EPA 600/R-07/058 I August 2008 I www.epa.gov/ada
United States
Environmental Protection
Agency
Assessment of Near-Stream
Ground Water-Surface Water
Interaction (GSI) of a Degraded
Stream before Restoration
1200800 1200900 1201000 1201100 1201200 1201300
Easting (UTM ft)
National Risk
atory, Ada, Oklahoma 74820
-------
-------
Assessment of Near-Stream
Ground Water-Surface Water
Interaction (GSI) of a Degraded
Stream before Restoration
Elise A. Striz
Robert S. Kerr Environmental Research Center
Paul M. Mayer
Project Officer,
Robert S. Kerr Environmental Research Center
Office of Research and Development
National Risk Management Research Laboratory, Ada, Oklahoma 74820
-------
Notice
The U.S. Environmental Protection agency through its Office of Research and
Development funded the research described here. This research report has been
subjected to the Agency's peer and administrative review and approved for publica-
tion as an EPA document. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
Striz, Elise A. and Paul M. Mayer. Assessment of Near-Stream Ground
Water-Surface Water Interaction (GSI) of a Degraded Stream before Restoration"
EPA/600/R-07/058. U.S. Environmental Protection Agency, 2008.
Front Cover photos:
A view of Minebank Run looking upstream at Transect 3 in July 2002.
Aerial view of Minebank Run taken from low-level blimp platform
(photo courtesy of Ken Jewell).
Regional ground water equipotentials on March 22, 2002.
Back Cover photos:
Composite topographic and aerial photograph of a portion of the heavily urban-
ized Minebank Run wtershed showing locations of piezometers (red dots) and
the 1-695 Beltway in the lower left (courtesy of Robert Shedlock).
Incision and infrastructure exposure at Minebank Run before restoration.
Location of monitoring well transects in relation to restoration design
(Figure courtesy of Maryland Department of Environmental Protection and
Resource Management).
-------
Foreword
The U.S. Environmental Protection Agency is charged by Congress with protecting the Nation's land, air, and water
resources. Under a mandate of national environmental laws, the Agency strives to formulate and implement actions
leading to a compatible balance between human activities and the ability of natural systems to support and nurture
life. To meet this mandate, EPA's research program is providing data and technical support for solving environmental
problems today and building a science knowledge base necessary to manage our ecological resources wisely,
understand how pollutants affect our health, and prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory is the Agency's center for investigation of technological
and management approaches for preventing and reducing risks from pollution that threatens human health and the
environment. The focus of the Laboratory's research program is on methods and their cost-effectiveness for prevention
and control of pollution to air, land, water, and subsurface resources; protection of water quality in public water
systems; remediation of contaminated sites, sediments and ground water; prevention and control of indoor air pollution;
and restoration of ecosystems. NRMRL collaborates with both public and private sector partners to foster technologies
that reduce the cost of compliance and to anticipate emerging problems. NRMRL's research provides solutions
to environmental problems by: developing and promoting technologies that protect and improve the environment;
advancing scientific and engineering information to support regulatory and policy decisions; and providing the technical
support and information transfer to ensure implementation of environmental regulations and strategies at the national,
state, and community levels.
This publication has been produced as part of the Laboratory's strategic long-term research plan. It is published and
made available by EPA's Office of Research and Development to assist the user community and to link researchers with
their clients. This report describes a case study of the influence of stream channel geomorphology on surface water
and ground water hydrology. The objective of the study was to characterize ground water-surface water interaction
(GSI) in Minebank Run (MBR), a degraded urban stream in Baltimore County, Maryland that was slated for restoration.
This study represents the first phase in quantifying the effects of stream restoration on GSI at MBR and is intended to
provide the basis for comparison of post-restoration GSI at MBR.
Stream restoration at MBR will utilize stream channel reconstruction methods designed to stabilize banks and reduce
erosion, an approach intended to protect property and sewer and drinking water infrastructure. A clear understanding
of GSI behavior under degraded conditions is necessary to quantify the effects of geomorphic methods of restoration
on GSI. In turn, GSI is known to influence stream function such as nutrient uptake, stream metabolism, and primary
production. Thus, the value of these baseline data extends to predicting the influence of geomorphic structure on GSI
behavior and on subsequent biogeochemical and biotic response.
Robert W. Puls, Acting Director
Ground Water and Ecosystems Restoration Division
National Risk Management Research Laboratory
-------
-------
Contents
Notice ii
Foreword iii
Figures vi
Tables vii
Acknowledgments viii
Abstract ix
1.0 Introduction 1
2.0 Background 3
3.0 Site Description 5
4.0 Methods 7
Study design and site characterization 7
Ground water-surface water interaction analysis 9
5.0 Results 13
Stream geomorphology 13
Stream geological setting 14
Regional surface water and ground water hydrology 18
Characterization of near-stream ground water-surface water interaction 21
Temperature verification of ground water-surface water interaction 31
Stream flow verification of ground water-surface water interaction 32
Storm surge ground water-surface water interaction 32
6.0 Discussion 37
7.0 Conclusions 39
8.0 References 41
-------
Figures
1. Scales of ground water surface water interaction (image adapted from
Stream Corridor Restoration Handbook; FISWRG, 1998) 4
2. Minebank Run watershed geographic location (adapted from "Watershed Characteristics
and Pre-Restoration Surface-Water Hydrology of Minebank Run, Baltimore County, Maryland,
Water Years 2002-04," Doheny et al., 2006) 5
3. Minebank Run Watershed (adapted from "Watershed Characteristics and Pre-Restoration
Surface-Water Hydrology of Minebank Run, Baltimore County, Maryland, Water Years
2002-04," Doheny et al., 2006) 6
4. Incision and infrastructure exposure at Minebank Run before restoration 6
5. Location of the Minebank Run Study Reach in Cromwell Valley Park (adapted from"Watershed
Characteristics and Pre-Restoration Surface-Water Hydrology of Minebank Run, Baltimore
County, Maryland, Water Years 2002-04," Doheny et al., 2006) 7
6. Location of monitoring well transects in relation to restoration design (Figure courtesy of
Maryland Department of Environmental Protection and Resource Management) 8
7. Transect monitoring well network design. Distances not to scale (image adapted from Stream
Corridor Restoration Handbook FISWRG, 1998) 9
8. Stainless steel piezometer used in piezometer nests 9
9. Location of monitoring wells and piezometer nests Minebank Run Study reach 9
10. Example vertical equipotential stream cross section with left bank and right bank
compartments on either side of the stream thalweg divide 10
11. Transect 1 at beginning of study in Fall 2001 13
12. Transect 2 at beginning of study in Fall 2001 13
13. Transect 3 at beginning of study in Fall 2001 13
14. Ground surface elevation cross sections at Transects 1, 2, and 3 at beginning of study 14
15. Geological cross section lines for Minebank Run study site 14
16. Minebank Run Thalweg Cross Section A-A' 15
17. Transect 3 Cross Section B-B' 15
18. Transect 2 Cross Section C-C' 16
19. Transect 1 Cross Section D-D' 16
20. Regional ground water equipotentials on March 22, 2002 19
21. Regional ground water equipotentials on August 22, 2002 19
22. Regional ground water equipotentials on September 23, 2003 20
23. Transect 1 vertical cross section equipotentials on March 22, 2002 22
24. Transect 1 vertical cross section equipotentials on June 24, 2002 22
25. Transect 1 vertical cross section equipotentials on August 22, 2002 (study low) 23
26. Transect 1 vertical cross section equipotentials on September 30, 2002 23
-------
27. Transect 1 vertical cross section equipotentials on December 17, 2002 23
28. Transect 1 vertical cross section equipotentials on September 23, 2003 (study high) 24
29. Transect 2 vertical cross section of equipotentials on March 22, 2002 24
30. Transect 2 vertical cross section of equipotentials on June 24, 2002 25
31. Transect 2 vertical cross section of equipotentials on August 22, 2002 (study low) 25
32. Transect 2 vertical cross section of equipotentials on September 30, 2002 25
33. Transect 2 vertical cross section of equipotentials on December 17, 2002 26
34. Transect 2 vertical cross section of equipotentials on September 23, 2003 (study high) 26
35. Transect 3 vertical cross section of equipotentials on March 22, 2002 27
36. Transect 3 vertical cross section of equipotentials on June 24, 2002 27
37. Transect 3 vertical cross section of equipotentials on August 22, 2002 (study low) 27
38. Transect 3 vertical cross section of equipotentials on September 30, 2002 28
39. Transect 3 vertical cross section of equipotentials on December 17, 2002 28
40. Transect 3 vertical cross section of equipotentials on September 23, 2003 (study high) 28
41. Transect 1 continuous stream bed piezometer temperatures 31
42. Transect 2 continuous stream bed piezometer temperatures 31
43. Transect 3 continuous stream bed piezometer temperatures 31
44. Steam discharge hydrograph in response to storm event on August 3, 2002 33
45. Transect 1 piezometer hydrograph response to storm event on August 3, 2002 33
46. Transect 2 piezometer hydrograph response to storm event on August 3, 2002 33
47. Transect 3 piezometer hydrograph response to storm event on August 3, 2002 34
48. GSI flow rates at Transect 1 during storm surge on August 3, 2002 34
49. Transect 1 piezometer temperature response to storm event on August 3, 2002 34
50. Transect 2 piezometer temperature response to storm event on August 3, 2002 35
51. Transect 3 piezometer temperature response to storm event on August 3, 2002 35
Tables
Table 1. Transect 1 Piezometer Conductivity 17
Table 2. Transect 2 Piezometer Conductivity 17
Table 3. Transect 3 Piezometer Conductivity 17
Table 4. Comparison of ground water levels on dates of March 22, 2002, August 22, 2002 and
September 23, 2003 21
Table 5. Ground water (GW) and surface water (SW) flow calculations for Transect 1 29
Table 6. Ground water (GW) and surface water (SW) flow calculations for Transect 2 30
Table 7. Ground water (GW) and surface water (SW) flow calculations for Transect 3 30
Table 8. Stream flow at each of the transects on specific dates 32
-------
Acknowledgments
We offer our sincere gratitude to the USGS researchers, Ed Doheny, Roger Starsonick, and Bob Shedlock for their
tremendous work in collecting surface and ground water hydrology data and preparing the pre-restoration surface water
hydrology analysis. We thank Baltimore County Department of Environmental Protection and Resource Management,
especially Don Outen, Steve Stewart, Karen Ogle, and Candy Crowell for their assistance in site selection and
continuing project support. We also thank Brad Scroggins, Ken Jewell, Steve Acree, Randall Ross, and Russell Neill
for invaluable field and technical assistance. Finally, we thank the Cromwell Valley Park personnel, especially Leo
Rebetsky, for supporting the project and providing unfettered access to the stream site at all times.
-------
Abstract
In Fall 2001, EPA undertook an intensive collaborative research effort with the US Geological Survey (USGS) and
the Gary Institute of Ecosystem Studies to evaluate the impact of restoration on water quality at a degraded stream in
an urban watershed using a before/after stream restoration study design. One objective was to evaluate if particular
stream restoration techniques improve ground water-surface water interaction (GSI) and if beneficial hydrologic
exchanges between the stream and riparian/floodplain may be enhanced to improve water quality. An essential piece
of this comprehensive study was to characterize, measure, and quantify near-stream (GSI) before and after stream
restoration at specific stream features and assess how the geomorphology and geology at each feature impact GSI.
This research report describes the pre-restoration study of GSI at specific stream features in a degraded urban stream
in Towson, MD. The study employed a comprehensive evaluation of the surface water hydrology, ground water
hydrology, geomorphology, and geology along a specific stream reach slated for restoration. Ground water level
measurements in piezometer nests in the stream bed and banks over time were found to be sufficient to characterize
the losing or gaining nature of near-stream GSI. Temperature measurements were used to verify these interactions.
The GSI was simply and effectively quantified using gradients calculated from the piezometer nest ground water
levels and Darcy's law in a simple compartment model. Flow was quantified and used to calculate residence times in
the sediments. These residence times may be used to quantify the mass removal of nutrients and other contaminants
if reaction kinetics are known. Results of the pre-restoration study reveal the highly variable nature of GSI on the
temporal and spatial scales of interest. Results also reveal how specific stream features and settings influence GSI.
Flow and residence time were found to be closely dependent on the stream feature geology and geomorphology.
Consequently, any restoration that impacts these features likely will strongly influence GSI. Results of this study
established the pre-restoration GSI. An identical study of the post-restoration GSI is underway. These results will be
compared to the pre-restoration state in a second report to evaluate the impact of stream restoration on GSI to determine
if improvements in water quality may be achieved.
Keywords: Baltimore, Maryland; geomorphology; ground water; ground water-surface water interaction; GSI;
hydrology; hyporheic; Minebank Run; restoration; temperature; urban stream
-------
-------
1.0
Introduction
Stream restoration is comprised of a group of techniques
that are intended to improve the physical, chemical and
ecological functions of a degraded stream (Falk et al.
2006). Physical functions include the stream's ability to
manage its flow and sediment load. Chemical functions
include the stream's assimilative capacity to process
nutrients and other contamination to maintain stream
water quality. Ecological functions include providing
aquatic and riparian habitats. All of these beneficial
stream functions are intimately tied to the stream flow
and the exchange of water in and out of the stream
bed and banks known as ground water-surface water
interaction (GSI).
GSI in streams is broadly defined as the exchange of
ground water and surface water of a stream at several
temporal or spatial scales. Watershed scale GSI is on the
order of miles and involves ground water discharge over
an entire watershed which supplies the stream base flow.
Near-stream scale GSI, on the order of feet and days,
captures the losing/gaining reaches of streams where
transport and processing of nutrients that support aquatic
and riparian habitats occurs. Sediment scale GSI refers
to exchanges in the stream beds and banks that occur on
the scale of inches and minutes. GSI in the near-stream
and sediment scale is often called hyporheic flow and is
critical for support of aquatic life.
One goal of stream restoration has been to improve
GSI in the belief that enhanced exchange will improve
the physical, chemical, and ecological functions of a
stream. A critical key to unlock the impact of restoration
on stream function is the ability to quantify GSI flow
in and out of the stream and evaluate how restoration
impacts this flow. In the case of water quality, GSI flow,
magnitude, and direction may be used to calculate mass
removal of nutrients and contaminants in the stream bed
and banks using the measured reaction kinetics before
and after restoration. Efforts to understand the physical
factors which influence GSI and to develop practical
methods to quantify it, however, have been hindered
by the inherent complexity of the multiple temporal
and spatial scales of the interaction. The challenges of
evaluating near-stream GSI at various stream features, in
particular, remain an impediment to assessing whether or
not particular stream restoration actions have improved
GSI and stream function.
In 2001, EPA/ORD/GWERD undertook an intensive
collaborative effort with the US Geological Survey
(USGS), Gary Institute of Ecosystem Studies and
Baltimore County Department of Environmental
Protection and Resource Management (DEPRM) to
evaluate the impact of restoration on a degraded stream
in an urban watershed using a before/after stream
restoration study design. The main objective was to
assess if stream restoration would improve the water
quality of a degraded stream by improving its physical,
chemical, and ecological functions. An essential piece
of this comprehensive study was to evaluate the effects
of particular stream restoration techniques on GSI. The
hypothesis was that a degraded stream has poor GSI and
if restoration could enhance GSI, water quality could
be improved through beneficial hydrologic interactions
between the stream and riparian/floodplain.
To assess the impact of restoration on GSI, it was
essential to characterize, measure, and quantify near-
stream GSI at specific stream features. Our study,
therefore, had two phases, one before and one after
restoration. The objectives of the pre-restoration study
which took place from Fall 2001 to Summer of 2004
were to:
1. Evaluate the stream geomorphology and geology
at specific stream features and assess how stream
geomorphology and geology at specific stream
features influences GSI.
2. Develop effective and simple methods to
characterize, measure, quantify and verify GSI at
specific stream features to establish baseline GSI in
the unrestored stream for comparison to the restored
state.
The objectives of the post-restoration phase, currently
underway, are to employ methods developed during
pre-restoration phase to evaluate water quality benefits
of specific stream restoration techniques due to effects
on GSI.
This research report describes the results of the pre-
restoration assessment of GSI at specific stream features
and methods to effectively and simply quantify it. The
second research phase will describe the post-restoration
GSI in relation to the pre-restoration state and quantify
the mass removal of nutrients and contaminants
which can be attributed to restoration impact on GSI.
This comparison is intended to assess the impact of
restoration on GSI and water quality due to specific
stream restoration techniques. The results should
provide restoration designers and stakeholders practical
methods to measure GSI to determine mass removal
of nutrients and contaminants in stream sediments and
identify specific restoration techniques which improve
GSI to enhance water quality and other beneficial stream
functions.
-------
-------
2.0
Background
Numerous hydrologic, chemical, and ecological benefits
to streams and riparian areas have been associated
with near-stream GSI (Boulton et al, 1998; Cirmo
and McDonnell, 1997; Hayashi and Rosenberry, 2002;
Kasahara and Hill, 2006; Poole et al., 2006). Often a
claim is made that stream restoration can enhance GSI
and benefit stream function. For example, restoring
incised streams in order to reconnect stream channels to
floodplains, may improve near-stream GSI. Interaction
may also be improved by incorporating permeable
sediments in stream beds and banks in place of low
conductivity strata. These changes may increase the
transport of nutrient laden water into the organic rich
soil profiles in the floodplain and riparian zones. These
nutrients may be utilized by riparian plants or denitrified
by microbial activity, thereby improving water quality
by reducing excess nitrogen and other nutrients. There
is also the potential for removal of contaminants through
adsorption in the stream sediments. In addition,
restoring stream channels may enhance a stream's ability
to manage its flow and sediment load, reducing sediment
transport. In this way, restoration which improves GSI is
predicted to also improve water quality.
To date, few studies have assessed the effects of
restoration on GSI (Kasahara and Hill, 2006). If an
objective of stream restoration is to improve GSI, it
is critical to characterize, measure, and quantify this
exchange in the near-stream environment. However,
GSI is notoriously difficult to evaluate because
interactions are hidden and occur at spatially and
temporally variable scales (Winter et al. 1998). In
addition, the literature has employed inconsistent
terminology to describe GSI.
Examples of GSI are demonstrated in Figure 1, which
represents a general depiction of the sources of water
to a stream in a watershed. Precipitation falling to the
earth may become surface runoff to the stream or enter
the subsurface ground water system. GSI includes the
ground water and shallow subsurface flow that moves
down gradient and discharges to the stream. GSI also
encompasses stream water exiting and re-entering
stream beds and banks through near-stream paths. The
scales of these interactions are spatially and temporally
variable and restoration will have a different impact on
each.
To simplify the scale issue, this study employed a
definition developed by Boulton et al. (1998) who
separated GSI into three main scale categories;
watershed, near-stream, and sediment scale GSI.
Boulton et al. described watershed scale GSI as the
regional perennial ground water flow to the stream
shown in Figure 1. Watershed GSI occurs on the scale of
miles and years and is typically defined as the base flow
of a stream which is measured by stream gaging. At this
scale, the ground water is assumed to discharge evenly
to the stream with no variation along its length. As
this discharge is a function of the watershed hydrologic
setting, including ground water storage characteristics
and climate, watershed GSI is typically not targeted and
is unlikely to be modified in any significant fashion by
stream restoration.
Near-stream GSI occurs on the scale of feet and days,
as stream and ground water flows into and out of stream
beds and banks (Figure 1). Near-stream GSI captures
the spatial variability of the interaction along the length
of a stream where reaches may be losing or gaining. In
a losing reach, water moves from the stream into the bed
and banks to ground water. In a gaining reach, ground
water moves into the stream, thereby providing a portion
of its flow. Stream restoration is highly likely to affect
near-stream GSI, especially if it impacts the stream
channel features, including stream depth, width, and bed
and bank materials.
Sediment-scale GSI involves the small exchanges on the
order of inches and minutes within the stream bed and
banks. This interaction is a function of various factors
including small changes in stream flow and stream bed
structure (Castro and Hornberger, 1991; Savant et al.,
1987). Stream restoration may greatly impact sediment-
scale GSI but designing stream restorations to target
these specific interactions may not be practical as they
are difficult to maintain. Near-stream and sediment
scale GSI are often referred to as hyporheic exchange,
a lumped term that we will avoid using to prevent
confusion.
For this study, near-stream GSI was of special interest
because the exchange of surface and ground water in
the stream bed and banks may be significantly impacted
by stream restoration. Current methods available to
characterize and measure near-stream GSI include
stream and ground water level measurements, flow and
temperature surveys, seepage measurements, numerical
flow and heat modeling, and stream and subsurface
tracer studies (Harvey and Wagner, 2000; Kalbus et al.,
2006). These approaches have been used alone or in
combination to measure and quantify GSI.
-------
water
table
saturated
overland
flow
GSI
Figure 1. Scales of ground water surface water interaction (image adapted from Stream Corridor Restoration Hand-
book; FISWRG, 1998).
One set of techniques to quantify GSI employs ground
water levels and stream flow measurements, usually
coupled with ground water/stream tracer studies and
analytical/numerical flow modeling. Bencala et al.
(1984) and Bencala (1984) performed comprehensive
tracer studies of streams to interpret the dynamic
physical and chemical properties that control tracer
transport in the stream and hyporheic zone. Harvey
and Bencala (1993) were able to simulate and assess
the impact of stream features on hyporheic GSI by
measuring near-stream ground water levels and tracer
studies. Wroblicky et al. (1998) also used water levels
and tracer studies in transient numerical flow simulations
to quantify the spatial and seasonal variation in reach
scale GSI. Harvey et al. (1996) evaluated the reliability
of tracer studies by testing their sensitivity to stream
flow conditions. Choi and Harvey (2000) identified the
need to address several storage zones in the stream to
account for tracer behavior. Harvey and Wagner (2000)
and Kalbus et al. (2006) described and summarized
methods to quantify near-stream GSI.
Another set of techniques to quantify GSI uses stream
and ground water temperature monitoring and heat
modeling. Silliman and Booth (1993) and Constantz
(1998) provided some of the first examples of the use of
temperature measurements to identify losing and gaining
portions of streams. Silliman et al. (1995) also provided
mathematical formulations to quantify flux across the
streambed for one-dimensional flow using measured
temperatures. Conant (2004) used ground water levels
and mapped streambed temperature to quantify stream
bed GSI. Becker et al. (2004) quantified ground water
discharge using stream flow measurements, stream
temperature surveys and heat transport modeling of
temperature gradients below the stream bed. Stonestrom
and Constantz (2004) and Stonestrom and Constantz
(2003) provide excellent descriptions of temperature
measurement and modeling methods to quantify GSI.
Once GSI is measured and quantified, it may be
possible to evaluate the factors that influence it and
those that may be modified through stream restoration.
Many studies have established the influence of stream
geomorphology and geology on GSI along a stream
reach (Savant et al., 1987; Castro and Hornberger,
1991; Eshelman et al., 1994; Fryar et al., 2000; Harvey
and Bencala, 1993; Poole et al., 2006; Gooseff et al.,
2006). For example, streams in high conductivity
alluvial settings composed of cobbles and sands offer
less resistance and therefore greater potential for GSI.
Streams flowing through heavy clays would be expected
to have less interaction. Highly incised streams in
consolidated bed rock would likely experience less GSI
than stream meandering through sandy point bars. Deep
pools with low velocities will likely interact differently
with the ground water system than swift flowing
reaches. In addition, geomorphology and geology
will affect the residence time in the stream sediments,
thereby influencing biological interactions in the near
stream environment. Longer residence times may allow
sustained microbial activity such as denitrification. It
is therefore, of interest to determine how geologic and
geomorphologic characteristics of the stream influence
hydrology and residence time.
-------
3.0
Site Description
The stream selected for this study was Minebank
Run (MBR), a second order stream located in a small
watershed in the Piedmont physiographic region of
Maryland in the south central section of Baltimore
County (Figure 2). The headwaters of MBR are on
the east side of Towson, MD and its outlet confluences
with lower Gunpowder Falls, a major tributary of the
Chesapeake Bay.
The choice of MBR was made after several
reconnaissance trips with Baltimore Department of
Environmental Protection and Resource Management
(DEPRM) personnel to streams slated for restoration
in Baltimore County. MBR had already undergone
successful restoration of its headwaters in 1999 and its
lower reaches were scheduled to be restored in 2004
and 2005. The unrestored reach possessed several
characteristics that made it ideal for the before/after
study, including a heavily degraded stream condition,
a comprehensive geomorphic and riparian restoration
design with significant manipulation of the stream
channel and banks along the entire length of stream, a
USGS stream gaging station, and convenient site access.
The MBR watershed covers 3.24 square miles and the
stream itself is three miles in length (Figure 3). Stream
channel slopes are <1 percent in most locations (Doheny
et al., 2006). Relief varies from 100 to 300 ft in the
watershed. There is about a 340 ft drop in elevation from
the headwaters of MBR to its outlet at Gunpowder Falls.
The watershed geology is composed of a complex series
of crystalline rocks including the Setters formation,
the Cockeysville Marble and the Loch Raven Schist
(Doheny et al., 2006). Underlying this group is the
Baltimore Gneiss. Maps of surface geology show that
the crystalline rocks are overlain by alluvial deposits
composed of cobbles, gravels, and sands (Crowley et
al., 1976). Some clay deposits were noted in the upper
reaches of the watershed. The stream flows through
exposed crystalline rocks in the upper reaches and
unconsolidated deposits of colluvium and alluvium in
lower reaches.
MBR is a classic example of stream that has undergone
massive changes as a consequence of urban development
in the watershed. When settled, the watershed was
historically agricultural, but is now about 80% urbanized
TlfX
HoH^vA-
±~ ,^r$L
^^:%i N w^htxW XA^NT
^.A^ 1) "oN^^ 2fr^=
S^AnJSL'^
YJ 4 ' ^
^ s T -4;
%7 AN NE\\ARU-N-C)EL '
w\^ryz$4
10 15 MILES
Figure 2. Minebank Run watershed geographic location (adapted from "Watershed Characteristics and Pre-Restoration
Surface-Water Hydrology of Minebank Run, Baltimore County, Maryland, Water Years 2002-04," Doheny etal., 2006).
-------
0.5 1 HLMiETER
Figure 3. Minebank Run Watershed (adapted from "Watershed Characteristics and Pre-Restoration Surface-Water
Hydrology of Minebank Run, Baltimore County, Maryland, Water Years 2002-04," Doheny et al., 2006).
with over 25% impervious surfaces (Doheny et al,
2006). The stream has several runoff outfalls along
its length including one from the 1-95 corridor. The
combination of fairly large stream slopes, significant
relief and runoff from impervious surfaces causes
the stream to experience pronounced flashy behavior
during storm events (Doheny et al, 2006). MBR has
experienced many of the geomorphic changes expected
with flashy urban flows and is deeply incised with
little to no riparian buffer. This incision has exposed
infrastructure such as sewer lines and caused extensive
bank failure, threatening personal property (Figure 4).
In addition, this incision contributes to heavy sediment
loads and diminished ecological condition.
As a consequence of its poor condition and threats to
property and infrastructure, Minebank Run was one
of the streams targeted by DEPRM for restoration. A
detailed restoration study was made of the stream and
comprehensive restoration designs for the entire stream
were prepared. In 1999, about 1.5 miles of MBR
from the headwaters downstream were restored. The
remainder of the stream was scheduled to be restored
in 2004. The headwater restoration included Natural
Stream Channel Design (NSCD) techniques such as
bank armoring with rip-rap, step pools, and meanders
(Rosgen, 1996). Riparian restoration included geotextile
bank stabilization and substantial riparian revegetation.
Although these efforts were intended solely to achieve
geomorphic stability to protect infrastructure and
property, we speculated that restoration may influence
GSI and water quality of MBR.
Figure 4. Incision and infrastructure exposure at Mine-
bank Run before restoration.
-------
4.0
Methods
Study design and site characterization
In the summer of2001,ashort section of the unrestored
reach of MBR in Cromwell Valley Park, Towson, MD
was selected for intensive study (Figure 5). A
continuous stream gauge, 01583980 Minebank Run
at Loch Raven, MD, had already been installed in
the park by the USGS at MBR in 2000 (Figure 5). A
new continuous stream gage, 0158397967 Minebank
Run near Glen Arm, MD, was installed upstream of
the study reach (Figure 5). A continuous rain gage,
392449076331100 Minebank Run Rain Gage, was also
installed to provide a real time precipitation record
(Figure 5).
In the Fall 2001, the locations of the three transects
were chosen for intensive study of the stream before
restoration (Figure 5). These locations were selected
based on the unrestored condition of the reach and the
restoration plans (Figure 6). The intent was to position
transects in locations where degraded stream features
would undergo a specific restoration technique. The
study was designed to measure baseline conditions
in the degraded stream condition and then measure
the response to the restoration. To that end, the
geomorphology of Minebank Run was evaluated by the
USGS and published in a recent report (Doheny et al.,
2007).
Transect 1, was located at a deeply incised pool where
a new stream channel was to be created (Figure 5).
Transect 2, was positioned 237 ft upstream of Transect
1 and placed across an incised narrow riffle and point
bar which was to be armored and raised (Figure 5).
Transect 3, was positioned 148 ft upstream of Transect 2
and located in a flat shallow terrace section which was
to undergo little modification (Figure 5). An ancillary,
3-piezometer nest was placed in the stream channel
approximately 100 ft downstream of Transect 1 was
occasionally sampled for hydrology.
Ttfmtr
39°25Kr
CROMWELL VALLEY PARK
EAND IDENTIFICATION NUMBER
> WELL TRANSECT (conslsffng of 2-Inch monitoring wells on
each flood plain, and 1-inch piezometer nests in the channel
bed and on each channel bank)
MINEBANK HUN STUDY REACH
Well transect No. 1—,
(Station 0158397971) \
Well transect No. 2 -,
(Station 0158397969) \
Well transect No. 3 A \
(Station 01583979681 \
Station x
0158397967
{Minebank Run
near Glen Arm, MD),
-«—12-foat diameter Loch Ravan-Montabello Tunnel
{Loch Raven Reservoir to Montabollo Filtration Plant!
Figure 5. Location of the Minebank Run Study Reach in Cromwell Valley Park (adapted from" Watershed Charac-
teristics and Pre-Restoration Surface-Water Hydrology of Minebank Run, Baltimore County, Maryland, Water Years
2002-04," Doheny et al., 2006).
-------
Unrestored Stream
Restored Stream Thalweg ^™
Restored Stream Banks ^™
Rock Structures
Existing Seweriine
Minebank Run
0 50 100 150 FEET
Figure 6. Location of monitoring well transects in relation to restoration design (Figure courtesy of Maryland Depart-
ment of Environmental Protection and Resource Management).
In November 2001, an EPA field team installed the
monitoring wells and piezometers at each transect at
the site. The transect monitoring design (Figure 7)
was composed of paired two-inch diameter monitoring
wells in the floodplain and nests of three piezometers
in the stream bed and banks. This design was intended
to capture the spatial and temporal scales of near-
stream GSI and the larger regional ground water flow
system. For each transect, nests of three piezometers
were installed in the stream bed and banks using a
Simco™ direct push rig. The piezometers were one-inch
diameter stainless steel with six inch long wire-wound
screens with 0.01 inch mesh (Figure 8). Piezometer
nests installed in the stream bed were driven to depths
of approximately two, four, and six feet below ground
surface in the thalweg. Next, piezometer nests were
installed on each banks close to the stream edge. Bank
piezometers were driven to depth to match the same
mean sea level elevations of the piezometers in the
stream. Throughout this report, the piezometers in the
stream bed and banks will be identified as shallow,
medium, and deep.
In addition to the piezometer nests, two-inch PVC cased
monitoring wells with five-foot screens were installed
in the floodplain near the stream. Pairs of wells were
placed approximately fifty feet perpendicularly from
the stream thalweg. One well of each pair was installed
just below the water table and the second well of each
pair was installed to the point of bedrock refusal. Single
monitoring wells were placed approximately one
hundred feet from the stream thalweg. After installation,
all piezometers and monitoring wells were developed to
ensure proper ground water sampling. The piezometer
and monitoring well locations and elevations were
surveyed using standard line of sight methods. The
aerial map shows the georeferenced locations of the
monitoring wells and piezometer nests (Figure 9).
As the monitoring wells were installed, 5 cm diameter
(ID) continuous cores were collected to refusal using
a Geoprobe™ direct drive system. The cores were
characterized to describe the lithology. In July 2004,
additional cores were collected at all of the stream bed
and bank piezometer nests and characterized. Slug
testing was done at all of the piezometers using falling
and rising head tests to determine conductivity (Bouwer
and Rice, 1976 and Bouwer, 1989).
Water levels were measured at all of the wells and
piezometers approximately every two weeks starting
in January 2002. To capture the continuous time scale,
many of the piezometer nests and some of the floodplain
wells were fitted with Solinst™ Model 3001 data loggers
in March 2002 to measure temperature and ground
water elevation every five minutes. These data loggers
were downloaded every three months. Ground water
data were collected until the wells were removed for the
restoration in July 2004.
-------
Shallow/deep
monitoring wells
Piezometer well
nests
Shallow/deep
monitoring wells
&
Figure 7. Transect monitoring well network design. Distanc- Figure 8. Stainless steel piezometer used in piezom-
es not to scale (image adapted from Stream Corridor Restora- eter nests.
tion Handbook, FISWRG, 1998).
0 10 20 30 40*
Figure 9. Location of monitoring wells and piezometer nests Minebank Run Study reach.
All of the site characterization and ground water level
data were measured in English units of feet for increased
resolution at the small scales involved in this study.
All of the calculations were also done in units of feet
to maintain consistency. Data and calculations were
reviewed and stored in electronic form in databases or
MS Excel spreadsheets according to the approved EPA
Quality Assurance Project Plan. The USGS stream
and rain gage data for MBR are available on line http://
waterdata.usgs.gov/nwis/.
Ground water-surface water interaction analysis
One factor known to impact near-stream GSI is the
geomorphological setting, the physical relationship
between the stream and the ground surface. The stream
geomorphological setting at each of the transects was
evaluated using field observations, stream profiles, and
stream velocity. The stream profile defines the cross
sectional area at each transect which, in turn, determines
the channel velocity for a given stream flow rate. The
flow velocity through each transect was evaluated to
determine if geomorphology created a stream hydrologic
setting that influenced the GSI.
-------
Another factor known to influence near-stream GSI is
the geological setting. Geological setting is important
because the lithology provides the stream bed location
relative to the geologic layers. Geological setting also
defines the conductivity of the sediments, a measure of
the resistance to ground water-surface water exchange
in the near steam sediments, with low conductivities
limiting exchange and high conductivities enhancing
exchange. If a stream is situated in highly conductive
alluvial sediments such as sands and gravels, water
will exchange easily in these zones. If fractures or
other preferential flow features are present, GSI may be
enhanced. If a stream is in silt, clays or incising into
rock, near-stream GSI is limited by these non-conductive
sediments. The geologic setting at MBR was defined
using cores to develop lithologic cross sections and slug
testing to assess hydraulic conductivity (Bouwer and
Rice, 1976 and Bouwer, 1989).
To describe the watershed scale GSI at MBR, the
regional stream and ground water hydrology were
assessed. The USGS evaluated the MBR site pre-
restoration stream hydrology data and published a
comprehensive surface water hydrology report (Doheny,
et al., 2006). The regional ground water flow system
was defined using traditional horizontal ground water
equipotential contours across the site developed from
the biweekly ground water level data. These contours
provided the ground water flow magnitude and direction
across the site and gave some insight into the interaction
of the regional ground water flow with the stream.
The near-stream GSI on the scale of tens of feet was
demonstrated at each of the transects using vertical
equipotential contours from piezometer water levels
measured biweekly (Figure 10). These contours allowed
the potential gradients under each of the stream transects
to be visualized and the flow directions inferred.
Although numerous methods were available to quantify
GSI, it was immediately apparent from the data that
most were not suited to the situation at MBR. Ground
water equipotentials evaluated both seasonally and for
the ground water high and low during the study at each
of the transects demonstrated a complex flow system
that was spatially and temporally variable. Traditional
tracer and modeling techniques (Harvey and Bencala
1993) typically assume steady state conditions and
were not suitable to capture the spatial and temporal
variation observed in the near-stream GSI at Minebank
Run. Tracer tests would need to be run each time the
flow field varied to capture its unique signature. Flow
and temperature modeling was also impractical as the
models require many boundary and hydrologic condition
assumptions and therefore, the true transient nature
of the near-stream GSI would be nearly impossible to
reproduce.
An objective of the study was to develop a transparent
and effective method to quantify the GSI before and after
restoration. We decided to employ a simple hydrologic
flow analysis which used measured water levels from the
piezometer nests. The vertical and horizontal gradients
were calculated from these measured values at each
of the transects. These gradients were used in Darcy's
Stream Cross Section
Ground Water Surface X Streamjhalweg Divide
is
to
a
-10 -50 5 10 15
Horizontal distance relative to stream thalweg location (ft)
Figure 10. Example vertical equipotential stream cross section with left bank and right bank compartments on either
side of the stream thalweg divide.
-------
law to determine the flow at each of the transects over
time by applying some simple assumptions. The first
assumption was that the thalweg of the stream acts as a
ground water divide at shallow depths near the stream
bed such that flow on one side of the stream is not
influenced by the flow on the other. In small headwater
streams with partial penetration and perennial base flow
such as MBR, this approach was strongly supported
by near-stream piezometer temperature and ground
water level measurements at the site. This allowed
the flow system to be split into a left and a right bank
compartment on either side of the stream thalweg divide
as shown in Figure 10.
The second assumption was the ground water seepage,
q, in each compartment can be represented by one
constant linear vector. This vector was determined
by calculating the vertical and horizontal gradients in
each compartment separately based on the water levels
measured in the stream bed and bank piezometers and
applying Darcy's law:
(i)
measured bulk densities at each transect. A volumetric
flow for the compartment was also determined:
,^.
= -k—j+—
dz
where k was the hydraulic conductivity measured by
slug tests at the site. The horizontal gradient, dhldx, was
calculated for each bank using the difference between
the piezometer water levels in the stream bank and
stream beds. Concurrently, the vertical gradient, dhldz,
was determined using the water levels in stream bed
piezometers. A negative horizontal or vertical gradient
was defined as flow in toward the stream. The resultant
seepage vector through the compartment on either side
of the stream center line was then defined by determining
its magnitude, dhlds, , and its direction counterclockwise
from horizontal, 6, using equations 2 and 3.
a/i
tan6 =
a*
(2)
(3)
Once the magnitude of the seepage vector and its
direction were known, it was possible to calculate
the residence time, /, along a path line, s, through the
compartment:
t = s/(q/n) (4)
where n is the porosity which was calculated from core
Q = qA
(5)
where A is the unit cross sectional area perpendicular to
the flow vector.
The GSI flow rates for each transect were calculated on
several dates to demonstrate the spatial and temporal
variation in the near-stream GSI at MBR. To verify the
results, the GSI rates were compared to the measured
stream flow rates to assess if they supported the losing
and gaining nature of the stream along the reach. In
addition, the continuous record of temperature obtained
from the Solinst™ Model 3001 data loggers installed in
the shallow and medium depth piezometers in the stream
beds was evaluated to verify the losing or gaining nature
of the GSI at each stream transect using the methods of
Stonestrom and Constantz (2003).
MBR experienced several storm surges during the
study. The impact of these events was captured by
the continuous data loggers which measured level
and temperature every five minutes. The continuous
temperature record during the storm surge in the
piezometers was evaluated using the methods of
Stonestrom and Constantz (2003) to help assess the
residence time of the stream water as the surge drove
water into the stream bed sediments. Water levels in
the stream bed and bank piezometers were plotted to
show how one particular storm surge event influenced
GSI. Darcy's law was applied as a first approximation
to make an estimate of the storm surge flow rates into
the stream bed which varied continuously over time.
The application of Darcy's law under these conditions
was not rigorous. Typically one may use the gradients
from these measurements and Darcy's law to assess the
seepage velocity into the stream bed as was done for the
vertical equipotentials for the bi-weekly water level data.
However in the case of a storm surge, the hydraulic
gradient in the stream bed and banks is continuously
changing as a function of time so flow is not steady.
The flow is also moving at high velocity, which means
it is turbulent, with Re>l. Therefore Darcy's law is not
truly applicable as it assumes steady laminar flow with a
Reynolds number, Re
-------
-------
5.0
Results
Stream geomorphology
At MBR, transects were strategically placed so
the unrestored stream physical features could be
used to provide insights into the impact of stream
geomorphology on stream flow behavior and the pre-
restoration GSI. Transect 1 was located at a deep pool
with severe incision on the left bank (top half of picture)
and a depositional point bar on the right bank (lower
half of picture) (Figure 11). Transect 2 was located at
a shallow, narrow riffle with an incised left bank and
a point bar on the right bank (Figure 12). Transect 3
was located at a shallow, wide terrace (Figure 13). The
piezometer nests can be seen in the stream bed and
banks (Figure 11). The stream bed piezometers were
replaced flush with the stream after a storm event bent
them (Figure 13).
Figure 11. Transect 1 at beginning of study in Fall 2001.
Figure 13. Transect 3 at beginning of study in Fall 2001.
The mean sea level ground surface elevation of the
stream profiles at Transects 1, 2, and 3 at the beginning
of the study is shown in Figure 14. At Transect 1,
stream velocity at base flow was measured as 0.19 ft/sec.
This low velocity was a consequence of incision at
this transect which had created a deep, wide pool with
large cross sectional area. This low velocity created a
relatively static setting for interaction with the ground
water in the stream bed and banks. At Transect 2,
the stream flowed through a narrow shallow cross
section with a measured base flow stream velocity
of 0.66 ft/sec. This high velocity created a dynamic
setting for interaction with the underlying ground
water. At Transect 3, flow moved through a large flat
cross sectional area defined by the stream terrace. The
measured base flow stream velocity was 0.29 ft/sec and
created a static environment for GSI.
•-•-,
Figure 12. Transect 2 at beginning of study in Fall 2001.
-------
Transect Stream Profiles
November 2001
226
Downstream
212
-20 -10 0 10
Distance from stream thalweg piezometer (ft)
Figure 14. Ground surface elevation cross sections at Transects 1, 2, and 3 at beginning of study.
Stream geological setting
At MBR, cores were collected and characterized to
create several cross sections to define the geological
setting (Figure 15). Cross section A-A' was defined
along the stream thalweg and shows the lithology
described by the core holes near the deep stream
bed piezometers, 146, 149 and 153 (Figure 16). The
lithology directly under the stream was gravelly clay
ranging from five to ten feet thick, underlain by a clay
layer. A thick layer of poorly sorted sand was found
under the clay. Sediment cores did not reach bed rock at
these sites.
0 50 100 150 FEET
0 10 20 30 40 METERS
Figure 15. Geological cross section lines for Minebank Run study site.
-------
GEOLOGIC CROSS SECTION
Figure 16. Minebank Run Thalweg Cross Section A-A'.
Cross section B-B' was constructed based on cores
extracted along Transect 3 (Figure 17). Piezometer 146
was in the thalweg of the stream, 162 was in the left
bank and 157 was in the right bank. The other cores
were located at the floodplain monitoring wells. The
lithology on the left side of the stream was a top layer of
sandy clay underlain by a gravelly clay that varied from
3-6 feet thick. On the right side of the stream, the top
layer was silty clay underlain by gravelly clay near the
stream bed and sandy clay farther away from the stream.
The clay layer was only found on the right side of the
stream and was a few feet thick. A continuous layer of
poorly sorted sand was encountered under both sides
of the stream. Limestone bedrock was encountered on
both sides of the stream but not reached directly under
the stream.
GEOLOGIC CROSS SECTION
Cross Section B-B1
Mine Bank Run, Towson, MD
TranectS
Figure 17. Transect 3 Cross Section B-B'.
-------
Cross section C-C' was constructed from cores collected
along Transect 2 (Figure 18). Piezometer 149 was in
the thalweg of the stream, 167 was located in the left
bank, and 171 was in the right bank. On the left side of
the stream, the lithology was sandy clay underlain by
the gravelly clay. On the right side of the stream, the
top layer was part of a point bar composed of gravels
and underlain by gravelly clay. The clay layer was only
encountered under the right side of the stream. A thick
continuous layer of poorly sorted sand was encountered
under both sides of the stream. Limestone bedrock was
encountered on the left side of the stream but was not
reached directly under the stream or on the right side.
GEOLOGIC CROSS SECTION
40 50 SO 70 BO 90 100 110 120 130 140 150 160 170 180 190 200
Figure 18. Transect 2 Cross Section C-C'.
Cross section D-D' was constructed from the cores col-
lected along Transect 1 (Figure 19). Piezometer 153 was
in the thalweg of the stream and piezometers 176 and
179 were located in the left and right bank, respectively.
Both sides of the stream exhibited a top layer of sandy
clay underlain by the gravelly clay. The gravelly clay
was about twice as thick on the right side of the stream
as on the left. The clay layer was evident under the right
bank of the stream but disappeared on the left bank. A
thick continuous layer of poorly sorted sand was en-
countered under both sides of the stream. The limestone
bedrock, located at shallow depths under the left side
dipped steeply before flattening out under the stream and
right hand side.
GEOLOGIC CROSS SECTION
Cross Section D-D'
Mine Bank Run, Towson, MD
Transect 1
~l iso
DISTANCE (feet)
Figure 19. Transect 1 Cross Section D-D'.
-------
These cross sections demonstrate that the stream and
floodplain were composed of thick sections of gravelly
clay, clay, and sand underlain by limestone bedrock. The
gravelly clay and sands were found on both sides of the
stream and the clay was located on the right side in the
floodplain. The limestone bedrock was encountered
at depth and cores indicated it was weathered and
fractured. However, as the limestone was so removed
from the stream bed, that the possibility of preferential
flow from the bedrock fractures to the stream was
dismissed.
Before this geological characterization, the MBR
watershed was described as being composed of strictly
heterogeneous alluvial deposits ranging from cobbles,
gravels and sands which were underlain by limestone
bedrock. A search was made to determine if a manmade
dam was used in the watershed to account for the clay
layers (Walter and Merritts 2008), but no historical
evidence for such a structure could be located. However,
historical records, including maps and photographs,
were discovered of extensive iron mining operations
active until the 1800s that stripped topsoil throughout the
watershed (Dorothy Merritts, personal communication).
This type of mining involved digging shallow pits along
the watershed, excavating a shipping canal parallel to the
current stream thalweg, and constructing limestone kilns
along the stream channel. Slag from historic smelting
operations can be seen in the stream channel today
and some unusual sediment layers observed in incised
banks of Minebank Run are probably derived from the
intensive mining activity of this era (Dorothy Merritts,
personal communication). Based on this information,
the geological setting for Minebank Run has clearly been
very disturbed.
To complete the description of the geological setting
and interpret its impact on GSI, conductivity, k, was
measured using slug tests at each of the transect
piezometers at MBR. The results are shown in Tables
1, 2 and 3, respectively. The measured values were in
general agreement with the geological characterizations
of the cores in the stream bed and banks. At Transect
1, the piezometers were screened in a heterogeneous
gravelly clay and the conductivity ranged from 0.01 to
310.0 ft/day. The low values in the deeper piezometers
indicate they were in a clay layer whereas the medium
depth stream bed and right bank piezometers were
most likely screened in cobble layer, perhaps part of
an abandoned channel not clearly identifiable from
the cores. At Transect 2, the conductivities were also
representative of a heterogeneous gravelly clay but had
less of a range from 0.07- 9.2 ft/day. At Transect 3,
the conductivities were consistent and higher because
the piezometers were screened in the sand, with the
exception of one deep piezometer in the clay. The range
was 10.0-58.0 ft/day.
Table 1. Transect 1 Piezometer Conductivity
Name
BAEe176
BAEe177
BAEe178
BAEe153
BAEe154
BAEe155
BAEe179
BAEe180
BAEe181
Location
T1 LB deep
T1 LB medium
T1 LB shallow
T1 Stream deep
T1 Stream medium
T1 Stream shallow
T1 RBdeep
T1 RB medium
T1 RB shallow
Estimated K
(ft/day)
0.01
6.90
3.70
0.14
310.00
ND
0.89
170.00
19.00
Table 2. Transect 2 Piezometer Conductivity
Name
BAEe167
BAEe168
BAEe169
BAEe149
BAEe150
BAEe151
BAEe171
BAEe172
BAEe173
Location
T2 LB deep
T2 LB medium
T2 LB shallow
T2 Stream deep
T2 Stream medium
T2 Stream shallow
T2 RB deep
T2 RB medium
T2 RB shallow
Estimated K
(ft/day)
8.00
9.20
0.07
2.70
5.10
ND
1.80
0.41
2.30
Table 3. Transect 3 Piezometer Conductivity
Name
BAEe162
BAEe163
BAEe164
BAEe146
BAEe147
BAEe148
BAEe157
BAEe158
BAEe159
Location
T3 LB deep
T3 LB medium
T3 LB shallow
T3 Stream deep
T3 Stream medium
T3 Stream shallow
T3 RB deep
T3 RB medium
T3 RB shallow
Estimated K
(ft/day)
46.00
38.00
49.00
0.14
14.00
47.00
58.00
10.00
32.00
The conductivity of the sediments in the stream bed
and banks have significant implications for GSI. The
magnitude of ground water flow in the bed and banks,
q, is directly proportional to the magnitude of the
conductivity, k, and the hydraulic gradient, h, as defined
in Darcy's law for flow in saturated sediments, q=kh.
For example, Transect 1 has high conductivities in the
shallow- and medium-depth stream bed and in the right
bank piezometers suggesting higher exchange of ground
water and surface water in response to the hydraulic
gradient created by the stream and regional ground water
elevations.
The left bank at Transect 1 had lower conductivity,
suggesting lower exchange rates under similar stream
-------
flow rates. All of the deep piezometers at Transect 1
had low conductivity, which suggested sediments at
this depth would limit ground water and surface water
exchange. Conductivity at Transect 2 was homogeneous
throughout all locations and depths, suggesting
homogeneous ground water exchange throughout all
depths across the channel. Transect 3, situated in sands
with conductivity an order of magnitude higher than
the other transects, would be expected to have greater
ground water-surface water exchange in response to
changes in the stream or regional ground water levels.
Regional surface water and ground water hy-
drology
The regional surface water hydrology of Minebank Run
was evaluated by the USGS and published in a recent
report (Doheny et al., 2006). According to the report,
the average annual precipitation for the Baltimore region
is 42 inches. Water year 2002 was considered a drought
with an annual precipitation of 32.95 inches. Water
year 2003 was a wet year with a total precipitation of
64.19 inches. In water year 2002, the low mean annual
discharge for MBR at the Glen Arm, MD, stream gage
just above Transect 3 was measured as 1.15 ftVsec. In
water year 2003, the high mean annual discharge of
4.34 ftVsec was reported at this same gage.
Using these values, it was possible to estimate a range
of the mean amount of ground water discharged to the
stream on a regional scale each day during years 2002
and 2003 of this study. Assuming that stream flow is
entirely base flow, a mean discharge of 1.15 ftVsec in
2002 translates into about 99,360 ftVday of ground
water entering the stream above the gage. For 2003,
the annual mean discharge of 4.34 ftVsec translates into
374,976 ftVday of ground water entering the stream
which represents a small portion of the ground water
storage in the 2.1 square miles drained above the gage.
At the watershed scale, assuming that this discharge
is equally distributed along the length of the stream
(ca. two miles for this gage), discharge ranges from
0.0036 ftVday/ft in 2002 to 0.0135 ftVday/ft in 2003.
These data demonstrate that little of the total ground
water enters the stream as base flow. Because MBR
only partially penetrates the thick sediments under the
thalweg as demonstrated in the geological cross sections,
the stream intercepts little ground water. The majority
of ground water moves slowly at depth down the valley
under the stream with very little interaction with the
stream because the ground water and surface water are
not defined by the same boundary.
This minimal interaction between the ground water and
stream system revealed by the stream hydrology at the
watershed scale can also be seen in the ground water
hydrology at MBR. Figure 20 displays the horizontal
equipotentials derived from the ground water level
measurements for each of the transects at the beginning
of the study on March 22, 2002 for a 400 x 400 ft
region around the stream reach. These equipotentials
show that ground water flow moves from roughly
west to east across the study site. The gradient ranges
from 2.0 - 4.0 ft/100 ft. With an estimated average
hydraulic conductivity of 20 ft/day, this translates into
a ground water seepage velocity of 0.4 - 0.8 ft/day.
The equipotentials on March 22, 2002 are not clearly
perturbed by the stream, which means that minimal
ground water was moving into the stream. These
data are corroborated by the stream gage flow of only
0.87 ftVsec on this date.
-------
: 14315900-
14315700^
1200800 1200900 1201000 1201100
Easting (UTM ft)
1201200 1201300
Figure 20. Regional ground water equipotentials on March 22, 2002.
In the summer of 2002, MBR experienced a severe
drought. Figure 21 displays the equipotentials on
August 22, 2002 for the ground water low created by
the drought. The flow was still from west to east. The
gaged stream flow for this date was 0.04 ftVsec. During
this time, the stream ceased to flow at Transects 1 and
2 and stream water began entering the ground water
system as recharge due to dropping ground water levels.
As a consequence, equipotentials show the complete lack
of interaction of the ground water with the stream.
14316100-
14316000^
14315900-
14315800-
14315700-
1200800 1200900 1201000 1201100 1201200 1201300
Easting (UTM ft)
Figure 21. Regional ground water equipotentials on August 22, 2002.
-------
In 2003, the MBR watershed received high rainfall
amounts. In response to this recharge, a ground
water level high was recorded on September 23, 2003
(Figure 22). The stream gage reflected this event with a
recorded flow of 7.7 ftVsec. The regional ground water
equipotentials, which clearly show the influence of the
stream, are strongly distorted by stream interaction at
both Transect 1 and Transect 3. The stream was acting
as the drain for the additional recharge.
1431610O
1431600O
1431590^
1431580O
1431570&
1200800 1200900 1201000 1201100 1201200 1201300
Easting (UTM ft)
Figure 22. Regional ground water equipotentials on September 23, 2003.
These equipotentials demonstrate the temporal
variability of the ground water levels and stream flow.
The ground water levels varied greatly over time at the
site, with water levels at some of the monitoring wells
differing by more than seven feet from the low to the
high (Table 4). The equipotentials also capture the
spatial variability in the ground water levels. The water
level variation in specific wells increases with increasing
distance from the stream; wells farthest from the stream
showed the greatest variation in water level while wells
close to the stream maintained a smaller range.
-------
Table 4. Comparison of ground water levels on dates of March 22, 2002, August 22, 2002, and September 23, 2003
Name
BAEel82
BAEel83
BAEel78
BAEel55
BAEel86
BAEe 181
BAEel91
BAEel92
BAEelTO
BAEe 175
BAEel69
BAEe 151
BAEel73
BAEel95
BAEel88
BAEel61
BAEel60
BAEe 166
BAEel64
BAEe 148
BAEe 159
BAEe 156
BAEe 145
Location
T1MW 100 ft Left bank
T1MW 5 Oft Left bank
Tl Left bank shallow
Tl Stream shallow
Tl Riffle shallow
Tl Right bank shallow
T1MW 5 Oft Right bank
T1MW 100 ft Right bank
T2MW 100 ft Left bank
T2MW 5 Oft Left bank
T2 Leftbank shallow
T2 Stream shallow
T2 Right bank shallow
T2 Point bar shallow
T2MW 180 ft Right bank
T3MW 110 ft Left bank
T3MW 150 ft Left bank
T3MW 5 Oft Left bank
T3 Left bank shallow
T3 Stream shallow
T3 Right bank shallow
T3MW 5 Oft Right bank
T3MW 100 ft Right bank
3/22/2002
msl (ft)
NA
216.83
215.06
215.42
214.49
215.24
214.35
214.36
217.78
217.48
217.48
216.98
217.43
217.48
216.25
219.82
220.50
NA
219.45
219.51
219.34
219.05
219.25
8/22/2002
msl (ft)
NA
215.30
NA
NA
212.12
NA
210.87
210.71
216.30
216.14
216.53
216.19
216.24
215.89
213.78
218.72
219.37
NA
218.49
218.66
218.53
217.56
217.08
9/23/2003
msl (ft)
219.93
218.96
216.97
NA
214.38
215.23
217.13
217.58
219.72
218.95
218.59
216.73
217.97
218.45
221.23
222.78
222.96
220.57
220.03
219.83
220.33
221.42
222.79
High vs. Low
msl (ft)
NA
3.66
NA
NA
2.26
NA
6.26
6.87
3.42
2.81
2.06
0.54
1.73
2.56
7.45
4.06
3.59
NA
1.54
1.17
1.80
3.86
5.71
Characterization of Near-Stream Ground Water-
Surface Water Interaction
Once the geomorphology, geology, stream and ground
water hydrology for MBR were described, the near-
stream GSI was characterized. Vertical equipotential
contours were created at each of the transects using
measured piezometer-nest water levels for several dates
to show the gradients under each of the stream transects
at a scale in tens of feet. The equipotentials were drawn
using an automatic contouring program for the dates
of March 22, 2002, June 24, 2002, August 22, 2002,
September 30, 2002, December 17, 2002 and September
23, 2003. These dates were selected to show seasonal
variation and capture the GSI on the ground water low
observed on August 22, 2002 and the ground water high
on September 23, 2003. The gaged stream flow on these
dates was 0.87, 0.35, 0.04, 0.12, 1.7, and 7.7 ftVsec,
respectively. For all vertical equipotential figures, the
black, solid line represents the stream bed profile at
the transects in Fall 2001. The blue line represents
the ground water level surface based on the shallow
piezometer measurements (e.g. Figure 23). Ground
water level is drawn across the stream for reference, but
it does not represent actual stream water level which is
a function of other factors besides ground water level.
Note that the automatic contouring can create artifacts
from edge effects outside the data points which should
be ignored. Flow is from high to low equipotential.
Transect 1 was a deep incised pool with a slow base
flow velocity and high conductivity. These features
would suggest it would experience good exchange
with the ground water system. Figures 23 to 28 show
vertical ground water equipotentials under Transect 1
for select dates. The March 22, 2002 equipotentials
indicate flow was moving from the stream into the bed
and banks, demonstrating that the stream was losing
water. The situation was the same for June 24, 2002 but
-------
the gradients were greater. These data suggest that water
level in the pool was sufficient relative to the ground
water to allow stream water to move continuously
into the stream bed and banks. Near the height of the
drought on August 22, 2002, the pool dried up and
the water levels in the piezometers dropped below the
shallow stream bed and bank piezometers. Flow then
moved under the bed, down gradient, and parallel to the
stream as part of the larger regional ground water flow
system. Therefore, ground water was not interacting
with surface water at this time. On September 30, 2002,
ground water levels had risen above the stream and there
was once more interaction between ground and surface
water. Consequently, stream flow moved downward and
outward to the left bank. During the ground water high
on September 23, 2003, Transect 1 flowed from the left
bank under the stream to the right bank. The stream bed
shallow piezometer was lost in a storm and could not
be used to create these contours. Although the ground
water levels had risen, the stream depth in the pool
was apparently sufficient to prevent ground water from
discharging upward into the stream. Overall, Transect 1
was a consistently losing reach.
.Q
-20 -10 0 10
Horizontal distance relative to stream piezometer location (ft)
30
Figure 23. Transect 1 vetical cross section equipotentials on March 22, 2002.
0)
I
£
a
.ID
Q
-30
-20 -10 0 10
Horizontal distance relative to stream piezometer location (ft)
30
Figure 24. Transect 1 vertical cross section equipotentials on June 24, 2002.
-------
-I
.92
CD
f 5
Q.
co
2
-------
-20 -10 0 10
Horizontal distance relative to stream piezometer location (ft)
Figure 28. Transect 1 vertical cross section equipotentials on September 23, 2003 (study high).
Transect 2 was a shallow narrow riffle with a steep
incised left bank and a point bar on the right bank with a
high base flow velocity. Piezometers in Transect 2 were
situated in low conductivity sediments with slow hydro-
logic exchange. Figures 29-34 show the ground water
equipotentials under Transect 2 for select dates. Ground
water flow was consistently from the stream banks into
the bed, indicating that Transect 2 was a gaining reach.
The vertical gradient ranged from a low of 0.02 ft/ft dur-
ing the drought to a high of 0.1 ft/ft on Sept. 23, 2003.
Stream width was narrow and surface flow was swift so
the area for GSI was small. This geomorphology was
likely the cause of a focused upwelling of ground water
discharge at this point. Rapid flow at this reach may also
have allowed for more discharge to the stream.
e.
i
0)
IB 5-
I 0
Q.
E
•8
| -5
i
2
-25
_ Piezometer Water Level
i? / \ *7±
«/217.^ ^ ^» 'Jj
217.32
\ «
-20 -15 -10 -5 0 5 10 15
Horizontal distance relative to stream piezometer location (ft)
20
25
Figure 29. Transect 2 vertical cross section of equipotentials on March 22, 2002.
-------
§ -25
a
-15 -10 -5 0 5 10 15
Horizontal distance relative to stream piezometer location (ft)
20
25
Figure 30. Transect 2 vertical cross section of equipotentials on June 24, 2002.
g
I
J2
a>
a.
to
3
2
§ -25
_ Piezometer Water Level
^216-47
-20 -15 -10 -5 0 5 10 15
Horizontal distance relative to stream piezometer beat ion (ft)
Figure 31. Transect 2 vertical cross section of equipotentials on August 22, 2002 (study low).
216.24
216.21
20 25
§ -25
-15 -10 -5 0 5 10 15
Horizontal distance relative to stream piezometer location (ft)
20
25
Figure 32. Transect 2 vertical cross section of equipotentials on September 30, 2002.
-------
e.
.SB 0
Q.
co
2
t8
-5
. Piezometer Water Level
8 -25
-20 -15 -10 -5 0 5 10 15
Horizontal distance relative to stream piezometer location (ft)
20
25
Figure 33. Transect 2 vertical cross section of equipotentials on December 17, 2002.
-25
-15 -10 -5 0 5 10 15
Horizontal distance relative to stream piezometer location (ft)
20
25
Figure 34. Transect 2 vertical cross section of equipotentials on September 23, 2003 (study high).
Transect 3 was a wide shallow stream terrace with low
stream velocity and high conductivity bed sediments.
Figures 35-40 show the ground water equipotentials
under Transect 3 for the chosen dates. Stream water
consistently moved from the stream into the beds and
banks at low gradients ranging from 0.005 to 0.02 ft/ft,
indicating that Transect 3 was a losing reach. On
December 17, 2002, however, the direction of flow
reversed and began moving into the stream. During
the ground water high level on September 23, 2003,
this reversal was complete, and the transect became
a gaining reach with a gradient of 0.022 ft/ft. The
vertical gradients in Transects 1 and 2 were of similar
magnitude, but gradients at Transect 3 were about an
order of magnitude less, likely a consequence of the
conductive stream bed lithology. The bed and banks at
Transects 1 and 2 were composed of the same gravelly
clay. Transect 3, however, had more sandy sediments
and a higher conductivity as shown in Table 1, indicating
that gradients may be lower. This high conductivity
may have allowed the flow direction to be more easily
reversed over a short time frame.
-------
. Piezometer Water Level
I
e
8
b
-20 -10 0 10
Horizontal distance relative to stream piezometer location (ft)
20
30
Figure 35. Transect 3 vertical cross section of equipotentials on March 22, 2002.
e.
I
S
J2
0)
"8
Piezometer Water Level
0 ^
Q.
(0
1
-20 -10 0 10
Horizontal distance relative to stream piezometer location (ft)
20
30
Figure 36. Transect 3 vertical cross section of equipotentials on June 24, 2002.
P Piezometer Water Level
-20 -10 0 10
Horizontal distance relative to stream piezometer location (ft)
20
30
Figure 37. Transect 3 vertical cross section of equipotentials on August 22, 2002 (study low).
-------
• Piezometer Water Level
-20 -10 0 10
Horizontal distance relative to stream piezometer location (ft)
20
30
Figure 38. Transect 3 vertical cross section of equipotentials on September 30, 2002.
g
I
S
-------
Using the gradients and Darcy's law as described in
Methods, near-stream GSI flow was calculated for
all transects on the selected dates. Flow results for
Transect 1 are calculated based on a five-foot wide by
five-foot deep compartment (Table 5). In the incised
pool on Transect 1, water consistently moved into the
ground water system at a high rate due to conductive
sediments and high vertical gradients, indicating that
this reach was a losing section of the stream. On
March 22, 2002, a volumetric flow rate, Q, of 6.32 and
8.06 ft3/day/ft2 moved into the ground water system on
the right and left banks, respectively. At this rate, the
ground water residence time, to cover a five foot path, s,
was 0.32 and 0.25 days, respectively. On June 24, 2002,
the rate was substantially higher with 35.56 ft3/day/ft2
entering the ground water system on the right bank and
37.17 ft3/day/ft2 entering the left bank. Higher rates
were a consequence of the encroaching drought which
lowered ground water levels below the stream bed. By
August 22, 2002, the drought was so severe that no
stream flow was present at Transect 1 and therefore
no interaction occurred between ground and surface
water. All surface water was moving into ground water
just upstream of Transect 2. The GSI returned to a rate
of 1.55 and 4.46 ftVday/ft2 to the ground water for the
right and left banks by December 2002. GSI at Transect
1 could not be calculated on September 23, 2003,
the ground water high because the shallow stream
piezometer was lost in a storm in June 2003. Overall,
the data demonstrated Transect 1 was in a losing reach
with high and variable volumetric flow rates and low
residence times in the stream bed sediments.
Table 5. Ground water (GW) and surface water (SW) flow calculations for Transect 1
Date
3/22/2002
6/24/2002
8/22/2002
9/30/2002
12/17/2002
9/23/2003
3/22/2002
6/24/2002
8/22/2002
9/30/2002
12/17/2002
9/23/2003
Compartment
Transect 1 Right Bank
Transect 1 Right Bank
Transect 1 Right Bank
Transect 1 Right Bank
Transect 1 Right Bank
Transect 1 Right Bank
Transect 1 Left Bank
Transect 1 Left Bank
Transect 1 Left Bank
Transect 1 Left Bank
Transect 1 Left Bank
Transect 1 Left Bank
dh/dz*
0.020
0.114
dh/dx**
0.005
0.008
dh/ds
0.020
0.115
K (ft/day)
310.00
310.00
q (ft/day)
6.32
35.56
A (ft2)
1.00
1.00
NO CONNECTION
0.035
0.005
0.020
0.114
0.035
0.005
0.000
0.000
0.035
0.005
310.00
310.00
10.79
1.55
1.00
1.00
SHALLOW PIEZOMETER DESTROYED
0.021
0.036
0.033
0.014
0.026
0.120
310.00
310.00
NO CONNECTION
0.048
0.014
310.00
310.00
8.06
37.17
14.82
4.46
1.00
1.00
1.00
1.00
SHALLOW PIEZOMETER DESTROYED
Q( ft'/d)
6.32
35.56
10.79
1.55
8.06
37.17
14.82
4.46
s(ft)
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
jjftftft
0.41
0.41
0.41
0.41
0.41
0.41
0.41
0.41
t (days)
0.32
0.06
0.19
1.32
0.25
0.06
0.14
0.46
Direction
SWtoGW
SW to GW
SW to GW
SWtoGW
SWtoGW
SWtoGW
SWtoGW
SW to GW
^Positive flow from stream: Negative flow to stream
^Positive flow out from stream: Negative flow in toward stream
:*Calculated from soil core bulk density
The GSI flow results for Transect 2 were calculated
based on a five-foot wide by five-foot deep compartment
(Table 6). In this high velocity riffle with an incised
left bank and a point bar on the right bank, flow
moved consistently from the ground water into the
stream. The low rates were a consequence of the less
conductive sediments and lower vertical gradients. On
March 22, 2002, a volumetric flow rate, Q, of 0.07 and
0.10 ft3/day/ft2 moved into the stream from the ground
water on the right and left banks, respectively. At this
rate, residence time to cover a five foot path was 21.03
and 15.17 days for the right and left banks, respectively.
On June 24, 2002 when flow rates were substantially
higher, 0.52 ft3/day/ft2 entered the ground water system
on the right bank and 0.51 ft3/day/ft2 entered the left
bank. By August 22, 2002, the drought was so severe
that no stream flow was present at Transect 2 and
therefore no GSI. All surface water moved into ground
water just upstream of Transect 2. The GSI remained in
this range through the rest of 2002 but returned to 0.90
and 0.91 ft3/day/ft2 on September 23, 2003, the date of
the ground water level high. Overall, the data indicated
that Transect 2 was a consistently gaining reach with low
and relatively consistent volumetric flow rates and high
residence times in the stream bed sediments.
-------
Table 6. Ground water (GW) and surface water (SW) flow calculations for Transect 2
Date
3/22/2002
6/24/2002
8/22/2002
9/30/2002
12/17/2002
9/23/2003
3/22/2002
6/24/2002
8/22/2002
9/30/2002
12/17/2002
9/23/2003
Compartment
Transect 2 Right Bank
Transect 2 Right Bank
Transect 2 Right Bank
Transect 2 Right Bank
Transect 2 Right Bank
Transect 2 Right Bank
Transect 2 Left Bank
Transect 2 Left Bank
Transect 2 Left Bank
Transect 2 Left Bank
Transect 2 Left Bank
Transect 2 Left Bank
dh/dz*
-0.010
-0.131
-0.091
-0.167
-0.227
-0.010
-0.131
-0.091
-0.167
-0.227
dh/dx**
-0.016
-0.025
-0.023
-0.028
-0.044
-0.024
-0.019
-0.0222
-0.0318
-0.0564
dh/ds
0.019
0.133
0.094
0.169
0.231
0.026
0.132
0.0935
0.169
0.234
K (ft/day)
3.90
3.90
q (ft/day)
0.07
0.52
A (ft2)
1.00
1.00
NO CONNECTION
3.90
3.90
3.90
3.90
3.90
0.37
0.66
0.90
0.10
0.51
1.00
1.00
2.00
1.00
1.00
NO CONNECTION
3.90
3.90
3.90
0.36
0.66
0.91
1.00
1.00
1.00
Q (ftVd)
0.07
0.52
0.37
0.66
1.80
0.10
0.51
0.36
0.66
0.91
s(ft)
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
5.00
jjftftft
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
0.31
t (days)
21.03
2.99
4.25
2.35
1.72
15.17
3.01
4.25
2.35
1.70
Direction
GW to SW
GWtoSW
GWtoSW
GW to SW
GWtoSW
GW to SW
GW to SW
GWtoSW
GWtoSW
GW to SW
^Positive flow from stream. Negative flow to stream
:*Positive flow out from stream, Negative flow in toward stream
^Calculated from soil core bulk density
The GSI flow results for Transect 3 were calculated
based on a five-foot wide by five-foot deep compartment
(Table 7). Results indicated a generally losing reach at
this wide, shallow terrace where water moved slowly
from the stream into the ground water system despite
sediments that were conductive. On March 22, 2002,
a volumetric flow rate, Q, of around 0.61 ftVday/ft2
passed to the stream from the ground water on the right
and left banks, respectively. At this rate, the residence
time, to cover a five-foot path was about 2.4 days.
On June 24, 2002, the rate was substantially lower;
0.15 ft3/day/ft2 entered the ground water system on the
right bank and 0.02 ft3/day/ft2 entered the left bank. On
August 22, 2002, the drought was severe and the stream
lost more surface water to the ground water system. The
stream started to recover and, by December 2002, ground
water and surface water flowed slowly downstream
parallel to the stream bed with little interaction. During
the ground water high on September 23, 2003, the GSI
had reversed direction and Transect 3 became a gaining
reach, receiving 1.34 fWday/ft2 from the left bank and
0.71 fWday/ft2 from the right bank. Overall, Transect 3
was a consistently losing reach with long residence times
which experienced a flow reversal to a gaining reach in
the wet year of 2003.
Table 7. Ground water (GW) and surface water (SW) flow calculations for Transect 3
Date
3/22/2002
6/24/2002
8/22/2002
9/30/2002
12/17/2002
9/23/2003
Compartment
Transect 3 Right Bank
Transect 3 Right Bank
Transect 3 Right Bank
Transect 3 Right Bank
Transect 3 Right Bank
Transect 3 Right Bank
dh/dz*
0.020
0.005
0.020
0.010
0.000
-0.005
dh/dx**
0.003
0.002
0.002
0.004
0.002
-0.022
dh/ds
0.020
0.005
0.019
0.010
0.002
0.022
K (ft/day)
30.50
30.50
30.50
30.50
30.50
30.50
q (ft/day)
0.61
0.15
0.58
0.31
0.06
0.67
A (ft2)
1.00
1.00
1.00
1.00
1.00
2.00
Q (ft'/d)
0.61
0.15
0.58
0.31
0.06
1.34
s(ft)
5.00
5.00
5.00
5.00
5.00
5.00
n***
0.29
0.29
0.29
0.29
0.29
0.29
t (days)
2.37
9.51
2.50
4.75
23.77
2.16
Direction
SWtoGW
SWtoGW
SW to GW
SW to GW
parallel to bed
GW to SW
3/22/2002
6/24/2002
8/22/2002
9/30/2002
12/17/2002
9/23/2003
Transect 3 Left Bank
Transect 3 Left Bank
Transect 3 Left Bank
Transect 3 Left Bank
Transect 3 Left Bank
Transect 3 Left Bank
0.020
0.005
0.020
0.010
0.000
-0.005
-0.001
-0.001
0.003
0.0008
0.000
-0.011
0.020
0.001
0.020
0.0099
0.000
0.012
30.50
30.50
30.50
30.50
30.50
30.50
0.61
0.02
0.61
0.30
0.00
0.36
1.00
1.00
1.00
1.00
1.00
2.00
0.61
0.02
0.61
0.30
0.00
0.71
5.00
5.00
5.00
5.00
5.00
5.00
0.29
0.29
0.29
0.29
0.29
0.29
2.39
95.08
2.38
4.80
NA
4.06
SWtoGW
SWtoGW
SW to GW
SW to GW
parallel to bed
GW to SW
""Positive flow from stream; Negative flow to stream
:*Positive flow out from stream; Negative flow in toward stream
'"""Calculated from soil core bulk density
-------
Temperature verification of ground water-
surface water interaction
Transect 1 exhibited strongly losing behavior. This
movement of surface water to the ground water was
corroborated by the continuous stream bed ground
water temperatures measured in the data loggers around
March 22, 2002 at both the shallow and medium depths
(Figure 41). These temperatures clearly exhibited the
expected diurnal temperature variation of the stream
surface water as it moved relatively quickly into the
highly conductive bed sediments at Transect 1. This
diurnal temperature swing in the stream bed piezometers,
which varied by several degrees Celsius, strongly
supported the conclusion that this deeply incised pool in
conductive sediments was consistently losing.
Date and Time
Figure 41. Transect 1 continuous stream bed piezometer
temperatures.
Vertical equipotentials mapped over time showed
Transect 2 to be a consistently gaining reach. Data
from the stream bed piezometer in Transect 2 on the
days around March 22, 2002 demonstrated relatively
consistent temperatures at the medium depth as ground
water discharged to the stream (Figure 42). The
temperature in the stream bed piezometers displayed
steady ground water temperature unaffected by the
diurnal variation in the stream water temperature. The
mild variation in the shallow piezometer temperature
over 0.5 degree Celsius range was most likely a function
of the sediment temperature and not stream water
movement. If stream water was moving into the bed, as
it did at Transect 1, the temperature would likely show a
diurnal variation of several degrees.
> T2 Medium Depth Piezometer •
T2 Shallow Depth Piezometer
Date and Time
Figure 42. Transect 2 continuous stream bed piezometer
temperatures.
Vertical equipotentials showed Transect 3 to be a
slowly losing reach with parallel flow under the bed.
Consequently, temperature in the stream bed piezometers
would be expected to display the diurnal variation of
the stream surface temperature. However, continuous
stream bed temperatures for Transect 3 for the days
around March 22, 2002 indicated that temperature in
the shallow piezometer varied diurnally but that the
medium-depth piezometer did not (Figure 43). This
temperature signature was closer to that of Transect 2, a
gaining reach, than that of the losing reach at Transect 1.
This may have been a consequence of the very slow flow
into the bed which was too small to strongly reflect the
diurnal variation of the stream surface temperature.
Figure 43. Transect 3 continuous stream bed piezometer
temperatures.
-------
Stream flow verification of ground water-
surface water interaction
The classification of the specific transects as losing
or gaining from the above near-stream GSI flow
calculations and temperature analysis were also verified
by examining the stream flow measurements at each
transect. By comparing these values, the flow loss
and gain between transects can be roughly assessed.
However, this approach may only be used if no
tributaries or springs are along the study reach which
was the case.
Table 8 shows measurements of surface stream flow
that were made at the USGS stream flow gage above
Transect 3, Transect 3, Transect 2, Transect 1, and just
below Transect 1 over time. Nearly all measurements
indicated loss of stream flow between the stream
gage and just below Transect 1, a finding supported
by the equipotential analysis. Table 8 shows little
difference in stream flow between Transects 3 and
2, corresponding to the low GSI calculated at these
transects. However, a consistent loss of flow between
Transects 2 and Transect 1 and between Transect 1 and
below corresponds to the GSI calculations which show
Transect 1 to be strongly losing.
Table 8. Stream flow at each of the transects on specific dates
Date
1/15/2002
3/05/2002
5/20/2002
7/15/2002
9/03/2002
11/07/2002
1/13/2003
3/10/2003
5/12/2003
7/07/2003
9/02/2003
1 1 700/9004
Q-Gage
(cfs)
0.385
0.489
0.726
0.220
0.166
0.691
3.40
1.62
2.68
1.16
0 809*
Q-Transect 3
(cfs)
0.396*
0.489*
0.711*
0.208*
0.179
0.643*
1.45*
3.15*
1.37*
2.53
1.04*
0 809*
Q-Transect 2
(cfs)
0.396*
0.489*
0.711*
0.208*
0.172
0.643*
1.45*
3.15*
1.37*
2.61
1.04*
Q-Transect 1
(cfs)
0.357
0.499
0.626
0.133
0.141
0.622
1.44
3.89
1.39
2.59
0.95
Q-Below Transect 1
(cfs)
0.226
0.397
0.478
0.073
0.014
0.579
1.38
3.59
1.34
2.29
0.83
*—One discharge measurement was made that approximately represents both locations.
Storm surge ground water-surface water
interaction
From November 2001 to July 2004, Minebank Run
stream experienced eighteen large storm discharges
in response to major precipitation events (Doheny et
al., 2006). Instantaneous peak discharges during these
events ranged from 247-1390 ftVsec compared to a
mean annual discharge of 1.15 ftVs-4.34 ftVs (Doheny
et al., 2006). All of these events were captured in real
time by the continuous data loggers installed in the
piezometer nests and monitoring wells.
A storm on August 3, 2002 with a duration of 1.0 hour
produced 1.18 inches of rainfall (Doheny et al., 2006).
The peak gage height measured 7.58 feet and discharge
in the stream increased from 0.05 ftVsec (drought
conditions) to an instantaneous peak discharge of 725
ftVsec (Doheny et al., 2006). The stream gage response
is shown in Figure 44. Remarkably, the storm surge
peaked and disappeared within about two hours.
-------
Date and Time
Figure 44. Stream discharge hydrograph in response to
storm event on August 3, 2002.
The shallow and medium depth piezometers in the
stream bed exhibited the expected rise in water level
with the storm surge (Figure 45). Strikingly, the rise and
fall of the hydrograph matched the timing of the stream
discharge almost exactly and covered some six feet of
water level fluctuation. The stream bed at Transect 1
had a high conductivity of 310 ft/day, which enabled
this almost instantaneous interaction. The right bank
piezometer had a high conductivity of 170 ft/day and
also reflected a similar but reduced and delayed rise and
fall in water level as a consequence of the head loss and
time lag as water moved through the sediments. The
piezometer in the left bank, positioned in sediments with
much lower hydraulic conductivity (7 ft/day), responded
with a lower and delayed peak.
Figure 45. Transect 1 piezometer hydrograph response
to storm event on August 3, 2002.
The response of the stream bed and bank piezometers at
Transect 2 to the August 2002 event is shown in Figure
46. The conductivity in the stream bed at Transect 2 was
3.9 ft/day, about two orders of magnitude less than at
Transect 1. Though the stream was gaining at Transect
2, the response of the stream bed piezometers was
similar to Transect 1. The height of the storm surge at
this transect was unknown but was likely less than the
stream gage peak height because this transect was
situated at a point bar which would have allowed over
bank flow, thus reducing the height of the storm surge.
The shallow depth stream bed piezometer data logger
failed so no data were collected. The right bank
piezometer located in the point bar near the stream
reflected about the same rise as the stream bed
piezometer. The left bank piezometer was located about
thirty feet from the stream bed piezometer and showed a
much smaller increase in water level because of the large
head loss through the gravelly clay sediments.
y y y y y y y
Date and Time
Figure 46. Transect 2 piezometer hydrograph response
to storm event on August 3, 2002.
The response of the stream bed and bank piezometers at
Transect 3 to the August 2002 event is shown in Figure
47. This transect was a shallow wide terrace with an
average stream bed conductivity of 30.5 ft/day. The
near-stream GSI equipotential analysis showed that
Transect 3 was a slowly losing reach. As the storm
surge passed through, water level in the stream bed
piezometers rose almost four feet. The medium depth
piezometer exhibited less of a rise, reflecting the head
loss as the water was pushed into the bed sediments.
The left bank stream piezometer was in conductive sands
and thus, showed a similar rise, muted by the distance
from the stream bed. The right bank piezometer,
although in similarly conductivity sediments, showed an
unexpected muted response.
-------
Figure 47. Transect 3 piezometer hydrograph response
to storm event on August 3, 2002.
The August 3, 2002 storm surge clearly drove water
rapidly into the stream beds and banks as shown by the
water levels. Despite its limitations for this unsteady,
turbulent flow system discussed in the methods section,
Darcy's law was used as a first estimate of the temporal
variation in GSI in the stream bed at Transect 1 on the
August 3, 2002 storm event. Figure 48 displays the flow
rates calculated using Darcy's law and the gradient in the
stream bed as a function of time during the storm surge.
The results reveal a flow that was strongly driven down
into the stream bed (positive), peaking at the height of
the storm surge, falling off, and finally reversing back
into in the stream (negative). This event demonstrates
how rapidly the storm surge affects the GSI in the stream
bed. Such events may significantly influence stream
function but have not typically been quantified in GSI
evaluations.
W3/02 d/3/02 d/3/02 d/3/02 d/3/02 d/3/02 d/3/02 d/3/02 d/3/02 d/3/02 8/3/02
0:00 0:15 0:30 0:45 1:00 1:15 1:30 1:45 2:00 2:15 2:30
Data and Tima
Figure 48. GSI flow rates at Transect 1 during storm
surge on August 3, 2002.
In addition to the change in water level, data loggers also
recorded changes in water temperature in the stream bed
piezometers during and after the storm event. Figure 49
shows the temperature response at Transect 1 from
midnight August 2 to midnight August 12. Just before
the storm event, Transect 1 was a strongly losing reach
and the stream bed piezometers reflected a diurnal
variation in both the shallow and medium depth
piezometers. As the storm surge moved through, stream
water was driven into the bed. The temperature response
showed a cooling of almost 8 degrees Celsius in the
medium depth piezometer. In the days following the
storm surge, the shallow depth piezometer showed a
muted diurnal temperature response. The medium depth
stream bed piezometer reflected a constant temperature
indicating that the water driven deeper into the bed by
the storm surge was cooling to match the ground water
temperature and no new water was moving down into
the bed. At this time, the losing nature of Transect 1 was
changed to gaining as stream water driven into the bed
water was flowing back into the stream. This reversal
lasted almost 8 days after which the flow was
reestablished as losing at Transect 1 when the diurnal
pattern resumed on August 11.
Date and Time
Figure 49. Transect 1 piezometer temperature response
to storm event on August 3, 2002.
Figure 50 shows the temperature response at Transect 2
from midnight August 2 to midnight August 12. The
shallow stream bed piezometer data logger failed, so no
data were available at this depth. Just before the storm
event, Transect 2 was a consistently gaining reach. The
stream bed and the nearby right bank piezometer in
the point bar reflected a corresponding ground water
temperature pattern. As the storm passed, a surge of
cool stream water was driven into the bed. Unlike the
response at Transect 1, the medium depth stream bed
piezometer showed a short dip in temperature at the
time of the storm surge and then quickly returned to the
-------
steady ground water temperature signature expected
for a gaining reach. The right bank piezometer did not
reflect this dip and the left bank piezometer showed no
response with the stream temperature during the entire
period. This limited response was likely a function of
low hydraulic conductivity at this transect (3.9 ft/day)
which precluded a large amount of water from moving
into the stream bed and banks.
Date and Time
Figure 50. Transect 2 piezometer temperature response
to storm event on August 3, 2002.
Date and Time
Figure 51. Transect 3 piezometer temperature response
to storm event on August 3, 2002.
Figure 51 shows the temperature response at Transect 3
from midnight August 2 to midnight August 12. Just
before the storm event, Transect 3 was a weakly
losing reach. The shallow piezometer showed the
diurnal variation of the stream temperature, while the
medium-depth piezometer reflected the steady ground
water temperature. As the storm passed, stream water
was driven into the bed and the temperature response
showed a cooling of about three degrees Celsius in the
shallow piezometer and one degree in the medium-depth
piezometer, a pattern similar to Transect 1, but reflecting
the lower conductivity of transect 3. The temperature
also did not return to a pre-storm temperature as did
Transect 1. The right bank and left bank piezometers
showed no discernable response to the event.
-------
-------
6.0
Discussion
The first objective of this pre-restoration study was to
assess the stream geomorphology and geology at specific
stream features and determine their influence on GSI
at MBR. Our results strongly demonstrated the impact
of geology and geomorphology on GSI. For example,
Transect 1 was located at a deep incised pool in highly
conductive sediments, creating a setting of low velocity
flow and high hydrostatic gradient between the stream
and the surrounding ground water. Consequently, the
stream was consistently losing. Transect 2 was located
at a narrow riffle with high velocity in low conductivity
sediments where the reach was consistently gaining at
a low rate due to the swift flow in a low conductivity
stream bed, creating a hydraulic gradient conducive to
flow into the stream. Transect 3 was located in a shallow,
wide terrace with slow velocity in conductive sediments,
a feature creating a mildly losing reach that reversed to
gaining when the ground water levels were high. The
tremendous variation in GSI at each stream feature
clearly demonstrated the impact of geomorphology and
geology.
The second objective of the pre-restoration study was to
develop effective and simple methods to characterize,
measure, and quantify near-stream GSI at specific stream
features. Our results demonstrated that using a network
of monitoring wells and piezometer nests to measure
water levels at defined time intervals was sufficient to
characterize and quantify the GSI at MBR at both the
regional and near-stream scale.
Ground water equipotentials defined by the bi-weekly
water levels at all of the site wells were used to assess the
ground water gradients at the regional watershed scale
at MBR. Our results demonstrated that ground water
flow varied greatly over space and time at the regional
scale at MBR. Given this variability, it was important to
be aware that numerous factors influence ground water
gradients and the interaction of the ground water with
the stream on a regional scale. The system was never
in a steady state and frequent sampling of ground water
levels and stream flow were necessary to characterize
hydrology.
The regional ground water horizontal equipotentials
were not very useful for interpreting the near-stream GSI
because they did not reveal the impact of the stream on
the ground water flow system. This was not unexpected,
as the majority of groundwater discharges at depth
from the MBR watershed and only a small amount is
intercepted by the stream as base flow. For MBR, the
regional scale should not be used to describe near-stream
GSI as it can not capture the spatial and temporal scale
at which this interaction occurs. The regional scale
horizontal equipotentials also do not address the vertical
gradients which define the exchange in and out of the
stream beds and banks at the near-stream scale.
The near-stream GSI at MBR was easily characterized
and quantified using the vertical equipotentials developed
from the stream bed and bank piezometer water levels.
On the near- stream scale, the losing or gaining nature
of the stream features at each transect could be seen on
these equipotentials. Equipotentials also showed the
near-stream GSI to be highly spatially and temporally
variable at each feature. Consequently, many of the
typical methods used to quantify GSI at this scale, such
as tracer tests or numerical heat and flow modeling would
be impractical or fail to capture this variation.
Our results showed that a simple model which split
the stream into two compartments at the thalweg was
sufficient to quantify the GSI flow at each stream
feature. The horizontal and vertical gradients for each
compartment were calculated using the water levels in
the stream bed piezometers and banks at each sampling
time. These gradients were then substituted into the 2-D
vector form of Darcy's law to determine the magnitude
and direction of the seepage vector along the resultant
flow path through the compartment on either side of the
stream at each transect. The residence time of the water
to pass through a five foot deep compartment was then
determined. This simple approach allowed the spatial
and temporal variation in the near-stream GSI to be
quantified biweekly using easily obtained ground water
level measurements. The results showed Transect 1 was
consistently losing but the magnitude of GSI was highly
variable, ranging from 1.55 to 37.17 ftVday/ft2 with a
residence time of 1.32 to 0.06 days over the two year
study period. Transect 2 was in a weakly gaining reach
with GSI flow ranging from 0.07 to 1.80 ftVday/ft2. The
residence time through a five foot depth was from 1.70
to 21.03 days. Transect 3 was in a losing reach that
reversed to gaining during a period of high ground water
levels. As a losing reach, GSI flow at Transect 3 reached
a maximum of 0.61 ftVday/ft2 with a residence time of
2.37 days. As a gaining reach, GSI flow at Transect 3
reached a maximum gain of 1.34 fWday/ft2 with a
residence time of 2.16 days to pass through five feet. The
alternating losing and gaining behavior was verified by
stream flow measurements and continuous temperature
data at each transect.
-------
The impact of a storm surge on GSI was also evaluated
at the transects using water level and temperature
readings collected at 5 minute intervals in the stream bed
and bank piezometers using continuous data loggers.
Our results showed that the storm surge drove water
into the stream bed at each feature to various degrees
depending on geomorphology and geology. GSI at
Transect 1 during the storm surge was estimated using
Darcy's law and showed flow entering the bed and then
reversing rapidly. Temperature data also confirmed the
impact of the storm surge on GSI at each transect and
helped assess the residence time of the surge in the bed.
The direction of the near-stream GSI at each of the
features is critical to the issue of water quality. If the
flow direction is consistently from the ground water to
the stream, as found at Transect 2, the water quality in
the stream is being impacted by ground water quality.
When no stream water enters the bed or banks, little
processing of stream contaminants in the sediments
to improve stream water quality occurs. If flow is
consistently from the stream into the stream bed and
banks, as at Transect 1 and Transect 3, beneficial
processing such as denitrification and adsorption may
occur (Kaushal et al. 2008). The magnitude of the
near-stream GSI flow at each feature is also critical.
High flows in high conductivity settings move more
fluid through the system as seen at Transect 1. Slower
flows, like at Transect 2 and 3, move less fluid but
increase residence time. Increasing residence time is
critical for increasing the mass removal of nutrients
and contaminants via biological processes and can only
be determined by knowing the GSI flow direction and
magnitude (Kaushal et al. 2008).
From the assessment of near-stream GSI, it is clear
that geomorphology and geological setting are major
factors that impact the magnitude and direction of
near-steam GSI. These factors will be greatly impacted
by restoration efforts which can change both the
geomorphology of the stream channel and the underlying
sediments. Restoration that changes geology and
geomorphology will alter GSI. GSI, in turn, influences
biological processes like nitrification and denitrification
which control the flux of nitrogen in water and,
therefore, dictate water quality. Therefore, the potential
exists to identify and select restoration techniques that
enhance GSI and improve water quality.
-------
7.0
Conclusions
1. Stream channel geomorphology and geologic
setting, especially stream channel width, depth, and
sediment lithology strongly influenced near-stream
GSI at MBR.
2. The regional ground water horizontal equipotentials
developed from bi-weekly ground water level
measurements at MBR varied substantially spatially
and temporally and did not provide the required
resolution for interpreting the near-stream GSI.
3. The characterization of near-stream GSI at MBR
stream features as losing or gaining was easily
demonstrated using cross-sections of vertical
equipotentials developed from water levels
measured in piezometer well nests in the stream bed
and banks.
4. Near-stream GSI flow rates were simply and
effectively quantified at MBR stream features using
a simple hydrologic analysis of stream bed and bank
piezometer nest water levels measured bi-weekly.
5. The direction and magnitude of near stream GSI at
MBR varied spatially and temporally among and
within stream features. Some features changed from
losing to gaining during the study.
6. Residence times derived from a stream bed
compartment model at each stream feature at MBR
were highly variable given their dependency on
GSI flow rate. Mass removal of nutrients and
contaminants in stream bed sediments is dependent
upon residence times.
7. The spatial and temporal variability of GSI at MBR
precluded the use of traditional stream tracer tests
and modeling to characterize and quantify GSI as
they are not suited to capture the highly variable
nature of the exchange.
8. Near-stream GSI losing and gaining behavior
was verified by stream flow measurements and
continuous data logger temperature data at each of
the stream features at MBR.
9. Continuous data loggers that measure water level
and temperature were successfully employed at
MBR to evaluate the impact of a storm surge on GSI
at stream features.
10. Given the dependency of GSI at MBR on stream
features and their geomorpholigical settings,
any restoration activity impacting these features
would likely also impact the GSI. The magnitude
and direction of GSI change could be easily
characterized using water levels in piezometer nest
in the stream bed and banks measured at regular
time intervals.
-------
-------
8.0
References
Becker, M. W., T. Georgian, H. Ambrose, J. Siniscalchi,
and K. Fredrick. Estimating flow and flux of
ground water discharge using water temperature
and velocity. Journal oj'Hydrology 296: 221-233
(2004).
Bencala, K. E., V.C. Kennedy, G.W. Zellweger, A.P.
Jackman, and RJ. Avanzino. Interactions of
solutes and stream bed sediment, 1, An experimental
analysis of cation and anion transport in a mountain
stream. Water Resources Research 20 (12): 1797-
1804 (1984).
Bencala, K.E. Interactions of solutes and stream
bed sediment, 2, A dynamic analysis of coupled
hydrologic and chemical processes that determine
solute transport. Water Resources Research 20,
(12): 1804-1814 (1984).
Boulton, A. J., S. Findlay, P. Marmonier, E. Stanley,
and H. M. Valett. The functional significance of
the hyporheic zone in streams and rivers. Annual
Review of Ecological Systems 29: 59-81 (1998).
Bouwer, H. The Bouwer and Rice slug test—an update.
Ground Water 27:304-309 (1989).
Bouwer, H. and R.C. Rice. A slug test method for
determining hydraulic conductivity of unconfined
aquifers with completely or partially penetrating
wells. Water Resources Research 12:423-428
(1976).
Castro, N.M. and G.M. Hornberger. Surface-
subsurface water interactions in an alluviated
mountain stream. Water Resources Research 27:
1613-1621(1991).
Choi, J.,J.W. Harvey, and M. Conklin. Characterizing
multiple timescales of stream and storage zone
interaction that affect solute fate and transport in
streams. Water Resources Research 36 (6): 1511-
1518(2000).
Cirmo, C.P. and JJ. McDonnell. Linking the hydrologic
and biogeochemical controls of nitrogen transport
on near-stream zones of temperate forested
catchments: a review. Journal of Hydrology 199:
88-120 (1997).
Conant, B. Delineating and quantifying ground water
discharge zones using streambed temperatures.
Ground Water 42 (2): 243-257 March-April 2004.
Constantz, J. Interaction between stream temperature,
streamflow and groundwater exchanges in alpine
streams. Water Resources Research 34 (7): 1609-
1615 (1998).
Crowley, W.P, J. Reinhardt, andE.T Cleaves. Geologic
Map of Baltimore County and City. Maryland
Geologic Survey (1976).
Doheny, E.J., RJ. Starsoneck, E.A. Striz, and P. M.
Mayer. Watershed Characteristics and Pre-
Restoration Surface-Water Hydrology ofMinebank
Run, Baltimore County, Maryland, Water Years
2002-04,USGS Scientific Investigations Report
2006-5179(2006).
Doheny, E.J., RJ. Starsoneck, P.M. Mayer, and E.A.
Striz. Pre-restoration geomorphic characteristics
ofMinebank Run, Baltimore County, Maryland,
2002-04. USGS Scientific Investigations Report
2007-5127(2007).
Eshleman, K., J.S. Pollard, and A.K. O'Brien.
Interactions between groundwater and surface
water in a Virginia Coastal Plain Watershed, I.
Hydrological Flowpaths. Hydrological Processes 8:
389-410 (1994).
Falk, D.A., M.A. Palmer, and J.B. Zedler. "Foundations
of Restoration Ecology." Island Press, Washington,
DC. 2006,364pp.
Federal Interagency Stream Restoration Working
Group (FISRWG), "Stream Corridor Restoration:
Principles, Processes, and Practices," GPO Item
No. 0120-A;SuDocsNo. A57.6/2:EN 3/PT.653.
ISBN-0-934213-59-3, October 1998.
Fryar, A.E., EJ. Wallin, and D.L. Brown. Spatial
and temporal variability in seepage between a
contaminated aquifer and tributaries to the Ohio
River. Ground Water Monitoring and Remediation
129-146, Summer 2000.
Gooseff, M.N., J. K. Andersen, S.M. Wondzell, J.
LaNier, and R. Haggerty. A modeling study of
hyporheic exchange pattern and the sequence, size
and spacing of stream bedforms in the mountain
stream networks, Oregon, USA. Hydrological
Processes 20: 2443-2457 (2006).
Harvey, J. W. and K.E. Bencala. The effect of streambed
topography on surface-subsurface water interaction
in mountain catchments. Water Resources Research
29 (1) 89-98 (1993).
Harvey, J.W, BJ. Wagner, and K.E. Bencala.
Evaluating the reliability of the stream tracer
-------
approach to characterize stream-subsurface water
exchange. Water Resources Research 32 (8): 2441-
2451 (1996).
Harvey, J.W. and BJ. Wagner. "Quantifying hydrologic
interactions between streams and their subsurface
hyporheic zones." Streams and Ground Waters,
Jones, J.AandMulholland, P. J. eds, Academic
Press, San Diego, 2000, 425 pp.
Hayashi, M. and D.O. Rosenberry. Effects of
groundwater exchange on hydrology and ecology
of surface water. Ground Water 40 (3): 309-316
(2002).
Kalbus, E., F. Reinstorf, and M. Schirmer. Measuring
methods for groundwater-surface water interactions:
a review. Hydrology and Earth System Sciences
10: 873-887 (2006).
Kasahara, T. and A.H. Hill. Effects of riffle-step
restoration on hyporheic zone chemistry in N-rich
lowland streams. Canadian Journal of Fisheries
andAquatic Sciences 63: 120-133 (2006).
Kaushal, S., P.M. Groffman, P.M. Mayer, E. Striz, and A.
Gold. Effects of stream restoration on denitrification
in an urbanizing watershed. Ecological Applications
18:789-804 (2008).
Poole, G.C., J.A. Stanford, S.W. Running, and C.A.
Frissell. Multiscale geomorphic drivers of
groundwater flow paths: subsurface hydrologic
dynamics and hyporheic habitat diversity. Journal
of North American Benthological Society 25 (2):
288-303 (2006).
Rosgen, D., "Applied River Morphology," Wildland
Hydrology, Pagosa Springs, CO, 1996.
Savant, S.A., D.D. Reible, and L.J. Thiobedeaux.
Convective transport within stable river sediments.
Water Resources Research 23: 1763-1768 (1987).
Silliman, S.E., J. Ramirez, and R.L. McCabe.
Quantifying downflow through creek sediments
using temperature time-series: one dimensional
solution incorporating measured surface
temperature. Journal of Hydrology 167:99-119
(1995).
Silliman, S.E. and D.F. Booth. Analysis of time-
series measurements of sediment temperature for
identification of gaining vs. losing portions of Juday
Creek, Indiana. Journal oj'Hydrology 46: 131-148
(1993).
Stonestrom, D. A. and J. Constantz. Heat as a Tool for
Studying Movement of Ground Water Near Streams,
USGS Circular 1260, 2003.
Stonestrom, D.A. and J. Constantz. Using Temperature
to Study Stream-Ground Water exchanges" USGSS
Geological Survey Fact Sheet 2004-3010, 2004.
Walter, R.C. and D. J. Merritts. Natural streams and the
legacy of water-powered mills. Science 319: 299-
304 (2008)
Winter, T.C., J.W. Harvey, O.L. Franke, and WM.
Alley. Ground Water and Surface Water, A Single
Resource, USGS Circular 1139, 1998.
Wroblicky, G.J., M.E. Campana, H.M. Valett, and C.N.
Dahm. Seasonal variation in surface-subsurface
water exchange and lateral hyporheic area of two
stream-aquifer systems. Water Resources Research
34 (3): 317-328 (1998).
-------
-------
<#3
:^Srr4^1^^^';^?
t p, --gSiiJ^^J* 0»-* -.-"*• -V,^-* v .Vi
•$&^$i^&r>>
. /^,^vg 1
^^"^ .ygk ik ^*v^.^^ \^^ i » "X i ' *• " -^,**"" •• j^ j n, \ ^t; .^TJ^^^B^^^B
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free
------- |