United States
Environmental Protection
Agency
Office Of Water
(WH-553)
EPA841-F-93-004
November 1992
Number 4
TMDL Case Study
Nomini Creek Watershed
Key Feature:
Project Name:
Location:
.- .- r • ',*, 'jf\ -. . •'.•:"„" ,:.-.
Scope/Size: '*';*.•••
Land Type: •-'•-
Type of Activity: "
Pollutants: ..-...;•--,.
TMDL Development:
Data Sources:
Data Mechanisms:
Monitoring Plan:
Control Measures:
Use of a geographic information system
and watershed models to identify areas of
critical nonpoint pollution.
Nomini Creek Watershed GIS Study
USEPA Region Ill/Westmoreland
-County, Virginia/Potomac River
- Small watershed, 1505 hectares
Ecoregion 65, Southeastern „
plains; -..•;.'
Agriculture ->. • •• .. .-":..
Nutrients, sediment ,
'
. .
State, locairfederal ~J :,-: ,-'
Modeling (SLOSS, PHOSPH); GIS
(VirGIS)
Yes, long-term BMP effectiveness
monitoring: ;:
BMPs :. , . V ."-...;;.
Summary: Using the Nomini Creek watershed (Figure 1),
Virginia's Department of Conservation and Recreation and its
Polytechnic and State University demonstrated how geographic
information system (GIS) technology can be used to (1) prioritize
and target waterbodies with multiple water quality concerns, and
(2) target BMPs to critical nonpoint source loading areas to meet
load allocations more effectively.
Nomjni Greek
Watershed
Chesapeake Bay
Inset
Area
FIGURE 1. Location of the Nomini Creek watershed
in Virginia , \ '
The Department of Conservation and Recreation selected the Nomini Creek watershed as an area in which to evaluate and
monitor the effectiveness of best management practices (BMPs) for the Chesapeake Bay Program. To identify the critical
phosphorus and sediment loading areas within the watershed so that BMPs could be sited effectively, the Division tested
the feasibility of integrating VkGIS, a state-run GIS, with two simple pollutant yield models (SLOSS and PHOSPH).
Because Virginia's data base was sufficiently large, VkGIS was able to provide the data requked to run the models. The
output from these models successfully identified critical areas of nonpoint source loading. BMPs were sited on these areas
and an intensive water quality monitoring program is currently in place to evaluate BMP effectiveness and to verify the
estimated pollutant loads. .
Contact:
Jf, Michael Flagg, Vkgiaia DCR-DSWC, 203 Governor Street, Suite 206, Richmond, VA 23219, phone
{804)786-3959 ; .- -"
-------
BACKGROUND
Programmatic Issues
EDITOR'S NOTE: Water quality management in the
Chesapeake Bay employs the essential elements of a
TMDL assessment. Stakeholders recognized water
quality problems, identified the causes and sources of
these problems, set an achievable target or endpoint, and
then targeted controls for specific point and nonpoint
pollution sources to decrease their contributions.
Technical efforts to address difficulties or obstacles to
effective water quality management in the Chesapeake
basin therefore have a unique place within the TMDL
Case Study series.
: la the early 1970s, water quality managers recognized
that the water quality of the Chesapeake had become
substantially degraded and prevented the Bay from fully
meeting its designated uses. Subsequent studies of the
Chesapeake Bay, sponsored by USEPA, identified
, nutrients as a primary cause of the depleted oxygen and
. other water quality problems. USEPA, the District of
Columbia, and the states of Pennsylvania, Maryland and
Virginia agreed that reducing total nutrient loads to the
Bay by 40 percent was an achievable goal and would
improve the health of the ecosystem. As the primary
source of nutrients in the basin, agricultural activities
were targeted for implementation of BMPs.
To begin to address these nonpoint source problems,
Virginia, Maryland, Pennsylvania, and the District of
Columbia have initiated watershed programs to
demonstrate and monitor the benefits of various BMPs.
The new information on BMP effectiveness will
eventually be used to encourage more widespread use of
BMPs throughout the Chesapeake drainage to achieve the
40 percent reduction in nutrient loads.
In 1985, the Virginia Department of Conservation and
Recreation, Division of Soil and Water Conservation
(DCR-DSWC or the Division) selected the Nomini Creek
watershed as an area in which to evaluate and monitor
BMP effectiveness. The watershed was chosen because
of its proximity to the Bay and because agriculture is its
predominant land use activity. The watershed is also
typical of row-cropped agricultural areas hi the Virginia
coastal plain. Agricultural loadings of nitrogen,
phosphorus, and sediment from Nomini Creek enter the
lower Potomac River, which is a main tributary to the
Chesapeake Bay.
To site BMPs effectively, the Division sought to identify
critical nutrient loading areas within the watershed. The
Virginia Geographic Information System (VirGIS), in
conjunction with simple pollutant loading models, was
proposed to accomplish this task. Although the GIS was
developed to facilitate these types of projects, no one had
used it yet for this purpose. The Division and the
Department of Agricultural Engineering at Virginia
Polytechnic Institute and State University therefore
conducted a sub-study to determine whether critical
nutrient loading areas could be properly identified when
VirGIS was linked with the models.
The Resource
The Nomini Creek watershed is located in Westmoreland
County in eastern Virginia. Its 3,719 acre watershed is
approximately 54 percent woodlands (there is a
significant amount of commercial forestry), 43 percent
croplands, and 3 percent homesteads and roads. There
are'no towns in the watershed. The primary agricultural
activity in the watershed is row cropping of com, barley,
and wheat (USDA SCS, 1979). Twenty-five farmers
cultivate the 1,646 acres of cropland in the watershed.
Thirty-three percent (546 acres) of the farmland is
farmed by the landowners/operators themselves, while
the remaining 67 percent (1100 acres) is leased for others
to farm (DCR-DSWC, 1986). One small beef cattle
operation also exists in the watershed.
Nomini Creek currently meets all of its designated uses.
Biological parameters indicate that water quality within
the watershed is good; however, preliminary water
quality sampling results indicate a high level of nutrients.
In 1986, gross soil erosion was calculated at 7,920
tons/year for the study portion of the watershed.
Approximately 1,584 tons of that sediment is delivered to
Nomini Creek, along with 10,700 pounds of phosphorus.
Eroding cropland is responsible for about 95 percent of
the total sediment and phosphorus that is delivered to the
creek (DCR-DSWC, 1986). The state, in conjunction
with the Virginia Forestry Association, currently
monitoring to determine how much forestry practices
contribute to the pollution load.
Hydrologically, the watershed contains first- and second-
order streams flowing through nearly level to gently
sloping topography. The soils are mostly fine sandy
loam to loam, with slopes that range from nearly level to
15 percent, except along stream banks where slopes
range from 15 percent to 50 percent (USDA SCS, 1979).
Approximately 40 percent of the cropland in the
watershed is classified as highly erodible.
ASSESSING AND CHARACTERIZING
PROBLEM AREAS
GIS and Modeling Tools
The data necessary for characterizing nutrient loading
patterns within the Nomini Creek watershed were
-------
available on VirGIS, a state-run GIS. A GIS is a
computerized system for storing and manipulating data
that have a spatial component (i.e., data for which
geographical location is important). VirGIS was initiated
in 1985 by the Division as a tool for developing
modeling and mapping procedures that could readily
identify land areas with nonpoint source pollution
potential.
can be modified to estimate potential soil erosion for
each hydrologically homogeneous cell in a watershed.
For Nomini Creek, a cell size of 1/9 hectare was used.
Three main equations are used in SLOSS.The first
equation computes soil loss per unit area of watershed
(A,). It is expressed as:
VirGIS is a" modular and highly interactive program that
consists of a large database coupled with nearly 500
special-purpose programs that manipulate and display
data. Basic data types, or layers, taken from 7.5-minute where
quadrangle maps, county soU, surveys, National High
Altitude Program color-infraredphotos, and U.S.
Geological Survey (USGS) digital elevation models '
include elevation, soils, land use, surface water, _;
watersheds, and county boundaries. The data are stored
primarily in raster form (i.e., as small area! units) with
each unit, or cell, representing from 1/9 to 1 hectare.
These units are joined to form data layers that cover
areas ranging from 7.3 million hectares up to 1.0.1
million hectares (the entire State of Virginia). VirGIS
can use the basic data layers to calculate "derived" data
layers (Figure 2) for input parameters required by
nonpoint source pollutant yield models.
Virginia Polytechnic and State University and the state
developed two simple nonpoint source pollutant yield
models, SLOSS and PHOSPH, to characterize the where
Nomini Creek watershed.
SLOSS is a simplified pollutant yield model designed to
estimate soil loss and sediment delivery to a stream. It is '
based on the Universal Soil Loss Equation (USLE) and
(i)
Q
P£
= maximum number of cells;
= soil credibility factor;
= topographic factor;
= land use/land cover management; and
= support practice factor.
The SLOSS model then calculates a sediment delivery.
ratio for each cell. This ratio relates the amount of
sediment lost, from a cell to the amount that will actually
be delivered to the stream channel. The delivery ratio
(DR) is expressed as:
(2)
b = land cover factor;
Lf = length of the flow path between cell i
and the channel outlet; and
Sf = slope function. ,.
Soils
Land Use
Elevation
Watersheds
Surface Waters
Boundaries
SCWD ^
Boundaries
BASE DATA LAYERS
Slope
Slope Length
Slope Length
Factor
Erodibility
Factor
Tolerance
Factor
Delivery
Factor
Water Quality
Index
Erosion
Index
DERIVED DATA LAYERS
FIGURE 2. Basic and derived data layers used in the VirGIS database
3
-------
Total sediment yield (L.) is then calculated in the
expression:
1-1
where the parameters are as previously defined.
PHOSPH is a simplified phosphorus loading model
developed to circumvent the intense data requirements of
more complex phosphorus models. The basic equation in
PHOSPH is:
(4)
l-i
where
TP, = total sediment-associated phosphorus
.delivered to the stream outlet;
Pq * average phosphorus content of the
surface soil layer for soil hi cell i;
L, s-sediment yield for each cell (eq. 3);
and
ER_ = phosphorus enrichment ratio. "
The phosphorus enrichment ratio is defined as the mass
of phosphorus in the eroded sediment per unit mass of
•phosphorus hi the surface soil layer. It is calculated hi
the equation:
-039
(5)
where CmiK and Cmln are the maximum and minimum
percent clay content of the soil in each cell.
Locating NFS Problem Areas
The first step in the modeling process was to obtain the
necessary data layers from the VirGIS system. This was
done by creating a data window for Nomini Creek to
obtain only the relevant data from the much larger
VirGIS data base. Once these data were extracted,
several VirGIS programs were used to convert the
.information into parameters that would be accepted by
the models (e.g., the topographic factor, LS, was
determined from VirGIS slope uiformation). The SLOSS
model was then used with these parameters to estimate
sediment loss. The sediment loss estimate then was
passed to the PHOSPH model, which calculated
sediment-associated phosphorus export. This process is
summarized in Figure 3.
The spatial resolution of the VirGlS data for Nomini
Creek was 1/9 hectare. In other words, each type of data
was recorded for blocks of land 1/9 hectare in area.
Consequently, individual sediment and phosphorus export
values were calculated for each of approximately 33,470
individual cells within the watershed. The numerical
results from both models were reprocessed by VirGIS
into maps that allowed easy comparison of loadings
throughout the watershed. Figure 4 shows the GIS-
generated jtnap for phosphorus yield in the watershed.
It was determined that estimated sediriient yield exceeded
the "high" threshold of 22.4 tons/hectare/year in
approximately 15 percent (227 hectares) of the Nomini
Creek watershed. Due to the lack of established values
hi the literature, the sediment threshold was arbitrarily
Bute Data
Layers
GIS
Derived Input
Parameters
Etevalkxi
Cover
Pollutant
Yield Maps
Lty
i '
1
J
^
— >•
"K", "R", "LS"...
— >
NPS Models
Sediment Loading
FIGURE 3. Conceptual framework for the integration of a GIS with water quality models
-------
; I?B/ACIYR
10.00 - 200.00
> 200.00
Seal* 1:30000 C«ISiz«. I/SNA(057AC)
FIGURE 4. Phosphorus yield in the Nomini Creek
watershed
determined,in order to/facilitate classification into high,
medium, or low categories. For phosphorus,
approximately 21 percent of the watershed (316 hectares)
exceeded the high threshold value of 1.12
kg/hectare/year. The phosphorus threshold was obtained
from extensive literature review.
SITING NFS CONTROLS
The critical area maps generated by VirGIS assisted in
siting BMPs where they were needed in the watershed.
The BMPs included no-till fanning, nutrient management
plans, grassed waterways, drop structures, diversions,
pasture management, and the removal of land from
production. The installation of the BMPs was
accomplished through the Virginia Agricultural BMP
Cost-Sharing Program, in cooperation with the U.S.
Department of Agriculture's Soil Conservation Service •
and Agricultural Stabilization arid Conservation Service,
the Virginia Department of Forestry, the Cooperative
Extension Service, and the Northern Neck Soil and
Water Conservation District.
FOLLOW-UP MONITORING
Intensive water quality monitoring, which is the primary
focus of the Division's efforts in the watershed, is
continuing and will be used to verify the estimated
loadings of sediment and phosphorus. This monitoring
includes both storm event and ambient monitoring. Storm
sampling is conducted whenever the stream level
increases by 2/10 of a fop.t Storm event parameters
include total nitrogen, total phosphorus, nitrate, ammonia,
phosphate, and total suspended solids. Ambient
monitoring occurs on a monthly basis for protozoans,
pesticides, and a suite of weather parameters. Colifprm
bacteria are sampled every 2 weeks, and land use data
are collected twice per year. Monthly groundwater
monitoring is also conducted for pesticides, nutrients, and
water table depth.
MANAGEMENT CONSIDERATIONS
The maps produced by the modeling process are
especially valuable as watershed management tools.
First, they highlight the portions of a watershed that are
critical hi terms of their sediment and;phosphofus loading
potential. (The critical loading-level can be ,set by the
user based on literature review or professional judgment)
This feature provides managers with a readily
understandable means for determining areas that are in
need of control measures. For states, this information
can assist in targeting areas to receive BMP cost-share
. funding, potentially increasing the cost-effectiveness'of"
existing state cost-share programs.
In addition, GIS output maps ate useful as an education
tool. Landowner cooperation is sometimes a difficult
obstacle hi the implementation of BMP cost-share
programs. These maps could be used by agricultural
field personnel as visual means for promoting program
cooperation. 7
Targeting high-priority waterbodies or watersheds for
TMDL development often involves more than just
technical factors. It may also involve the evaluation of
factors related to recreational, economic, and ecological
values such as the risk to human health and aquatic life;
the degree of public interest and support in protecting a
waterbody; the recreational, economic, and aesthetic
importance of the waterbody; and its vulnerability or
fragility as aquatic habitat Many of these factors
contain spatial components and can be displayed on
maps. Overlaying these maps on maps of potential
pollutant yield would illustrate which waterbodies are of
special concern. Coupled with professional judgement,
-------
these maps could facilitate the prioritization and targeting
of watersheds with the greatest need for TMDL
development.
When sufficient data are collected, the Nornini Creek
Study will provide useful information on the
effectiveness of BMPs in reducing NFS pollution on an
entire watershed in the coastal plain. Furthermore, if the
statistical analyses show a significant instream
phosphorus reduction, this experiment will have provided
a new tool for managers to use in achieving the
Chesapeake Bay 40 percent nutrient reduction goal. It is
important to note, however, that the success of this
technique in the coastal plain will not necessarily make it
a viable modeling tool in other parts of the Chesapeake
drainage. The ten-year monitoring period for Nomini
Creek will be complete in 1995.
REFERENCES
DCR-DSWC. 1986. Nomini Creek Watershed Plan
Westmoreland County, Virginia. Virginia Department of
Conservation and Recreation, Division of Soil and Water
Recreation. Revised 6/10/86.
Omernik, J.M. 1987. Ecoregions of the conterminous
United States. Annals of the Association of American
Geographers 77(1):118-125.
Shanholz, V.O., C.J. Desai, N. Zhang, J.W. Kleene, and
C.D. Metz. 1990. Hydrologic/water quality modeling in
a GIS environment. Written for presentation at the 1990
International Summer Meeting sponsored by the
American Society of Agricultural Engineers, Columbus,
OH, June 24-27, 1990. Paper No. 90-3033.
USDA SCS. 1979. Westmoreland county soil survey.
Soil Conservation Service National Cooperative Soil
Survey, U.S. Department of Agriculture, Washington,
DC.
USEPA. 1991. Guidance for water quality-based
decisions: The TMDL process. U.S. Environmental
Protection Agency, Office of Water, Washington, DC.
EPA 440/4-91-001.
Tb4s case study was prepared by Research Triangle
Institute, Research; triangle Park, KG, in coajuttttioA with
USEPA, Office of Office of Wetlands, Oceans, and
Watersheds, Watershed Mattag0ment: Section, To obtain
copies, contact your EPA Regional 303(d)/lMDL
Coordinatofv
------- |