METHODS FOR ESTIMATING NONPOINT
SOURCE CONTAMINATED GROUND WATER
DISCHARGE TO SURFACE WATER
Reference Manual
September 8, 1993
U.S. Environmental Protection Agency
Office of Water
Ground Water Protection Division
Washington, O.C.
-------
TABLE OF CONTENTS
Chapter I. Introduction 1-1
A. Overview of This Reference Manual 1-1
B. Rationale for Conducting Assessments 1 -2
C. Assessment Methods Identified by EPA 1 -3
Chapter II. Theory and Application of First-Tier Methods 2-1
A. Overview 2-1
B. Introduction to First-Tier Methods 2-1
C. Base Flow/Hydrograph Separation Theory tor Estimating Ground Water
Discharge 2-3
1 ..-Base Flow and Streamflow Theorvi 2-3
2.1 Theory of Hydrograph SeparationJ 2-8
3. Using Base Flow/ Hydrograph Separation Theory to Obtain
Estimates of Ground Water Discharge 2-10
4. Data Requirements 2-12
5. Data Resources 2-12
6. Data Evaluation 2-12
7. Applicability of the Method and Accuracy of Results 2-13
D. Ground Water Monitoring Well Theory for Estimating Ground Water
Discharge —. 2-18
[ 1. Darcy'sLaw [ 2-18
2. Data Requirements 2-20
3. Data Resources 2-20
4. Data Evaluation 2-21
5 Applicability of tie Method and Accuracy of Results 2-21
E. Use of Ground Water Wells for Estimating Pollutant Concentrations in
the Discharge 2-23
1. Theory 2-23
2. Data Requirements 2-23
3. Data Resources 2-24
4. Data Evaluation 2-24
5. Applicability of the Method and Accuracy of Results 2-25
Chapter III. Methods 3-1
A. Overview 3-1
B. Hydrograph Separation Method 3-2
C. Water Level Measurements and Darcy's Law 3-16
D. Ground Water Loading Measurements 3-24
Chapter IV. Case Studies 4-1
A. Overview 4-1
B. Case # 1 Base Flow/Hydrograph Separation 4-1
C. Case # 2 Use of Ground Water Monitoring Wells for Estimating
Discharge and Contaminant Loading 4-16
Appendix A. Additional Assessment Methods A-1
Appendix B. Glossary of Terms B-1
Appendix C. Checklists C-1
Appendix D. Reference Graph and Table for Hydraulic Conductivity and Permeability D-1
Appendix E. Research Paper on Methods for Estimating Recharge E-1
-------
CHAPTER I. INTRODUCTION
A. Overview of this Reference Manual
The U.S. Environmental Protection Agency (EPA) Ground Water Protection Division has
prepared this Reference Manual as a resource document to describe two approaches for
estimating contaminated ground water discharge to surface water. This manual is intended
for use by regulatory officials and water quality planners who are interested in assessing
contaminant source loads in water bodies of limited quality. EPA identified the methods
described in this report based on their documented use in the technical literature, their
relatively low cost of application, and their ability to generate data suitable for nonpoint
source management studies.
The manual is the second volume of a larger study completed by EPA. The first volume,
entitled "A Review of Methods for Assessing Nonpoint Source Contaminated Ground-Water
Discharge to Surface Water," {EPA 570/9-91-010) comprises a summary of the technical
literature describing seven general assessment methods. As part of this study, EPA prepared
a detailed annotated bibliography describing over 120 technical papers identified in the
literature. This literature review formed the technical basis for the methods presented in this
manual.
This manual is organized in four chapters. In this first chapter, we describe the rationale for
conducting analyses of nonpoint source load to surface water bodies and briefly present the
various techniques available for assessing contaminated ground water discharge to surface
water. In the second chapter, we present the theory and application of two of these
approaches: base flow/hydrograph separation techniques and ground water monitoring
techniques. We describe the step-by-step methodology for applying these two methods in
Chapter 3. Finally, in Chapter 4, we present case studies that illustrate the applications of
these methods in determining the magnitude of nonpoint source contamination associated
with ground water discharge to surface water bodies.
-------
1-2
B. Rationale for Conducting Assessments of Contaminated Ground Water
Discharge
There is a growing awareness that no/ipoint sources contribute significant contaminant loads
to surface water bodies. In fact, nonpoint sources, such as agriculture or urban runoff,
account for the majority of contaminant loading in many surface water bodies. These
nonpoint source loads can enter surface waters through overland flow, soil moisture
movement, and ground water discharge. This training manual focuses on this third
component of nonpoint source loading, contaminated ground water discharge to surface
water.
Ground water can enter surface waters in the form of springs or bank discharge. According
to the U.S. Geological Survey (USGS), 40 percent of the average annual streamflow
nationwide is from ground water. In humid areas such as the eastern seaboard and north
central region, ground water may contribute as much as 90 percent of streamflow during
some seasons.1 If this ground water is contaminated by disperse nonpoint sources, the
discharging ground water may affect the water quality of receiving waters over a wide area.
The importance of contaminated ground water discharge to surface water bodies has been
recognized in isolated areas, but the nationwide significance of this problem has not yet been
fully documented.
Programs are emerging at the Federal and State levels to begin controlling nonpoint source
loads to surface water bodies. For example, Section 319 of the Clean Water Act required
States, by August 4,1968, to identify those water bodies that are not expected to attain or
maintain their designated uses because of point or nonpoint source loads. Section 319 also
calls for an identification of programs to achieve implementation of best management
practices (BMPs) for controlling the nonpoint source categories identified under the Act.
Currently, many States do not have certified or established procedures for assessing the
magnitude of nonpoint source load to their surface water bodies, particularly for assessing
contaminant load associated with ground water discharge. In addition, many States that are
1 USGS, 1988. National Water Summary 1986 - Hydrologic Events and Ground-Water
Quality. USGS Water-Supply Paper 2325.
-------
1-3
presently assessing nonpoint source loads have not established procedures to determine the
type or degree of controls for managing nonpoint sources needed to stay within the Total
Maximum Daily Load2 limits of receiving water bodies. This training manual was designed to
assist regulatory officials and water quality planners in developing programs to assess and
manage nonpoint source contamination associated with ground water discharge to surface
water.
The methods presented below have been applied in a variety of settings throughout the
United States. They are described here in detail to promote their application in areas that are
experiencing surface water quality degradation due to nonpoint source contamination.
C. Assessment Methods Identified by EPA
In a companion volume to this training manual, EPA presents detailed descriptions of seven
general techniques that have been used to estimate the magnitude of nonpoint source
loading to surface waters through contaminated ground water discharge. These general
methods include:
• Base flow analysis/hydrograph separation, regression analysis, or mass
balance approaches to estimate the contribution of ground water to stream
flow.
• Ground water quality sampling and measurements of ground
water flux to estimate loading of contaminants to surface water.
• The use of seepage meters or mini-piezometers to measure
ground water discharge to surface water through the stream or
lake bed.
• Geophysical techniques (seismic waves/ground water
interactions).
2 Total Maximum Daily Load (TMDL) is defined as the assimilative capacity of a water
body, which is the sum of tie individual Waste Load Allocations (WLAs) for point sources,
Load Allocations (LAs) for nonpoint sources, and natural background (see 40 CFR 130.2 (h)),
plus a safety factor. A Load Allocation is the portion of a receiving water's loading capacity
that is allocated to one of its existing or future nonpoint sources of pollution or to natural
background sources (see 40 CFR 130.2 (f)). In sum, the TMDL should encompass the
contaminant waste loads from both point sources arid nonpoint sources.
-------
1-4
• Numerical models of surface-water/ground water interactions,
• Empirical functions estimating ground water nonpoint source
loading to surface water for various land use types.
• Environmental isotopic methods to estimate the contribution of
ground water to stream flow.
Each of these seven methods may be suitable for different locations, physical settings,
contaminant types, or regulatory/programmatic needs. In addition, these methods require
different amounts of time and resources, and result in the collection of different types of data.
As a result, the goals of the nonpoint source assessment program and available program
resources should be evaluated before identifying the appropriate methods to evaluate
contaminated ground water discharge to surface waters. A number of considerations that
should be evaluated when developing an appropriate assessment approach include: time
and resource constraints; level of data specificity needed to support management decisions;
and ability to characterize temporal and spatial variations in nonpoint source contaminant
loading patterns. The ability of each of the seven methods to address these considerations is
briefly outlined in Figure 1 -1.
In general the methods described above should be used as tools to determine the relative
contributions of contaminant inputs to surface waters from different sources, rather than
absolute numbers. Knowledge of the relative contributions from surface runoff, ground water
discharge, or point discharges will assist planners and officials in prioritizing
protection/remedial activities.
This manual emphasizes two methods, base flow hydrograph separation and ground water
monitoring, that require low to moderate resources. For additional information on the other
assessment techniques, refer to the companion document3, and Appendix A of this manual.
The base flow/hydrograph separation method and the ground water monitoring method
3 U.S. EPA/OGWP, 1990. A review of methods for assessing nonpoint source
contaminated ground-water discharge to surface water. EPA 440/6-90-007.
-------
Figure 1-1
Limitations and Requirements of the Seven Methods for Assessing
Nonpoint Source Contaminated Ground-Water Discharge to Surface Water
1
1
Base Row/Hydrograph
¦1
Darcy's Law
Estimates
¦
Seepage Meter/
Mini-piezometer
II
Numerical
Models
¦
Loading
Functions
Geophysical
Methods
¦I
Isotopic
Methods
'
¦
W;
I Time and
1 Resource
| Requirements
low
PS
tow
high
high
¦
low
high
high
11
Data
Requirements
11
m
moderate
moderate
moderate
high
low
high
high
¦
Spatial
Variation
Characterization
no
¦
no
possibly, with
multiple sample
points
yes
no
yes
no
P
I
1 •
i
i
Temporal
Changes
Characterization
J
yes
*~1
yes
i
|
yes, with
repeated
samples
yes
no
yes
yes
T**.
i
i
!m"
II
Accuracy
of Results
J
moderate
moderate
high
u
high
y
low
moderate
to
i
moderate
to
high
-------
1-6
should be thought of as "first-tier" approaches for providing an overview of the magnitude of
nonpoint source pollution loading within a watershed. These methods may in turn be
supplemented with more resource-intensive techniques to further characterize the watershed
or gather more site-specific information. With the lower level of resources required for the
first-tier methods, there is a corresponding Increase in the variability of the estimate of
contaminant loads that enter surface waters through ground water discharge. These first-tier
approaches should, therefore, only be used to provide rough estimates of the magnitude of
contaminant loading from ground water discharges. In addition, due to the uncertainty and
limitations of the methods, results should be presented as a range, and where possible, more
than one estimation technique should be utilized to confirm the results.
-------
CHAPTER II. THEORY AND APPLICATION OF FIRST-TIER METHODS
A. Overview
In this chapter, we outline the general concept of the first-tier approaches, along with the
theories and limitations of the two methods. A step-by-step description of method
applications is presented in Chapter 3. in addition, several examples are provided in
Chapters 3 and 4 to illustrate the uncertainties that may arise as a result of the varying
hydrologic conditions of watersheds. These examples are provided for illustrative purposes.
Specific approaches in dealing with the uncertainties in estimating contaminated ground water
discharge are left to the user to develop on the basis of the limitations described in this
chapter.
The two selected methods are not necessarily appropriate tor every watershed, and one
particular method may be more appropriate than the other under specific circumstances. It is
therefore important to identify the water quality problems of concern and to define the goals
of the study before starting the assessment process. The specific problems and goals
associated with each particular situation will determine the best approach and methodology
for the study.
B. Introduction to First-Tier Approaches
Both of the first-tier approaches discussed in this chapter have three major steps (Figure 2-1).
The first step consists of estimating the volume or rate of ground water discharge to the
surface water body. The second step involves estimating the pollutant concentration in the
discharging ground water. Sampling ground water wells near the surface water body for
pollutants is presented as a method for estimating the concentration of pollutants in the
ground water. Based on steps one and two, the average loading rate associated with ground
water discharge to the surface water body can be calculated by multiplying the average
ground water discharge rate by the average concentration of contaminants in ground water.
-------
FIGURE 2-1
Overview of First-Tier Approaches
to Estimate Contaminant Loads from
Ground Water to Surface-Water Bodies
WATER QUANTITY x WATER QUALITY = CONSTITUENT LOAD
ESTIMATE GROUND
WATER DISCHARGE
BASE FLOW/
HYDROGRAPH
SEPARATION
ESTIMATE GROUND WATER
CONTAMINANT CONCENTRATIONS
GROUNDWATER
WELL SAMPLING
AND
ANALYSIS
ESTIMATE
TOTAL
CONTAMINANT
LOADING
TO SURFACE
WATER BODY
-------
2-3
Ground water discharge rates to surface water calculated using base flow/hydrograph
separation techniques and ground water well monitoring may not be identical because of the
uncertainties associated with these methods. At best, the discharges and loading rates
calculated using the two methods are expected to be comparable if they are computed for
the same season or time interval.
C. Base Flow/Hydrograph Separation Theory For Estimating Ground Water
Discharge
1. Base Flow and Streamflow Theory
Precipitation entering a watershed may travel to a stream as surface runoff, interflow, and
ground water discharge (see Figure 2-2). The contribution of ground water to a stream or
river is commonly referred to as base flow. In the humid portions of the U.S., base flow is
generally present in a channel during both storm and non-storm conditions since ground
water is continuously discharging to the river. Base flow, on the other hand, is rare for
ephemeral streams and streams in the arid portions of the U.S. where river channels often act
as recharge areas to ground water.
The contribution of soil moisture movement to streamflow is known as interflow. Interflow is
usually negligible except during wet or storm periods.
The amount of water contributed to a stream by each of these processes is reflected in the
shape of the stream hydrograph. A hydrograph is simply a graph of stream discharge
against time for a particular point along the stream. Figure 2-3 presents a typical stream
hydrograph associated with a short duration precipitation event. This individual storm
hydrograph is part of a larger hydrologic record made up of other storm hydrographs
interconnected by streamflow during non-storm periods.
The stream hydrograph can be divided into a number of segments. In Figure 2-3, Segment
A-B represents a period of rising stream water level and discharge in response to the
precipitation event. The steep slope of this rising limb is due to the relatively rapid movement
of surface runoff to the discharge point. Thus, the slope of the rising limb is primarily
-------
FIGURE 2-2. Hydrologic Components Contributing
to Stream Flow
-------
FIGURE 2-3. Typical Storm Hydrograph Associated
with a Short Duration Precipitation Event
B
TIME
WHERE:
Segment A-B represents the rising limb.
B is the maximum total discharge as a result of the precipitation event.
Segment B-C is the recession limb.
Segment C-D is the non-storm period base flow.
-------
2-6
controlled by the rate of surface runoff. The crest of the hydrograph is located at point B and
represents the maximum stream discharge as a result of the precipitation event. This point
may vary depending on the size and configuration of the watershed. Segment B to C shows
the period of decreasing stream discharge and is referred to as the recession limb. The early
part of the recession limb is controlled by the rapid decrease in surface runoff. However, as
time progresses the slower processes of interflow and base flow begin to dominate as the
amount of surface runoff continues to decrease. Extending the line beyond point C to the
next storm hydrograph represents base flow during non-storm periods.
Figure 2-4 graphically presents the relative contributions of surface runoff and ground water
discharge to an ideal storm hydrograph. Interflow and base flow are not differentiated in this
figure. The total stream flow during the storm event can be calculated from the area under
the hydrograph. In order to estimate the contribution of ground water discharge to the storm
flow, the line P-Q-R must be found. The volume of ground water discharge can then be
estimated by determining the area under the line P-Q-R. The method for finding the line P-Q-
R and estimating the volume of discharge is called the hydrograph separatfon technique. The
general theory of this technique is presented in the following section, and the step-by-step
methodology is discussed in Chapter 3.
The peak discharge and total volume of streamfiow during a precipitation event also depends
on the soil moisture conditions existing prior to the event (i.e., antecedent soil moisture
conditions). For precipitation events of a given size, the flow processes (i.e., surface runoff,
interflow, and base flow) will deliver a larger volume of storm flow to the channel more rapidly
under wet antecedent soil moisture conditions than under dry antecedent conditions, thereby
increasing both the volume and rate of stream discharge. Surface runoff is greater during wet
periods because of the reductions in both soil infiltration rates and surface depressional
storage. Interflow increases during wet periods because the soil matrix is more conducive to
soil water movement. Base flow also increases because elevated water tables may create
steeper hydraulic gradients, thus promoting water movement in aquifers and because there
will be larger saturated areas from which ground water can discharge into the channel.
Wet periods are generally seasonal, and are usually caused by variations in precipitation and
-------
FIGURE 2-4.
The Relative Contributions of Surface
Runoff and Ground Water Discharge
for an Ideal Storm Hydrograph
i /»,•<
TIME
-------
2-8
evapotranspiration. Figure 2-5 shows the hydrologic record at a gaging station along the
Spokane River for a four year period. Higher peak flows and base flows are observed for this
river during the winter and spring months. This increase in stream flow is probably caused by
an increase in snow melt. It is therefore critical to take seasonal variation into consideration
when studying streamflow and base flow at a particular locality. In order to assess ground
water and surface water relationships during both dry and wet periods it may be necessary to
collect stream flow measurements over as much as six months.
The simple hydrograph separation theory may not be applicable where snow melt contributes
significantly to streamflow. Under such circumstances, base flow estimates for periods with
significant snow melt should be made from empirical estimates of the percentage of base flow
in streamflow. These empirical estimates are regional in nature. Local USGS offices should
be consulted to obtain such estimates for the stream or region of interest. Care must be
taken with this approach. Base flow processes are much slower than surface runoff
processes, and estimating base flow as a percent of streamflow can be misleading.
Estimates using empirical estimates of base flow are not time specific. Thus, base flow ratios
should only be used to estimate the total volume of base flow for the entire snow melt period.
2. Theory of Hvdroaraph Separation
The general procedure for the hydrograph separation technique consists of three parts:
• Locating the rising limb base flow line and the base flow recession line (i.e.,
line segments P to Q and Q to R in Figure 2-4, respectively);
• Finding the slope of the lines segments and developing regression equations;
and
• Integrating the regression equations for the line segments to obtain estimates
of the area under the lines (i.e., base flow volume).
The regression equations developed for the line segments can also be used to estimate
ground water discharge at any point in time. For example, the shaded area between times T
and U in Figure 2-4 represents the volume of ground water discharge for the storm estimated
by integrating the regression equations, while the discharge rate associated with point S
-------
FIGURE 2-5
40 t
Four Year Hydrograph of Spokane River
Gaging Station at Spokane, Washington
30
o *8
*§ S
20
w o
G {5
10
HV01/85 03/30/86 09/26/86 03/25/87 09/21/87 03/19/88 09/15/88 03/14/89 09/10/89
12/30/85 06/28/86 12/25/86 06/23/87 12/20/87 06/17/88 12/M/88 06/12/89
October 1,1985 to September 30,1989
-------
2-10
represents an estimate at a time that can be calculated from the regression equations.
Drawing the stream hydrograph on semilogarithmic paper (stream discharge on the
semilogarithmic axis) will aid in identifying the slopes of the line segments. When drawn on
semilogarithmic paper, an ideal hydrograph will have three identifiable line segments of
different slopes along the recession limb as is shown in Figure 2-6. The line segment U-V
has the steepest slope and represents the rapid depletion of surface runoff. When surface
runotf storage is depleted, the slope of the recession curve flattens slightly. Line segment V-
W represents the contribution of interflow and base flow to streamflow. After interflow storage
is depleted, the slope of the recession curve continues to flatten. Line segment W-X
represents the contribution of ground water to streamflow. The slope of line segment W-X is
known as the ground water or base flow recession constant for the watershed located above
the discharge point. If line segment W-X, representing base flow, is extended back to time Y,
which is located directly under the crest of the hydrograph, the corresponding maximum base
flow rate as a result of the precipitation event can be estimated. The line segment
representing the rising base flow {i.e., line Z-Y) can be established by connecting point Y to a
point (i.e., point Z) representing base flow prior to the runoff event.
When performing hydrograph separation, one should understand the inherent uncertainties.
For example, hydrographs can vary greatly due to seasonal variations in weather conditions,
watershed size, and watershed characteristics. The slopes representing base flow, interflow,
and overland flow depend upon surface and soil characteristics, the magnitude of
precipitation, and the location of precipitation in a watershed. Recession limbs will vary for
different storm events occurring in large basins having geologic variability, or when rainfall
that produces runoff occurs over only a portion of a watershed. In addition, recession limbs
may distort when multiple storm events occur over a short time period. These limitations will
be discussed tn greater detail in later sections.
3. Ustna Base Flow/Hvdroaraph Separation Theory to Obtain Estimates of Ground
Water Discharge
There are a number of potential uses of ground water discharge data derived Irom the base
flow/hydrograph separation theory. One is to aid in determining waste load allocations for
-------
FIGURE 2-6.
Semilogarithmic Stream Hydrograph
/ 6°° Oirb**
WHERE:
Segment U-V represents recession kimb contributions of
dream flow, interflow and base flow.
Segment V-W represents contributions of interflow and base flow.
Segment W-X represents contributions as a result of
ground water storage depletion
Y is the maximum base flow discharge as a result of the precipitation event.
Line Z-Y represents the rising base fbw.
-------
2-12
permitting activities. Waste load allocations are generally completed for iow flow periods of
the stream discharge record. Base flow estimated from hydrologic records during non-storm
periods can be used to estimate ground water discharge during low flow. For example, the
discharge associated with point D on the hydrograph shown in Figure 2-3 can be used to
represent low flow. However, the entire annual streamflow record should be reviewed to
make sure that the base flow estimated reflects low flow periods for the year. Ideally, several
years of record should be evaluated to obtain an average low flow, so that the estimate is not
biased by hydrologic extremes.
4. Data Requirements
Hydrographs should be prepared by plotting stream discharge versus time. Continuous
discharge records, obtained from a chart recorder, should be used to prepare the
hydrograph. The discharge rates are expressed in terms of a volume of water per unit time
(e.g., cubic feet per second [cfs] or liters per second [l/s]).
5. Data Resources
U.S. Geological Survey publications provide the main source of streamflow data in the United
States. Detailed streamflow data, pertaining to the creek, stream, or river of interest, can
usually be obtained by writing to the agency or one of its field offices located throughout the
country. Other Federal agencies such as the U.S. Army Corps of Engineers, U.S. National
Weather Service, and U.S. Soil Conservation Service make occasional observations on stage
and/or flow. Additionally, many States have agencies that collect streamflow measurements.
When existing stations are not available and a new station must be installed USGS general
procedures should be used to set up the new station.4
6. Data Evaluation
The gaging station from which flow data were obtained should be located in proximity to the
study area. This ensures that the estimate of base flow using hydrograph separation
techniques will be representative of actual ground water discharge rates. If flow data are
obtained from a gaging station located upstream of the study area, predicted ground water
4USGS. 1969. Techniques of Water Resources Investigations, Chapter A-6, General
Procedure of Gaging Streams.
-------
2-13
discharge and loading rates for the watershed may be less than the actual rates. Conversely,
if flow data are obtained from a gaging station located downstream of the study area,
predicted ground water discharge and loading rates may be greater than the actual rates.
The relationship between the accuracy of the base flow and loading predictions and the
distance between the gaging station and the study area is a function of the characteristics of
the watershed being studied. The representativeness of the base flow prediction using flow
data from distant gaging stations can be evaluated by installing a temporary gaging station at
the study area. Flow rates at the temporary gaging station can be compared with flow rates
at the distant gaging station. If needed, a permanent gaging station can be installed at the
study area.
To estimate base flow using hydrograph separation techniques, the hydrograph must have
sufficient resolution so that the time of peak discharge and the shape of the recession limb
can be identified. It is possible to identify time of peak discharge and base flow if continuous
discharge records from chart recorder are used to prepare the hydrograph. However, stream
discharge measurements are often collected at twelve hour or greater intervals at gaging
stations. Depending on the size and characteristics of the watershed, and the characteristics
of the precipitation event, it may not be possible to obtain sufficient temporal resolution for
applying the base flow/hydrograph separation technique.
7. Applicability/ of the Method and the Accuracy of the Results
This section addresses the limitations and the magnitude of the uncertainty inherent in the
application of the base flow/hydrograph separation method. While the level of uncertainty will
vary with each application and the quality of the available data, the following discussion
provides an overview of the factors that analysts should consider in using the base
flow/hydrograph separation method to estimate nonpoint source loading rates.
Base flow estimates obtained from non-storm periods may not be appropriate for
estimating discharge over extendsd periods of tlms. As can be seen in Figure 2-4, the
ground water discharge to a stream increases during precipitation events. Base flow
estimated from a non-storm period does not include this "fast" response associated with
ground water discharge from precipitation events. Therefore, determining annual discharges
-------
2-14
using base flow estimated from non-storm periods could underestimate the total base flow
contribution.
Seasonal variations in base flow can raise uncertainties when estimating ground water
discharges over extended time periods. If seasonal variation is significant, base flow
estimates must be collected during the seasons of interest. Hence, estimates of ground
water discharge on an annual basis or other extended time periods wall require discharge
estimates from a combination of storm and dry weather conditions. Evaluation of every day
of an annual record is probably too labor intensive.5 Thus] some type of representative
sampling approach should be developed to select dates for evaluation that wilt reflect both
the seasonal variation and the variation in discharges between storm conditions and dry-
weather conditions.
Furthermore, the hydrograph separation technique is constrained by a limited time
period of the hydrologic record. The time interval between points T and U in Figure 2-4
represents the window of time over which a single estimate using hydrograph separation can
apply, if there is significant variation between ground water discharge during precipitation
events and ground water discharge during non-storm periods, the estimate from the
hydrograph separation may overestimate the average ground water discharge if it is used as
an average rate over an extended time period. However, hydrograph separation can provide
estimates of ground water discharge for specific storms or at specific-times during a storm.
The base flow portion of the recession limb of the semllogarlthmic hydrograph is not
always characterized by a single slope or a straight line. The major factors influencing the
shape of the recession limb are bank storage, aquifer heterogeneity, and reservoir storage.
When the stage of a stream increases, some of the water flowing In the stream may move
into stream banks; and when the stream stage declines, this flow is reversed. The effect of
bank storage on hydrograph shape Is to lower the hydrograph peak and elongate the
recession limb of the hydrograph (Figure 2-7). The recession constant, Kp calculated from a
hydrograph of a stream with significant bank storage, will be less than the true base flow
5 Unless computer graphics and digitizers are available.
-------
FIGURE 2-7.
©
&
(0
o
CO
0
1
s
CO
Time
Bank Storage Effects
Flood hydrograph with no bank storage
Flood hydrograph with bank storage
-------
2-16
recession constant for the stream. Base flow estimates made with a recession constant that
includes bank storage effects will be smaller than the actual base flow of the stream, and
produce underestimated loading rates to the stream.
Streamflow, in both storm conditions and nonstorm conditions, can be significantly
altered by reservoirs. During storm conditions, a reservoir may prevent determination of the
base flow component of the hydrograph if the reservoir temporarily stores a significant
volume of flood water. Temporary storage of flood waters will reduce the peak stream
discharge rate of the precipitation event, and will elongate and increase flow in the recession
limb of the hydrograph. This increased flow in the recession limb will be composed of both
surface runoff, which has been delayed by temporary storage in the reservoir, and base flow.
Since the recession limb in this case reflects both surface runoff and base flow it cannot be
used to estimate base flow. Thus, the base flow/hydrograph separation method is not
applicable where streamflow is highly regulated by reservoirs.
In general, first tier methods are not applicable for fractured geological systems and
karat systems. Base flow during low flow periods is a direct measure of the ground water
discharge to a stream and should reflect the discharge from fractured systems during low
flow periods. However, as discussed in more detail in the section on ground water well
sampling, contaminant concentrations in the ground water of fractured systems cannot be
reliably estimated.
Although base flow/hydrograph separation results are directly applicable to riverine
systems having reliable stage or gage measurement Information, the results have limited
application for lakes. Base flow/hydrograph separation methods can only be applied to the
riverine systems supplying lakes as a means of estimating contaminated ground water
discharge to lakes. The base flow/hydrograph separation method wilt provide an incomplete
picture where lakes receive significant amounts of inflow by seepage through the lake bed.
The base flow/hydrograph separation method estimates base flow contributed by the
entire watershed above the point of Interest In order to calculate ground water discharges
for specific stream reaches, multiple gaging stations will be necessary. Utilization of stations
-------
2-17
above and below the reach of interest will enable the researcher to account for the difference
between inputs and outputs and quantify ground water inputs to the reach of interest. In
addition, base flow at the point of interest can be complicated by upstream withdrawals from
streamflow. Significant withdrawals can take place for irrigation and through recharge zones
from the stream back to ground water systems. If withdrawals for irrigation can be estimated,
the effects of these withdrawals on base flow can be assessed. However, estimating the
effect of recharge zones from streams to ground water requires a high level of data specificity
and more comprehensive methods such as numerical modeling. Thus, the base
flow/hydrograph separation method is generally not appropriate where there are significant
withdrawals from streamflow by losses to ground water.
Small watershed systems generally have a higher day to day variation in discharge than
larger watershed systems. Thus, estimated base flow for small watersheds, approximately
100 square miies or less, using the base flow/hydrograph separation technique will vary to a
greater extent than the larger systems.
The limitations and considerations discussed above affect the utility of the base
flow/hydrograph separation method, and should be evaluated prior to using the method. A
summary list of these considerations is given below:
• applicability to lakes and/or riverine systems;
• effects of bank storage;
e effects of seasonal and annual variations in base flow;
• effects of snow melt;
• effects of recharge zones from the riverine system bade to the ground water
system;
• applicability to fractured and karst hydrogeoiogic systems;
-------
2-18
• effects of upstream withdrawals for irrigation;
• effects of reservoir storage; and
• effects of the size of the watershed.
The existence of some of these conditions may dictate the use of different methods. Other
conditions can be addressed through careful planning and analyses, and through a
reafizatkjn of fhe limitations of ttre method, and the possible variability of the results.
D. Use of Ground Water Monitoring Wells for Estimating Discharge
1. Darcy's Law
Ground water discharge from an aquiler to surface water is controlled by the physical
properties of the aquifer and the distribution of water level across the aquifer. By obtaining
water-levei measurements from monitoring wells and piezometers in a watershed, the
discharge rate of ground water to adjoining surface water can be estimated using Darcy's
Law.
Darcy's Law states that the average velocity of ground water is equal to the product of the
hydraulic conductivity of the aquifer medium and the hydraulic gradient. The discharge rate is
determined by muttfpfying the average velocity by the cross-sectional area of the aquifer
perpendicular to the ground water flow direction. The hydraulic conductivity of an aquifer is a
proportionality constant, commonly expressed in length over time (e.g., cm/sec), which
reflects the ability of the aquifer medium to conduct water. For most natural aquifers, values
of hydraulic conductivity fall In the range of 0.001 to 1 cm/sec. The discharge rate of ground
water is proportional to the hydraulic conductivity of the aquifer.
The hydraulic gradiant between tocatfons /n a watershed is derived by taking the difference in
the height of the water levels found In ground water wells and then dividing the water level
difference by the distance between the wells. In order to obtain a comprehensive picture of
-------
2-19
the water level over the entire watershed, water level measurements from monitoring wells
and piezometers in the watershed are often used to construct a water level map. This water
level map can then be used for the derivation of hydraulic gradients. In natural systems, the
movement of ground water is in the direction of the declining water level elevations, and the
discharge rate is proportional to the steepness of the gradient.
The cross-sectional area of an aquifer is a function of the geometry of the aquifer. A review
of lithologic information from boring logs, well logs, and geophysical surveys are necessary to
determine the basic geometry of the aquifer. In conjunction with information describing flow
direction (i.e., based on the water level map), a cross-sectional area perpendicular to the
ground water flow direction can be determined for the aquifer.
Four major assumptions are made in the application of Darcy's Law to a aquifer discharge
rates. First, the aquifer is assumed to be homogeneous (i.e., physical properties such as
hydraulic conductivity are the same throughout the entire aquifer). Second, the aquifer is
isotropic (i.e., physical properties such as hydraulic conductivity are the same in all directions
within the aquifer). Third, the flow of ground water in the aquifer is primarily one dimensional.
And fourth, the entire cross section of the aquifer perpendicular to the channel is assumed to
discharge to the water body. In general, these assumptions are almost never fully met in
natural systems. Discharge from the entire aquifer may not always occur. Depending on the
regional and local geology some of the water in the aquifer may flow under the water body
without discharging to the surface water. Underflow may be identified by measuring the
hydraulic gradients in the aquifer below the water body. If a downward vertical head gradient
is found at or below the water body, ground water may actually pass beneath the surface
water body in a process referred to as underflow. However, to simplify the method and
present it as a first tier method we will assume that the entire aquifer cross section
contributes to surface water discharge for the following discussion. By assuming the entire
aquifer contributes, the method produces a high end or conservative estimate. Thus, with
careful delineation of aquifers and relaxation of these assumptions, Darcy's Law can be used
to approximate the discharge of ground water from a watershed.
-------
2-20
2. Data Requirements
In applying Darcy's Law to an aquifer, it is necessary to obtain information on the hydraulic
conductivity of the aquifer medium. The hydraulic conductivity can be measured in situ via
field tests (e.g., aquifer test) or estimated on the basis of the predominant rock type of the
aquifer. Field tests can produce the most reliable hydraulic conductivity values, but are costly
and time consuming. In most situations, "typical" values of hydraulic conductivity for various
rock types can be obtained from publications and textbooks. (See Appendix D for a range of
values of hydraulic conductivity for selected geologic materials.) Furthermore, information
from previous field tests can sometimes be found and used in deriving a more reliable
estimate for the hydraulic conductivity of the aquifer.
A water-level map of the aquifer can be generated using water-level measurements from
monitoring wells and piezometers in the watershed. To assure that the water-level
measurements are representative of the actual conditions in a watershed, wells and
piezometers used for the measurements should be evenly distributed throughout the
watershed and be screened at the appropriate depth. In addition, well water level monitoring
data should be obtained over an extended time period and examined for any pronounced
temporal patterns of variability.
Using the water-level map, flow directions and hydraulic gradients can be determined for the
aquifer. In addition, this water-level map can be used, in conjunction with lithologic
information {e.g., wells logs, boring logs, and geophysical surveys), to approximate the cross-
sectional area of the aquifer.
3. Data Resources
Existing monitoring wells and piezometers located within a watershed can be used to gather
water level data. Old test wells and abandoned wells can also be used to provide
supplementary data tor the water level map. In addition, water level data may be obtained
from the U.S. Geological Survey in the form of water level maps, open file reports, and ground
water monitoring data files. When existing wells do not provide reliable information regarding
the aquifer, new wells can be installed to properly characterize the aquifer.
-------
2-21
The Geological Survey is also an excellent source of information on the lithologic and
hydraulic properties of aquifers. Well logs, boring logs, and geophysical surveys are often
available from the Geological Survey, State geological and water offices, and local
engineering companies to supplement readily available information on aquifer geometry and
hydraulic properties of the aquifer.
4. Data Evaluation
Whenever possible, available lithologic and hydraulic properties of the aquifer should be
evaluated in detail. If the hydraulic conductivity of the aquifer exhibits high degrees of
variability, or if the aquifer geometry is highly complicated with impermeable layers or variable
discharge areas, it may be reasonable to select an alternative method to estimate ground
water discharge.
The construction details of existing wells and piezometers must be evaluated to determine
whether they are representative of the aquifer under study. Wells and piezometers screened
at different depths in an aquifer, or in different aquifers, may not provide representative water-
level measurements for ground water discharge estimations.
Measurements of water level must be taken over space and time. If the coverage of water
level data is inadequate (i.e., due to a lack of monitoring wells over a substantial portion of
the watershed), it may not be possible to obtain reliable estimates of hydraulic gradients. It is
also important to examine the locations of pumping wells and their associated pumping rates.
A pumping well that withdraws a substantial quantity of ground water can lower the local
water level and affect water levels of nearby monitoring wells and piezometers. Furthermore,
water-level measurements over an extended time period should be examined to uncover any
temporal variations. K the water level exhibits a pronounced seasonal pattern, it would be
important to select representative conditions for estimating ground water discharge rates. In
general, the frequency of water-level measurements should increase as watershed size
decreases. The shorter residence time of ground water in small watersheds is generally
associated with water levels that are more variable over time.
-------
2-22
5. Applicability of the Method and the Accuracy of the Results
The reliability of using Darcy's Law to estimate ground water discharge is dependent on a
number of factors.
Jn natural systems, the water JeveJ fluctuates according to the seasons, recharge
conditions, and artificial withdrawals. During wet seasons, higher ground water levels are
associated with larger cross-sectional areas and steeper hydraulic gradients, thus promoting
the discharge of ground water at a higher rate. In contrast, lower discharge rates are
common during dry seasons when lower water levels predominate. It is, therefore, essential
to examine the temporal variability of ground water level over an extended period in order to
select representative periods for estimating ground water discharge from a watershed.
The reliability of estimated ground water discharge is also affected by the variability of
hydraulic characteristics or the aquifer and complexity of the aquifer geometry. In
general, the use of water level measurements and Darcy's Law are not applicable to aquifers
that are not homogeneous and isotropic (i.e., heterogeneous and anisotropic). In such
aquifers, ground water flow is highly unpredictable and predicted flow path can differ
significantly from the actual flow path. In these cases, tie derived hydraulic gradient will not
be representative of the actual hydraulic gradient, For these more complex situations,
numerical models are commonly used to provide more reliable estimates, in addition,
because hydraulic conductivity varies spatially and directionally within the aquifer, accurate
estimations for hydraulic conductivity and cross-sectional area may be impossible to
determine.
This method has been used in both marine and fresh water environments, in riverine and lake
systems, and for glacial and dolomite aquifers. However, this method Is not generally
applicable for geologic formations that are heterogeneoue and anisotropic due to
fracturing or channeling.
The limitations and considerations discussed above concern the use of Darcy's Law to
estimate ground water discharge rates to surface water, and should be Investigated prior to
using this method. A summary list of these considerations Includes the following:
-------
2-23
• applicability of assuming aquifer homogeneity;
• applicability of assuming isotropic conditions;
• applicability of assuming one dimensional flow;
• effects of seasonal variations in ground water levels;
• applicability to fractured and karst hydrogeologic systems; and
• effects of withdrawals at pumping wells on ground water levels.
E. Use of Ground Water Wells for Estimating Pollutant Concentrations in the
Discharge
1. Theory
Ground water samples taken from wells and piezometers within a watershed can be used to
characterize the spatial distribution of ground water contaminants and to estimate nonpoint
source contamination load. The infiltration and percolation of precipitation, along with
irrigation water and return flow of domestic wastewater, can transport contaminants to the
ground water. In areas of porous soils and highly fractured bedrock, infiltration and
percolation can be very rapid. When these areas coincide with the presence of nonpoint
source loading (e.g., agricultural areas), concentrations of certain contaminants (e.g., nitrates
and pesticides) may become highly elevated in the ground water. By taking ground water
qualify measurements and multiplying the concentrations by the rate of ground water
discharge, a broad range of contaminant loading rates to surface water through ground water
discharge can be estimated. The wells and piezometers from which ground water
contaminant concentrations are obtained should be screened at the appropriate depth and
distributed to represent the water quality of the discharge to the surface water body.
2. Data Requirements
in order to obtain representative nonpoint source loading rates, it is essential to collect and
-------
2-24
analyze water samples that represent the quality of the water entering the surface water body.
As a result, the wells and piezometers for water quality measurements should be screened at
the depth of the discharging ground water formation and should be located as close to the
surface water body as possible. In addition, if land use characteristics are highly variable
within the watershed, wells should be selected to characterize the various land uses and
associated contaminant loading patterns.
3. Data Resources
Existing monitoring wells and piezometers located within a watershed can be used to gather
water quality data. Old test wells, abandoned wells, drinking water welts, irrigation wells, and
livestock watering wells can be used to provide supplementary water quality data. In
addition, water quality data may also be available from the U.S. Geological Survey in the form
of open file reports and ground water monitoring data files, and from local health departments
in the form of water testing reports. When existing wells are inadequate in providing reliable
information regarding water quality of the aquifer, new wells can be installed to properly
characterize the aquifer.
4. Data Evaluation
If the water quality measurements are concentrated in a small portion of the watershed, the
calculated loading rate for a given constituent may be biased. For example, if water quality
measurements were only obtained from wells located adjacent to feed lots, the predicted
overall water quality for the entire watershed would be worse than the actual overall condition.
Conversely, if water quality measurements were obtained only from wells located in woodland
areas, the predicted overall water quality for the watershed would be better than the actual
overall condition.
The construction history and driller's log of each sample well should be reviewed carefully to
obtain information on screen depth and construction methods. A well constructed without a
proper annular seal may provide a conduit for surface runoff to enter the aquifer. Water
quality measurements from such a well may not be representative of the actual condition of
the watershed. Wells screened at different depths in an aquifer, or in different aquifers, may
not provide representative water quality measurements. If a well is only screened in the lower
-------
2-25
portion of the aquifer, contaminants at the upper portion will not be detected in water samples
from the weJI.
5. Applicability of Method and Accuracy of Results
The reliability of water quality measurements ia affected by the location of the sample
wells. Sample wells located near the discharging area can be used to represent the water
quality of the ground water discharging to surface water. In other cases, sampie wells
should be distributed evenly throughout the watershed across different land use types. If the
distribution of wells is biased toward a particular type of land use, the overall water quality of
the watershed may either be over- or underestimated. For example, if the majority of the
wells are located in agricultural areas, the constituent loading rate to the surface water may
be overestimated.
Ground water quality In any aquifer can be expected to vary seasonally. Measured
contaminant concentrations in the shallow portions of the aquifer should be higher during the
growing season because of higher fertilizer and pesticide usage during the growing season.
The relation between contaminant concentration In the shallow portion of the aquifer and time
will be a function of depth to the water table, soil properties, and the properties of the
contaminant of Interest. If only one set of ground water quality measurements are used to
characterize the overall ground water quality of the watershed, the calculated constituent
loading rate may not represent conditions throughout the year. To ensure a proper
representation of ground water quality, a sampling program should include more than one
round of ground water quality measurements to account for seasonal variations in
contaminant loadings. This seasonal data will better represent the average annual constituent
loading rate to adjoining surface water,
The reliability of water quality measurements is further affected by the size and the
geology of the watershed. In general, the residence time of constituents In small
watersheds is short. In these small watersheds, the frequency of water quality measurement
should be increased to compensate for the higher variability of the measurements. Finally,
because of the rapid and less diffusive flow pattern of ground water in fractured and karst
systems, water quality measurements from these aquifer are extremely variable over both
-------
2-26
space and time. Therefore, it is very difficult to obtain a representative measurement of water
quality in aquifers in fractured or karst formations.
The limitations and considerations discussed above concern the use of water quality
measurements from monitoring wells and piezometers to estimate contaminant discharge
rates from ground water to surface water and to calculate surface water loadings. These
limitations and considerations should be evaluated carefully before using the obtained water
quality measurements. A summary list of these considerations is given below:
• effects of well construction and well conditions;
• effects of well-screen depth;
• effects of well locations;
• effects of seasonal variations in water quality;
• effects of watershed size; and
applicability to fractured and karst hydrogeologic systems.
-------
CHAPTER III. METHODS
A. Overview
This chapter describes the step-by-step directions for applying the hydrograph separation and
ground water level monitoring methods for estimating contaminated ground water discharge
to surface water. In addition, a standard approach for collecting ground water quality
information and incorporating it into the contaminated load assessments is described. The
estimation of base flow for low flow periods using stream discharge records is not discussed
in detail since its estimation and application is relatively straight forward.
The methods are organized into several steps, and an example problem is presented along
with each of the method steps. For each step, critical questions and approaches involved in
applying the method are outlined. The steps presented are not necessarily all-inclusive;
additional steps or information may be necessary depending on the scope of the study. A
flow chart of steps is included at the end of each method description as a review, and the
steps are outlined in the Appendix C checklists.
-------
3-2
B. Hydrograph Separation Method
This method is described in the following six steps.
STEP 1: Determine Watershed Characteristics and Availability of Gaging Station Data
Question
What are the watershed
characteristics?
Will data from the nearest
gaging station provide
reasonable estimates of
stream discharge at the
point of interest?
Do the gaging station data
have sufficient resolution?
Are the gaging station data
in the correct units?
Are other methods more
appropriate?
Approach
Using topographic maps identify the surface water bodies
of interest, measure their size, delineate and measure the
watershed drainage area, and determine the proximity of
the point of interest to the nearest gaging station.
Determine if the nearest gaging station is close enough
to the point of interest to provide reasonable estimates of
discharges by comparing the drainage areas of the
watershed above the gaging station to the drainage area
above the point of interest.
Determine the gaging station identification number, and
obtain the discharge records from USQS or other
appropriate source. Make certain that the discharge
records for the station have sufficient resolution to identify
individual storm hydrographs. Some gaging stations only
record daily peak discharges.
In addition, make sure the data are in the form of volume
per time. When the data are in the form of river stage
(i.e., water surface elevation) per time, obtain the stage-
discharge tables for the gaging station. Stage-discharge
tables contain the estimated relationship between the
volume of flow and water surface elevations. Convert the
recorded stage elevations to discharges using the stage-
discharge tables.
If it is determined that no gaging station data are
available to reasonably represent the point of interest, or
if the available data do not have sufficient resolution,
other options exist such as: installing a gaging station to
obtain the appropriate data, or utilizing other appropriate
methods described in the companion document to
determine the volume of the ground water discharge.
-------
3-3
EXAMPLE 1:
Watertown has a surface water intake located along the Little Sugar River. Local and
regional authorities are concerned about risks posed by pollutants in the river to the
population served by the intake. The authorities suspect that pollutants are reaching the
river from both ground water and surface water discharges. In order to set priorities for
future studies, facility upgrades, or remedial actions, the authorities would like to have a
rough estimate of the pollutant load from nonpoint sources that enter the surface water via
ground water discharges.
Step 1:
The watershed drainage area to the intake was measured using a topographic map and
planimeter as 100 km . The nearest gaging station is located approximately 2 km
upstream from the water intake. The records for the gaging station report the drainage
area to the station as 90 km . The station and the intake have similar drainage areas (i.e.,
they dfffer only by 10 percent). Therefore, it is reasonable to use the gaging station data
to represent river discharges at the water intake.
The initial review of the gaging station data revealed that there is sufficient resolution to
draw individual storm hydrographs, since discharges are recorded in six-hour increments,
in addition, the station data (Table 3-1} are in the form of volume per time. Thus, there
are sufficient data available to complete a first-tier estimate using the hydrograph
separation technique.
-------
TABLE 3-1. Stream Discharge Data
TIME (days)
DISCHARGE (Us)
0.00
30.0
0.25
28.9
0.50
27.8
0.75
26.8
1.00
25.8
1.25
24.8
1.50
23.8
1.75
23.0
2.00
22.2
2.25
50.0
2.50
100.0
2.75
150.0
3.00
200.0
3.25
162.0
3.50
131.2
3.75
106.2
4.00
86.0
4.25
82.8
4.50
79.8
4.75
76.8
5.00
74.0
5.25
71.2
5.50
68.6
5.75
66.1
6.00
63.6
6.25
61.3
6.50
59.0
6.75
56.8
7.00
54.7
TIME (days)
DISCHARGE (l/sj
7.25
52.7
7.50
50.7
7.75
48.9
8.00
47.0
6.25
45.3
8.50
43.6
8.75
42.0
9.00
40.5
9.25
39.0
9.50
37.5
9.75
36.1
10.00
34.8
10.25
33.5
10.50
32.3
10.75
31.1
11.00
29.9
11.25
28.8
11.50
27.8
11.75
26.7
12.00
25.7
12.25
24.8
12.50
23.9
12.75
23.0
13.00
22.1
13.25
21.3
13.50
20.5
13.75
19.8
14.00
19.0
-------
3-5
STEP 2: Select Storm Hvdroaraphs for Analysis
Question
What time periods are of
interest?
Can representative
hydrographs be identified
from the gaging station
records?
Approach
Identify the time period used for the gaging station data,
and the seasonal time periods of interest for the study.
Review the gaging station data and select several
representative storm events for the time periods of
interest. For each storm event selected, draw a
hydrograph with stream discharge on the Y-axis and time
on the X-axis. Review the hydrographs and select
individual hydrographs to be analyzed for base flow.
EXAMPLE 1: Step 2
Watertown has determined that they are interested in estimating the total annual loading.
However, financial resources are limited. Therefore, the town has elected to determine the
average daily loading rate from base flow estimates for a single representative storm
event. The town reasons that this is appropriate since there is little seasonal variation in
precipitation for the region surrounding Watertown.
When selecting the hydrograph to be analyzed the town considered two items:
1) The town made sure that the precipitation event represented by the selected
hydrograph reflected frequently occurring conditions, since a large precipitation
event with a long recurrence interval would not be representative of "average" or
frequent conditions. Pr?f ipttatten riatfl obtained from the nearest weather station
showed that the selected event was of shorter duration and was less intense than
the one-year, 24-hour precipitation event for the area.
2) The town also made sure that the hydrograph indicated a peak discharge.
Discharge measurements for the selected time period are presented in Table 3-1. The
hydrograph for the 14-day period is shown in Figure 3-1. The hydrograph shows that
significant increases in stream discharge started on Day 2 and the peak discharge, as a
result of the precipitation event, occurred on Day 3.
-------
FIGURE 3-1. Hydrograph
DISCHARGE (l/s)
-------
STEP 3: Determine the Hvdroaraph Slope Seoments for the Receding Limb
Question Approach
Are the different slope Graph the stream discharge on semilogarithmic paper
segments identifiable for the with the log10 of discharge on the Y-axis and time on the
receding limb? X-axis. If the hydrograph does not generate two distinct
slopes, use the time step (N), estimated from the
following equation, as a guide for a time step to the point
where base flow predominates:
N = A0-2 (1)
where A is the watershed drainage area in square miles
and N is the time step in days from the start of the
precipitation event to when base flow predominates.
Is the time step fairly Compare the time step when base flow predominates to
constant? several hydrographs. Also ensure that the time step
agrees with tie hydrographs and is not over- or
underestimating the base flow contribution.
EXAMPLE 1: Stop 3
Figure 3-2 shows the hydrograph for the Little Sugar River on semilogarithmic scale.
Inspection of the hydrograph shows that distinct slope segments are identifiable, and the
time step when base flow predominates is approximately two days. Thus the time step
calculated in Step 3 does not have to be used. The slope segment A-B in Figure 3-2 is
the slope segment representing base flow.
Note that the time step will not always be graphically obvious and that equation 1 will aid
in identifying the time step in these cases. For this watershed, the use of equation 1
yields:
N = (100 km2 x 0.386 Sq Mi / km2)0,2 * 2.1 days.
I, j)/o+- - A)tj —
X hi - (Ai/-€^y ~ .2,/ f .
-------
FIGURE 3-2. Semilogarithmic Hydrograph
DISCHARGE (l/s)
TIME (days)
-------
3-9
STEP 4: Determine the Base Flow Recession Constant
Approach
Use the time step (N) and the slope segment identified in
Step 3 to develop the base flow recession line. Extend
the lower line equivalent to line A-B in Figure 3-3) back to
the time ol peak discharge {point C). The Y value of
point C is the maximum baseline discharge to streamflow.
Find an additional point on the lower line of the semilog
hydrograph. Determine the base flow recession constant,
Kj., using the maximum baseline discharge and some
base flow rate at a known time following the maximum
discharge. This equation is:
Kr = (Qj/Q0)1/t (2)
where Q0 = maximum baseline discharge (volume per
time);
t = time after maximum base flow rate (time
usually days; t = tj -10;
Qj = base How at time t (volume per time; see
Figure 3-3).
EXAMPLE 1: Step 4
Line A-B in Figure 3-3, represent the base flow contribution to streamflow. By extending
line segment A-B back in time to the time of peak discharge, the maximum base flow as a
result of the precipitation event (i.e., Q0 in equation 2) can be estimated. The maximum
base flow as a result of the precipitation event, as determined from Figure 3-3, is 100 l/s.
The base flow recession constant for the storm can be determined for the recession limb
of the semilogarithmic hydrograph and the maximum base flow. For this storm on the
Little Sugar River the base flow recession constant is:
K, = (19 l/s /100 l/s)in1 days = 0.86 l/day.
Question
What is the base flow
recession constant?
-------
FIGURE 3-3. Hydrograph Separation
DISCHARGE (l/s)
300
200
Q. a
/qV a
-
I A* oP <)
D
Ci|Q|)
i i i
i i
(,i°P
100
50
30
20
10
0
6 8 10
TIME (days)
12
14
16
-------
3-11
STEP 5: Determine the Base Flow Constant for the Rising Limb of the Storm
Question Approach
What is the base flow Extend the line from the point of maximum base flow to a
constant of the rising limb? point where base flow predominates at the start of the
storm (equivalent to line segment C-D, Figure 3-3).
Determine the rising base flow constant using the
maximum base flow discharge and the base flow
discharge at a point where base flow predominates
immediately prior to the precipitation event (point D,
Figure 3-3). The equation is:
Kl-(Q0/Q|),/l (3)
where Q0 maximum baseline discharge;
t = time after initial point where base flow
predominates to point of maximum base
flow (i.e., t = t0 -1|); and
Qj = base flow prior to the precipitation event
(Figure 3-3).
EXAMPLE 1: Step 5
Extending a line from the maximum base flow discharge, point C to point D in Figure 3-2
represents base flow contributions to streamftow during the rising portion of the storm.
The base flow constant for this rising portion of the storm is:
K, = (100 l/s / 22 l/s)1/(3 days '2 day8) = 4.54 l/day.
-------
3-12
STEP 6: Determine the Total Volume of Base Flow for the Precipitation Event
Question Approach
What is the total volume of The line B-C-D in Figure 3-2 represents the base flow
base flow? contribution to streamflow. The area under the base flow
line equals the total volume of base flow.
The area under the lines can be determined graphically
or mathematically. To graphically determine the area
under the line segment use a planimeter or determine the
average discharge rate for both the rising and receding
line segments, and multiply by the number of days
represented by each line segment. Summing the two
areas will provide an estimate of the volume of base flow
for the entire storm. If using a planimeter, measure the
areas using the normal hydrograph, not the hydrograph
plotted on semilogarithmic paper.
To mathematically estimate the area under each line, the
regression equations for the separate line segments must
be integrated. The integral of the recession line segment
with respect to time is:
fr\ -\ Q0K,'dt = Q(/j_l\«-,0>-_!_KI0l (4)
^ ^ln(K,) J InfK,)
Similarly, the integral of the rising line segment is:
^ OK,1 dt = Q, UL K,<10"10 - _J_ Kj°] (5)
ln(KD In(Kj)
The sum of these two areas equals the total area under
the line segments which is an estimate of the volume of
base flow from time t( to time 1 (Figure 3-2).
-------
3-13
EXAMPLE 1: Step 6
The volume of the base flow during the receding portion of the precipitation event on the
Little Sugar River is:
"11 Q0K/dt = 100 l/s [1 / ln{0.86 l/d) x (0.86 l/d)11 days -1 / ln(0.86 l/d) x {0.86 l/d)0days]
° = (100 l/s) x (86400 s/d) x (5.37 days) = 4,6 x 107 I.
The volume of base flow during the rising portion of the precipitation event is:
j"1 Q K*dt = 22 l/s [1 / ln(4.54 l/d) x (4.54 l/d)1 day -1 / ln(4.54 l/d) x (4.54 l/d)° days]
° = (22 l/s) x (86400 s/d) x (2.3 d) = 4.4 x 106 I.
The total volume of base flow during the precipitation event equals:
4.6 x 107 I + 4.4 x 106 I = 5.0 x 107 I.
This period covers 12 days. Thus, the average daily base flow discharge is 4.2 x 106 l/d
or 49 l/s. This results in an annual base flow discharge of 1.5 x 109 l/year.
If base flow for low flow at 14 days is used as a comparison, the average daily base flow
discharge is:
19 l/s x 86,400 s/d « 1.6 x 106 l/day.
This results in an annual low flow base flow discharge of 6.0 x 108 l/year. These two
estimates of base flow, the storm flow and the low flow, can be used to represent the
range of the annual base flow discharge. The estimates multiplied by the average
concentration of pollutants determined using the ground water well sampling method
described later give an estimated range of the nonpoint source contaminated ground
water discharge to Little Sugar Creek.
In addition to estimating the volume of base flow discharge, as described above, the
hydrograph separation technique can be used to estimate base flow at specific times during
the precipitation event. This is done using the regression equations developed for the line
segments C-D and C-B shown In Figure 3-3. The equations for the two line segments are:
Q, = QQ^im, and
0, = Q,K,(,-t0
-------
3-14
for the receding and rising line segments, respectively, where t is the time at the point of
interest. A flow chart summarizing the steps in the hydrograph separation method is
presented as Figure 3-4.
-------
FIGURE 3-4.
Flow Chart of the Hydrograph
Separation Methodology
STEP 4: Determine the base flow recession constant.
J
-------
3-16
C. Water Level Well Measurements and Darcy's Law
The water-level monitoring method is described in the following three steps.
STEP 1: Determine Watershed and Aquifer Characteristics and the Availability of
Sampling Wells
Question
What are the watershed
characteristics?
What are the aquifer
characteristics?
What is the spatial coverage
of monitoring wells and
piezometers?
Are well logs, boring logs,
geophysical surveys, and
construction history
available for the wells and
piezometers?
Approach
Using topographic maps, identify the surface water body
of interest, measure its size, delineate and measure the
watershed drainage area.
Using geologic maps (e.g., structure, stratigraphy, and
lithology maps), identify the rock types and hydraulic
properties of the aquifer underlying the watershed.
Determine whether the aquifer is homogeneous and
isotropic.
Using hydrogeologic maps (i.e., locations of wells and
piezometers), determine the availability and spatial
coverage of sample points. Information on well locations
can be requested from developers, engineering
companies, public and private water organizations, and
the U.S. Geological Survey. If the spatial coverage of the
wells is inadequate, installation of additional welis may be
required.
Determine if ttthologic and construction information are
available for the monitoring wells and piezometers.
Are other methods more
appropriate?
If it is determined that the underlying aquifer is far from
homogeneous and isotropic, it may be necessary to u$e
other options such as those described in Appendix A.
-------
3-17
EXAMPLE 2:
^AquavilleXas a surface water intake located along the Big Salt Creek as shown in Figure
y5r~t6cal and regional authorities are concerned about risks posed by pollutants in the
river to the population served by the intake. The authorities are suspicious that pollutants
are reaching the river from both ground water and surface water discharges. In order to
begin assessing this problem, the authorities conducted a study to estimate ground water
discharge tc^BT^SaJT Creek^"1
Step 1: f ^ \
The watershed above the water intake was measured ai^Ooo krrrjdsing topographic
maps and a planimeter. An initial review of geologic informSnJCTfon the aquifer suggested
that the aquifer is(hamageneous^and isotropic^ ^ ^ tCefs
Based on information from the U.S. Geological Survey, private and public water
organizations, and local engineering companies, a total of 12 monitoring wells and
piezometers are available for water level measurements (see Figure 3-5). An adequate
spatial coverage was provided by these 12 wells over most of the watershed. In addition,
elevation information is available lor all 12 wells.
-------
FIGURE 3-5. Big Salt Creek Watershed
-------
3-19
STEP 2: Determine Sampling Periods and Obtain Water Level Measurements
Question
What time periods are of
interest?
Can a water-level map be
generated?
Approach
Using historical well level measurements, identify the time
periods and seasonal time periods that are representative
of the hydrologic conditions of the watershed.
Obtain depth to water level measurements using steel
tape and chalk or electric well sounder. Subtract the
depth to water level measurements from the elevations of
the wells to obtain water level elevation data. Generate
contours for the water level measurements using manual
interpolation methods or automated contouring
packages.
When well elevations are unknown it may be necessary
to conduct a vertical survey.
EXAMPLE 2: Step 2
Aquaville's water quality planners have determined that they are interested in estimating
the total annual nonpoint source loading from ground water discharge. Because of
limited financial resources, the town has elected to determine the average daily loading
rate from ground water discharge during a time period when the water level elevation is
about average. Historical water level measurements frorrj two of the 12pveils were
examined to identify time periods that can be used for water level sampling.
Depth to water level measurements were made for all 12 wells. These measurements
were subtracted from the surveyed elevations of the wells. A water level contour map was
generated by manually interpolating to obtain water level elevations (see Figure 3-6).
-------
FIGURE 3-6. Water Table Elevation Map
Big Salt Creek Watershed
-------
3-21
STEP 3: Applying Darcv's Law
Question Approach
What is the hydraulic
conductivity of the Aquifer?
What is the hydraulic
gradient?
What is the cross-sectional
area of the aquifer?
What is the discharge rate of
the ground water to the
surface water body?
Determine the hydraulic conductivity of the aquifer by
means of field tests or on the basis of aquifer materials,
Typical hydraulic conductivity values can be found in
hydrogeology textbooks and publications.
Using the water level contour map, generate flow lines
between water level contours. These flow lines should be
at right angles to the water level contours. Determine the
maximum water fevef efevafion and the minimum water
level elevation for each of the flow lines, identify their
locations and the horizontal distance between these
locations. Calculate the hydraulic gradient by dividing
the difference between maximum and minimum water
level elevations by the horizontal distance between them.
Using iithofogic maps and well logs, determine the
average thickness of the surfictal aquifer. The thickness
of the aquifer is then multiplied by the ength of the
region jn which ground water and surface water are in
contact to obtain the cross-sectional area of the aquifer.
Using the estimated hydraulic conductivity, hydraulic
gradient, arid cross-sectional area of the aquifer, use
Darcy's Law to calculate the discharge rate from the
aquifer to the surface water body:
Q - K(dh/dl)A (8)
where: Q = discharge to the stream,
K = hydrauffc conductivity,
dh/dl = horizontal hydraulic gradient, and
A = cross-sectional area of the aquifer
parallel to the stream.
-------
3-22
EXAMPLE 2: Step 3
Using geologic maps and site investigation reports, the hydraulic conductivity (K) of the
aquifer underlying the watershed of Big Salt Creek was determined to be 10"4 cm/sec.
Based on the lithologic data for the 12 wells located in the watershed, silty sand is found
to be the primary aquifer material. The hydraulic conductivity of silt sand is estimated to
be 10"4 cm/sec. __
Flow lines were drawn at right angles to the water level contours on the water level
elevation map (Figure 3-6). The maximum and minimum water level elevations are 300
and 100 meters respectively. The horizontal distance measured between these locations
is 20 kilometers. The hydraulic gradient was calculated as dh/dl, where dh is the
difference between the maximum water level elevation and minimum water level elevations
The derived hydraulic gradient is:
(300 m -100 m)/(20km x 1000 m/km) = 0.01 m/m. H — Cl OJ
Lithologic data for the 12 wells located in the watershed suggest that the average
thickness of the aquifer is 100 meters. The section of Big Salt Creek above the intake
was measured as 50 kilometers long. The cross-sectional area (A) of the aquifer in cm2
is:
A = (100 m x 100 cm/m) x (50 km x 1000 m/km x 100 cm/m)
= 5 X1010 cm2. /t ^ r X M
By substituting the obtained values in Darcy's Law, the calculated discharge rate of
ground water (Q) to Big Salt Creek from each half of the watershed is:
Q = K (dh/dl)A o
= (10*4 cm/sec) x (0.01 m/m) x (5 x 1010 cm2)
= 50,000 cm3/sec . ,
- 50liters/sec. (j£) j/S'
isi) ~ zQ-< Y!
i/j
efetst
11
The total ground water discharge to the Big Salt Creek from both sidec of the wotorsfaedj
assuming equal ground water contributions from both sides, is 100 liters/sec.
A flow chart summarizing the steps for using water level well measurements and
Darcy's Law to estimate ground water discharge to surface water is presented in
Figure 3-7.
-------
FIGURE 3-7. Flow Chart for Using Water
Measurements and Darcy's
Level
Law
-------
3-24
0. Ground Water Loading Measurements
The approach for estimating ground water contaminant loading of surface water bodies is
described in the following three steps.
STEP 1: Evaluate the Spatial Coverage of Sampling Wells in the Watsrshed
Question
What is the spatial coverage
of monitoring weils and
piezometers?
Are well logs, boring logs,
geophysical surveys, and
construction history
available for the monitoring
wells and piezometers?
Are measurements from
sampled wells and
piezometers representative
of the watershed?
Approach
Using hydrologic maps (i.e., locations of wells and
piezometers), determine the availability and spatial
coverage of sample points. Information on well locations
can be requested from developers, engineering
companies, public and private water organizations, and
the U.S. Geological Survey. II the spatial coverage of
wells is inadequate (e.g., a lack of wells in ctose proximity
to the water body), installation of additional wells may be
required.
Determine if lithologic and construction information are
available for the monitoring wells and piezometers.
Determine if the wells and piezometers are properly
constructed and screened at the appropriate depth to
represent the water quality of the discharging aquifer.
-------
3-25
EXAMPLE 3:
Aquaville has a surface water intake located ori the Big Salt Creek as shown in Figure 3-5.
In order to set priorities for future studies, facility upgrades, or remedial actions, the
authorities would like to have a rough estimate of the pollutant load from nonpoint
sources that enter the surface water via ground water discharges.
Step 1:
Based on information from the U.S. Geological Survey, private and public water
organizations, and local engineering companies, a total of 12 monitoring wells and
piezometers are available for water level measurements (see Figure 3-5). An adequate
spatial coverage was provided by these 12 wells over most of the watershed. Additional
review of well logs indicates that water-quality measurements are representative of the
watershed.
-------
3-26
STEP 2: Determine Temporal and Spatial Variations of Water Quality
Measurements and Estimate Average Contaminant Concentrations
Question
Approach
What time periods are of
interest?
How variable is the sampled
water-quality data across the
watershed and what is the
estimated contaminant
concentration?
Using historical well water-quality measurements, identify
the time periods and seasonal time periods that are
representative of the water-quality conditions of the
watershed. If ground water quality is expected to vary
seasonally, a sampling program involving multiple time
periods should be implemented to better represent the
average annual contaminant loading rate to the surface
water.
Examine spatial variations of water quality from sampled
wells in the watershed. If variations do not exist, an
average value of each specific contaminant can be
utilized. If variations exist, identify the contaminant
concentrations that best represent the loading to the
surface water. Methods to estimate total ground water
quality across the watershed include:
1. Using only contaminant concentrations of wells
and piezometers closest to the surface water
body; or
2. Segmenting the watershed by land use
categories, estimating concentrations based on
water quality data for each land use type, and
aggregating contaminant loadings lor all land use
categories to obtain the total load.
-------
3-27
EXAMPLE 3: Step 2
Because of limited financial resources and a lack of historical water-quality d£ta, Aquaville
officials have elected to determine the average daily loading rate from groundwater
discharge during a time period when the water level elevation is aboutjavefrage~] Historical
water level measurements from two of the 12 wells were examined to ia&niify time periods
that can be used for water level and water quality sampling.
Water-quality measurements were made foftnTTiTwells/ These measurements are
presented in\T^ble 3-2. / The average contaminant concentration in ground water in the
watershed is mg/liter. The average contaminant concentration found in the six wells
closest to the Big Salt Creek is 7 mg/liter. The effect of land use on the average
contaminant concentration in the watershed was evaluated and no correlation was
observed between land use and measured contaminant concentrations at the well heads.
However, the contaminant concentrations of the 6 wells closest to Big Salt Creek appear
to provide the/most representative estimate of the immediate contaminant loading rate to
the river. /
-------
TABLE 3-2
Water Quality and Water Level Elevation Data
WELL#
CONTAMINANT
WATER LEVEL
CONCENTRATION
ELEVATION
(mg/l)
(m)
1
7
100
2
6
100
3
2
200
4
3
200
5
1
300
6
6
100
7
8
100
8
1
300
9
4
200
10
3
200
11
6
100
12
7
100
Average Contaminant Concentration across the watershed: 4.7 mg/l
Average Contaminant Concentration in wells closest to the surface water: 7.0 mg/l
-------
3-29
STEP 3: Determine Contaminant Loading Rate to Surface Water
Question
What is the ground water
discharge rate to the surface
water?
Approach
Using hydrograph separation or water tevel measurement
techniques, determine the ground water discharge rale to
the surface water.
What is the constituent
loading rate to the surface
water?
Multiply the constituent concentration in the discharging
ground water by the ground water discharge rate to the
surface water.
EXAMPLE 3: Step 3
The constituent loading rates to the Little Sugar River (Example 1) and Big Salt Creek
(Example 2) are determined by multiplying the constituent concentrations in the
watersheds by the ground water discharge rates to the surface waters.
For the Little Sugar River, using the hydrograph separation technique, the average base
flow discharge is estimated to range between 49 l/s and 19 l/s. The average constituent
concentration from sampled wells and piezometers is 7 mg/liter, The calculated annual
constituent loading rate is:
Loading Rate (LR) = Discharge X Constituent Concentration
= 49 Itters/sec X 7 mg/liter = 340 mg/sec
= 19 liters/sec X 7 mg/liter = 130 mg/sec.
Thus, the annual load of the constituent of interest contributed to Little Sugar Creek by
ground water discharge ranges from 10,700 kg to 4,100 kg.
For Big Salt Creek, water-quality measurements from the six wells closest to the river were
used to provide an estimation of the immediate constituent loading rate to the river. Using
water level measurements, the ground water discharge rate is estimated as 100 liter/sec.
The calculated annual average constituent loading rate is:
Loading (LR) = Discharge rate X Constituent Concentration
= 100 liter/sec X 7 mg/liter
= 700 mg/sec.
Thus, the average annual load of the constituent of interest contributed to Big Salt Creek
is 22,000 kg.
-------
3-30
A flow chart summarizing the steps for using well sampling and analysis to estimate
ground water contaminant concentrations and contaminated ground water discharge
loads to surface water is presented in Figure 3-8.
-------
FIGURE 3-8. Flow Chart for Ground Water
Quality Measurements
-------
This page intentionally left blank.
-------
CHAPTER IV CASE STUDIES OF METHODS APPLICATIONS
A. Overview
This chapter presents two case studies to illustrate the application of the methods described
in Chapter III. These case studies are based on information gathered from the technical
literature. Data sources are identified in the text. Wherever necessary, data are
supplemented in the case studies to better illustrate the application of the methods. These
case studies are not intended to highlight all relevant applications for the methods, nor do
they illustrate all method limitations. Instead these case studies are presented to provide the
reader with a fundamental understanding of the manner in which the methods can be applied.
B. Case Study No. 1: Base Flow/Hydrograph Separation6
Hydrograph separation methods were used on data from the Spokane River Basin in
southwestern Washington State. The goal of the study was to estimate the contaminated
ground water discharge to the Spokane River for several days following storm conditions.
The specific contaminant of interest was nitrate.
Part 1: Estimate the Ground Water Discharge using the Hydrograph Separation
Method
Step 1: Determine the Watershed Characteristics and the Availability of Gaging
Station Data
In the upper portions of the Spokane River in Idaho, the river loses water to ground water. In
contrast, the studied portion of the river both gains and loses water at different times of the
year. Additional withdrawals and recharges occur through irrigation extractions for rangeland
and other agricultural activities. These extractions reduce the total contribution of ground
®This case study is based upon data reported in C.R. Patmont, G.J. Pelletier, LR.
Singleton, et. al., 1987, The Spokane River Basin: Allowable Phosphorus Loading. WDOE
Contract Number C0087074, September 1987, and personal communication with Luis Fusts,
USGS.
-------
4-2
water discharge to streamflow since water is lost from the system above the gaging station.
The base flow estimate can be considered as conservative since it will reflect losses from the
system.
The study area consists of metamorphosed sediments, gneiss, and granitic intrusives.
Relatively impervious bedrock occurs close to the ground's surface. The riverbed itself
contains glacial deposits, which range greatly in their permeability. As a result, the river may
not be continuously recharged by ground water. Washington Department of Ecology
studies7 indicate that upper portions of the river have nearly negligible ground water input,
but the portion of the Spokane River from the Idaho border to the City of Spokane is
influenced by a riverine aquifer. This should not significantly affect the ability to predict base
flow, but may affect the ability to estimate the concentration of contaminants in the
discharging ground water since only portions of the watershed discharge directly to the
stream.
The river flow is also affected by an impoundment. Fourteen miles into Idaho, the Spokane
River is dammed to form Coeur D'Alene Lake (R.M. 111.7), thereby influencing water levels in
this portion of the Spokane River.8
The USGS maintains gaging stations at two locations along the Spokane: Post Falls, at river
mile 100.5, and the Spokane station in the City of Spokane. This case study focuses on data
collected at Spokane, as this location is downstream of and closer to the point of interest.
Daily surface water gaging data (flow in cfs) and ground water quality data for this study were
collected from October 1985 to September 1989. The hydrograph for the four years of data (s
presented in Figure 4-1.
7C.R. Patmont, G.J. Pelletier, LR. Singleton, et. al., 1987.
8Officials in charge of the Lake were contacted and they indicated that the lake had a
fixed discharge spillway system that was dependent on inflow, and that flood storage effects
to the downstream hydrograph at the point of interest were minimal.
-------
FIGURE 4-1. Four Year Hydrograph of Spokane River
Gaging Station at Spokane, Washington
12/30/85 06/23/86 12/25/86 06/23/87 12/20/87 06/17/88 12/14/88 06/12/89
October 1,1985 to September 30,1989
-------
4-4
Step 2: Select Storm Hvdroqraphs for Analysis
As Figure 4-1 indicates, the majority of the basin's annual discharge occurs in the spring.
These seasonal increases in streamflow are probably caused by an increase in discharge
from snow melt. To ensure that the case study represented the ground water discharging to
the stream, hydrographs following the wet (snow melt) season were chosen. This portion of
the water year will show significant bank storage effects, providing a conservative estimate of
ground water contributions to surface water quality.
Two hydrograph peaks were identified. The first period occurred between July 29 and August
12, 1989. The second period occurred between May 21 and June 11, 1986. The
hydrographs for these streams are graphed in Figures 4-2 and 4-3, respectively. It should be
noted that the May-June hydrograph is actually a double hydrograph, and this may cause the
ground water discharge estimates to be higher. This hydrograph also occurred closer to the
rainy season, which may also account for any higher discharges. The peak discharge for the
1986 data was 13,200 cfs and occurred between May 29 and 30. The peak discharge for the
1989 data was 1,660 and occurred between August 3 and 4.
Step 3: Determine the Hvdroaraph Slope Segments for the Receding Limb
Each hydrograph was graphed on a semi-log scale, to facilitate the identification of any
changes in slope of the descending leg of the hydrograph. Figures 4-4 and 4-5 represent the
logarithmic graphs of the 1989 and the 1986 hydrographs, respectively. Although the shapes
of the hydrographs differ, the slopes for both recession limbs appear to change between 4
and 5 days following peak discharge. Although semilog hydrographs for this reach provide
well-defined slopes, we analyzed additional hydrographs from the data set and found that
four to five days following a single rain event indicated a period where base flow
predominates surface water discharge. Thus, calculation of the time period to the
predominance of base flow using equation (1) was not necessary.
Step 4: Determine the Base Flow Recession Constant
We extended the recession lines on each semilog hydrograph to the time of peak discharge
to obtain estimates of the maximum base flow. The maximum discharge for ground water on
May 30,1986, was estimated as 7,903 cfs. The maximum discharge for ground water for
-------
1800
1700
1600
1500
1400
1300
1200
1100
1000
900
800
700
4-2. Single Event Hydrograph of Spokane River
Gaging Station at Spokane, Washington
/29J89 07/31/89 08/02/89 08/04/89 08/06/89 08/08/89 08/10/89 08/12/89
07/30/89 08/01/89 08/03/89 (18/05/89 08/07/89 08/09/89 08/11/89
July 29,1989 to August 11,1989
-------
FIGURE 4-3. Single Event Hydrograph of Spokane River
Gaging Station at Spokane, Washington
16
14
12
^ (A
A 13
VH f-
o s
V 3 1°
O S
E £
8
j.. i
L J I
1 '
I I I I
I I
05/21/86 05/25/86 05/29/86 06/02/86 06/06/86 06/10/86
05/23/86 05/27/86 05/31/86 06/04/86 06/08/86 06/12/86
May 22 to June 15,1986
-------
FIGURE 4-4. Logarithmic Single Event Hydrograph of Spokane River
Gaging Station at Spokane, Washington
07/29/89 07/31/89 08/02/89 08/04/89 08/06/89 08/08/89 (18/10/89 08/12/89
07/30/89 08/01/89 08/03/89 08/05/89 08/07/89 08/09/89 08/11/89
July 29,1989 to August 11,1989
Straight line indicates the ground—water regression using 8/7 to 8/12 data, and regressing lo the peak at 8/4
Regression equation is flow(cfs) = 1030 — 14(Jays)
-------
FIGURE 4-5. Log Single Event Hydrograph of Spokane River
Gaging Station at Spokane, Washington
100000
©
% 10000
o
E
1000
05/21/86 05/25/86 05/29/86 06/02/86 06/06/86 06/10/86
05/23/86 05/27/86 05/31/86 06/04/86 06/08/86 06/12/86
May 22 to June 15,1986
Straight line represents regression of flow(cfs) = 9831 — 214.3(days)
Data used for regression from 6/3 to 6/8, and regressed to discharge peak at 5/29/89
-------
4-9
August 4, 1989, was 932 cfs. These very different values reflect different periods in the water
year for the river, as well as the assumptions noted above. The flow in this reach of Spokane
ranges from 500 cfs to 24,000 cfs over the course of a year.
The base flow recession constant is part of a function that is used to estimate the base flow
for any time following peak flow. Base flow recession requires the use of the estimated
maximum discharge, and base flow at a point along the recession line where base flow
predominates. We used discharge at 5 days following the peak to identify a point that
estimates base flow predominance in the hydrograph:
May 1986 hydrograph: peak = 7903 cfs
5 days following peak = 5000 cfs
August 1989 hydrograph: peak = 932 cfs
5 days following peak = 862 cfs
These data were entered into the following equation to determine the base flow recession
constant:
K, - (Qj/Qo)1"
where Q0 = maximum baseline discharge (volume per time)
t = time after maximum base flow rate (time, usually days)
Qpj = base flow at time t (volume/time)
The recession constant for the May 1986 hydrograph was 0.902. The recession constant tor
the August 1989 data was 0.985.
The base flow for any time (t) after peak flow is = QqK^.
Step 5: Determine the Base Flow Constant for the Rising Limb of the Storm
The rising limb base flow constant was calculated for a line connecting the maximum base
flow with a point reflecting base flow prior to the storm. The base flow rate prior to the slorm
-------
4-10
for the May 1986 hydrograph is difficult to estimate since it is a double hydrograph. However
a point at the start of the first hydrograph was utilized to estimate base flow prior to the
storm. The base flow rates estimated prior to the storms were 5700 cfs and 890 cfs for the
May 1986 and the August 1989 storms respectively. The data was entered into the following
equation to determine the rising limb base flow constant:
The calculated rising limb base flow constants were 1.04 and 1.05 for the May 1986 and the
August 1989 storms respectively.
The base flow at any time (t) between the point prior to the storm and the time of peak flow
is:
Step 6: Determine the Total Volume of Base Flow for the Storm Event
To estimate the base flow to the Spokane River System for the ten days following the peak
flow, the area under the recession curve must be estimated. The integral of the above
function for the 1986 storm is:
Kj = (Q0/Ql)1'1
where Q0 = maximum base line discharge (volume/time),
t = time from base flow rate identified prior to the storm to time of
maximum base flow (time, usually days)
Q(j) = base flow rate prior to the storm.
= QfflK,'
'10 Q0K/ dt = Q0 [_!_ K,'0 - -L Kr°l
, MK,) ln(K,)
thus, the base flow volume estimates for the time period from the time of peak flow to ten
days following peak flow are:
-------
4-11
May 1986 hydrograph ten-day ground water contribution after peak flow
total flow = 7903 cfs (86,400 s/d)[1/!n(0,902)(0.9021°) -
1/ln(0.902)(0.902°)]
= 42.6 X 10B cubic feet.
August 1939 hydrograph ten-day ground water contribution after peak flow
total flow = 932 cfs {66,400 s/d)[1/ln(0.985){0.9851°) -
1/ln(0.965}(0.9850)]
= 7.5 X 108 cubic feet.
The base flow from the start of the storm to the time at peak flow must be estimated from the
area under the rising base flow curve. The integral of this function for the 1986 storm is:
f6 QjK* dt = Qj LJ_ Kj8 - _1_ Kj°]
4, In(K^) InfK,)
thus, the base flow volume estimates for the rising limb of the hydrographs are:
May 1986 hydrograph rising limb ground water contribution
total flow = 5700 cfs (86,400 s/d)(t/ln(1,04)(1.04®) -
1/ln(1.04)(1.04°>]
= 46.3 X 1Q8 cubic feet.
August 1989 hydrograph rising limb ground water contribution
total flow = B90 cfs (86,400 s/d)[1/ln(1.05)0 .OS5) -
1/ln(1.05)(1.0S°)J
= 4.4 X 10s cubic feet.
Total base flow for the storms from the time just prior to the storm to ten days following the
tfme to peak discharge equals the sum of the base flow estimates for the rising and receding
-------
4-12
portions of the storm. The estimated total base flow for the May 1986 storm covers 18 days
and equals:
42.6 X 108 cubic feet + 46.3 X 108 cubic feet = 88.9 X 108 cubic feet;
while the total base flow for the August 1989 storm covers 15 days and equals:
7.5 X 108 cubic feet + 4.4 X 108 cubic feet = 11.9 X 10® cubic feet.
The average daily flow rates for the two storms were 4.9 X 108 cubic feet and 7.9 X 107 cubic
feet for the 1986 and 1989 storms respectively.
Part 2. Estimate Ground Water Quality
Step 1: Evaluate the Soatial Coverage of Sampling Wells in the Watershed
Ground water quality data collected from eight wells in the watershed were used to assess
nitrate contaminant load. A map of well locations is provided in Figure 4-6. These sample
point are not evenly distributed throughout the watershed, but are all located in one aquifer
which is known to discharge to the river. These wells were established to monitor the water
quality of the aquifer. Therefore, it is assumed that the wells were constructed to obtain
representative water quality samples of the aquifer.
Step 2: Determine the Temporal and Spatial Variations of Water Quality Measursmontl
and Establish Average Constituent Concentration
Ground water data was sampled every three months from October 1986 to July 1989. This
case study used the August 1989 hydrograph recession and ground water data from July
1989.
Data on the northern shore of the Spokane River indicates lower nitrate levels than the
southern shore. This may be the result of the very limited well data at the northern side, but
-------
FIGURE 4-6. Spokane River Watershed Map Below Lake Coeur D'Aline
-------
4-14
for this exercise we averaged the data for the southern sides of the river separately. In
addition, we deleted one sample point because of its much greater distance from the river
than the other data points. The northern well data (two wells) for July 1989 average 1,04 mg/|
nitrate. The southern well data (five wells) average 1.18 mg/l nitrate for July 1989,
Step 3. Determine Contaminant Loading Rate to Surface Water
Using the average contaminant concentration data, the loading at any time could be
estimated by dividing the estimated flow in half (assuming the flow from each half of the river
bed is equal), and multiplying this value by the ground water concentrations averaged above
and adding these two values together for the total flow. For the peak flow into the river,
estimated as 932 cubic feet per second, the load totals:
Peak Daily Load = 932 cfs((1.04+1.18)/2)mg/l)(28.321 l/cf)
= (29,287mg/s)(86,400s/day)(1 O^kg/mg)
= 2,530 kg/day nitrate
From this calculation, the peak day following the August rain event will have a loading from
the watershed aquifer of 2,530 kilograms of nitrate a day.
We estimated the total nitrate load from ground water discharge to surface water for the
fifteen day period represented by the August 1989 storm as:
Fifteen day Load = 11.9 X 108 cf (1.11 mg/l)(28.321 l/cf)(10"6 kg/mg)
= 37,400 kg/15 days following the peak storm.
Thus over 37,000 kg of nitrate are estimated to enter the surface water from ground water
during the fifteen day storm period for an average of 2,490 kg/day.
Using the same water quality data with the May 1986 storm ground water discharge estimate
gives an average nitrate loading estimate of 15,500 kg/day.
These estimates show a high degree of variability. This example was chosen to show the
-------
4-15
high variability and limitations of the method. The application of the hydrograph separation
method was hindered by significant seasonal variations in streamflow, withdrawals from
streamflow for irrigation, and recharge to ground water. In addition, the limited distribution of
sample wells affected the utility of the hydrograph separation method. The sample wells were
located to represent discharge from a single aquifer, not the entire watershed or reach of the
river. The hydrograph separation technique estimates ground water discharge for the entire
watershed above the gaging station. A different method such as well level measurement and
Darcy's Law could be applied to estimate discharge for the individual aquifer.
-------
4-16
Case Study No. 2: Use of Ground Water Monitoring Wells for Estimating
Discharge and Contaminant Loading
Water-level measurements and hydrogeological properties were obtained for a subwatershed
of East Mahantango Creek, a tributary of the Susquehanna River, which is about 25 miles
north of Harrisburg, Pennsylvania.9 The watershed is designated WE-38. The goal of this
study is to estimate the contaminant loading of nitrate to surface water through ground water
discharge.
Part 1: Estimating Ground Water Discharge using Water Level Measurements and
Darcy's Law
Step 1: Determine Watershed and Aquifer Characteristics, and the Availability of
Sampling Wells
The WE-38 watershed is a rural upland watershed in the non-glaciated, folded, and faulted
Appalachian Ridge and Valley Province. The surface elevations of the watershed range from
750 to 1,600 feet. The length of the river within the watershed was measured as 9,500 feet.
The WE-38 watershed is underlain by two geologic formations, Trimmers Rock (Late
Devonian) and Catskill (Late Devonian-Early Mississippi). The Trimmers Rock Formation is
predominantly shale, outcropping at the watershed outlet (south) in a near-horizontal position
and increasing in dip at its contact with the overlying Catskill Formation. The Catskill consists
of interbedded shales, siltstones and sandstones, becoming increasingly coarse-grained from
south to north. The dip of the Catskill Formation increases to where a relatively pure quartz,
sandstone-conglomerate outcrops to form the northern watershed divide. Less than 5 feet of
shallow residual soils, predominantly silt loams, cover virtually all of the bedrock.
The ground water flow in the watershed appears to be controlled by a two-layered geologic
9Gburek, W.J., R.R. Schnabel, and S.T. Potter. Modeling the Effects ol the Shallow
Weathered Fracture Layer on Nitrate Transport. Northeast Watershed Research Center,
USDA-ARS University Park, Pennsylvania.
-------
4-17
system: a deeper, moderately fractured regional aquifer and a shallow, weathered, intensively
fractured aquifer. Ground water flow paths in the watershed are directly related to
topography, and subsurface water divides generally coincide with major surficial water
divides. Below a depth of approximately 300 feet, well yield becomes negligible due to low
hydraulic conductivity (e,g., 0.1 to 0.3 feet/day) and low specific yield (e.g., 0.001).
Nine observation wells were available for ground water level measurements in the deeper
regional aquifer. The spatial coverage by these observation wells is adequate for generating
a water-level map. Because these observation wells were constructed to eliminate the effects
of the shallow ground water, their data are not representative of the shallow aquifer.
Therefore, this study will only examine the ground water flow of the deeper regional aquifer.
The actual flow dymanics of this watershed are somewhat more complicated and require data
from the shallow aquifer for a more comprehensive assessment of flow. For a more detailed
discussion of the shallow aquifer see Gburek et al. 1988.
Step 2: Determine Sampling Periods and Obtain Water Level Measurements
Well-level measurements from March 1983 to December 1987 were available for the nine
observational wells. In this study, discharges and contaminant loading rates are calculated
on a quarterly basis from 1983 to 1987.
A potentiometric map (i.e., a water level map) of the WE-38 watershed is presented in Figure
4-7. The data for this potentiometric map is from the fourth quarter of 1983.
Step 3: Applying Darcv's Law
The hydraulic conductivity of this deeper regional aquifer was determined to be 2.3x10"6
ft/sec. Based on the water-level contour maps (Figure 4-7), the length from the highest data
point to the lowest data point is approximately 4,050 feet (i.e., between well AB-37-55-45D
and well AD-37-22-85I). The difference between the highest and the lowest water levels
within the watershed ranged from 155 to 170 feet during the 1983 to 1987 period.
Using lithologic information, the thickness of the regional aquifer was estimated to be
approximately 300 feet. Given that the river length in the watershed is approximately 9,500
-------
FIGURE 4-7. Fourth Quarter 1983 Potentiometric Surface Map
12000
12000
- 10000
- 8000
6000
4000
4000
4000
0000
0000
10000
12000
-------
4-19
feet, the calculated cross-sectional area is 2,850,000 square feet. To account for discharge
from both sides of the watershed, this cross-sectional area is doubled to total 5,700,00 square
feet.
Ground water discharge rates on a quarterly basis from 1983 to 19B7 were calculated using
Darcy's Law and they are listed in Table 4-1. The discharge values were calculated in cubic
feet per second (cfs) and then converted to cubic meters per second (cms). The calculated
discharge values ranged from 0.014 to 0.015 cms. A plot of discharge rate against time is
presented in Figure 4-8.
Total discharge for the watershed, measured at a stream gaging station {Table 4-1), ranges
from 0.057 to 1.24 cms. The contribution ol ground water to total discharge ranges from
1.2% to 100% (reported as 250%) of the total discharge. This high variability is caused by the
complicated hydrogeologic environment and the effects of the shallow aquifer on the total
flow system. In addition, the high ground water discharge values are probably a function of
an overestimation of the portions of the aquifer contributing to streamflow. In other words, all
portions of the aquifer may not converge with the surface water. As a result, there is a
reduction in ground water flow to the stream caused by the convergence of the ground water
flow paths from the total aquifer cross section to the portion of the aquifer intersecting the
stream channel. The high percentage values are generally associated with the fall quarter
when the shallow fractured aquifer is thought to be drained and has the (east effect upon the
regional flow system.
Part 2: Application of Ground Water Quality Measurements
Step 1: Evaluate the Spatial Coverage of Sampling Wells in the Watershed
Water quality measurements were available from nine observation wells. The spatial coverage
is adequate for estimating constituent concentration of ground water in the deeper regional
aquifer.
-------
TABLE 4-
1
Quarterly data from 1983 to 1987
DATE
dh
dl
dh/dl
k
Area
x 2
Q
Q
Total Q
Q % of
Concentration
Load
(feet)
(feet)
(ft/ft)
(ft/sec)
(Sq ft)
(cfs)
(cms)
(cms)*
Total Q
mg/l
mg/s
3/18/83
160.49
4050
0.0396
0.0000023
2850000
2
0.5195
0.0147
0.0910
16.18
2.00
29.44
6/1/83
162.34
4050
0.0401
0.0000023
2850000
2
0.5255
0.0149
0.0915
16.27
2.00
29.78
9/1/83
156.34
4050
0.0386
0.0000023
2850000
2
0.5061
0.0143
0.0187
76.68
2.00
28.68
12/10/83
162.78
4050
0.0402
0.0000023
2850000
2
0.5269
0.0149
0.1879
7.95
2.00
29.86
3/1/84
164.27
4050
0.0406
0.0000023
2850000
2
0.5317
0.0151
0.2457
6.13
2.00
30.14
6/1/84
170.22
4050
0.0420
0.0000023
2850000
2
0.5510
0.0156
0.4029
3.87
2.00
31.23
9/1/84
156.89
4050
0.0387
0.0000023
2850000
2
0.5079
0.0144
0.0133
108.06
2.00
28.78
12/1/84
157.97
4050
0.0390
0.0000023
2850000
2
0.5114
0.0145
0.2196
6.60
2.00
28.98
3/1/85
160.33
4050
0.0396
0.0000023
2850000
2
0.5190
0.0147
0.0910
16.17
2.00
29.41
6/1/85
158.25
4050
0.0391
0.0000023
2850000
2
0.5123
0.0145
0.0765
18.97
2.00
29.03
9/1/85
155.92
4050
0.0385
0.0000023
2850000
2
0.5047
0.0143
0.0108
132.62
2.00
28.60
12/10/85
158.14
4050
0.0390
0.0000023
2850000
2
0.5119
0.0145
0.0646
22.45
2.00
29.01
3/1/86
164.41
4050
0.0406
0.0000023
2850000
2
0.5322
0.0151
0.1371
11.00
2.00
30.16
6/1/86
160.22
4050
0.0396
0.0000023
2850000
2
0.5186
0.0147
0.0507
28.97
2.00
29.39
9/5/86
155.39
4050
0.0384
0.0000023
2850000
2
0.5030
0.0143
0.0632
22.56
2.00
28.51
12/1/86
160.95
4050
0.0397
0.0000023
2850000
2
0.5210
0.0148
0.1377
10.72
2.00
29.53
3/1/87
160.25
4050
0.0396
0.0000023
2850000
2
0.5187
0.0147
1.2420
1.18
2.00
29.40
6/1/87
157.78
4050
0.0390
0.0000023
2850000
2
0.5107
0.0145
0.0258
56.13
2.00
28.95
9/1/87
155.09
4050
0.0383
0.0000023
2850000
2
0.5020
0.0142
0.0057
251.02
2.00
28.45
12/1/87
162.33
4050
0.0401
0.0000023
2850000
2
0.5255
0.0149
0.4554
3.27
2.00
29.78
* From stream gaging station.
-------
FIGURE 4-8. Quarterly Ground-Water Discharge Estimates
Watershed WE-38
0.0158
0.0156
0.0154
0.0152 -
« 0.015
O 0.0148
0.0146
0.0144
0.0142 -
0.014
i—i—r
3/18/83 9/1/83
i i r
3/1/84 9/1/84 3/1/85 9/1/85 3/1/86 9/5/86
QUARTERLY DATES
i—r
3/1/87 9/1/87
-------
4-22
Step 2: Determine Temporal and Spatial Variations of Water Quality Measurements
and Estimate Average Constituent Concentration
Water-quality measurements from the nine observation wells were available from March 1983
to December 1987 on a quarterly basis. Variations of constituent concentration on a
seasonal basis are captured in this quarterly data set.
The locations of the nine observation wells cover a full range of land use types across the
watershed (e.g., forest, cropland, and meadow). Average concentration of nitrate in the
deeper aquifer is lowest beneath forest (e.g., 0.1 mg/L) and highest under cropland (e.g., 4.8
mg/L). A representative nitrate value of 2 mg/L was used to calculate contaminant loading
rates for the deeper regional aquifer.
Step 3: Determine Contaminant Loading Rate to Surface Water
Contaminant loading rates on the quarterly basis were calculated using ground water
discharge rates and the representative nitrate concentration. The calculated quarterly nitrate
loading rates ranged from approximately 28 to 32 mg/sec (see Table 4-1).
-------
APPENDIX A. ADDITIONAL ASSESSMENT METHODS
Overview of the Assessment Methods Identified by EPA
In the companion Volume to this report10, EPA presented summary descriptions of seven
general techniques that have been used to estimate the magnitude of nonpoint source load
to surface waters through contaminated ground water discharge. The following discussion
summarizes several of these methods. The base flow/hydrograph separation method and the
ground water level measurement method are not included below since they are discussed in
detail in the body of this manual. The reader is referred to the accompanying volume and
annotated bibliography for more information concerning these methods.
Seepage Meters and Mini-piezometers
Seepage meters and mini-piezometers provide a simple, direct means of measuring the
quantity and quality of ground water discharge to surface water. These methods are often
used to field verify other estimation techniques. Seepage meters and mini-piezometers may
be used to provide "point measurements" of ground water discharge and water quality.
Researchers often use 55 gallon drums cut in half as seepage meters. These drums are
partially submerged in sediment, and a vent hole with an attached plastic bag is used to
collect water samples over a measured time. The collected volume of water divided by the
cross sectional area of the drum provides a groundwater discharge velocity. Mini-
piezometers are generally constructed from pipes manually inserted into the sediment. Water
levels from mini-piezometers provide estimated hydraulic head data. Discharge rates across
the sediment-water interface are then calculated using Darcy's Law. Limitations to these
methods include the assumption that sampling points are representative of the sediments
spatially and temporally. In addition, these methods require extensive manpower to use and
maintain. Harsh seasons and strong currents may require equipment modifications.
10A Review of Methods for Assessing Nonpoint Source Contaminated Ground-Water
Discharge to Surface Water, EPA 440/6-90-007.
-------
A-2
Geophysical Methods
Geophysical methods offer a spatial survey of the hydraulic properties of sediments in a
surface water body. This type of survey offers a continuous characterization of bottom
sediments by seismic, electrical charge, resistivity, and temperature and electrical
conductance measurements. These methods require little previous information on the water
body, and generate a range of information, including estimates for aquifer thickness, seepage
zones, or conductivity changes across sediment boundaries.
Geophysical methods offer the ability to characterize properties of larger water bodies that
would be difficult to measure directly in open waters. Seismic methods offer information on
sediment type, thickness, and sequence from the travel patterns and reflection of seismic
waves sent from a ship. Electrical conductance methods charge the sediment with current,
then shut off the power to measure the rate of decay. The rate of decay indicates day
content, from which the vertical hydraulic conductivity may be estimated. Resistivity
measurements also use a source of current which enters the subsurface through point
electrodes. Resistivity measurements employ receiving electrodes as well, measuring the
potentials of the flow field and identifying the sediment's composition, which is used to
estimate the vertical hydraulic conductivity. Temperature and electrical conductance methods
use a sediment probe towed behind a boat that measures differences in temperature and
conductance between the sediment and the ground water. Anomalies indicate ground water
discharge zones.
Geophysical methods are limited by the uncertainties associated with indirect determination of
sediment properties. They offer qualitative rather than quantitative estimations and require
ground-truthing to verify results.
Regression Analysis, and Mass Balance Approaches
Regression analysis, and mass balance approaches provide estimates for the distribution of
ground water flux to an area. These methods represent an established way to determine
ground water discharge to surface water in a watershed and require accurate data on the
drainage area of the watershed, precipitation records, water holding capacity of the soil
classes in the watershed, and drainable porosity measurements. The methods use graphical
-------
A-3
analysis, established equations, and mass balance methods to estimate the ground water
loading to surface water. Base flow determined using these methods can be used in
conjunction with ground water quality measurements obtained in wells located adjacent to the
surface water body to estimate the ground water loading rate to surface water. Limitations on
regression analyses and soil moisture balances are due to their basis on one uniform region's
basin characteristics.
Numerical Models
Numerical models simulate aquifer or watershed systems using equations governing flow and
mass balance. Computer models provide a useful screening tool to assess general effects of
regulatory scenarios on ground water quality issues. Aquifer characteristics, chemical
concentrations, the geology, and hydrology of the study site require proper characterization
prior to modelling. Input requirements include the spatial distribution of hydraulic
conductivity. Modelling is a cumulative process that involves data gathering and modet
verification to ensure an accurate depiction for real world situations in the computer model.
Computer modelling requires a large amount of data for calibration. Data acquisition may
involve considerable time, expertise, and expense.
Loading Functions
This method is useful for general estimates of loadings from ground water sources using land
use descriptions, soil type descriptions, precipitation, temperature, and stream data. Loading
functions do not have many management applications due to the difficulty of predicting the
effects of changes in the input components on loading rates. Loading functions use empirical
equations to estimate runoff and ground water discharge. The method assumes that
empirical loading rates are representative lor actual conditions and that transport is not scale
related. Significant uncertainty exists in both the input parameters and in results generated.
Isotooic Methods
Isotopic tracers from naturally or anthropogenically occurring activities exist in natural waters
and can be used to identify water origin and movement. This method offers the ability to
determine actual runoff generation In a range from the macropore level to catchment slope
study areas and generates a physically based runoff separation analysis which indicates
-------
A-4
contributing source of water and travel paths. The model used to identify these data
assumes that water is a mixture of old versus new water, the former coming from the current
precipitation event, the latter existing in the catchment prior to the event. Disadvantages to
using tracers include the need for conditions where old and new water does not significantly
differ and the high cost of sample analysis.
-------
APPENDIX B. GLOSSARY
Anisotropic - A condition in which the hydraulic properties of an aquifer are dissimilar in
different directions.
Antecedent conditions - Initial soil moisture conditions.
Aquifer - A geologic formation that is saturated and that can store, transmit, and yield
significant quantities of water.
Base flow - Streamflow due to soil moisture or ground water.
Darcy1s Law - A formula for the flow of fluids through porous media based on the
assumptions that the medium is isotropic, homogeneous, and the flow is one dimensional.
Drainable porosity - Fraction of the soil volume that drains as the water table descends.
Ephemeral stream - A stream or portion of a stream which flows only in direct response to
precipitation.
Fractured system - Geologic system broken by interconnecting cracks. May be due to
intense folding or faulting. Also a common structure in limestone oil reservoirs.
Heterogeneous - A characteristic of a medium which signifies that the medium has properties
which vary with position within the medium.
Homogeneous - A characteristic of a medium which signifies ttiat the physical property of
every element of volume has the same value regardless of location.
Hydrograph - A graph of stream discharge versus time.
-------
B-2
Infiltration -- Movement of water from the surface into the soil.
Interflow -- Flow of water through the upper soil zones to a stream.
Isotropic - A condition in which hydraulic properties of an aquifer are equal in all directions.
Karst - A type of topography that is formed over limestone, dolomite, or gypsum by
dissolving or solution, and that is characterized by closed depressions or sinkholes, caves,
and underground drainage.
Percolation -- Drainage through the unsaturated zone.
Piezometer - Small diameter well that measures water table or confined aquifer pressure.
Recession Curve - Portion of the hydrograph where runoff is contributed by base flow.
Saturated hydraulic conductivity - A proportionality constant indicating the permeability of a
porous media.
-------
APPENDIX D. Reference Graph and Table for Hydraulic Conductivity
and Permeability
-------
This page intentionally left blank.
-------
Range of Values of Hydraulic Conductivity and Permeability
Rocks
a
s
2
- [
¦a "q I
2 e «
at ojt
£ *» y
»- 3 2
« o
1 « U
.? a-O
"O O O
2 E w •
a2"
S V
y4»Sl?
o
w
U.
c
o
o
§
l/l
c
a
01
"O y
c
e
in
CO
w
«
o
(A
"S
I|8
T. a. «»
llil
c § 8?
=fi
I
J i
fi2
£ UO
is
k k K K K
(darcy) (cm2) (cm/s) (m/s) (gal/day/ft2)
-to5
-f0"3
-to2
-»
-I04
-to"4
-10
-10"'
-10®
-103
-to*5
-t
-I0~2
-105
MO2
-I0*6
-10"'
-10"5
-104
-10
-10-7
-10*2
-I0"4
-103
¦ t
-I0'B
-f0"s
-10'9
-102
-10"'
-10"9
-10"4
-I0"6
-10
V
o
l_
-10-'°
- 10"S
-10-7
-1
¦10"S
-10""
-10"*
-10"*
-10"'
-10"4
M
T
0
1
-10"7
-10"*
-10"2
-I0"5
-10"»
-10"*
-JO"10
- 10"3
-10"*
-10"14
-10**
- 10"u
-10"4
-10-7
-10"15
-10"0
-I0"tt
-10"9
* to"*
-to"**
-10*M
¦10"'1
-10"#
10
-7
Source: Freeze, R.A. and J.A. Cherry. 1979. Groundwater. Englewood
Cliffs, New Jersey: Prentice-Hell. p. 29.
-------
Conversion Factors for Permeability and Hydraulic Conductivity Units
Permeability, k* Hydraulic conductivity. K
cm2
ft2
darcy
m/s
ft/s
U.S. gal/day/fl2
cm1
1
1.08 x 10->
1.01 x 10«
9.80 x 10*
3.22 x 10'
1.85 x 10»
ft2
9.29 x 102
1
9.42 x I0*»
9.11 x 10*
2.99 x I0«
1.71 x I0»2
darcy
9.87 x 10-'
1.06 x 10-"
1
9.66 x 10"«
3.17 x 10-3
1.82 x 10'
nvs-
1.02 x I0">
1.10 x 10~«
1.04 x I0»
1
3.28
2.12 x 10«
ft/s
3.11 x 10~«
3.35 x 10-'
3.15 x 104
3.05 x 10"'
1
6.46 x 10'
U.S. gal/day/5.42 x 10~»®
5.83 x 10-"
5.49 x 10"2
4.72 x 10"7
1.55 x 10"«
1
~To obtain k in ft2, multiply k in cm2 by 1.08 x 10"3.
Source: Freeze, R.A. and J.A. Cherry. 1979. Groundwater. Englewood Cliffs, New Jersey: Prentice-Halt,
p. 29.
-------
APPENDIX E. Research Paper on Methods for Estimating Recharge
The following research paper describes three computerized methods used by the U.S.
Geological Survey to estimate long-term mean total recharge and effective recharge (i.e., the
contribution to streamflow from ground water discharge) using daily mean streamflow
records. This paper is included in this manual with the permission of the American Water
Resources Associaton.
-------
This page intentionally left blank.
-------
Regional Aquifer Systems of the United States
AQUIFERS OF THE
SOUTHERN AND EASTERN STATES
Edited by
William R. Hotchkiss
Office of the Regional Hydrologist
U.S. Geological Survey, WRD
Lakewood, Colorado 80225
and
A. Ivan Johnson
A. Ivan Johnson, Inc.
7474 Upham Court
Arvada, Colorado 80003
Papers presented at
AWRA 27th Annual Conference and Symposium
September 8-13,1991
New Orleans, Louisiana
AWRA MONOGRAPH SERIES NO. 17
AMERICAN WATER RESOURCES ASSOCIATION
5410 Grosvenor Lane, Suite 220
Bethesda, Maryland 20814-2192
Telephone: (301) 493-8600
-------
This page intentionally left blank.
-------
Regional Aquifer Systems of the United States
AQUIFERS OF THE SOUTHERN AND EASTERN STATES
AMERICAN WATER RESOURCES ASSOCIATION
METHODS OF USING STREAMFLOW RECORDS FOR ESTIMATING TOTAL AND
EFFECTIVE RECHARGE IN THE APPALACHIAN VALLEY AND RIDGE,
PIEDMONT, AND BLUE RIDGE PHYSIOGRAPHIC PROVINCES
A.T. Rutledge 1
ABSTRACT: The U.S. Geological Survey's Appalachian Valleys-Piedmont Regional Aquifer-System Analysis study uses
records of daily mean streamflow to estimate long-term mean total recharge and effective recharge in basins where
streamflow is unregulated. Three computerized methods employed are (1) the streamflow-partitioning method, (2) the
recession-curve displacement method, and (3) the ground-water-level rating-curve method. Streamflow partitioning, a
method of estimating a daily record of ground-water discharge on the basis of the daily record of streamflow, yields
an estimate of effective recharge (total recharge minus riparian evapotranspiration). The recession-curve
displacement method yields an estimate of total recharge by subtraction of the total potential ground-water
discharge before a recharge event from the total potential ground-water discharge after the event. The ground-
water-level rating-curve method, which is based on an empirically derived relation between streamflow during times
of negligible direct runoff and water level in a nearby well, yields an estimate of total recharge if a single
rating curve for winter data is used or an estimate of effective recharge if a separate rating curve for summer data
also is used.
INTRODUCTION
In the U.S. Geological Survey's (USGS) regional aquifer-system analysis of the Appalachian Valley and Ridge,
Piedmont, and Blue Ridge physiographic provinces (APRASA) (Swain and others, 1991), streamflow records are used to
estimate total and effective ground-water recharge. Streamflow records are being analyzed to (1) demonstrate the
role of hydrogeologic factors in controlling the processes of recharge, discharge, and ground-water flow; and (2)
estimate regional components of the ground-water budget. Data from approximately 200 streamflow-gaging stations
(fig. 1) were selected for analysis. The general criteria for selection of gaging stations are: (1) drainage area
less than 500 mi1 (square miles), (2) period of record that includes the entire 1980's, (3) most of the record clas-
sified as "good" in USGS data reports, and (4) no significant regulation or diversion of flow. All drainage areas
exceed 2 mi1 and 93 percent exceed 10 mi1.
The principal components of the ground-water budget in these basins are total recharge, effective recharge, and
riparian evapotranspiration. In the APRASA study area, most recharge results from downward percolation of water
through the unsaturated zone during or immediately after precipitation. Recharge also can result from the infiltra-
tion of melting ice and snow and, in some localities, by infiltration of water from losing streams. Most discharge
from the saturated zone takes place at discharge points that are at or near streams and eventually contributes to
streamflow as ground-watar discharge (base flow). The contribution to streamflow from ground-water discharge is
thus referred to as effective recharge. Effective recharge is less than total recharge because of the loss to the
atmosphere of water from stream channels and from the saturated zone near stream channels. This loss is referred to
as riparian evapotranspiration.
It is assumed that interbasin ground-water flow, underflow near the streamflow-gaging station, deep flow to
regional flow systems, consumptive use of ground water, and recharge to the aquifer through the streambed arc
negligible. The time periods of analysis are long enough (several years) so that changes in aquifer storage are
considered negligible. Calculations of total or effective recharge in units of specific discharge (for example,
inches per year) require that the drainage area be known and that drainage and ground-water divides coincide.
1 Hydrologist, US. Geological Survey, 3600 West Broad Street, Room 606, Richmond, VA 23230
59
-------
LAKE
(MICHIGAN,
Figure 1.—The study area of the regional aquifer-system analysis of the Appalachien
Valley and Ridge, Piedmont, and Blue Ridge physiographic provinces and drainage
areas contributing to flow at gaging stations for which streamflow records are
analyzed.
60
-------
The purpose of this paper is to describe three methods that are being used by APRASA: the streamflow-
partitioning method, which gives an estimate of effective recharge; the recession-curve-displacement method, which
gives an estimate of total recharge; and the ground-water-level rating-curve method, which gives estimates of total
and effective recharge. Although use of these methods entails calculation of ground-water discharge for each day or
for each event, the intended use is for the determination of the long-term (multiple year) mean. Application to a
smaller time scale may result in substantial errors, The results represent the integration of processes throughout
the entire drainage area and thus will give the mean for the basin.
METHODS FOR ESTIMATING TOTAL AND EFFECTIVE RECHARGE
The methods described in this paper are executed by Fortran-77 computer programs that read data files of daily
mean streamflow. Before the programs are executed, the streamflow data (in addition to ground-water-level data in
the case of one method) must first be obtained from USGS records and assembled in a standard format. Streamflow
data must consist of complete calendar years of record (365 values per year), but ground-water-level records can be
intermittent. Some of the programs create simple graphical representations on the computer monitor that do not
require "graphics hardware or software" because they consist of alphanumeric characters. All programs create
summary output files and detailed output files that allow the user to check calculations and create detailed
graphics of results by means of separate graphics software.
Streamflow-Partitioning Method
Streamflow partitioning is a method of estimating a continuous record of ground-water discharge, or base flow,
under the streamflow hydrograph. The integration of this record over a period of several years provides an estimate
of effective recharge.
Previous Work
Many different techniques, most of which involve considerable subjectivity in their application, have been used
to separate streamflow into components, and have been referred to as "hydrograph separation." Horton (1933)
described a method of shifting a "normal depletion-curve" horizontally across a hydrograph, noting that segments of
the hydrograph that coincide with this curve represent periods during which streamflow is equal to ground-water
discharge, then estimating ground-water discharge during periods of surface runoff by timply connecting the points
where the hydrograph departs from the normal depletion curve. Barnes (1939) separated surface flow, storm seepage,
and base flow by assigning to each a distinct "depletion factor." Kulandaiswamy and Seetharaman (1969) pointed out
that Barnes' method may not be reliable for separating streamflow into three parts, but that the hydrograph can be
separated into direct runoff and base flow. In some applications, investigators have used characteristic base-flow-
recession curves of the stream in combination with records of precipitation, snowfall, temperature, and ground-water
levels (Olmsted and Hdy, 1962).
In many cases, the method can be desoibed as two steps: (1) locating time periods of negligible surface runoff
and designating that ground-water discharge equals streamflow, and (2) interpolating ground-water discharge between
these periods. The declaim that surface runoff is negligible can be baaed on antecedent recession. Linsley and
others (1982) used the empirical relation
N-A*-* (1)
for determining the time base of surface runoff (NMhc number of days after a peak in the hydrograph whan suiface
runoff (including the bulk of interflow) oases is a function of the drainage area (M, in square mflas. For a
given day, the antecedent recession requirement is met if recession has been continuous for N days or more preceding
the day. Other workers use linearity of recession, on the graph of logarithm of streamflow as a function of time,
as the indicator that ground-water discharge equals streamflow (Ineson and Downing, 1964). There are many methods
for executing step 2 (Snyder, 1939; Chow, 1961; Pettyjohn and Henning, 1979; and Nathan and McMahon, 1990); some are
based on the assumption that the ground-water discharge peak is concurrent with the streamflow peak, and others are
based on the assumption that the recession of ground-water discharge continues after the time when surface runoff
61
-------
begins—that is, the response of ground-water discharge is delayed relative to surface runoff. Some methods involve
simple linear interpolation to estimate ground-water discharge between the start and the end of surface runoff.
Hydrograph separation has been applied in the APRASA study area for a stream in the Piedmont physiographic
province in Pennsylvania (Olmsted and Hely, 1962), one stream in the Blue Ridge physiographic province in Georgia
(Stewart and others, 1964) and two streams in the Valley and Ridge physiographic province in Pennsylvania (Becher
and Root, 1981).
Computer techniques have been used to separate hydrographs and have allowed for increased speed of analysis and
repeatability of results because of the removal of subjectivity. Hewlett and Hibbert (1967) used a computer program
for separating "quick flow" from "delayed flow" in streams of the eastern United States. Pettyjohn and Henning
(1979) used a method that involves searching for the minimum streamflow along an interval of the record that is 2N
days in duration and assigning values of daily base flow by three different techniques. A method developed by the
Institute of Hydrology (1980) includes searching for the minimum streamflow along 5-day nonoverlapping periods
searching the series of minimums for values that are less than 0.9 times the two outer values, and defining such
values as turning points. The base-flow hydrograph is constructed by connecting all the turning points. Thjs
method and a digital filter method used in signal analysis (Lyne and Hollick, 1979) are discussed by Nathan and
McMahon (1990).
A method called streamflow partitioning (Knisel and Sheridan, 1983; Shirmohammadi and others, 1954.
Shirmohammadi and others, 1987) requires daily records of streamflow and precipitation and the designation of a
threshold daily precipitation. For a given day, base flow equals streamflow if, on that day and all days that
precede it by N days or less, precipitation is less than the threshold value. Base flow on all other days is
estimated by linear interpolation.
Description of Computer Program
The method of streamflow partitioning described here includes two steps: (1) Ground-water discharge is set
equal to streamflow during times of negligible surface runoff, and (2) ground-water discharge between these periods
(during surface runoff) is interpolated. The method is similar to that of other investigators (Knisel and Sheridan
1983; Shirmohammadi and others, 1984; Shirmohammadi and others, 1987) in that daily values of streamflow are used
and that step 2 is executed by linear interpolation. It differs in the method of step 1; although the previous
investigators used antecedent precipitation to indicate that ground-water discharge equals streamflow, the author
uses antecedent streamflow recession.
A flow diagram showing the steps in the streamflow-partitioning program is shown in figure 2. The program
fills a one-dimensional array of daily mean streamflow data, then searches this array for days that fit an antece-
dent recession requirement. On each of these days, ground-water discharge is designated equal to streamflow as lone
as it is not followed by a daily decline of more than 0.1 log cycle. It can be inferred from Barnes (1939) that a
daily decline more than 0.1 log cycle could indicate interflow (stormflow) or surface flow. The program searches
the array again, determining by linear interpolation the ground-water discharge on the remaining days. (For some
streamflow records, this interpolation may cause the calculated ground-water discharge to exceed streamflow for a
few days in the record. The last step corrects for this.)
After the array of daily ground-water discharge has been filled, the program user can review the results. The
program creates graphical representations and tables on the computer monitor that show streamflow and ground-water
discharge for each day in the period analyzed. If the user chooses, an output data file can be created that lists
streamflow and ground-water discharge for each day. Plots of output data files, which are hydrographs of daily
streamflow and daily estimated ground-water discharge for two streams in the study area, are shown in figure 3
Final output includes the addition of one line to the bottom of a file. This line is a summary of the session
including stream data file name, station number, drainage area, and estimated effective recharge.
Rfgarion.Ctirve-Di«pla«ment Method
The recession-curve-displacement method is based on the upward shift in the streamflow-recession curve that
occurs as a result of a recharge event The basis of the method is that a recharge event will increase the total
potential ground-water discharge, V; which is the total volume of water that will drain from the system if allowed
to do so for infinite time without further recharge.
62
-------
EXPLANATION
The antecedent recession requirement is ait for eh* day in question if,
for the part of eha daily Man atraaaflow record that includaa all days
that precede tha day in quaatlon by N daya or laaa, tha atraaaflow on
aach of thasa daya la graatar than or equal to tha atraaaflov on tha day
that follows it. (N- tiae basa of aurfaca runoff.)
Tha antlra proeadura ia axaeutad for thraa valuaa of N: ona la tha naxt
intagar analler than tha raault of aquation 1, and tha othar two ara
tha naxt two integers that ara largar than tha raault of aquation 1.
Curvilinaar interpolation gives tha final eatiaata of basa flow that
corresponds to the preciae rasult of aquation 1.
Figure 2-- Flow diagraa ahoving tha proeadura of atraaaflow partitioning.
(Baaa flow la considered to be ground-water diacharge.)
63
-------
10.000
Sewee Creek neer Decatur, Tenneaeee
(Valley and Ridge physiographic province)
Drainage area -117 square mile*
1.000
a
z
o
o
UJ
CO
a
UJ
&
w
UJ
Ui
u.
100
' \ Estimated dally nun
ground-water discharge
10
MAY
JUNE
JULY
AUGUST SEPTEMBER QCTcTbTr"
£D
q 10,000
T
Ui
O
a
<
x
o
CO
Q
Indian Crtak noar Laboratory, North Carolina
(Piedmont physiographic province)
Drainage area - 69.2 square milea
1.000 -
Oaiiy moan
'Streamflow
100 V
10
Figure 3.--Hydrographs showing the results of streamflow partitioning. (Where th#
curves coincide, ground-water discharge equals streamflow. For both
hydrographs. the time base of surface runoff is 3 days.)
64
-------
Theory and Previous Work
Meyboom (1961) expressed V in the following expression, which is based upon a linear relation between the
logarithm of ground-water discharge and time:
V = OxK ¦ (2)
2.3026
where
V is total potential ground-water discharge (L5)
Q is ground-water discharge at initial time (L' /T), and
K is the storage delay factor (Singh and Stall, 1971), which is the time for ground-water
discharge to decline through one log cycle (T).
Rorabaugh (1964) expressed ground-water discharge to a stream as a complex function of time after the recharge
event; he found, however, that the function can be approximated after "critical time" by an equation that expresses
the logarithm of ground-water discharge as a linear function of time. Critical time is expressed by the following
equation (Rorabaugh, 1964):
CT = 0.2xa' xS, (3)
where
CT is critical time (T),
a is the average distance from the stream to the hydrologic divide (L),
S is the storage coefficient, and
T is transmissivity (L1 /T).
A formulation that gives critical time as a function of the streamflow recession rate can be obtained by com-
bining equation 3 with the following equation from Rorabaugh and Simons (1966):
a'xSs K (4)
T 0.933
to obtain
CT * 0-2144xK. (5)
The formulations of Glover (1964) and Rorabaugh (1964) show that the total potential ground-water discharge at
critical time is equal to approximately one-half of the total volume of water that recharged the system during the
event This finding, combined with the principle of superposition, is the basis for the method: at critical time
after an event, the potential ground-water discharge minus the potential ground-water discharge had the event not
occurred is equal to one-half of the recharge that occurred during the event Total recharge is thus calculated by
use of the following equation:
R«2y(Q2-Ql)*K, (6)
2.3026
where
R is total recharge due to the event (L *),
Q1 is ground-water discharge at critical time as extrapolated from the pre-event
streamflow recession (L1 /T), and
Q2 is ground-water discharge at critical time as extrapolated from the post-event
streamflow recession (L' /T).
65
-------
The recession-curve-displacement method can be executed by the following steps, as shown in figure 4 (from
Bevans, 1986): (1) determining the storage delay factor (K) from the hydrograph, (2) calculating critical time (CT)
from equation 5, (3) using critical time to determine the time on the hydrograph to which streamflow recessions will
be extrapolated, (4) determining the hypothetical ground-water discharge at critical time by extrapolation of the
pre-event recession curve, (5) determining the hypothetical ground-water discharge at critical time by extrapolation
of the post-event recession curve, and (6) applying equation 6. In practice, the user identifies the position of
the post-event recession curve soon after surface runoff ceases so that the effects of riparian evapotranspiration
are minimized; thus, the result is an estimate of total recharge- The recession-curve-displacement method has beer,
applied in the APRASA study area for three streams in the Piedmont physiographic province in North Carolina (Daniel
and Sharpless, 1983) and for a stream in the Valley and Ridge physiographic province in Pennsylvania (Gerhart and
Lazorchick, 1988).
Description of Computer Program
The computer program locates days in the streamflow record that fit an antecedent recession requirement (eq l
rounded to the next larger integer). Consecutive days that fit this requirement are considered periods of ground-
water-flow recession. The program defines a peak as the largest streamflow between two consecutive periods of
ground-water-flow recession. Each peak is considered a recharge event. The total recharge is calculated for each
peak by use of procedures diagrammed in figure 5.
A critical aspect of the method is the extrapolation (in time) of ground-water discharge from periods of
ground-water-flow recession to points in time that are outside of these periods of recession. Before critical time
this equation is used: '
dQ = _£_, {7)
J3T
where
dQ is the difference between the ground-water discharge and the ground-water discharge
that would have occurred at the same time in the absence of the recharge event,
C is a constant, the value of which is dependent on the magnitude of the recharge event, and
dt is the time since the recharge event.
Equation 7, which is a simplified version of Rcrrabaugh's equation 4 (1964), quantifies the margin of flQw
resulting from the recharge event, which is superimposed on the flow that would have occurred in the absence of the
event Equation 7 is used to extrapolate the margin of flow at critical time by solving for C for each day 0f
period of ground-water recession that occur* before critical time, obtaining the average, and substituting this C
and critical time back into the equation. After critical time, extrapolation is done by use of an equation of th
form:
0 v,n(-dt/K)
-------
10,000
1,000 -
o
o
u>
CO
IE
UJ
&
H
UJ
Ul
u.
u
s
o
— 100 -
ni of Interest
J
O
<
UJ
S
h*
trs
z
<
UJ
2
<
a
1.0
PROCEDURE
1. Compute storage delay factor,
K (32 d/log cycle)
2. Compute critical time, CT
(0.2144 *K, or 6.86 d)
3. Locate time that is
6.B6 days after peak
4. Extrapolate pre-event
recession (5 ft 3/s)
5. Extrapolate post-event
recession (23 ft 3/sl
6. Compute total recharge,
2 '(IB ft 3/s3- 32 d 8 64 0 0 a
2.3026 ' 1 d
- 4.32 • 107 ft3
10 20
MARCH 1974
31
EXPLANATION
Daily streamflow
—— Extrapolated ground-
water discharge
ft Fe«t
s Seconds
d Days
Figure 4.—Procedure for using the recession-curve displacement method to estimate total
recharge in response to a recharge event [modified from Bevans, 1986, fig. 5).
67
-------
Figura 5.—Flow diagram (right) and schematic of hydrograph (abova)
showing tha proeaduraa axacutad for aach paafc in tha
raeaasion-curva diaplacamant mathod.
68
-------
EXPLANATION
Storage daisy factor (time par log cycle of recession after critical tims)
Critical tima
Paak counter— Tha abova calculations pertain to the "present peak," which
ia peak I
The time of occurrence of peak I
A constant (fox peek I) that ia used in the aquation for the margin of flow
due to peak Z up to tha critical time after peak I
The flow and tiaa at critical tine after the last paak <2-1) that would
have oecurad ia tha absence of all peaks subsequent to the last paak— one
exception is that if 1-1, than QA(X) and TA(I) are tha flow and tiaa on the
lest day of tha recession period before tha first paak
Tha flew and tiaa at critical tiaa after tha present paak (Z) that would
have oecurad in tha absence of tha present paak and all subsequent peaks
The flew and tisw at critical tiaa after the present peak (Z) that would
have oecurad in tha absence of all paaka subsequent to tha present paak
Difference between the Measured flow during the recession period after
peak X and tha flew that would have occured in the absence ef peak Z
Tiaa increment between tbe paak and a day daring the recession period
Margin ef flew at eritieal tiaa after peak Z that oeeurs because ef peak Z
only
On the hydro graph, the flow on a day during the period of ground-water
flow recession that follows peak Z
K
CT
I
TP (Z)
C(X)
OA(I), TA(Z)
OB (I), TBC(Z)
QC, TBC(I)
dQ
dT
AQ
69
-------
Final program output includes a file'that lists details about each set of calculations of total recharee
including most of the variables shown in figure 5. A summary of the results of the program session is added to the
bottom of another file that includes summaries for all other sessions that have been run.
Ground-Water-Level Rating-Curve Method
The ground-water-level rating-curve method is a means of estimating a daily record of ground-water discharge b
obtaining a rating between ground-water level and ground-water discharge. Data used for constructing the rating ai?
obtained for time periods when surface runoff can be considered negligible.
Previous Work
In a study of a basin in the Piedmont province in Maryland, Han-old (1934) found that, for short periods of
time, it is possible to make a close estimate of the discharge of the stream from the water level in a well nearb
but found that the relation between ground-water level and streamflow varies with the seasons. Rasmussen and
Andreason (1959) applied the method for a stream in the Coastal Plain of Maryland using the weekly average ground
water levels measured in 25 wells in the basin. A similar investigation of a stream in Dlinois (Schicht anJ
Walton, 1961; Walton, 1962) was based on daily mean ground-water levels in five observation wells and revealed tw
distinct ratings: one for periods of small evapotranspiration and one for periods of appreciable evapotrans -
ration. The mean daily ground-water evapo transpiration was estimated from the difference between the two curv
lri results from the Maryland and Illinois investigations, the relation between ground-water level and ground-wat
discharge is curvilinear: as ground-water level increases, the increment of ground-water discharge per increment
ground-water level increases.
An investigation in the APRASA study area in the Piedmont of Pennsylvania involved the selection of three ind
wells based on comparison of ground-water-level hydrographs with the streamflow hydrogTaph {Olmsted and Hely, 190?
Using winter data, the investigators found a linear relation between the monthly composite ground-water level and
the monthly average ground-water discharge. The investigators state that the relation could have been nonlinear f
different wells were selected, and they devote considerable discussion to the shape of the relation line betw
ground-water level and ground-water discharge, pointing out that other studies have shown curvature. Olmsted
Hely estimated monthly evapotranspiration from ground water by subtracting the ground-water discharge computed K
hydrograph separation from the ground-water discharge computed from ground-water levels using a ma the ma h ^
expression for the straight line representing the winter data. They found that data points representing suirun
months also lie near a straight line that is similar in slope but displaced (decreased ground-water discharge)
respect to the winter data.
Description of Computer Programs
The method described here for applying the ground-water-level rating-curve method involves the conaecuti
execution of two computer programs. The first program reads a data file of daily mean streamflow and another ftu
of daily ground-water level in a well that is in or dose to the drainage area. The program locates days in u!!
streamflow record that fit an antecedent recession requirement (eq. 1, rounded to next larger integer). For thas*
days, the streamflow (ground-water discharge) and the ground-water level (if it exists) are written to an outn.
file. ^
Before execution of the second program, the user must review the output data file created by the first program
and determine the best-fit aquation relating ground-water discharge to ground-water levels. An example equation
might be
Q»AxL* +BxL + C,
where
Q is the ground-water discharge (L5 /T),
L is the ground-water level in the well (L), and
70
-------
A, B, and C are coefficients, the value* of which are determined by least-squares regression. A convenient
method for obtaining this equation is to use a graphics software package that enables the user to screen the data:
before determining the best-fit equation, data points that are "outliers" can be eliminated. These outliers can
exist because the water level in the well can be affected by intermittent ground-water pumpage from nearby wells.
After the best-fit equation is found, coefficients A and B are written to a data file to be read by the second
program.
The second program reads the data file of ground-water-level data and the file containing the coefficients of
the best-fit equation. The program then calculates ground-water discharge for each day of record.
The ground-water-level rating-curve method can be used to estimate total and effective recharge. If the first
program and the subsequent determinations of the best-fit equation are executed twice—once for summer data and once
for winter data-then two sets of coefficients can be written to the data file that is read by the second program.
Two rating curves are thus available: one obtained from winter data for times of negligible riparian evapotranspi-
ration and one obtained from summer data for times of peak riparian evapotranspiration. The second program can
estimate total recharge by calculating ground-water discharge for each day of ground-water-level data solely on the
basis of the winter rating curve. The second program can estimate effective recharge if the winter rating curve is
used for winter ground-water levels and the summer rating curve is used for summer ground-water levels; both rating
curves are used for spring and fall ground-water levels, and the two sets of results are averaged.
Final program output includes a summary of the session that shows the estimated total and effective recharge
and the coefficients that were used. This output file also shows the corresponding results from all other program
sessions.
The introduction states that the methods described in this paper are to be used to determine total or effective
recharge only on a long-term basis. This is particularly the case for the ground-water-level rating-curve method
because this method is based on a correlation that exhibits considerable scatter. Errors evident at a small time
scale tend to cancel out when a large time scale is considered.
The ground-water-level rating-curve method cannot be applied to all streamflow-gaging stations because of the
long period of analysis required for APRASA (usually more than 10 years) and the scarcity of suitable ground-water-
level data. Many ground-water-level records cannot be used because the location of the well is too far from the
basin of interest or because of the short record at ground-water-level data. The method nonetheless can serve as a
verification of results of other methods, where it is applicable.
FUTURE STUDIES
Estimates of total and effective recharge will be used in APRASA for assessing the hydrogeologic controls
on basin water budgets. In addition, these estimates will help to determine the regional water budget for the
APRASA area.
The use of three different methods will allow for verification of results. Because the ground-water-level
rating-curve method give* estimates of both effective and total recharge, it* results can be compared to those of
the streamflow-partitioning or recession-curve-displacement methods. Although the streamflow-pamboning and
recession-curve-diiplacement methods give estimates of different components of the ground-water budget (effective
recharge and total recharge, respectively), results can be compared. The difference between effective recharge and
total recharge, which is riparian evapotranspiration, can be estimated from the difference between the estimates of
total and effective recharge from the ground-water4evd rating-curve method. Another estimate of riparian
evapotranspiration em be obtained from analyst of the difference between summer and winter ttntaSow recession.
Daniel (1976) quantified ground-water evapotranspiration from the departure from linearity of the summer recession
curve relative to the linear winter recession curve. An empirical method for obtaining the mathematical expression
for strcamflow recession (Rutledge, 1991) may be employed If summer and winter rwceseton curves are both nonlinear
LITERATURE CITED
Barnes, B5., 1939. The Structure of Discharge Recession Curves. Transactions of American Geophysical Union 20:
pp. 721-725.
71
-------
Becher, Albert E., and Samuel I. Root, 1981. Groundwater and Geology of the Cumberland Valley, Cumberland County
Pennsylvania. Pennsylvania Geological Survey, 4th series, Water Resources Report 50,95 pp.
Bevans, Hugh E., 1986. Estimating Stream-Aquifer Interactions in Coal Areas of Eastern Kansas by Using Streamflow
Records. In Selected Papers in the Hydrologic Sciences, Seymour Subitzky (Editor). US. Geological Survev
Water-Supply Paper 2290, pp. 51-64.
Chow, V.T., 1964. Handbook of Applied Hydrology. McGraw-Hill, New York, NY.
Daniel, Charles C. HI, and N. Bonar Sharpless, 1983. Ground-Water Supply Potential and Procedures for Well-Site
Selection in the Upper Cape Fear River Basin, North Carolina. North Carolina Department of Natural Resources
and Community Development and U.S. Water Resources Council in cooperation with U.S. Geological Survey, 73 pp
Daniel, James F.( 1976. Estimating Groundwater Evapotranspiration from Streamflow Records. Water Resources
Research 12 (3): pp. 360-364.
Fenneman, N.M., 1946. Physical Divisions of the United States. U.S. Geological Survey Map, scale 1:7,000,000.
Gerhart, James M. and George J. Lazorchick, 1988. Evaluation of the Ground-Water Resources of the Lower Susquehanna
River Basin, Pennsylvania and Maryland. U.S. Geological Survey Water-Supply Paper 2284,128 pp.
Glover, R.E., 1964. Ground-Water Movement. U.S. Bureau of Reclamation Engineering Monograph Series No in
pp. 31-34.
Harrold, L.L, 1934. Relation of Stream-Flow to Ground-Water Levels. Transactions of American Geophysical Uni
15: pp. 414-416
Hewlett, John D. and Alden R. Hibbert, 1967. Factors Affecting the Response of Small Watersheds to Precipitation '
Humid Areas. In International Symposium on Forest Hydrology, W.E. Sopper and H.W. Lull (Editors). Pereamo
New York, NY, pp. 275-290. '
Horton, Robert E. 1933. The Role of Infiltration in the Hydrologic Cycle. Transactions of American Geophvsi 1
Union 14: 446-460.
Ineson, J. and R.A. Downing, 1964. The Ground-Water Component of River Discharge and its Relationahin to
Hydrogeology. Journal of die Institution of Water Engineers 18 (7): pp. 519-541.
Institute of Hydrology, 1980. Low Flow Studies. Research Report 1, Institute of Hydrology, Wallingford, U. K.
Knisel, W.G. and J.M. Shofdan, 1983. Procedure for Characterizing Hydrologic Processes in the Coastal PUin Qf ^
Southeastern United States. In Proceedings of the UNESCO International Program on Hydrology on u.
Flatlands, Buenoa Aim, Argentina, Volume 1: pp. 191-211.
Kulandaiswamy, V.C. and S. Seetharaman, 1969. A Note on Barnes' Method of Hydrograph Separation. Journal «#
Hydrology 9: pp. 222-229.
Linsley, Ray K., Max A. Kohler, and Joseph LH. Paulhus, 1982. Hydrology for Engineers (3d ed.) McGraw-Hill w-,
York, NY, 906 pp.
Lyne, V. and M. Hollick, 1979. Stochastic Time-Variable Rainfall-Runoff Modelling. In Hydrology and Water Resources
Symposium, Perth, 1979, Proceedings: National Committee on Hydrology and Water Resources of the In*itution of
Engineers, Australia, pp. 69-92.
72
-------
Meyboom, P., 1961. Estimating Ground-Water Recharge from Stream Hydrographs. Journal of Geophysical Research 66
(4): pp. 1203-1214.
Nathan, R.J. and T.A. McMahon, 1990. Evaluation of Automated Techniques for Base Flow and Recession Analysis.
Water Resources Research 2617): pp. 1465-1473.
Olmsted, F.H. and A.G. Hely, 1962. Relation Between Ground Water and Surface Water in Brandywine Creek Basin
Pennsylvania. U.S. Geological Survey Professional Paper 417-A, 21 pp.
Pettyjohn, Wayne A. and Roger Henning, 1979. Preliminary Estimate of Ground-Water Recharge Rates, Related
Streamflow and Water Quality in Ohio. Ohio State University Water Resources Center, Project Completion Report
No. 552, Columbus, Ohio, 323 pp.
Rasmussen, William C. and Gordon E. Andreasen, 1959. Hydrologic Budget of the Beaverdam Creek Basin Maryland. U.S.
Geological Survey Water-Supply Paper 1472,106 pp.
Rorabaugh, M.I., 1964. Estimating Changes in Bank Storage and Ground-Water Contribution to Streamflow.
International Association of Scientific Hydrology, Publication no. 63, pp. 432-441.
Rorabaugh, M.I. and W.D. Simons, 1966. Exploration of Methods Relating Ground Water to Surface Water, Columbia
River Basin-Second Phase. U.S. Geological Survey Open-File Report, 62 pp.
Rutledge, A.T., 1991. A New Method for Calculating a Mathematical Expression for Streamflow Recession. In National
Conference on Irrigation and Drainage, W.F. Ritter (Editor). Irrigation and Drainage Division of the American
Society of Civil Engineers, Honolulu, Hawaii, 1991, pp. 337-343.
Schicht, R.J. and W.C. Walton, 1961. Hydrologic Budgets for Three Small Watersheds in Illinois. Illinois State
Water Survey Report of Investigation 40.
Shirmohammadi, Adel, W.G. Knisel, and J.M. Sheridan, 1984. An Approximate Method for Partitioning Daily Stream/low
Data. Journal of Hydrology 74; pp. 335-354.
Shirmohammadi, Adel, J.M. Sheridan, and W.G. Knisel, 1987. Regional Application of an Approximate Streamflow
Partitioning Method. Water Resources Bulletin 23(1): pp. 103-111.
Singh, Krishan P. and John B. Stall, 1971. Derivation of Base Flow Recession Curves and Parameters. Water
Resources Research 7 (2): pp. 292-303.
Snyder, Franklin F., 1939. A Conception of Runoff-Phenomena. Transactions of American Geophysical Union 20:
pp. 725-738.
Stewart, J.W., J.T. Callahan, R. F. Carter, and others, 1964. Geologic and Hydrologic Investigation at the Site of
the Georgia Nuclear Laboratory, Dawson County, Georgia. U.S. Geological Survey Bulletin 1133-F, 90 pp.
Swain, L.A., E.F. Hollyday, C.C. Daniel m, and OS. Zapecza, 1991. Plan of Study for the Regional Aquifer-System
Analysis of the Appalachian Valley and Ridge, Piedmont, and Blue Ridge Physiographic Provinces of the Eastern
and Southeastern United States, with a Description of Study-Area Geology and Hydrogeology. US. Geological
Survey Water-Resources Investigations Report 91-4066,44 p.
Walton, William C., 1962. Selected Analytical Methods for Well and Aquifer Evaluation. Illinois State Water Survey
Bulletin 49,81 pp.
73
-------
WORKSHOP CASE STUDY # 1 -- Part 1: Hydrograph Separation
Background
A number of studies have been completed by USGS regarding the water resources and
geohydrology of Eastern Chester County, Pennsylvania.1,2 Portions of these studies are
used in this case study to demonstrate the application of the base flow / hydrograph
separation method.
The Valley Creek watershed is in eastern Chester County in southeastern Pennsylvania (see
Figure 1), and is a tributary to the Schuylkill River. This area of Pennsylvania is in the
Piedmont physiographic province.
The watershed occupief 22.6 square miles,/and 68 percent of the basin is underlain by
Cambrian and Ordovician limestone and dolomite. The center of the basin is underlain
mostly by this easily eroded limestone and dolomite, which form Chester Valley. The
northern and southern parts of the basin consist of resistant quartzite and phyllite that
respectively compose the North Valley and South Valley Hills (see Figure 2). Of these
formations, the carbonate rocks form a complex, heterogeneous water table aquifer within the
valley. Although most of the ground water system is under water table conditions, ground
water can be confined locally by the relatively impermeable sides of a fracture or solution
channel.
The Chester Valley ground water system more closely resembles a fractured rock system
than that of a classic karst condition, which is generally characterized by sinkholes, dry
valleys, and cavern development. Most ground water flow in the Valley Creek area is local
with discharge to nearby streams. Regional ground water flow in eastern Chester County is
to the Schuylkill River.
Stream gaging station records for 1984 (USGS gaging station 01473169 located at the
Pennsylvania Turnpike) are available along with more detailed stream records for three storms
that occurred during 1984. The drainage area of Valley Creek above this gaging station is
20.8 square miles. See Figure 3 for the complete stream flow and base flow hydrographs for
calendar year 1984.
It is important to note that three major quarry operations extract carbonate rocks from the
deposits underlying the basin. Also, annual precipitation averages between 45 to 50 inches
in this part of Pennsylvania.
Additional information about the generalized direction of ground water flow and the
stratigraphic section for this area are provided in Figures 4 & 5, respectively, at the end of
Part 1 of the Case Study.
1 USGS, 1990. Geohydrology And Simulation Of Ground-Water Flow In The Carbonate Rocks Of The Valley
Creek Basin, Eastern Chester County, Pennsylvania. Water-Resources Investigations Report 89-4169.
2 USGS, 1987. Effect of Urbanization on the Water Resources of Eastern Chester County, Pennsylvania
Water-Resources Investigations Report 87-4098.
-------
Figure 1: Location of Valley Creek Watershed
76* 40' 20' 7S»
Source: USGS, 1990. Geohydrology and Simulation of Ground Water Flow in the
Carbonate Rocks of the Valley Creek Basin, Eastern Chester County,
Pennsylvania. Water Resources Investigation Report 89-4169.
-------
Problem Statement
Given the hydrograph information and chemical analysis of nearby wells, you are to estimate
the annual loading of VOCs and N02 + N03 from ground water discharge in the Valley Creek
watershed. In this part of the Case Study, you will estimate the ground water discharge rate
in the watershed.
In Part 2 of the Case Study, you will estimate contaminant loading to surface water through
ground water discharge.
Tasks to be Completed:
To assist in estimating the ground water discharge, three individual storm hydrographs and
data are presented on the following pages. For reference, these storm events are labeled on
the hydrograph in Figure 3. Using the checklist provided in Appendix C as a guide to the
base flow/hydrograph separation method, select the most appropriate individual storm
hydrograph and estimate the volume of ground water discharge to Valley Creek.
Discussion Questions:
How much confidence do you have in the accuracy of the ground water discharge estimate''
II you had more time and resources, what additional information would you like to have
available and how would you obtain a more accurate estimate?
What influences do seasonal variations have on this method as applied to this basin?
Several active quarries are located in the north central portion of the basin. How could these
affect the ground water discharge?
If more resources were available, how could you quantify the impact of the quarries on
ground water discharge?
-------
120
m bo
O
Hydrograph for Storm 1, Valley Creek Watershed
at Pennsylvania Turnpike Bridge, Pennsylvania
March 1984
Source: USGS, 1990. Geohydrology and Simulation of Ground-Water Flow in the Carbonate flocks of the Valley Creek Basin,
Eastern Chester County, Pennsylvania. Water Resources Investigation Report 89-4169.
-------
Figure 5: Stratigraphic Section of the Valley
Creek Basin Area
SYSTEM
MiD
¦JtA
Crdovician
Caabriin
Lata inti
Middle
Ftocaroioie
SERIES
Upper Trtissic
Lower Crdovician
Upper
Canbrj.an
Middle
Ctabciitt
Lower
Caabrlan
CEOUQCIC UNIT
Lycile ano
Epstein (196')
Stockton FonJicion
ierg and otners
C1986 J
Stockton Faraatiot*
Conticoj* Linaitont
Elbrook
Formation
Ledger
Doloaica
Kinztr*
Formation
vincaga
Dolc»ice
Mitiecaa
Quarts ice and
Harpers
Fhylllta.
undivided
ClUekiaa
Quartxlte
Laucocraclc
and
InceraadLace
Felaic Cnatil
I
o
«
w
O
Conestoja Formation
Elbrook Fonaacion
Ledger Fc mat ion
timer*
formation
V incase
Formation
Anclecaa
Formation
tad
Formation,
undivided
Chicklee
Formation
«
-C
U
V)
A
e
O
ke
n
h*
O
w
u
O
A
m
Cn«i««
Source:
USGS, 1990. Geohydrology and Simulation of Ground Water Flow in the
Carbonate RocKs of the Valley Creek Basin, Eastern Chester County,
Pennsylvania. Water Resources Investigation Report 89-4169.
-------
WORKSHOP CASE STUDY #1 -- Part 2: Estimating Contaminant Load through Ground
Water Discharge
This portion of the case study completes the contaminant load analysis begun under Part #1.
In this exercise, combine the developed ground water discharge estimate with the
contaminant information presented below to derive an estimate of contaminant load in the
Valley Creek Watershed.
Background
Much of the Valley Creek watershed has been developed with suburban/residential, industrial,
and commercial uses as Chester County is located in the greater Philadelphia metropolitan
area and numerous transportation networks {e.g., Pennsylvania Turnpike) traverse the County
and the Valley Creek watershed. Although most of Chester Valley is already developed, small
areas of contiguous open space do exist in the valley.
Problem Statement
You have been asked to obtain an estimate of the annual loading of VOCs and N02 + N03
from ground water discharge in the Valley Creek watershed to the Schuylkill River. Ground
water quality data from approximately 20 wells within the basin are available. A map showing
the sample well locations is provided in Figure 1, and a table listing the ground water quality
data is attached. The map also shows the watershed boundary of Valley Creek to its mouth
at the Schuylkill River and to the USGS gaging station at the Pennsylvania Turnpike.
Tasks to be Completed:
Using the check list in Appendix C as a guide to the contaminant load estimation technique,
identify the wells with the most representative ground water contaminant concentrations and
use those values to estimate contaminant load to the river in terms of kg/day.
Discussion Questions:
How much confidence do you have in the accuracy of the loading estimate?
If you had more time and resources, what additional information would you like to have
available, and how would you obtain a more accurate estimate?
What are the major assumptions involved in using this method?
-------
Ground Water Quality Data Valley Creek Watershed, Pennsylvania1
WELL #
VOC | no2 + no3
Sample Date
Micrograms/L
Sample Date
mg/L
207
08-84
25
08-84
2.7
251
08-81
20
08-81
1.8
1969
06-83
30
-
-
1976
06-84
32
05-83
0.87
1983
06-80
<5
06-80
4.7
2089
06-84
300
06-84
0.13
2136
08-81
223
06-82
1.8
2197
06-83
<5
06-83
0.62
2198
06-83
<5
06-83
4.4
2438
06-81
40
06-81
2.6
2444
08-81
180
08-62
7.8
2445
08-81
1,210
08-81
2.1
2466
06-82
<5
06-82
4.2
2478
08-82
313
08-82
0.94
2494
06-83
<5
-
-
2545
08-64
<5
08-84
2.7
2569
04-65
B.8
-
-
2613
06-64
<5
06-B4
5.3
2672
05-85
17,400
-
-
2676
06-B4
6B0
06-84
a.14
2677
06-84
<5
06-84
3.1
2749
06-84
24
06-84
0.54
n =
13
18
minimum =
8.8
0.13
maximum =
1,210
7.8
median =
40
24
mean =
237
2.6
standard
deviation -
350
2-1
- Indicates no data
<5 Indicate* no VOC detection (assumes a detection limit of $ »g/L)
non-detacta and concentration for Wall #2672 (17,400 jig/L) not Included In atatlsttcs for VOCa
1 Sources:
USGS, 1990. Geohydrology And Simulation Of Ground-Water Flow In The Carbonate Rocks of The Valley Creek
Basin, Eastern Chester County, Pennsylvania. Water Resources investigation Report 89-4169.
USGS, 19B7. Effects of Urbanization on the Water Resources of Eastern Chester County, Pennsylvania Water
Resources investigations Report 87-4098.
-------
ADDITIONAL WORKSHOP CASE STUDIES
In addition to the previous case study on hydrograph separation and the use of ground water
wells to estimate discharge and contaminant loading, this workshop also includes two case
studies provided by speakers invited by U.S. EPA Region 3. Abstracts and complete text of
two articles describing one of these case studies are provided on the following pages.
Biographies of the invited speakers are provided below.
Michael DiMatteo
Michael DiMatteo is employed with the Pennsylvania Department of Environmental Resources.
Dr. George M. Simmons, Jr.
George Simmons is a member of the Biology Department at Virginia Polytechnic Institute and
State University (VPI&SU) in Blacksburg, Virginia. He has been a member of that faculty
since 1971. Dr. Simmons is an aquatic ecologist and has worked in a variety of aquatic
habitats including ice-covered lakes fn Antarctica and continental shetf waters off the
Carolinas and Key Largo. Dr. Simmons' current research efforts are focused on ground water
discharge to coastal marine environments in the southern Chesapeake Bay. Most recently he
has been investigating the reasons for elevated fecal coliform densities in tidal inlets of the
Chesapeake Bay. He holds the rank of Alumni Distinguished Professor.
Dr. William Q. Reay
William Reay is a Research Associate in the Department of Civil Engineering at Virginia
Polytechnic Institute and State University in Blacksburg, Virginia. Dr. Reay received a Ph.D. in
Biology from (VPI&SU) in 1992, an M.A. in Marine Science from the College of William and
Mary, School of Marine Science, and a B.S. in Biology from George Mason University. His
research interests include upland, wetland, and offshore sediment nutrient dynamics and
hydrology. Geographic areas of research have included the Chesapeake Bay and coral reef
environments in Bermuda and Florida. More recently, research interests include subsurface
bacterial transport, watershed best management practices designed to reduce ground water
nutrient loadings to estuarine environments, and the design of remote controlled
instrumentation to aid in offshore sediment hydrodynamic studies.
-------
VOL. 28. NO 6
WATER RESOURCES BULLETIN
AMERICAN WATER RESOURCES ASSOCIATION
DECEMBER 1992
GROUNDWATER DISCHARGE AND ITS IMPACT ON SURFACE WATER
QUALITY IN A CHESAPEAKE BAY INLET1
William G. Reay, Daniel L. Gallagher, and George M. Simmons, Jr.2
ABSTRACT: Surface water, groundwater, and groundwater dis-
charge quality surveys were conducted in Cherrystone Inlet, on
Virginia's Eastern Shore. Shallow groundwater below agricultural
fields had nitrate concentrations significantly higher than inlet sur-
face waters and shallow groundwater underlying forested land.
This elevated nitrate groundwater discharged to adjacent surface
waters. Nearshore discharge rates of water across the sediment-
water interface ranged from 0.02 to 3.69 liters•m"2,hr1 during the
surveys. The discharge was greatest nearshore at low tide periods,
and decreased markedly with increasing distance offshore. Vertical
hydraulic heads, E),, and inorganic nitrogen flux in the sediments
followed similar patterns. Nitrate was the predominant nitrogen
species discharged nearshore adjacent to agricultural land use,
changing to ammonium farther offshore. Sediment nitrogen fluxes
were sufficient to cause observable impacts on surface water quali-
ty; nitrate concentrations were up to 20 times greater in areas of
groundwater discharge than in the main stem inlet water. Based on
DIN:DIP ratios, nitrogen contributions from direct groundwater
discharge and tidal creek inputs appear to be of significant ecologi-
cal importance. This groundwater discharge links land use activity
and the quality of surface water, and therefore must be considered
in selection of best management practices and water quality man-
agement strategies.
(KEY TERMS: groundwater seepage; estuaries; nonpoint source
pollution; nutrients, Chesapeake Bay; seepage meters; groundwa-
ter surface water interactions.)
INTROPUCTION
Nonpoint source nutrient loadings to the
Chesapeake Bay and its tidal tributaries are signifi-
cant and contribute to the degradation of water quali-
ty (U.S. EPA, 1988). Pollution control programs
designed to reduce such loadings have concentrated
on Best Management Practices (BMPs) which include
activities and land uses that minimize sediment loss
and surface runoff. Such BMPs, including conserva-
tion tillage, contouring and terracing, sedimentation
ponds, and porous pavement, can reduce nutrient
inputs to adjacent waters by overland flow, but can
potentially increase nutrient contamination of shal-
low aquifers (Dillaha, 1990).
The shallow unconfined Columbia aquifer in
Virginia's coastal plain displays characteristics that
are conducive to groundwater contamination and sub-
sequent transport of contaminants into aquatic envi-
ronments. These characteristics include: shallow
water table depths, low topographic relief (which
reduces surface water runoff), and highly permeable
sandy substrates. Research has shown that a strong
relation exists between land use and shallow ground-
water quality within the Chesapeake Bay Watershed
(Reay and Simmons, 1992). Combined with high-risk
aquifer characteristics, intensive agricultural prac-
tices, and urban land uses within the coastal plain
increase the potential for aquifer degradation.
In general, unconfined coastal aquifers are charac-
terized by flow in the direction of adjacent surface
waters in response to elevated hydraulic heads within
upland regions. As fresh groundwater moves into
nearshore coastal environments, flow paths are
sharply deflected upward by salinity density gradi-
ents. Consequently, most groundwater discharge from
unconfined aquifers occurs adjacent to the land-water
interface and represents a mixture of fresh groundwa-
ter and interstitial seawater (Bear, 1979).
Groundwater discharge of freshwater to estuarine
systems can be substantial and has been implicated
in nitrogen enrichment of coastal systems (Sewell,
1982; Johannes and Hearn, 1985; Lee and Olsen,
'Paper No. 92076 of the Water Rttources Bulletin. Discussions are open until August 1,1993.
Respectively, Research Associate and Assistant Professor, Department of Civil Engineering; and Professor, Department of Biology;
Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061.
WATER RESOURCES BULLETIN
1121
-------
Reay, Gallagher, and Simmons
1985; Valiela et al., 1990). Although significant effort
has been invested in studying surface and groundwa-
ter quality in the Chesapeake Bay watershed, it has
been only recently that research has focused on the
significance of groundwater discharge to this particu-
lar Bay system (Simmons, 1989; Libeloe* a/., 1991).
This study was conducted to evaluate water quality
impacts of groundwater discharge to a shallow estuar-
ine embayment. The specific research objectives were
to examine groundwater discharge patterns in a
nearshore region, to assess the degree of coupling
between shallow groundwater quality and nearshore
sediment nutrient flux, and to evaluate the potential
consequences of groundwater nutrient loadings in
terms of dissolved nitrogen and phosphorus ratios.
STUDY SITE DESCRIPTION
Research was conducted in Cherrystone Inlet
(37*19 'N, 75'59"W)( located in the southern tip of the
Delmarva Peninsula on the Eastern Shore of Virginia
(Figure 1). Land use for Virginia's Eastern Shore has
been estimated at 40 percent agricultural, 50 percent
forested, 8.5 percent pasture, and 1.5 percent urban
(U.S. EPA, 1988). Cherrystone Inlet was chosen for
this study because it possesses the high-risk upland
and aquifer characteristics discussed earlier. It is
approximately 6.4 km2 in area with maximum depths
on the order of three tp four meters. The inlet is a
polyh aline system with water column salinities on the
order of 20 ppt. The inlet receives freshwater inflow
from five principal creeks. Mean tidal range is 0.7
meters with a spring range of 0.85 meters (NOAA,
1990). Annual rainfall for 1991, monitored at Painter,
Virginia (-30 kilometers north of the inlet), was 89.5
centimeters, 83 percent of the 50-year average
(Virginia Agric. Ext. Service, personal communica-
tion, 1991).
The shallow unconfined groundwater system is the
Columbia aquifer and is underlain by the Yorktown
confined aquifer approximately 25 meters below Bea
level (Meng and Harsh, 1988). Surrounding upland
soils are predominantly well-drained Bojac fine and
Munden sandy loams which are characterized by
moderately rapid (5-15 cm^hr1) and rapid (15-120
cm*hrl) infiltration rates, respectively (Cobb and
Smith, 1989). Surface soils are underlain by pleis-
tocene-holocene sediments which also exhibit rapid
infiltration rates. Agricultural crops within the water-
shed consist of grain, vegetables, soy beans, and cot-
ton. Typical annual fertilizer applications for grain
crops are on the order of 25 kilograms per hectare for
phosphorus and 120 kilograms per hectare for
nitrogen (Eyreville Farms, Cheriton, Virginia, person-
al communication, 1992).
METHODOLOGY
Upland Shallow Groundwater Hydrology and
Quality
Monitoring wells were established at three loca-
tions adjacent to Cherrystone Inlet (Figure 1) to col-
lect water samples and to measure water table
elevations. These wells penetrated approximately one
meter into the Columbia aquifer. Shallow groundwa-
ter flow direction was determined using an equipoten-
tial contour, assuming flow to be perpendicular to the
contour in the direction of decreasing hydraulic head
(Heath, 1983). Samples for chemical analysis were
taken after three well volumes of water were removed
by a peristaltic pump.
Sediment Characteristics and Pore Water Chemistry
To assess the ability of nearshore sediments to
transmit water, vertical hydraulic conductivities (Ky)
of surficial sediments (upper 20 cm) were determined
on undisturbed replicate (N=3) cores using a falling
head permeameter (Klute, 1965). Vertical hydraulic
conductivities were measured at five sites comprising
three sandy and two silt-clay sediment type locations.
In addition, replicate (N=3) core measurements of
vertical hydraulic conductivity were taken along a
transect extending from a low tide mark to 50 meters
offshore at the Eyreville site (Figure 1).
Vertical Eh profiles were used as an indicator of
oxidation-reduction potential of nearshore sediments.
Sediment cores were collected at the low tide mark, at
five meters offshore, and ten meters offshore of the
Eyreville site. Measurements of sediment Eh were
determined using the technique of Bailey and
Beauchamp (1971) and referenced to a standard
hydrogen cell at 20*C. Diffusion controlled samplers,
designed and prepared according to Simon et al.'
(1985), were used to collect water samples at the sedi-
ment-water interface and interstitial pores to a depth
of 25 centimeters. Diffusion samplers were allowed to
equilibrate foT 14 days prior to collection. Immediate-
ly following retrieval (-1 minute), diffusion samplers
were placed and maintained in an inert nitrogen gas
environment to diminish gas exchange effects. Deeper
pore water samples (>25 cm) were collected by mini-
piezometers described by Lee and Cherry (1979).
WATER RESOURCES BULLETIN
1122
-------
Groundwater Discharge and Its Impact on Surface Water Quality in a Chesapeake Bay Inlet
Chesapeake
Bay
Legend
* Seepage Metes
A Surface Water Samples
• Upland Wells
General Groundwater
Flow Direction
Figure 1. General Schematic of Cherrystone Inlet, Surface Water and Groundwater Discharge
Sampling Station*, Upland Wells and General Groundwater Flow Directions.
Seepage Patterns and Sediment Nutrient Flux
Seepage meters were used to measure water dis-
charge and inorganic nitrogen flux across the sedi-
ment-water interface at four locations (Figure 1).
These seepage meters were constructed of nalgene
and transparent plexiglass and followed the basic
design of Lee and Cherry (1979). They enclosed 0.25
m2 of sediment with a head space of six liters.
Discharge and sediment solute fluxes were deter-
mined from replicate seepage meters (N=3-6) placed
immediately below the low tide mark. In order to
address spatial variations in discharge rates of water
across the sediment-water interface, replicate (N=6)
seepage meters were also placed in a transect extend-
ing offshore at the Eyreville site. Incubation time
periods ranged from three to six hours and were con-
ducted starting near high tide and continuing to low
tide conditions. Discharge rates (liters*m-2*hrl) were
corrected for anomalous short-term influx of water
into the collection bladders (Shaw and Prepas, 1989).
Bulk sediment solute flux measurements accounted
1123
WATER RESOURCES BULLETIN
-------
Fteay, Gallagher, and Simmons
for volume changes due to seepage, but did not correct
for biotic water column effects. To identify tidal varia-
tions in hydraulic head differences between the water
column and nearshore sediments, manometers were
used in conjunction with mini-piezometers.
Manometers followed the basic design and installa-
tion of Winter et ai. (1988).
Surface Water Quality Survey
The upper portion of Cherrystone Inlet (3.7 km2)
was subdivided into transects which incorporated 73
sampling stations (Figure 1). The experimental design
concentrated sampling effort in nearshore regions
where groundwater discharge studies were being con-
ducted. Samples were taken at a uniform depth of 10
centimeters below the water surface during slack low
water conditions. Three sampling stations within the
main stem of Cherrystone Inlet incorporated vertical
sampling stations of approximately one meter inter-
vals. Sampling stations for the watershed stream
headwaters were located near the central ridge of the
peninsula which demarcates the groundwater flow
divide. Headwaters of watershed streams were sam-
pled in concert with inlet water quality surveys.
Surveys were conducted in February, April, and July
1991. Contour plots were generated by Systat/
Sygraph version 5.0 (Systat, Inc., Evanston, Illinois).
Analytical Chemistry
Ammonium (NH4+) was determined by a modified
phenate method (Strickland and Parsons, 1972). To
eliminate hydrogen sulfide interference in pore water
samples, ammonium samples were acidified, purged
with nitrogen gas, and readjusted with respect to pH
prior to analysis. Nitrite (N02 ) was determined by
diazotizing with sulfanilamide and coupling with N-
(l-naphthyl)-ethylenediamine to form an azo dye
(U.S. EPA, 1983; Method 354.1). Nitrate (N03*) was
determined after reduction to nitrite by use of Cu-Cd
columns (U.S. EPA, 1983; Method 353.3). Dissolved
inorganic phosphorus (DIP) was determined by the
ascorbic acid method using a single combined reagent
(U.S. EPA, 1983; Method 365.2). Chloride was deter-
mined by titration with silver nitrate using a chro-
mate end point (APHA, 1989; Method 4500 B).
Salinity was determined by a Beckman Industrial
induction salinometer Model RS-10 with a precision of
0.003 ppt.
RESULTS
Upland Groundwater Hydrology and Quality
Groundwater flow patterns for the Eyreville and
Scott sites generally followed surface relief toward the
shoreline (Figure 1). Depth to the water table was
shallow, on the order of one to two meters. Mean hori-
zontal hydraulic gradients were on the order of 0.001
and 0.003 m'nr1 at the Scotts and Eyreville sites,
respectively. Horizontal hydraulic gradients increased
at the shoreline up to 0.023 m^nr1 at the Eyreville
site.
Descriptive statistics for groundwater and water-
shed stream nutrient chemistry are given in Table I.
Nitrate accounted for -99 percent of the dissolved
inorganic nitrogen (DIN = NH^ + NOj' ~ N03")
TABLE 1. Upland Groundwater and Watershed Creek Dissolved Inorganic Nitrogen and Phosphorus Concentrations
Standard error of the mean and sample size are denoted within parentheses.
Adjacent
NOs*
NOj-
nh4+
DIP
Sit*
Land Uh
M-mol« liter1
uax>l •liter1
limol* liter1
mnol*Ut*H
Eyreville
Agricultural
652.2
1.0
2.5
06
(32.2,52)
<0.4,52)
(0.2,51)
(0.1,46)
Eyre hall
Agricultural
602.7
0.3
2.8
0 4
(32.5,13)
(0.1,14)
(0.5,14)
(0.2.11)
Steel mar
Agricultural
222.0
0.4
9.0
0.3
(19.2,10)
(0.1,9)
(0.5,8)
(0.1.91
Old Castle
Forested
6.6
1.8
12.9
0.2
(2.0,16)
(1.1,17)
(3.3,18)
(0.03.6)
Scott
Forested
81.8
1.7
5.5
0.2
(28.3,7)
(1.5,7)
(1.3,5)
(0.03,6)
Watershed Stream Headwaters
Mixed
135.0
2.3
23.9
0.4
(N«5)
(31.0,15)
(0.6.15)
(6.8,15)
(0.2,10)
WATER RESOURCES BULLETIN
1124
-------
Groundwater Discharge and Its Impact on Surface Water Quality in a Chesapeake Bay Inlet
species found in groundwater underlying agricultural
fields. Nitrate was also the dominant inorganic nitro-
gen species found within the watershed's stream
headwaters. Dissolved total inorganic nitrogen levels
in shallow groundwater underlying forest sites were
significant!y lower (p = <0.001) than those beneath
agricultural lands. While mean difference was only
0.3 (imol'liter-1, groundwater dissolved inorganic
phosphorus (DIP) levels were significantly higher
underlying agricultural lands (p=0.001) as compared
to forested lands.
Sediment Characteristics and Pore Water Chemistry
Intertidal and nearshore sediments within the
main stem of Cherrystone Inlet were predominantly
sandy substrates, whereas sediments in the more
protected coves and upper creek reaches were a silt-
clay mix. Nearshore sandy sediment vertical
hydraulic conductivities averaged 9.42 cm^hr*1 (N=9),
whereas the silt-clay sediments averaged 0.09 cm*hr-
1 (N=6). As shown in Figure 2a, vertical hydraulic
Dtotano* oflMtort (m)
OMane* oflMwra On)
Figure 2. (a) Sediment Vertical Hydraulic Conductivity
-------
Reay, Gallagher, arid Simmons
conductivities decreased with distance offshore at the
Eyreville site, with mean values ranging from 13.40
cm*hr-1 at the low tide mark to 0.97 cm*hr-1 50
meters offshore.
Vertical profiles of Eh in the nearshore sediments
are shown in Figures 3a and 3b. In general, sediment
reducing conditions increased with distance offshore.
The December sampling resulted in more oxic condi-
tions overall. A sharp redox gradient occurred within
the upper few centimeters of sediment at all sites dur-
ing the July sampling, whereas only the five and ten
meter sites displayed this pattern during the
December sampling. These patterns could be the
result of decreased microbial respiration due to colder
temperatures and/or increased oxic groundwater
discharge associated with winter and early spring
seasons. Vertical profiles of pore water chloride levels
indicated a gradient of fresh groundwater input to the
nearshore region with freshwater fluxes decreasing
with increasing distance offshore (Figure 4a). Sharp
pore water chloride concentration gradients in the
upper ten centimeters of sediment indicated mixing
with the overlying water column. Elevated pore water
nitrate levels (Figure 4b) were associated with low
chloride levels and nearshore regions where oxic sedi-
ments prevailed. In comparison, the more reduced off-
shore pore waters were dominated by ammonium ions
(Figure 4c).
Seepage Patterns and Sediment Nutrient Flux
Potentiometric head differences between the sedi-
ment and water column at the Eyreville site are
shown in Figure 5. Hydraulic heads of the interstitial
waters at the low tide mark were inversely correlated
with tidal elevation. A similar, although greatly
reduced, pattern was seen 10 meters offshore.
Negative heads, indicating surface water flow into the
sediments were observed during high tide for the
nearshore site. Thus, observed seepage at the next
low tide would be a mixture of both fresh groundwa-
ter and saline water from previous inflows.
Seepage discharge measurements and sediment
nutrient flux values are given in Table 2. Measurable
nearshore discharge occurred at all sites, with higher
levels in April and July. Decreases in salinity levels
were observed in 75 percent of the discharge samples,
resulting from dilution with less saline groundwater.
The overall mean discharge rate within Cherrystone
Inlet was 0.35 liters^m-2'hr*1 (Se = 0.09, N=47) for
the three surveys. Results at Eyreville showed that
discharge of water across the sediment-water inter-
face decreases with distance offshore (Figure 2b), fol-
lowing the same general pattern as the vertical
hydraulic conductivities and potentiometric heads.
WATER RESOURCES BULLETIN
Thus, for Cherrystone Inlet, groundwater discharge to
overlying surface water appears to be a nearshore
phenomenon with elevated rates occurring during low
tide conditions.
Figure 3. (a) Sediment Profiles for July 1991 and (b) December
1990 with Distance Offshore at the Eyreville Site.
Nitrate was the predominant inorganic nitrogen
ion being discharged in the nearshore sandy environ-
ment. Nitrite and ammonium fluxes were generally
much lower. Ammonium fluxes on occasion suggested
biotic uptake. Overall mean sediment inorganic nitro-
gen fluxes immediately below the low tide mark were
7.5 pmol*nr2*hr-l (S,=2.9, N=40) for ammonium,
154.3 umol*m-2*hr1 (Se=68.7, N=36) for nitrate, 1.6
jimol*m-2#hrl for nitrite (Se=0.8, N=45), and 172.1
jimol«m-2»hri (Se=73.9, N=34) for DIN. Nitrate flux
decreased with distance offshore, while the ammoni-
um flux increased (Figure 2c). This was consistent
with the Eh profiles shown in Figure 3, which showed
a more reduced environment offshore. Considering
that the N03'-N02* coupled stabilizes Eh at about
+ 100 to +200 mVolts (Bailey and Beauchamp, 1971),
the transport of groundwater-derived nitrate across
1126
-------
E
o
CL
a
m
CO
O
c
30
O
m
w
09
c
F
3
0 100 200 300 400 500
CI (mmol/L)
120
0 200 400 600 600
N03(umol/L)
100
120
0 50 100 150 200
NH4(umol/L)
Figure 4. (a) Interstitial Chloride, (b) Nitrate, and (c) Ammonium Concentration Profiles
with Distance Offshore at the Eyreville Site. Sample Date: July 17, 1992.
O
3
c
3
O-
«
to
-t
D
8
zr
-1
HQ
3
U-
3
in
c
I
(T>
»
£
n
e
3
n
3"
m
u
o
*
X*
rp
00
»
A
-------
Reay, Gallagher, and Simmons
the sedimeiit-water interface would be relatively con-
fined to the nearshore region. More reduced forms of
nitrogen, in particular ammonium, would dominate
DIN sediment flux further offshore where Eh mea-
surements are on the order of 0 to -100 mVolts.
Surface Water Quality Survey
Surface water quality parameters for the three sur-
veys are given in Table 3. Seasonal salinity patterns
were consistent with elevated late winter-spring
freshwater inflows and decreased summer contribu-
tions. The inlet appeared well mixed vertically.
Vertical variations of surface and bottom water salini-
ties were generally less than 1 percent. Vertical varia-
tions in DIN levels of surface and bottom waters
varied between 9 and 26 percent. Temporal and spa-
tial variations were observed for dissolved inorganic
nitrogen species, both in terms of quantitative
amounts and relative percent of each species. Mean
DIN was lowest in July. As a percentage of DIN, NO3"
decreased while NH4* increased from winter to sum-
mer conditions. Summer is a period of decreased sur-
face and groundwater inflow and NO3' flux, and a
period of more reducing environment in the sedi-
ments.
Groundwater discharge caused a measurable
change in surface water quality. Nitrate contour plots
for the three sampling periods are shown in Figures
6a-6c. Elevated nitrate in most watershed creeks was
evident, as well as large gradients at Eyreville and
occasionally Old Castle, where groundwater with ele-
vated nitrate was entering the inlet. Contour plots of
DIN:DIP ratios for Cherrystone Inlet are given in
Figures 7a and 7b. Based on weighted surface areal
cover, DIN:DIP ratios above 16 accounted for 100 and
66 percent areal coverage for the upper portion of the
Inlet for the February and July sampling, respective-
ly-
Figure 5. (a) Surface Water Tidal Elevations and (b) Hydraulic
Head Differences Between Surface Waters and Interstitial
Waters (1 meter in depth) at the Eyreville Site.
Sample Date: October 10, 1990.
TABLE 2. Subsurface Seepage Rates and Sediment Inorganic Nitrogen Fluxes
(negative values denote uptake by sediment).
Sit*
Adjacent
Land Un
Data
(1M1)
Discharge
liter*
NOj' flux
NO,' flux
^unol«m"a,hr"1
nmol •
Eyreville
Agricultural
February
0.11
55.4
0.5
0.2
April
1.06
1506.3
21.2
14.8
July
1.46
60.5
0.1
7.3
Eyrehall
Agricultural
February
0.07
11.3
0.1
-0.2
April
0.34
38.1
2.7
2.1
July
0.28
50.1
0.0
3.1
Steelmans
Agricultural
February
0.02
-
-
April
0.61
43.6
1.7
10.9
July
0.1S
1.1
0.3
2.7
Old Cattle
Forested
February
0.07
10.7
0.1
-0.2
April
0.18
24.9
0.4
62.3
July
0.15
2.1
0.1
4.S
WATER RESOURCES BULLETIN
1128
-------
Groundwater Discharge an(J Its Impact on Surface Water Quality in a Chesapeake Bay Inlet
TABLE 3. Cherrystone Inlet Surface Water Quality Summary
concentrations are expressed in ^moMiter').
Month Temperature Salinity Percent of DIN
(1991) CC)
-------
Figure 6. (a) Low Tide Nitrate Contours for Cherrystone Inlet for February, (b) April, and
-------
Groundwater Discharge and Its Impact on Surface Water Quality in a Chesapeake Bay Inlet
A. February
B. July
Figura 7. (a) Low Tida DIN-.DIP Ratio Contour* for Charryttona Intet for February and (b) July 1991 Sampling Farioda.
Shaded portions rapreaant area< with OIN:OIP rmtioi of 16:1 or baJow.
1131
WATER RESOURCES BULLETIN
-------
Reay."Gallagher, and Simmons
Elevated DIN levels in surface waters were associ-
ated with tidal creek inputs or nearshore areas under
the influence of groundwater discharge. In some
cases, nitrate levels in nearshore regions influenced
bv groundwater discharge from agricultural land uses
exceeded main stem concentrations by over an order
of magnitude. Nutrient enriched, lower salinity water
could be distinguished approximately 50 meters off-
shore of regions exhibiting elevated groundwater dis-
charge.
Salinity (ppt)
Figure 8. Nitrate Concentrations Versus Salinity at the
Eyreville Site During the April Sampling Period.
Assuming the molar uptake ratios of nitrogen and
phosphorus by marine phytoplankton are relatively
constant (DIN.DIP = 16:1; Redfield et ai, 1963), water
column DIN:DIP ratios can provide insight into pri-
mary productivity limitation. On a percent areal basis
during low tide conditions, the entire upper inlet dis-
played DIN:DIP ratios greater than 16:1 for the
February sampling. The lowest concentrations of DIN
and the highest rates of pelagic primary productivity
occur during the summer in the Chesapeake Bay sys-
tem (D'Elia et ai, 1986; U.S. EPA, 1982), suggesting a
potential for nitrogen limitation of primary productiv-
ity. The main stem and nearshore regions at high tide
of Cherrystone Inlet exhibit summer DIN:DIP ratios
on the order of 10:1 and net primary productivity lev-
els on the order of 20 p.mol Og'Hter^hr1 (Reay,
1992) and thus could be nitrogen limited. However, in
contrast to main stem inlet waters, over half of the
upper inlet waters displayed DIN:DIP ratios greater
than 16:1 during July 1991. Elevated DIN:DIP ratios
were associated with upper inlet areas under the
influence of direct groundwater discharge or tidal
creek inputs. As a result of these nitrogen inputs,
surface waters may not show DIN:DIP ratios that
suggest nitrogen limitation dunng summer months.
This additional source of nitrogen during periods of
peak productivity may lead to seasonal water quality
problems associated with eutrophic waters.
Given the importance of extensive nutrient load-
ings, BMP selection should consider the direct inter-
actions between surface and groundwater. Land use
practices that reduce the availability of nitrate in
soils and its subsequent migration to the water table
are of paramount importance in limiting aquifer
degradation, By virtue of their location between
uplands and surface waters and soil conditions con-
ducive for denitrification, riparian forests and wet-
land communities may provide a viable and
ecologically sound management option for restoration
of groundwater quality prior to its discharge into sur-
face waters. Because monitoring programs are
designed to identify and incorporate nutrient sources
to surface waters, programs should include a sam-
pling protocol that addresses groundwater contribu-
tions in regions that display high risk characteristics.
Neglecting groundwater discharge as a nutrient
source may lead to misinterpretation of data and
error in water quality management strategies.
ACKNOWLEDGMENTS
The authors woyld like to thank several people for their field
assistance and technical support. Scott Smedley, Howard Nippert,
Karen Kelly Reay. Mike Robinson, Eduardo Miles, Andrea Dietneh.
and Bryan Salley provided assistance in the field and/or with water
chemical analysis. Vessel support was provided by Bob Carpenter,
Charlie Moore, and Jack Brady. The authors would also like to
thank Roger S. Buym, Carlton and Annebell Scott, and Furlong
Baldwin Tor allowing access to the study sites and providing a gen-
uine concern for water resource protection in Virginia. This
research was supported by Virginia's Water Resources Research
Center, the Virginia Division of Soil and Water Conservation, and
the National Science Foundation. This manuscript benefited by
comments from three anonymous reviewers.
LITERATURE CITED
APHA, 1989. Standard Methods for the Examination of Water and
Wastewater (17th Edition). American Public Health Association,
Washington, D.C.
Bailey, L. D. and E. G. Beauchamp, 1971. Nitrate Reduction and
Redox Potentials Measured with Permanently and Temporarily
Placed Platinum Electrodes in Saturated Soils. Canadian
Journal of Soil Science 51:51-50.
Bear, J., 1979. Hydraulics of Groundwater. McGraw-Hill Book
Company, New York, New York, 5S9 pp.
Capone, D. G. and J. M. Slater, 1990. Inter annual Patterns of
Water Table Height and Groundwater Derived Nitrate in
Nearshore Sediments. Biogeochemistry 10:277-288.
Cobb. P. R. and D. W. Smith, 1989. Soil Survey of Northampton
County, Virginia. U.S, Dept. of Agriculture, Soil Conservation
Service, and Virginia Polytechnic Institute and State University,
94 pp.
WATER RESOURCES BULLETIN
1132
-------
Groundwater Discharge apd Its Impact or Surface Water Quality in a Chesapeake Bay Inlet
2,500
cm' 2,000
£
O 1.500
E
D
1,000
X
D
LL 500
(1)
2? o
z
-500
N03 Flux = 776. B(Freshwater Discharge) - 20.2
¦ r2= 0.85, N=56
—
~
-
~
~
~ a
~
D ~
~
a ~
CD
~
l i I i I
0.5
1.5
2.5
Freshwater Discharge (L/m2-hr)
Figure 9. Sediment Nitrate Flux Venus Fresh Water Discharge at the Eyreville Site. Freshwater discharge rates
were calculated based on volume and salinity changes in seepage meters.
Sampling date interval includes November 1990 to January 1993.
Cushing, E. M„ I. H. Kantrowitz, and K. R. Taylor, 1973. Water
Resources or the Delmarva Peninsula. U.S. Geological Survey
Professional Paper 822, Washington, D.C., 56 pp.
DTlia, C. F., J. G. Sanders, and W. R. Boynton, 1986. Nutrient
Enrichment Studies in a Coastal Plain Estuary: Phytoplankton
Growth in Large-Scale, Continuous Cultures. Canadian Journal
of Fisheries and Aquatic Sciences 43:397-406.
Dillaha, T. A., 1990. Role of Best Management Practices in
Restoring the Health of the Chesapeake Bay: Assessments of
Effectiveness. In: Perspectives on the Chesapeake Bay, 1990:
Advances in Estuarine Sciences, M. Haire and E. C. Krome
(Editors). U.S. EPA Chesapeake Bay Program CBP/TRS41/90,
pp. 57-81.
Hamilton, P. A. and R. J. Shedlock, 1992. Are Fertilizers and
Pesticides in the Ground Water? A Case Study of the Delmarva
Peninsula, Delaware, Maryland, and Virginia. U.S. Geological
Survey Circular 1080,16 pp.
Heath, R. C., 1983. Buic Groundwater Hydrology. U.S. Geological
Survey Water-Supply Paper 2220,85 pp.
Johannes, R. E. and C. J. Hearn, 1986. The Effect of Submarine
Groundwater Discharge on Nutrient and Salinity Regimes in a
Coastal Lagoon Off Perth, Western Australia. Estuarine,
Coastal and Shelf Science 21:789-800.
Johnston. R. H., 1971. Base Flow as an Indicator of Aquifer
Characteristics in the Coastal Plain of Delaware. U.S.
Geological Survey Professional Paper 750-D, Washington, D.C.,
pp. D212-D215.
Klute, A., 1965. Laboratory Measurement of Hydraulic
Conductivity of Saturated Soil. In: Methods of Soil Analysis.
Part I. Physical and Mineralogical Properties. Including
Statistics of Measurement and Sampling, C. A. Black (Editor).
Amer. Soc. Agronomy, Inc., Publisher. Madison. Wisconsin, pp.
210-221.
Lee, D. R. and J. A. Cherry, 1979. A Field Exercise on Ground
Water Flow Using Seepage Meters and Mini-Piezometers.
Journal of Geologic Education 27:6-10.
Lee, V. and S. Olsen, 1985. Eutrophication and Management
Initiatives for the Control of Nutrient Inputs to Rhode Island
Coastal Lagoona. Estuaries 8:191-202.
Libelo, L., W. G. Maclntyre, and G. H. Johnson, 1991. Groundwater
Nutrient Discharge to the Chesapeake Bay: Effects of Nemrshore
Land Use Practices. In: New Perspectives in the Chesapeake
Bay System: A Research and Management Partnership, J. A.
Mihursky and A. Chanty (Editor*] (December 4-6, 1990,
Baltimore. Maryland). Chesapeake Research Consortium
Publication 137:613-622.
Meng, A. A. and J. F. Harsh, 1988. Hydnlogk Framework of the
Virginia Coastal Plain. U.S. Geological Survey Professional
Paper 1404-C, Washington, D.C., 55 pp.
NOAA, 1990. East Coast of North and South America, Including
Greenland: Tide Tables 1991 High and Low Water Predictions.
U.S. Dept. of Commerce, National Oceanic and Atmospheric
Administration.
1133
WATER RESOURCES BULLETIN
-------
Rfeay, Gallagher, and Simmons
Reay, W. G. and G. M. Simmons, Jr., 1992. Groundwater Discharge
in Coastal Systems: Implications for Chesapeake Bay. In:
Perspectives on Chesapeake Bay, 1992: Advances in Estuarine
Sciences, S. Nelson, P. Elliot, B. Farquhar, and C. McManus
(Editors). Chesapeake Research Consortium Publication 143:17-
44.
Reay, W. G., 1992. Eituarine Sediment Nutrient Exchanges: The
Importance of Physical Transport Mechanisms and Benthic
Micro-Communities. Ph.D. Dissertation, Virginia Polytechnic
Institute and State University, Blacksburg, Virginia, 263 pp.
Redfield, A. C., B. H. Ketchum, and F. A. Richards, 1963. The
Influence of Organisms on the Composition of Sea-Water. In:
The Sea, Vol. 2, M. Hill (Editor). Intersciences Publishers, New
York, New York, pp. 26-77.
Sewell, P. L., 1982. Urban Groundwater as a Possible Nutrient
Source for an Estuarine Benthic Algal Bloom. Estuarine,
Coastal and Shelf Science 15:569-576.
Shaw, R. D. and E. E. Prepas, 1989. Anomalous Short-Term Influx
of Water Into Seepage Meters. Limnology and Oceanography
34(7):1343-1351.
Simmons, G. M., Jr., 1989. The Chesapeake Bay's Hidden
Tributary: Submarine Groundwater Discharge. In: Proceedings
of Groundwater Issues and Solutions in the Potomac River
Basin/Chesapeake Bay Region. National Water Well Associa-
tion, Dublin, Ohio, pp. 9-29.
Simmons, G. M., Jr., W. G. Reay, S. Smedley, and M. T. Williams,
1991. Assessing Submarine Groundwater Discharge in Relation
to Land Use. In: New Perspectives in the Chesapeake Bay
System: A Research and Management Partnership, J. A.
Mthursky and A. Chaney (Editors) (December 4-6, 1990,
Baltimore, Maryland). Chesapeake Research Consortium
Publication 137:635-644.
Simon, N. S., M. M. Kennedy, and C. S. Massoni, 1985. Evaluation
and Use of a Diffusion-Controlled Sampler for Determining
Chemical and Dissolved Oxygen Gradients at the Sediment-
Water Interface. Hydrobiologia 126:135-141.
Strickland, J. D. H. and T. R. Parsons, 1972. A Practical Handbook
of Seawater Analysis (Second Edition). Fisheries Research
Board of Canada Bulletin, Ottawa, Canada.
U.S. EPA, 1982. Chesapeake Bay Program: A Synthesis. U.S. EPA,
Region 3, Philadelphia, Pennaylvania.
U.S. EPA, 1983. Methods for Chemical Analysis of Water and
Wastes. EPA-600/4-79-020.
U.S. EPA, 1988. Chesapeake Bay Nonpoint Source Programs. U.S.
EPA, Region 3, Chesapeake Bay Liaison Office, Annapolis,
Maryland, 122 pp.
Valiela, J. M„ J. Costa, K. Kenneth, J. M. Teal, B. Howes, and D.
Aubrey, 1990. Transport of Groundwater-Boroe Nutrients from
Watersheds and Their Effect on Coastal Waters.
Biogeochemistry 10:177-197.
Winter, T. C., J. W. Labaugh, and D. 0. Rosenberry, 1988. The
Design and Uae of a Hydraulic Potentiomanometer for Direct
Measurement of Differences in Hydraulic Head Between
Ground Water and Surface Water. Limnology and Oceanography
33(5):1209-1214.
Zimmermann. C. R., 1991. Submarine Groundwater Discharge to
the Patuxent River and Chesapeake Bay. In: New Perspectives
in the Chesapeake Bay System: A Research and Management
Partnership, J. A. Mihursky and A. Chaney (Editors) (December
4-6, 1990, Baltimore, Maryland). Chesapeake Research Consor-
tium Publication 137:663-668.
WATER RESOURCES BULLETIN
1134
-------
PROCEEDINGS OF THE
FIRST INTERNATIONAL CONFERENCE ON
GROUND WATER ECOLOGY
Edited by
Jack A. Stanford
Flathead Lake Biological Station
The University of Montana
311 Bio Station Lane
Poison, Montana 59860
and
John J. Simons
Ground Water Protection Division
U.S. Environmental Protection Agency
401M Street S.W.
Washington, D.C 20460
Sponsored by the
United States Environmental Protection Agency
American Water Resources Association
5410 Grosvenor Lane, Suite 220
Bethesda, Maryland 20814-2192
and the
Ecological Society of America
2010 Massachusetts Avenue N.W., Suite 420
Washington, D.C 20036
Tampa, Florida
April 26-29,1992
Although the information in this document has been funded wholly or in part by the United
States Environmental Protection Agency under assistance agreement X9951S0-01 to the
American Water Resources Association, It may not necessarily reflect the views of the
Agency and no aifidal endonement should be inferred.
-------
Literature citation for this volume:
Stanford, Jack A. and John J. Simons (Editors), 1992. Proceedings of the
First International Conference on Ground Water Ecology. American
Water Resources Association, Bethesda, Maryland, 420 pp.
AMERICAN WATER RESOURCES ASSOCIATION TECHNICAL PUBLICATION SERIES
TPS-92-2
LIBRARY OF CONGRESS CATALOG CARD NUMBER: 92-7014*
1992 COPYRIGHT BY THE AMERICAN WATER RESOURCES ASSOCIATION
All rights reserved. No put of this book may bo reproduced in any form or by any mechanical mwu without
permission from the publish*. Theee proceedings were published by the Amanean Witer Resources Association,
5410 Grosvenor Lane, Suite 220, Bethesda. Maryland 20814-2192. The views and statements advanced in this
publication are solely those of the authors and do not uyi—ill offldal views or policies of the editors or of the
American Water Resources Association, the United States Environmental Protection Agency, and the Ecological
Society of America. Commuracatiorts in reprd to this publication should be sent to the Circulation Department of
the American Water Raeouices Association. 5410 Gwwut Lane, Suite 220, Bet hearts. Maryland 20814-2192. (301}
49W600.
-------
FIRST INTERNATIONAL CONFERENCE ON GROUND WATER ECOLOGY
U.S. Environmental Protection Agency
April American Water Resources Association 1992
SUBMARINE GROUNDWATER DISCHARGE QUALITY IN RELATION TO
LAND USE PATTERNS IN THE SOUTHERN CHESAPEAKE BAY
G.M. Simmons. Jr.,1 E. Mites/ W. Reay,1 and D. Gallagher
ABSTRACT: Coastal environments are an place to observe tbe linkage between upland
groundwater quality and its discharge into adjacent surface waters. This discharge is known as
Submarine Groundwater Discharge (SGWD) and is a mixture of recycled coastal seawater and freshwater
from upland areas due to the advecove flow of groundwater.- The efflux of SGWD is primarily at the
land/water interface and represents a type of nonpoint discharge. Three major land use types were
investigated with respect to the quality of SGWD. These types were natural areas off extensive wetlands
(Reference), agricultural areas, and urban areas. Ten sites were selected for study in.the southern
Chesapeake Bay. Tbe average SGWD rate was ~0.8 L m4 hr\ but ranged between 0.0 to
7.9 L m1 hr1. The quality of water being discharged is a minor image of groundwater quality in
adjacent upland areas. The importance of biological buffers, or filtration such as wetlands,
cannot be minimiMtl and desewre entMemaiinw tn pnwrt m—> «iw«i» «w«w«m>nw
KEY TERMS: Chesapeake Bay, groundwater, iwpM* sources,
INTRODUCTION
The movement of water across sedimem/waier faces due to advectivc flow is termed submarine
groundwater discharge (SGWD) (sp, specific discharge (Freeze and Cherry, 1979)). In coastal marine
environments it is usually a mixture of freshwater and teawater, and is driven by an hydraulic bead from
adjacent upland regions. Tbe resulting SGWD has variable salinity and tends to ester tbe surface water
environment at the land/water una face and region (Fig. 1).
The purpose of this paper is 1) to that groundwater movement can link land and
surface water eavyaems, 2) to dnmamiate that groundwater discharge can reflect land use activities
adjacent to surftce water todies; and 3) to suggest that groundwater and its quality can be
hy nttng ifpfWUhCt.
Historically, groundwater discharge into surface waw was determined by calculation (Really and
Goodman, 1985). »» Mljn| miu|miihiii ami mini piiwofirn
as designed by Lee (1977) and Lee and Cherry (1978), has been made in a variety of habitats. Among
some of the studies are those from lakes and ftreams by Lee and Hynes (1978), Lock and John (1978),
Brock a at. (1982), Befamger and Mikuset (1985), Cherkauer and McBride (1988), and Shaw, a ll>
1 Biology Department, Virginia Polyicdaiic Institute and State Univenity, Blaukibuig, Virginia 24061
s Civil CngrnrrruigDepaiuueut, V'uguha Polytechnic Institute and Stam University, Blncksburg, Virginia
24061
341
-------
Zan* el
\9»o»n i
Figure 1. SctaMric di^ini of ttefaah ml nmwtfwncr interface im email envm*—.
(from Joannes 1980).
a990V ia esniiries by Bokaaiewicz (1980), ZiniiBeniiai ttll. (1985), Midmyre a it
MQOTVZimiMniMfl(1991), md Sim^g *<)*»;,»
Bty, has Nrr*1 oampkttd by Rety and Simmnw (1992).
TK<.neenngMeofSGWD is iniponiniecoiog»Mllybec«o>etitliproc«ireprMenn» vefaicte by
a'trtite^Sana'af"5GWDla»yrvtmUrtolto
nutnem Dndces. o4 potadm (Maan* 19W)
» yr^TaJ> «""?.*?»
^S^<»i«^d««toSOWDpn^W(hx«a»-aIo<^Cotantai«M«ora.
Mid-Atlmoc Cot»> PW"-
Figm2 stowfthi taabooof am dot km baanMl it ite pnM SBMty. SitedMGripanm,
indudiaf iliw uwifw of the tool nib, tare tea ilmiiiwi by Shrawm a it (1991). It be
STUDY STIES
342
-------
v •
_ Chesapeake - Bay—
Pigurs 2. Study situ for SGWD in thm southern cftssspasks Bay.
343
-------
Urban sites selected uuiizsd sepoc systems or recently developed central sewage systems. The Lr S
Environmental Protechon Agency (1988) esmnued the overall land use pattern within the Chesapeake
Bay watershed to be 54% forests. 36% agricultural lands, 7% urban areas, and 3% wetlands.
METHODS
Eflafrlithing ant) Sampling the Field Site
Field sites were selected on the basis of similar srrtimmt types, groundwater flow direction,
adiacent land use activity, and accessibility after aerial surveys and groundtnitfa inspections. Each site
has several land-tased wells logged sad placed immediately in front of the potential seepage study site.
Well screens were set into the top one nwff of the unconfined atjuifer.
At the time of sampling, wells were pumped 6* salinity and nutrient chemistry, seepage meters
were arrayed at the low tide marie, piezometers were placed far other water sampling and/or hydraulic
head mn-nr-~™ ambient water samples were aken, and sediment samples were taken. Samples of
SGWD were tt**" « high tide and again at low tide. Bach lueammnrnB were used to calculate
discharge, but only those samples taken at low tide were used for nutrient analyses. Seepage meters
were deoloved for only one tidal cycle because experience bas shown this tune period to be sufficient
to obtain evidence of SGWD coupling without an appreciable change in oxygen concentrations in the
seepage metes. GenmHy, SGWD measurements were replicated with four seepage «««*» «y one
area Hydraulic heads were measured from |jit iiimftrn and manometer* according to die technique of
Winter CI li- (198*)-
i ahn—»wy Methods:
All water samples were collected in acid-washed, 3X rinsed in diflilledydiomMd water. Water
samples for nutneut analyses are filtered through pte-rimed 0.45pm Nodeopore filters and 2-3 replicates
are analyzed for each nutrient for each water sample. Nutriea analyses, except fcr ammoomm,
fniiiwai the methodology of the U.S. Environmental Protection Agency (U-S.E.P.A., 1983).
Ammonium wai analyzed according to Stridden* and Parsons (1972). Salinity was measure in the
laboratory with a p-*™— Industrial Induction Salinoninrr (RoaemoA Analytical, Model RS-IO).
XomeiWTO was maaaurrt with a ealibiweA loot smm thermometer. Oxygen was measmed by using
a macro- or miciivWtflJder technique and titration with the appropriate thwsulfate normality (American
Public Health Assoc. tt ti, 1976).
Surficial samples (upper 20 cm> were analyzed accordiag to gram size, vertical
. , , ha fiYfT ml ueroent organic nmner from undisturbed cans. Gram size analysis and per
Qciioito atoibMWKaecniiMKl by
w« «l I** »U™
-------
C x V,) - (Co X Vo)
Flux =
T x A
where.
C,C, =
Solute concentration within seepage meter system at Qme i and 0, respectively
(fimoie L )
v„v. =
Volume of water within seepage meter system at time t and 0, respectively in
Liters (L)
T »
Time interval from 0 to i (tar)
A ¦
Bag effects were also accounted for in caimlirion of discharge volumes (Shaw and Prepnx, 1989).
A Kroskall-Wallis amlyns of variance by tank test was used to if SGWD and anient flux
vaned between sues. Data were pooled according to i—d use activities, and a nonpar*"1**"'* multiple
comparison test for unequal sample sizes was used to. detenmne which land use sites differed
significantly at a level of a « 0.03 (Zar, 1984).
RESULTS AND DISCUSSION
Descriptive mtimcs for study site properties are given in Table 1. In study
were dominated by sandy srriiiiMi (•! to 44) with relatively low organic t (<2.515).
Table 1. Descriptive siaiwirs of ^""tn propenies for ench study site.
345
-------
Exceptions to this occured at Reference site 1 and Agricultural site 1, where sediments exhibited elevated
silt-clay and gravel size fractions, respectively. Mean vertical Hydraulic conductivity ranged between
IC,J8 and 1(T 71 cm sec1, and was strongly influenced by the stfc-day sediment size fractions (^-0.70).
With the exceptions of Reference site 1 and Urban site 2, were conducive to wtier
transmissions displaying vertical hydraulic conductivities similar to clem sands and very fine suds.
Decreased salinities were observed in seepage mens at all sites, indicating the discharge of
fresher water. Based on chloride mass balances, submarine groundwater discharge of fresh water
accounted for 4% (STD=6, N-97) of the total volume collected in seepage meters. Submarine
groundwater discharge rates varied from 0.00 to 7.9 L m4 hr't with a grand men of 0.8 L m* hr1
(STD" 1.0,H-27l). Although mean (pooied iccordmg to apitad bad-use) SOWD rater were relatively
similar (Table 3), discharge from the urban sites were sigmfirantty 19.85, p« <0.001) higher
than forested or agricultural sites. Utilizing Darcy's Law and permeabilities, hydmlic head
differences between pore waters (one meter is depth) and the overlying water column need only be on
the order of milBmrtrrs to ceatimeten to explain observed discharge rates. Exceptions to this occarad
at sites characterized by elevated silt-clay wriiment fractions in which sediment compaction during core
collection may have resulted in decreased vertical hydraulic conductivities (i.e., reference site 1, urbnn
sites 2 and 3). Measured hydraulic head differences between pore waters ('1 m depth) and overlying
surface waters ranged between 0.2 and 8.0 centimeters.
Descriptive statistics for upland shallow groundwater nutrient chemistry are given in Table 2.
Mean mmte uonucuutkw varied from W.6 to 602.8 umof f, and represented n|imuia—m|j 95%
of the inorganic aiuogca faond in the groundwater unda lying igncultarsl fields. Aamomum
contributions to the inorganic nitrogen pool were negligible despite the fact that ammrwinni nitsafce is
a commonly applied fertilizer. The dissolved oxygen concentration in upland agricultural wells were
always near iauimiun values and k is jammed that all aiimaaaai was convened to nitrate by the
nrierobiMi process of akrificitioa. "it tfir irrhw vtn aaimrniiiiinrss rhr ifrnrinmr jnrrpmr nifriifi u
species and mean concentrations varied from 65.1 to 1749.0 umoi L'1. Mean nitrate «nd nitrite
coocentmioos were below 3.0 umol L"' for all urban sites. The icfaencc stations were rlmi m u i iTr»j
by low ornate and ammonium levels as competed id the agiienhmal and urban ttttintri, respectively.
Mean dissolved iiHwganir uiuugen levels in shallow gmuudwaier underlying foiested wedaad did not
exceed 25 umol L '. Overall mean inorganic phosphate levels were 0.4 and 0.6 umol L"1 at the
agricultural and reference sites, respectively. At the nrbm ntec. the nhrrnhati* ion wm
one order of magnimrin greater than that observed at the agricultural and reference sites. Thtaegeaesal
relationships, bciwtun land-use activities and shallow uminrtwus quality, concur with a recent US
Geological Survey conrhrrrri in Virgmia's coastal plain (Hanriltnn and Denver 1990).
values indicate uuuient uptake by the sediments. In general, the uuiriem composition of SGWD
reflected the quality of shalkm groundwater underlying atgacent uplands. Ammonium flnxee varied
from -87.8 to 197.9 umol nr1 hr' at the reference sites, -11.0 to 220.1 mnol m4 hr1 at the agricultural
sites, and 0.0 to 8186.7 mnol m4 hr1 at the urban sites. Phnsphatr fha varied from -10.610 12.8
umol m* hr1 at the infeinun atax, <0.71019.2 moot nr1 hr1 at the ayricnhural sites, and -1.2 to
132.5 umol or4 hr' at the urban sites. Nearshoce ammnniiim (H4Mi«i»35 J3, p«< 0.001) and
phosphate (H^.4^-27.83, p-<0.001) fluxes were significantly higher at stteaatfracetn to urban land
uses u compared to agricultural and reference land uses, ft should be noted the ammaamm aid
phosphate fluna at tnfaan site 3 snongty skewed the data. Name flax varied from -66.3 to 35 J
umol m4 hr' at the reference sites, -9.3 to 4346.0 umol nr* hr4 at the agricuknral sites, and-9.4 to
346
-------
Table 2. Summnr of aument conceuunious (junoi L') in upland wells.
ww j nob
« 1 ~
rnmmm
sm
u
1AM
u
(US
u
IIU
u
(1JL4)
U>l
1U»
iisjiis
u
IT1.1S
U
(CMS
¦4
(•AS
u
WAS
ti
(US
u
1US
U
(US
4J§
OJM
u
(IAS
u
~us
i*a«n
SHI
A*
fUl*
rtauwi
1U
KM. 1. Ml
u
«MI
mi
M
IUI
1«U
(I1US
M
(UJI
u
(US
an
t.«
*T.T)
nu
(MTJLS
U
(US
u
NbMt
M 4
U
HXJS
(i*u5e~
u
(uan
M
MMI
*¦
U
(Mb*
3TM
nu4
1U
naui
M
ftUI
tttai
Ski
a*
(US
1j»
MM
(u*
MS
tu
-------
5.4 umol m; hr ' at the urban sites. Nitrate fluxes 59.59, p= <0.001) from sites adjacent
to agricultural land-uses were significantly higher than those observed at the urban and reference sites.
SUMMARY
1. Sediments in coastal plain areas are conducive to water transport because they are highly permeable.
2. SGWD is a common hydrological property of all sites studied in the southern Chesapeake Bay and
is probably an important property in similar coastal marine environments.
3. Low salinity water could be detected at all SGWD study sites with a mean freshwater component
of "4%.
4. The quality of SGWD reflects adjacent land use activity. In general, elevated nitrate levels could be
detected off unbuffered agricultural areas, and and dissolved phosphate could be detected
off unbuffered urban areas.
5. Buffer zones, such as wetlands, are important biological zones for reducing nutrient input to coastal
environments and deserve maihimifn protection.
6. Seepage metm, or modifications thereof, in conjunction with upland wells and other hydrological
mouuieinenta can provide guide and retiaMecstiinaBac of nonpoint discharge sources and their respective
water quality properties.
ACKNOWLEDGEMENTS
It is important to recognize the agencies that have supported this research. These are the Virginia
Division of Soil and Water Conservation; the Environmental Protection Agency; the Eastern Share of
Virginia National Wildlife Refttge; the Nature Conservancy at Brownsville, Virgin*; and, the
Department of Biology, Virginia Polytechnic limimiw and Stale University, Blacksburg, Virginia.
We also are grateful to all the indivktaais who allowed us access to their property for rexarchpurpoaes.
REFERENCES CITED
American FuUk Health Association, American Water Worts Asndation, Water Pollutaoa Coam
Federation, 1976. Smdawl methods for the f lainaiiiMi of water and wastewater. Founnaah
Edition. American Public Health Aiaociitinn, Wadmgwn, D.C. xaax + 1193 pp.
Belanger, T.V. and D.F. Mikntei, 1985. On tbe use of seepage meters to estimate groundwater nmneat
totting to lake*. Wtr. Res. Bull. 21:265-272.
Bokunkwicz, H., 1980. Groundwater seepage into Great South Bay, New York. Estnar. Cans. Mar
Sci. 10:437-444.
Brock, T.D., D.R. Lee, E. James, and D. Wiaek, D., 19S2. Groundwater seepage ss a nutrient source
to a drainage lake; Lake Mendoca, Wisconsin. Water Ret. 16:1255-1263.
Qiertsner, D.A. and J.M. McBride, 19W. A remotely operated seepage meter for use in large tafcu
and rivers. Ground Water 26:165-171.
348
-------
D'Elia. C.F., K.L. Webb, and J.W. Porter, 1981. Nitrate • neb groundwater inputs to Discovery Bay.
Jamaica: a figPlfir»nT source of N to Local reefs? Bull. Mar. Sci. 31:903-910.
Folk, R.L.. 1980. Petrology of Sedimentary Rocks. Hemphill Publishing Company. Austin. Texas.
182 pp.
Freeze. R.A. and J.A. Cherry, 1979. Groundwater. Prentice-Hall, Inc., Englewood Cliffs, N.J.
07682. 604. pp.
Hamilton, P. A. and J.M. Denver, 1990. Effects of land use on ground-water flow and shallow ground-
water quality, Delmarva Peasinsula, Maryland, and Virginia. Groundwater, Sept/Oct. Vol. 38.
Johannes, R.E., 1980. The ecological significance of the submarine discharge of ground water. Mar.
Ecol. Prog. Ser. 3:365 - 373.
Klute, A., 1965. Laboratory measurement of hydraulic conductivity of saturated soil, pp. 210-221.
In: C.A. Black (ed.), nf<^l i-hmi Part 1. Phvwfl Pmnatiw.
InrhwKny nf mtt wirf Sitnntino AffleT. SOC. Of AgTOn., IllC., MldlSOO, WlSC.
Lee, D. R.t 1977. A device for measuring seepage flux in lakes and estuaries. Limnol. Oceanogr.
22:140-147.
Lee, D.R. and J.A. Cherry, 1978. A field exercise on groundwater flow using seepage meters and
mini-piezometers. J. Geoi. Ed. 27:6-10.
Lee, D.R. and H.B.N. Hynes, 1978. Identification of groundwater discharge zones in a reach of
Hillman Creek in Southern Ontario. Poll. Res. Can. 13:121-133.
Lewis, J.B., 1987. Measurements of groundwater seepage flux onto a coral reef: spatial and temporal
variations. Limnol. Oceanogr. 32:1165-1169.
Lock, M.A. and P.J. John, 1978. The tifufffrffi—» of groundwater mm a lake by a direct
method. Int. Revue ges. Hydrobiot. 63:271-275.
Maclntyre, W.G., G.H. Johnson, W.G. Reay, and O.M. Simmons, Jr., 1989. Groundwater non-point
sources of untriens to the southern Chrsapfalm Bay. In: Proceed. Groundwater Issues and
Solutions in the Potomac River Basin/Chesapeake Bay Region. Co-sponsored by the Assoc. of
Ground Water Srirnrists, et tl., George Washington University, Washington, D.C., p. 83-104.
Reay, W.G. and G.M. Simmons, Jr., 1992. Groundwater discharge in nramtl systems: implications
for Chenpcnke Bay, pp. 17-44. la PermeLiiw^ m rh~~~.iT* gg, ^222: AdUDGSS in
PqiMrwi* CRC Publication No. 143. p—Research Consortium, inc., P.O.
Box 1280, Solomons, Maryland 20688.
Reilly, T. E. and A.S. Goodman, 1985. Quantitative analysis of saltwater - freshwater relationships in
groundwwr • % htfawyfi penpecove. J. Hydrot. 80:123 -160.
Shaw, R.D. and E.E. Prepas, 1989. Anomalous, short-term influx of water into seepage meters.
Limnol. Oceanogr. 34:1343-1351.
Shaw, R.D., J.F.H. Shaw, H. Fricker, and E.E. Prepas, 1990. An « approach to quantify
groundwaw mniport of phu^hmns to Narrow Lake, Alberta. Limnol. Oceanogr. 35:870-886.
Simmons, Jr., G.M., sad J. Netberton, 1987. Groundwater discharge in a deep coral reef habitat -
evidence for a new biogeochenbcal cycle? In: C.T. Mitchell (ed.)Proc. Amer. Assoc.
Underwater Sci., 947 Newhall St, Costa Mesa, CA 92627, p. 1-12.
Simmons, G.M., Jr., 1992. The imporance of submarine groundwater diachacfe and seawater cycling
to material flux across sediment/water interlaces in mat inecnvuuumeats. Mar. Ecol. Prog. Ser.
84:173-184.
Simmons, G.M., Jr., W. Reay, S. Smedley, and M. Williams, 1991. A wearing submarine groundwater
discharge in idabon to land use, pp. 633-644. In: J.A. Mihunky and A. Chancy (eds.), lies
PHUMuimi in ibg SxSEB* Proceedings of the Chesapeake Bay Research
Conference, Dee. 4-9, 1990, Baltimore, MD. CRC Pobl. 137.
349
-------
Strickland. J.D.H. and T.R. Parsons, 1972. A Practical Handbook of Seawmter Analysis. Bull. Fish.
Res. Bd. Canada. 310 pp. j
U.S. Environmental Protection Agency, 1983. Methods for chemical analysis of water and wastes.
EPA-600M-79-020.
U.S. Environmental Protection Agency, 1988. Chesapeake Bay nonpotmsource programs. Annapolis.
Md..U.S;E.P.A. Chesapeake Bay Program.
Winter, T.C., J.W. LaBaugh, and D.O. Rosenberry, 1988. The design and use of a hydraulic
pocestioaianometer for direct measurement of differences m hydraulic bead between groundwater
and mi»face water. Limnoi. Oceanogr. 1209-1214.
Zar, J.H., 1984. Bfasawcal Analysis. Second Edition. Ptwtice- Hall, Inc., Englewood Cliffs, NJ.
Zimmerman, C.F., J.R. Montgomery, and P.R. Carlson, 1985. Variability of dissolved reactive
phrniphaar fim rates is nearsbore estnarme sediments: effects of groundwater flow. Estuaries
8:228-236.
Zimmerman, C.F, 1991. Ground water discharge® the Patnxeat River and Chesapeake Bay, pp. 663-
668. In: I.A. Mifaorsky and A. Chaney (eds.), NCff FrnnrrfiYTS in Ac nminrifcr Sygffll
Proceedings of the Chesapeake Bay Research Conference, Dec. 4-9, 1990, Baltimore, MD.
CRC Pnbl. 137.
350
------- |