total maximum daily loads (tmdls) conowingo creek ... · tmdl endpoints established for this...

Total Maximum Daily Loads (TMDLs) Conowingo Creek, Lancaster County

Prepared by: Susquehanna River Basin Commission Water Quality and Monitoring Program

March 2, 2001

i

TABLE OF CONTENTS EXECUTIVE SUMMARY............................................................................................................. 1 INTRODUCTION........................................................................................................................... 3 TMDL ENDPOINTS ...................................................................................................................... 7 Nutrient Loads and Organic Enrichment in Stream Systems.......................................................... 7 SELECTION OF THE REFERENCE WATERSHED................................................................... 8 DATA COMPILATION AND MODEL OVERVIEW................................................................ 11 GIS BASED DERIVATION OF INPUT DATA.......................................................................... 12 WATERSHED ASSESSMENT AND MODELING.................................................................... 15 TMDL COMPUTATIONS FOR PHOSPHORUS AND SEDIMENT ........................................ 17 TMDL Computation...................................................................................................................... 17

Margin of Safety................................................................................................................ 19 Load Allocation................................................................. Error! Bookmark not defined. Phosphorus ........................................................................ Error! Bookmark not defined. Sediment............................................................................................................................ 22

CONSIDERATION OF CRITICAL CONDITIONS ................................................................... 24 CONSIDERATION OF SEASONAL VARIATIONS................................................................. 25 RECOMMENDATIONS .............................................................................................................. 25 PUBLIC PARTICIPATION ......................................................................................................... 25 LITERATURE CITED ................................................................................................................. 26

LIST OF TABLES

Table 1. TMDL Endpoints for the Conowingo Creek Watershed ................................................. 1 Table 2. List of Impaired Streams with Designated Allocation Units........................................... 4 Table 3. Comparison Between Conowingo Creek and North Branch Muddy Creek Watersheds....................................................................................................................................................... 11 Table 4. GIS Data Sets Used by the AVGWLF Model ............................................................... 13 Table 5. Header Information Contained in Tables 6 and 7 .......................................................... 16 Table 6. Existing Loading Values for Conowingo Creek ............................................................ 16 Table 7. Existing Loading Values for North Branch Muddy Creek ............................................ 17 Table 8. Unit Area Loads for the Conowingo and North Branch Muddy Creek Watersheds ..... 17 Table 9. TMDL Computation for Conowingo Creek................................................................... 18 Table 10. TMDL Allocations for Conowingo Creek ................................................................... 18 Table 11. TMDLs for Conowingo Creek ..................................................................................... 18 Table 12. Conowingo Creek Load Allocation for Phosphorus by Land Use/Source .................. 22 Table 13. Conowingo Creek Load Allocation for Sediment by Land Use/Source ...................... 24

- ii -

LIST OF FIGURES

Figure 1. Location Map for Conowingo Creek Watershed............................................................ 2 Figure 2. Map Showing Impaired Streams and Allocation Units for Conowingo Creek .............. 6 Figure 3. Map Showing the Land Use for Conowingo Creek........................................................ 9 Figure 4. Map Showing the Land Use for North Branch Muddy Creek...................................... 10

APPENDICES

Appendix A. GWLF Users Manual ................................................................................................ Appendix B. Strategy for Conducting Nutrient Related TMDL Assessments for Streams in Pa. . Appendix C. Introduction to Watershed Hydrology, Simulation, and Pollutant Transport ........... Appendix D. EMPR Methodology ................................................................................................. Appendix E. Model Input and Results ........................................................................................... Appendix F. Comment and Response (Final Document Only) .....................................................

1

EXECUTIVE SUMMARY The Conowingo Creek Watershed in Lancaster County has been scheduled for TMDL development. The watershed covers 34 square miles and is located south of the City of Lancaster (Figure 1). From Drumore, Pennsylvania, state route 272 parallels Conowingo Creek until it crosses the Pennsylvania/Maryland border. The protected use of the watershed is aquatic life. The aquatic life designation for the main stem Conowingo Creek is cold water fishes, with the tributaries designated as high quality, cold water fishes. The Susquehanna River Basin Commission (SRBC) developed Total Maximum Daily Loads, or TMDLs, for Conowingo Creek Watershed to address the impairments noted on Pennsylvania’s 1996 and 1998 303(d) lists and the 2000 305(b) report. Excess nutrient and sediment loads from agriculture are causing the impairments. The nutrient portion of the TMDL focuses on control of phosphorus. The nitrogen/phosphorus ratio for Conowingo Creek is approximately 13. Since phosphorus is limiting when the N/P ratio is greater than 10, only phosphorus loadings were addressed in this TMDL. Pennsylvania does not currently have water quality criteria for sediment or nutrients. For this reason, Pennsylvania’s Department of Environmental Protection (Pa. DEP) developed a reference watershed approach to identify the TMDL endpoints, or water quality objectives, for phosphorus and sediment in the impaired segments of the Conowingo Creek Watershed. By comparison to a similar non-impaired watershed, it has been estimated that the amount of phosphorus loading that will meet the water quality objectives for Conowingo Creek is approximately 21,338 lbs/yr (pounds per year)(Table 1). Sediment loading must be limited to 22,514,862 lbs/yr (Table 1). When these values are met, Conowingo Creek will support its aquatic life uses.

Table 1. TMDL Endpoints for the Conowingo Creek Watershed

Pollutant Current Loading (lbs/yr) TMDL (lbs/yr)

Percent Reduction in Loads Needed to Meet

TMDL Phosphorus 52,375 21,338 59% Sediment 27,301,404 22,514,862 18%

The TMDLs are allocated to runoff from agriculture and developed areas (Load Allocations - LAs), with 10% of the allowable loading reserved as a margin of safety (MOS). The TMDLs cover a total of 31 stream miles within Conowingo Creek Watershed. The TMDL establishes a reduction for phosphorus loading of 59% from the current yearly loading of 52,375 pounds, and a reduction in sediment loading of 18% from the current yearly loading of 27,301,404 pounds. More complete discussions of the Conowingo Creek TMDL and TMDLs in general are contained in the Information Sheet and the body of this document.

- 2 -

Figure 1 – Location map for Conowingo Creek Watershed.

- 3 -

INTRODUCTION Total Maximum Daily Loads, or TMDLs, were developed for the Conowingo Creek Watershed to address the impairments noted on Pennsylvania’s 1996 and 1998 303(d) lists, and the 2000 305(b) report (Table 2). Figure 2 shows the impaired segments within the Conowingo Creek Watershed. The main stem of Conowingo Creek was placed on the 1996 303(d) List based on data compiled for the 1996 305(b) report. A total of 16 miles were listed as impaired due to agriculture. The cause code indicated excess amounts of nutrient and suspended solids to be a problem. In 1999, as part of Pa. DEP’s Unassessed Waters Program, an additional 15 miles were added to the year 2000 305(b) report. All the biological surveys included kick screen sampling of benthic macroinvertebrates, and habitat surveys. Benthic macroinvertebrates were identified to family in the field. The biological surveys indicated impairment due to excessive amounts of sediment and nutrients, organic enrichment, and low dissolved oxygen (DO). Agricultural land use in the watershed is the cause for the violations of the aquatic life use. The primary method adopted by Pa. DEP for evaluating waters changed between the publication of the 1996 and 1998 303(d) lists. Pa. DEP is now using a modification of U.S. Environmental Protection Agency’s (U.S. EPA) Rapid Bioassessment Protocol II (RPB-II) as the primary mechanism to assess Pennsylvania’s unassessed waters. The assessment method requires selecting stream sites that would reflect impacts from surrounding land uses that are representative of the stream segment being assessed. The biologist selects as many sites as necessary to establish an accurate assessment for a stream segment. At each site, a biological assessment is conducted using the modified RBP II method. The length of the stream segment assessed can vary between sites. There are several factors that determine site location and how long a segment can be. These factors include distinct changes in stream characteristics, surface geology, riparian land use, and the pollutant that is causing impairment. For the purpose of TMDL development, it is often necessary to aggregate 303(d) listed stream segments. The primary reason to address multiple segments is compatibility with data used in TMDL analysis. For these TMDL analyses, the primary data sources are geographic information system (GIS) derived data. The land cover data set used for this analysis is represented by 30-meter squares. If the stream segment area for TMDL development is too small, error is introduced by using the data beyond its capability. For this reason, we have aggregated segments listed in the Conowingo Creek Watershed. This results in completing TMDLs for several segments, although the model analysis was completed as one watershed area. Neither Pennsylvania nor the U.S. EPA currently has water quality criteria for sediment or nutrients. Therefore, Pa. DEP developed a reference watershed approach to identify the TMDL endpoints or water quality objectives for nutrients and sediment in the impaired segments of the Conowingo Creek Watershed. The nutrient portion of the TMDL for this watershed addresses phosphorus, since it has been identified as the limiting nutrient. The Conowingo Creek Watershed TMDL Information Sheet that is attached to this document provides a primer for TMDLs (What are they and why are we doing them?) and water quality

- 4 -

standards (What makes up a water quality standard?). Appendices A and B provide information on the method being used by Pennsylvania for establishment of TMDLs.

Table 2. List of Impaired Streams with Designated Allocation Units Segment

ID Stream Code

Year Listed

Stream Name (Designated

Use) Source Code Cause Code Miles Degraded

Allocation Unit

6326 7162 1996 Conowingo Creek (CWF) Agriculture Nutrients

Suspended Solids

12.64 6

6326 7171 1998

Conowingo Creek/Unnamed Tributary (HQ-CWF)

Agriculture Nutrients

Suspended Solids

0.12 1

990629-1155-BPG

7171 2000


Crop/Grazing Related

Agriculture

Nutrients Organic

Enrichment/Low D.O.

Siltation

1.45 1

6326 7172 1998


Agriculture Nutrients

Suspended Solids

1.16 1

990629-1155-BPG

7173 2000



Agriculture

Nutrients Organic

Enrichment/Low D.O.

Siltation

1.33 1

990629-1155-BPG

7174 2000



Agriculture

Nutrients Organic

Enrichment/Low D.O.

Siltation

0.56 1

990707-1105-BPG

7186 2000


Agriculture

Nutrients Organic

Enrichment/Low D.O.

Siltation

2.77 5

990707-1105-BPG

7187 2000


Agriculture

Nutrients Organic

Enrichment/Low D.O.

Siltation

1.61 5

990629-1420-BPG

7176 2000

Little Conowingo

Creek (HQ-CWF)

Agriculture

Nutrients Organic

Enrichment/Low D.O.

Siltation

5.81 3

990629-1420-BPG

7177 2000

Little Conowingo

Creek (HQ-CWF)

Agriculture

Nutrients Organic

Enrichment/Low D.O.

Siltation

1.59 3

- 5 -

Table 2. List of Impaired Streams with Designated Allocation Units Segment

ID Stream Code

Year Listed

Stream Name (Designated

Use) Source Code Cause Code Miles Degraded

Allocation Unit

990629-1420-BPG

7178 2000

Little Conowingo

Creek (HQ-CWF)

Agriculture

Nutrients Organic

Enrichment/Low D.O.

Siltation

1.05 3

990629-1420-BPG

7179 2000

Little Conowingo

Creek (HQ-CWF)

Agriculture

Nutrients Organic

Enrichment/Low D.O.

Siltation

0.82 3

- 6 -

Figure 2. - Map showing impaired streams and allocation units for Conowingo Creek.

- 7 -

TMDL ENDPOINTS The TMDLs developed for the Conowingo Creek Watershed address sediment and phosphorus. Phosphorus was determined to be the nutrient limiting plant growth in Conowingo Creek Watershed. Because neither Pennsylvania nor EPA has water quality criteria for sediments or phosphorus, a method was developed to determine water quality objectives for these parameters that would result in the impaired stream segments attaining their designated uses. The method employed for these TMDLs is termed the “Reference Watershed Approach.” The Reference Watershed Approach compares two watersheds, one attaining its uses and one that is impaired based on biological assessment. Both watersheds must have similar land cover and land use characteristics. Other features such as base geologic formation should be matched to the extent possible; however, most variations can be adjusted in the model. The objective of the process is to reduce the loading rates of nutrients and sediment in the impaired stream segment to a level equivalent to, or slightly lower than, the loading rates in the reference stream segment. This load reduction will allow the biological community to return to the affected stream segments. The TMDL endpoints established for this analysis were determined using North Branch Muddy Creek as the reference watershed. The North Branch Muddy Creek Watershed lies within the lower Susquehanna River Basin, in York County, Pennsylvania. These endpoints are discussed in detail in the “Selection of the Reference Watershed” section. Nutrient Loads and Organic Enrichment in Stream Systems As indicated earlier, Conowingo Creek was listed as being impaired due to problems associated with nutrient loads and suspended sediments. In stream systems, elevated nutrient loads (nitrogen and phosphorus) can lead to increased productivity of plants and other organisms (Novotny and Olem, 1994). Additional problems can also occur if nutrient loads are not reduced. Typically, the quantities of trace elements are plentiful in aquatic ecosystems; however, nitrogen and phosphorus may be in short supply. The nutrient that is in the shortest supply is called the limiting nutrient, because its relative quantity affects the rate of production (growth) of aquatic biomass. If the nutrient load to a water body can be reduced, the available pool of nutrients that can be utilized by plants and other organisms will be reduced (Novotny and Olem, 1994). In most efforts to control eutrophication processes in water bodies, emphasis is placed on the limiting nutrient. In some instances, this may not always be the case. For example, if nitrogen is the limiting nutrient, it still may be more efficient to control phosphorus loads if the nitrogen originates from difficult to control sources such as nitrates in ground water. The ratio of the amount of nitrogen (N) to the amount of phosphorus (P) is often used to determine which nutrient is limiting (Thomann and Mueller, 1987). If the N/P ratio is less than 10, nitrogen is limiting; if the N/P ratio is greater than 10, phosphorus is the limiting nutrient. A ratio equal to 10 indicates neither phosphorus nor nitrogen is limiting. In the case of Conowingo Creek, the N/P ratio is approximately 13. Since the N/P ration is 13, only phosphorus was addressed by the TMDL. Controlling the phosphorus loading to Conowingo Creek will limit plant growth and result in raising the dissolved-oxygen levels.

- 8 -

SELECTION OF THE REFERENCE WATERSHED The reference watershed approach was used to estimate the appropriate reduction of phosphorus and sediment loading necessary to restore healthy aquatic communities to the Conowingo Creek Watershed. This approach is based on selecting a non-impaired watershed (“reference”) and determining its current loading rates for the pollutants of interest. The objective of the process is to reduce loading rates of those pollutants identified as causing impairment to a level equivalent to the loading rates in the reference watershed. Achieving the appropriate load reductions should allow the return of a healthy biological community to affected stream segments. In general, three factors should be considered when selecting a suitable reference watershed. The first factor is to use a watershed that has been assessed by the Department using the Unassessed Waters Protocol and has been determined to attain water quality standards. The second factor is to find a watershed that closely resembles Conowingo Creek Watershed in physical properties such as land cover/land use, physiographic province, and geology. Finally, the size of the reference watershed should be within 20-30% of the impaired watershed area. The search for a reference watershed that would satisfy the above characteristics was done by means of a desktop screening using several GIS coverages including the Multi-Resolution Land Characteristics (MRLC) Landsat-derived land cover/use grid, the Pennsylvania’s 305(b) assessed streams database, and geologic rock types.

The watershed used as a reference for the Conowingo Creek Watershed is the North Branch Muddy Creek Watershed. This watershed is located in the Piedmont Province in State Water Plan (SWP) Basin 7I, York County. The digitized (reference) watershed is referred in this report as "North Branch Muddy Creek Watershed". Table 3 compares the two watersheds in terms of their size, location, and other physical characteristics. All of the North Branch Muddy Creek stream segments have been assessed and were found to be unimpaired. The analysis of value counts for each pixel of the MRLC grid revealed that land cover/use distributions in both watersheds are fairly similar. The agricultural land use, which is the source of impairment in Conowingo Creek Watershed, accounts for 83% of the total land area as compared to 63% in North Branch Muddy Creek Watershed (Figures 3 and 4). North Branch Muddy Creek Watershed has significantly less agricultural lands, however, the availability of unimpaired watersheds with similarly high agricultural land use was limited. The surficial geologies of the Conowingo Creek and North branch muddy watersheds were also compared and are a perfect match. The geology of both watersheds consists entirely of igneous/metamorphic rock. The bedrock geology affects primarily surface runoff and background nutrient loads through its influences on soils and landscape as well as fracture density and directional permeability. A look at these attributes in Table 3 indicates that these watersheds also compare very well in terms of average precipitation and soil K factor. The portion of North Branch Muddy Creek Watershed selected for the analyses is approximately 44 square miles, comparable to the 34 square miles of the Conowingo Creek Watershed.

- 9 -

Figure 3. – Map showing the land use for Conowingo Creek.

- 10 -

Figure 4. – Map showing the land use for North Branch Muddy Creek.

- 11 -

Table 3. Comparison Between Conowingo Creek and North Branch Muddy Creek Watersheds

Attribute

Conowingo Creek

North Branch Muddy Creek

Physiographic Province

Piedmont

Piedmont

Area (square miles) 34 44

Predominant Land Use Agriculture (83%) Agriculture (63%)

Geology Igneous/Metamorphic Igneous/Metamorphic

Soils

Dominant HSG B (98%), C (2%) B (100%) K Factor 0.32 0.32 20-Year Average Rainfall (in)

36.6

36.6

20-Year Average Runoff (in)

1.3

1.3

DATA COMPILATION AND MODEL OVERVIEW The TMDLs were developed using the GWLF model. The GWLF model provides the ability to simulate runoff, sediment, and nutrient (N and P) loadings from given variable-size source areas (e.g., agricultural, forested, and developed land). It also has algorithms for calculating septic system loads, and allows for the inclusion of point source discharge data. It is a continuous simulation model that uses daily time steps for weather data and water balance calculations. Monthly calculations are made for sediment and nutrient loads, based on the daily water balance accumulated to monthly values. GWLF is a combined distributed/lumped parameter watershed model. For surface loading, it is distributed in the sense that it allows multiple land use/cover scenarios. Each area is assumed to be homogeneous in regard to various attributes considered by the model. Additionally, the model does not spatially distribute the source areas, but aggregates the loads from each area into a watershed total. In other words, there is no spatial routing. For subsurface loading, the model acts as a lumped parameter model using a water balance approach. No distinctly separate areas are considered for subsurface flow contributions. Daily water balances are computed for an unsaturated zone as well as a saturated subsurface zone, where infiltration is computed as the difference between precipitation and snowmelt minus surface runoff plus evapotranspiration. GWLF models surface runoff using the Soil Conservation Service Curve Number (SCS-CN) approach with daily weather (temperature and precipitation) inputs. Erosion and sediment yield are estimated using monthly erosion calculations based on the Universal Soil Loss Equation (USLE) algorithm (with monthly rainfall-runoff coefficients) and a monthly composite of KLSCP values for each source area (e.g., land cover/soil type combination). The KLSCP factors are variables used in the calculations to depict changes in soil loss erosion (K), the length slope factor (LS), the vegetation cover factor (C), and conservation practices factor (P). A sediment delivery ratio based

- 12 -

on watershed size and a transport capacity based on average daily runoff are applied to the calculated erosion to determine sediment yield for each source area. Surface nutrient losses are determined by applying dissolved N and P coefficients to surface runoff and a sediment coefficient to the yield portion for each agricultural source area. Point source discharges can also contribute to dissolved oxygen losses to the stream and are specified in terms of kilograms per month. Manured areas, as well as septic systems, can also be considered. Urban nutrient inputs are all assumed to be solid phase, and the model uses an exponential accumulation and washoff function for these loadings. Subsurface losses are calculated using dissolved N and P coefficients for shallow groundwater contributions to stream nutrient loads. The subsurface sub-model only considers a single, lumped-parameter contributing area. Evapotranspiration is determined using daily weather data and a cover factor dependent upon land use/cover type. Finally, a water balance is performed daily using supplied or computed precipitation, snowmelt, initial unsaturated zone storage, maximum available zone storage, and evapotranspiration values. All of the equations used by the model can be viewed in Appendix A, GWLF Users Manual. For execution, the model requires three separate input files containing transport-, nutrient-, and weather-related data. The transport (TRANSPRT.DAT) file defines the necessary parameters for each source area to be considered (e.g., area size, curve number, etc.) as well as global parameters (e.g., initial storage, sediment delivery ratio, etc.) that apply to all source areas. The nutrient (NUTRIENT.DAT) file specifies the various loading parameters for the different source areas identified (e.g., number of septic systems, urban source area accumulation rates, manure concentrations, etc.). The weather (WEATHER .DAT) file contains daily average temperature and total precipitation values for each year simulated. GIS BASED DERIVATION OF INPUT DATA The primary sources of data for this analysis were geographic information system (GIS) formatted databases. A specially designed interface was prepared by the Environmental Resources Research Institute of the Pennsylvania State University in ArcView (GIS software) to generate the data needed to run the GWLF model. The GWLF model was originally developed by Cornell University. The new version of this model has been named AVGWLF (ArcView Version of the Generalized Watershed Loading Function) In using this interface, the user is prompted to identify required GIS files and to provide other information related to “non-spatial” model parameters (e.g., beginning and end of the growing season, the months during which manure is spread on agricultural land, and the names of nearby weather stations). This information is subsequently used to automatically derive values for required model input parameters, which are written to the TRANSPRT.DAT, NUTRIENT.DAT and WEATHER.DAT input files needed to execute the GWLF model (see Appendix A). For use in Pennsylvania, AVGWLF has been linked with statewide GIS data layers such as land use/cover, soils, topography, and physiography; location-specific default information such as background N and P concentrations and cropping practices are also included. Complete GWLF-

- 13 -

formatted weather files are also included for eighty-eight weather stations around the state. Table 4 lists the GIS data sets and provides an explanation of how they were used for development of the input files for the GWLF model.

Table 4. GIS Data Sets Used by the AVGWLF Model

Censustr Coverage of census data including information on individual home septic systems. The attribute susew_sept includes data on conventional systems, and su_other provides data on short circuiting and other systems.

County The county boundaries coverage lists data on conservation practices that provide C and P values in the Universal Soil Loss Equation (USLE).

Gwnback A grid of background concentrations of N in groundwater derived from water well sampling. Landuse5 Grid of the MRLC that has been reclassified into five categories. This is used primarily as a

background. Majored Coverage of major roads. Used for reconnaissance of a watershed. MCD Minor civil divisions (boroughs, townships and cities). Npdespts A coverage of permitted point discharges. Provides background information and cross check

for the point source coverage. Padem 100-meter digital elevation model. This is used to calculate slope and aspect. Palumrlc A satellite image derived land cover grid that is classified into 15 different landcover

categories. This dataset provides landcover loading rate for the different categories in the model.

Pasingle The 1:24,000 scale single line stream coverage of Pennsylvania. Provides a complete network of streams with coded stream segments.

Physprov A shapefile of physiographic provinces. Attributes rain_cool and rain_warm are used to set recession coefficient

Pointsrc Major point source discharges with permitted N and P loads. Refwater Shapefile of reference watersheds for which nutrient and sediment loads have been calculated. Soilphos A grid of soil phosphorous loads which have been generated from soil sample data. Used to

help set phosphorus and sediment values. Smallsheds A coverage of watersheds at the 1:24,000 This coverage is used with the stream network to

delineate the desired level watershed. Statsgo A shapefile of generalized soil boundaries. The attribute mu_k sets the k factor in the USLE.

The attribute mu_awc is the unsaturated available capacity, and the muhsg_dom is used with land use cover to derive curve numbers.

Strm305 A coverage of stream water quality as reported in the Pennsylvania’s 305(b) report. Current status of assessed streams.

Surfgeol A shapefile of the surface geology used to compare watersheds of similar qualities. T9sheds Data derived from a PA DEP study conducted at PSU with N and P loads. Zipcode A coverage of animal densities. Attribute aeu_acre helps estimate N & P concentrations in

runoff in agricultural lands and over manured areas. Weather Files Historical weather files for stations around Pennsylvania to simulate flow.

As described in the Data Compilation and Model Overview section, the GWLF model provides the ability to simulate surface water runoff, as well as sediment and nutrient loads from a watershed based on landscape conditions such as topography, land use/cover, and soil type. In essence, the model is used to estimate surface runoff and nonpoint source loads from different areas within the watershed. If point source discharges are identified, and the corresponding nutrient loads are quantified, these loads are summed to represent the total pollutant loads for the watershed.

- 14 -

In the GWLF model, the nonpoint source (or “background”) load calculated is affected by terrain conditions such as amount of agricultural land, land slope, and inherent soil erodibility. It is also affected by farming practices utilized in the area, as well as by background concentrations of nutrients (i.e., N and P) in soil and groundwater. Various parameters are included in the model to account for these conditions and practices. Some of the more important parameters are summarized below: Areal extent of different land use/cover categories: This is calculated directly from a GIS layer of land use/cover. Curve number: This determines the amount of precipitation that infiltrates into the ground or enters surface water as runoff. It is based on specified combinations of land use/cover and hydrologic soil type, and is calculated directly using digital land use/cover and soils layers. K factor: This factor relates to inherent soil erodibility, and affects the amount of soil erosion taking place on a given unit of land. LS factor: This factor signifies the steepness and length of slopes in an area and directly affects the amount of soil erosion. C factor: This factor is related to the amount of vegetative cover in an area. In agricultural areas, the crops grown and the cultivation practices utilized largely control this factor. Values range from 0 to 1.0, with larger values indicating greater potential for erosion. P factor: This factor is directly related to the conservation practices utilized in agricultural areas. Values range from 0 to 1.0, with larger values indicating greater potential for erosion. Sediment delivery ratio: This parameter specifies the percentage of eroded sediment that is delivered to surface water and is empirically based on watershed size. Unsaturated available water-holding capacity: This relates to the amount of water that can be stored in the soil and affects runoff and infiltration. It is calculated using a digital soils layer. Dissolved nitrogen in runoff: This varies according to land use/cover type, and reasonable values have been established in the literature. This rate, reported in mg/l, can be re-adjusted based on local conditions such as rates of fertilizer application and farm animal populations. Dissolved phosphorus in runoff: Similar to nitrogen, the value for this parameter varies according to land use/cover type, and reasonable values have been established in the literature. This rate, reported in mg/l, can be re-adjusted based on local conditions such as rates of fertilizer application and farm animal populations. Nutrient concentrations in runoff over manured areas: These are user-specified concentrations for N and P that are assumed to be representative of surface water runoff leaving areas on which manure has been applied. As with the runoff rates described above, these are based on values

- 15 -

obtained from the literature. They also can be adjusted based on local conditions such as rates of manure application or farm animal populations. Nutrient build-up in non-urban areas: In GWLF, rates of build-up for both N and P have to be specified. In Pennsylvania, this is estimated using historical information on atmospheric deposition. Background N and P concentrations in groundwater: Subsurface concentrations of nutrients (primarily N) contribute to the nutrient loads in streams. In Pennsylvania, these concentrations are estimated using recently published data from USGS. Background N and P concentrations in soil: Since soil erosion results in the transport of nutrient-laden sediment to nearby surface water bodies, reasonable estimates of background concentrations in soil must be provided. In Pennsylvania, this information is based on literature values as well as soil test data collected annually at Penn State University. These values can be adjusted locally depending upon manure loading rates and farm animal populations. Other less important factors that can affect sediment and nutrient loads in a watershed are also included in the model. More detailed information about these parameters and those outlined above can be obtained from the GWLF Users Guide provided in Appendix A of this document. Specific details in this guide that describe equations and typical parameter values used can be found on pages 15 through 41. Additional descriptions of hydrologic functions and pollutant transport processes that operate within a watershed can be found in Appendix C. WATERSHED ASSESSMENT AND MODELING The AVGWLF model was run to establish existing loading conditions for the North Branch Muddy Creek Watershed, and the Conowingo Creek Watershed. Adjustments to specific GWLF-related parameters were made based on information gathered from county conservation districts and field observations. North Branch Muddy Creek

• reset c for cropland to 0.16 from 0.21 • reset p for cropland to 0.30 from 0.45; contour farming and the presence of natural

riparian buffers found along the entrenched creek • nitrogen concentrations in groundwater adjusted to 4.0; based on well data provided with

the AVGWLF model Conowingo Creek

• reset c for cropland to 0.3 from 0.26 • reset p for cropland to 0.55 from 0.45; lack of riparian buffers and significant stream

bank erosion • nitrogen concentrations in groundwater adjusted to 7.5; based on well data provided with

the AVGWLF model

- 16 -

Table 5 presents an explanation of the header information contained in Tables 6 and 7. The 20-year means for these parameters for Conowingo Creek are shown in Table 6. Table 7 shows the results for the reference watershed, North Branch Muddy Creek.

Table 5. Header Information Contained in Tables 6 and 7

Land Use Category The land cover classification that was obtained from the MRLC database Area (acres) The area of the specific land cover/land use category found in the watershed. Total P The estimated total phosphorus loading that reaches the outlet point of the watershed that is

being modeled. Expressed in lbs/year. Unit Area P Load The estimated loading rate for phosphorus for a specific land cover/land use category.

Loading rate is expressed in lbs/acre/year Total N The estimated total nitrogen loading that reaches the outlet point of the watershed that is

being modeled. Expressed in lbs/year. Unit Area N Load The estimated loading rate for nitrogen for a specific land cover/land use category. Loading

rate is expressed in lbs/acre/year Total Sed The estimated total sediment loading that reaches the outlet point of the watershed that is

being modeled. Expressed in lbs/year. Unit Area Sed Load The estimated loading rate for sediment for a specific land cover/land use category. Loading

rate is expressed in lbs/acre/year

Table 6. Existing Loading Values for Conowingo Creek

Land Use Area (acres) Total P (lbs/yr) Total N (lbs/yr) Sed Load (lbs/yr) HAY/PAST 8212.87 3449.41 13068.21 1708190.44 CROPLAND 9947.29 47448.57 104509.72 25528849.13 CONIF_FOR 542.41 4.40 16.31 2235.13 MIXED_FOR 481.91 3.60 13.23 1862.61 DECID_FOR 4447.81 71.92 163.14 33154.41 DEVELOPED 76.05 144.50 11821.06 27112.17 GROUNDWATER N/A 1252.28 530201.89 0.00 POINT SOURCE N/A 0.00 0.00 0.00 TOTAL 23708.34 52374.68 659793.56 27301403.89

- 17 -

Table 7. Existing Loading Values for North Branch Muddy Creek

Land Use Area (acres) Total P (lbs/yr) Total N (lbs/yr) Sed Load (lbs/yr)

HAY/PAST 4909.88 1098.33 6210.58 1173108.67 CROPLAND 12676.23 21608.17 93724.38 24495458.31 CONIF_FOR 303.93 2.87 16.31 3118.41 MIXED_FOR 358.30 3.75 20.50 4027.80 DECID_FOR 9382.39 614.86 2358.70 714034.89 TRANSITION 4.94 27.78 106.92 31579.13 DEVELOPED 214.98 55.11 481.93 27112.83 GROUNDWATER N/A 1662.49 394857.31 0.00 POINT SOURCE N/A 0.00 0.00 0.00 SEPTIC SYSTEMS N/A 107.14 16339.39 0.00 TOTAL 27850.65 25180.50 514116.02 26448440.04 The unit area load for each pollutant in each watershed was estimated by dividing the mean annual loading (lbs/year) by the total area (acres). Results for the unit area loadings can be found in Table 8. Unit area loads for phosphorus, nitrogen, and sediment in Conowingo Creek watershed are 2.21 lbs/acre/yr, 27.83 lbs/acre/yr, and 1,151.55 lbs/acre/yr respectively. Unit area loads for phosphorus, nitrogen, and sediment in the North Branch Muddy Creek Watershed are 0.90 lbs/acre/yr, 18.46 lbs/acre/yr, and 949.66 lbs/acre/yr respectively.

Table 8. Unit Area Loads for the Conowingo and North Branch Muddy Creek Watersheds

Watershed Unit area load for P (lbs/acre/yr) Unit area load for N

(lbs/acre/yr) Unit area load for Sediment

(lbs/acre/yr) Conowingo 2.21 27.83 1151.55 North Branch Muddy Creek 0.90 18.46 949.66

TMDL COMPUTATIONS FOR PHOSPHORUS AND SEDIMENT The TMDLs established for Conowingo Creek consist of a load allocation (LA) and a margin of safety (MOS) for phosphorus, nitrogen, and sediment. The wasteload allocation (WLA) accounts for point source discharges located within the watershed. No point sources are located within the Conowingo Creek Watershed, so a WLA is not necessary for computing the TMDL. TMDL Computation The load reduction calculations in Conowingo Creek are based on the current loading rates for phosphorus, nitrogen, and sediment in the North Branch Muddy Creek, the reference watershed. Based on biological assessment, North Branch Muddy Creek is attaining its aquatic life uses. North Branch Muddy Creek is designated as a cold water fishery (CWF). The nutrient and sediment unit area loading rates were computed for North Branch Muddy Creek using the

- 18 -

AVGWLF model (Table 8). These unit-area loading rates were then used as the basis for establishing the TMDLs for Conowingo Creek. The TMDL value for each pollutant was determined by multiplying the unit area loading rates for North branch muddy Creek by the total watershed area of Conowingo Creek. Table 9 presents this information. Table 10 shows the allocations made for each land use.

Table 9. TMDL Computation for Conowingo Creek

Pollutant Unit Area Loading Rate in North branch muddy

Creek (lbs/acre/yr)

Total Watershed Area for Conowingo Creek

(acres)

TMDL Value (lbs/yr)

Phosphorus 0.90 23708.34 21337.51 Sediment 949.66 23708.34 22514862.16

Table 10. TMDL Allocations for Conowingo Creek

Phosphorus Sediment

Land Use Area (ac)

Unit Area Load (lbs/ac/yr)

Total (lbs/yr)

LA (lbs/yr)

% Reduction

Unit Area Load

(lbs/ac/yr)

Load (lbs/yr) LA (lbs/yr)

% Reduction

Hay/Past 8212.87 0.42 3428.81 1863 45 207.99 1708190.44 1173212 31% Cropland 9947.29 4.77 47478.49 15889 66 2556.41 25528849.13 19027942 25% Conif_For 542.41 0.01 4.63 4.63 0 4.12 2235.13 2235.13 0% Mixed_For 481.91 0.01 3.75 3.75 0 3.86 1862.61 1862.61 0% Decid_For 4447.81 0.01 48.49 48.49 0 7.45 33154.41 33154.41 0% Developed 76.05 1.26 95.91 80 17 356.50 27112.17 24970 8% Ground-water N/A -- 1314.6 1314.6 0 -- -- -- --

Point Source N/A -- -- -- -- -- -- -- --

TOTAL 23708.34 2.21 52374.68 19203.47 63 1151.55 27301403.89 20263376.15 26% For the purpose of allocating loads in an impaired stream segment, the TMDL equation is as follows: TMDL = WLA + LA + MOS The WLA (wasteload allocation) portion of this equation is the total loading that is assigned to point sources. The LA (load allocation) is the portion of this equation that is assigned to nonpoint sources. The MOS (margin of safety) is the portion of loading that is reserved to account for any uncertainty in the data and computational methodology used for the analysis, represented by 10% of the TMDL value. Table 11 presents the TMDLs for Conowingo Creek.

Table 11. TMDLs for Conowingo Creek

Pollutant TMDL (lbs/yr) WLA (lbs/yr) LA (lbs/yr) MOS (lbs/yr)

- 19 -

Phosphorus 21338 -- 19204 2134 Sediment 22514862 -- 20263376 2251486

The individual components of the TMDLs are discussed in detail below. Margin of Safety The Margin of Safety (MOS) for this analysis is explicit. Ten percent of each of the TMDLs was reserved as the MOS. Using ten percent of the TMDL load is based on professional judgement and will provide an additional level of protection to the uses of the waterbody.

Conowingo Creek Phosphorus – 21338 x 0.1 = 2134 lbs/year

Sediment - 22514862 x 0.1 = 2251486 lbs/year

- 20 -

Load Allocation The load allocation (LA) for each subbasin was computed by subtracting the margin of safety value from the TMDL value. Individual load allocations were then assigned to land uses/sources that are shown in Table 12. Not all land use/source categories were included in the allocation because they are difficult to control, or they provide an insignificant portion of the total load. Loading values for land uses/ sources that were not part of the allocation were carried through at their existing loading value. Observations made in the field showed significant runoff originating from both agricultural land and residential/urban development. Since best management practices (BMPs) such as riparian buffers would not discriminate between reductions in either nutrients or sediment, land uses associated with these activities were included in the reduction scenario. Jackson Run, although unimpaired, was included in the allocation table since it flows directly into an impaired portion of the Conowingo Creek. Phosphorus The MOS was subtracted from the TMDL value.

Conowingo Creek LA = 21338 (TMDL) – 2134 (MOS)

LA = 19204 lbs/yr Since the impairments are believed to be primarily caused by agricultural activities and runoff from developed areas, only the loads associated with these land uses (HAY/PAST, ROW_CROPS, and DEVELOPED) were considered in the reduction scenario. The remaining loads were subtracted from the LA value. Conowingo Creek

Adjusted LA = 19204 – 1380 Adjusted LA = 17824 lbs/yr

The allocable load is the portion of the load that is available to allocate among the contributing sources.

It is important that the TMDL target load for each segment be achievable. For this reason, the subbasins were further divided into allocation units. Conowingo Creek has six allocation units. These allocation units provide a specific target load for tributaries within the unit boundaries. The unit area loading rates determined by the model were used to calculate the load allocations based on the land use distribution within each allocation unit. The following section discusses the load allocation process in detail.

- 21 -

EMPR is carried out in the following manner. Each land use/source load is compared with the allocable load to determine if any contributor would exceed the allocable load by itself. The evaluation is carried out as if each source is the only contributor to the pollutant load entering the receiving waterbody. If the contributor exceeds the allocable load, that contributor is reduced to the allocable load. This is the baseline portion of EMPR. After any necessary reductions have been made in the baseline, the multiple analyses are run. The multiple analyses will sum all of the baseline loads and compare them to the allocable load. If the allocable load is exceeded, an equal percent reduction will be made to all contributors’ baseline values. After any necessary reductions in the multiple analyses, the final reduction percentage for each contributor can be computed. A detailed description of the EMPR method can be found in Appendix D.

The results of the Load Allocations for Conowingo Creek are presented in Table 12. The load allocation for each land use is shown along with the percent reduction necessary for each source. The impaired segment, as listed on Pennsylvania’s 303(d) list, can be matched with the allocation unit using Table 2 and Figure 2.

- 22 -

Table 12. Conowingo Creek Load Allocation for Phosphorus by Land Use/Source

Land Use Acres Current

Loading Rate (lbs/acre/yr)

Allowable Loading Rate (lbs/acre/yr)

Current Load(lbs/yr)

Load Allocation

(lbs/yr)

Percent Reduction

Allocation Unit 1 – Conowingo Creek and Unnamed Tributaries Hay/Past 1625.45 0.42 0.23 682.69 373.85 45Row Crops 1752.66 4.77 1.60 8360.19 2804.26 66Developed 4.45 1.26 1.05 5.61 4.67 17

Allocation Unit 2 – Conowingo Creek Hay/Past 1953.03 0.42 0.23 820.27 449.20 45Row Crops 2764.53 4.77 1.60 13186.81 4423.25 66Developed 8.90 1.26 1.05 11.21 9.35 17

Allocation Unit 3 – Little Conowingo Creek Hay/Past 1816.04 0.42 0.23 762.74 417.69 45Row Crops 1901.66 4.77 1.60 9070.92 3042.66 66Developed 1.55 1.26 1.05 1.95 1.63 17

Allocation Unit 4 – Jackson Run Hay/Past 541.74 0.42 0.23 227.53 124.60 45Row Crops 400.75 4.77 1.60 1911.58 641.20 66Developed 4.45 1.26 1.05 5.61 4.67 17


Allocation Unit 6 – Conowingo Creek Hay/Past 1150.87 0.42 0.23 483.37 264.70 45Row Crops 1589.42 4.77 1.60 7581.53 2543.07 66Developed 38.69 1.26 1.05 48.75 40.62 17 Sediment 1. The margin of safety value was subtracted from the TMDL value. This quantity represents

the load allocation (LA). Conowingo Creek

LA = 22,514,862 (TMDL) – 2,251,486 (MOS) lbs/year LA = 20,263,376 lbs/year

- 23 -

2. Again, only loads associated with agricultural activities or nonpoint urban runoff were considered in the reduction scenario. The remaining loads were subtracted from the LA value.

Conowingo Creek

Adjusted LA = 20,263,376 – 37,252 Adjusted LA = 20,226,124 lbs/year

The allocable load is the portion of the load that is available to allocate among the contributing sources.

3. This quantity was allocated among the four remaining land use/sources. The allocation

method used was Equal Marginal Percent Reduction (EMPR). The allocation method is discussed in the phosphorus section.

4. The results of the load allocations for Conowingo Creek are presented in Table 13. The load allocation for each land use is shown along with the percent reduction necessary for each source. The impaired segment, as listed on Pennsylvania’s 303(d) list, can be matched with the allocation unit using Table 2 and Figure 2.

- 24 -

. Table 13. Conowingo Creek Load Allocation for Sediment by Land Use/Source

Land Use Acres Current

Loading Rate (lbs/acre/yr)

Allowable Loading Rate (lbs/acre/yr)

Current Load (lbs/yr)

Load Allocation

(lbs/yr)

Percent Reduction


Allocation Unit 2 – Conowingo Creek Hay/Past 1953.00 207.99 142.85 406204.5 278986.9 31Row Crops 2764.50 2566.41 1912.81 7094840.4 5287963.9 25Developed 8.90 356.50 328.33 3172.9 2922.2 8

Allocation Unit 3 – Little Conowingo Creek Hay/Past 1816.00 207.99 142.85 377709.8 259416.4 31Row Crops 1901.70 2566.41 1912.81 4880541.9 3637591.2 25Developed 1.60 356.50 328.33 570.4 525.3 8

Allocation Unit 4 – Jackson Run Hay/Past 541.70 207.99 142.85 112668.2 77382.1 31Row Crops 400.80 2566.41 1912.81 1028617.1 766654.3 25Developed 4.50 356.50 328.33 1604.3 1477.5 8


Allocation Unit 6 – Conowingo Creek Hay/Past 1150.90 207.99 142.85 239375.7 164406.6 31Row Crops 1589.40 2566.41 1912.81 4079052.1 3040220.6 25Developed 38.70 356.50 328.33 13796.6 12706.5 8 CONSIDERATION OF CRITICAL CONDITIONS The AVGWLF model is a continuous simulation model that uses daily time steps for weather data and water balance calculations. Sediment and nutrient loads are calculated monthly, based on a daily water balance. Therefore, all flow conditions are taken into account for loading calculations. Because there is generally a significant lag time between the introduction of sediment and nutrients to a waterbody and the resulting impact on beneficial uses, establishing these TMDLs using average annual conditions is protective of the waterbody.

- 25 -

CONSIDERATION OF SEASONAL VARIATIONS The continuous simulation model used for this analysis considers seasonal variation through a number of mechanisms. Daily time steps are used for weather data and water balance calculations. The model requires specification of the growing season, and hours of daylight for each month. The model also considers the months of the year when manure is applied to the land. RECOMMENDATIONS

The pollutant reductions in the TMDLs are allocated to agricultural and residential/urban development activities in the watershed. Implementation of best management practices (BMPs) in the affected areas should achieve the loading reduction goals established in the TMDLs. Substantial reductions in the amount of sediment reaching the streams can be made through the planting of riparian buffer zones, contour strips, and cover crops. These BMPs range in efficiency from 20% to 70% for sediment reduction. Implementation of BMPs aimed at sediment reduction will also assist in the reduction of phosphorus. Additional phosphorus reductions can be achieved through the installation of more effective animal waste management systems and stone ford cattle crossings. Other possibilities for attaining the desired reductions in phosphorus and sediment include streambank stabilization and fencing. Further ground truthing will be performed in order to assess both the extent of existing BMPs, and to determine the most cost-effective and environmentally protective combination of BMPs required to meet the nutrient and sediment reductions outlined in this report. Collaborative efforts between several state, federal, and local agencies have identified segments for implementation of BMPs. However, no funding has been allocated to the Conowingo Creek Watershed (WRAS, 1999). Efforts could be improved through the establishment of a watershed association. PUBLIC PARTICIPATION A public meeting to discuss and accept comments on proposed TMDLs was held on January 25, 2001 beginning at 7:00 p.m., in the main auditorium of the Farm and Home Center in Lancaster . Public notice of the draft TMDL and the public meeting was published in the Pennsylvania Bulletin and the Lancaster Intelligencer. Notice of final plan approval will be published in the Pennsylvania Bulletin.

- 26 -

LITERATURE CITED

Haith, D.A., and L.L. Shoemaker, 1987. Generalized Watershed Loading Functions for Stream Flow Nutrients. Water Resources Bulletin, 23(3), pp. 471-478. Nizeyimana, E., B.M. Evans, M.C. Anderson, G.W. Petersen, D.R. DeWalle, W.E. Sharpe, J.M. Hamlett and B.R. Swistock, 1997. Quantification of NPS Pollution Loads within Pennsylvania. Environmental Resources Research Institute, The Pennsylvania State University, Pub. No. ER97-08, 61 pp. Novotny, V., and H. Olem, 1994. Water Quality: Prevention, Identification, and Management of Diffuse Pollution. Van Nostrand Reinhold, New York.

Samuels, W., 1998. Case Studies: Solving Watershed-Based Problems Through the Use of GIS, Internet and EPA National Data Bases. In: Watershed Management: Moving from Theory to Implementation, Water Environment Federation, Denver, Colorado, pp. 1175-1182. Thomann, R.V., and J.A. Mueller, 1987. Principles of Surface Water Quality Modeling and Control. Harper & Row, New York. Watershed Restoration Action Strategy (WRAS) for Subbasin 07K, 1999. Pennsylvania Department of Environmental Protection, Bureau of Watershed Conservation, 15 p.

CONOWINGO CREEK TMDL COMMENTS

General Comment: TMDL proposals should cite the specific use of water that is impaired,

thereby requiring the TMDL’s development. This is especially the case when the impairing pollutants have no numerical water quality criteria. In the case of Trindle Spring Run and Conowingo Creek, these streams are considered impaired by sediment and nutrients when, in a separate Department of Environmental Protection action, they are acknowledged as supporting Class A Wild Trout Fisheries. If the impairment is not sufficient to prevent the stream from supporting the top functional fishery classification, is the use of the stream as a Cold Water Fishery really impaired? (3)

Response: It is possible for a stream to have a very good fish population and at the

same time demonstrate impairments to other aquatic life, water supply, or recreational uses. Most of the entries on the PA 303(d) list are the result of biological surveys conducted as part of the Department’s Unassessed Waters Program. A rapid biological screening protocol is used to evaluate numerous aspects of in-stream or riparian physical habitat and macroinvertebrate community structure, and make a determination of attainment or non-attainment of water quality standards.

The section of Trindle Spring Run that was added recently to the list of

Class A wild trout streams is currently on the 303(d) list because of impairment from priority organics. This will result in a fish consumption advisory for that stream segment. The un-named tributary and upper section of the main stem of Trindle Spring Run as well as Conowingo Creek are currently on the 303(d) list for impairment related to sediment and nutrients based on physical habitat and benthic macroinvertebrate community impacts.

Comment: Please clarify how the 1996 303(d) listing for 14.44 miles of the Conowingo

Creek is represented in Table 2. Conowingo Creek can be found on Pennsylvania’s section 303(d) list of

impaired waters as stream code #7162 listed for impairments from nutrients and Turbidity/suspended solids due to agricultural sources. On the 1998 303(d) listing for the Conowingo, the segment code of 6326 was added. Table 2 shows 3 seemingly separate listings for 1996. Although all have a segment code of 6326, there are 3 different stream codes assigned # 7162, #7171 and #7172. (2)

Response: The additional codes under segment 6326 were listed in 1998. The table

was corrected to reflect that information. Reference Watershed Section Comment: Please explain why having significantly less agricultural lands in the

reference watershed than in the impaired Conowingo watershed is protective of the Conowingo watershed.

This section states that the reference watershed, the North Branch Muddy

Creek, has significantly less agricultural lands but that the difference is protective of the Conowingo watershed. (2)

Response: The statement was removed to prevent confusion. Comment: PBA is very concerned about the use of reference watersheds to establish

TMDLs. Under federal requirements, loading capacity represents the maximum concentration of a pollutant at which a stream can remain in attainment of water quality standards. A TMDL should equal loading capacity plus a quantitative margin of safety. In establishing the TMDL for the streams in question, DEP fails to establish their respective loading capacities. Additionally, the specific selection of reference streams seems to be significantly flawed. Warm Water Fisheries are referenced against Cold Water Fisheries (Mains and Gum Runs compared to Griers Hollow, et.al. as well as Pequea and Chickies Creeks compares to Conococheague Creek). Of even greater concern is the issue that numerous impaired streams are compared against High Quality (HQ) streams (Yellow Breeches Creek, Letort Spring Run). Since a HQ stream (Pennsylvania equivalent to federal Tier II) represents a condition where ambient water quality exceeds the water quality necessary to support existing uses, the use of HQ streams as referenced for non-HQ streams will result in a TMLD that is overly restrictive. Finally, the ad-hoc subdivision of a watershed for use as a reference is highly subjective, and to PBA’s understanding, is not supported by any forma scientific review. (3)

Response: In order to establish a loading capacity for an impaired stream segment

where no numerical water quality criteria exist, Pennsylvania has developed a reference watershed approach. The allowable loading rate for an impaired stream is established by evaluating the loading rate of a non-impaired watershed selected based on matching the land use distribution, surface geology, and size of the impaired watershed. The modeling methods used for these analyses are sensitive to land use characteristics, geology, known

nutrient soil concentrations, rainfall and drainage area. A good match for a reference watershed based on these characteristics over-rides concerns about matching use classifications of the streams in making our selections. The important common feature of the reference watersheds is that their biological communities are unimpaired. The reduction in loads projected in the TMDL should, therefore, restore the biological condition of the impaired water to an unimpaired level. However, the degree of recovery will be controlled, and in some cases limited, by numerous physical habitat issues. Impaired, non-HQ or EV waters, will not be expected to “recover” to antidegradation levels as the result of TMDL implementation.

As far as selecting portions of a watershed to use as a reference the

following rule was applied; only upstream headwater stream segments could be cut out for the purpose of a reference (this means that no downstream impaired segment could be cut out, also that no portion of the reference watershed should drain into any section where an impairment is present). There could be exceptions to this practice, however, there must be very good justification in order to deviate from the rule.

Comment: Generally, the reference watershed selected is appropriate. However, the

draft TMDL fails to discuss whether the differences in agricultural practices and condition of riparian buffers and streambanks reasonably account for North Branch Muddy Creek (reference watershed) meeting water quality standards while Conowingo Creek Watershed does not, or whether the difference in groundwater nitrogen concentration between the two watersheds also plays a significant role. (1)

Response: North Branch Muddy Creek meets water quality standards based on

Pennsylvania’s biological assessment standards, while Conowingo Creek does not. The 1996 and 1998 303(d) lists specifically identify agriculture as the source of use impairment in the Conowingo Creek watershed. Nutrients and siltation are identified as the causes of impairments. The North Branch Muddy Creek watershed used in this report has been assessed using the Department’s Unassessed Waters Program protocol and was determined to be supporting its designated uses. These two watersheds are a fairly close match in terms of land cover/land use, physiographic province, geology, and size. The extent of agricultural land uses in each watershed is similar. The primary difference between the two watersheds at the time of the section 303(d) listing of Conowingo Creek was the presence of natural riparian buffers in the North Branch Muddy Creek watershed. As for groundwater nitrogen concentrations, water-quality data collected by the US Geological Survey from wells within York and Lancaster Counties indicate that nitrogen concentrations are not significantly different.

Methods Comment: The PBA appreciates the efforts of the Department in translating narrative

water quality criteria into a quantitative TMDL. Further, PBA conceptually approves of the modeling techniques used to develop the TMDL. PBA further recommends that the Department consider developing numeric water quality criteria for phosphorus and sediment. (3)

Response: We are currently working with EPA to develop nutrient criteria. Comment: Please include a Table illustrating the total watershed, including landuse

types by acreage and the reductions assigned to each landuse, as is found in similar TMDLs

The TMDL includes a Table by sub-watershed allocation units but does not

include a Table illustrating the watershed and the landuse types before it was broken into allocation units to assigned reductions. (2)

Response: The requested table was added to the document. Comments: The proposed TMDL fails to establish total maximum daily loads. It

establishes only a yearly limit, whereas the Clean Water Act requires total maximum daily loads. DEP has not explained why setting a yearly limit, which presumably allows for daily, weekly, or monthly fluctuations in loads as long as the yearly totals are not exceeded, adequately protect water quality on a daily basis. Congress clearly intended that water quality standards be met every day, not just most days or on an annual basis. In addition to failing to meet statutory requirements, setting only annual loads is inadequate for performance monitoring and regulatory enforcement. For these purposes, daily loadings and streamflows should be calculated for one or several critical or frequently encountered seasonal weather conditions. Such daily loading and streamflow values could be easily extracted from mass and water balance calculations already performed internally by ArcView Version of the Generalized Watershed Loading Function (AVGWLF). They would be more readily useful measures for monitoring of loads and enforcement of the TMDL. (1)

Response: The CWA requirement for total maximum daily loads allows for the

expression of a TMDL in units of mass per time, toxicity, or other appropriate measures. DEP in consultation with EPA has determined that annual loadings are more appropriate for expression of nonpoint source TMDLs for nutrients and sediment.

Comment: The TMDL fails to establish nitrogen loading limits without sufficient

justification for not doing so. The TMDL states that the calculated ratio of N:P is greater than 10:1, and therefore Phosphorus is assumed to be the limiting nutrient. However, the N:P ratio was calculated from the yearly

ratio of nitrogen to phosphorus loading. It can be expected that this ratio will vary with seasonal differences in precipitation and surface runoff and contributions of various land uses to the nutrient concentrations in the water, a fact which the TMDL ignores.

Groundwater is a significant source of nitrogen pollution in the watershed.

As a result, during dry periods when groundwater is the predominant source of streamflow and nutrient loading, the N:P ratio will be high and phosphorus can be assumed to be the rate-limiting nutrient. In contrast, nutrient loading from surface sources such as hay/pasture and cropland will occur primarily during storm events or wet weather periods. Nutrient contributions from groundwater during these high-flow periods are likely to decline, and the N:P ratio will decline accordingly. If the N:P ratio becomes low enough, nitrogen will become the rate-limiting nutrient rather than phosphorus.

Simple calculations based on the yearly loading data illustrate how

misleading it is to base the calculation of the N to P ratio on the total yearly loading from all sources. If the N:P ratio is calculated from total yearly loads including groundwater as a source, the ratio is about 13:1, above the 10:1 threshold on which DEP bases its limiting nutrient determination. However, if the N:P ratio is calculated from nutrient loads from surface nutrient sources alone, the ratio becomes significantly lower: approximately 2:1. Because the latter ratio is below 10:1, nitrogen may be the rate-limiting nutrient rather than phosphorus when surface runoff becomes the major source of streamflow and nutrient loading.

Because of the uncertainty over which nutrient is limiting at any given time

during the year, the Commentor strongly urges DEP to establish a nitrogen TMDL in addition to the phosphorus TMDL. (1)

Response: Although ground water contributions of N will be highest relative to

overland runoff contributions during the summer months (May through September), total nitrogen loads will normally be lowest in these months due to low flows and increased plant uptake. Phosphorus does enter the stream through overland flow (runoff); however, periods of high P exports correspond to periods of high soil loss. During the wet winter months, there is normally enough ground cover to dissipate the erosive energy of precipitation. Total P loads, on a unit area basis, are typically highest in the fall (after harvest when more bare soil is exposed) and in the spring (more intense rainfall events on fields being prepared for planting). However, TN loads are also higher in the fall and spring such that the N:P ratio remains greater than 10.

Comment: The TMDL fails to meet the Clean Water Act requirement for establishing a

maximum daily load for impaired waters that reflects seasonal variations.

First, the draft TMDL fails to specify whether water quality is impaired all year, or whether water quality is only impaired during particular time periods. While the document specifies that aquatic biological surveys were conducted to determine whether Conowingo Creek Watershed is impaired, it does not indicate whether this assessment method accounts for seasonal variations in water quality. In addition, the draft TMDL indicates that the reference watershed, North Branch Muddy Creek, is not impaired for nutrients or sediment, but does not specify what testing was done to reach this conclusion and whether this testing accounted for seasonal variations. To be an appropriate reference watershed, North Branch Muddy Creek must meet water quality standards throughout the year.

Second, DEP asserts that the yearly TMDL load accounts for seasonal

variations because the models use daily time steps for weather data and water balance calculations, and consider growing seasons, hours of daylight, and the months when manure is applied to the land. While DEP appropriately used this seasonal data to estimate the total yearly loads in the subject and reference watersheds, DEP failed to use the data to calculate a daily TMDL that assures water quality standards are met in all seasons of the year. The CWA requires the TMDL to be established “at a level necessary to implement the applicable water quality standards with seasonal variations.”33 U.S.C. §1313(d)(C). If water quality impairment varies due to seasonal differences in weather and agricultural activity, then TMDLs should be established for each relevant season or time period.

Third, the TMDL fails to account for seasonal variations in the N:P ratio, as

is discussed more fully in the previous subsection of these comments. As a result, the TMDL inappropriately leaves out limits on nitrogen loading. (1)

Response: TMDLs for nonpoint sources of pollution are developed to protect the

stream from impacts that occur at “critical” conditions. Critical conditions for nonpoint sources are times of runoff usually associated with precipitation. Similar to the way TMDLs protect waters from point source pollution at the critical low flow condition ensures protection at other less critical periods, TMDLs developed to protect the stream from impact of nonpoint sources during runoff ensure protection under all other conditions.

Comment: DEP has made a reasonable allocation of the loads among non-point sources

in the watershed. The Commentor commends DEP for making this allocation in the TMDL, as the TMDLs established by other states often fail to do so. (1)

Response: Thank you for the supportive comment.

Comment: DEP fails to provide a rationale for selecting 10% as the margin of safety. The margin of safety should be based on the inherent uncertainty of the models used rather than the undefined “best professional judgment.” (1)

Response: The margin of safety used in these TMDLs does take into consideration the

inherent uncertainty of the AVGWLF model. The “best professional judgment” referred to in the report includes information from those individuals who developed, calibrated, and currently maintain the AVGWLF model. Inclusion of the 10% margin of safety provides an additional level of protection to designated uses.

Comment: Please clearly explain the reasons for the difference in the values found in

the Tables compared to attaining those values through the calculations described

Several values used to determine phosphorus and sediment loading in the

Conowingo Creek seemed to have been averaged or rounded. Although these values vary only slightly, the differences compound as the number is carried through several calculations. Additionally, the acreages vary from Table to Table. The reasons for the variation in these values should be clearly stated in the TMDL. We have listed some examples of the differences found in the TMDL values below.

For example, it is explained that the TMDL value in for each pollutant was

determined by multiplying the loading rate by the acres in the watershed. The values shown in Table 9 however are not those determined directly through multiplication. Table 9 shows a TMDL value of 19,400lbs/yr phosphorus, when multiplied directly the value is 19,350lbs/yr. For sediment, Table 9 shows the value of 20,400,000lbs/yr. When multiplied directly the TMDL value is 20,425,000lbs/yr for sediment.

From this value the Margin of Safety is determined and then subtracted to

get the load allocation. Again, the direct calculations are not the same as the values shown in Table 10 (19,400lbs/yr-1,940lbs/yr = LA of 17,500lbs/yr vs. 17,460lbs/yr for a direct subtraction for phosphorus). After subtracting the loads not reduced, the adjusted load allocation for phosphorus stated in the TMDL is 16,100lbs/yr. If calculated directly it is 15,985lbs/yr. This is not a large difference but it is larger than some of the allocations made to low and high intensity development. The same is true for sediment, where for example, in the EMPR computations (Page 21)the LA of 18,400,000lb/yr minus the LNR of 37,940lbs/yr equals an adjusted LA of 18,400,000lb/yr of sediment, the same LA value before the LNR was subtracted.

The allocations units for phosphorus by landuse shown in Table 11, can be

added to equal 15,952.5lbs/yr. Not the 16,100lbs/yr determined in the

TMDL, but much closer to the 15,985lbs/yr determined by directly calculating the numbers.

Table 11 also denotes the acreage of cropland. If the acres in each of the

sub-basins are added they result in 9947.4 acres of cropland, Table 6 states there is 9,800 acres of cropland a difference of 147 acres. Also, the acreage of low and high intensity development is different than denoted in Table 6. The acreage of high intensity development in Table 11 is 25.6acres compared to the 40 acres stated in Table 6, a difference of 36%. (2)

Response: The values were re-calculated without rounding/averaging. The

discrepancies have been removed by doing so. Comment: If the difference in TMDL values and acreage (discussed above) are due to

rounding, then the load reductions assigned to low and high intensity development must be re-evaluated.

If the margin of error in determining the total LA is larger than the

reductions assigned to the low and high intensity development landuse types, these reductions must also be questioned. These landuse types are very small in comparison to the watershed size, and it may be more accurate for the model to combine the ‘low and high’ intensity development into one ‘developed land’ category. (2)

Response: Nutrient loads from low/high intensity development and septic systems were

lumped together under the category “developed”. Comment: EPA has several concerns regarding the 93% sediment reduction from row

crops. These concerns, discussed below are (1) explanation of the modification made in the model to the ‘c’ and ‘p’ factors to adjust for cropland (2) the large difference between the required reductions from Hay and Pasture land and those assigned to row crops and (3) the reasonable assurance that the TMDL can be implemented when a 93% reduction is called for from an agricultural sources.

Please explain the reasons to adjusted the ‘c’ and ‘p’ factors to the numbers

used. In the reference watershed ‘p’ is reduced to account for contour farming

and the presence of natural riparian buffers. The ‘c’ factor, which accounts for crop type and cultivation practices, is also reduced. In the impaired Conowingo Creek both of these factors are then increased to adjust for these factors, or lack there of.

The differences in cropland are adjusted for factors in both watersheds. EPA is concerned these adjustments may have been over compensated for, resulting in a higher than needed reduction requirement for cropland, especially when compared to the other contributing landuse sources which require only a 9% reduction. EPA believes that although the values themselves seem to be in a reasonable range, the reasons to adjust the ‘p’ and ‘c’ factors to these values for both watersheds must be further explained. Default values should be used unless these values can be explained.

Please discuss any factors including watershed characteristics, soil erosion,

livestock density and animal farming practices that may have contributed to the large differences between the required sediment reductions assigned to row-crops and those assigned to pasture land.

The reasonable assurance that the TMDL can be implemented must also be

questioned when a 93% reduction is called for from an agricultural source. The TMDL states that Best Management Practices (BMPs) range in efficiency from 20-70% removal rates for sediment reduction. It would seem even with all BMPs in place, the reductions would still fall short of the TMDL requirement from cropland sources. (2)

Response: There was an error made entering a loading value into the EMPR

spreadsheet. This error produced the extremely large (and incorrect) reduction needed for “cropland”. The tables in the document now reflect the correct reductions needed for the agricultural land uses.

Recommendations/BMPs Comment: From the limited information presented on the effectiveness of possible

implementation plans, the Commentor does not believe the TMDL provides reasonable assurance that the required reductions in phosphorus and sediment loadings will be met. In addition, the document does not specify if Best Management Practice (BMP) implementation is planned for the whole watershed or just for impaired areas, and the document fails to consider the expected BMP compliance rate for landowners. Another concern is that no funding has been allocated to the watershed for BMP implementation. Without funding or adequate support from the state, there can be no reasonable assurance the TMDL will be met. (1)

Response: TMDLs developed under section 303(d) of the CWA are not intended to be

a step-by-step description of how to restore an impaired watershed. Federal law requires establishment of a pollutant load that will ensure attainment of water quality standards and an allocation of that load among point and nonpoint sources. These TMDLs have established pollutant loads, along

with allocations of those loads, which will ensure attainment of water quality standards. Implementation plans, including assurances of specified load reductions, are not currently required as part of the TMDL under section 303(d). Information on potential remediation activities, including BMPs, was provided as an indication that the identified load reductions were achievable. The information should prove helpful to those developing plans to meet the specified reductions. While the Department insures compliance with all applicable laws and regulations, the most effective and achievable means of meeting the goals set forth in these TMDLs will come from the local level. The Department will also provide organizational, technical and financial assistance to watershed groups who undertake implementation. Please contact the Department if you want further information.

Comment: The Commentor also urges DEP to make appropriate use of the AVGWLF

model in developing site-specific implementation plans. The AVGWLF model could be a powerful tool for selection of management practices to achieve the watershed loading objectives. Various reduced combinations of values for AVGWLF’s crop management factor (C) and conservation practices factor (P) could be used in the model to simulate the watershed loading objectives for Conowingo Creek Watershed. For each acceptable combination of C-factor and P-factor values, there will be an associated array of site-specific crop management and conservation practices. These practices are tabulated for different crops, pasture, and woodland in the GWLF Users Manual. Thus, AVGWLF could generate a sophisticated array of acceptable, location-specific management practices from which land owners/managers could select. (1)

Response: The model does not allow for the direct input of BMPs on the landscape to

predict reduction values. GWLF is a lumped parameter model; therefore, assessment of reductions due to BMP implementation in impaired sub-watersheds must be done external to the model. The Department will provide information and documentation on the AVGWLF model upon request.

Comment: The Commentor also recommends that a monitoring program be part of any

implementation plans to determine if the BMPs are having their intended effect on water quality or if other remedial measures are required. (1)

Response: The Department agrees with this comment. The Department will continue

to assess water quality and designated use attainment in the Conowingo Creek watershed through its ongoing assessment activities.

Comment: Any remediation measures to address identified water quality problems may benefit certain threatened and endangered species in certain watersheds by improving water quality. However, in some instances, these measures have the potential to adversely affect federally listed species; therefore, further consultation will be necessary to identify and address these cases as described above. (4)

Response: Detailed remediation and implementation plans are not required as part of the TMDL submittal and have not been completed at this time. All current regulations will be followed and threatened and endangered species will be protected in developing a remediation plan for the watershed.

LIST OF COMMENTORS

1. James M. Stuhltrager Susan D. Mack Mid-Atlantic Environmental Law Center c /o Widener University School of Law 4601 Concord Pike P.O. Box 7474 Wilmington, DE 19803

2. U.S. Environmental Protection Agency Region III Headquarters Philadelphia, PA

3. Mark Maurer Assistant Director of Government Affairs PA Home Builders Association Lemoyne, PA

4. U.S. Fish & Wildlife Service

EXECUTIVE SUMMARYTable 1. TMDL Endpoints for the Conowingo Creek Watershed

INTRODUCTIONTMDL ENDPOINTSNutrient Loads and Organic Enrichment in Stream Systems

SELECTION OF THE REFERENCE WATERSHEDDATA COMPILATION AND MODEL OVERVIEWGIS BASED DERIVATION OF INPUT DATAWATERSHED ASSESSMENT AND MODELINGTMDL COMPUTATIONS FOR PHOSPHORUS AND SEDIMENTTMDL ComputationMargin of SafetyLoad AllocationPhosphorusSediment

CONSIDERATION OF CRITICAL CONDITIONSCONSIDERATION OF SEASONAL VARIATIONSRECOMMENDATIONSPUBLIC PARTICIPATIONLITERATURE CITEDconowingo_cr.pdfReference Watershed SectionMethods

Recommendations/BMPsComment: Any remediation measures to address identified water quality problems may benefit certain threatened and endangered species in certain watersheds by improving water quality. However, in some instances, these measures have the potential to advResponse: Detailed remediation and implementation plans are not required as part of the TMDL submittal and have not been completed at this time. All current regulations will be followed and threatened and endangered species will be protected in develoLIST OF COMMENTORS

total maximum daily loads (tmdls) conowingo creek ... · tmdl endpoints established for this...

Documents