United States
                        Environmental Protection
                        Agency
                                             Office Of Water
                                             (WH-553)
EPA841-F-93-005
November 1992
Number 5
4>EPA       TMDL   Case   Study
                        Albemarle/Pamlico  Estuary
     Key Feature:
     Project Name:
     Location:
     Scope/Size:
     Land Type:
     Type of Activity:
     Pollutant(s):
     TMDL Development:
     Data Sources:
     Data Mechanisms:
     Monitoring Plan:
     Control Measures:
                    A nutrient screening approach that
                    uses GIS technology to model
                    watersheds within a large, multibasin
                    area.

                    Albemarle/Pamlico Nutrient Budgets
                    EPA Regions HI and IV, North
                    Carolina and Virginia,
                    Albemarle/Pamlico drainage
                    Multibasin estuary, 30,880 mi2
                    Ecoregion 65, Southeastern plains;
                    Ecoregion 63, Middle Atlantic coastal
                    plain
                    Agriculture, forestry, urban
                    Nitrogen and phosphorus
                    PS, NFS
                    State, federal
                    Modeling, GIS (ARC/INFO)
                    No
                    No
                                                           FIGURE 1. Albemarle/Pamlico study area
Summary: Over the past decade, North Carolina's Albemarle
and Pamlico (A/P) Sounds (Figure 1) have experienced
increasing water quality problems ranging from fish kills to
declining populations of aquatic vegetation.  In response to these
and other problems, the A/P Estuarine Study was initiated to characterize the A/P basins and determine potential
management strategies.

Excessive nitrogen and phosphorus loadings were identified early in the study as key factors impairing the health of the
estuary. As a first step to address this problem, the North Carolina Department of Environment, Health, and Natural
Resources, Division of Environmental Management (NCDEM) developed a modeling approach that is currently being used
to screen the A/P watersheds for areas contributing excessive nutrients to surface waters. The approach uses LANDSAT
imagery and geographic information system (GIS) technology to analyze land use within the basins, and then uses a
combination of export coefficients and mass balances to calculate nutrient loadings from the 68 North Carolina watersheds
and 44 Virginia watersheds within the study area.  The output from this model allows relatively rapid graphical
comparison of per-hectare loading values for various levels of spatial aggregation (basin, sub-basin, etc.). North Carolina
intends to use this information to develop TMDLs as part of its new Basinwide Planning system.  The process and
technology used in the A/P Estuarine Study can help other states in the identification, prioritization, and targeting of
critical areas for the TMDL process.                                    .
       Contact;   Randall, Dodd, Research Triaagle Institute, P.O. Box 12194, Research Triangle Park, NC 27709-2194,
                phone (919)541-6491                             -                  	'

-------
 BACKGROUND

 North Carolina's Albemarle and Pamlico basins comprise
 the second largest estuarine system in North America
 (Figure 1). With a total watershed area of 30,880 mi2,
 the A/P basins are home to nearly 2 million permanent
 residents. The area is prized for its many recreational
 opportunities, including swimming, boating, fishing, and
 shellfish harvesting. The system is also a significant
 nursery area for East Coast fisheries.  The A/P fisheries
 industry constitutes a substantial portion of the coastal
 and state economy, with an estimated value of processed
 fisheries products of $64.7 million in  1988.  Table 1
 summarizes the estimated commercial landings of the
 predominant estuarine-dependent finfish, crustaceans, and
 shellfish in the A/P area.

 In recent years, the quality of this valuable resource has
 been declining because of human waste contamination,
 draining of wetlands, increased near-stream development,
 and agricultural and urban runoff.  Effects are being seen
 in the form of declining populations of submerged
 aquatic vegetation, decreasing shellfish yields, skin and
 shell diseases on aquatic organisms, and extensive algal
 blooms.

 In response to these problems, and in  light of the
 system's high recreational and commercial values, the
 A/P system was designated as an estuary of national
 significance in 1987 and was selected to be studied as
 part of EPA's National Estuary Program.  The resulting
 A/P Estuarine Study was initiated as a cooperative
 program between EPA and the State of North Carolina's
 Department of Environment, Health, and Natural
 Resources (DEHNR).  Its purpose is to evaluate the
nature of the basins' environmental problems and to
determine how the estuaries can best be preserved and
managed.
ASSESSING AND CHARACTERIZING
THE PROBLEM

The 1991 Status and Trends report for the A/P Estuarine
Study stated that accelerated eutrophication resulting
from nitrogen and phosphorus loadings is a significant
cause of water quality degradation in many tributaries to
the A/P estuary. In particular, the study found that
nuisance algal blooms, anoxic events,  and some fish kills
have been associated with nutrient loading to the estuary
(Steel, 1991). As a first step toward controlling nutrient
loadings, the A/P Study developed a screening approach
designed to narrow the focus of future nonpoint source
(NPS) control efforts.  The approach consists of
developing nutrient budgets within the 68 sub-basins
composing the North Carolina portion of the A/P study
area (Research Triangle Institute,  1992a).  These budgets
will then serve as a tool for screening out critical areas
of high nutrient loading.  The information gained from
this process is valuable for TMDL development and for
targeting BMP cost share funding to areas causing the
greatest nutrient loadings.,

Calculating the Nutrient Budget

The approach, which uses export coefficients and mass
balance models  to estimate nitrogen and phosphorus
budgets for A/P sub-basins, can be separated into point
and nonpoint source calculations.
TABLE 1. Commercial landings of estuarine-dependent finfish, crustaceans, and shellfish in North Carolina, 1986-1990, in
thousands of pounds (Steel, 1991)
Species
Finfish
Atlantic croaker
Flounder
Atlantic Menhaden
Striped bass
Crustaceans
Blue crab
Shrimp
Shellfish
Clams
Oyster
Bay scallops
1986
9,425
.8,845
66,378
189
23,755
6,162
1,356
745
306
1987
7,289
7,984
55,499
262
32,424
4,416
1,207
1,426
155
1988
8,434
10,265
73,716
116
35,604
8,139
940
913
39
1989
6,824
7,555
66,750
101
34,725
8,923
1,295
530
84
1990
5,731
5,137
71,647
114
38,002
7,802
1,334
323
62

-------
  TABLE 2. LANDSAT Land Use/Cover Types

          Water
          Developed
          Agriculture
          Low density vegetation
          Pine forest
          Bottomland swamp
          Mixed and hardwood
          Shadow
          Marsh
          Atlantic white cedar
Point source discharge data for the A/P study area were
obtained from Discharger Monitoring Reports through
NCDEM and the Virginia Water Control Board.  Recent
(1989-90) nitrogen and phosphorus inputs were
calculated by using the median value of monthly records
of flow and concentration data from  point source
dischargers. The data were transferred to a digital map
in a GIS by using the latitude/longitude values included
for each discharger.  By overlaying watershed boundaries
onto this map, point source discharges of nitrogen and
phosphorus were assigned to each of the 68 A/P sub-
basins.

Nonpoint source loadings were calculated in two ways.
First, export coefficients were used to calculate loadings
from each land use in each sub-
basin of the study area.  This was
accomplished by overlaying data
from a 1987/88 LANDSAT land
use/cover study with digitized
watershed boundaries to determine
the area of each  land use within
each watershed.  A literature review
of export coefficients was then used
to determine low, most likely, and
high values of nutrient export fen-
each land use/cover category.
Multiplying land use areas by the
appropriate export coefficients
yielded nitrogen  and phosphorus
loading values by land use for each
sub-basin.  Table 2 summarizes the
land use/cover categories used in the
study.  Figure 2  shows a GIS-
generated map of relative nitrogen
loadings for the Tar-Pamlico Basin.
                         limitations, this approach was used only for agricultural
                         areas in the 16 gaged watersheds of the A/P study area.
                         Inputs for the mass balance included fertilizer,
                         precipitation, livestock wastes, and nitrogen fixation.
                         Outputs included the nutrients in harvested crops, soil
                         fixation, denitrification, loss to swamp forests, and river
                         export.

                         Numerous data sources and several assumptions were
                         required to prepare the mass balance for the agricultural
                         areas. Fertilizer inputs were assumed to be equal to the
                         amount recommended by the Agricultural Extension
                         Service for a particular crop.  The distribution of each
                         crop type in each sub-basin was determined by
                         overlaying county and watershed boundaries onto the
                         LANDSAT cover information and then combining this
                         information with county-level agricultural statistics. In
                         each county,  the proportion of LANDSAT agricultural
                         land attributed to any crop was determined by the
                         average proportion of county agricultural land planted
                         with that crop during  the period 1987-89.  In areas where
                         one county is in multiple sub-basins, the crop acreages
                         were allocated to each sub-basin assuming the crops
                         were uniformly distributed within the county.

                         Point estimates of atmospheric nitrogen loading were
                         obtained from EPA's  Regional Atmospheric Deposition
                         Model (RADM). The GIS was used to map these points
                         and interpolate contours for wet and dry deposition.
A mass balance approach was also
used to calculate nonpoint source
loadings.  This approach attempts
to account for and balance the input,
output, and storage of nutrients in
the system.  Because of data
                     TOTAL NONPOINT N LOADING (KG/HA*Y)
FIGURE 2. GIS display of relative nitrogen loadings in the Tar-Pamlico basin

-------
 These maps were overlaid on the A/P sub-basin maps,
 and area-weighted averages were calculated for each
 deposition type.

 For determining livestock inputs, county livestock
 inventory data were averaged for the period 1987-1989.
 Per-aniinal nutrient generation values were then used to
 calculate total production.  Based on literature review
 and professional judgment, the values 3, 5, and 10
 percent were chosen as the low, most likely, and high
 scenarios for the proportion of animal waste nutrients
 actually entering the stream.  A  uniform distribution of
 livestock was assumed in allocating county loading
 values to each sub-basin.

 Nitrogen fixation rates (inputs) for soybeans and peanuts
 (the dominant leguminous crops) were determined from
 the literature. Denitrification (output) was assumed to be
 15 percent of applied fertilizer nitrogen, also according to
 the literature. Nutrients lost to harvesting were
 determined by multiplying  county crop production data
 by the nutrient content of each crop as found in the
 literature.  Allocations to sub-basins were made as
 discussed for fertilizer and livestock.

 Calculation of losses to wetlands was limited to the
 portion of agricultural land whose drainage passes
 through wetlands before entering a stream.  Based on the
 literature, wetlands were assumed to drain areas 3, 5, or
 11 times their size. These  areas  were calculated for each
 sub-basin and the portions of agricultural nutrients from
 these areas were reduced by 64 percent for nitrogen and
 43 percent for phosphorus (Kuenzler and Craig, 1986).

 Average annual total nitrogen (TN) export and total
 phosphorus (TP) export were estimated for gaged
 watersheds. These estimates were based on STORET
 daily average flow and concentration data for the 1987-
 89 water years. Average loading (flux) was calculated
 according to the equation:
where

        W   = average flux (kg/y);
        w   as mean of measured flux";
        Q   = mean of daily average flow; and
        q    = mean of measured flow*.

        * on days when nutrients were sampled

To complete the agricultural mass balance, a storage term
was calculated as the difference between total inputs and
outputs.  The nutrient mass contained in the storage term
 was assumed to be associated with soil, groundwater, and
 biomass.

 To reduce GIS programming time and enable modelers
 to use conventional personal computers, both GIS
 (ARC/INFO) and spreadsheets (Excel) were used to
 perform file manipulations and data analysis. The
 modelers chose to use spreadsheet programs whenever
 possible and reserved the GIS for spatial analyses that
 involved watershed and county boundaries and land use
 cover data.

 Targeting and Prioritizing

 Nutrient loadings for each of the 68 sub-basins in the
 North Carolina portion  of the A/P study area were
 summarized into charts depicting the data at various
 levels of spatial aggregation. Graphical analyses of
 nutrient loadings at each level were then used to compare
 relative impacts between basins, sub-basins, or land uses
 within a sub-basin.

 Basin Scale

 Figure 3 depicts per-hectare nonpoint nitrogen loadings
 for each basin in the North Carolina portion of the A/P
 study area. This presents a spatial perspective that the
 state can use to identify critical areas for the TMDL
 process. Specifically, for a state implementing basinwide
 water quality management, this information allows basins
 to be prioritized according to their relative loadings.
 This would help ensure that resources such as monitoring
 staff, modelers, and funding to implement best
 management practices (BMPs) are distributed in the most
 cost-effective manner. In this example, the Chowan,
 Tar-Pamlico, and Neuse basins show relatively high area!
 loading values.  These basins can then be selected for
 analysis at the sub-basin, or  watershed, scale.

 Watershed Scale

 Figure 4 shows the nonpoint nitrogen and phosphorus
 loading from each of the 68  sub-basins within the North
 Carolina portion of the A/P study area. Since the units
 have been converted to per hectare values, they allow
 comparison of the relative impacts of each sub-basin
 within the study area. For the TMDL process, this type
 of information can assist states in targeting critical
 watersheds on a regional basis.  In essence, this can help
 to ensure that the areas selected for TMDL development
 are critical not only on a local scale, but on a regional
 scale as well.

Figure 5 shows nitrogen loadings from sub-basins within
 the Chowan Basin.  As a TMDL prioritization tool, this

-------
                          Nonpoint Nitrogen Loadings (kg/ha-yr) by Basin
                                 For the North Carolina Portion of the A/P Study Area
     c   K ..
     Q)   0 T
     Ol
     2   A ••

     *   2-

         0
               ID
                                I
3
3
                                                      Basin
                                                                         Oi
                                                                        Q.
!
cc
I
                           FIGURE 3. Comparison of nitrogen loadings among basins
perspective allows ranking of the severity of watershed
loadings within a basin that has already been identified
as having critical loadings.  Chowan Sub-basin 4, for
example, shows a particularly high loading rate. Moving
one spatial level lower, a State could then target the
specific land uses within this watershed for BMP
implementation.
         Figure 6 shows the total NFS nitrogen loads from each
         LANDSAT cover category in Chowan Sub-basin 4. This
         type of information is useful in targeting land uses for
         management practices and justifying controls to the
         affected public (e.g., row-crop farmers).
Nonpoint N and P loadings by Sub-Basin (kg/ha-yr)
1J> 0
T Potential _ |j T
Target Sub- ^H k n t 1>B

x-v Basins a B^H 8 1 "H-6
%T H H H B H 1
ff B-L i 1 | 1 I 1 I • I I
S5 ell If L 1 1 1 n. 1 • 1. I BBMiill IE.B! 1 1 il t!'2
r 6 11 nrj Is i i IB bill hlB Ii5ill 1 R Hi nl i It1
S SB BpinB HllnHnn HH HB MHH 3ElE:l»MH NWHHHti H»»HE4nHMH H SHfclHH 0HR i 9R 1
i1 JI 1 II ISlliliill ill 1! Illlllll illlSlIlSiSS IiilII 1 111 I0-8
u -" r i jT^T^rr"!^T^'p rn rf^^r« r* » ™™ri r (J
Sub-Basin #
^ Nitrogen H Phosphorous

Phosphorous (kg/ha-yr)
              FIGURE 4.  Identifying critical sub-basins within the Albemarle/Pamlico study area

-------
                               Nitrogen Loadings in Sub-Basins of the Chowan Basin
                                                      (kg/ha)
                                                   3           4
                                                    Sub-Basin f
                      FIGURE 5. Comparison of sub-basin loadings within a targeted basin
FOLLOW-UP

As part of its new Basinwide Planning process, North
Carolina intends to use the information obtained from the
A/P screening to target areas for TMDL development
Watersheds showing high nitrogen and phosphorus
loadings, especially those nearest the estuaries, will be
selected for special management attention.  The Tar-
Pamlico Basin, in the central portion of the study area,
has already been targeted for PS/NPS trading of nitrogen
loads.

Once critical watersheds have been identified, other
small-scale nutrient and sediment models can be used to
                                         assist in determining the optimum BMP scenarios to
                                         reach desired reduction limits.  The Nomini Creek
                                         TMDL Case Study (#4) discusses the use of watershed-
                                         scale models with a GIS in order to quantify loads and
                                         identify critical nonpoint source areas.

                                         The A/P modeling tools are currently being used to
                                         develop and test nutrient control and population growth
                                         scenarios across the A/P basins.  Additional GIS tools
                                         are being developed by NCDEM to help decision makers
                                         analyze the water quality impacts of nonpoint source
                                         controls.  For example, GIS technology is being used to
                                         provide maps highlighting areas of critical nitrogen and
                                         phosphorus loadings.  For  TMDL targeting, these maps
                    2,500,000 -•


                    2,000,000 -


                    1.500,000 -•
                                   Point and Non-Point Nitrogen Loadings by Landuse
                                                    Chowan Sub-Basin #4
                  S
                  Dl
                  g

                  z
o 1,000.000 -•
                      500.000 -
                                 DIRECT     AGRICULTURE     FOREST &
                              ATMOSHERIC                   OTHER
                                                      POINT       DEVELOPED
                                                     SOURCES
                        FIGURE 6. Comparison of loading sources in a targeted sub-basin

-------
 will be combined with maps showing other variables of
 concern, such as public interest, aesthetic value, or
 fragility of a particular system.  A GIS can then generate
 a master targeting map, showing areas where multiple;
 water quality concerns overlap (Figure 7). Similar •* ':, I
 critical area maps are being developed for the A/P study
 using a variety of GIS data layers.

 GIS technology is also expected to play a role in
 illustrating the need for controls to the agricultural
 community and municipal dischargers. In addition, the
 GIS will be used to study the use of riparian buffer
 protection and restoration to supplement on-the-fann
 BMPs,  which alone may not achieve the desired nitrogen
 and phosphorus reductions in A/P watersheds.
                       N Loadings
                       P Loadings
                      Recreational
                        Values
                      Public Interest
                      High QualltyX.
                      Nursery Areas
                          Fish
                      Consumption
                       Advisories
                      MASTER
                     TARGETING
                        MAP
 FIGURE 7.  Creation of a master targeting map

The A/P databases and modeling results are being
incorporated into "sub-basin profiles," which summarize
the characteristics of each sub-basin (e.g., land use, point
sources, crop statistics, livestock operations, pollutant
loading characteristics). This information has also been
compiled into a PC data base, which will allow NCDEM
users to have desktop access to watershed data that were
previously available only through a mainframe or a GIS.
 for its ability to handle a large multibasin area, while still
 providing output for screening at the sub-basin level. It
 is important to note that the A/P nutrient screening
 approach estimates annual loadings, a feature that is
 suitable within a long-term context for nutrient
 management of basins draining to estuaries and lakes.
 Data requirements are such that small watersheds or
 large basins can be screened and management scenarios
 can be tested (e.g., combinations of BMPs and PS
 control strategies).

 The A/P approach is, however, more data-intensive than
 some screening models. Where nationally available data
 were out-of-date or unreliable (e.g.,  land use/cover
 information, some agricultural statistics) the latest data
 were acquired (e.g., the 1987-88 LANDSAT imagery and
 state-maintained crop statistics) to ensure that screening
 results would be credible to the decision makers and to
 other stakeholders as well.
REFERENCES

Kuenzler, E.J., and NJ. Craig. 1986.  Land use and
nutrient yields of the Chowan River watershed.  In
Watershed research perspectives, ed. D.L. Correll,
Smithsonian Institution Press, Washington, DC.

RTI.  1992a. Watershed planning In the Albemarle-
Pamlico Study Area, Report 1: Annual nutrient budgets.
Research Triangle Institute, Research Triangle Park, NC.

RTI.  1992b. Watershed planning in the Albemarle-
Pamlico study Area, Report 2: Mass balances for gaged
watersheds.  Research Triangle Institute,  Research
Triangle Park, NC.

Steel, J., ed.  1991. Albemarle-Pamlico Estuarine,
System, technical analysis of status and trends. A/P
Estuarine Study Report 90-01.
OSace of
                            a*
                   
-------

-------