. Appendices A and B (figs. A8 and B4)
provide examples of water-table maps in fractured-rcck settings from the two study areas in
Wisconsin.
: ' ' .. . "' , ' ^ . ' ' '" '
Once a water-table map is constructed, flow lines are drawn perpendicular to the
water-table elevation lines. These flow lines begin at the well and extend upgradient to the
ground-water divide. Using a water-table map to determine ground-water flow lines assumes
an isotropic aquifer, which is not always the case in fractured-rock settings. In simple
hydrogeologic settings (without major faults, fades changesvetc.), the ZOC delineated by the
flow-system mapping method takes into account the ground-water flow system geometry. It
neither includes downgradient areas that do not contribute water to the well nor excludes
upgradient areas that do contribute water to the well. This method tends to be conservative in
the sense that it usually overestimates the true ZOC for a given well.
Advantages of the flow-system mapping method are
1) the method is simple and requires only limited training in hydrogeology;
2) the method cam be used with various types of data ranging from office data to
detailed field data;
3) the method uses mappable hydrogeologic boundaries. ,
Disadvantages of the flow-system mapping method are
1) the method assumes a uniform, two-dimensional aquifer that approximates a uniform
porous medium;
2) the method can produce unaccepjably large ZOC estimates if the protected well is
located far from a ground-water divide;
3) errors in the water-table map can cause large errors in ZOC delineation.
Results of the application of the flow-system mapping method at the Junction City and
Sevastopol sites are discussed below together with this results of the flow-system/TOT
method, and shown in figures 10, 11, and 12.
Flow-System Mapping with Time of Travel Calculations
A water-table map can be used to estimate the horizontal hydraulic gradient. Using the.
gradient in combination with estimates ,bf hydraulic conductivity; and aquifer porosity,
ground-water velocity can be calculated according to:
~v = Ki/n (D
33
-------
where T is the average linear velocity of the ground water, K is the horizontal hydraulic
conductivity, i is the horizontal hydraulic gradient, and n is the porosity. The velocity, in
combination with a specified time of travel, can be used to limit the WHPA to that portion of
the ZOC that will contribute water to the well within a specified amount of time.
Determination of the position of the TOT line incorporates the assumption that contaminants
in ground water will move in the same direction and at the same velocity as the ground
water.
Calculation of the TOT boundary is based on:
d="vs ',(2)
where d is the upgradient distance from the well to the TOT line, ₯ is the average linear
velocity across the ZOC (Calculated'using Eq. 1 above), and t is the desiwsd time of travel.
Note that the hydraulic gradient, i, in Eq. 1, is calculated as the total change in water-table
elevation from the upgradient ZOC boundary to the well divided by the horizontal distance
from the upgradient ZOC boundary to the well. This is clearly a simplification of reality
because, in most cases, i will not be uniform over the basin. However, w most cases, the
error in the location of the TOT line will be small.
Advantages- of combining flow-system mapping with the TOT criterion are
1) the TOT criterion provides a way to limit the WHPA in areas where the ZOC
delineated from flow-system boundaries is unacceptably large;
2) adding the TOT criterion requires little additional work once the flow-system method
has been completed;
3) the method,requires only elementary mathematics.
Disadvantages of combining flow-system mapping with the TOT criterion are
1) errors in estimates of porosity or hydraulic conductivity can cause large errors in the
TOT calculation and thus in WHPA delineation;
2) the method assumes a uniform, two-dimensional aquifer that approximates a uniform
porous medium;
3) the presence of a highly conductive fracture zone could cause very large errors in the
TOT calculation and in the resulting WHPA.
Application of the flow-system/TOT method is shown with examples! from the Junction
City and Sevastopol sites.. The examples use only the most accurate aquifer parameter
estimates derived from pumping tests. The use of the flow-system mapping method at both
sites required the assumption that the fractured rock behaves as a uniform porous medium at
the WHPA scale.
Example 1: Junction City «Ye.--Detailed site investigation conducted at the Junction City
well field provided a water-table map that clearly indicated a cone of depression around the
village well (hatchured contours, fig. 10). The cone creates ground-water divides between
34
-------
I _-
SCALE 1:24 000
FEET 500 0 500 1000 2000
*-n4o Water-table contour
(interval 10 ft)
Ground-water divide
Village well (JC-9)
Zone of contribution
Time of travel .
Figure 10. ZOC delineation in crystalline rocks using a field-measured water-table map. A,
B, and C are points where hydraulic gradients and ground-water velocities were calculated.
Velocities were calculated using the hydraulic conductivity detennined from the pumping test.
35
-------
flow lines reaching the village well and flow lines bypassing it. Outlining these ground-water
divides and extending the divides upgradient to the regional water-table divide produces the
ZOC shown in figure 10.
Once the ZOC has been delineated, Eq. 1 can be used to estimate .
velocities needed to delineate that portion of the ZOC within a specific TOT threshold. The
horizontal hydraulic gradient (« in Eq. 1) represents the decrease in total hydraulic head over a
horizontal distance. The hydraulic gradients at points A, B, and C on fiijjure 10 are,
respectively, 0.0063, 0.013, and 0.05. Hydraulic conductivity estimated from a pumping test
of the village well (Appendix A) was 8.3xlQ-5 ft/sec (about an order of magnitude higher than
that estimated from slug tests and specific capacity data). Using Eq. 1, with a fractured-rock
porosity of 0 10 (Freeze and Cherry, 1979), calculated ground-water velocities at points A, B,
and C are about 165,340, and 1300 ft/yr. .;Note-that= ground-water velocities increase^
significantly near the village well as the ground-water flow paths convCTgett) enter the well.
Because of the variability of ground-water velocities over the study area TOT lines based on
this analysis are only approximate. Figure 10 shows approximate TO1 distances for 1 year, 4
years, and 13 years.
Example 2 Sevastopol sire.~The Sevastopol site example demonsilrates use of the
flow-system mapping method at a site where the production well is not yet in use and no
mappable cone of depression yet exists. In addition, water-level measurements at multilevel
piezometer nests revealed that the ground-water system at the Sevastqxd site was more
complex than that at Junction City and required a three-dimensional analysis. The system can
be best characterized by two separate flow systems, one shallow and one deep, which .
correspond to zones of intense fracturing, separated by zones of less-fractured dolomite. Hie
water-table map for the shallow system and the potentiometric-surface map for the deep
system and their respective ZOCs are shown in figures 11 and 12, respectively. Because no
cone of depression exists around test well MW1, the downgradient boundaries of the ZOCs
were squared off and flow lines were started arbitrarily 500 ft on either side of the well for
the shallow system (fig. 11) and 1000 ft on either side of the well for the deep system (fig.
12). The flow lines were then drawn upgradient to the ground-water divide.
From the perspective of implementation, there is a significant problem with a narrow
WHPA, as shown by the shaded zone in figure 12. In such areas, the delineation of the ZOC
depends on the position of the potentiometric contours, and a relatively small error in the
potentiometric map can result in a major shift in the ZOC. This problem is most critical
when the ZOC is long and narrow.
One method to help insure that the WHPA covers the actual ZOC in such areas is to
delineate a buffer zone around the mapped ZOC (U.S. EPA, 1987X In figure &*****«
zone was extended outward from the mapped ZOC boundary by 250 ft for every 1000 ft of
distance upgradient from the well.
36
-------
_Q_:_ _. 1 '-
^eao Water-table contour
(interval 10 ft) , "'-
o o o Ground-water divide
, * Test well (MW-1) '
Zone of contribution
SCALE 1:24 000
FEET 600 0 500 1000 2000
^760 * «,
\ ^ »«. *' yo
Y nep"^<- °
I
I i ^ *
-1- ^ *
\ ^ -
U
I
If
Figure 11. ZOC delineation in a shallow ground-water system in dolomite using a.
water-table map.
, ' -. .31 ..'-. -- '
-------
Water-table contour
.(interval Sit)
Ground-water divide
Test well (MW-1)
Zone of contribution
Time of travel
Butter zortt
SCALE 1:24 000
500 0 5001(1)00 2000
Figure 12. ZOC delineation in a deep ground-water system in dolomite using a
potentiometric-surface map.
38
-------
In the
(derived from
' ftJrt' «!d AfTOT^m'tiie3^nd-^^vidTio the well (a distance of about 650 ft) is
S5 '*££& IKM S5^
. . ' *»_:,.? «>»iry* ft/**r ^inh ft/vr"i Based on this velocity, TOT lines
velocity of the deep system is J.AXIU iwscc \.*w" wy*j- **»»»»»
have been drawn on figure 12.
Delineating an adequate ZOC at the Sevastopol site required combining the shaUow and
ZOCs into one overall ZOC. A conservative.^umpnon for^pmpo^ Qf
combined ZOC.
The combined ZOC was drawn by first overlaying the individual ZOCs from the shallow
and deep systems, and then adding an additional area where water ^£*£? *
could move and to enter the ZOC of the deep system. Points A and B on figure 13
AeLctionsof ground-waterrmvement in this combined ZOC Point Alies*>**
the deep system. Water entering the deep system at point A will foUow Abdicated
the well Point B does not lie in the ZOC for the deep system, but it does lie in that part of
me Sow system that could contribute to the deep system ZOC. Water entenng the shallow
system at point B could move southeast through the shallow system, downward to the deep
system, then southwest through the deep system to the well.
Flow-System Mapping with Uniform Flow Equation
Description.--^ construction of a water-table map allows the application of the uniform
flow equation (Todd, 1980) to define the ZOC to a pumping well in a sloping water able
(fie K) The input requirements are the same as for combining flow-system mapping wur,_
the TOT criterion. The uniform flow equation assumes a uniform porous medium and can be
expressed as:
-Y/X = tan(2nKbiY/Q) <3)
where Y is the distance from the Well parallel to the pre-pumping equipotentid Unes,:JfJiAe
distance from the well perpendicular to the pre-pumping equipotennal hnes, K is the hydraulic
conductivity, b is the saturated thickness of an aquifer, i is the pre-pumping hydraulic
39
-------
o
o
o
o
o
26
ooo Ground-water divide (shallow system)
Ground-water divide (deep system)
* Test well (MW-1) -v
Zone of contribution
j%%4 Shallow system [JH^ Deep system
Area where shallow system may
contribute to ZOC of deep system
SCALE 124 000
FEET SOO 0 SOO 1000 2000
Figure 13. ZOC delineation in dolomite using water-table and potentiometric-surface maps.
Arrows show hypothetical travel paths for water particles originating in the deep system
(Point A) and in the shallow system (Point B).
40
-------
ORIGINAL
PIEZOMETRIC
SURFACE
GROUND
0 /SURFACE
DRAWDOWN QURVE
IMPERMEABLE
t T-T
E
CONFINED
AQUIFER
i tine Dime A Rl P
>
r
(a)
EQUjPOTENTIAL LINES S
GROUNDWATER
I
UNIFORM-FLOW
EQUATION
X, - -
2Kbi
DISTANCE TO
DOWN-GRADIENT
NULL POINT
BOUNDARY
LIMIT
LEGEND:
Pumping Well
Where: -, ,.
Q « Well Pumping Rate
K - Hydraulic Conductivity
b * Saturated Thickness
i« Hydraulic Gradient
NOT TO SCALE
Figure 14. ZOC delineation using the uniform flow equation.
41-
-------
gradient, and Q is the well pumping rate. This equation leads to two equations that delineate
the ZOC of a well:
and
XL = -Q/(2nKbi)
YL = ±QI(2Kbi)
(4)
(5)
where XL is the distance from the well to the prc-pumping downgradient null or stagnation
point, and YL is the distance to the transverse boundary limits from the uipgradient boundary
center (fig. 14).
Advantages of this method are
1) the method accounts for some of the effects of pumping on the ZOC without detailed
mapping of a cone of depression, which reduces the amount of required field work;
2) the method is simple and requires only limited training in hydrogeolpgy;
3) the method uses data derived from a water-table map.
- , I- , i :
Disadvantages of this method are
1) the method assumes a uniform, two-dimensional aquifer that approximates a uniform
porous medium;
2) the method ignores the effects of hydrologic boundaries (except {ground-water
divides), aquifer heterogeneities, and non-uniform recharge;
3) the method can produce unacceptably large ZOC estimates if the protected well is
located far from the ground-water divide;
4) errors in the water-table map or in estimates of porosity or hydraulic conductivity
can cause large errors in ZOC delineation. |
Example 1: Junction City site.Using the water-lable map and the hydraulic
conductivity estimate based on the pumping test, the following values were assigned to the
variables in the uniform flow equation:
K = 8.3xia5 ft/sec, |
b= 160ft,
i = 0.0047,
Q = 0.072 ftVsec. , i
These values resulted in a downgradient null point (XL) of -180 ft and a transverse null
point (Yd of +580 ft (fig. 15). The ZOC extending to the ground-water divide in this
example is narrower than the ZOC delineated from the water-table map (illg. 10), because the
mapped cone of depression is larger than would be predicted by using the uniform flow
equation.
42
-------
SCALE 1:24 000
FEET 500 0 500 1000 2000
Ground-water divide
Village well (JC-9)
Zone of contribution
11
Figure IS. ZOC delineation in crystalline rocks using the uniform flow equation.
'- ''' ' . 43 " ' ' '
-------
Example 2: Sevastopol site.--On the basis of aquifer evaluation, geophysical logging,
and the potentiometric-surface map of the deep system (see Appendix B), the following
parameters were assigned to the deep aquifer system:
K= 1.4X104 ft/sec,
b = 410ft, i
i = 0.0046,
Q = 100 gpm (0.22 ftVsec).
The calculated XL and YL are -130 ft and ±420 ft for the hypothetical municipal well.
The ZOC drawn with these limits is shown in figure 16. Note that the ZOC has been curved
to follow the regional direction of ground-water flow.
Residence-Time Approach
Description.-Thc residence-time approach utilizes water chemistry and isotopes to
identify ground-water travel paths and flow rates. Geochemical parameters; (for example,
mineral concentrations and saturation indices) can help indicate the source area of ground
water. Environmental isotopes (tritium, oxygen-18) in ground water can be used to estimate a
minimum age of water produced by a well. Such analyses are relevant to iiZOC and TOT
analyses in three ways. First, relative age determinations can provide a check on travel-time
estimates obtained by the hydraulic approaches described above. Second, in areas where the
water produced by a well can be shown to be hundreds or thousands of years old, the
potential ZOC of a well might be so large that local wellhead protection might not be
appropriate or effective. Third, in areas where the geochemical and isotopic signatures of
ground water vary radically from place to place, these variations can be used to differentiate
.zones of rapid recharge from zones of less rapid recharge. For example, a well located near a
river that produces water having geochemical and isotopic contents similar to the river water
might be directly connected to the river through the fracture network; a well adjacent to a
river that produces water with a different geochemical and isotopic _ content might hot be
directly connected to the river.
Tritium (3H) is a radioactive isotope of hydrogen that is naturally present at low levels in
the earth's atmosphere, but tritium in the atmosphere increased dramatically following
atmospheric atomic weapons testing from 1952 to the mid-1960s. During this time, all
recharging ground water was enriched with tritium, and ground water that has entered aquifers
since 1952 generally contains elevated tritium levels. The half-life of tritium (12.3 years) is
relatively short, making it an excellent indicator of recent ground-water recharge and relative
ground-water age (Egboka and others, 1983; Knott and Olimpio, 1986), where age is defined
as the time since the water was in contact with the atmosphere. Heridry (1988) summanzed
the general qualitative interpretations of ground-water age on the basis of tritium in
ground-water (table 3). Tritium analyses are reported in tritium units--a ratio of tritium atoms
(3H) to the much more common 'H atoms. One tritium unit, or TU, repreasents one tritium
atom per 10U hydrogen atoms.
44
-------
640 Water-table contour
(interval 5 ft)
- Ground-water divide
Test well (MW-1)
\ | Zone of contribution
SCALE 1:24 000
FEET 600 0 500 1000 2000
Figure 16, ZOC delineation in a deep ground-water system in dolomite using the uniform
flow equation. {
..:'.'--/ .. 45 ."' ' '...." - / :,.
-------
Table 3. Qualitative interpretations of tritium concentrations in ground water
(fromHendry, 1988).
Concentration (TU) Interpretation
> 100 Average ground water likely recharged during
thermonuclear testing between 1960 Mid 1965.
10-100 Average ground water less than 35 years old.
2-10 Average ground water at least 20 yeiiirs old.
< 2.0 Average ground water older than 30 years.
< 0.2 Average ground water older than 50 years.
Oxygen-18 ("O) is a naturally occurring isotope of oxygen present as low concentrations
in air and water. The ratio of "O to the more common "O is a function of climate, season,
latitude, and weather patterns. In general, the 18O/16O ratio becomes lower in more northerly
latitudes and colder climates, and so the "O content of ground water has been used as an
indicator of the climate at the time the water recharged. Bradbury (1985) ?wd Desaulniers
and others (1981) have used the "O content of ground water in clayey fractured tills in
Wisconsin and Ontario, respectively, to suggest that the ground water rechiirged in a colder
climate than present, possibly as much as 10,000 years ago. In addition, lhe-llO/I^O ratio in
precipitation varies seasonally, and variations in the ratio are often used to distinguish ground
water originating from different recharge areas. .As with tritium, 1SO values are reported as a
ratio deviation from a standard. For 18O the standard is known as Standard Mean Ocean
Water (SMOW); results are reported as per mil deviations from this standaiid.
The residence-time approach requires the collection of high-quality ground-water samples
from pumping wells, monitoring wells, and discharge points such as springs and streams.
These samples are tested for a full range of inorganic cations and anions as well as such field
parameters as pH, conductivity, dissolved oxygen, and temperature. In adclition, the water
should be analyzed for 3H and "O. This method also requires information about the
mineralogy and geochemistry of rocks supplying water to the well.
Advantages of the residence-time approach are
1) the assumption of a uniform porous medium is not necessary;
2) the method can give information about relative ground-water age, which can be
46
-------
useful in determining the appropriateness of WHPA delineation;
3) the method helps confirm TOT estimates made by other techniques; .
4 the method does not require detailed measurements of aquifer parameters, although
knowledge of such parameters increases the method's usefulness.
Disadvantages of the residence-time approach are _.,;,,.
na* method requires skill and experience in geochemical and isotopic interpretation,
2) the method is not applicable to all settings, and results are sometimes ambiguous,
3) eeochemical and isotopic analyses can be expensive; " -x^ «-
4) thTmethod may not produce a mappable ZOC, but it .can help confirm a ZOC and
TOTs delineated by some other method.
1: Junction City «re.~Ground- water samples
pezor in and around the Junction City well field between
(fig. 17). The samples were analyzed for all major cations and amons and for *H and O.
Analytical results are presented in Appendix A (tables A4 and A5).
Using the qualitative tritium interpretations in table 3, ground water currently produced
by the village well contains about 21 TU and is therefore less than 35 years old. This is
relatively young ground water, which indicates that ground-water flow paths from the
recharge area to the village well are relatively short.
Most other bedrock wells and piezometers near the village well also produced water less
than 35 years old, with three exceptions. Shallow piezometer JC5, bedrock well JC12 and
deep piezometer JC23A contain, respectively, 5.4. 3.4, and 7.7 TU (see inset map 5g. 17;
table A5, Appendix A), and ground water at these locations iis ^*f " *c_°*^rSv
Piezometer JG5 is finished in the thickest section (greater than 50 ft) of unlithified silty clay
present in the well field. Ground water at JC5 has taken « least 20 years jo^e >ss than
50 ft downward, implying that the bedrock aquifer probably receives very little recharge in
i?e vic^r/oTlCS. Well JC12 is several thousand feet west of the viUagrweU and is outside
Ae cone of depression. The tritium data confirm that JC12 is outside the ZOCof the village
well (fig. 17). Finally, piezometer JC23A is finished deep in the poorly fractured^section of
the bedrock aquifer. Even though the village well (JC9) extends to an equal depth as JC23A
S 17) the ^screpancy in tritium values between JC9 and JG23A indicates that die deep
poorly fractured pan of the aquifer probably contributes little or no water to the village weU.
This finding implies that the ZOC for the well may be small, and that the well receives little
or no water from a regional flow system.
The "O values for the Junction City area range from -9.6 to -11.6 per mil (table A5,
Appendix A), which is about the range of variation expected for "O in recent precipitanon^in
cS Wisconsin. The oxygen data are thus consistent with the tntium data (ground water
£s ihan 35 years old), and the small variations between wells in the study area are probably
not significant.
47
-------
1200
West A
12
1100
1000
1000
,27.3 Tritium content
<35 Estimated age (yr)
Approximate ground-water Bow path
Figure 17. ZOC verification using tritium (iata at Junction City, Wisconsin.
48
-------
Results of chemical analyses (table A4, Appendix A) lead to these conclusions about
ground-water flow to the village well. ., .
1) Because the chemistry of the village well (JC9, 325 ft deep) is very similar to the
chemistry of surrounding test wells (JC10, JCll, and JC24, about 100 ft deep) and is
not similar to the chemistry of piezometer JC23A (300 ft deep), it seems likely that
most of the water produced by the village well moves into the well from the upper,
highly fractured bedrock. The lower part.of the bedrock supplies little water to the
village well.
2) The village well chemistry is different from the chemistry of a surface-water sample
collected from the swale west of the village well (SW01). This suggests that there is
ho direct conduit carrying surface water to the village well through the fracture
system. . . . . .
3) The village well chemistry is clearly different from water chemistry in the vicinity of
piezometer JC5, screened at the base of a thick layer of unUthified materials.
Apparently, little recharge enters the aquifer from areas of thick, clayey surficial
materials. .
4) Water qualityin the upper, fractured bedrock is relatively uniform, supporting the
concept of a diffuse-flow system (porous medium) rather than conduit-flow system
for this aquifer. ,
Taken together, the isotope and chemical data suggest that water produced by the village
well is well-mixed, flows mainly through the upper fractured zone, and has a residence time
of less than 35 years. These results suggest that wellhead protection measures based on the
porous-media assumption will be effective at Junction City.
Example 2: Sevastopol site.--Ground-water samples were collected from 16 wells and
piezometers during November 1988, and analyzed for major ions and for 3H and I8O.
Analytical results are presented in Appendix B (tables B4 and B5).
As with the Junction City example, the residence time of ground water in the fractured
dolomite of the Sevastopol site is generally less than 35 years based on qualitative tritium
interpretations. The "O data are consistent with this age interpretation (table B5, Appendix
B). Isotopically "young" water is expected in Door County because of the high degree of
fracturing of the dolomite and the thin soil cover, which allow rapid recharge. The similarity
of isotope results from all piezometers is consistent with the assumption of porous-medu
behavior for the dolomite aquifer.
Major-ion ground-water chemistry is uniform with depth and position at the Sevastopol
site. The homogeneity of bedrock at the Sevastopol site probably masks any subtle chemical
changes associated with varying source areas. The lack of major variations in major-ion
concentrations, electrical conductivity, mineral-saturation indices, and other parameters
supports the assumption that the dolomite aquifer behave* y a porous medium (diffuse-flow
system) at the WHPA scale.
49
-------
Numerical Flow/Transport Models
Description.-A. WHPA can be delineated using computer models thai approximate
ground-water and/or solute transport equations numerically. Such delineation is usually a
two-step process: simulating a flow system followed by calculation of contaminant flow
paths within that system. Where the hydrogeologic setting is complex, roodels can be
particularly useful because they allow simulation of a wide variety of coalitions and
ground-water flow boundaries. Modeling of a flow system involves discretization of either a
two- or three-dimensional problem domain into nodes. Such discretization can account for
spatial variability of aquifer parameters, thus enabling the inclusion of aquifer heterogeneity
and anisotropy in the model simulation. Most ground-water flow models also allow for
temporal variation,bf.many parameters. -The flexibility of computer models allows for
variation of recharge rates, pumping rates, thickness of aquifer layers, storativity, and
hydraulic conductivity. Models such as the widely-used U.S. Geological Survey (USGS)
Modular Three-Dimensional Finite-Difference Ground-Water Flow Model (McDonald and
Harbaugh, 1988) are able to simulate pumping wells, rivers, drains, recharge, and
evapotranspiration.
Most numerical models in the public domain at the present time (19B9) simulate
ground-water flow using the governing equations of porous-media flow. Such models are
adequate for wellhead protection studies in fractured-rock aquifers if the aquifer behaves as a
porous medium at the scale of the study. Although models that can simulate flow through
fracture networks using the governing equations for fracture flow exist, few, if any, such
models are currently in die public domain. In addition, the enormous data requirements for
such models limit their use to very sophisticated studies.
Once a flow-system model is calibrated so that the simulated head distribution
approximates the field heads, transport computer programs can simulate the probable flow
paths that contaminants may follow and the TOTs of these contaminants;. A ZOC can be
delineated by starting these paths at various locations within the flow system and noting
which flow paths terminate at the pumping well. Model-produced TOTs along these travel
paths can further refine the ZOC using the TOT criterion.
Model simulations are only as reliable as their input parameter values. The cost and
technical expertise needed for adequate data collection can be quite high md such collection
can require substantial field investigations. Input parameters requiring some degree of field
measurement can include aquifer transmissivity, porosity, arid the thickness of various layers.
Characterization of these layers requires a high degree of geological background and skill.
Building, running, and calibrating the model are also complex tasks requiring skilled
personnel and a large time investment. In general, if the modeled system is an accurate
portrayal of the real system, the resulting ZOC represents the most accurate delineation
possible. Changes in the ZOC delineation resulting from natural or man-made effects can
also be predicted. The accuracy and adaptability of the model to so many types of
hydrogeologic settings make this method desirable, but it is usually the most costly method to
50
-------
imolement in tarns of time, money, and personnel. Numerical modeling is probably be'st
£pS » utTsituauons where the accuracy required or the now system complex
warrant such a costly approach.
Attvaniaves of numerical simulation are . . .
$rS£ available numeric models am Simula* aquifers in thr«°»I»'1
can include most of the inhomogeneity. anisotropy, and transient behavior observed
'in the field; , . .
2) if properly discretized, numerical models can simulate discrete fraqture zones;
3 because numerical models give an integrated solution over the model domain,
ground-water flow paths and travel times can be determined withmuch greater
precision than with other methods;
4) adequate numerical codes are widely available.
Disadvantages of numerical simulation are ^^ .
1) most practical models require a porous-medium assumption at some scale,
2) models require significantamounts of dateforproper calibration, verification, and
prediction; ' , u i
3) modeling is often very expensive and time-consuming because it requires substantial
amounts of data and expertise.
Example 1: Junction City ste-Modeling of the Junction City area was performed using
the USGS modular model (McDonald and Harbaugh, 1988). Field work performed for model
parameter estimation consisted of borehole drilling and logging, piezometer installation,
piezometer and deep well slug tests, a pumping test, surveying, water-level measurements.
and characterizing bedrock exposures for lithology and fracture-trace orientation (Appendix
^A) Using the information from the field investigations, the flow system was divided into
layers, assigned hydrologic boundaries, spatially discretized, and assigned appropriate aquifer
parameter values. Details of the modeling are given in Appendix A.
Once the head distribution was adequately simulated, PATH3D, a three-dimensional
particle-tracking program (Zheng and others, in press), was used to delineate a ZOC for the
village well PATH3D uses the model-produced head distribution to predict the path of
hypothetical particles of water within a flow system; it deals only with the process of
advection and ignores the effects of contaminant dispersion and diffusion. By using the
steady-state model simulation, particles can be tracked backward from any specified point to
the particle origin in the ground-water system. The easiest way to delineate a ZOC is to
place the particles in a small circle surrounding the pumping welland n^their paths ^
backwards to either the water table or a ground-water flow divide. PATH3D was also used to
find the downgradient extent of the ZOC. Trid-and-error placement of particles at the water
table near the village well was used to establish the downgradient null point (the divide
between the water escaping the well's influence and the water moving backward toward the
well) Model results indicated that particles beginning 310 or more feet downgradient
(pre-pumping) of the village well continued to move away from the well. -Particles beginning
51
-------
less than 305 ft downgradient of the village well were within the ZOC of the well. The final
ZOC and approximate TOTs predicted by numerical modeling are shown to. figure 18.
Exdinple 2: Sevastopol sire.--Numerical modeling of the Sevastopol site was performed
similarly to the Junction City modeling, requiring a similar intensity of field work (Appendix
B). The modeling and calibration of the Sevastopol site was more complex, however, due to
the existence of the shallow and deep flow systems, greater aquifer anisotuxupy and
heterogeneity, and greater head fluctuations associated with seasonal or recharge events.
PATH3D was again used to delineate a ZOC. Hypothetical particles were placed around
the pumping well and their paths were tracked backward until they emergiBd at the water
table. The particle paths delineate a thin elliptical. ZOC for the well shown in figure 19.
Calculated TOTs are also shown. The model predicts small TOTs relative to estimates from
other methods due to the incorporation in the model of a thin but highly conductive fracture
zone about 180 ft below the land surface. The ability of the model to simulate such a
fracture zone makes it superior to previously discussed methods that assiuioe a homogeneous
aquifer.
WHPA Delineation Methods for Fractured Rocks
That Do Not Behave As Porous Media
Fractured-rock aquifers that do not behave as porous-media aquifers generally fall into
two categories. The first category includes aquifers with numerous interconnected fractures.
These aquifers contain discrete zones or regions of intense fracturing, large fracture apertures,
or fractures widened by solution that are significantly more permeable than the surrounding
fractured rock. The second category of fractured-rock aquifers that do noit behave as
porous-media includes rocks with very sparse and poorly connected fractwres in a
low-permeability matrix. Such situations are probably most common in structurally
homogeneous igneous and metamorphic rocks such as granite and quamite. In such aquifers,
obtaining adequate yield for production wells can be difficult and usually involves completing
the well to intersect one or two major water-bearing fractures that act as conduits and storage
reservoirs for ground water.
In both cases, ZOCs based on well hydraulics or the uniform flow equation will be
incorrect because they ignore the system heterogeneity. However, porous-media-based
numerical models may be able to simulate some of these systems by treating the permeable
fractured zones as permeable model layers or a series of nodes in a less-permeable matrix.
The model at the Sevastopol test site (Appendix B) demonstrated this method by simultting a
high permeability horizontal fracture zone in the dolomite aquifer. The model showed that
the presence of this zone had a profound effect on ground-water flow patllis and travel times
in the WHPA.
52
-------
Junction City /"' .
SCALE 1:24 000
FEET 600 0600 10002000
. Ground-water divide
* Village well (JC-9)
gT] Zone of contribution
TOT Time of travel
11
Figure 18. ZOC predicted by numerical modeling for a well in crystalline rocks.
53
-------
7*.
26
Ground-water divide
V Test well (MW-1)
| | Zone of contribution
TOT Tirrie of travel
SCALE 1:24 000
FEET SCO 0 500 11000 2000
CJ'ltviUt. .,
35
LAKC
r
Figure 19. ZOC predicted by numerical modeling for a well in dolomite.
54
-------
Vulnerability mapping and hydrogeologic mapping are also possible methods for
fracrured-rock aquifers that dp not behave as porous media. Vulnerability mapping may be
used to identify areas particularly vulnerable to ground-water contamination, and these areas
could form the basis for the delineation of a WHPA using the arbitrary fixed radius method
or the simplified variable shapes method (U.S. EPA, 1987). Hydrogeologic mapping (U.S.
EPA, 1987) uses geologic contacts, structural features, and water-table maps to determine
ground-water basin boundaries. In some cases, the ground-water basin may function as the
ZOC for a given well; in cases where the basin is small enough, the entire basin could be
delineated as the WHPA.
The residence-time approach is useful in settings where the porous-media assumption
does not hold, this approach can,be used to establish the age and geochemical origin of
water produced by the well 10 be protected. -The residence-time approach alone cannot be
used to determine a WHPA and it should be used in combination with the hydrogeologic
mapping method or the vulnerability mapping method.
' f " '.
WHPA Comparative Analysis
Cost Analysis
Exact prediction of the costs inherent in each of the methods is difficult because the
amount of field work required for each method depends on how much information is already
available, the complexity of the problem area, and the degree of accuracy desired by the
wellhead protection program. For example, the work required for aquifer parameter ' >
estimation can range from the least costly and least accurate method of simply citing the
average values of parameters (such as hydraulic conductivity or porosity) found in literature
to the most costly and most accurate method of performing a pumping test. The latter
requires extensive field work and many hours of technical data analysis. The parameters
necessary for each ZOC delineation method are included in table 4. Table 5 summarizes
some of the work, time, skill, and approximate cost requirements of performing the individual
tasks for each of the methods.
The cost estimates are primarily based on an hourly rate that represents the actual salary
to an individual at a particular skill level. It does not include general overhead, benefits,
taxes, profit, or the amortization of equipment; If a consultant were employed to perform the
same tasks, s/he would usually charge three times the hourly costs that are listed in table 5.
Six ground-water consultants or consulting firms in Wisconsin were contacted regarding
typical costs for this type of work. They would charge $40 to $100/hr for tasks with skill
levels IV, V, and VI. The cost of the flow-system mapping method with calculations (but
without drilling or monitoring-well installation) probably would be between $10,000 and
$20,000. " '
55
-------
Table 4. Data requirements for WHPA delineation methods for fractured crystalline and
dolomitic rocks. .
Method
Data Requirements
K V Q n
Hydrologic
R Boundaries
Aquifer
, Geometries
Vulnerability
Mapping
Flow-System Mapping
Flow-System Mapping
with TOT
Flow-System Mapping
with Uniform Row
Equation
Residence Time
Numerical
Flow/Transport
Model
Geologic, soils, and water-table maps
Field mapping of surficial features
X XX
X X X X X
Water sampling and analyses
X
X
X
X
X X X X
X X X
Explanation:
K - hydraulic conductivity
V - vertical leakance
Q - well pumping rate
n - porosity
i - hydraulic gradient
b - aquifer thickness
S - storativity
R - recharge rate ;
Assessment, Comparison, and Selection of Methods
In deciding which of the ZOC delineation methods is most appropriate for a given
wellhead protection program, many factors need to be considered. These include relative
accuracy, level of technical expertise needed, and costs. Although the best method might
differ from one setting to another, a comparison of the results from the Junction City and
56
-------
Table 5. Estimated work, time, skill, and cost requirements for selected WHPA delineation methods.
Method
Vulnerability
Mapping
t
^
Row-System
Mapping
Row-System
Mapping with
TOT calculations
Work Requirements
Field
- Location of measurable wells
- Depth to (round water
measurements
- Location of bedrock
outcrops
Office
- Interpretation of soil
surveys
- Constructing depth to
ground water map
- Constructing depth to
bedrock map
- Map compilation
Field
- Location of measurable wells
- Surveying well elevations
- Piezometer installations
- Water-level measurements
Office
- Collection and plotting
of well logs
- Hand contouring of maps
- Computer contouring
- ZOC delineation
All work listed under
flow-system mapping method
plus:
Approx. Time
Requirement1
2 days
several days
several days
several days .
2 days
2 days
2 days
t ,-
2 days
3 days
3 days
1 1 day
2-3 days
' 2 days
2 days
4hrs
Skill
Level2
I
rv
m.iv
soils
expert
m
m
m
i
IV
iv. m
iv
m
m
v
IV
Approx.
Costs ($)3
300
600
J500
2400.
500
400
400
400
* . « wjut
1700
4100
100
400 ,
2000
120
2620
240
160
800
22
1260
3880
wan
- ' '.' '
Field
- Aquifer paramecer esnmation
Office
- Interpretation of
hydraulic gradients
2-3 days
1 hr
iv. m
1000
20
57
-------
Method
Work Requirements
Approx. Time
Requirement1
Skill
Level2
Approx.
Costs ($)3
Flow-System
Mapping with
Uniform Flow
Equation
Residence-Time
Approach
- Hydraulic conductivity
estimates (from
literature, specific
capacity estimates.
analysis of field data)
. Application of {round-
water velocity equation
to establish TOTs ,
All work listed under
flow-system mapping method
plus:
Field
- Aquifer parameter estimation
Office
- Interpretation of
hydraulic gradients
- Hydraulic conductivity
estimates
- Application of uniform
flow equations
Field
- Water sampling
Laboratory
- Sample analyses
Office
- Data interpretation
2-3 days
5hrs
2-3 days
Ihr
2-3 days
4hr
2 days
VI
IV. DI
2 days
IV. EH
n
VI
n
IV
cnem lalj
VI
1500
50
1570
6450
3880
1000
20
1500
60
1580
6460
30A
3000
1200
4500
Numerical
Modeling
Might include:
- Location of measurable wells
- Surveying well elevations
- Water level measurements
- Piezometer installation
- Borehole drilling and
logging ,
- Geophysical logging
- Video logging
- Slug tests
- Aquifer pumping test
- Location of bedrock outcrops
2 days
3 days
several days
3 days
1 week.
several days
Iday
1-2 weeks
2 days
2 days
I
rv
IV
iv. m
rv. m
rv. m
rv
IV
TV. ffl
iv. m
100
400
500
2000
15000
5000
1000
1500
2000
500
28000
58
-------
Method
Work Requirements
Approx. Time
Requirement1
Skill
Level2
Approx.
Costs ($)3
Office
- Water-table mapping
- Bedrock surface
elevation mapping
- Analysis of field data:
-Shig tests
-Geophysical logs
-Pumping test
Borehole drilling.,.
- Initial model construction:
-Parameter selection
-Boundaries
-Spatial discretization
(horizontal and vertical)
-Dei* management
- Model calibration
- Transient simulations
- Application of particle
tracking program to
delineate ZOC
- Prepartibn of graphic
output
1 week
2 days
2-3 weeks
3 weeks
2 weeks
1 week
2 weeks
1 week
ffl
m
VI
m
VI
m
VI
VI
VI
VI
V
500
200
5000
9000
6000
3000
6000
1 Time requirements depend on the scale and complexity of the problem.
1 Skill Level:
I . Little or no technical expertise required
fl - Some knowledge of hydrogeology helpful
IE - Training in hydrogeology and/or mapping required
IV - Training in hydrogeologic field methods required
V - Computer expertise required '
VI - Requires combination Of computer and hydrogeologic expertise
* Costs dp not include overhead, equipment, travel expense, and other administraldve charges.
Hourly
Rate (S)
5
7.50
10
15
50
75
59
-------
Sevastopol sites is illustrative of some of the advantages and disadvantages of each method
and of which methods are the most likely candidates for wellhead protection programs in
fractured-rock settings. ,
The various WHPA delineation methods applied in this study resulted in areas of greatly
varying size and shape and in TOTs of greatly varying lengths. For the pur][X>se of
comparison, it is assumed that the ZOCs delineated by numerical modeling, as shown m
figures 18 and 19, are the most accurate because they incorporated the greatest amount of
site-specific field data. The ZOCs delineated with the flow-system mapping, the flow-system
mapping with the uniform flow equation, and the numerical modeling methods at the Junction
City and Sevastopol sites are compared in figures 20 and. 21, respectively.
With the exceptions of vulnerability mapping.and^the residence.time approach, the
WHPA delineation methods tested require the assumption that the fractured-rock aquifer
behaves similarly to a porous medium at the scale of the wellhead protection effort.
However, numerical models, when carefully discretized, may be able to simulate some
high-hydraulic conductivity fracture zones. Although the porous-media assumption can
introduce some error, it will be reasonable in many, if not most, fractured-rock settings.
Assuming that porous-media models can accurately simulate fractured-irock aquifers in
WHPA studies, wellhead protection analyses in fractured rocks differ from iilnalyses in porous
media in several important respects. First, fractured-rock aquifers tend to be highly
anisotropic. Second, fractured-rock aquifers tend to be heterogeneous in three dimensions.
Third, production wells in fractured-rock .aquifers are often constructed with long open-hole
intervals. Pumping of such wells draws water from long vertical sections of the aquifer and
tends to result in larger ZOCs. For all three reasons, WHPA determinations in fractured-rock
aquifers usually require a more sophisticated analysis than in porous media. Numerical
modeling is currently the only practical means of carrying out this type of sophisticated
analysis, incorporating anisotropy, heterogeneity, and three-dimensional flow.- More
sophisticated methods of analysis, such as discrete fracture models, fractal analysis, and
stochastic models, offer promise but are currently in the research stage and are not viable
options for community wellhead protection projects at this time.
Vulnerability im^p/iif .-Vulnerability mapping can help delineate land surface areas most
susceptible to contamination including open fractures, exposed or shallow tjedrock, and
permeable soils. In addition, if fracture conduits can be identified in. the subsurface (for
example, by using geophysical logs), well casing should be placed so that it isolates the well
from major conduits. The results of vulnerability mapping can be used for WHPA delineation
using either the arbitrary fixed radius method or simplified variable shapes methods (fig. 9).
In areas where the aquifer does not behave as a porous medium, vulnerability mapping offers
a relatively cost-effective basis for delineating a WHPA. If geologic, soils, and water-table _
maps are available for the site in question; the only field work necessary is field mapping of
surficial fractures. :
60
-------
£.1
-
E'* IX E
SCALE 1:24 000
FEET 500 0 500 1000 2000
Ground-water divide
* Village well (JC-9)
----- Flow-system mapping
. Uniform flow equation
. Numerical modeling
>c*,-5i
""" F~ " ;:
11
Figure 20. ZOC comparative analysis for a well in crystalline rocks.
61
-------
25
Ground-water divide
* Test well (MW-1)
Flow-system mapping
Uniform flow equation
Numerical modeling
SCALE 1:24 000
FEET 500 0 500 1000 2000
r
35
Figure 21. ZOC comparative analysis for a well in dolomite.
62
-------
Flow-system mapping and flow-system mapping combined with TOT criterion.-** bob
sites flow-system mapping produced die largest ZOC estimates and included themajonty
ofthe ZQC area delineated by numerical modeling. General similarity between the
flow-system mapping and the computer-modeled ZOCs is expected because the . u
hydraulic-head distributions generated by the computer simulations are calibrated against the
water-table maps. Differences between the ZOCs occur in part due to the consideration or
heterogeneity and anisotropy by the modeling method but also due to less than perfect model
calibration. How-system mapping therefore might offer the most protection for the least cost.
In general, the amount of field work involved in flow-system mapping depends on how many
water-table measuring points exist in the well vicinity. If no additional wells are needed, the
only field work required is measuring water levels in existing wells and surveying the
elevations of those; wells. Using data from existing well logs saves substantial amounts of
time and should suffice outside the area immediately surrounding the production well.
The flow-system mapping approach is especially useful.when the ground-water flow
boundaries are close to the well. In the case of the Junction City well, the ground-water
divide is only about 1 mile away. At the Sevastopol site, the shallow system ground-water
divide is only about 1000 ft from the well; the deep system ground-water divide is about 1.5
miles from the well. In cases where ground-water flow boundaries are farther from the well,
the resulting ZOC can be extremely large and the TOT criterion can be used to limit the
WHPA to that portion of the ZOC that will contribute water to the well within a specified
1 length of time.
Using various estimates of hydraulic conductivity at the Junction City site resulted in
different TOT estimates. The most accurate method for measuring hydraulic conductivity (the
pumping test) resulted in the shortest TOT estimates (fig. 10). These TOT estimates suggest
that the whole ZOC be used as the WHPA unless dealing with contaminants that degrade to
non-toxic levels in less than 5 years. It should be noted, however, that the ZOC and TOT
estimates from this estimate differ from those calculated by the numerical modeling. The
computer-delineated ZOC was smaller and extended farther east than the water-table map
ZOC (fig. 20). The difference in TOTs occurred in part because the computer model
simulates flow paths in three dimensions and accounts for flow paths through deeper zones of
. varying hydraulic conductivity. .
At the Sevastopol site, the flow-system mapping approach was more complicated due to
the existence of shallow and deep ground-water systems, but the results were more compatible
with the results of numerical modeling than at the Junction City site." At the Sevastopol site,
the flow-system mapping ZOC is slightly larger than the computer-delineated ZOC (fig. 21)
and only a small area ofthe computer-delineated ZOC is not included in the flow-system
mapping ZOC. The TOT estimates based on the potentiometric-surface map (fig. 12) were_
much longer (100 years compared to 1 year) than those predicted by numerical modeling (fig.
19). The modeling TOTs were short because the model accounted for a very thin but very
conductive zone about 180 ft below the land surface. ,
63
-------
Disadvantages of using the flow-system mapping method in fractured-rock settings are
that the method does not account for three-dimensional flow and for the heterogeneity and
anisotropy inherent in many fractured-aquifer systems. At the Junction City si«5, the
hydraulic conductivity decreases dramatically with depth as the number of water-transmitting
fractures diminishes. At the Sevastopol site, a highly fractured conductive zone dramatically
reduces travel times.
Flow-system mapping with analytical equations.-Applying the uniform flow equations
after a water-table map is constructed requires little additional time or cost and provides a
good analytical basis for ZOC delineation. It is recommended that the uniform flow
equations be added to any ZOC delineation made from flow-system mapping. The two
methods used in tandem can serve as checks on one another's accuracy. The uniform flow
equation ZOCs, calculated.with the hydraulic^conductivities.measured during the pumping
tests, were the smallest of the three ZOCs compared in figures 20 and 21. At the Junction
City site, the size of the cone of depression of the village well (fig. 10) indicates that the
uniform flow equation ZOC (fig. 20) is probably too small to adequately protect the village
water supply. At the Sevastopol site, the ZOC delineated with the uniform flow equation
corresponds closely with the flow-system mapping and the numerical modeling ZOCs, but it
is still the smallest (fig. 21). "
Residence-time approach.~Thc residence-time approach by itself does not delineate a
ZOC and is therefore only useful when used in combination with one of the other approaches.
It serves as a useful check on the TOTs estimated by the other methods because TOTs based
on the residence-time approach do not depend on measurements of aquifer parameters.
Because it requires no assumptions about the aquifer, it is useful in providing TOTs for the
flow-system mapping method and can be used in conjunction with this method for WHPA
delineation. It is also useful in verifying modeling results and in model calibration.
The residence-time approach will be helpful in settings where the fractured-rock aquifer
does not behave as a porous medium. In some areas, the pattern of fracture conduits is so
complex that delineation of individual conduits will not be possible, and ground-water flow
paths will be unpredictable. In such settings, isotopic and geochemical information can be
useful for delineating ground-water ages, travel times to the well, and recharge areas.
Occasionally, isotopic data will show that water reaching the well to be protected is hundreds
or thousands of years old. In such situations, wellhead protection practices might not be
practical because the ZOC for the well could be extremely large and the effect of wellhead
protection practices would not be seen for hundreds of years. In other areas, the
residence-time approach might indicate particular bedrock types or areas of the landscape
most likely to contribute water to the well.
Numerical modeling.-Numerical modeling requires substantially more time and expertise
than the other methods considered (table 5), but it produces smaller, more accurate ZOCs.
Being able to delineate a smaller area might be very important in some setting:!!. The amount
of effort and resources required might also be justified where there is need for a high degree
64
-------
of accuracy. In areas where the ground-water flow system and hydrblogic boundaries are
complex, numerical models might provide a significant improvement over other *°*^
Although numerical models are based on analytical equations that assume a mnform porous _
medium, they can take into account heterogeneities and anisotropy that are ^^^^
in a fractured-rock aquifer. The magnitude of improvement that numerical modeling bnngs to
ZOC delineation will vary from site to site! Greatest improvemem will occur at more ^
complex sites. The following site characteristics may justify the added expense and time
required for numerical modeling:
1) significant discrete fractures or fracture zones,
2) significant anisotropy,
3) significant spatial variations in various hydrogeologic parameters (T, K, recnarge,
4) significant vertical movement of water and/or significant variation in total hydraulic
; head with depth,
5) significant changes in water levels seasonally or through time.
Numerical modeling is potentially the most accurate WHPA delineation method even in
settings that do not behave as porous media. Although most widely available numerical codes
are based on porous-media physics, these codes can often simulate flow in fracture Conduits
and can capture the spatial variability and anisotropy often present in fractured-rock aquifers.
For example, the numerical model of the Sevastopol site (Appendix B) adequately simulates a
highly conductive horizontal fracture zone and shows that this zone has a significant influence
on ground-water movement toward the well to be protected.
At the Junction City site, the numerical model produced a ZOC that is smaller.than the
flow-system mapping ZOC (fig. 20). Increased precision of the numerical model method, due
to inclusion of heterogeneities and anisotropy, probably accounts for the difference, although .
the lack of perfect model calibration may also play a role. Part of the computer-delineated
ZOC extends beyond the eastern boundary of the flow-system mapping ZOC. Ttusare a
might be the result of a less-than perfect model calibration at the site. Using the ZOC
delineation as the sole criterion, the logical conclusion would be that the modeling produced
only marginal improvement considering the time requirements and complexity of the task.
However, the TOT lines resulting from the particle-tracking program (fig. 18) were ' - .
substantially different from those resulting from the flow-system mapping approach (fig. 1U).
Whereas modeling predicted that water moving from the ground-water divide to the village
well might 'take 50 to 100 years, the mapping approach predicted a TOT of less than 13
years. The order-of-magnitude difference occurred mostly because the model considers a
three-dimensional path through a nonhomogenepus system; the other methods assume a
homogeneous, two-dimensional system.
The hydrogeologic setting at the Sevastopol site was somewhat more complex than at the
Junction City site. The existence of a shallow and a deep ground-water system and the
existence of a thin, highly conductive fractured layer make ' :s site a mucjrmore jkely
candidate for numerical modeling. The TOTs predicted with numerical modeling (fig. 19)
65
-------
were quite different from those calculated using the potentiometric-surface map (fig. 12). The
numerical modeling ZOC, however, was only somewhat smaller than the ZOC predicted with
the flow-system mapping method (fig. 21).
In some cases, numerical models may produce ZOCs significantly different from those
produced by flow-system mapping. These differences are usually due to details of anisotropy,
heterogeneity, and a three-dimensional system that are included in the model'but not included
in the simpler flow-system mapping method. In such cases, the model-produced ZOC is more
reliable. In general, numerical modeling will be required when the aquifer is extremely
anisotropic or heterogeneous, or when the aquifer exhibits significant changes in hydraulic
head over short periods of time.
Conclusions !
At the Junction City site and the Sevastopol site,, methods less sophisticated than ,
flow-system mapping clearly did not encompass enough of the site characteristics to result in
an accurate ZOC determination. The flow-system mapping approach was sufficient for
conservative (more protective) estimates of the ZOC. With the adequate field data obtained
from the piezometers and existing wells, numerical modeling was not warranted for ZOC
delineation but contributed more accurate TOT estimates. At both sites, the TOT estimates
were more sensitive to the various delineation methods than was the ZOC delineation itself.
When possible, the entire ZOC should be used as a WHPA. However, in some situations the
resulting ZOC will be extremely large and in these cases the TOT criterion may be used to
limit the size of the WHPA if implementation concerns arise.
At each of the two test sites in Wisconsin, the fractured-rock aquifers behaved as porous
media at the wellhead protection scale. For a fractured-rock aquifer that does not act as a
porous medium, WHPA delineation may be accomplished by a combination of Mislnerability
mapping, hydrogeologic mapping, the residence-time approach, tracer tests, or numerical
modeling. Flow in such aquifers can occur mainly through discrete fracture conduits. It is
particularly important to identify areas where such conduits intersect either the land surface or
the production well because such conduits can offer direct and rapid pathways for
contamination to travel from the land surface to the well.
66
-------
Chapter IV
IMPLEMENTATION POSSIBILITIES
Applicability and Usefulness of Selected Methods
The studv of two fractured-rock settings in Wisconsin demonstrated that the standard ,
I tests, configuration of the water table,
, and no dramatic fluctuations of water levels
or water chemistry.
methods The flow-system mapping method offered the most accuracy for the leas t cost In
SneS the cost of this method will depend on the number of existing wells available for
in- the » of study. If additional wells need to be drilled.
we - .
be ST Applying the uniform flow equations after a water-table map is ;
time or cost and improves the results gained by mapping
alone.
The residence-time approach using isotope and '^
.information on recharge areas and ground-water flow paths.
ZOC for a well but it can verify ZOCs and TOTs determined -by other
relatively good agreement between TOT estimates using analytical methods «jdTOT_
estimates using tnTresidence-time approach shows that the porous-media methods were
adequate for WHPA delineation at the two test sites in Wisconsin.
Numerical modeling requires substantially more time and
methods considered, but it offers the highest degree of accuracy M\ both
67
-------
The evaluation of WHPA delineation methods addressed the question of the degree of
technical sophistication necessary for WHPA delineation in fractured rocks. The least
sophisticated and least expensive, fixed radii methods are clearly inadequate for accurately
delineating WHPAs and would generally be considered primarily as a first step in a wellhead
protection program. On the other hand, the most sophisticated and also the most expensive
method, the computer-based numerical model, cannot eliminate all uncertainties, and the
improved accuracy in WHPA boundaries may be minimal in some settings. In addition,
numerical modeling is generally beyond the means of local governments. Therefore, the best
methods for a wellhead protection program in fractured rocks are fairly simple approaches
producing reasonable results, such as the flow-system mapping method and the residence-time
approach.
The limitation of thii study is that the selected methods were tested only in one type of
fractured-rock environment-aquifers with closely spaced fractures, me geological condition
found at both test sites. The fractured crystalline aquifer consisted of 35 ft of extensively
weathered rock underlain by 90 ft of competent rock with vertical fractures spaced every 0.5
to 2 ft. In the fractured dolomite, horizontal fractures occurred on the order of every 1 to 10
ft and prominent vertical fractures were spaced approximately 10 to 20 ft apart. Under these
conditions, the fractured rocks behaved as uniform, porous media at the scale of field'tests.
In general, the size of most ZOCs would be large enough that the porous-media approach
may be adequate for WHPA delineation in many areas.
For fractured rocks that do not behave as porous media, the following techniques and
methods may be applicable:
1) hydrogcologic mapping; . .
2) mapping of areas vulnerable to contamination, in combination with arbitrary fixed
. radius or a simplified variable shape; .
3) residence-time approach; and
4) numerical flow/transport models. .
Management Strategies for WHPAs
Even though management implications of WHPA delineation were not in the scope of
this study, it seems appropriate to at least mention the main problems communities may face
in delineating WHPAs and give references to publications that can help communities select
appropriate management tools and strategies. These very questions were the ones most often ,
asked at public meetings conducted during the study at the two sites.
The ZOC determined by hydrogeologic analysis can be used as the basis for the areal
delineation of a WHPA, in which potentially contaminating uses and practices sire limited and
protection measures implemented. Ideally, the area within the WHPA would include the
entire ZOC--all of the ground-water flow system that contributes to a well or wellfield.
However, no matter how important hydrogeologic factors are in delineating WHPAs, other
68
-------
non-hydrogeologic concerns can influcnce-and in some cases detennine-the establishment
and ultimate configuration of the final WHPA. In reality, WHPA boundaries may have to
relate to political and administrative boundaries and physical factors, as well as hydrogeologic
factors. This means "squaring up" the WHPA boundaries using a large-scale topographic map
or aerial photographs so that the WHPA follows boundaries that can be easily traced on the
ground (roads, water bodies, fence lines, and similar features). Political boundaries were
considered by local officials at the public meetings as a major threat for implementing
WHPAs because difficulties may arise if the WHPA crosses two or more governmental units.
the establishment of WHPAs generally will be a compromise between the desirable and
the feasibte-a compromise between the socioeconomic and public health interest WHPA
regulations could have adverse economic effects on a community if an undue amount of land
were placed into a wellhead protection district On the other hand, if the delineated area is
too small, under-protection may occur. This dilemma is difficult to solve, and requires
patient negotiations between the opponents. Even in Europe, where wellhead protection has
been practiced for more than 15 years, this problem remains to be solved European countries
use this general "rule" for determining the extent of wellhead protection districts: "on
economic and planning grounds the protection zones must be as small as possible, while on
public health grounds they should be as large as possible" (Headworth, 1986).
Once a community decides to implement a wellhead protection program, they must
decide which management techniques will be effective. Techniques that can be used in local
programs can be categorized as regulatory and nonregulatory, although in practice, most
programs are a mix of these (Bom and others, 1987).
Regulatory approaches involve placing a system of legal constraints on land uses or on
particular activities that have a potential to contaminate the ground water. The delineation of
a WHPA does not take place in a vacuum. Every state has a number of regulatory and
management programs intended to control potential contaminating activities and sources of
contamination; these can be used for implementing protection measures within the WHPA.
Therefore, wellhead protection programs always should be considered in the context of
existing state or local ground-water management programs.
In addition to regulatory tools, there are numerous nonregulatory tools that can
complement regulations and'government efforts to curb ground-water contamination.
Nonregulatory approaches include activities such as public education, voluntary best
management practices (BMPs), governmental coordination, inspection and training programs.
emergency spill response plans, and monitoring to identify water-quality problems.
Some of the regulatory and nonregulatpry tools are summarized in table 6. More details
on the listed tools and other examples can be found in Born and others (1987), Jackson and
others (1987), U.S. Environmental Protection Agency (1989), Yanggen and Amrhein (1989).
and Zaporozec (1985).
69
-------
Table 6. Potential management tools for wellhead protection
(source; Brrn and others, 1987; U.S. EPA, 1989).
Regulatory
Nonregulatoiy
Zoning Ordinances. Zoning ordinances typically tre
comprehensive land-use requirements designed to direct the
development of an area. Many local governments have
used zoning to restrict or regulate, certain land uses within
wellhead protection areas.
Subdivision Ordinances...Subdivision ordinancesJTC*.,*.-.,
applied to land that is divided into two or more subunits
for sale or development. Local governments use this tool
to protict wellnead areas in which ongoing development is
causing contamination.
Site Plan Review. Site plan reviews are regulations
requiring developers to submit for ipproval plans for
development occurring within a giver, area. This tool
ensures compliance u-iih regulations cr other requirements
made within' a wellhead protection area.
b-iign Standards. Design standards typically are
regulations that apply to the design and construction of
buildings or structures. This tool can be used to ensure
that new buildings or structures placed within a wellhead
protection area are designed so as not to pose a threat to
the water supply. -
Operating Standards. Operating standards are regulations
that apply to ongoing land-use activities to promote safety
or environmental protection. Such standards can minimize
the threat to the wellhead area from ongoing activities such
as the application of agricultural chemicals or the storage
and use of hazardous substances.
Source Prohibitions. Source prohibitions are regulations
that prohibit the presence or use of chemicals or hazardous
activities within a given area. Local governments can use
restrictions on the storage or handling of large quantities of
hazardous materials within a wellhead protection area.
Inspection and Testing. Local governments can use their
statutory home rule power to require more stringent control
of contamination sources within wellhead protection areas
than given in federal or state rules. '
Purchase of Property or Development Rights. The
purchase of property or development rights is a tool used
by tome localities to ensure complcti control of land uses
in or surrounding a wellhead area. This tool may be
preferable if regulatory restrictions oil land use are not
politically feasible and the land puicliiase is affordable.
Public Education^ Public education often consists of
brochures, pamphlets, or seminars designed to present
wellhead area problems and protection efforts to the public
in an understandable fashion. This tix>l promotes the use
of voluntary protection efforts and balds public support, for
a community protection program.
Waste Reduction. Residential hazardous waste
management programs can be designed to reduce the
quantity of household hazardous was le being disposed of
improperly. This program has been rwed in localities,
where municipal landfills potentially threaten ground water
due to improper household waste disposal in the wellhead
area.
Bes' Management Practices. BMPat are vohmtar. actions
thai have a long tradition of being wed, especially in
agriculture. Technical assistance for fanners wishing ic
apply them is available from local Extension and SCS
offices.
Training and Demonstration. The! e p: ..grams can
complement many regulations. For cxamp' raining
underground storage tank inspectors and local emergency
response teams or demonstration of iigricultural BMPs.
Ground-Water Monitoring. Ground-water monitoring
generally consists of sinking a series of lest wells and
developing an ongoing water quality testing program. This
tool provides for monitoring the quality of the
ground-water supply or the movement of a contaminant
phone.
Contingency Planning. Local governments can develop
their own contingency plans for emergency response to
spills and for alternative water supply in case of con.,
tamiiiaikm * >v!e existing supply.
70
-------
Transferability of Results
(by William J. McCabe, U.S. EPA)
1. '. i
Introduction
The WHPA delineation methods tested in Wisconsin can be applied,to similar settings
The ^^A^Swl States. Unconfined crystalline-rock .and carbonate-rock aquifers are
WlA*t*">+, **» _ .^_ ; - .^«.ff_. *!_-. .--«m^vi«*nat^ .
ond it« iwssessions The text and figures 22 and 23 give a general
Tme te^rTques discussed in this document may be applicable^ Because
reconSeS ^qufs.may*!not tetp^opSe fo? all fr^aured-rock aquifers at greater
depth (more than about 300 ft below land surface) (table 7).
Furthermore, the techniques have not been tested in fractured ^dstones, ^n karst, and ,in
cavernous volcanic settings; they probably will not be applicable in such settings.
Acknowledgments
The author thanks the numerous scientists who orally provided information f01 ^s
secticT Any errors, or misinterpretations that appear here, however are those of Author
The scientists contacted during the preparation of this section were irorrrtne L
Survey (USGS) District Offices and Environmental Protection Agency ^A;
Offices The USGS scientists are Charles Avery, Allen Bewsher, Stephen Blumer,
Brown, Robert Buchmiller, Marvin Crist, Dan Davis, Jeffrey Deroche, Patrick
Robert,Faust, Ector Gann, Joseph Gates, Michael Gaydos Roy Glass, Robert
Greemen, James Harrill, John Havens, John Helgesen, Jeff ^Patrick
Hollett, William Horak, Ivan James, Ron Kobel, JamerKrohelski, Larry
MacNish, Angel Martin, John Miller, Joseph Moreland, Kathy Peter, John PwelI Stanley
Robson, Michael Shulters, Jeffrey Stoner, Donald Vaupel, John Vccchioh Richard ^
Whitehead, John Williams, and Allen Zack. The EPA scientists are Kenneth Wenz and Mike
Wireman.
' ' v .- " . . ' --
Mapping Criteria
On the basis of geologic characteristics, the area of the United States and its Passions
was classified into three categories: 1) unconfined, fractured, nonvolcanic aquifer areas, 2)
actured, volcanic aquifer areas; and 3) other. The techniques described in this
applicable in most of category 1 and 2. In category 1, techniques may not be
71
-------
"j Unconlined fractured dolomite, limestone.
_J and crystalline aquifer areas
DO NOT USF THIS MAP OR FNLARGEMENT FOR SITE SPECIFIC PURPOSES
Figure 22. Areas of unconfined fractured rock aquifers in the contiguous united States (compiled by WJ. McCabe, U.S. EPA, 1989).
-------
EXPLANATION
Unconlinod Fisclutsd
SI Crystalline Aqurter
tX> Not Use This Map lot
Sit* Specific Puiposes
FiKure 23. Areas of unconfined fractured-rock aquifers in Alaska, Hawaii, Puerto Rico.U.S. Virgin Islands, and Guam
(after Moody and others, 1985).
-------
Table 7. Common depth ranges of unconfmed fractured-rock aquifers in the United States
and its possessions (data from Moody and others, 1985).
State
Depth to Top
of Aquifer (ft)
State
Depth to Top
of Aquifer (ft)
ALABAMA N/D
ALASKA 50-500
ARIZONA 50-1000
ARKANSAS N/D
CALIFORNIA - volcanics 75-200
- mountains & foothills 50-300
COLORADO - foothills 100-250
CONNECTICUT N/A
DELAWARE 40-100
FLORIDA N/A
GEORGIA 4O-600
GUAM N/A
HAWAII - Hawaii 20-900
-Kauai 100-1100
- Maui 90-500
- Oahu 30-1100
IDAHO 100-3000
ILLINOIS 50-500
INDIANA N/A
IOWA 50-500-
KANSAS N/A
KENTUCKY N/A
LOUISIANA N/A
MAINE 20-800
MARYLAND & DC 3(MOO
MASSACHUSETTS N/A
MICHIGAN 25-200
MINNESOTA -' 120-1300
MISSISSIPPI N/A
MISSOURI N/D
MONTANA 1CJO-300
NEBRASKA N/A
NEVADA 100-1200
NEW HAMPSHIRE N/A
NEW JERSEY 35-800
NEW YORK 10-300
NEW MEXICO N/D
NORTH CAROLINA 70-200
NORTH DAKOTA N/A
OHIO N/A
OKLAHOMA - southeast 50-400
OREGON 100-600
PENNSYLVANIA 75-150
PUERTO RICO N/D
RHODE ISLAND N/A
SOUTH CAROLINA 50-300
SOUTH DAKOTA N/D
TENNESSEE 50-150
TEXAS N/D
UTAH N/D
VERMONT N/A
VIRGINIA 50-300
VIRGIN ISLANDS (all) 100-150
WASHINGTON - north 2:0-200
- south 50-750
WEST VIRGINIA N/D
WISCONSIN 50-180
WYOMING 50-250
Explanation:
N/A SB Not applicable.
N/D SB Not determined.
74
-------
. techniques inay not be applicable in
areas of pseudpkarst volcanics
Essisas KS SSMSSSS
permeability for volcanies in these areas to be considered aquifers ......
Only aquifers of drinking-, stock-, or irrigation-water quality are mapped on
W^ttd of Appalachia, no extensive water-well drilling programs, which would help
better define, aquifer boundaries and quality, have taken place.
Areas Where Techniques of This Document May Be Applicable
For ease of discussion, aquifers of the continental United States are grouped into four
midcontinent section, consisting of the Interior Plains with «^*£££^
and Interior Highlands with uplifts and folded eroded strata; and 4) the ««mporton
United States, consisting of the Rocky and Pacific Mountain systems and Intennontane
Plateaus and Basins (Fenneman, 1946). .
fracturing at shallow or moderate depths. %
Physiographic Region 2. -Unconfined fractured-rock aquifers are present in northeastern
Minneso" noS.ern Wisconsin, western Upper Peninsula *r *M* '
eastern Pennsylvania, nonhern New Jersey, Maryland, Wguua parts o
North and South Carolinas, and northern Georgia and Alabama (fig. 22).
75
-------
LAURENTIAN UPLANDS
-vl
ON
r\
__ | INTERIOR PLAINS
INTERIOR HIGHLANDS
EXPLANATION
PHYSIOGRAPHIC REGION BOUNDARY
Figure 24. Major physiographic regions of the continental United States (based on Fenneman, 1946).
-------
fractured igneous and metamofphic rocks overlain by
fractured-rock aquifers (Moody and others, 1985).
into southwestern Minnesota.
R
STwodc-nountain ranges and in volcanic rocks that characterize this region,
The volcanic lavas of ^^^
of the west are also areas of unconfmed fractured-rock aquifers (Moody and
others, 1985). '
Alaska -More than one half of the state contains unconfined frac tured-rock aqu« ers and,
the upland areas of Alaska (fig. 23).
77
-------
Such areas contain large volumes of ground water and, with respect to the storage and flow of
ground water, are karst-like. The use of the techniques presented in this document is not
appropriate in such areas.
US. possessions.The indurated, fractured volcanic deposits of Puerto Rico's central
core (fig. 23) serve as the island's minor aquifer (Heath, 1984). The volcanic: rock aquifers
arc present oh each of the Virgin Islands (Moody and others, 1985). On the islands of St.
John and St. Thomas, the fractured lava flows serve as principal aquifers (fig. 23). On the
island of St. Croix, unconfined, fractured volcanic rock aquifers occur in the: western and
mid-eastern sections (Moody and others, 1985). The lava flows in northern Guam have low
permeability and no appreciable amounts of water (Moody and others, 1985).
78
-------
REFERENCES CITED
Attig JW Clayton, L., Bradbury, K.R., and Blanchard,M.C. 1987. Coitfirmation of tundra
polygons and shore-ice corpse trenches in central Wisconsin and other applications
of ground-penetrating radar (abs.): Geological Society of America Abstracts with
Programs, v. 19, p. 187.
Bates, R.L. and Jackson, J.A. (eds.). 1980 (2nd ed.). Glossary of geology: American
Institute of Geology, Falls Church, VA, p. 244.
Bebel D J 1989: A ground resistivity technique-for locating fractures in buried Precambrian
basement, central Wisconsin: fo: Proceedings, 35th Annual Institute on Lake
Superior Geology; v. 35, p. 3. ;
Bell, E.A. and Sherrill, M.G. 1974. Water availability in central Wisconsin - an area
of near,surface crystalline rock: U.S. Geological Survey Water-Supply Paper 2022,
'.32P.- .._ - .,:-.":. , ' ;:. ;- "-.;. * ;; v
Blanchard, M.C. 1988. Investigation of the shallow fractured dolomite aquifer in Door
County, Wisconsin: M.S. Thesis - Geology, Univ. of Wisconsin, Madison, 186 p.
: . f ( ," - -
Blanchard, M.C. and Bradbury, K.R. 1987. A comparison of office-derived versus
field-checked water table maps in a sandy unconfmed aquifer Ground Water
Monitoring Review 7/2, p. 74-78. ;
Bom, S.M., Yanggen, D.A., and Zaporozec, A. 19871' .A guide to groundwater quality
planning and management for local governments: Wis. Geological and Natural
History Survey Spec. Rep;. 9, 91 p.
Bouwer, H. 1978. Groundwater hydrology: McGraw-Hill, Inc., New York, p. 132-133.
Bradbury, K.R. 1982. Hydrogeologic relationships between Green Bay of Lake^Michigan
and onshore aquifers in Door County, Wisconsin: Ph.D. Thesis - Geology, Univ. of
Wisconsin, Madison, 287 p. .
Bradbury, K.R. 1985. Major ion and isotope geochemistry of groundwater in clayey till.
northwestern Wisconsin, U.S.A.: In: Proceedings, First Canadian/American^
. Conference on Hydrogeology-Practical Applications of Ground Water peocherrustry
(B. Hitchon and E.I. Wallick, eds.), National Water Well Associanon, Dublin, OH,
p. 2S4-289.
79
-------
Bradbury, K.R., Blanchard, M.C., and Muldoon, M.A. 1988. Hydrogeology imd groundwater
geochemistry in fractured dolomite, northeastern Wisconsin: In: Proceedings,
International Association of Hydrogeologists Symposium on Hydrogeology of
Fractured Rocks, Atlanta, GA, 1988, p. 23-27.
Bradbury, K.R. and Rothschild, E.R. 1985. A computerized technique for estimating the
hydraulic conductivity of aquifers from specific capacity data: Ground Water 23/2,
p. 240-246. ,
Clayton, L. 1986. Pleistocene geology of Portage County, Wisconsin: Wis. Geological and
Natural History Survey Information Circular 56, 19 p.
Cooleyi R.L. 1977. A method of estimating parameters and assessing reliability for models
of steady-state groundwater flow. I. Theory and numerical properties: Water
Resources Research 13/3, p. 603-617. , ,
Cooper, H.H., Jr., Bredehoeft, J.D., and Papadopoulos, I.S. 1967. Response of a .
finite-diameter well to an instantaneous charge of water: Water Resources Research
3, p. 263-269.
Davis, S.N. and DeWiest, R.J.M. 1966. Hydrogeology: John Wiley & Sons,, Inc., New
York, NY, 463 p.
» . . - '
Desaulniers, D.E., Cherry, J.A., and Fritz, P. 1981. Origin, age, and movement of pore water
in argillaceous Quaternary deposits at four sites in southwestern Ontario: Journal of
Hydrology 50, p. 231-237.
Egboka, B.C.E., Cherry, J.A., Farvolden, R.N., and Frind, E.O, 1983. Migration of
contaminants in groundwater at a landfill: a case study. 3. Tritium as an indicator
of dispersion and recharge: Journal of Hydrology 63, p. 51-80.
Emmons, P.J. 1987. An evaluation of the bedrock aquifer system in northeastern Wisconsin:
U.S. Geological Survey Water-Resources Investigations Report 85-4199, 48 p.
Fenneman, N.M. 1946. Physical divisions of the United States: U.S. Geological Survey,
Physiographic Committee, one-page map.
Freeze, R.A. and Cherry, J.A. 1979. Groundwater Prentice-Hall, Inc., Englewood
Cliffs,"NJ, 604 p. ' . '
Gale, J.E. 1982. Assessing the permeability characteristics of fractured rock:
In: Narasimhan, T.N. (ed.), Recent Trends in Hydrogeology, Geological Society of
America Special Paper 189, p. 163-182. " .
80
-------
Grecnberg, J.K. and Brown, B.A. 1986. Bedrock geology of P^geCounty, Wisconsin:
Wis. Geological and Natural History Survey Map 86-4, scale 1:100,UUU.
Headworth HG 1986. Groundwater protection policies m European countries:
In- Proceedings, Seminar on Groundwater Protection, Policy and Management,
Strasbourg, March 20-21, 1986, European Institute for Water, Strasbourg, France,
p. 1-20.
Heath, R.C. 1984. Ground-water regions of the United States: U.S. Geological
Survey-Water Supply Paper 2242, 78 p.
Hendry.MJ, 1988. Do isotopes have,a place in ground-water studies?: Ground Water 26/4,
p. 410-415.
Hickey, J.J. 1984. Field testing the hypothesis of Darcian How through a carbonate aquifer
Ground Water 22/5, p. 544-547.
Hsieh, PA., Neuman, S.P., Stiles, O.K., and Simpson, E.S. 1985. Field determination of the
three-dimensional hydraulic conductivity tensor of amsotropic media. 2. .
Methodology and application to fractured rocks: Water Resources Research 21/11,
p. 1667-1676.
Hutasoit, L.M. and Davidson, D.M., Ir. 1989. Hydrogeology of me Precambrian bedrock.
Wood, Portage, and Marathon Counties, central Wisconsin (abs.): Abstracts 13th
Annual Mtg., Amer. Water Resources Assoc. Wis. Section, March 2 & 3, 1989,
Madison, Univ. of Wis. Water Resources Center, p. 41.
Hvorslev, M.J. 1951. Time-lag and soil permeability in groundwater observations: U.S.
Army Corps of Engineers, Waterways Experiment Station Bulletin No. 36,
Vicksburg, MS, 50 p.
Jackson, G. and others. 1987. Agricultural management practices to minimize ground-water
contamination: Univ. of Wis. Extension Environmental Resources Center, Madison,
WI, 115 p.
Knott, J.F. and Olimpio, J.C. 1986. Estimation of recharge rates to the sand and gravel
aquifer using environmental tritium, Nantucket Island, Massachusetts: U.S.
Geological Survey Water-Supply Paper 2297, 26 p.
Lippelt, I.D. and Hennings, R.G. 1981. Irrigable lands inventory-Phase 1. Groundwater and
related information: Wis. Geological and Natural Hi,;oryp"rvey Miscellaneous
Paper 81-1, 10 maps. ,
81
-------
Long, J.C.S., Rerner, J.S.. Wilson, C.R., and Witherspoon, P.A. 1982. Porous media
equivalents for networks of discontinuous fractures: Water Resources Research
18/3, p. 645-658. .
McDonald, MG. and Harbaugh, A.W. 1988. A modular three-dimensional finite-difference
ground-water flow model: Chapter Al, Book 6. Modeling techniques, Techniques
of Water-Resources Investigations of the U.S. Geological Survey, chapters paginated
separately.
Moody, D.W. and others. 1985. National water summary 1984-Hydrologieal events,
selected water-quality uends, and ground-water resources: U.S. Geological Survey
Water Supply Paper 2275, 467 p.
Nauta R 1987. A three-dimensional groundwater flow model of the Silurian dolomite
aquifer of Door County, Wisconsin: M.S. Thesis - Geology, Univ. of Wisconsin,
Madison, 105 p.
Oner, A.J. and Fiala, W.D. 1978. Soil survey of Portage County, Wisconsin: USDA Soil
Conservation Service, 96 p. + maps.
Papadopoulos, I.S. 1965. Nonsteady flow to a well in an infinite anisotropk aquifer:
|n: Proceedings, Dubrovnik Symposium on Hydrology of Fractured
Rocks, International Association of Scientific Hydrology, Dubrovnik, Yugoslavia, p.
21-31.
Quilan, J.F. and Ewers, R.O. 1985. Ground water flow in limestone terrenes: Strategy
rationale and procedure for reliable, efficient monitoring of ground water quality in karst
areas: Im Proceedings, National Symposium and Exposition on Aquifer Restoranon and
Ground Water Monitoring, National Water Well Association, Worthington, OH, p. 197-
234.
Rosen, CJ. 1984. Karst geomorphology of the Door Peninsula, Wisconsin: M.S.
Thesis - Geology, Univ. of Wisconsin, Milwaukee, 119 p.
Saines, M. 1981. 'Errors in interpretation of ground-water level data: Ground Water
Monitoring Review 1/1, p. 56-59.
Schuster, W.E., Bachhuber, J.A., and Stieglitz, R.D. 1989. Groundwater pollution potential
and pollution attenuation potential in Door County, Wisconsin: Door County Soil
and Water Conservation Dept., Sturgeon Bay, WI, 5 maps, scale 1-inch = 2640 ft
Shapiro, A.M. and Nicholas, J.R. 1989. Assessing the validity of the channel model of
fracture aperture under field conditions: Water Resources Research 23/5, p.
817-828.
82
-------
Sherrill, M.G. 1975. Ground-water contamination in the Silurian dolomite of Door County.
Wisconsin: Ground Water 13/2, p. 209-213.
Sherrill M G. 1978. Geology and ground water in Door County, Wisconsin, with emphasis
on contamination potential in the Silurian dolomite: U.S. Geological Survey
Water-Supply Paper 2047, 38 p.
Shuster, E.T. and White, W.B. 1971. Seasonal fluctuations in the chemistry of limestone
springs: a possible means for characterizing carbonate aquifers: Journal of
Hydrology 14, p. 93-128.
'Smith E D. and Vaughan, N.D. 1985. Experience with aquifer testing and analysis in
fractured low-permeability sedimentary rocks exhibiting nonradial pumping
response: In: Proceedings, Hydrogeology of Rocks of Low Permeability^
InternationaT Association of Hydrogeologists 17th Congress, Tucson, AZ, Memoirs,
v. XVn.p. 137-149.
Snow. D.T. 1969. Anisotropic permeabUity of fractured media: Water Resources Research
5, p. 1273-1289.
Socha, B.J. 1983. Fracture-trace analysis for water-well site locations in
igneous and metamorphic rock in central Wisconsin: Wis. Geological and Natural
History Survey Miscellaneous Paper 83-5, 37 p,
Stephenson, D.A., Fleming, AJL, and Mickelson, D.M. 1988 Glacial deposits:
In- Back, W., Rosenshein, J.S., and Seaber, P.R. (eds.), Geological Society of
America, The Geology of North America, v. 0-2, Hydrogeology, p. 301-314.
Stoeru, M.W. and Bradbury, K.R. 1989. Mapping recharge areas using a ground-water flow
model--a case study: Ground Water 27/2, p. 220-229. N
Theis C V 1935. The relation between the lowering of the piezometric surface and the rate
,sind duration of discharge of a well using ground-water storage: Transactions Amer.
Geophysical Union, v. 16, pt. H, p. 519-524.
Thwaites, F.T. and Bertrand, K. 1957. Pleistocene geology of Ae Dc»r Peninsula,
Wisconsin: Bulletin of the Geological Society of America 68, p. 831-S8U.
Todd,D.K. 1980 (2nd ed.). Groundwater hydrology: John Wiley & Sons, Inc., New York,
NY, 535 p. , ^ ... f . .,;- ,
Trescott, P.C. and Larsoni S.P. 1976. Supplement to Open-File Report 75-438: U.S.
Geological Survey Open-File Report 76-591, 21 p.
83
-------
U.S. Environmental Protection Agency. 1987. Guidelines for delineation of wellhead
protection areas: U.S. EPA, Office of Ground-Water Protection, Washinjjjton, DC,-
chapters paginated separately.
U.S. Environmental Protection Agency. 1989. Wellhead protection programs: tools for local
governments: U.S. EPA, Office of Ground-Water Protection, Washington, DC,
50 p.
Weeks, E.P. 1969. Determining the ratio of horizontal to vertical permeability by
aquifer-test analysis: Water Resources Research 5/1, p. 196-214.
Wiersma, J.H., Stieglitz, R.D., Cecil, D.L., and Metzler, G.M. 1984. Characterization of the
. shallow groundwater system in an area with thin soils and sinkholes:
Springer-Verlag, New York, NY, Environmental Geology Water Sciences;, v. 8, no.
1/2, p. 99-104.
Yanggen, D.A. and Amrhein, L.L. 1989. Groundwater quality regulation: existing
governmental authority and recommended roles: Columbia Journal of
Environmental Law 14/1, p. 1-109.
Zaporozec, A. (ed.). 1985. Groundwater protection principles and alternatives for Rock
County, Wisconsin: Wis. Geological and Natural History Survey Spec. Kept. 8,
p. 57.
Zaporozec, A. and Cotter, R.D. 1985. Major ground-water units of Wisconsin: Wis.
Geological and Natural History Survey Educational Series 28, p. 6 and 8.
Zheng, C., Bradbury, K.R., Anderson, M.P. (in press). A computer model for calculation of
groundwater paths and travel times in transient three-dimensional flows: Wis.
Geological and Natural History Survey Information Circular.
84
-------
APPENDIX A
JUNCTION CITY SITE, PORTAGE COUNTY, WISCONSIN
SITE SELECTION
The first demonstration area for testing wellhead protection methods in a fractured-rock
aquifer is located in central Wisconsin (see fig. 1). Crystalline rock of Precambrian age is at
or near land surface throughout much of central Wisconsin. .It is,either exposed or covered
by residual soil, a thin discontinuous layer of Pleistocene deposits, or small patches of
Cambrian sandstone (Bell and Sherrill, 1974). Several communities in central Wisconsin rely
on fractured Precambrian crystalline rock as their sole source of water supplies (Zaporozec
and Cotter, 1985). In choosing a site for this study, we identified several municipal water
supplies that were dependent on bedrock wells finished in crystalline rock. Utilizing
published water-table maps (Lippelt and Hennings, 1981), we looked for a site in a simple
hydrogeologic setting located near a ground-water divide. The Village of Junction City,
located approximately 12 miles northwest of Stevens Point in Portage County, met these
criteria. Additional advantages of this site included several existing test wells completed in
the crystalline rock, which could be used in the study, as well as cooperative town officials
and landowners. ,
INVESTIGATION METHODS
Preliminary site information was obtained from well constructors' reports on file at the
Wisconsin Geological and Natural History Survey. Data from 45 existing wells were plotted
on 1:24,000 topographic maps. The locations of the wells were digitized and entered into a
computer database along with water-level, geologic, and specific^capacity data. Reid
checking identified 15 additional bedrock wells (JC9, the village well; JC10 - JC14; JC16 -
JC19; JC22; and JC201 - JC204). The locations of the village well, the 14 existing bedrock
wells, bedrock holes drilled as part of this study (JC23 - JC24), and well logs used to
construct a geologic cross section are shown on figure Al.
Water-level measurements were taken on the 15 bedrock wells to assist in construction
of a detailed water-table map. Well JC18 is an observation well included in the statewide
water-level observation network under the number PT-82, for which water-level records are
available from 1951. Geophysical logging was conducted on four of the existing bedrock
wells (JC10, JC11, JC13, and JC14) as well as on core hole JC23 and test well JC24
(described below), which were specifically constructed for this project. Parameters measured
included natural gamma radiation, resistivity, spontaneous potential, hole diameter, and
temperature. A temperature log was also taken on the village well (JC9),
85
-------
P ,L
33
34
.1
I f'
I
JC-201I
_
JC-204
JC-202
JC-203
.-.I
10
SCALE 1:24 000
FEET 500 0 500 1000 2000
Villagewell (JC-9)
Existing bedrock well
A A' Cross-section line
Figure Al. Location of v/ells in a portion of the Junction City study area, Portage County,
Wisconsin.
86
-------
T0
field, eight piezometers (JCl - JC8 ^eremsui - JG4 and
aterials, and to conduct Pie^^gh!Lulic conductivity.
and JC22 provided estimates of bedrock hyora nC23V was
(jcv JC8)
^^
JC24) and in the pumped well (JG9).
*
-------
250 500 FEET
Village well (JC-9)
Bedrock welts
O Shallow piezometers
"*. Swale
r^ Road
II 'I Railway
V Surftetiwatur sample
Figure A2. Location of wells and piezometers in the well field at Junction City.
-------
Table Al. Well and piezometer data! Junction City study area, Portage County, Wisconsin.
Well
No.
Well"
Depth
Open
Interval
(feet below measuring point)
Average Depth
to Water
Measuring Point"
Elevation
(feet above msl)
JC1
JC2
JC3
JC4
JC5
JC6
3C7
JC8
JC9
JC10
JCM
JC12
JC13
JC14
JC16
JC17
JC18
JC19
JC22
JC23A
JC23B
JC24
JC31
JC201
JC202
JC203
JC204
32
30
43
30
61
35
26
18
321
128
98
143
96.
162
50
130
36
21
173
297
120
100
101
186
29^32
27-30
40-43
27-30
58-61
32-35
23-26
15-18
48-321
18-128
30-98
21-143
47-96
30-162,
18-50
40-130
10-36
10-21
12-173
248-297
30-120
10-100
24.7
23.9
25.8
23.7,
27.5
28.4
9.2
10.6
27.4
21.0
21.5
3.7
7.8
7.3
7.0
7.5
6.0
4.5
33.8
18.9
21.9
4.9
26.0
24.0
21.0
27.9
19.0
1153.5
1159.8
1154.4
1154.4
1167.9
1169.9
1172.8
1178.5
1159.4
1152.6
1157.8
1152.7
1147.5
1151.7
1149.1
1150.6
1151.0
1128.6
1164.8
1151.5
1151.5
1149.1
1183.0
1160.7
1151.2
1141.6
1151.7
Numbers JC1 - JC8 are 1 1/4-in diameter PVC standpipe piezometers with 3-ft plotted 1 screens
Numbers JC9 - JC204 are 6-in diameter wells cased into the bedrock and uncased belo*
the bedrock surface.
" Measuring point elevations range from 1/2 to 3 ft above ground surface. ^
Sve to measuring point elevation which is measured in feet above mean sea level (msl
- Indicates well depth or open interval unknown. ,
89
-------
sand, 7 to 63 percent silt, and 10 to 66 percent clay. A depth-to-bedrock map (fig. A3) was
constructed with data from bedrock exposures, mineral exploration cores, well constructors'
reports, and the eight piezometer boreholes installed as part of this study. The piezometer
boreholes indicated that the bedrock surface is deeper or more deeply weathered below
surface swales than in the areas between surface swales; resistivity surveys over the swales
(Bebel, 1989) confirm this.
The bedrock geology in Portage County has been mapped by Greenberg and Brown
(1986) and a detailed map of the Junction City area has been compiled by Brown for this
project (see fig. 2). In the well field, the dominant lithology is a dark gray Precambrian
mafic to intermediate metavolcanic rock. The rock consists of chlorite, quartz, and iron
oxides. Some relict, highly weathered-feldspar is visible; in fresh rock at. depth. Iron oxides
and quartz partially fill fractures in the upper, weathered zone; calcite is the dominant fracture
filling at depth. Continuous core recovered from JC23 (fig. A4) indicated that the top 35 ft
of bedrock is extensively weathered and heavily fractured (shown as cross-hatches on log).
The next 90 ft contained open, nearly vertical fractures spaced approximately 0.5 to 2 ft apart.
Below a depth of approximately 155 ft, visible fractures were nearly absent in the core.
The presence of three distinct bedrock zones under the surficia! deposits was
corroborated by the natural gamma logs from JC10, JC11, JC12, JC23, and JC24. Figure A5
shows the correlation of the gamma logs for these holes. Zone 1 consists of clayey residuum
and reworked hillslope deposits. Zone 2 is a highly weathered rock zone, approximately 20
to 40 ft thick. Zone 3 consists of rock with open fractures. The transition from a zone of
open fractures (zone 3) to a zone with few to no fractures (zone 4) at a depth of about 165 ft
is suggested by the gamma log from JC23 (fig. A5) and also by temperature logs from the
village well (JC9) and JC23 (fig. A6). The temperature log for the village well indicates a
change in temperature between 150 and 160 ft. (The temperature spike an 155 ft is caused by
the motor of the submersible pump in this well.) The change in temperature occurs because
little water enters the well below 160 ft; the zone of open fractures contributes significant
amounts of water to the village well; the zone with few fractures contributes little. The
temperature log of JC23 also shows a change at about 170 ft. Below this depth the borehole
temperature is constant, indicating that little or no water enters the well in the lower zone.
Information from the geophysical logs was combined with data from existing well
constructors' reports to generate a geologic cross section of the Junction City area (fig. A7).
RESULTS OF INVESTIGATIONS
Water-Table Mapping
A water-table map was essential for wellhead protection studies in the Junction City area.
A preliminary water-table map was constructed using surface-water features and
90
-------
L_
P :L - E'' I
SCALE 1:24 000
FtET 500 0 500 1000 2000
-20- Depth to bedrock contour
(interval 10 ft)
Village well (JC-9)
, i
Figure A3,. Depth to bedrock in the Junction City area.
91
-------
DESCRIPTION
£
0>
I
W
o
!
o>
5
0
o>
JD
£
O
0
?0
40
60
80
100
120
140
160
180
200
220
240
260
280
snn
WSA
(xW
WvS
SS="
5=»- '
^^^P^-^p^"*
i
_.___.
*T/
>-^^
53^-
' I
^ «r^:
\ 1
=====
r'l'
"i / '
y^af^L,
$^
^
Unlithified" clay and silt
Highly weathered and fractured rock
* Fe seam
^~ Fracture
^. Good unweathered mafic volcanic rock
Abundant calcite veins
Broken and brecciated rock
Some vertical calcite veins "
"""~ Minor calcite veins
"""^ Very firm rock
Parallel foliation
Clay zone :
Calcite vein '
Vertical foliation fractures \
Calcite coated fracture
Very few fractures
'j
Shear zone filled with pyrite
/
Broken zone
Broken zone
Calcite veins
Solid rock
Very broken rock
.
Some calcite
Pyrite veins
i
Figure A4. Geologic log for core hole JC23 at Junction City.
.92
-------
JC-23
JC-24
JC-11
. a §
JC-10
1*0
110
1M
I
.as -
«"
»
x
s
X
c
'
>
L
^
$"
^
f
/
.--. j*-«
-
s
^
I
L
*
/^
^
« ^
!>
?
>
^-
Zone 1 - clay«y
residuum
Zone 2 highly _^
weathered
rock
Zone 3 - rock
with open
fractures
Zone 4 - rock
with closed
fractures
Figure A5. Correlation of natural gamma logs for wells in
log the horizontal axis is counts per second (note different scales) and
d° pth
-------
JC-9 Temp (°C)
7.0 7.5 8.0 8.5 9.0
JC-23 Temp (°C)
7.0 7.5 8.0 8.5 9.0
u
« I
(
1
v
/
/
i
I
t
\
\\
Lt _
\f
I
~Zt-
!
<
, \
t
'
i
i
4
«
<
1
"
.
Figure A6. Temperature logs for the village well-(JC9) and the'nearby, dleep core hole
(JC23). $oth logs were run while the wells were being pumped. The vertical axis is depth in
ft from the land surface. Logs are slightly offset to correct for differing land surface
elevations. "
94
-------
Io fi o 11
S a s 2
« S 9 *
lisui J«oq« ul uoitASia
7 ' ""
i~^^_ . V . B
T^-^-r^-
1 ~ -~. WPC r*'-
UV , MV
- . - : - . ' . ' - '
_ . E : . =
' ' , . ~ ' ' ^- ' u
^ *" ' tj?-- = ;^^:^=^^'] - w. ^-i£^^,_
""I \"~\~" \
- , MV « MV MV .''..
; . . vtnci' >C«* «i«ggt>nM lOwnes
L ^r-^ -^:: - - - sc:: - t".: - . -.owu.
SCALE . ''"'-- ' ' ' . ' ' ' . -
. MC 0 MO 1C
t\pp\ox.n\\aws v
SC Unlithified silty
WPC Weathered Pr
MS Metasediment
MV Mafic volcanic
MO IMC »000« ... ".-.-'..'
s c. " r. .
/ater-table
clay . ,- "
ecambrian
ary rocks I Unweathered '
s . v ) Precambrian ' .
Figure A7. Geologic cross secdon for the Junction City area.
95
-------
depth-to-water data from well constructors' reports. The elevations of the surface-water
features and the measuring point elevations of the wells were determined from 1:24,000
topographic maps. Due to sparse data in the vicinity of the well field, this office-derived map
did not delineate a cone of depression around the village well.
' The final water-table map (fig. A8) is based on the above data plus water levels from the
15 field-located bedrock (JC9, JC10 - JC14, JC16 - JC19, JC22, and JC201 - JC204) wells
and the 8 piezometers (JC1 - JC8) measured in the spring of 1989. The elevations of these
measuring points were determined by rod and level surveying from known benchmarks. The
final water-table map depicts a clear cone of depression around the Junction City village well
(hatchured contours, fig. A8).
Water-Level Measurements
Vertical Distribution of Hydraulic Head
Total hydraulic head decreases with depth in the Junction City well field, indicating that
ground-water recharge occurs in the vicinity of the village well. Nested piezometers at four
sites provided estimates of the vertical hydraulic gradients. Piezometer pairs JC3 and JC4,
and JC5 and JC6 were finished in the unlithified surficial materials. At JC5 and JC6
hydraulic head decreased downward, with a gradient of 0.036, as calculated from water levels
measured over a 6-month period. At piezometers JC3 and JC4, the gradient was 0.011
downward. ,
Piezometers (JC23A and JC23B) placed in the core hole indicated a relatively strong
upward gradient of 0.016 between the deep portion of the bedrock and the upper zone with
open fractures. Piezometer JC23A was open from 904 to 855 ft above msl (approximate
depth of 250 to 300 ft below land surface) and piezometer JC23B was open from 1126.5 to
1051.5 ft above msl (approximate depth of 30 to 120 ft below land surface). The significant
head difference between these two bedrock piezometers is probably due to shallow drawdown
near the pumping village well.
Water-Level Fluctuations -
Water levels were recorded approximately monthly in the 8 piezometers and in wells JC9
to JC13. Water levels in piezometers within the mapped cone of depression were influenced
by pumping. Piezometers JC2 to JC4 exhibited fluctuations ranging from 9.6 ft for JC2, to
14.83 ft for JC3, and 13.83 ft for JC4. Wells JC10 and JC11, also located within the cone of
depression, fluctuated 13.74 ft and 6.86 ft, respectively. The wells and piezometers outside
the cone of depression exhibited smaller water-level fluctuations. Piezometers JC5 to JC8 all
fluctuated less than 6 ft; wells JC12 and JC13 fluctuated 2 ft or less.
96
-------
(-' -*<. ''<
SCALE 1:24 000
FEET 600 0 600 1000 2000
Water-table contour
(interval 10 ft)
Ground-water divide
Village well (JC-9)
Data point
Measuring point
Figure A8. Portion of the water-table map of. the Junction City area.
;' '- . 97 . .." .. ' .'
-------
Aquifer Tests
A variety of methods was employed to determine the transmissivity (T) and hydraulic
conductivity (K) distributions of the geologic materials in the Junction City area. These
methods ranged from simple office calculations of hydraulic conductivity based on data
available in well constructors* reports to a field-work intensive multi-well pumping test.
Specific Capacity Data
Specific capacity data from well constructors' reports were used to estimate the
transmissivity and hydraulic conductivity distribution of the fractured rock using the computer
program TGUESS (Bradbury and Rothschild, 1985). The calculated values for hydraulic
conductivity ranged from 1.2x10"* to 9.2x10"' ft/sec and are approximately log-normally
distributed. The geometric mean conductivity, which best represents the average conductivity
of the log-normally distributed values, was 4.3x10"*- ft/sec. A computer-generated contour
map of log hydraulic conductivity values for the Junction City area is shown in figure A9.
Slug Tests
Falling- and rising-head slug tests were conducted on the eight piezometers completed in
the unlithified deposits and on six existing bedrock wells. Results were aiisdyzed by the
Hvorslev (1951) and the Cooper and others (1967) methods. Table A2 contains a summary
of the slug test results. For each well, falling- and rising-head values were arithmetically
averaged to obtain the values reported in table A2. The geometric mean hydraulic
conductivity of the fractured crystalline rock based on slug tests was approximately 9.8xl06
ft/sec. The hydraulic conductivity of the overlying silt and clay was appnwdmately an order
of magnitude lower, about 6.6xlO"7 ft/sec.
Table A2. Results of slug tests at the Junction City site.
Piezometers
K (ft/sec) Wells
K (ft/sec)
JCl
JC2
JC3
JC4
JC5
JC6
JC7
JC8
1.9x10''
5.7X10-*
3.4x10'
6.2x10'
1.4x10"'
13x10-*
2.0xlO>7
4.0x10'
JC10
JC11
JC12
JC13
JC14
JC22
43x101
2.9x10 '
5.7x10*
7.0x10 '
5.9x10-*
3.1x10-*
98
-------
SCALE 1:24 000
FEET 500 0 500 1000 2000
-s.o Log of hydraulic conductivity contour
(interval .25)
V Village well (JC-9)
Measuring point
Figure A9. Computer-generated map of hydraulic conductivity values for the crystalline rocks
at Junction City. Hydraulic conductivity values were determined from specific capacity data.
99
-------
The installation of piezometers JC23A and JC23B and observation well JC24 allowed
hydraulic testing of the upper and lower parts of the bedrock aquifer. Single-well tests were
conducted on the core hole prior to piezometer installation and on the observation well. After
piezometers JC23A and JC23B were installed in the core hole, a slug test was conducted on
the deeper piezometer (JC23A). These results (summarized in table A3) show that the
hydraulic conductivity of the upper pan of the bedrock aquifer (represented by JC23B) is
approximately four orders of magnitude greater than the hydraulic conductivity of the lower
pan of the aquifer (JC23A).
Table A3. Results of single-well tests on core hole piezometers at Junction City.
Well or
Piezometer
Tested
JC23
JC23A
JC23B
Method
Pumping
Slug
Pumping
Open Interval
(ft below
measuring pt.)
30 - 297
248 - 297
30 - 120
Open Interval
(ft above mean
sea level)
1122-855
904 - 855
1122 - 1032
Hydraulic
Conductivity
(ft/sec)
1.6 x 10'5
1.0 x 10 8
1.4 x 10"
Pumping Test
On June 1 and 2, 1989, a 24-hour pumping test was conducted in the village well field,
using the village well (JC9) as the pumping well. During the multi-well pumping test most
wells were hand-measured; however, the pumped well (JC9) and two piezometers (JC23A and
JC23B) were measured using pressure transducers connected to a recording datalogger.
Pressure transducer readings were calibrated using hand-measured water levels recorded
throughout the pumping test. The village well was turned off for 24 hours prior to the test to
allow water levels to stabilize. Pre-pumping water-level trends were measured and used to
adjust the drawdowns recorded at each well The corrected drawdowns at each observation
point were analyzed using the Theis nonequilibrium method (Theis,1935),
Measurable drawdowns ranged from 19 ft at the pumped well to about 1 ft at
observation wells located approximately 450 ft away. Figure A10 shpws the extent of the
cone of depression at the conclusion of pumping. The cone is slightly elliptical, but the
general uniformity of drawdown around the pumped well suggests that the fractured rock at
Junction City behaves as a porous medium at the scale of the pump test.
100
-------
/,
,/
v ^
\
'f Drawdown, ft
v
' ' '-- VJG 1 '
. \,2\ />v\ 'r
JC-3. JC-4
« JC-10
' v*. 1-^
N *-^ jfe-
je-24
Seel*
0 100 200 f»
Figure AID. Drawdown for the Junction City pumping test.
°-°13
Transmissivity, ft /t'«c
J6-I«
Secl«
o too 200n
Figure All. Transmissiviry distribution in the Junction City
weUfield based on the pumping test data.
101
-------
On the basis of bedrock wells, transmissivity ranged from 0.011 to .0.022 frVsec, with a
geometric mean of approximately 0.013 frVsec. The specific yield ranged from 0.003 to
0.019, with a geometric mean of 0.0049. The transmissivity map (fig. All) shows a possible
zone of higher conductivity east and south-southwest of the pumped well. This zone may
correspond to swales in the landscape, and is consistent with previous observations that such
swales may correspond to zones of bedrock fractures (Socha, 1983) or zones of more deeply
weathered bedrock.
The method of Weeks (1969) was used to calculate the horizontal to vertical anisotropy
ratio (K/K,) of the bedrock aquifer. This method uses die principle that partially penetrating
observation wells generally show less drawdown than fully penetrating wells during a
pumping test in an anisoffopic aquifer.-"The departure of the partially penetrating well from a
predicted drawdown curve is related to the K,/^ ratio. Applying this method to the pumping
test data yields a K^/Kv ratio of 0.0006. This ratio seems anomalously low, but it is indeed
possible that the aquifer might be more transmissive in the vertical than in the horizontal
direction because the fractures and bedrock foliation are predominantly vertical.
Water Chemistry and Isotope Analyses
Water Chemistry
Ground water in the vicinity of the Junction City village well is currently of very good
quality. As shown in table A4, concentrations of all constituents tested are relatively low and
do not exceed drinking water quality standards. In particular, nitrate (NO3"), a common
ground-water contaminant in agricultural regions, was not detected in any water samples.
Ground water at Junction City is chemically a calcium-magnesium-bicairbonate water.
The chemistry of water produced by the village well (JC9) is similar to the chemistry of
water from nearby piezometers, suggesting that the source of water for the village well is
geochemically similar to shallow aquifer materials in the well field area. The deep
piezometer (JC23A) finished at 297 ft below the surface produces water with significantly
higher concentrations of HCO3" than the village well, suggesting that water produced by the
village well enters the well near the upper pan of the open interval.
Saturation indices for calcite and dolomite are negative for most water samples (table
A4), indicating that the ground water is undersaturated with respect to these minerals, as
would be expected for relatively young ground water in crystalline rocks. The deep core
piezometer (JC23A) shows oversaturation with respect to these minerals, and ground water in
the interval open to JC23A (248-297 ft deep) may be cor.sid'My older than the water
currently produced by the well. Positive saturation indices at shallow piezometer JC1, located
just downslope to the east of the village well, may be due to the use of road de-icing salt on
102
-------
Table A4.
Chemlca, ,,yScs of Junction «y .*« «"' ««- are in mimgr, per Ifar unfess cth^se indicated)
_ s,^.r.»inn Index
Simple-
in
JC 01
JC 02
JC 03
JC 04
JC 05
JC Ofi
jr o?
JC OB
jr o<»
jr cw
jr ii
JC 12
JC 13
JC 23A
JC 24
SH 01
Smupl*
!)«»«
' ~
10/28/88
10/28/B8
IO/2B/88
IO/2B/BB
10/2B/BB
IO/2B/88
10/2B/B8
IO/28/Bfl
IO/2B/B8
06/22/89
IO/2B/B8
IO/2B/B8
10/2B/B8
07/06/89
06/22/89
IO/2B/BB
T«p.
«C
7.5
B.O
7.8
8.0
7.3
8.0
10.0
9.0
10 3
10.5
7.5
9.0
9.5
8.2
13,0
3.0
Cond.
mho
_
390.3
241.4
197.3
233.9
215.5
513.0
299.9
27B.B
347.1
40.3
205.4
246.1
369. 1
DIM-
pH Oxygen
8.40
7.41
8.92
7.90
8.17
7 BO
7.71
7.0o
7.29
7.50
6.BB
6.99
6.90
8.15
7.40
9.1
6.3
5.6
6:3
5.3
B.7
7,6
99
2.0
4.8
7.0
7.1
2.0
3.8
CB ,'
32.1
23.2
16.1
2B.B
14.3
44,5
27.1
23.7
35.2
40.4
25.9
20.0
20.1
30 5
23.5
31.7
MK
19.1
11. 8
B.1
15.3
15.4
29.4
IB. 8
13.3
IB. 6
21. B
12.3
11.9
14.3
16.9
12.1
17.0
NH
13.1
5.0
7.0
7.0
2.5
12.fi
9.4
11.4
6.2
3.1
4.5
1.9
4.5
30.1
1.2
10. B
K
9.3
1.5
12.3
3.2
3.9
7.5
4.2
4.9
1.2
1.5
0.9
3.0
2.4
3.6
O.B
22.0
Fn
0.05
0.05
0.06
0.06
0.07
O.OR
0.07
0.04
<0.01
0.03
0.04
0.13
0.21
C«/M«
1,68
1.97
1.91
1.89
0.93
1.51
1.44
1.78
1.89
1.86
2.11
1.68
1.40
1.81
1.94
1.86
Calritt!
-
0.41
-0.71
0.47
-0.05
-0.20
-0.19
-0.20
-1.18
-0.54
-0.33
-1.31
-1.26
-1.36
0.25
-0.81
-0.56
Dolowitc
0.66
-1.63
0.74
-0.30
-0.31
-0.47
-0.44
-2.51
-1.24
-0 80
r2.86
-2.64
2.75
0.35
-1.74
-1.40
' - '~
PCOj
3.33
-2.34
-4.02
-2.76
3.14
-2.B1
-2.50
-2.12
-2.12
2. 38
-1.92
-2.01
-1.95
-2.94
-2.48
-2.42
-------
U.S. Highway. 10. Higher concentration of Cl' in this piezometer is cluuacteristic of road salt
contamination. . .
Isotopes
Ground-water samples for tritium (3H) and oxygen-18 (ISO) were collected from 15 wells
and piezometers in and around the Junction City well field. Table A5 contains the isotope
results.
As discussed in Chapter HI, most of the ground water in the Junction City area contains7
between 20 and 30 tritium units and is thus less than about 35 years old, including waiter
produced by the village well (JC9): 'One shallow piezometer (JC5), a deep well (JC12), and a
deep piezometer (JC23A) yield water with significantly less tritium, and the ground water in
the vicinity of these points is probably older.
- i1
Table AS. Isotope results for the Junction City area (estimated age is based on
qualitative interpretations summarized in table 3, chapter HI).
Well
JC1
JC2
JC3
JC4
JC5
JC6
JC9
JC9
JC10
JC11
JC12
JC13
JC23A
.JC23B
JC24
Sample
Date
10/28/88
10/28/88
10/28/88
10/28/88
10/28/88
10/28/88
10/28/88
06/22/89
10/28/88
10/28/88
10/28/88
10/28/88
07/07/89
06/22/89
06/22/89
Tritium
20
<35
<35
<35
<35
<35 :
>20
<35
>20
<35
<35
UQ
(per mil)
-11.55
-11.02
-9.58
-9.78
-9.67
-10.07
-9.90
-9.59
-9.60
-9.78
-9.65
-9.63
-9.75
-9.98
,
104
-------
NUMERICAL MODELING
v . . . * . ' ' '
Model Selection
basic model code and utility modules.
Conceptual Model
The first step in modeling any F^^^'Sw^ of the field-measurea wro-
the ground-water now system. On the basis orthe tie»«""?", Q
6 .. ^ u.-;» «o« KI. rf^ineated that includes the Junction uiiy
-
wherc the streams arc only intenmnem. -iiin.w
nodes and streams were modeled as constant-head .f^jSld.wiler drains
discharge points. Nonboundary intermittent streams were as rou
using the modular model's drain package.
105
-------
E AU RLE I
SCALE 1:100000
KILOMETRES 1 p 1 2 3 A
MILES 1 0; 1 2
Village well (JC-9)
Hydrologic boundary
Figure A12. Hydrogeologic boundaries for the Junction City site.
106
Site locat'O"
in
-------
Model Grid Design
ith the smallest discretization near the village well.
depth-u-bedrockmap (fig. A3)
overlies the fracnred crystaltae ^. this dep* the
and overlies aPI^*"5 $££££ oflUT bedrock characteristics, *e model was
bedrock contams few ^^.^"'^^^uihaied material and highly weathered
-
U«s rbyeTISknesscs and input parameters assigned to each layer.
Table A6. Layer characteristics for Junction City model.
Model Layer
Layer Type
Conceptual
Representation
Thickness KH X*v Average
(ft) (ft/sec) Ratio Porosity
Unconfined
Confined
Confined
Uhlithified
material and
highly weathered
bedrock
s
Bedrock with
open fractures
Bedrock with
few fractiires
40-90 3.6X10-6
115
150
10
2.0x10-*
to
9.5xlO'5
6.4xlO:8
1:1
1:1
0.10-0.27
0.05
0.03
highly
Model Layer 1 - Unlithified Material and Highly Weathered Bedrock
107
-------
O
oo
1
11
0
or
''
';
*
T
- ^.
-
\
^
i;
\
v
..
*>>
.N
\
.'
'T
"^
O
:°
\
V
-1
1
«
VN
\
0
^
o
,
0
N,"
V
,,*
O
o
0
o
0
-H-
~w
k .
'o
\
\
o
1
It
o
P:
o
^
,
o:
o
o
,
.
,
4
1
(
n
\
V
\.
11 9
0
V
..'
r^"
\
I'
e
I*
\"
\
,,
G^
v
\
*
T"
, A . .
x'»
\ ^'_
"'%
r>^
^-
V
\
.-
^
,
v
?.
,\
*
if
l
:
.-4.
. , -i
[
|
f<
i
\
\
. \
»
_
,
\
\
\
\
\
- ,\
"v
-\
--J-
.'**
,.,<
,
1 \
V \
v
- V
\
. "v
j
v
. .'.
. t
^:
* ^
'\
1 \
__
r
v
(iL
'ls
S
\
\
\
j^
1
?"<
\
1
V
^
\
^
,'0*
>*
'.
/i
*'f
''^.
^
y
"\
\
)
xfc
>
*
3r
i, .1
v.'
v
\
X,
\
.*
-A_
^ ^
X
X
%k
,,»
-
4 Village w*ll
Constant h*ad nod*
o No-flow nod*
F**t
17
COLUMNS
Figure A13. Areal view of grid and spatial discretization used in the numerical simulation of
the junction City ground- water flow system.
-------
1200
0*1000-
C
O
950-
>
UJ 900
Layer 2
Layer 3
Water Table
850
Figure A14. Configuration of model layers used in numerical model of ground-water flow at
the Junction City site (represents area near the village well).
109
-------
each node using saturated thickness-weighted averages oif the conduct!vines of each of the
two composing units (unlithified material and highly weathered bedrock).
Hydraulic conductivity of the unlithified material was based on shallow piezometer slug
tests performed near the village well (table A2). Spatial variation of hydraulic conductivity
throughout the rest of model area was based on the soil type given in the soil survey of
Portage County, Wisconsin (Otter and Fiala, 1978). Piezometer drillholes suggest that the
unlithified materials consist of Pleistocene hillslope deposits that grade iniio weathered
Precambrian rock. In areas where no other data were available, the soil material was assumed
to represent the composition of the entire thickness of unlithified sediment Hydraulic
conductivity values were adjusted according jo soil type; values ranged from 4.6x10"7 ft/sec in
the silty clay soils to 6.6xl(Jf ft/sec in sandy soils. ;«>,-">
Hydraulic conductivity of the highly weathered bedrock was based qp aquifer
transmissivity determined from the pumping test. Hydraulic conductivities along the swales
near the well field were increased by a factor of 2.25 on the basis of the pumping test results.
The swales are believed to correspond to areas of more highly weathered bedrock. Hydraulic
conductivity away from the well field was adjusted relative to the hydraulic conductivities
shown on the map generated from specific capacity data (fig. A9). Model conductivities
varied spatially by as many as three orders of magnitude.
The elevation of the bottom of layer 1 was calculated for each grid mode by subtracting
the depth to bedrock plus 35 ft (thickness of the highly weathered bedrock) from the surface
elevation. The drill core recovered from well JC23 contained numerous vertical fractures in
the top 35 ft, suggesting that vertical hydraulic conductivity could be greater than horizontal.
Because model calibration with vertical hydraulic conductivities greater than horizontal
proved difficult, vertical hydraulic conductivity was set equal to the horizontal for model
layer 1.
Model Layer 2 Bedrock With Open Fractures
Layer 2 was simulated as a confined layer 115 ft thick, treated as unconfined if the water
table were to drop below the top of layer. The pump test indicated that hydraulic
conductivity of the upper bedrock was about 7.5xlO"s ft/sec. For best model calibration, the
hydraulic conductivity of layer 2 was varied spatially from 2.0x10"* to 9.5xlO'5 ft/sec, and
vertical conductivity was equal to the horizontal conductivity.
Model Layer 3 - Bedrock With Few Fractures
Layer 3 was simulated as a confined layer with a thickness of 150 ft. On the basis of
slug test results from piezometer JC23A, the hydraulic conductivity of the lower bedrock
layer was assumed to be much less than that of the upper bedrock layers and was lowered
110
-------
ft/sec. Transmissivity of layer 3 was therefore set at
was se, equal to horiz^ta, eor^ucrivitv.
Model Calibration
lr!« ,o 20 inTw iThigh recharge areas. TTie average recharge rate over the ennre model
areas to 20 myyr ui mgn rec ^ ^ calculated recharge rate based on
e^±io£t a' US Seolo£eal Survev, long-term observation weU jus,
Ae s^y area (FT-82. which was numbered JC18 for tos study).
gprn). .,
'- . . . ' '".'.'.-" ''I''.': ' ,'-.-''
WHPA Simulations
Steady-State Flow Simulation
The model simulation produced a water-table map frg. A-15) that was a reasonable
match of the one produced with field measurements (fig. A8). The only substantial _
SscrepaLV between the simulated and measured water-table elevanons occurred m the ^
vicStyTthe village well where the model simulation did not predict as much drawdown in
the cone of depression as observed, probably due to discretization effects.
' ' . ' ' ' ' - ' ' ' ' ' Y
Transient Flow Simulation
Simulation of transient ground-water conditions at Junction City «^J|j"«£l
aquifer storaage, coefficients for each model layer. In addinon, recharge and WeU pumpage
were varied according to realistic seasonal fluctuations.
Estimates of storage coefficient were based on pumping test o
e a bedrock storage%oefficient of 0.0049. To account for the lowei: storage cc* ficiem^
unlithified materials, the storage coefficient of the ennre model layer 1 was set to 0.0013
Ill
-------
SCALE 1:24 000
FEET 500 0 500 1000 2000
-1 uo* Water-iable contour
(interval 10 ft)
Village well (JC-9)
Figure A15. Simulated steady-state water-table elevations in the Junction City area.
112
-------
o.ooooi.
Records of water levels in a USGS long-term observation well "umbered JC18
specific year, was being simulated, the recharge rates were calibrated by matching the
magnitude rather than the timing of the water-level fluctuations.
Simulated pumping rates were varied according to the 1988 ^
records. For each year simulated, the pumping rate ranged from 0.056 frVsec in April to
0.098 ftVsec in June.
The transient simulation was run for a period of 10 years to V*^****
head distributions with which to run PATH3D, the pamde-tracking program. The transient
simulation results differed little from the steady-state simulation results.
Particle-Tracking Simulations
PATH3D (Zheng and others, in press) is a particle-tracking program that makes use of
head dtsnfbution created by a flow-model simulation. O"-"*^^^^-'?^
for every node and every time-step and panicle paths are based on these vel«*ne^ PATH3D
tracks particle movement due to advection only; it ignores W**F^J?%^ -5 _L, -
Sfu ion, retardation, and other chemical processes. If a steady-state *****^£&
PATH3D can trace a particle's movement backward as well as forward in time. This model
ca^bmtySJows ZOC delineation by placing hypothetical particles at the well in question and
tracking them back to their origins in the ground-water system.
fa addition to the files used for the modular model simulations, PATH3D also requires
aquifer porosity for each model layer in order to calculate velocities. Poros^ °^*e ,
high end ofthe porosity range for fractured crystalline rock "^^^g^^
(1979). The saturated thickness-weighted average for layer 1 ^^^^toms^
the soU type and thickness of the unlhhified material and ranged from 0.1 to 0.27. Porosity
for layers 2 and 3 was set at 0.05 and 0.03, respectivelv (table A6).
Using the steady-state model simulation, particles at several deP^7 ^tes ^
backward in time from the well to delineate the upgradient ZOC. Figure A16 illustrate? me
113
-------
ZOC delineated when panicles are placed at a depth of 53 ft (elevation of 1106 ft above
msl)^presenting a point 5 ft below the bottom of the village well casing. The particles were
tracked backwards through time until they reach the water table. The end points of the
simulated paths therefore represent the origins in the ground-water system of particles that
would eventually reach the village well at the specified depth. TOT lines shown in figure
A16 are drawn to indicate the times involved for particles to travel from the water table to the
well due to advection. A larger ZOC is created when the particles are ssarted at a depth of
123 ft (approximately 75 ft below the bottom of the village well casing) at an elevation of
1036 ft above msl (fig. A17).
Because the ground-water flow system at Junction City is three-dime nsional, TOT lines
based on an assumption of horizontal flow can be misleading. Figure Al8 shows five panicle
paths (1 - 5) in profile view, along with the total travel time of each particle. The farther
upgradient a particle enters the water table, the deeper it sinks into the saturated zone before
reaching the well (compare particle paths 1 and 5). Particles that were sunned in the zone of
bedrock with few fractures (model layer 3) have much longer TOTs due to the substantially
lower hydraulic conductivity of this zone.
Particles reaching the village well at a depth of 162 ft (elevation of S'97 ft above msl)
entered the water table near the ground-water divide. An elevation of 1000 ft was therefore
used to delineate the largest predicted ZOC, extending almost from the ground-water divide to
the well. Hypothetical particles were placed around the well at 1000 ft elevation and tracked
backwards to the water table near the model boundary. The ZOC and associated TOTs are
shown in figure A19. This is the most accurate ZOC for the well because, it takes into
account the full depth of the three-dimensional flow system terminating at the well.
PATH3D was also run with the 10-yr transient simulation. Particles were placed at
points upgradient from the well along flow paths established in the steady-state runs. The
paths tracked in the transient runs were identical to those in steady-state. When dealing with
a site such as Junction City, in which the water-table fluctuations are relatively small, the
transient simulations appear to be an unnecessary addition to the steady-state runs.
114
-------
30-YR TOT
20-YR
10-YR TOT
SCALE 1:24 000
FEET 500 0 500 1000 2000
Village well (JC-9)
TOT Time of travel
Figure A16, Travel paths and TOTs for particles reaching the Junction City village well at a
depth of 53 ft below the land surface or 1106 ft above msl.
115
-------
I
L E'"' IN . E
o
I
I.
33
SCALE 1:24600
FEET 600 0 SCO 1000 2000
Ground-water divide
Village well (JC-9)
TOT Time of travel
11.
Figure A17. Travel paths and TOTs for particles reaching the Junction City village well a: a
depth of 123 ft below the land surface or 1036 ft above msl.
116
-------
Vtllag* Well
1150
Path
Path
Path
Path
Path
250
500
1000
1900 tobo MOO sooo
Dlstane* Upgrodisnt from Well (ft)
MOO
4000
4500
Figure A18. Profile view of travel paths and TOTs for hypothetical particles reaching
Junction City village well. Solid lines indicate particle travel paths; the time included at the
endof each line is the total TOT for a given particle path. Dashed lines indicate the 10-yr
and 20-yrTOTs.
117
-------
I ,
L_
1 X E
I
I .
I.---
v,,-
<
33
34'
?' -35
«
SCALE 1:24 000
FEET 600 -0 500 1000 2000
Ground-water divide
Village well (JC-9)
TOT Time of travel
11
Figure A19. Travel paths and TOTs for panicles reaching the Junction City village well at
depth of 162 ft below the land surface or 997 ft above msl.
118
-------
APPENDIX B
SEVASTOPOL STUDY ABLEA, DOOR COUNTY, WISCONSIN
SITE SELECTION
The second demonstration area for testing wellhead protection methods in firacwred-rock
aquifers is in the central part of Door County (see fig. 3), which forms a peninsula between
Lake Michigan and Green Bay in northeastern Wisconsin.
Door County is ideal for studies of fractured carbonate-rock aquifers for severaTreasons.
First the Silurian dolomite is highly fractured and provides the sole source of ground water
for mo* county residents (Zapofozec and Cotter, 1985). Second, because there.is only^t *in
veneer of soil covering the peninsula, the fractured rock is very near the land surface and is
comparatively easy to study. Third, dolomites of Silurian age cover a large area of the .
northern midwest/extending south along the Lake Michigan shore to Chicago and wrapping
abound the Michigan Basing occur in pans of Michigan, New York and Ontano Finally, a ,
great deal of prior hydrogeologic research provides a good database for use in wellhead
Protection studies. Thwaites and Bertrand (1957) described the geology of the Door
Peninsula, and Shemll (1978) summarized the hydrogeology of the county. More recently,
Bradbury (1982) and Nauta (1987) modeled ground-water flow in portions of the county,.
Schuster and others (1989) constructed detailed maps of soils and surficial features in part of
the county, and rated the soils as to pollution attenuation potential.
' This area has a history of elevated nitrate, chloride, bacteria, and occasional^', lead_ _
- levels in ground-water samples collected from private and public wells (Blanchard, W88.
Wiersma and others, 1984). Such contamination is believed to be a direct result of
agricultural and other land-use practices in areas where thin soils overlie the fractured
dolomite. , ;
Recent work by the WGNHS has focused on details of the hydrogeologic systenn at
several research sites in the county. Blanchard (1988) and Bradbury and others (1988) -
instrumented two research sites near the center of the peninsula in the town of Sevastopol
about 10 miles north of the city of Sturgeon Bay (see fig. 3, fig. Bl). Existing test we s and
piezometers at one of the sites, referred to as the Sevastopol test site, provided an excellent
starting point for this wellhead protection study. Ongoing ground-water monitoring at another
site, referred to as the barnyard research site, provided additional data points.
Door County is predominantly rural, and only two municipalities in the county are served
by community wells. Neither of these communities was judged to be suitable as a research
site due to complexities resulting from urbanization and proximity to surface-water bodies
Therefore, the wellhead protection methods demonstrated in Door County were applied to a
119
-------
Culsvltlf
ii
! .' X
TawnbwM.
SCALE 1:100 000
T2f N
OwtcUMRd.
Sevastopol Test
(Sec 2, T 28N,
o
o
£
c
o
©MW6A-C *MW
© MW5A-C
© MW4A-B
©MW3
©MW2A-D
t
z 4 Test Well MW1
© Piezometer Nest
^^P^^f^m^
0 50 ft
Site
R 26E)
1
1
© MW7A-C
Sevastopol test site
Barnyard research site
Barnyard Research Site
(Sec 36, T 29N, R 26E)
©H3A-
F««dlot
Bom
©H1A-B
QH2A-C
Townlln* Roija
50 ft
Figure Bl. Research sites in Door County, Wisconsin. Generalized locations are shown in
the top diagram. Detailed site diagrams are shown below.
120
-------
test well located at the Sevastopol test site described above. The hydrogeologic setting of this
site is characteristic of much of central Door County, and far more background data were
available at this site than at other locations in the county.
INVESTIGATION METHODS
Hydrogeologic studies conducted at the Sevastopol test site from 1986 through 1989^
examined, in detail, the vertical and horizontal movement of ground water through a small
arlTof fractured dolomite, by determining the position of the water table, measuring vertical
hydraulic gradients, measuring aquifer parameters, and sraPto«^rad.wt^atJ"^'
depths below the surface in a ground-water recharge area. In addition, detailed borehole
geophysical studies identifiedhorizontal fracture zones within the bedrock.
'seven monitoring wells (MW1 - MW7) were installed at the Sevastopol test site (fig.
Bl) using air-rotary drilling. Piezometer nests were installed in five of the wells. ,
Piezometers in each well are designated by the letters A, B C, etc J*W*-f»
construction details for all wells and piezometers. Five of the wells (MW1 - MW5) are
oriented approximately along a ground-water flow line and also along a majoi'^^
feature. Two of the wells (MW1 and MW2) reach a depth of approximately 240 ft, the_
common depth of newly constructed domestic wells in the area. ^^OW7?L^W3'
MW4, and MW5) were installed on a line between the two deep wells (MW1 and MW2).
Two additional wells (MW6 and MW7) are oriented at right angles to the bne formed by
MW1 through MW5, Piezometer nests installed in wells MW2, MW4, MW5, MW6, and
MW7 were used to investigate changes in hydraulic head and water chemistry in relation to
depth The annular space between piezometers was sealed with a mixture of bentonitt and
cement grout, and the piezometers were developed using compressed air. The resulting «ray
of 15 piezometers and two wells was used for measurements of the vertical distribution of
hydraulic head in the dolomite, for slug tests, for two pumping tests, and for obtaining
high-quality ground-water samples that were analyzed for major cations and anions as^well as
isotopes >H and "O. Seven additional piezometers (H1A-B, H2A-G, and H3A-BV at the
barnyard research site, located approximately 1.5 miles northeast of the Sevastopol test sue
(fig. Bi), provided additional hydrogeologic data. The construction of these piezometers was
similar to the construction of piezometers at the Sevastopol test site. In addition, water levels
were measured in approximately 50 domestic and irrigation wells in the area surrounding the
sites.
Geophysical logs, including three-arm caliper, spontaneous potential, singlejointand
normal resistivity, natural gamma radiation, borehole temperature, and borehole fluid flow
were obtained at most of the monitoring wells prior to casing installation. In addition,
television logs provided a visual inspection of fractures and other features^inside four
boreholes. A gVound-penetrating radar (GPR) survey of the site (Amg and others, 1987) gave
details on depth to bedrock and delineated shallow fractures.
121
-------
Table Bl. Well data, Sevastopol Study Area, Door County, Wisconsin.
Well
No.
Sevastopol Test Sit
. ; :*>MW1- />»=
MW2A
MW2B
MW2C
MW2D
MW3
MW4A
MW4B
MW5A
MW5B
MW5C
MW6A
MW6B
MW6C
MW7A
MW7B
MW7C
Barnyard Research
H1A
H1B
H2A
H2B
H2C
H3A
H3B
Well Open Average Depth
Depth Interval to Water .
(feet below measuring point)
c
<>, 240 -.=.-.:
42
161
147
78
60
44
31
24
21
19
61
40
20
185
153
106
Site
62
37
63
50
29
63
37
, . . . At\L')Af)
: ' fU-^fc^HJ
227-237
152-157
134-144
70-75
20-60
41-44
27-30
23-24
20-21
17-18
56-61
35-40
13-18
177-182
145-150
101-106
56-61
31-36
58-63
44-49
23-28
58-63
31-36
140.9
134.6
123.0
115.7 ''
70.8
36.5
30.4
17.7
19.5
17.4
14.2
48.2
31.0
16.2
154.1
137.0
105.3
36.6
29.8
37.1
32.5
25.1
38.8
30.3
Mflsasuring Point
Elevation
(feet above msl)
i
798.2
794.7
794.7
794.7
794.7
794.8
796.0
796.0
797.6 .
797.5
797.5
794.9
794.9
794.9
797.3
797.3
797.3
720.2
720.2
715.2
715.2
715.2
719.9
719.9
122
-------
HYDROGEOLOGIC SETTING
bed(seefig.4)
/ppOTnfly. *^±S2rS?.S^?*S2 in *. p^toninan. joint «.
'"0i^rn4S^W^»ftwS »«r the «s«m*site. Aerial photopaphs from
clayey or silty sediment.
evidence of fracn« o
included loss of drilling fluids ., «^M.^.«St^-ft^Sl (221 and 216 ft
MW7 and 235 ft in MW2
. A sui« of .ogs from
^^^
sK5^
Temperature increases at the offsets suggest that et
otKer elevations where temperature changes were not
123
-------
200 FEET
Figure B2. Expression of bedrock fractures in alfalfa field at, the Sevastopol site, Door
County, Wisconsin. Top: Oblique air photo showing vigorous alfalfa growth over fractures.
Bottom: Highlighted locations of fracture traces.
124
-------
to
BOOn
SP Resistivity Gamma Galiper Temp TV
Flow
-------
have been enlarged by dissolution. The television log confirmed the presence of a permeable
zone between 650 and 640 ft above msl (approximately 150 ft deep); this zone has a
dissolved, "Swiss cheese" appearance rather than the appearance of a discrete fracture. The
spinner flow meter detected significant borehole flow at an elevation of about 615 ft above
msl, a place in which the television log showed numerous vugs. The spontiineous potential
and resistivity logs can be easily correlated among boreholes at the site, and appear to be
xelated more to lithologic changes than to fracture locations.
RESULTS OF INVESTIGATIONS
Water-Table Mapping
Wellhead protection studies require up-to-date water-table and potentiometric-surface
maps for the ZOC delineation. Although detailed ground-water elevations were available at
the Sevastopol test site, the construction of reliable water-table maps of the surrounding area
required additional field measurements. During the summer of 1989, field personnel
measured ground-water levels in approximately 50 domestic and irrigation vi/ells in the area
surrounding the site. After corrections for land-surface elevation, these data, in conjunction
with piezometer data and surface-water elevations, allowed the construction of two contour
maps of ground-water elevations (fig. B4). Note that the water table near the Sevastopol test
site lies about 40 ft deep at an elevation of about 760 ft above msl (dashed contours, fig. B4),
and the Sevastopol test site is just west of the ground-water divide for the shallow system.
The potentiometric surface (solid contours, fig. B4) is more than 120 ft lower than the water
table with a hydraulic head of about 640 ft above msl near the Sevastopol ttsst site. The
ground-water divide for the deep system lies about 2 miles northeast of the Sevastopol test
site.
Water-Level Measurements
Vertical Distribution of Hydraulic Head
Vertical hydraulic gradients at the Sevastopol site are steeply down ward and are much
greater than horizontal gradients, suggesting that the aquifer is highly anisoaopic. Figure B5
shows the distribution of total hydraulic head in the subsurface in March 1989. The water
table at the Sevastopol site fluctuates over elevations of about 760 to 780 ft above msl. or
about 20-40 ft below the land surface as shown by the heads in "the shallow piezometers. The
continuous presence of water in the shallow piezometers (MW5A-C and MW6A-C) was
unexpected on the basis of previous research (Bradbury, 1982; Shenill, 1978), which placed
the water table in the area about 150 ft below the land surface. Below the-water table.
hydraulic heads decreased significantly with depth. A change in the vertical hydraulic
126
-------
Shallow
.-740-',
Hydraulic head contour
(10 ft.*-interval-* 5 ft)
Ground-water divide
Test well (MW-1)
Barnyard research site
Measuring point
SCALE 1:24 000
FEET 600 0 600 1000 2000
Figure B4. Distribution of hydraulic head in the shallow and deep ground-water systems in
the Sevastopol study area, August 1989.
'' ' ' ' . 127 : .''.'..''.''/'".
-------
oo
a
1 . '
* Piezometer Elevation (ft above msl)
>D (ft O» O» Nj *J Oi
.'if w .. . o (ft o (ft o
d 0,0 o o o o
Eh A
&* " ^*
o O o
M, _«.
CT Q
"§ ""' -
ess «_»
n" *< -m
"tr ~" O
g. 3°
S
s- cT
CO T- .
o i >4
en f^ "
S CL
a
** . t *
M ^"^
S |^S
g- ^^ oi
B. o-l
*
i i i i i i i i i 1 i i "i i i i i i i 1 I I I i i i i i i 1 i i i i i i i i i 1 i i i i i i i i i 1
;
^_ '
s . ' 5 ' s x " -
^ ^ ^ x
> > s ^
\ .
i x
>4 \ ' "
O ^ ,_
' ' 35 ' ' "
^ ^^ S^ ^M ^M
£V J ^^
V ' ^^» " ^^ » ^^* ^^*
v m O» Ol
S > CD
^ >N^-B
- 0» -^^B, . ^
' > ^ r S
" ^C ~^2 ^^
£ . 0 - . - "^ 5fe '^'
-«- -- - -- - - - - ------- ----- ----- - - --. - - ^ 21 cn
-S . > S o
sr ' . . ., _ -
** '
oo . - .-...'
-------
gradient occurred at about elevation 650 ft above msl, the approximate position of a
prominent horizontal fracture zone.
Ss^sr^rssj^^icstf- >
l^S wa*r show tha, a relatively shallow water table occurs at the sne. m addmon to
the deeper potentiometric surface.
Water-Level Fluctuations
Significant temporal water-level fluctuations occurred at the Sevastopol testsite, and _ - .
period.
The rapid response of the shallow and deep piezometers to spring snowmeU. suggest; that
both systems are directly connected to the land surface. The two systems respond dofferemly
J..^__ j_. ~~~*c TH* «,a«.r level in MW3 (shallow system) drops sharply ana tnen
level in MW2A
conductive fracture zones not present in the shallow system.
Aquifer Tests
^S±S^S^
Additional data obtained from other ground-water monitoring projects in the area
supplemented the Sevastopol test site measurements.
129
-------
BUU
-
r ^
cn
^ 750
o
O
D 700
-*
M
"S 650
D
-------
Specific Capacity Data
2S±±S«.253SyES±W
=B=^^^^^
r^re" m *e blstSate of the conductivity over the entire thickness of the aquifer
(Bouwer, 1978). ,
Slug Tests
Results of 14 slug tests (table B2) gave the distribution of hydraulic conductivity in the
dolomite Slug tests consisted of instantaneously changing the water level in a well or
piezometer by inserting or removing a solid slug of inert matend;and measunng wate - .
recovery using a recording datalogger. Slug tests were analyzed ^ *e "^7^^^
Cooper and omers (1967) methods. Hydraulic conductivities ranged from 3.2x10^ ft/sec to
3 STt/seTwUh a geometric mean of 3.2x10'' ft/sec (table B2). The highest hydrauhc
conductivity,3.6x10- ft/sec, was recorded in piezometer MW7A, which * screen* across a
fractal zone at an elevation of 617 ft above msl (depth 180 ft, see table Bl). Very low
hydraulic conductivities at piezometers H2B, MW5A, arid MW6B suggest that these
piezometers intersect very few fractures, and these results are probably charactenstic o1
hydraulic conductivity of unfractured dolomite blocks or massive units with few fractures.
Pumping Tests
Multi-well pumping tests suggested that the shallow and deep pans of the aquifer have
^f^l^K^s. o2 test, on we£ finished in the shallow^^J^S
for 24 hours at a pumping rate of about 0.05 ftVsec (22 gpm);^ ^ te« ^ well^=rm shed
thp rieen zone was conducted for 19 hours at a pumping rate of 0.07V it /sec (^ gpni>.
Drawdowns were measured in adjacent wells and piezometers and in the pumped wells,
^ssswrati ^=j^
\r£air^KiSnpr^^^
i« of ground water. Third, the aquifer is significantly anisotropic vertically and
horizontally. ..-.
131
-------
Table B2. Results of slug tests at the Sevastopol
test site and Barnyard research site.
Well or
Piezometer
MW2A
MW2B
MW2D
MW5A
MW6A
MW6B
MW7A
H1A
H1B
H2A
H2B
H2C
H3A
H3B
Mtximum
Minimum
Geometric Mean
K (ft/sec)
i 1.6X104
LOxio5
4.1x10*
6.9x10*
4.3xlOJ
5.5x10*
3.6x10'
5.3x10*
1.8x10*
1.4x10-*
3.2x10*
l.OxlO'
2.9x10*
2.4x10*
3.6x10'
3.2x10*
3.5x10*
Table B3. Results of pumping tests at the Sevastopol site.
Test Zone:
Date of Test-
Test Duration:
Pumped Well:
Pumping Rate:
Observation Wells:
Transmissivity:
Storage Coefficient:
Shallow. 0-60 ft
March 1988
24 hours
MW3
0.05 fiVsec (22 gpm)
MW4. MW5. MW2D
l.lxias ft»/sec
0.04
Deep, 150-240 ft
May 1989
19 hours
MW1
0.079 tf/t*c (35 gpm)
MW2A.MW7A
5.9xlCrz fr'/sec
0.0014
132
-------
An analysis of directional transmissivity using the method of Papadopoulos (1965).
yielded the following values for the deep zone:
Tu; 0.15 ftVsec; azimuth N33 E^
0.037 ftVsec; azimuth N123°E
K T nd'^are the principal directions of the transrrussivity tensor, tod by defmition
where Tu and Tn are the P"^ u is based on only ^ observation wells, so the ^
are perpendicula^ 10 ****££ ^^^Ay. However, this result is consistent with
at the site.
- '- ' . , ' '
Water Chemistry and Isotope Analyses
Water Chemistry
n£a\; ^content was consistently below the 10 mg^ dnnkmg water standard. , ....
Isotopes
rapid recharge.
133
-------
Table B4.
Chemical analyses of Door County samples (all values are in milligrams per liter unless otherwise indicated)
Simp IP Sn»plr T*mp.
IP Onlr °C
MWI
MWI
MW2A
HW2A
MW2D
MW2n
HW3
HW.1
M»4A
HW4A
HW6A
MWT.C
MW7A
HIA
HIA
HIA
HIM
HIM
HIM
H?A
H2A
H2R
HZ8
H2C
H2C
N3A
H3A
H3A
H3B
H38
II/02/B8
03/22/B9
II./02./BB
03/22/R9
il/02/Bfl
03/22/R9
1 I/02/B8
03/22/89
11/02/88
03/22/B9
II/02/B8
03/22/89
11/02/88
03/22/89
03/22/B9 x
II/07/8R
II/02/KR
03/20/R9
Ofi/OR/IW
1 I/02/R8
01/20/fW
06/08/B9
1 I/02/R8
06/08/89
11/02/88
06/08/89
11/02/88
06/08/89
11/02/88
03/20/89
06/08/89
! 1/02/88
03/2C/S3
06/08/89
7.9
7.9
7.6
7.6
8.1
8.2
8.1
8.8
8.2
8.4
8.1
B.2
9.1
8.7
5.8
7 8
7.7
B.S
8.3
9.0
8.6
8.3
7.7
8.2
8.5
8.8
7.4
10.1
6.9
8.3
8.2
10.7
S.I
S.O
7.9
Cond.
umhn
612.9
736.9
549.4
531.1
615.4
465. 1
615.4
532.7
588.2
578.9
571.8
664.7
660.0
947.0
>I303.4
607 0
525.0
683.0
927.7
953.8
704.8
912.7
928.3
718.7
1115.9
870.6
'1105.2
726.4
1090.2
481.8
487.7
476.6
10SS. 3
955.3
1180.3
pH
7.10
7.80
7.21
6.95
6.99
7.35
6.90
7.45
7.21
7.20
7.02
7.55
7.55
7.35
7.85
7.05
8.00
7.91
7.40
6.40
7.99
7.15
6.50
7.10
6.30
7.20
6.30
7.10
6.40
7.60
7.50
6.20
7.51
7. IS
6.30
DISK.
OxyKrn Cn Kg
4.4
12.5
1.9
3.5
3.5
10.0
3.9
8.1
2.4
3.0
1.5
5.5
5.0
5.0
9.4
I.S
2.4
6.3
4.8
7.8
2.8
3.1
2.9
2.7
5.7
6.4
8.3
2.3
73.1
60.5
66.4
59.5
75.2
59.3
77.4
60.1
68.7
64.6
67.7
57.1
70.5
74 5
86.9
72.6
60.8
65.2
71.8
71.7
77.4
75.8
86.1
70.7
90.3
77.8
78.7
82.1
90.9
49.7
52.2
49.5
63.2
88. 6
81.1
35.4
30.6
32.4
27.6
36.6
29.8
38.8
30.3
34.0
33.3
32.6
27.1
33.5
37.1
50.1
35.6
30.9
35.6
39.6
37.4
40.8
42.8
47.3
36.5
50.0
41.7
43.6
43.7
50.7
25.5
28.2
26.5
36.3
49.6
48. 0
Nn
1.8
1.2
1.5
3.3
1.0
2.1
1.6
1.9
6.9
7.7
2.1
1.5
23.3
13.5
57.8
3.3
1.6
20.8
20.0
12.4
5.3
3.2
3.6
7.4
16.2
11.8
52.3
6.5
15.5
3.3
5.2
20.0
60.1
10.0
30.8
K
2.4
1.2
1.1
2.8
2.8
2.4
3.3
2.7
2.6
1.2
1.4
1.8
1.9
1.0
2.7
2.9
1.6
7.1
5.8
5.5
2.4
2.4
4.1
2.7
18.8
3.8
15.4
3.5
16.2
1.4
0.8
1.7
4.6
3.2
20.0
F«
0.02
<0.0)
0.03
<0.0i
0.55
<0.01
0.08
<0.01
0.31
0.10
0.02
-------
Table B5. Isotope results for Door County.
Well
MW1
MW2A
MW2D
MW4A
MW5A
MW6A
HIA
H1B
H2A
H2B
H2C
H3A
H3B
Sample
Date
tritium
20
<35
-11.18
-11.19
-11.15
-9.58
-9.70
-11.23
NUMERICAL MODELING
Justification and Code Selection
of the problem is larger than jhe scale of fracturing. This » *? « » ^^^^ the
The wellhead protection modeling reported here builds on
investigations that successfully simulated ground-water flow m pans of Door Counry using
135
-------
numerical codes. Bradbury (1982) simulated ground-water movement at a site in northern
Door County using the two-dimensional parameter estimation inverse code of Cooley (1977)
and a three-dimensional finite difference forward code (Trescott and Larson, 1976). Nauta
(1987) extended this work to a three-dimensional model covering all mainland Door County.
Both investigations focused on simulating ground-water discharge to Green Bay and Lake
Michigan, and the models included only limited detail in the center of the county. Emmons
(1987) included the fractured dolomite of Door County in a regional model simulating flow
through the entire bedrock system in eastern Wisconsin, but his model was mot detailed
enough for wellhead protection studies.
The purpose of the numerical modeling study at the Sevastopol site is lira simulate as
accurately as possible .the ground-water flow .system at the site with the goal of delineating
the ZOC for a hypothetical municipal well. The modular ground-water flow code of
McDonald and Harbaugh (1988) provides for three-dimensional simulations incorporating
vertical and area! anisotropy and heterogeneity. This is a widely available and generally
accepted computer code, which is in the public domain and can be adapted lo run on personal
computers. The PATH3D particle-tracking code (Zheng and others, in press) links directly to
the flow model and was the main tool used in ZOC delineation in this study.
Conceptual Model
The Door County model assumes a completely saturated three-dimensional ground-water
system with hydrogeologic characteristics based on the results of field investigations in and
around the Sevastopol study area. The observation of significant changes in total hydraulic
head with depth, combined with the knowledge that extreme variations in hydraulic
conductivity occur with depth, make a three-dimensional model essential for accurate
simulation of ground-water flow paths from the land surface to a well. As previously
described, an unsaturated zone occurs under certain conditions beneath the upper saturated
zone at the test site. Because the flow model cannot simulate unsaturated flow, this zone was
treated as completely saturated. The resulting simulations are thus probably overprotecme
for wellhead protection purposes because ground water moves more rapidly through a
saturated system than through an unsaturated system.
Boundaries of the ground-water model are conceptually simple, and consist of constant
heads where the dolomite aquifer intersects Green Bay on the western side and Lake
Michigan on the eastern side of the county, constant heads along the Sturgeon Bay Ship
Canal to the south, zero-flux boundaries along the ground-water divide north of the study
area, and a zero-flux boundary at the base of the model where the underlying Maquoketa
shale forms a regional aquitard. Additional constant-head boundaries occur where surface
streams, such as Lilly Bay Creek and Donlans Creek, are in continual communication with
the aquifer. The upper boundary of the aquifer is open to recharge, which varies spatially
136
-------
Model Grid Design
23
The finite-difference grid used in the Door County model (fig. B7) cantos °^ 24 rows,
olumrTiid 4 layers, for a total of 2208 nodes, of which approximately 75 percent are
dvtm n^ spacing is irregular, and ranges from 5000 ft at *^*££« "^
500 ft in the area of interest around the test site. The gnd is oriented atan anmutf > of N30 E
so that the rows and columns in the model approximately parallel the principal directions of
the transmissivity tensor estimated from the pumping test.
For modeling purposes the ground-water flow system is divided into four horizontal
layers based on observed depth variations (fig. B8). The layers are numbered one to four, -
wfth layeM b^ing the shallowest «id layer 4 ^deepest, In model simulations ground water
cTm^e only horizontally within layers and only vertically between layers. ^ layer, ,
cover increasingly larger areas of the finite-difference grid, with layer I covering the smallest
arTanTlayer I covering the largest area. Table B6 summarizes the layer thicknesses,
characteristics, and input parameters. In all layers, a horizontal amsotropy of jW was _
assumed between the northeast (along model columns) and northwest (along model rows),
corresponding to the findings of the pumping test conducted at the test site.
Model Calibration
Steady-state model calibration consisted of adjusting model parameters until the model
suitably reproduced the field-measured water-table and potentiometric-surface maps of late
August 1989 (fig. B4). The model proved very sensitive to recharge rates and to the vertical
, , . ' » -."'
Table B6. Layer characteristics for Door County model.
Model Layer
Layer Type
1
2
Unconflned
Semiconflned
Semiconfined
Semiconfined
Conceptual
Representation
Upper system 0-75
Semiconfining 0-80
bed
Narrow fracture 0-5
zone
Lower system 0-365
Thickness
(ft) (ft/sec)
1.8x10-'.
83x10*
1.5xlO-2
1.0xlO-Jto
S-OxlO-1
Ratio
Average
Porosity
1:1
10:1 to
1000:1
1:1
1:1
0.01
0.01
0.05;
0.05
137
-------
CO
o
DC
4 Sevastopol site
Constant head node
o No-flow node
COLUMNS
Scale
I
0
I
10,000ft
Figure B7. Finite-difference grid used in numerical model of ground-water flow at the Sevastopol
sue.
138
-------
Water Table
Potentiornetric
Surface
Layer 3
o
1.400
Layer 4
300
FigureB8. Configuration of .nodcl layers used in numerical model of grounder flow a.
the Sevastopol site.
139
-------
hydraulic conductivities of each layer. Previous models of Bradbury (1982) and Nauta (1987)
provided baseline data for both these parameters. In the absence of detailed field data on the
variability of recharge, an initial estimate of the recharge distribution was obtained using the
inverse procedure of Stoertz and Bradbury (1989). Adjustment and smoothing of the resulting
recharge matrix, using the surficial materials maps of Sherrill (1978) and Schuster and others
(1989), gave the recharge distribution used in final model calibration.
Calibration to hydraulic heads alone generally does not yield a unique solution to a
ground-water flow problem. Calibration of the Door County model considered four criteria
for which field data were available:
1) reasonable match of the hydraulic head maps for the upper and lower layers,
2) accurate simulation of vertical hydraulic gradients measured at piezometer nests,
3) accurate simulation of ihelocation of ground-water divides,
4) accurate simulation of the ground-water flux measured as base flow in surface
streams.
Figure B9 shows simulated late-summer hydraulic heads for a "best-calibration" run in the
area around the Sevastopol test site (compare to field data in fig. B4). Although differences
exist between the simulated and measured heads, the match is considered acceptable for
wellhead protection simulations. .
WHPA Simulations
Steady-State Flow Simulation
A steady-state simulation produced the delineation of a reasonable ZOC at the
Sevastopol test site. The simulation includes a hypothetical pumping well at node (13 12)
the location of the Sevastopol test site (fig. B7). Currently, only two communities in the
county own wells, and most county residents obtain their water from individual private wells
Under current development pressure, however, more community wells may be installed in the
county in the future. There is interest in ZOC delineation for public and private wells.
Therefore, the simulation included a hypothetical community well pumping at 100 gpm With
current 170-ft well-casing restrictions, water enters the pumping node only in layers 3 and 4
Due to the high transmissivity of layers 3 and 4 and to the averaging of pumping stress over
an entire node, the pumping simulation produced only a small cone of depression in the
dolomite aquifer, such results are commonly observed in field pumping tests in the county.
Particle-Tracking Simulation
Coupling the PATH3D particle-tracking code (Zheng and others, in press) with the
three-dimensional head distribution produced by the pumping simulation, gives a picture of
140
-------
730" Head in shallow system
660 Head in deep system
V Test, well (MW-1) ,
SCALE 1:24 000
FEET soo o soo woo 2000
Figure B9. Simulated hydraulic head distribution at the Sevastopol site.
141
-------
the three-dimensional ZOC for the hypothetical well under stressed conditions. In the
steady-state simulations, PATH3D moves particles upgradient from the well until the particles
emerge at the water table. The final particle distribution at the water table represents the land
surface area where recharge eventually becomes ground water produced by the well; the
panicle paths represent three-dimensional ground-water flow paths.
Figure BIO shows the results of the PATH3D simulations in areal view. The simulation
used eight particles arranged around the hypothetical well (pumping 100 gpm) in layer 3 (for
clarity, fig. BIO shows only seven particle paths). The ZOC is narrow and elliptical,
extending about 2 miles northeast of the hypothetical pumping well (see fig. 16). Figure: BIO
also shows simulated travel times from the recharge area to the hypothetical well.
The numerical model also provides & picture of ground-water movement in the vicinity
of the hypothetical pumping well. Figure El 1 is a cross-section through the ZOC for the
hypothetical pumping well and shows how a thin, permeable fracture zone (at about 180 ft
below the land surface) affects ground-water movement. In this figure, psnrticles were started
at various depths along the well casing and moved upgradient to their recharge, points. Notice
that the horizontal fracture zone is the main conduit for particle movement to the well.
Ground water does not move with a constant velocity along the flow path but instead moves
most rapidly in a horizontal direction in layer 3 (about 180 ft below the hind surface) and
much slower vertically between layers 1 and 2.
Figure Bll also shows how the depth of casing of the pumping well can affect the
well's ZOC. Particles recharging farthest 'from the well enter the well bore at lower
elevations than particles recharging near the well (path 4). Thus, the more deeply the well is
cased, the larger the ZOC becomes.
142
-------
26
. *(-
35
5 ,
I
1-mon
* .test well (MW-1)
TOT Time of travel
SCALE 1:24 000
FEET SOO 0 500 1000 2000
1-yr TOT
. <-.. . *~-.
Figure BIO. Simulated particle paths using the PATH3D code, Sevastopol site.
. ; - " ' ' 143
-------
SEVASTOPOL SITE PARTICLE TRACKING PROFILE
I
Tes Wall MW-1
800-T
700-
w
E
o
&
o
500-
c
O
300-
200
1
water table
pofentlometrlc surface
-»-- Path 1; 5 yr
»***« Path 2; 1 yr
* * * * Path 3; 2 yr
>MMMW* Path 4; 37 yr
rO
-100
200 o
x
H300 «
O
400 ^
/N
-*»
*+~s
r-500
600
2000 4000 6000 8000 10000
Upgradient Distance from Test Well MW-1 (ft)
Figure Bll. Cross section of the Sevastopol site showing vertical ground-water movement from the surface to various depths
along a well casing.
' ' * "
------- |