&EPA
United States
Environmental Protection
Agency
Delineation of Source-water Protection
Areas in Karst Aquifers of the Ridge
and Valley and Appalachian Plateaus
Physiographic Provinces: Rules of Thumb
for Estimating the Capture Zones of
Springs and Wells
-------
United States
Environmental Protection Agency (4606M)
EPA816-R-02-015
www.epa.gov/safewater
July 2002
Printed on Recycled Paper
-------
Delineation of Source-water Protection Areas in Karst
Aquifers of the Ridge and Valley and Appalachian Plateaus
Physiographic Provinces: Rules of Thumb for Estimating
the Capture Zones of Springs and Wells
-------
The United States Environmental Protection Agency recognizes
the co-authorship of this document by:
Marilyn Ginsberg, Ph.D.
U.S. Environmental Protection Agency
Office of Water
Washington, DC 20004
and
Arthur Palmer
Professor of Hydrology
Department of Earth Sciences
State University of New York
Oneonta, NY 13820
-------
Acknowledgements
The authors offer their sincere appreciation to Jeffrey Imes, Hydrologist, US Geological Survey,
Rolla, Missouri, who provided valuable guidance, extensive discussion, draft and pre-draft
document and diagram review; to E.F. "Pat" Hollyday for his review and the generous
contribution of his time for consultation on technical questions; and to
Professor I Van Brahana, State University of Arkansas at Fayetteville, who provided guidance
and valuable suggestions, particularly in the formative stages of the Rules of Thumb.
The authors want to thank the reviewers of this document. Each provided significant
information, assistance and insight into its development: Dr. Kenneth Bradbury, Research
Hydrogeologist, Wisconsin Geological and Natural History Survey; Jeffrey A. Helmuth,
Hydrogeologist, Source Water Protection Team Leader, Wisconsin Department of Natural
Resources; Dr. Jack Hess, Executive Director, Geological Society of America; Bruce Kobelski,
Geologist, US Environmental Protection Agency, Washington, DC; Ryan McReynolds,
Environmental Scientist, US Environmental Protection Agency, Washington DC;
Wendy Melgin, Hydrologist and Deputy Chief, Watersheds and Wetlands Branch, Region 5,
US Environmental Protection Agency; Glenn E. Moglen, Asst. Professor, Dept. of Civil and
Environmental Engineering, University of Maryland at College Park; Bruce M. Olsen, PG,
Supervisor, Source Water Protection Unit, Minnesota Department of Heath; Randall Orndorff,
Geologist, US Geologcial Survey, Reston, Virginia; Foster Sawyer, Hydrologist, South Dakota
Geological Survey; Dr. Michael Smith, Research Hydrologist, National Oceanographic and
Atmospheric Administration/National Weather Service; Robert E. Smith, Geologist,
US Environmental Protection Agency, Washington, DC; Mike Wireman, Hydrogeologist,
Region 8, US Environmental Protection Agency; and Lloyd Woosley, Assistant District Chief,
US Geological Survey, Austin, Texas.
-------
Table of Contents
Acknowledgements i
List of Figures iv
List of Tables vi
Introduction 1
Historical background 1
Purpose 2
Vulnerability of karst aquifers 3
Problems in predicting karst ground-water patterns 4
Nature of karst ground-water flow in the Appalachians 4
Rule-of-Thumb topics 7
Part I: Rules of Thumb for estimating capture zones of karst springs
in the Ridge and Valley and Appalachian Plateaus Provinces 9
Introduction 9
Traditional methods 9
Resources needed to perform the Rules of Thumb for springs 10
Rule of Thumb 1.1: Dominantly bedded karst aquifers 11
Rule of Thumb 1.2: Dominantly fractured karst aquifers 16
Field tests 17
Part II: Rules of Thumb for estimating the capture zones for wells in karst
aquifers of the Ridge and Valley and Appalachian Plateaus Provinces 20
Introduction 20
Structural characteristics used in the Rules of Thumb 20
Resources needed to apply the Rules of Thumb 22
Assumptions 22
11
-------
Rule of Thumb 2.1: Steeply dipping aquifers (either dominantly
fractured or dominantly bedded) 23
Rule of Thumb 2.2: Gently dipping, dominantly bedded aquifers 27
Rule of Thumb 2.3: Gently dipping, dominantly fractured aquifers 27
List of References 31
Appendix 1: Water Fact Sheet Hydrologic Hazards in Karst Terrain 33
Appendix 2: Clarification of karst flow patterns 35
Appendix 3: Simple method for measuring spring discharge 37
in
-------
List of Figures
Figure 1: Ridge and Valley and Appalachian Plateaus Physiographic Provinces 1
Figure 2: Generalized karst ground-water flow in the Appalachians 4
Figure 3: Idealized view of water flow through dipping, dominantly bedded rocks 5
Figure 4: Flowchart for general Rules-of-Thumb process for estimating the capture
zones for karst springs and wells in the Ridge and Valley and Appalachian Plateaus
Provinces 8
Figure 5: Spring capture zone in plan and three-dimensional views 13
Figure 6: Cross section view of a spring capture zone where the updip projection
of the karst aquifer extends beyond the ridge, and the updip valley is floored by
carbonate rock 15
Figure 7: Cross section view of spring capture zone where the updip projection of
the karst aquifer extends beyond the ridge, and the updip valley is separated from
the underlying carbonate rock by an intervening thick shale 15
Figure 8: Graham (Plum) Springs, Mammoth Cave area of Kentucky. The capture
zone estimated with RT 1.1 is shown superimposed on the capture zone determined
by extensive dye traces and potentiometric data from well logs
(Quinlan and Ray, 1981, in White and White, 1989) 18
Figure 9: Davis Spring, Greenbrier County, West Virginia. The capture zone
estimated with RT 1.1 is shown superimposed on the capture zone estimated by
Jones (1997) on the basis of dye tracing 19
Figure 10: Capture zone of a well, located far enough from the headward
boundary of the spring's capture zone that there is no interference, in a steeply
dipping aquifer, regardless of whether the aquifer is dominantly fractured or
dominantly bedded 25
Figure 11: Capture zone of a well, located near the headward boundary of the
spring's capture zone, in a steeply dipping aquifer, regardless of whether the aquifer
is dominantly fractured or dominantly bedded 26
Figure 12: Capture zone of a well, located near the headward boundary of the
spring's capture zone, in a gently dipping, dominantly bedded aquifer 28
IV
-------
Figure 13: Capture zone of a well, located far enough from the headward boundary
of the spring's capture zone that there is no interference, in a gently dipping,
dominantly bedded aquifer. 29
Figure 14: Capture zone of a well, located far enough from the headward boundary
of the spring's capture zone that there is no interference, in a gently dipping,
dominantly fractured aquifer 30
Figure 15: Cross section through a typical capture zone showing the basic premise
behind the Rules of Thumb 35
Figure 16: a) Position of floats at the beginning of a stream-velocity measurement;
b) Position of floats at the end of a stream-velocity measurement 38
Figure 17: Cross section of a stream, with locations of depth-measurement points 39
-------
List of Tables
Table 1: Rules of Thumb for delineating approximate capture zones for springs
and wells in Appalachian karst aquifers 7
Table 2: Structural characteristics critical to use of the Rules of Thumb 21
VI
-------
Introduction
The goal of this document is to help delineate the approximate capture zones for springs and
public water-supply wells in karst aquifers of the Ridge and Valley and Appalachian Plateaus
Physiographic Provinces (Figure 1). These Rules of Thumb (RTs) are relatively low-cost
approaches that can be used by ground-water technical personnel, and which are likely to be
more accurate than fixed-radius methods. The initial intent of the US Environmental Protection
Agency (EPA) was to develop RTs applicable to all karst areas in the US, but regional
differences among karst settings limit the RTs described here to the Appalachians.
This document is written for ground-water technical professionals in agencies that implement
either the state Wellhead Protection Program (WHPP) or the Source Water Assessment and
Protection Program (SWAPP), or those who are interested in supporting the goals of these
programs. With the aid of ground-water
technical professionals, this document can
also be used by those with no technical
background.
Historical background
Section 1453 of the Safe Drinking Water Act
(SDWA) 1996, as amended by Congress in
1996, established the Source Water
Assessment and Protection Program, which
required all states to submit individual Source
Water Assessment Plans (SWAPs) to EPA for
approval and subsequent implementation by
the state. SDWA 1996 made it clear that
these programs were established "for the
protection and benefit of public water
systems", and further required all states to
make complete assessments of their public
drinking-water supplies by May 2003. Under
these SWAPs, each state and participating
Indian Tribe will: delineate the boundaries of
areas in the state (or on Tribal lands) that
supply water to each public water supply
(PWS), identify significant potential sources
of contamination, and determine the
susceptibility of each system to those sources
of contamination. State analysis of these
assessments will indicate the susceptibility of
each PWS in a source-water protection area
(SWPA) to the inventoried sources of
contamination in that area. „. , _., ,,,„ ,. , , • ™ .
Figure 1. Ridge and Valley and Appalachian Plateaus
Physiographic Provinces.
1
-------
The EPA published its national Source Water Assessment and Protection Program guidance
document on August 6, 1997 (EPA, 1997). The document outlines in detail the minimum
program elements required for EPA approval for meeting the stated goals of the program - that
is, the protection and benefit of each PWS. The EPA's review and approval of these programs
revealed a number of consistencies in states' approaches. In most cases, the states conducted
contaminant-source inventories and susceptibility analyses for each PWS within individual
SWPAs, rather than for PWS clusters within area-wide SWPAs. Most state SWAPs rely heavily
on EPA-approved WHPPs for protection of ground-water sources of drinking water and have
essentially met the SDWA 1996 requirements under their existing state WHPPs.
For various reasons, mainly for convenience in the absence of effective guidelines, many states
have used arbitrary or calculated fixed radii to delineate wellhead protection areas (WHPAs)
and SWPAs. Most fixed-radius approaches, while meeting minimal requirements for
public-health protection, are problematic even in hydrogeologic settings with the lowest
probable risk of contamination. The fixed-radius approach has significant deficiencies that are
particularly troubling in karst settings. The approach is flawed by the assumption that water
flows equally from all directions, and can significantly underestimate wellhead and spring
capture zones. This is a particular problem in karst, because of its high vulnerability to
contamination, high ground-water velocities and limited opportunity for in-situ remediation. In
karst settings, the calculated fixed-radius approach, based on a porous-media assumption, is not
a significant improvement over the arbitrary fixed-radius approach. The porous-medium
assumption does not apply to karst.
To strike a balance between enhancing public-health protection and minimizing cost, EPA has
developed several RTs for voluntary use by states that would like more protection at relatively
low cost. These RTs, described below, provide simple ways to delineate WHPAs and SWPAs
for springs and wells in karst aquifers of the Appalachians (both the Ridge and Valley and
Appalachian Plateaus Provinces). The original intent was to develop RTs applicable to other
karst settings as well. However, regional differences between the Appalachians and other karst
settings, including the patterns of conduits, fractures, and flow directions in relation to geologic
structure, are too great to allow this. The EPA hopes that these RTs will encourage ground-
water technical experts in other karst settings to look for commonalities in flow characteristics
that might be used to develop RTs helpful in their settings.
Purpose
This document provides RTs for approximating the capture zones of springs and PWS wells in
the karst aquifers of the Appalachians. These RTs are relatively low-cost approaches that are
likely to be far more accurate than fixed-radius methods. States in the Appalachians may
consider using these RTs in their source-water assessments, to obtain a better balance between
degree of public-health protection and cost. The EPA urges Appalachian states not only to
consider using these RTs in place of fixed-radius methods, but also to consider moving beyond
these RTs to more-accurate delineations through more site-specific delineation approaches.
-------
These RTs require more effort than fixed-radius methods, but are fairly inexpensive, and in
nearly every case, provide more-accurate estimates of capture zones for wells and springs (and,
therefore, significantly more protection to PWSs). However, the RTs are not foolproof.
Many sophisticated methods, such as computer modeling, hydrogeologic mapping, and tracer
testing, are available to analyze complex ground-water systems. However, many data are
needed to make the methods work, particularly in highly heterogeneous systems. Computer
modeling rarely works in karst regions unless the models account for turbulent-flow conduits.
Even dye tracing, which is usually considered the most convincing tool for delineating
ground-water basins in karst, is not physically or economically feasible in some cases, and
rarely gives more than an outline of the major conduit flow. Geophysical surveys may help to
delineate the local geologic framework and major conduits, but the surveys cannot determine
detailed flow patterns and divides in karst. However, all these methods provide significant
information and analysis, and EPA encourages their use to supplement the RTs in this
document. All these methods must be performed by experienced professionals.
Simple guidelines such as these RTs greatly over-simplify the complex ground-water flow
systems in Appalachian karst, "... karst groundwater protection is much more complicated
than the protection of porous aquifers, because the karst systems heterogeneity requires
generally much time and combination of different research methodologies." (Biondic, 2002).
Nevertheless, the RTs apply fairly well to Appalachian karst aquifers. Examination of 149 dye
traces in karst of West Virginia described by Jones (1997) shows that more than 75% could have
been predicted by applying the techniques described here. Nearly all other traces were in
massive carbonate rocks with complex and poorly mapped structure. Some of the predictable
traces covered distances as much as 15 miles. The techniques also work well in areas in well-
bedded carbonate rocks in low-relief plateaus such as those of the Mammoth Cave area in
Kentucky (see dye traces by Quinlan and Ray, 1981, in White and White, 1989). These two
field studies are the only American examples widely available in published form.
These RTs are least applicable in those areas of the Appalachian Plateaus where there is a thick
cap of insoluble rock and limestone is exposed as only a narrow band at the bottoms of deep
valleys. They also may fail to apply in massive, intensely faulted carbonates, for example, parts
of the thick Cambrian-Ordovician carbonates in the Valley of Virginia.
Vulnerability of karst aquifers
Karst aquifers are among the most highly vulnerable to contamination, particularly where the
overlying soil is thin (Appendix 1). This vulnerability results from: (1) the presence of point-
source recharge features such as sinkholes, (2) solutionally widened flow paths, and (3) rapid
velocities of ground water and contaminants. These characteristics reduce the time and matrix
surface areas for in-situ, water-quality improvement by water-soil-rock interactions and
chemical or microbial breakdown of contaminants. Point-source recharge features and wide
flow paths limit natural filtration, although a thick regolith can provide some in-situ, natural
remediation of water quality.
-------
Problems in predicting karst ground-water patterns
It is difficult to identify ground-water flow paths and divides in karst aquifers. This difficulty
arises from the extreme heterogeneity and anisotropy that typify karst aquifers, and from
changes in ground-water patterns with different stages of flow. For example, ground-water flow
paths, divides, and basin boundaries can shift in response to rising ground-water levels during
and after major precipitation events. Also, conduits in the unsaturated zone can behave rather
independently, causing ground-water basins to interfinger and overlap at any given time.
Nature of karst ground-water flow in the Appalachians
Water flows overland from ridge tops, then enters the ground in upland regions, primarily
through recharge features (for example, sinkholes, sinking streams, fissures), and exits at
springs in low areas, mainly stream valleys (Figure 2). Diffuse infiltration can also take place
through the soil or through a caprock of permeable material such as sandstone. Where
insoluble rock overlies soluble rock in an upland area, recharge features tend to develop along
the contacts. On steep slopes that do not readily develop sinkholes, diffuse infiltration can
occur through the soil or into bedrock fissures. In the Appalachian karst, small valleys are
often dry, as they lie above the local potentiometric surface. The only perennial surface streams
are major entrenched rivers that serve as outlets for ground water, as well as for streams that are
perched on insoluble, low-conductivity rocks.
Strike arc,: Dip
I I I Dijwndip [law in
* jrtsaturaied z.cne
i '
i i
Poiiiiun
1'lort -Tersects rhs Aatef tebte
i-assurnad ro be at appro maiely
Ihn xiirne ciev;ilit;in ;IK Ihe s;|jr ficj I
Figure 2. Generalized karst ground-water flow in the Appalachians.
-------
CIO
From recharge points, the pattern of individual ground-water flow paths tends to have a strong
downdip component in the unsaturated (vadose) zone and a strong tendency to follow the strike
in the saturated (phreatic) zone (Figure 3). In karst areas, solution conduits develop along the
routes of greatest vadose and phreatic water movement. These conduits carry high-velocity
turbulent flow, and they include caves that are large enough to explore. In the Appalachians,
some of these conduits reach tens of meters in diameter. The statements about preferred flow
routes in this section are supported by the mapping of accessible conduits.
The reason for the downdip tendency of vadose water is that gravitational water follows the
steepest available paths through the unsaturated (vadose) zone. Vertical or steeply dipping
fractures are not always able to transmit all the water that enters from karst recharge features.
This forces water to overflow along the next-steepest openings, which are generally extensive
bedding-plane partings. In this way, much of the water follows the dip of the strata, jogging
downward across the beds wherever a
sufficiently wide fracture is
encountered that offers a steeper path.
Major flows of karst ground water,
thus, follow stair-step patterns to the
water table with a strong lateral offset
in the direction of the stratal dip. This
tendency is most common in well-
bedded rocks, especially where there
are insoluble interbeds. It also
applies to well-fractured rocks, but to
a smaller degree.
In steeply dipping rocks, vadose
water drains along steep paths to the
water table. In gently dipping rocks,
vadose water can remain perched
along bedding-plane partings for
distances up to several kilometers
before reaching the water table; thus,
there can be a great horizontal
dislocation between where the water
infiltrates and where it ultimately
reaches the water table.
Figure 3. Idealized view of water flow through dipping, dominantly
bedded rocks. Vadose water tends to flow down the dip of the beds,
following bedding-plane partings or perched on thin insoluble
strata. Where it reaches the water table, the flow tends to follow
approximately the strike of the beds to the most efficient outlet.
Most phreatic flow follows shallow paths because fractures and
partings become narrower with depth. (The strike-oriented pattern
is favored because it is the intersection between the water table and
the favorable dipping bed that delivers the incoming vadose water.
In places the vadose water may jog downward along discordant
fractures from one bed to another, but then resumes its downdip
trend.)
Where the vadose flow reaches the water table, the dip of the strata loses its influence, because
gravity is more or less offset by the increasing hydrostatic pressure with depth. Instead, the
water follows the most "efficient" path to the nearest available surface outlet - that is, the path
that provides the least resistance for a given amount of flow. Because the incoming vadose
water is already following favorable partings and fractures, those same openings continue to
serve as the optimum paths for flow at, and just below, the water table, in the saturated
(phreatic) zone. The preferred phreatic paths are, therefore, along the strike of the dominant
structures, at their intersection with the low-gradient water table that is typical of
-------
high-permeability karst aquifers. In places, the major flow routes usually loop down below the
water table, but the overall low-gradient, strike-oriented trend dominates, especially in steeply
dipping rocks. Appendix 2 clarifies the patterns of flow within a spring capture zone and
provides a simplified description of the basic premise behind the RTs.
Fractures and partings become narrower with depth, owing to the pressure of overlying rock,
and as a result, the most favorable flow routes are at shallow depth, at or just below the water
table. This trend is disrupted to some extent in strongly faulted and folded rocks, but even in
these areas the majority of ground-water flow is close to the water table.
For the purpose of delineating capture zones, it is important to consider two relationships:
1) The position of a strike-oriented conduit is determined by the location of the most
abundant sources of incoming vadose flow. Most of these are located updip from the
strike-oriented conduit.
2) The location of the strike-oriented conduit, in turn, determines the general location of the
spring outlet. Thus, the general pattern of underground drainage to the spring can be
anticipated.
The prevalence of these patterns was noted in statistical summaries by Palmer (1986, 1999a) on
the basis of geologic mapping of karst conduits, with supporting evidence from dye traces.
These two articles (short entries in symposium volumes) were intended only to draw attention to
a promising approach, rather than to recommend immediate application. However, the
reliability of this approach in predicting approximate spring capture zones came to the attention
of EPA for use in this document.
Turbulent-flow conduits are surrounded by large areas of bedrock that contain only diffuse
laminar flow along narrow fractures, partings, and intergranular pores. This diffuse seepage
drains into the conduits in much the same way that it flows into surface streams. Thus, if the
pattern of conduits were known, the pattern of the surrounding diffuse flow could be predicted.
Linear zones of low head, as detected by piezometric data from wells, generally indicate the
approximate location of phreatic conduits. Vadose conduits cannot be detected in this way.
Both vadose and phreatic conduits can cross under topographic divides. This is least likely
where soluble rock is overlain by a thick cover of insoluble rock, such that the soluble rock is
exposed only near the bottoms of stream valleys. In these areas, underground karst drainage
usually follows paths roughly parallel to the stream valleys, because that is where fractures are
most likely to have widened by stress release. The capture zone for a spring extends beyond the
outcrop area of the soluble rock, if surface runoff is contributed from neighboring insoluble
rocks.
-------
Water in karst aquifers is highly susceptible to contamination for the reasons stated above.
However, most of the water-quality problems are limited to conduits, which represent only a
tiny volume of the entire aquifer, despite the fact that they carry the majority of the aquifer's
discharge. The surrounding zones of laminar flow are less susceptible to contamination, since
their recharge is rather diffuse. In addition, their ground-water flow is tributary to the conduits,
except for brief intervals during rising floods, when conduit flow is temporarily reversed into
the surrounding network of pores and fissures, as a form of bank storage. The presence of a
regolith does not affect the utility of the RTs.
Rule-of-Thumb topics
The RTs described in this document are outlined in Table 1.
Rules of Thumb for estimating the capture
zones for springs in Appalachian karst
aquifers:
Rules of Thumb for estimating the capture
zones for wells in Appalachian karst
aquifers:
RT 1.1: For dominantly bedded aquifers
RT 2.1: For steeply dipping aquifers, either
dominantly fractured or dominantly bedded
RT 1.2: For dominantly fractured aquifers
RT 2.2: For gently dipping, dominantly
bedded aquifers
RT 2.3: For gently dipping, dominantly
fractured aquifers
Table 1. Rules of Thumb for delineating approximate capture zones for springs and wells in
Appalachian karst aquifers.
Figure 4 is a flowchart summarizing the general RT process.
Estimating a spring's capture zone requires information on geologic structure, topography, and
the hydrologic budget for the spring's actual ground-water basin. With the RT methods,
calculated capture zones for neighboring springs may overlap slightly, because the techniques
are designed to overestimate the areas slightly. Estimating a well's capture zone requires
knowing the spring capture zone in which the well is located, the directions of maximum and
minimum hydraulic conductivity, and the size of the cone of depression. The geologic setting
of a well is similar to that of the spring capture zone in which the well is located.
In the section titled "Field tests", RT 1.1 was used to estimate the capture zones for two large
karst springs. The results were very close to published estimates of the capture zones based on
dye tracing.
-------
Obtain discharge for spring of Interest
1
Determine total capture /nnc area necessary
to support discharge of spring
1
Review availahk1 geological inforiiiiitUm
\
>isil field «
1
Plo< rechargefrutwres, PWS wells, and
iv informal ion, «n (upturn plik map
Draw strike-
1
In increments, miniate caplure mnt by extendi
updip lini'-H Ki >urt'LU'i' from Mriki.1 line
I
Add area draining to recharge fciturcs
us appropriate
1
With t-u'h incrt'nii.'nt, comp^fk* si/i
/one to tlm' si/t- of tlir ari>a
fruiii the\priri)4
1
In ilnniiti^nlli, fi'D(,'turi'd riivk, jdjii>r (In-
ujMiirt /wni1 h> v^ttridln^ its
bi 541^'b. in (lit1
1
Add any adjaccftt arfa$ drain ing to
included k»rsl recharge features
Figure 4. Flowchart for general Rules-of-Thumb process for
estimating the capture zones for karst springs and wells in the
Ridge and Valley and Appalachian Plateaus Provinces.
-------
Part I: Rules of Thumb for estimating capture zones of karst springs
in the Ridge and Valley and Appalachian Plateaus Provinces
Introduction
The RTs for approximating the capture zone of a spring are described below. The EPA used the
following criteria in developing these RTs: (1) they should apply to the majority of springs in
Appalachian karst settings, (2) they should be conceptually simple, (3) they should be relatively
low cost, (4) it must be feasible for them to be performed by ground-water specialists at state
SWAPP-implementing agencies, or by local hydrogeology graduate students supervised by
someone versed in karst hydrology, and (5) estimates of parameters needed to use the RTs are
available from state geological surveys or US Geological Survey (USGS) offices; nearby
university geology departments may also be sources of relevant information. Engineers and
geoscientists knowledgeable about karst hydrology can also apply these techniques. Some
states may require that such work be performed by licensed professionals. The EPA estimates
that the delineation of a spring capture zone with the RTs below can be completed within
several days by professionals or within a week or two by graduate hydrogeology students. (The
majority of the time required to estimate a well's capture zone is spent delineating the spring
capture zone in which the well is located.)
Traditional methods
There are three traditional methods for delineating the capture zone of a karst spring. Method 1,
described below, is no more time consuming than the RTs, but it is difficult to apply in
extensive karst areas. The last two described below can be costly and time consuming. The
RTs described later are suggested if the resources are not available to apply these traditional
methods. However, the RTs can also provide initial estimates of capture zones that facilitate
applying the traditional methods. The traditional methods can be used alone or in combination.
They are summarized here, because they provide insight into the nature of ground-water flow in
karst.
1. Many spring capture zones can be approximated from the distribution of recharge features
(sinkholes, sinking streams, fissures, etc.). Karst springs are fed primarily by such features,
many of which are visible on topographic maps. The capture zone for a karst spring can
often be roughly outlined by identifying clusters of sinkholes and sinking streams in the
upland surrounding the spring, and including any surface drainage into them. This
qualitative approach is best used only as a preliminary step in applying one or more of these
traditional methods.
To constrain the size of the spring's capture zone, the spring's discharge is measured during
a period of high base flow. Assuming that recharge is uniform over its capture zone, the
size of the capture zone of a spring is proportional to the spring's discharge. This
relationship can be estimated from USGS stream-flow records, which include stream basin
area and discharge for every monitored drainage basin (available in hardcopy, and also
online at http://water.usgs.gov/nwis/sw). High base-flow discharge per unit area of the
major river basin in the area is calculated from stream-flow records and basin-size
-------
information. The value of high base-flow discharge per unit area is then used in
combination with spring-discharge rate, to determine the size of the spring's capture zone.
Example: If the high base flow for the major river in an area is 0.5 feet-Vsecond/mile2, and
the discharge of the spring, measured during high base flow, is 3.0 feet3/second, then the
size of the basin that feeds the spring will be approximately 6 miles2.
The discharge per unit area of a karst spring is usually slightly larger than that for
neighboring surface streams, because evapotranspiration is smaller in basins in which most
of the flow is underground. This technique does not indicate the location of capture-zone
boundaries. This procedure for estimating the size of the spring's capture zone is part of the
RTs described later. However, the RTs provide a more valid delineation of spring capture
zones where large areas of karst drain to several different springs, and where the divides
between capture zones cannot be defined by the extent of karst recharge features.
2. Ground-water drainage divides can be roughly delineated from water levels in wells.
However, the ground-water divides detected with wells rarely indicate the full capture zones
of karst springs, because vadose water can easily cross over water-table divides. Divides
can also shift with changes in ground-water stage. It is necessary to make all measurements
during periods of similar base flow. All wells must penetrate the karst aquifer in question,
but they must not extend through it into lower aquifers.
3. Dye tracing is the most reliable way to delineate spring capture zones, if there are abundant
recharge features into which the dye can be introduced. This procedure must be performed
only by those with extensive dye-tracing experience. State regulations on the use of tracers
must be observed. The technique uses small concentrations of non-toxic dye (for example,
fluorescein), which ideally emerge at springs in non-visible concentrations of typically a few
parts per billion. Positive traces are identified by dye breakthrough curves from continuous
or intermittent monitoring. Tracer studies can be costly, time-consuming, and sometimes
confusing (for example, where subsurface drainage divides overlap at different levels, or
shift laterally as water levels change), and they do not always provide positive results (that
is, the dye may not be detected after injection). Poor technique or background noise from
other sources of fluorescence can lead to false positives. In addition, dye tracing does not
reveal the exact pattern of conduits between injection and detection sites.
Where time or resources are not sufficient to apply the traditional methods, the following Rules
of Thumb can provide estimates of spring capture zones with relatively little cost or effort.
Resources needed to perform the Rules of Thumb for springs
A small amount of basic hydrogeologic information is needed to perform the RTs. Some data
may have to be obtained independently, if not readily available from other sources. Although
dye tracing and potentiometric mapping would help to refine delineations based on the RTs, it
is usually because these data are not available that the RTs are employed. Required information
is:
10
-------
1. Discharge records for streams in the area, to obtain an approximate rate of recharge to the
aquifer.
2. The spring's discharge rate, measured at high base flow, at least several days after a
significant precipitation event. The most appropriate flows generally occur in February in
the southern part of the Appalachians and in March in the northern part. Although spring
discharge data may be available from other sources, the reader can refer to Appendix 3,
which describes a simple method for measuring spring discharge.
3. Determination of whether the carbonate rock is dominantly fractured or dominantly bedded,
and whether steeply dipping (>5 degrees) or gently dipping (<5 degrees). "Dominantly
fractured" in this context means that there are prominent joints and/or faults, discordant to
the bedding. "Dominantly bedded" means that beds and bedding-plane partings are more
conspicuous than discordant fractures. In most cases, the distinction will be clear. Where it
is not, a compromise between the RTs can be applied. That is, the capture zone would
consist of the superimposed capture zones for both settings.
4. A geologic map of the area, preferably at 1:250,000 scale or larger (that is, more detailed),
from which the strikes and dips in the spring's capture zone can be determined, and which
provides information on the elevation of formation contacts. Geologic/hydrogeologic
descriptions of the area are helpful. These include, but are not limited to, published or
unpublished: reports, isopach maps, structural maps, and well-log descriptions.
5. Topographic maps, preferably at the 1:24,000 or 1:100,000 scale; aerial photographs and
county soil-survey maps at these or similar scales are also helpful. Walking the streams and
estimated capture zone can reveal significant features not recorded on maps and photos.
Rule of Thumb 1.1: Dominantly bedded karst aquifers
This RT assumes that the hydraulic head in the strike-oriented main conduit is not substantially
higher than the head at the spring to which it drains. The elevation of a spring is, therefore, a
reasonable approximation of the elevation of the water level in the conduit draining to the
spring. In the Appalachians, these conditions appear to be the norm in carbonate rocks that are
dominated by bedding-plane partings. Palmer (1999a) estimates that in predominantly bedded
carbonate rocks about 75% of the main conduits will fall within 10 degrees of the strike line,
and that about 90% of the vadose conduits that feed them will be oriented within 10 degrees of
the dip direction. At least 80% of the vadose conduits should have no more than a 10-degree
downward discordance across the strata.
The Rule of Thumb consists of the following steps:
Step A: Obtain the high base-flow discharge for the spring of interest. This may be available
from the USGS, state geological survey, or nearby universities. However, if this
information is lacking, spring discharge can be estimated by the simple method
described in Appendix 3.
11
-------
Step B: Determine the total surface area of land necessary to supply the discharge to the spring.
Consult USGS stream-flow records to estimate the discharge per unit area for surface
streams (see Traditional Method 1, above). Assume that this value is roughly the same
as that for springs (it will usually be somewhat smaller for streams than for springs).
Use equation 1, below, to estimate the size of the capture zone needed to supply the
spring discharge:
Equation (1):
spring discharge (Iength3/time)
= size of spring capture zone (length2)
ground-water recharge rate (length/time)
Make sure the units in the equation are compatible, so they cancel to produce the
desired units on the right-hand side. If the ground-water recharge rate is not available,
use the discharge per unit area from the published records for nearby streams:
spring discharge (feet-Vsecond)
= size of spring capture zone (miles2)
discharge per unit area (feet-Vsecond/mile2)
This equation should be applied over as long a period of data as possible, to reduce the
effect of discharge fluctuations. However, even for shorter periods, it provides a valid
first approximation.
Step C: Review available geologic information to determine the strike and dip of the carbonate
rocks in the upland above the spring. Examine topographic maps and aerial photos for
likely recharge features (sinkholes, etc.).
Step D: Visit the field site to look for recharge features not shown on maps and aerial photos.
If feasible, walk upstream and downstream a few thousand feet along the stream into
which the spring of interest drains, in a search for additional springs. Some may
function as overflows for the main spring, but they will usually be dry during periods
of low flow
Step E: Plot the field information (Figure 5) on a large-scale topographic map, for example, a
scale of 1:24,000 or 1:100,000. Where multiple springs are in close proximity (within
about 1000 feet), consider them to be one large spring with a combined discharge,
located at the outlet that is at the lowest elevation. On the map, mark the carbonate
rock's strike and dip directions, and its dip angle (Figure 3). Plot any PWS wells on the
map (it may be necessary to estimate capture zones for several springs in order to find
the one in which a given well is located.)
12
-------
Plan View
Mote: the dofrn&p width
equals SGV,> of Ihe ddwodtp'widlh of the
o* ihe
It is £dd§dJn-tr6*ninantly Cradured settings
- ,1-ij
"
~| Hprirc Capluro^cro
Figure 5. Spring capture zone in plan and three-dimensional views.
13
-------
Step F: From the spring, draw a line parallel to the average strike of the beds, into the karst area
that feeds the spring. This line is a rough approximation of the main solution conduit
that feeds the spring. (This line extends into the region of recharge features identified
on the map.) The length of this line depends on how large a capture zone is needed to
supply the spring's discharge (Figure 5).
Step G: From the spring, extend a line up the dip of the beds until it intersects the surface. This
requires knowing the dip angle (from the geologic map or similar information) and
elevations of the land surface (from the topographic map).
Example: If the spring elevation is 600 feet, assume that the main solution conduit
feeding it is at the same elevation (represented by the strike line). If the dip is
10 degrees to the east and the local land surface lies at 900 feet, a line projected upward
along the dip from the strike line will intersect the surface at a point equal to
[(900-600)/tan(10°)] = 1700 feet west of the strike line.
Proceeding along the strike line, away from the spring, draw additional updip lines at
intervals until they, too, intersect the land surface. Connect the points where the updip
lines intersect the surface. With each additional step, calculate the total accumulated
area enclosed by the lines, and compare it to the estimated capture zone of the spring
determined in Step B. Stop when the two areas are equal. As each new area is added,
include any additional land area that lies outside the zone defined above, but which
contributes surface runoff to it. Do not draw a boundary through a cluster of recharge
features; rather, include the entire cluster. Do not extend the estimated capture zone into
obviously non-karst regions in which surface drainage is away from the spring. The
capture zone can extend beneath an insoluble, low-conductivity caprock (Figures 6 and
7), provided that the soluble aquifer containing the spring in question extends beneath it,
and that the contact between the two lies above the spring level. The capture zone
should not cross entrenched rivers with perennial flow, although it can extend across
small streams perched on insoluble, low-conductivity caprock (Figure 7).
The resulting shape is roughly a quadrilateral (only approximately, because, 1- the
irregularity of the land surface prevents the updip boundary from being a straight line,
and 2- additional land area, as noted above, may contribute surface runoff). Figure 5
shows the updip projection from the strike line, reaching the surface without passing
beneath the topographic divide. Alternative settings are described at the end of these
Steps.
Step H: If the enclosed area is too small to account for the spring discharge, expand the
estimated capture zone into adjacent areas that contain karst recharge features. Expand
the boundaries incrementally, in equal steps, into all of these karst areas, regardless of
direction, until the size of the capture zone is sufficient to account for the spring
discharge. Do not allow the capture zone to cross perennial streams, unless they are
perched on insoluble low-conductivity strata above karst aquifers. This procedure is
especially appropriate where the dip is steep, because the capture zone delineated by
14
-------
area cori1nE«jt'ig
In spring
area contributing to
dcwndip recharge
Taaluf&s
intermittent awarn
Ihiil will drain to
ckwndip
Shale
Limestone
Figure 6. Cross section view of a spring capture zone where the updip projection of the karst aquifer extends
beyond the ridge, and the updip valley is floored by carbonate rock.
area dnanny
W rectusnje feglwes
an updip sidn of slrnsm
within thfl sfes tf this
that is flonrart bfy
. ii t/ a *r*.Lji"i-l« Ilieie a •'••.:
the stream receives minor seepage from the downdip
-------
Step G will be a very narrow strip that is unlikely to be large enough to deliver the
necessary recharge to the spring.
If the updip projection in Step G extends beneath a topographic divide, drainage within the
capture zone is slightly more complex, as shown in Figures 6 and 7. If any adjacent areas
contribute surface runoff to this delineated capture zone, the capture zone should be expanded
to include these areas. An example of such an area is an updip valley floored by carbonate rock
(Figure 6). Such a valley is unlikely to contain a major surface stream. Any area of surface
runoff that does not contribute water to the area in Step G should be excluded from the spring
capture zone. An example of such an area is an updip valley separated from the underlying
carbonate rock by an intervening thick shale (Figure 7).
Some springs may be located in such a position that the strike line could be drawn in either
direction. For example, if the spring is located at the northernmost bend in a river, and the
strike is east-west, the strike line could obviously be drawn in either direction - east or west -
and still extend into the upland karst area. It is unlikely that the capture zone extends in both
directions, although field examples are known in which this is the case. In some cases, the
appropriate direction may be clear from the distribution of karst recharge features. In the
absence of a clear distribution, the safest approach would be to extend a strike line in both
directions, following the instructions above, until the necessary capture-zone area is obtained in
each direction. This would give a total capture zone twice the necessary size to account for the
spring discharge.
The capture zone for a large spring may include areas drained by smaller, perched springs. If
the spring of interest is itself a perched spring, the procedure for delineating its capture zone
will be the same as described in Steps A-H above.
Rule of Thumb 1.2: Dominantly fractured karst aquifers
In dominantly fractured rocks, the flow pattern is roughly similar to that described above, but
the dip-and-strike control is less apparent. Water more readily follows fractures across the
bedding. As a result, the downdip portion of flow will be less extensive, and the main conduit
feeding the spring is less likely to be along the strike. The general relationships still hold, but
the estimated capture zone is less likely to be valid and should be modified. (Palmer [1999a]
estimates that although the average trend of phreatic conduits centers around the strike of the
beds, only 50% of them will fall within 50 degrees of the strike. Eighty percent of vadose
conduits have trends that fall within 50 degrees of the dip direction, and 55% of them have less
than 10 degrees downward discordance across the beds.)
Thus, in these settings it is best to start with RT 1.1 and then modify the capture zone by adding
a strip in the downdip direction about half the width of the original quadrilateral (Figure 5).
(There is no solid quantitative basis for this value; it is simply a safety factor that takes into
account known examples of updip flow in fracture-dominated aquifers.) In addition, any
adjacent areas containing karst recharge features should be added. In expanding the boundaries
in these two ways, no perennial surface streams should be crossed. The resulting area is likely
to be larger than necessary to account for the spring discharge, but this simply reflects the
greater uncertainty of the method.
16
-------
Field tests
The approaches above were tested at two sites by the senior author, who was unfamiliar with
either site. The first site was Graham (Plum) Springs, one of the largest in the Mammoth Cave
area of Kentucky. Although this area is not in the Appalachians, it contains prominent bedding,
and the local strike and dip are well mapped, making it a likely candidate for RT 1.1. The
capture zone estimated with this RT is shown superimposed on the capture zone determined by
extensive dye traces and potentiometric data from well logs (Figure 8; see Quinlan and Ray,
1981, in White and White, 1989).
The second site was Davis Spring, in Greenbrier County, West Virginia, the largest spring in the
state. In Figure 9, the estimated capture zone defined by RT 1.1 is shown superimposed on the
capture zone estimated by Jones (1997) on the basis of dye tracing. The strike line extending
from Davis Spring follows the axis of a syncline, and so the updip projection was made in both
directions away from the strike line.
It is clear that these RT delineations provide a far better match than would a fixed-radius
delineation. Materials for performing the necessary steps were limited to those available from
the USGS library in Reston, Virginia. That is, maps and literature were available, but aerial
photos were not used and no field work was performed. Each of these delineations was
performed in 2 to 3 days. However, the author was saved about 1 day of effort for each,
because the sizes of the spring capture zones were already known without resorting to
measurements and calculations of spring discharges. The author found that the second
delineation was performed more rapidly than the first, because the technique quickly became
routine.
17
-------
Figure 8. Graham (Plum) Springs, Mammoth Cave area of Kentucky. The capture zone estimated withRT 1.1
is shown superimposed on the capture zone determined by extensive dye traces and potentiometric data from
well logs (see Quinlan and Ray, 1981, in White and White, 1989).
18
-------
;
•* f m£~~ • 3t*M-p
p?*1
-'
Capture Zone as delermined by Rule of Thumb
Capture Zone as dBlentiined by Dye Tracing
Figure 9. Davis Spring, Greenbrier County, West Virginia. The capture zone estimated with RT 1.1 is shown
superimposed on the capture zone estimated by Jones (1997) on the basis of dye tracing. (The strike line
extending from Davis Spring follows the axis of a syncline, and so the updip projection is made in both
directions away from the strike line.)
19
-------
Part II: Rules of Thumb for estimating the capture zones for wells in karst
aquifers of the Ridge and Valley and Appalachian Plateaus Provinces
Introduction
There are three well RTs presented in Part II. These are listed in Table 1 (page 7).
Most wells in Appalachian karst obtain their water from narrow fissures (fractures and
partings), many of which are enlarged only slightly by dissolution. Conduits are sparse, and
relatively few wells obtain their water from them. Conduit-fed wells can be identified by
negligible drawdown or by turbid water during high flow (for example, after heavy rains).
Wells draw an insignificant amount of water from matrix blocks in the dense Appalachian
limestones. Therefore, the RTs below take into consideration only conduits and narrow
fissures.
Ground-water travel time is a useful criterion in WHPA or SWPA delineation in porous media.
However, in karst aquifers travel time can be so short that it generally is not a meaningful
criterion for estimating capture zones for wells.
The areas delineated by these RTs may be quite extensive. The estimated capture zone for a
well in a karst aquifer may be closer in size to that of a surface-water basin than to the capture
zone of a well fed by laminar flow in a porous medium (unless that well is conjunctively
delineated to include the surface-water catchment area). Capture zones in karst aquifers may be
overestimated. This overestimation is justified by, and results from: (1) high-velocity flow
through fissures and solution conduits over large distances, (2) uncertainty in identifying exact
ground-water flow paths, and (3) shifting of ground-water divides in response to storm events.
To obtain a first approximation for a well capture zone (that is, its WHPA/SWPA boundary) in
Appalachian karst, the approach to estimating a spring's capture zone (see Part I) will be
combined with standard approaches to delineating well capture zones. Although the resultant
capture zone likely will be very large, there will be critical areas within it that would be the
focus of protection-management activities. These critical areas are uphill from, and
surrounding, recharge features such as sinkholes and sinking streams. (In the downhill
direction, EPA informally suggests a critical-area setback at least equal to the state setback
[often called the "health setback" or WHPA "inner zone"] for protecting a PWS well against
direct spills. Generally, this setback ranges from 100 to 400 feet, depending on the state. If a
state is without such a setback, EPA informally suggests a minimum of 100 feet). Recharge
features can frequently be identified on large-scale (for example, 1:24,000) topographic maps
and aerial photographs, but it is more reliable to identify them in the field.
Structural characteristics used in the Rules of Thumb
There are two critical structural characteristics that must be evaluated in order to use the RTs
for wells. These characteristics are (1) steepness of the dip and (2) dominant type of
permeability, that is, bedding-plane partings vs. fractures discordant to bedding. In plateaus, the
dip is usually less than about 5 degrees. At the other extreme, where rocks are folded (as in the
20
-------
Ridge and Valley Province), most dips are steeper. Table 2 is a 2 x 2 matrix showing the
hydrologic conditions produced by these end members. (Kmax and K^^ are maximum and
minimum hydraulic conductivity, respectively.)
Dip Angle
Flow is Dominated by
Bedding-Plane Partings
(Dominantly Bedded)
Flow is Dominated by
Fractures Discordant to
Bedding Planes
(Dominantly Fractured)
Gently dipping
(< 5 degrees)
Horizontal anisotropy is
small, (Kmax «Kmm),
although there is much
strike-oriented flow for the
same reason as in the box
below. (Vertical anisotropy
can be great, causing
perching of vadose water
and strong downdip flow.)
Horizontal anisotropy is
great, usually about 10X
(Kmax«10Kmm),withKma:
oriented roughly parallel to
the dip. Perching of vadose
water is not common.
Steeply dipping
(> 5 degrees)
Apparent horizontal
anisotropy can be great,
because dipping bedding-
plane partings restrict
horizontal water movement.
Effective Kmax is in the
direction of the strike,
usually about 10 Kmm.
Vadose water follows steep
routes to the water table
along bedding.
Horizontal anisotropy typically
withKmax«JOKmm,
with K^x in the direction of
strike. This can be disrupted
by complex folds and faults,
but is fairly consistent. Most
vadose water takes steep
paths to the water table along
fractures, with little
perching.
Table 2. Structural characteristics critical to use of the Rules of Thumb
The user must determine whether the rock is dominantly bedded or dominantly fractured.
Although there is a continuum between the two, most Appalachian karst aquifers fall near one
of the two end members. Geologists or hydrogeologists at the USGS or state geological survey
are likely to be able to make this distinction. Otherwise, outcrops of the karst rock should be
examined to determine if many beds are visible or if joints are more prominent. Even where
there are many small fractures, unless they cut through large sections of strata (for example,
tens of feet), bedding-plane partings probably dominate over fractures. This is especially true
where there are numerous interbeds of relatively insoluble material, such as shale.
The user must also determine whether beds are gently dipping (<5 degrees) or steeply dipping
(>5 degrees). This can either be measured in the field or determined from a geologic map or
geological report.
21
-------
Resources needed to apply the Rules of Thumb
Once the spring's capture zone is delineated (RTs 1.1 and 1.2, above), the only additional
information required to determine the well's capture zone is the well's discharge (preferably
average-annual, or average long-term, discharge).
Assumptions
The authors have made several assumptions regarding ground-water flow to wells in the
Appalachian karst. Failure of these assumptions will lead to errors in delineating capture zones.
For that reason, the RTs are designed to overestimate somewhat, capture-zone areas. The
assumptions are:
1. The boundaries of a spring's capture zone are not distorted significantly by the pumping of a
well. A high-discharge well located near the headward boundary (that is, the boundary
farthest along strike from the spring) of a spring's capture zone is likely to distort that
boundary. However, the scale of the distortion is much smaller than that of the error in
estimating the position of the boundary and can usually be ignored. If desired, it is possible
to adjust for this distortion by modifying the boundary as discussed in "Example with an
alternative well position", below. A well located near the strike-line boundary of a spring's
capture zone is not likely to distort that boundary.
2. Aquifer tests in wells that do not intersect a conduit have response curves rather similar to
those for porous media. Although it might seem possible to use standard laminar-flow well
equations to estimate cones of depression, the great heterogeneity of karst aquifers and the
uncertainty of the hydraulic parameters make this approach unreliable.
3. In wells that tap conduits, pumping causes no reversal of gradient, and so the capture zone
for the well will be essentially the same as it would be for a spring located at that same
point.
4. Where anisotropy is likely (see Table 2), the cone of depression for the well will be
elongate. The reader can use the following guidelines:
If a well is sited in a gently dipping, dominantly bedded Appalachian karst aquifer, then
the aquifer is likely to behave in an isotropic manner, and the cone of depression will
not be significantly elongate.
22
-------
If a well is sited in a steeply dipping aquifer, regardless of whether it is dominantly
bedded or dominantly fractured, then Kmax is approximately along the strike direction,
and
K
max
d
max
dmax /dmin
J -3 J
umax ~ J umin>
where dmax and d^ are the lengths of the major and minor axes, respectively, of the
cone of depression. Thus, when Kmax ^ 10 K^^ the ratio of the major and minor axes
of the cone is approximately 3:1 and the cone is an ellipse.
Although there is a gradation between them, the transition between the Kmax and K^
ratios of about 1:1 and about 1:10 is rather sharp in the Appalachian karst.
Rule of Thumb 2.1: Steeply dipping aquifers (either dominantly fractured or dominantly
bedded)
Step A: Determine the size of the well's capture zone:
Equation 2 is used to estimate the size of the contribution area needed to supply a well over the
long term:
Equation (2):
average annual well discharge rate (Iength3/time)
ground-water recharge rate (length/time)
= size of well capture area (length2)
This equation is valid only over long periods of measurement, because of fluctuations in
recharge rate and in well discharge rate. However, even for shorter periods, the equation tends
to give a reasonable first approximation of the size of the capture zone. The estimated recharge
rate can likely be provided by the USGS and/or state geological survey.
Step B: Define the capture-zone ellipse:
Once the size of the capture zone is estimated, the Kmax and Kmin ratios (see Table 2) are used
to define the capture-zone ellipse (see assumption 4 above); in doing so, it is assumed that the
well is not pumping from a conduit.
23
-------
Step C: Determine separately, and then superimpose, the capture zones for both of the
following scenarios:
Scenario 1: The well draws from a fissured part of the aquifer, rather than from a
conduit. The direction of Kmax approximately coincides with the strike direction. The
capture-zone ellipse, determined in the previous Step, is drawn on a map. The ellipse is
oriented such that its major axis is parallel to the direction of Kmax. The well is located
essentially in the center of the ellipse. (Figure 10 depicts the capture zone of a well,
located far enough from the headward boundary of the spring's capture zone that there is
no interference, in a steeply dipping aquifer.) Kmax ^10 Kmin, regardless of whether
the aquifer is dominantly fractured or dominantly bedded. In dominantly fractured
settings, the major fracture set that accounts for Kmax is roughly parallel to the strike of
the beds (Deike, 1969). The direction of Kj^ is determined by the minor fracture set
perpendicular to Kmax.
Although the authors wish to avoid the use of a fixed radius, they suggested for added
protection, that the minimum capture-zone radius be no less than EPA's informal
suggestion of 0.5 mile for fractured-bedrock aquifers.
Scenario 2: The well dominantly withdraws from a conduit. Solution conduits are long
and narrow, and some tributaries are likely to extend to the updip boundary of the
spring's capture zone. However, for practical purposes, the location of individual
conduits is considered to be unknown.
To estimate the approximate capture zone that feeds the conduit, follow these steps:
(1) Estimate the extent of the spring capture zone in which the well is located
(useRT 1.1 or 1.2).
(2) Estimate the size and shape of the cone of depression, as shown above.
(3) From the edge of this cone closest to the spring, draw a line in the updip
direction as far as the edge of the capture zone defined for the spring
(Figure 10).
(4) From the downdip edge of the cone, draw a second line parallel to the
strike, in the direction away from the spring, to the far end of the spring's
capture zone. The probabilities that flow will deviate from the dip and strike
directions, as described in Part I, can be taken into account in drawing these
lines. However, this deviation likely will not cause significant errors in the
capture-zone boundary and an adjustment is probably unnecessary,
particularly if EPA's informally suggested 0.5-mile radius circle is used. If
the 0.5-mile radius circle extends further downdip than the calculated cone,
instead draw the line from the downdip edge of the circle. In addition, add
any adjacent areas draining to contained recharge features.
Thus, the well's potential capture zone includes that portion of the spring's capture zone
that lies both updip and upgradient along the strike from the well. There is little chance
of recharge to the well from beyond the cone of depression in either the downdip
24
-------
direction or in the direction of the spring. (Considering that the capture-zone boundary
is an estimate, the authors suggest, in order to be protective, that if the cone of
depression or the 0.5-mile radius circle extends beyond the strike line, the capture zone
includes all of the cone or circle, respectively, that does not extend beyond a perennial
Plan View Steeply Dipping Dominantly Bedded or
Steeply Dipping Dominantly Fractured
Direction of strike
and Kma:K
: Fit distort^!*' eautn*>J try
projecting a circJe OT ellipw onto
BID vralnr lafita, is nal shawm
®Sping
Boundary of Spring's
Ne'e: the dawndip width of this area
equals 50% of the dawndip width of the
calculated capture zone of (he well.
It is added in dominantly fractured settings.
/ Updlfi Pro|ec*on
o
L*
^ N RcLharpc fcalurea
LS
Qipi'Slnke
S Well
Rartui-DrcJg
SlHp-e-CaplurB-2ana Elkpsa
Well Capture Zone
Figure 10. Capture zone of a well, located far enough from the headward boundary of the spring's capture zone that
there is no interference, in a steeply dipping aquifer, regardless of whether the aquifer is dominantly fractured or
dominantly bedded.
25
-------
stream. If both cross the strike line, the authors suggest including all of the cone
rather than all of the circle [see Figures 12 and 13, below].) Any land areas draining to
the contained recharge features are added to the well's capture zone.
Consider now that the well in Figure 10 is instead located near the headward end of the spring's
capture zone (Figure 11). As in Figure 10, the well is sited in a steeply dipping karst aquifer.
Plan View Steeply Dipping Oominantly Bedded or
Steeply Dipping Dominantly Fractured
Direction ol dip
and Kmin
, ihe distortion -caused by
projecting a car-tie or eJlip-se owo
Die w,-jlr.r tahte, is not shown
Spring
eotwtfary of Spring's
Unslne&wd Capture Zone
***" Recharge Features-
Note: the efowndlp width of this area
equals 50% of the downdip w»dlh of the
calculated cap4ure zoo& of the well.
It is added in dominantly fractured settings.
Up<*p Prajocflon
i Cone
1 -Ajti*ary-Fajcid
RaOus Orcte
L_S
'.Vftll
o
f^ y Slnp.e-Cap!urfl-2flna hll pu:
Well Capture Zans
Figure 11. Capture zone of a well, located near the headward boundary of the spring's capture zone, in a
steeply dipping aquifer, regardless of whether the aquifer is dominantly fractured or dominantly bedded.
26
-------
The well can draw water from outside the spring's capture zone if the cone of depression
extends beyond it, although the additional area is rarely significant, especially in view of the
uncertain nature of the spring's capture-zone boundary. In this case, the capture zone for the
well is estimated simply by drawing lines updip from the two ends of the cone nearest and
farthest from the spring (Figure 11). These lines terminate at the updip boundary of the spring's
estimated capture zone. Any land areas draining to the contained recharge features are added to
the well's capture zone. Note that, in this example, because the well was located near the
boundary of the spring's capture zone, the width of the well's capture zone is relatively narrow.
Step D: In dominantly fractured settings, modify the capture zone by adding a strip in the
downdip direction about half the dip-direction width of the original well capture zone. In
addition, any adjacent areas draining to contained recharge features should be added. In
expanding the boundaries in these two ways, no perennial surface streams should be crossed.
The resulting area is likely to be larger than necessary to account for the well's discharge, but
this simply reflects the greater uncertainty in dominantly fractured settings.
Rule of Thumb 2.2: Gently dipping, dominantly bedded aquifers
The same approach as that described above is used in this setting. However, here, because
[K-max ~ Kmjn], the ellipse above collapses into a circle. Its area is determined as described
above. The well is located at the center of the circle (Figure 12). In this example, the well is
sited near the headward end of the spring's estimated capture zone. Figure 13 shows a similar
situation, except that the well is not located near the headward end of the spring's capture zone.
Rule of Thumb 2.3: Gently dipping, dominantly fractured aquifers
In this setting, the same approach is used as that described in RT 2.1, above. However, here
K-max ~ 10 Kmin, and Kmax is in the dip direction. Figure 14 shows a well distant from the
headward end of the spring's estimated capture zone. (In this example, the spring's capture-
zone boundary abuts a cluster of recharge features. The spring's capture-zone boundary has
already been adjusted to include the surface area draining to these features; thus, the well's
capture zone includes this area, too.)
27
-------
I Gently Dipping
N Dominantly Bedded
Plan View
'e Z-3i"«
fhfl ilF,ii>1irn fjS
ng a cirr.A r
ti aa:c;r lanln is nnl
SSei W «I)-1 - Arfci ilf-Sf >'-Fixt«J
o
Wei :japr,ira I
Figure 12. Capture zone of a well, located near the headward boundary of the spring's capture zone, in a gently
dipping, dominantly bedded aquifer.
28
-------
Gently Dipping
N Dominantly Bedded
Plan View
\
\
Note: trie alsioflon caused
t?y prqjBrJmg ft cirde cnlo
lha wstar lahlR ft nrrt shrnvr
Spring
.' Updip Prajectian
-'•'JT Ctll'f
o
Spnng'c
Unstraxscd Coplirc Zanc
Frwturns
Well
H^liUS Circle (Vi
Wei Capture Zone
Figure 13. Capture zone of a well, located far enough from the headward boundary of the spring's capture zone that
there is no interference, in a gently dipping, dominantly bedded aquifer.
29
-------
Gently
Dominantly Fractured
Plan View
Direction of slr*e.
minor fracture SB(
and
lhe.downaii> width or this
area equals 50% of the downdip
widih or ihft calculated
zone of the well.
Direction
major Fracture set
and
Note: Iht dnlalion cauHeO \
ay proo:6ro a :r:c-onlc
1he -.v-ilri Nl !>• i
o
S car* kt-1-Art n Ira ry-
Knrlur. CirclE I'.j iril:
Bcundarv cri Spnng s
Rfn:harqn Fftah.fHS
Li LitiKulum
S^
Figure 14. Capture zone of a well, located far enough from the headward boundary of the spring's capture zone that
there is no interference, in a gently dipping, dominantly fractured aquifer.
30
-------
List of References
Biondic, Bozidar, January 8, 2002, electronic communication, Institute of Geology, Croatian
Geological Survey.
Brahana, I Y, John Thrailkill, Tom Freeman, and W. C. Ward, 1988, Chapter 38 Carbonate
Rocks, The Geology of North America, Vol. O-2, Hydrogeology, Geological Society of
America, pages 333-352, in Hydrogeology, William Back. Joseph S. Rosenshein and
Paul R. Seaber, eds., 1988, Geological Society of America.
Deike, R. G., 1969, Relations of Jointing to Orientation of Solution Cavities in Limestones of
Central Pennsylvania, American Journal of Science, v. 267, pages 1230-1248.
Dingman, S. Lawrence, 1993, Physical Hydrology, Prentice-Hall, Inc., Englewood Cliffs, NJ.
Jones, William, K., 1997, Karst Hydrology Atlas of West Virginia, Karst Waters Institute,
Special Publication 4, pages 90-91.
McDowell, Robert, C., George J. Grabowski Jr. and Samuel L. Moore, 1981, Geologic Map of
Kentucky, US Geological Survey in cooperation with the 10th Kentucky Geological Survey and
the llth Kentucky Geological Survey, map, l:250,000-scale, sheet 1 of 4.
Palmer, A. N, 1999a, A statistical evaluation of the structural influence on solution-conduit
patterns, Karst Waters Institute, Charles Town, WV, Special Publication 5, pages 187-195.
Palmer, A. N, 1999b, Anisotropy in carbonate aquifers, Karst Waters Institute, Charles Town,
WV, Special Publication 5, pages 223-227.
Palmer, A. N, 1986, Prediction of contaminant paths in karst aquifers: Proceedings of First
Conference on Environmental Problems in Karst Terranes and Their Solutions, Dublin, Ohio,
National Water Well Association, pages 32-53.
Price, Paul H. and E. T Heck, 1937, Greenbrier County showing general and economic
geology, West Virginia Geological Survey, map, l:62,500-scale, east sheet (Map II).
Rantz, S. E., and others, 1982, Measurement and computation of streamflow; Volume 1.
Measurement of stage and discharge; Volume 2. Computation of discharge;
US Geological Survey Water Supply Paper 2175.
Quinlan, James F, and Ralph O. Ewers, 1989, Subsurface Drainage in the Mammoth Cave
Area, in Karst Hydrology Concepts from the Mammoth Cave Area, William B. White and
Elizabeth L. White, eds., 1989, Van Nostrand Reinhold, NY NY, pages 76, 78, 85.
31
-------
Quinlan, James E, and R. A. Ray, 1981, Groundwater basins in the Mammoth Cave Region,
Kentucky, Friends of the Karst Occasional Publication Number 1, map (Plate 1 in Karst
Hydrology Concepts from the Mammoth Cave Area, William B. White and Elizabeth L. White,
eds., 1989, Van Nostrand Reinhold, NY NY) .
US Environmental Protection Agency, 1997, State Source Water Assessment and Protection
Programs Guidance, Final Guidance, EPA 816-R-97-009, 152 pages.
US Geological Survey, 1986, Water Fact Sheet Hydrologic Hazards in Karst Terrain, Open-File
Report 85-677, 2 pages.
US Geological Survey, with Tennessee Valley Authority, 1986, Campbellsville Kentucky
l:100,000-scale metric topographic base, map.
US Geological Survey, with Tennessee Valley Authority, 1985, Bowling Green Kentucky -
Tennessee, l:100,000-scale metric topographic base, map.
White, William B. and Elizabeth L. White, eds.,1989, Karst Hydrology Concepts from the
Mammoth Cave Area, Van Nostrand Reinhold, NY NY, 346 pages.
32
-------
Appendix 1
W&TitR
T SHEET
U.S. GEOLOGICAL SURVEY. DEPARTMENT OF THE INTERIOR
HYDROLOGIC HAZARDS
IN KARST TERRAIN
Fi|jurB% im 'Im Fit I Slwvl mtr lwv-rVn«i»HH-ul ami <|i|wir Hi thiM rw*.> jnd MjrtlOll miterul tlcdting cm ejh* «{9
ThiV vim ilirw h> *ni|ihjuip lh*" runn-pl i^l in i.iprn flnw vyU^ni. In nmlity. the Hhird irnenHWl woultf th£H* rh* imtWful
.-ill Itflinv Iliv idivr vSrfjir--. fci ]»• cixinw IH:! ir |:-.i: r •.
Wrt*.T IS KARST3
tenm IWPT. ta a
!. dokm«?e, cf gvpium trv'
bv rait» ami underground wKerr ana a char*ctenz«l bv closed
ni. or lirikholn., IWJ underjjr^und
jubsiirfjce for paths Jire enl*g«f ow^r tinw. water mi,ivfi
in rhr *qu«lfr channel chwarter m one wtH-nr gn.mnrf-»»!er
flwii wai mit«illv Irtnivjn jnrnill. sciH?r«) tjpenin([j. In Ih* rwk
i -he fkiw path* trmtwiue to enUrj^. c. ji-^ mwf bv
lonr»ed and ih* jrouixiiwitmr lable nnty drop- hfkiw rtie l*w?l
of Hirfaee Jlwamv wrfscs itrcieij BH^ ihen br(in u» kne *«ier
1o Ihe MilwuriBce. A), moie c^ tT« wrfmis w^er i> dnrrnr^
undcqnTund. wrfice streams, ind Mrwm wrftys bctxinw • In?
coriipicunui feriure of the l»mj iwKm, ind 4rv «?pi*c«i by
tlosrd bninj. Funnels of circular depression cill«di i«i!i;kjl«i
tiftm d««*np it Hjrne place* in Ihc kr* points of 1K«* rktwd
tusins.
SINKItOU rlOQMNC
irt xinklnjBTi Af •HIFnK* MtUlT IPfrm f UMT1H li t ftta
at \hv ruftirri hydiukiijii >y%>in in kirtl l#gi£in(. PlMUfiAg
. KI urs iluiinj pciHxdt putMT.-. ?hr driinjRF. irjpjcltVOl slnUlolAt, Wdlnr
in MtoVr^nmnd conduits cm bac4 up inm (InMtol^ whvn. the
rale of slrjrm water ir*owp ocBSds the dfiinig* LJ|M iiv cit «ihfT
lUt ctye «trmm or Hie rerrtving. surtiiru ^irpun.
Urtun nruciurM are dhen bu* rn-ii -Mnthtilnn when? fading
problems mjy be jggut^ud tn- 11; im -H»-j^r r«tn- "j* rurxjtl trom
Iti; £j; un^d jrius i.i| det'eajed s&K
sinfe.lnl
iir.kh.:ilL- ji ji-iri^ jn
!IV llf!!im iTTEl %lll
HOW DO SI**ltr»tE« DCVUOPI
Sinkhole* develop as i ttiuti cil ihu L i Jl^i y? erf wrlne v netr-
tilttiCf (TiiliifilJ Tn«r» AM twci hii)ic lyp^j. ut
AACivl ObtftBUH-jts LVtii List L:L.S•«,:: i
r*K«:ii ::«jst L HY »«• n WATER TABLE
33
-------
*: ID limestone coll»ps« and QJ overiuirrten
into underlying limeslone cwntn. LJnesnne collapiei
Itrnerally occur due la the enlargement ol cave passages in
iriei.lone. ihe enlargement cum* the roan above die puugn
In weaken and eventually collapse to create sinkholei Over-
burden collapse: are more- common lhan limestone cullifti«-i
jrd genera > ccso : ir cn*f burden ilumolng inio opariing& 01
cavtlMf i.-i ihe iindcnying 1 1 mm ore.
In arctt where ilu MOM* latik- it uiuary ate** Ac owesbuidm-
brwi rone conuci, ixtllafuHiv oli*n ixcm wlwi ih* w^irr
dropt MM* 1h« uiTTlnjidiMi-lirrwv1m-»- :.imlji I. m1hH>r
during hfh-w.ilum? (Hjrnnin|g Phyv ii.-jlly I
i» i.H'iH. HI-HI *.I%IH| by lo» *jf bLKjyjn< wppgrl
M h*-s imnr i:fTHrciin|(> in Ihf l
fti l;?y tkiuuj'iiriK cil urfiiridt*d tnvrfairfan ckwyn t
gimj. Ihf
(h* lirtd (uriJC
- ttu-i cnvr
!j||:M[«s whwE »he
jnd Ipm) >rw ^Kurqppv lhj|l
iiiH «ni! iir^HM^ nilmntiK mi-y
l im,i.w dvnltf Into Itx- ^u
horn HnkhdM and dluppMfiAii ^i^pnij. In many
inip*Y si.irf+i.r itrearni thai,
. ft» ihnwgh vAiuntace conduin «i nnp-
ptar Ji ipvingi wlwrn ihcy b«c«iw surface urtdm* iiin
CKl«. 1TC prtSTWUl pfDbttflH *1 kjr jl -fer-
rjjnh. L~.iml«niin.)nt5 j-,VTi.ijl,^d >irfl urtun HOfm H-J[»T runoff.
tut I- jv li-^il. i rir:iirium. IK I and pease, and bacMM htjm pet-
iniii! j| wj\iHh. in.ii- i)? t tKreal to people -using wai*r wt^plwi
in k»!^1 iHn.Him ind In c-jvp jquutK IllE. Oye-Erii rtg IK hniquei
luiv iKiwn ilwt wpuc *mk ertluent cafl iravcf Ai^uf{h ihc Mim
«H!H iKii trv ih^rKtvnvtic of n*c«1 karst t« rains Inio th*> j:|inhT
inri (o !prmjgi in only 4 fev* hours. Water samptas, CDllHird «1
MM-if '•firmni Ji:Jlr>win[j heavy 'a Mil contain tUCHtiii Iliil jjrei-
FH Hnrrvrij! f mjm hunun wast», indlcatung ItUI rm hirgr
lyjlvrr. eHluemi fronn septic lank dram firti*, min the
imeilone aquifer.
CnnUnuniAiof* pfob*em5 are iggravned In Laric inruim by
Int- ( nmmnn pr«i.lkv of dnpO!ing of 1O«I VHJ liquid n-ilWS in
tiiikhhilnt wltHri- iK«ry m»y br wajhcd direct^1 irrto the aquif*-!.
The JrrL-lijpnpf nr anil wkiV»nri»d iwr vf hazardous mMcrjIi
tin riKfeawd ih* Ar«di fiorn ihi* prmlkt
Leaks, spilli, or ddibtraie dumping ij H^k -or
chemlrali- are a pantcularly ttncujt hiijrd tor k^rii
Chcrmcil] iKit ic»« r^oTi turletf tanlu nuy b* CM,
>ng. the>- may become highly {cmcMiirMNd in the
mtuphere and rise through IractunK in ih*av*rlyirig liir
Hi errler inhibted ftTuctumon the turtacc. Occaciondy httrm^
i -i urlMii trvf, miiHh hr n'4CVMed because tlMKI *1 baae-nMTH
The- degiH at CDnianiiiuiiun irf vh>lkw aquifers in karfl rer-
rain depend! primarily Upon whf-lht • -|TH*K ^qui^r) rfl-rm;
distnbubed or conccncraud r*cfvji^*.. jnd u(Hsn iN> pri>ximit>
and types of sourcesat conuminaiioi'i 5pnrtg« MI-K! wninr V%H||S
i«i Sarjf lerr*n. if supplied entirely by dluAiiud rMti*>g.*
ihick overburden or pi ni.Hfliilc W.J|FT HovMwr, rnvry sprtngs and water welli
in fc^ryt iHri-»«n rtv nn-p concentnAcd fechirsje from! a nea'by area
where route* of t nntarnnnttnr are present.
For addrtnxul inrormathon write w.
U.S.
, Itoom
Pnrfetior Hcholat C.
Cemer tar Cave and
We*«:rn Kentucky UViii.*r ijty
CrMn, KflfttuL-ky *31Ot
Dpon F«Je fteport
34
-------
Appendix 2: Clarification of karst flow patterns
This appendix is designed to clarify the patterns of flow within the spring capture zones
described in Part I. The description here is simplified to show the basic premise behind the
Rules of Thumb.
A cross section (Figure 15) is shown through a typical capture zone, viewed in the direction of
the strike (that is, strike is into/out of the page). The main solution conduit (A) that drains to
the spring is shown, and is assumed to lie at the same elevation as the spring.
H
B
Water
Figure 15. Cross section through a typical capture zone, showing the basic premise behind the Rules of Thumb.
(View is in the direction of the strike, that is, strike is into/out of the page.)
The capture zone for the spring includes the region updip from the strike-oriented conduit. A
line is drawn up the dip from the strike line (that is, the conduit) until it intersects the surface
(point B). Within zone C (extending between A and B), vadose water (in the unsaturated zone)
drains essentially down the dip, cutting across the beds in places where it follows steep paths
along fractures. The pattern of flow is shown by the descending arrows. Major flows (such as
D) form solution conduits tributary to the strike-oriented conduit. Minor seepage (E) through
fissured parts of the aquifer may reach the water table and follow the hydraulic gradient into
conduits (mainly the strike-oriented one).
Any areas of surface runoff that drain into zone C are included in the spring's capture zone (not
shown).
Zone F lies beyond the estimated capture zone in the updip direction. Infiltration into this zone
is limited. Vadose seepage drains more or less down the dip, reaches the water table, and flows
down the hydraulic gradient into the stream valley to the left.
35
-------
By jogging downward along fractures, much of the vadose water that infiltrates within zone C
reaches the water table and drains toward the right into the strike-oriented conduit (especially in
dominantly fractured aquifers). However, some vadose water may pass above the
strike-oriented conduit as downdip flow (G). This water reaches the water table to the right of
the conduit but then flows against the dip to the conduit. The delineation of the spring's capture
zone remains valid.
However, in dominantly fractured rocks, some water is able to infiltrate to the right of zone C
and still reach the conduit by flowing against the dip (zone H). Field experience suggests that
the width of zone H is typically less than half the width of zone C. A width equal to 50% of the
width of zone C is suggested to give an ample safety margin. This additional strip will increase
the area of the estimated capture zone beyond what is needed to supply the spring's discharge,
but it is appropriate to do so, in view of the uncertainty of flow paths in highly
fractured karst aquifers.
Piezometric measurements in wells can help to delineate the water-table divides. However, this
information does not take into consideration the lateral-flow component of vadose flow. Vadose
flow commonly crosses water-table divides, especially in dominantly bedded aquifers.
36
-------
Appendix 3: Simple method for measuring spring discharge
Introduction
The goal of this appendix is to provide a simple method for estimating spring discharge rate, a
critical parameter value needed in the Rules of Thumb described in the preceding sections.
Numerous hydrology texts discuss standard methods for obtaining this value. The
US Geological Survey (USGS), among others, has published considerable information on the
subject, much of it available at the USGS website www.usgs.gov.
The method below is a simplified (though less accurate) method for obtaining spring discharge,
incorporating information from Rantz, et al (1982) and Dingman (1993). The accuracy of the
method should be sufficient for the purposes of the Source Water Assessment and Protection
Program.
Most karst springs in the Appalachians form discrete streams of their own, which flow on the
surface for a short distance to a surface stream. Measuring the discharge in the surface channel
of such springs is not difficult. However, a small number of Appalachian springs discharge
directly into the beds of surface streams; measuring their discharge is difficult and beyond the
scope of this method. For the vast majority of Appalachian springs, the method below is useful.
Water issuing from karst springs typically forms shallow streams that are fairly easy to measure
without danger. Do not attempt to enter streams containing rapidly flowing water or water more
than knee deep.
Resources needed
Users will need: hip boots or chest waders, floats (for example, wood chips), stopwatch, meter
stick or yardstick, tape, string and a half dozen stakes.
Method for estimating spring discharge
Spring discharge should be measured during the period of high base flow in nearby surface
streams. In the southern Appalachians, high base flow typically occurs in February, several
days after a significant precipitation event. In the northern Appalachians it typically occurs in
March, several days after a significant precipitation event.
Spring discharge is determined in the following steps: (1) measure the velocity of the water in
the channel issuing from the spring, (2) measure the area of a cross section through which the
spring discharge flows, and (3) obtain the discharge by multiplying the velocity by the cross-
sectional area.
Velocity of flow in the channel
Velocity varies with depth within a stream and with distance from shore. A commonly used
approximation is that the average velocity of a vertical column of water in a stream is equal to
0.85 times the velocity at the surface (Dingman, 1993). The generalizations in this assumption
37
-------
are acceptable for the purposes of Source Water Protection. Velocity should be measured over a
length of channel approximately 10 times the channel width (Dingman, 1993).
Step 1: Draw a string across the channel, perpendicular to the channel axis, and stake the string
at both ends. The string traverse should be along a straight reach of the channel.
Measure the width of the channel at the traverse. Similarly, draw a second string across
the channel at a distance upstream of the first approximately equal to 5 times the
channel width. Repeat with a third string approximately 5 channel widths downstream
of the first string. (If the channel is not long enough, these distances can be shortened,
with a slight decrease in accuracy.) Measure the length of the stream channel between
the upstream and the downstream traverses. The upstream traverse marks the starting
location for the velocity measurement, and the downstream traverse marks the end.
The middle traverse is where the cross-sectional area is calculated.
Step 2: Along the upstream traverse, gently toss the floats, all at once, onto the stream surface
(Figure 16a). It will likely take a little practice to distribute the floats across the
channel, so the user should have a large supply of floats available.
Figure 16: a) Position of floats at the beginning of a
stream-velocity measurement.
Figure 16: b) Position of floats at the end of a stream-
velocity measurement.
Step 3: Measure the time required for the floats to travel to the downstream traverse
(Figure 16b). The first arrival time is used, because it yields the largest (that is, the
most protective) estimate of the spring capture zone.
38
-------
Step 4: Repeat this travel-time measurement several times. Average the results. Calculate the
resulting velocity by dividing the distance between the upstream and downstream
traverses by the average first-arrival travel time.
Step 5: Determine the average velocity of the most rapid vertical column of water by
multiplying the surface velocity obtained in Step 4 by 0.85. It might appear that Steps
4 and 5 defeat the purpose of using the maximum reading in Step 3. However, this
method assures not only that the spring capture zone is not underestimated, but also that
it is not grossly overestimated. In a smooth, uniform channel with rapid flow, the
resulting discharge estimate will be fairly accurate.
Cross-sectional area of the stream
The stream cross-sectional area is the width multiplied by the average depth.
Step 1: At the middle traverse, divide the stream channel into uniform intervals. For example,
in a 10-foot-wide channel, mark the string at one-foot intervals (Figure 17).
-
Channel WicWt-
1/10
1/10 1MO 1/10 1/10 1MO 1/10 1MO
.
. '
"
„ f • I • ' . S
'
• ^f - •. • • .- • I. •. ;
•
Depth Measuremerl Point
Figure 17: Cross section of a stream, with locations of depth-measurement points.
Step 2: At each mark, measure the stream depth with a thin meter stick or yardstick. Minimize
the standing wave by facing the thin edge of the measuring stick into the current.
Include one zero depth (to represent one edge of the stream - but not both).
Step 3: Average the depths and multiply the result by the width of the stream. This gives the
cross-sectional area of the stream. Be sure to use homogeneous units. For example, if
channel width and length are measured in feet, convert the average stream depth to feet.
39
-------
Calculate the spring discharge with equation 3
Equation 3:
Spring Discharge = (Stream Velocity) x (Stream Cross-sectional Area)
Calculate the size of the spring capture zone
Calculate the size of the spring capture zone with equation 1, in Part I, using an estimate of
ground-water recharge (which can usually be obtained from the USGS or the state geological
survey), and the discharge rate estimated with equation 3.
Hypothetical example:
In this example, an estimate of 0.8 feet/year of ground-water recharge is supplied by the USGS
District Office.
Velocity of flow in the channel:
Stepl: Strings are placed across the channel marking three traverses, as described
above. Channel width at the middle traverse is measured to be 6.0 feet. The
stream-channel distance between the upstream and downstream traverses is
measured to be 60 feet.
Step 2: Floats are released onto the surface of the channel along the upstream traverse.
Step 3: The time elapsed between releasing the floats at the upstream traverse and the
first float arrival at the downstream traverse is measured to be 53 seconds.
Step 4: Repeating Step 3 gives times of 51 seconds and 55 seconds. The average time
is, thus, 53 seconds.
Step 5: Divide the length (60 feet) by the time (53 seconds) to obtain the velocity
(1.13 feet/second). Multiplying the velocity by 0.85 yields 0.96 feet/second.
Cross-sectional area of the stream:
Step 1: The middle string is marked every 0.6 feet and channel depth is measured at
every mark, in this example, 0.0 feet, 1.2 feet, 1.9 feet, 2.2 feet, 2.4 feet, 2.1 feet,
1.8 feet, 1.3 feet, 0.8 feet and 0.6 feet. Note that the zero depth at only one
streambank is included in the measurements.
Step 2: The average stream depth at this traverse is, therefore, (0.0 + 1.2 + 1.9 + 2.2 +
2.4 + 2.1 + 1.8 + 1.3 + 0.8 + 0.6) - 10 = 1.43 feet.
40
-------
Step 3: The cross-sectional area of the stream at the traverse is: 1.43 feet x 6.0 feet
= 8.6 feet2 (rounded off to nearest tenth).
Spring discharge:
As shown in equation 3, the discharge is the adjusted velocity calculated above
(0.96 feet/second) times the cross-sectional area (8.6 feet2) = 8.3 feet-Vsecond. Since
one year (365.25 days) = 31,557,600 seconds, the measured spring discharge is
equivalent to about 262,000,000 feet3/year (rounded off to reflect uncertainty).
Size of spring capture zone:
The size of the spring capture zone is calculated with equation 1, in Part I, using the
recharge estimate of 0.8 feet/year and 27,878,400 feet2/mile2:
262.000.000 feet3/year = 327.500.000 feet2 =11.8 mile2
0.8 feet/year 27,878,400 feet2/mile2
41
------- |