MILL RUN WATERSHED TMDL Blair and Cambria Counties

Prepared for:

Pennsylvania Department of Environmental Protection

January 12, 2007

TABLE OF CONTENTS

INTRODUCTION ...... 1 LOCATION ...... 1 SEGMENTS ADDRESSED IN THIS TMDL ...... 3 CLEAN WATER ACT REQUIREMENTS ...... 3 SECTION 303(D) LISTING PROCESS ...... 4 BASIC STEPS FOR DETERMINING A TMDL ...... 5 SAMPLING RATIONALE ...... 5 WATERSHED BACKGROUND...... 6 METALS METHODOLOGY ...... 6 TMDL ENDPOINTS...... 8 TMDL ELEMENTS (WLA, LA, MOS)...... 9 TMDL METALS ALLOCATION SUMMARY...... 9 MILL RUN SEGMENT WATERSHED SILTATION TMDL...... 10 Reference Watershed Approach ...... 13 Selection of the Reference Watershed...... 13 WATERSHED ASSESSMENT AND MODELING ...... 16 TMDLS...... 17 Background Pollutant Conditions...... 17 Targeted TMDLs ...... 17 Margin of Safety ...... 18 Load Allocation ...... 18 Adjusted Load Allocation...... 18 TMDL ...... 19 CALCULATION OF SEDIMENT LOAD REDUCTIONS ...... 19 CONSIDERATION OF CRITICAL CONDITIONS ...... 20 CONSIDERATION OF SEASONAL VARIATIONS...... 20 RECOMMENDATIONS FOR IMPLEMENTATION (METALS AND SEDIMENT)...... 20 PUBLIC PARTICIPATION ...... 21 GIS DATA SETS...... 24 REFERENCES ...... 32

TABLES

Table 1. Mill Run Segments Addressed...... 1 Table 2. Applicable Water Quality Criteria...... 9 Table 3. Summary Table–Mill Run Watershed ...... 9 Table 4. Comparison Between Mill Run Segment and Racket Brook...... 14 Table 5. Existing Sediment Loads for Mill Run Segment ...... 16 Table 6. Existing Sediment Loads for Racket Brook...... 17 Table 7. Targeted TMDL for the Mill Run Segment Watershed...... 18 Table 8. Load Allocations, Loads Not Reduced, and Adjusted Load Allocations for Mill Run Segment...... 19 Table 9. TMDL, MOS, LA, LNR, and ALA for Mill Run Segment...... 19

Table 10. Sediment Load Allocations & Reductions for Mill Run Segment Error! Bookmark not defined. FIGURES

Figure 1. Mill Run Watershed...... 2 Figure 2. Location Map of the Mill Run Segment Watershed ...... 12 Figure 3. Location Map of the Reference Watershed, Racket Brook ...... 15

APPENDICES

Appendix A. AVGWLF Model Overview & GIS-Based Derivation of Input Data ...... 22 Appendix B. AVGWLF Model Inputs for the Mill Run Segment Watershed and Racket Brook Reference Watershed ...... 25 Appendix C. Equal Marginal Percent Reduction Method...... 27 Appendix D. Equal Marginal Percent Reduction Calculations for the Mill Run Segment TMDL ...... 28 Appendix E. Information Sheet for the Mill Run Watershed TMDL...... 29

ATTACHMENTS

Attachment A. Excerpts Justifying Changes Between the 1996, 1998, and 2002, Section 303(d) Lists and 2004 and 2006 Integrated Lists...... 33 Attachment B. TMDLs By Segment...... 36 Attachment C. Water Quality Data Used in TMDL Calculations...... 42 Attachment D YSI Continuous Sonde Data ...... 44 Attachment E. Comment and Response...... 46

TMDL1 Mill Run Watershed Blair and Cambria Counties, Pennsylvania

INTRODUCTION

This report presents the Total Maximum Daily Load (TMDL) developed for stream segments in the Mill Run Watershed (Figure 1). This was done to address impairments noted on the 1996, 1998, and 2002 Pennsylvania Section 303(d) lists, and the 2004 and 2006 Integrated Lists required under the Clean Water Act and covers two segments on this list (Table 1). High levels of metals and siltation caused these impairments. The TMDL addresses the two metals (iron and aluminum) identified as impairing the stream from urban runoff and combined sewer overflows (CSOs) and sediment associated with small residential runoff.

Table 1. Mill Run Segments Addressed

State Water Plan (SWP) Subbasin: 11-A Little Juniata River and Frankstown Branch Segment Year Stream Stream Source Cause Miles ID Listed Name Code Combined Sewer Overflow 6564 1996 Mill Run 16403 Unknown 3.9 Urban Runoff/Storm Sewers 20000327- 2002 Mill Run 16403 Small Residential Runoff Siltation 0.8 1001-TAS See Attachment A, Excerpts Justifying Changes Between the 1996, 1998, and 2002 Section 303(d) lists and the 2004 and 2006 Integrated lists. The use designations for the stream segments in this TMDL can be found in PA Title 25 Chapter 93.

LOCATION

Mill Run flows about 8.6 miles from its headwaters to its confluence with Beaverdam Branch. Mill Run is impaired from its mouth upstream 4.7 miles, at which point, the stream no longer flows through an urban landscape (Figure 1). The watershed is located predominately in Blair County with a small section in Cambria County and drains approximately 13 square miles in State Water Plan Subbasin 11A. The aquatic life use for Mill Run is high quality cold water fisheries from its source to the Allegheny Reservoir and warm water fisheries from the reservoir to the mouth.

1 Pennsylvania’s 1996, 1998, and 2002 Section 303(d) lists were approved by the U.S. Environmental Protection Agency. The 2004 and 2006 Integrated Lists have been approved by the U.S Environmental Protection Agency. The 1996 Section 303(d) list provides the basis for measuring progress under the 1996 lawsuit settlement of American Littoral Society and Public Interest Group of Pennsylvania v. EPA.

Figure 1. Mill Run Watershed

Mill Run flows through the lower portion of Altoona City before it enters Beaverdam Branch. The stream is channelized in areas as it travels through Altoona. The Mill Run Watershed can be reached traveling on the PA Turnpike to exit 146 (US 220/I-99). Upon exiting the Turnpike, merge onto US220/I-99 N and follow this to Altoona City. Several major roads intersect the watershed including: PA-36 and PA-764 (Figure 1).

SEGMENTS ADDRESSED IN THIS TMDL

The Mill Run Watershed is affected by pollution from CSOs, urban runoff/storm sewers, and small residential runoff. This pollution has caused high levels of metals and siltation throughout the lower half of the stream. Urban runoff/storm sewers impair Mill Run from the mouth to about four miles upstream. The small residential runoff affects Mill Run for about a mile above the urban impairment. The upper half of the watershed is forested and is attaining its designated use.

CLEAN WATER ACT REQUIREMENTS

Section 303(d) of the 1972 Clean Water Act requires states, territories, and authorized tribes to establish water quality standards. The water quality standards identify the uses for each waterbody and the scientific criteria needed to support that use. Uses can include designations for drinking water supply, contact recreation (swimming), and aquatic life support. Minimum goals set by the Clean Water Act require that all waters be “fishable” and “swimmable.”

Additionally, the federal Clean Water Act and the U.S. Environmental Protection Agency’s (USEPA) implementing regulations (40 CFR Part 130) require:

• States to develop lists of impaired waters for which current pollution controls are not stringent enough to meet water quality standards (the list is used to determine which streams need TMDLs);

• States to establish priority rankings for waters on the lists based on severity of pollution and the designated use of the waterbody; states must also identify those waters for which TMDLs will be developed and a schedule for development;

• States to submit the list of waters to USEPA every two years (April 1 of the even numbered years);

• States to develop TMDLs, specifying a pollutant budget that meets state water quality standards and allocate pollutant loads among pollution sources in a watershed, e.g., point and nonpoint sources; and

• USEPA to approve or disapprove state lists and TMDLs within 30 days of final submission.

Despite these requirements, states, territories, authorized tribes, and USEPA have not developed many TMDLs since 1972. Beginning in 1986, organizations in many states filed lawsuits against the USEPA

for failing to meet the TMDL requirements contained in the federal Clean Water Act and its implementing regulations. While USEPA has entered into consent agreements with the plaintiffs in several states, many lawsuits still are pending across the country.

In the cases that have been settled to date, the consent agreements require USEPA to backstop TMDL development, track TMDL development, review state monitoring programs, and fund studies on issues of concern (e.g., AMD, implementation of nonpoint source Best Management Practices, etc.). These TMDLs were developed in partial fulfillment of the 1996 lawsuit settlement of American Littoral Society and Public Interest Group of Pennsylvania v. EPA.

SECTION 303(D) LISTING PROCESS

Prior to developing TMDLs for specific waterbodies, there must be sufficient data available to assess which streams are impaired and should be on the Section 303(d) list. With guidance from the USEPA, the states have developed methods for assessing the waters within their respective jurisdictions.

The primary method adopted by the Pennsylvania Department of Environmental Protection (PADEP) for evaluating waters changed between the publication of the 1996 and 1998 Section 303(d) lists. Prior to 1998, data used to list streams were in a variety of formats, collected under differing protocols. Information also was gathered through the Section 305(b)2 reporting process. PADEP is now using the Unassessed Waters Protocol (UWP), a modification of the USEPA Rapid Bioassessment Protocol II (RPB-II), as the primary mechanism to assess Pennsylvania’s waters. The UWP provides a more consistent approach to assessing Pennsylvania’s streams.

The assessment method requires selecting representative stream segments based on factors such as surrounding land uses, stream characteristics, surface geology, and point source discharge locations. The biologist selects as many sites as necessary to establish an accurate assessment for a stream segment; the length of the stream segment can vary between sites. All the biological surveys include kick-screen sampling of benthic macroinvertebrates, habitat surveys, and measurements of pH, temperature, conductivity, dissolved oxygen, and alkalinity. Benthic macroinvertebrates are identified to the family level in the field.

After the survey is completed, the biologist determines the status of the stream segment. The decision is based on the performance of the segment using a series of biological metrics. If the stream is determined to be impaired, the source and cause of the impairment are documented. An impaired stream must be listed on the state’s Section 303(d) list with the documented source and cause. A TMDL must be developed for the stream segment. A TMDL is for only one pollutant. If a stream segment is impaired by two pollutants, two TMDLs must be developed for that stream segment. In order for the process to be more effective, adjoining stream segments with the same source and cause listing are addressed collectively, and on a watershed basis.

2 Section 305(b) of the Clean Water Act requires a biannual description of the water quality of the waters of the state.

BASIC STEPS FOR DETERMINING A TMDL

Although all watersheds must be handled on a case-by-case basis when developing TMDLs, there are basic processes or steps that apply to all cases. They include:

1. Collect and summarize pre-existing data (watershed characterization, inventory contaminant sources, determination of pollutant loads, etc.); 2. Calculate TMDL for the waterbody using USEPA approved methods and computer models; 3. Allocate pollutant loads to various sources; 4. Determine critical and seasonal conditions; 5. Submit draft report for public review and comments; and 6. Obtain USEPA approval of the TMDL.

This document will present the information used to develop the Mill Run Watershed TMDL.

SAMPLING RATIONALE

Pre-existing water quality data were not available for this TMDL. The Susquehanna River Basin Commission (SRBC) established a sampling procedure for the Mill Run Watershed that included collecting grab samples during storm events to establish urban runoff/storm sewer and CSO contributions. Four samples (first flush, rising limb, peak, and falling limb) were taken throughout the storm and a single sample was taken out of the four composited samples. Storm samples were collected during a spring, summer, and fall storm event. Single grab samples were also collected during baseflow and median flow conditions to determine baseline conditions in Mill Run. YSI sondes were placed in the stream during storm events to continuously record instream data.

Four points were sampled on Mill Run during storm events. BVDM5.0 was the most upstream point located above urban influences. BVDM6.0 was located just upstream of the CSO on Mill Run and BVDM7.0 was located below the mixing zone downstream of the CSO. BVDM8.0 was located at the mouth of Mill Run just upstream of Burgoon Run. Base and median flow samples were only collected at BVDM5.0 and BVDM8.0 because during these flow regimes, the CSO is not contributing to the stream flow. See Figure 1 for sampling locations. Raw data is located in Attachment C, Water Quality Data used in TMDL Calculations.

Based on the water quality results from the storm events compared to the baseline results, the CSO is not contributing enough pollutants to impair the stream. Nutrient levels remained below critical limits downstream of the CSO. Biological oxygen demand (BOD) levels did not show a dramatic increase from upstream to downstream of the CSO. BOD levels above the Altoona West CSO average 7.3 mg/l and below the CSO, they average 11.2 mg/l. DO levels did not fall below water quality standards while the CSO was discharging (Attachment D).

One grab sample on Mill Run above and below the CSO did show a significant increase in BOD, however, this was during an isolated event in which the CSO was discharging into Mill Run at baseflow conditions. This occurred because of isolated storms in the upper part of Altoona, while the Mill Run Watershed did not receive rain.

Urban runoff/storm sewers do contribute pollutants to Mill Run. Metal concentrations increase to critical levels during storm events. Iron and aluminum are the metals of concern based on the concentrations from the grab samples. These two metals will be addressed in the TMDL in addition to sediment from small residential runoff.

WATERSHED BACKGROUND

The northern portion of the Mill Run Watershed is located within the Allegheny Front section of the Appalachian Plateaus Province and the southern portion of the watershed is located in the Appalachian Mountain section of the Ridge and Valley Province. Mill Run is typical of watersheds in these sections of the Appalachian Plateaus and Ridge and Valley Provinces. The local relief for this area is moderate to high with the highest elevations being located in the northern part of the watershed area.

The rock type in the watershed is sedimentary; interbedded sedimentary rock comprises over 58 percent of the watershed. Shale spans over 17 percent of the watershed and carbonate rock comprises 13.5 percent of the area. The remaining rock type is sandstone.

The Berks-Weikert-Bedington series is the predominant soil type comprises 43 percent of the watershed. The series is moderately deep and well-drained, with moderate permeability. Runoff potential is negligible to rapid. Major land uses associated with this series are cropland and forest. This series has also been developed, as seen with Altoona City. The Leck Hill-Calvin-Klinesville and Hazleton- Dekalb-Buchanan series encompass approximately 20 percent and 19 percent of the watershed, respectively. These soil series are deep and well drained, with moderate to moderately rapid permeability. Runoff potential is negligible to rapid. Common land uses include: forest and cropland.

The remaining soil series, in order of abundance are: Morrison-Hazleton-Clymer and Chenango-Pope- Holly. These are deep and well drained soils with moderate to moderately rapid permeability. Runoff rates vary from negligible to rapid. The common lands use is cropland and forest.

Based on GIS datasets created in 2001, land use values were calculated for the Mill Run Watershed. Forested land is the dominant land use at approximately 52 percent; it is primarily located in the headwaters area. Developed land comprises over 36 percent of the watershed. The development in the watershed leads to the numerous storm sewer pipes and a combined sewer overflow (CSO) flowing into the stream channel during storm events. The remaining land uses in the watershed include agriculture, disturbed, and water (8.4 percent, 2.5 percent, and 0.6 percent, respectively).

METALS METHODOLOGY

A two-step approach is used for the TMDL analysis of urban runoff/storm sewer impaired stream segments. The first step uses a statistical method for determining the allowable instream concentration at the point of interest necessary to meet water quality standards. This is done at each point of interest (sample point) in the watershed. The second step is a mass balance of the loads as they pass through the watershed. Loads at these points will be computed based on average annual flow.

The statistical analysis described below can be applied to situations where all of the pollutant loading is from nonpoint sources, as well as those where there are both point and nonpoint sources. The following defines what are considered point sources and nonpoint sources for the purposes of our evaluation; point sources are defined as permitted discharges or a discharge that has a responsible party, nonpoint sources are then any pollution sources that are not point sources. For situations where all of the impact is due to nonpoint sources, the equations shown below are applied using data for a point in the stream. The load allocation made at that point will be for all of the watershed area that is above that point. For situations where there are point source impacts alone, or in combination with nonpoint sources, the evaluation will use the point source data and perform a mass balance with the receiving water to determine the impact of the point source.

Allowable loads are determined for each point of interest using Monte Carlo simulation. Monte Carlo simulation is an analytical method meant to imitate real-life systems, especially when other analyses are too mathematically complex or too difficult to reproduce. Monte Carlo simulation calculates multiple scenarios of a model by repeatedly sampling values from the probability distribution of the uncertain variables and using those values to populate a larger data set. Allocations were applied uniformly for the watershed area specified for each allocation point. For each source and pollutant, it was assumed that the observed data were log-normally distributed. Each pollutant source was evaluated separately using @Risk3 by performing 5,000 iterations to determine the required percent reduction so that the water quality criteria, as defined in the Pennsylvania Code, Title 25 Environmental Protection, Department of Environmental Protection, Chapter 93, Water Quality Standards, will be met instream at least 99 percent of the time. For each iteration, the required percent reduction is:

PR = maximum {0, (1-Cc/Cd)} where (1)

PR = required percent reduction for the current iteration

Cc = criterion in mg/l

Cd = randomly generated pollutant source concentration in mg/l based on the observed data

Cd = RiskLognorm(Mean, Standard Deviation) where (1a)

Mean = average observed concentration

Standard Deviation = standard deviation of observed data

The overall percent reduction required is the 99th percentile value of the probability distribution generated by the 5,000 iterations, so that the allowable long-term average (LTA) concentration is:

LTA = Mean * (1 – PR99) where (2)

LTA = allowable LTA source concentration in mg/l

3 @Risk – Risk Analysis and Simulation Add-in for Microsoft Excel, Palisade Corporation, Newfield, NY, 1990-1997.

Once the allowable concentration and load for each pollutant is determined, mass-balance accounting is performed starting at the top of the watershed and working down in sequence. This mass-balance or load tracking is explained below.

Load tracking through the watershed utilizes the change in measured loads from sample location to sample location, as well as the allowable load that was determined at each point using the @Risk program.

There are two basic rules that are applied in load tracking; rule one is that if the sum of the measured loads that directly affect the downstream sample point is less than the measured load at the downstream sample point it is indicative that there is an increase in load between the points being evaluated, and this amount (the difference between the sum of the upstream and downstream loads) shall be added to the allowable load(s) coming from the upstream points to give a total load that is coming into the downstream point from all sources. The second rule is that if the sum of the measured loads from the upstream points is greater than the measured load at the downstream point this is indicative that there is a loss of instream load between the evaluation points, and the ratio of the decrease shall be applied to the load that is being tracked (allowable load(s)) from the upstream point.

Tracking loads through the watershed gives the best picture of how the pollutants are affecting the watershed based on the information that is available. The analysis is done to insure that water quality standards will be met at all points in the stream. The TMDL must be designed to meet standards at all points in the stream, and in completing the analysis, reductions that must be made to upstream points are considered to be accomplished when evaluating points that are lower in the watershed. Another key point is that the loads are being computed based on average annual flow and should not be taken out of the context for which they are intended, which is to depict how the pollutants affect the watershed and where the sources and sinks are located spatially in the watershed.

Information for the TMDL analysis performed using the methodology described above is contained in the “TMDLs by Segment” section of this report.

TMDL ENDPOINTS

One of the major components of a TMDL is the establishment of an instream numeric endpoint, which is used to evaluate the attainment of acceptable water quality. An instream numeric endpoint, therefore, represents the water quality goal that is to be achieved by implementing the load reductions specified in the TMDL. The endpoint allows for comparison between observed instream conditions and conditions that are expected to restore designated uses. The endpoint is based on either the narrative or numeric criteria available in water quality standards.

Because of the nature of the pollution sources in the watershed, the TMDLs component makeup will be load allocations that are specified above a point in the stream segment. All allocations will be specified as long-term average daily concentrations. These long-term average daily concentrations are expected to meet water quality criteria 99 percent of the time. Pennsylvania Title 25 Chapter 96.3(c) specifies that a minimum 99 percent level of protection is required. All metals criteria evaluated in this TMDL are specified as total recoverable. Pennsylvania does have dissolved criteria for iron; however, the data

used for this analysis report iron as total recoverable. Table 2 shows the water quality criteria for the selected parameters.

Table 2. Applicable Water Quality Criteria

Parameter Criterion Value (mg/l) Total Recoverable/Dissolved Aluminum (Al) 0.75 Total Recoverable 1.50 30-day average; Total Recoverable Iron (Fe) 0.3 Dissolved

TMDL ELEMENTS (WLA, LA, MOS)

A TMDL equation consists of a wasteload allocation (WLA), load allocation (LA) and a margin of safety (MOS). The WLA is the portion of the load assigned to point sources. The LA is the portion of the load assigned to nonpoint sources. The MOS is applied to account for uncertainties in the computational process. The MOS may be expressed implicitly (documenting conservative processes in the computations) or explicitly (setting aside a portion of the allowable load).

TMDL METALS ALLOCATION SUMMARY

Information for the TMDL analysis using the methodology described above is contained in the TMDLs by Segment section in Attachment B.

As changes occur in the watershed, the TMDL may be reevaluated to reflect current conditions. Table 3 presents the estimated reductions identified for all points in the watershed. Attachment B gives detailed TMDLs by Segment analysis for each allocation point.

Table 3. Summary Table–Mill Run Watershed

Existing Load Allowable Load Load Reduction Station Parameter Percent Reduction (lbs/day) (lbs/day) (lbs/day BVDM5.0 Mill Run above development Fe 148.1 31.7 116.4 79 Al 101.0 12.5 88.5 88 BVDM6.0 Mill Run above Altoona West CSO Fe 434.9 68.6 249.9 78 Al 259.8 26.3 145.0 85 BVDM7.0 Mill Run below Altoona West CSO Fe 470.6 85.4 18.9 18 Al 241.3 40.5 0 0 BVDM8.0 Mouth of Mill Run Fe 619.5 42.9 191.4 82 Al 268.9 23.4 44.7 66

MILL RUN SEGMENT WATERSHED SILTATION TMDL

Summary of Small Residential Runoff Impaired Segment of Mill Run

1. The impaired stream segment addressed by this total maximum daily load (TMDL) is located in Logan Township, western Blair County. The segment drains approximately 6.75 square miles of Mill Run (Figure 2) and its headwaters, as part of State Water Plan Subbasin 11A. The aquatic life existing use for Mill Run, including its tributaries, is high quality cold water fishery above the Allegheny Reservoir, and a warm water fishery below the Allegheny Reservoir under §93.9f in Title 25 of the Pa. Code (Commonwealth of Pennsylvania, 2001).

2. This Mill Run Segment TMDL was developed to address use impairments caused by sediment. Pennsylvania’s 2004Integrated List identified 0.8 miles of Mill Run as impaired by siltation emanating from small residential activities in the basin. In order to ensure attainment and maintenance of water quality standards in this segment to Mill Run, mean annual loadings of total sediment will need to be limited to 2,568,123.81 lbs/yr.

The major components of the Mill Run Segment are summarized below:

Sediment Components (lbs/yr) TMDL (Total Maximum Daily Load) 2,568,123.81 MOS (Margin of Safety) 256,812.38 LA (Load Allocation) 2,311,311.43

3. Mean annual sediment loadings are estimated to be 5,303,044.2 lbs/yr. To meet the TMDL, the sediment loadings will require a 52 percent reduction.

4. There are no point sources to address in this TMDL. Load Allocations (LA) for sediment were made to the following nonpoint sources: hay and pasture lands, croplands, coniferous forest, mixed forest, deciduous forest, developed areas, streambanks, groundwater and septic systems.

5. The adjusted load allocation (ALA) is the actual portion of the LA distributed among nonpoint sources receiving reductions, or sources that are considered controllable. Controllable sources receiving allocations are hay/pasture, cropland, developed lands, and streambanks. The sediment TMDL includes a nonpoint source ALA of and 2,046,450.03 lbs/yr. Sediment loadings from all other sources, such as forested areas, were maintained at their existing levels. Allocations of sediment to controllable nonpoint sources, or the ALA, for this segment of Mill Run are summarized below:

Adjusted Load Allocations for Sources of Sediment Adjusted Load Current Loading Allocation Pollutant (lbs/yr) (lbs/yr) % Reduction Sediment 5,303,044.20 2,046,450.03 62

6. Ten percent of the Mill Run Segment sediment TMDL was set-aside as a margin of safety (MOS). The MOS is that portion of the pollutant loading that is reserved to account for any uncertainty in the data and computational methodology used for the analysis. The MOS for the sediment TMDL was set at 256,812.38 lbs/yr.

7. The continuous simulation model used for developing the Mill Run Segment TMDL 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. The combination of these actions accounts for seasonal variability.

Figure 2. Location Map of the Mill Run Segment Watershed

Reference Watershed Approach

The TMDL developed for the Mill Run Segment watershed addresses sediment. Because neither Pennsylvania nor the U.S. Environmental Protection Agency (USEPA) has instream numerical water quality criteria for sediment, a method was developed to implement the applicable narrative criteria. The method for these types of TMDLs is termed the “Reference Watershed Approach.” Meeting the water quality objectives specified for this TMDL will result in the impaired stream segment attaining its designated uses.

The Reference Watershed Approach compares two watersheds, one attaining its uses and one that is impaired based on biological assessments. Both watersheds ideally have similar land use/cover distributions. Other features such as base geologic formation should be matched to the extent possible; however, most variations can be adjusted for in the model. The objective of the process is to reduce the loading rate of pollutants in the impaired stream segment to a level equivalent to the loading rate in the nonimpaired, reference stream segment. This load reduction will result in conditions favorable to the return of a healthy biological community to the impaired stream segments.

Selection of the Reference Watershed

In general, three factors are considered when selecting a suitable reference watershed. The first factor is to use a watershed that the PADEP has assessed and determined to be attaining water quality standards. The second factor is to find a watershed that closely resembles the impaired watershed in physical properties such as land cover/land use, physiographic province, and geology/soils. Finally, the size of the reference watershed should be within 20-30 percent of the impaired watershed area. The search for a reference watershed for the Mill Run Segment 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 streams database, and geologic rock types.

A stream named Racket Brook was selected as the reference watershed for developing the Mill Run Segment TMDL. Racket Brook is located east of the city of Carbondale, in eastern Lackawanna County and western Wayne County, Pennsylvania (Figure 3). The watershed is located in State Water Plan subbasin 05A, and protected uses include aquatic life and recreation. The stream is currently designated as cold water fishery under §93.9z in Title 25 of the Pa. Code (Commonwealth of Pennsylvania, 2001). Based on PADEP assessments, Racket Brook is currently attaining its designated uses. The attainment of designated uses is based on sampling done by the PADEP in 1997, as part of its State Surface Water Assessment Program.

Drainage area, location, and other physical characteristics of this segment of Mill Run were compared to the Racket Brook reference stream (Table 4). Forested is the dominant land use category in both Racket Brook (89 percent) and the small residential runoff impaired section of Mill Run (83 percent). The geology, soils, and precipitation in both are also similar (Table 5).

Table 4. Comparison Between Mill Run Segment and Racket Brook

Watershed tribute Mill Run Segment Racket Brook Physiographic Ridge and Valley (60%) Ridge and Valley (98%) Province Appalachian Plateaus (40%) Appalachian Plateaus (2%) Area (mi2) 6.75 4.90 Land Use Agriculture (8%) Agriculture (6%) Development (5%) Development (4%) Forested (83%) Forested (89%) Geology Interbedded Sedimentary (78%) Interbedded Sedimentary (10%) Sandstone (17%) Sandstone (90%) Shale (5%) Soils Hazelton-Dekalb-Buchannon (35%) Arnot-Oquaga-Dystrochrepts (80%) Leck Kill-Calvin-Klinesville (35%) Udorthents-Urban Land-Volusia (15%) Berks-Weikert-Bedington (30%) Wellsboro-Oquaga-Morris (5%) Dominant HSG Hazelton-Dekalb-Buchannon Arnot-Oquaga-Dystrochrepts A (2%) A (21%) B (45%) B (0%) C (53%) C (48%) D (0%) D (31%)

Leck Kill-Calvin-Klinesville Udothents-Urban Land-Volusia A (0%) A (2%) B (32%) B (5%) C (44%) C (18%) D (24%) D (75%)

Berks-Weikert-Bedington Wellsboro-Oquaga-Morris A (0%) A (0%) B (13%) B (0%) C (52%) C (95%) D (35%) D (5%) K Factor Hazelton-Dekalb-Buchannon (0.18) Arnot-Oquaga-Dystrochrepts (0.21) Leck Kill-Calvin-Klinesville (0.23) Udorthents-Urban Land-Volusia (0.17) Berks-Weikert-Bedington (0.24) Wellsboro-Oquaga-Morris (0.25) 20-Yr. Ave. 45.8 43.8 Rainfall (in) 20-Yr. Ave. 2.88 3.47 Runoff (in)

Figure 3. Location Map of the Reference Watershed, Racket Brook

WATERSHED ASSESSMENT AND MODELING

The TMDL for the Mill Run Segment watershed was developed using the ArcView Generalized Watershed Loading Function model (AVGWLF) as described in Appendix A. The AVGWLF model was used to establish existing loading conditions for the Mill Run Segment watershed and the reference Racket Brook watershed. All modeling inputs have been attached to this TMDL as Appendix B.

The AVGWLF model produced information on watershed size, land use, and sediment loading. The sediment loads represent an annual average over a 20-year period (1978 to 1998). This information was then used to calculate existing unit area loading rates for the Mill Run Segment and Racket Brook reference watersheds. Sediment loading information for both the impaired watershed and the reference watershed are shown in Tables 5 and 6, respectively.

Table 5. Existing Sediment Loads for Mill Run Segment

Sediment Unit Area Pollutant Mean Annual Loading Source Acreage Loading (lbs/yr) (lbs/ac/yr) HAY/PAST 158.1 101,912.20 644.61 CROPLAND 185.3 3,219,369.80 17,373.83 CONIF_FOR 9.9 167.00 16.87 MIXED_FOR 93.9 7,410.00 78.91 DECID_FOR 3,464.4 257,284.40 74.27 QUARRY 103.8 296,117.20 2,852.77 COAL_MINE 9.9 4,947.20 499.72 TRANSITION 59.3 917,551.80 15,473.05 LO_INT_DEV 232.3 75,134.40 323.44 Streambank 423,150.20 Groundwater Point Source Septic Systems TOTAL 4,316.9 5,303,044.2 1,228.44

Table 6. Existing Sediment Loads for Racket Brook

Sediment Unit Area Pollutant Mean Annual Loading Source Acreage Loading (lbs/yr) (lbs/ac/yr) HAY/PAST 76.60 52,238.40 681.96 CROPLAND 98.80 853,842.80 8,642.13 CONIF_FOR 232.30 17,694.60 76.17 MIXED_FOR 336.10 18,243.40 54.28 DECID_FOR 2,231.40 146,414.40 65.62 QUARRY 17.30 268,412.60 15,515.18 UNPAVED_R 2.50 48,375.80 19,350.32 TRANSITION 19.80 301,730.60 15,238.92 LO_INT_DEV 116.10 41,141.20 354.36 HI_INT_DEV 7.40 87.80 11.86 Streambank 118,789.55 Groundwater Point Source Septic Systems Total 3,138.30 1,866,971.15 594.90

TMDLS

The targeted TMDL value for this segment of Mill Run watershed was established based on current loading rates for sediment in the Racket Brook reference watershed. Biological assessments have determined that Racket Brook is currently attaining its designated uses. Reducing the loading rate of sediment in this segment of Mill Run watershed to levels equivalent to those in the reference watershed will provide conditions favorable for the reversal of current use impairments.

Background Pollutant Conditions

There are two separate considerations of background pollutants within the context of this TMDL. First, there is the inherent assumption of the reference watershed approach that because of the similarities between the reference and impaired watershed, the background pollutant contributions will be similar. Therefore, the background pollutant contributions will be considered when determining the loads for the impaired watershed that are consistent with the loads from the reference watershed. Second, the AVGWLF model implicitly considers background pollutant contributions through the soil and the groundwater component of the model process.

Targeted TMDLs

The targeted TMDL value for sediment was determined by multiplying the total area of the Mill Run segment watershed (4,316.90 acres) by the appropriate unit area loading rate for the Racket Brook

reference watershed (Table 7). The existing mean annual loading of sediment to Mill Run (5,303,044.20 lbs/yr) will need to be reduced by 52 percent to meet the targeted TMDL of 597,865.70 lbs/yr.

Table 7. Targeted TMDL for the Mill Run Segment Watershed

Unit Area Loading Rate Area Racket Brook Reference Watershed Targeted TMDL for Mill Pollutant (ac) (lbs/ac/yr) Run (lbs/yr) Sediment 4,316.90 594.90 2,568,123.81

The targeted TMDL value was used as the basis for load allocation and reduction in the Mill Run Segment watershed, using the following two equations:

1. TMDL = LA + MOS 2. LA = ALA + LNR where: TMDL = Total Maximum Daily Load LA = Load Allocation (nonpoint sources) ALA = Adjusted Load Allocation LNR = Loads not Reduced

Margin of Safety

The MOS is that portion of the pollutant loading that is reserved to account for any uncertainty in the data and computational methodology used for the analysis. For this analysis, the MOS is explicit. Ten percent of the targeted TMDL sediment was reserved as the MOS. Using 10 percent of the TMDL load is based on professional judgment and will provide an additional level of protection to the designated uses of this segment to Mill Run. The MOS used for sediment TMDL was 256,812.38 lbs/yr.

MOS (sediment) = 2,568,123.81 lbs/yr (TMDL) x 0.1 = 256,812.38 lbs/yr

Load Allocation

The LA is that portion of the TMDL that is assigned to nonpoint sources. The LA was computed by subtracting the WLA and MOS values from the targeted TMDL values. LA for sediment is 456,816.14 lbs/yr.

LA (sediment) = 2,568,123.81 lbs/yr (TMDL) – 256,812.38 lbs/yr (MOS) = 2,311,311.43 lbs/yr

Adjusted Load Allocation

The ALA is the actual portion of the LA distributed among those nonpoint sources receiving reductions. It is computed by subtracting those nonpoint source loads that are not being considered for reductions (loads not reduced or LNR) from the LA. Sediment reductions were made to the hay/pasture, cropland,

developed areas (sum of LO_INT_DEV, HI_INT_DEV and septic systems), and streambanks. Those land uses/sources for which existing loads were not reduced (CONIF_FOR, MIXED_FOR, DECID_FOR, and groundwater) were carried through at their existing loading values (Table 8). The ALA for sediment was 2,046,450.03 lbs/yr.

Table 8. Load Allocations, Loads Not Reduced, and Adjusted Load Allocations for Mill Run Segment

Sediment (lbs/yr) Load Allocation 2,311,311.43 Loads Not Reduced 264,861.40 CONIF_FOR 167.00 MIXED_FOR 7,410.00 DECID_FOR 257,284.40 Groundwater -- Adjusted Load Allocation 2,046,450.03

TMDL

The sediment TMDL established for the Mill Run Segment watershed consists of a LA and a MOS. The individual components of the TMDL are summarized in Table 9.

Table 9. TMDL, MOS, LA, LNR, and ALA for Mill Run Segment

Sediment Component (lbs/yr) TMDL (Total Maximum Daily Load) 2,568,123.81

MOS (Margin of Safety) 256,812.38

LA (Load Allocation) 2,311,311.43

LNR (Loads Not Reduced) 264,861.40

ALA (Adjusted Load Allocation) 2,046,450.03

CALCULATION OF SEDIMENT LOAD REDUCTIONS

ALAs established in the previous section represent the annual total sediment loads that are available for allocation between contributing sources in the Mill Run Segment watershed. The ALA for sediment was allocated between agriculture, developed areas, and streambanks. LA and reduction procedures were applied to the entire Mill Run Segment watershed using the Equal Marginal Percent Reduction (EMPR) allocation method (Appendix C). The LA and EMPR procedures were performed using MS Excel and results are presented in Appendix D.

To meet the sediment TMDL, the current loading from controllable sources will require a reduction to 2,046,450.03 lbs/yr. This is achievable through sediment load reductions of 23 percent for hay/pastures, developed lands, and streambanks, and a 51 percent for cropland (Table 10).

Table 10. Sediment Load Allocations & Reductions for Mill Run Segment

Unit Area Loading Rate Pollutant Loading Pollutant (lbs/ac/yr) (lbs/yr) % Source Acres Current Allowable Current Allowable (LA) Reduction Sediment Hay/Pasture 158.10 644.61 498.42 101,912.84 78,800.93 23 Cropland 185.10 17,373.83 8,548.70 3,215,895.93 1,582,363.65 51 Developed 232.00 323.44 250.09 75,135.11 58,095.70 23 Streambanks 0.00 423,150.20 327,189.76 23

CONSIDERATION OF CRITICAL CONDITIONS

The AVGWLF model is a continuous simulation model, which uses daily time steps for weather data and water balance calculations. Monthly calculations are made for sediment loads, based on the daily water balance accumulated to monthly values. Therefore, all flow conditions are taken into account for loading calculations. Establishing the TMDL using average annual conditions is protective of the waterbody.

CONSIDERATION OF SEASONAL VARIATIONS

The continuous simulation model used for these analyses 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 combination of these actions by the model accounts for seasonal variability.

RECOMMENDATIONS FOR IMPLEMENTATION (METALS AND SEDIMENT)

TMDLs represent an attempt to quantify the pollutant load that may be present in a waterbody and still ensure attainment and maintenance of water quality standards. The Mill Run TMDL identifies the necessary overall load reductions for sediment and metals currently causing use impairments and distributes those reduction goals to the appropriate nonpoint sources. Reaching the reduction goals established by the TMDLs will only occur through BMPs.

BMPs that would be effective in lowering the amount of metals that flow into Mill Run include: riparian buffer strips, curb elimination, proper lawn care, bioretention areas, and wetland complexes, among other urban BMPs. Riparian buffer strips stabilize streambanks and provide habitat for wildlife. Proper lawn care includes not mowing to the edge of the stream channel and mulching any open soil areas. Curb elimination allows runoff to be spread out over adjacent vegetated areas instead of being channelized to the stream. Bioretention areas and wetlands decrease the velocity of runoff and help to

remove pollutants. Implementation of these BMPs will help to lower the pollutant load that reaches Mill Run and achieve the goals of the TMDLs.

Since the siltation impairment is located in a rural part of the watershed, BMPs that would be helpful in lowering the amount of sediment reaching Mill Run include: riparian buffer strips, proper lawn care, mulching, sodding, and water cisterns, among many others. Riparian buffer strips are most effective if planted using native vegetation. Sodding is the practice of stabilizing soil, with grass sod, in areas that have been disturbed; this is an immediate stabilization until seeding grass. Water cisterns prevent the rapid discharge of water from roof tops through gutters. The implementation of these BMPs, along with others, will assist in lowering the sediment loads to Mill Run and help reach the goals of the TMDLs.

PUBLIC PARTICIPATION

In the beginning stages of the Mill Run Watershed TMDL, an early notification letter was sent to inform stakeholders and interested parties that a TMDL would be completed in their watershed and offer them the opportunity to submit information for TMDL development. The PADEP considered all the information submitted and all pertinent information was included in the report.

Public notice of the draft TMDL was published in the Pennsylvania Bulletin on January 13, 2007, and The Mirror on February [Day], 2007, to foster public comment on the allowable loads calculated. A public meeting was held on February 6, 2007, at the Holiday Inn Express in Altoona, Pa., to discuss the proposed TMDL.

Appendix A. AVGWLF Model Overview & GIS-Based Derivation of Input Data

The TMDL for this segment of Mill Run was developed using the Generalized Watershed Loading Function or GWLF model. The GWLF model provides the ability to simulate runoff, sediment, and nutrient (nitrogen and phosphorus) loadings from watershed 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, which 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 homogenous 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 on watershed size, transport capacity, and average daily runoff is applied to the calculated erosion for determining sediment yield for each source area. Surface nutrient losses are determined by applying dissolved nitrogen and phosphorus coefficients to surface runoff and a sediment coefficient to the yield portion for each agricultural source area. Point source discharges also can contribute to dissolved 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 nitrogen and phosphorus coefficients for shallow groundwater contributions to stream nutrient loads, and the subsurface submodel 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 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.

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, which was 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 then written to the TRANSPRT.DAT, NUTRIENT.DAT and WEATHER.DAT input files needed to execute the GWLF model. For use in Pennsylvania, AVGWLF has been linked with statewide GIS data layers such as land use/cover, soils, topography, and physiography; and includes location-specific default information such as background nitrogen and phosphorus concentrations and cropping practices. Complete GWLF-formatted weather files also are included for 80 weather stations around the state.

The following table lists the statewide GIS data sets and provides an explanation of how they were used for development of the input files for the GWLF model.

GIS Data Sets DATASET DESCRIPTION Censustr Coverage of Census data including information on individual homes septic systems. The attribute usew_sept includes data on conventional systems, and sew_other provides data on short-circuiting and other systems. County The County boundaries coverage lists data on conservation practices, which provides 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. Used to calculate landslope and slope length. Palumrlc A satellite image derived land cover grid that is classified into 15 different land cover categories. This dataset provides land cover 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 nitrogen and phosphorus loads. Refwater Shapefile of reference watersheds for which nutrient and sediment loads have been calculated. Soilphos A grid of soil phosphorous loads, which has been generated from soil sample data. Used to help set phosphorus and sediment values. Smallsheds A coverage of watersheds derived at 1:24,000 scale. 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 PADEP 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.

Appendix B. AVGWLF Model Inputs for the Mill Run Segment Watershed and Racket Brook Reference Watershed

Mill Run Segment Transport

Racket Brook Transport

Appendix C. Equal Marginal Percent Reduction Method

The Equal Marginal Percent Reduction (EMPR) allocation method was used to distribute Adjusted Load Allocations (ALAs) between the appropriate contributing nonpoint sources. The load allocation and EMPR procedures were performed using the MS Excel and results are presented in Appendix F. The five major steps identified in the spreadsheet are summarized below:

1. Calculation of the TMDL based on impaired watershed size and unit area loading rate of the reference watershed.

2. Calculation of Adjusted Load Allocation based on TMDL, Margin of Safety, and existing loads not reduced.

3. Actual EMPR Process.

a. Each land use/source load is compared with the total ALA to determine if any contributor would exceed the ALA by itself. The evaluation is carried out as if each source is the only contributor to the pollutant load of the receiving waterbody. If the contributor exceeds the ALA, that contributor would be reduced to the ALA. If a contributor is less than the ALA, it is set at the existing load. This is the baseline portion of the EMPR. b. 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 ALA. If the ALA 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.

4. Calculation of total loading rate of all sources receiving reductions.

5. Summary of existing loads, final load allocations, and percent reduction for each pollutant source.

Appendix D. Equal Marginal Percent Reduction Calculations for the Mill Run Segment TMDL

Appendix E. Information Sheet for the Mill Run Watershed TMDL

What is being proposed? Total Maximum Daily Load (TMDL) plans have been developed to improve water quality in the Mill Run Watershed.

Who is proposing the plans? Why? The Pennsylvania Department of Environmental Protection (PADEP) is proposing to submit the plans to the U.S. Environmental Protection Agency (USEPA) for review and approval as required by federal regulation. In 1995, USEPA was sued for not developing TMDLs when Pennsylvania failed to do so. PADEP has entered into an agreement with USEPA to develop TMDLs for certain specified waters over the next several years. This TMDL has been developed in compliance with the state/USEPA agreement.

What is a TMDL? A TMDL sets a ceiling on the pollutant loads that can enter a waterbody so that it will meet water quality standards. The Clean Water Act requires states to list all waters that do not meet their water quality standards even after pollution controls required by law are in place. For these waters, the state must calculate how much of a substance can be put in the water without violating the standard, and then distribute that quantity to all the sources of the pollutant on that waterbody. A TMDL plan includes waste load allocations for point sources, load allocations for nonpoint sources, and a margin of safety. The Clean Water Act requires states to submit their TMDLs to USEPA for approval. Also, if a state does not develop the TMDL, the Clean Water Act states that USEPA must do so.

What is a water quality standard? The Clean Water Act sets a national minimum goal that all waters be “fishable” and “swimmable.” To support this goal, states must adopt water quality standards. Water quality standards are state regulations that have two components. The first component is a designated use, such as “warm water fishes” or “recreation.” States must assign a use, or several uses to each of their waters. The second component relates to the instream conditions necessary to protect the designated use(s). These conditions or “criteria” are physical, chemical, or biological characteristics such as temperature and minimum levels of dissolved oxygen, and maximum concentrations of toxic pollutants. It is the combination of the “designated use” and the “criteria” to support that use that make up a water quality standard. If any criteria are being exceeded, then the use is not being met and the water is said to be in violation of water quality standards.

What is the purpose of the plans? Mill Run is impaired due to siltation from small residential runoff and pollutants from combined sewer overflows and urban runoff/storm sewers. The plans include a calculation of the loading for metals and sediment to correct the problems and meet water quality objectives.

Why was Mill Run selected for TMDL development? In 1996, PADEP listed Mill Run under Section 303(d) of the federal Clean Water Act as impaired due to causes linked to unknown pollutants from urban runoff/storm sewers and combined sewer overflows, and siltation from small residential runoff. .

What pollutants do these TMDLs address? The proposed plans provide calculations of the stream’s total capacity to accept sediment and metals (iron and aluminum).

Where do the pollutants come from? The siltation and unknown impairments in Mill Run come from nonpoint sources of pollution, small residential and urban runoff.

How was the TMDL developed? PADEP used a reference watershed approach to estimate the necessary loading reduction of sediment that would be needed to restore a healthy aquatic community. The reference watershed approach is based on selecting a nonimpaired watershed that has similar land use characteristics and determining the current loading rates for the pollutants of interest. This is done by modeling the loads that enter the stream, using precipitation and land use characteristic data. For this analysis, PADEP used the AVGWLF model (the Environmental Resources Research Institute of the Pennsylvania State University’s Arcview based version of the Generalized Watershed Loading Function model developed by Cornell University). This modeling process uses loading rates in the nonimpaired watershed as a target for load reductions in the impaired watershed. The impaired watershed is modeled to determine the current loading rates and determine what reductions are necessary to meet the loading rates of the nonimpaired watershed. The reference stream approach was used to set allowable loading rates in the affected watershed because neither Pennsylvanian nor USEPA has instream numerical water quality criteria for nutrients.

Monte Carlo simulation techniques were used to determine the long-term average daily metal loads that a stream segment could accept and still meet water quality criteria 99 percent of the time. Monte Carlo simulation allows for the expansion of a data set based on its statistical makeup. Since there was no one critical flow condition where criteria were exceeded, the 50th percentile flow value was used to weight the pollutant concentrations. All analyses were started at the headwaters of each stream segment.

How much pollution is too much? The allowable amount of pollution in a waterbody varies depending on several conditions. TMDLs are set to meet water quality standards at the critical flow condition. For a free flowing stream impacted by nonpoint source pollution loading of nutrients, the TMDL is expressed as an annual loading. This accounts for pollution contributions over all stream flow conditions. PADEP established the water quality objectives for nutrients by using the reference watershed approach. This approach assumes that the impairment is eliminated when the impaired watershed achieves loadings similar to the reference watershed. Reducing the current loading rates for nutrients in the impaired watershed to the current loading rates in the reference watershed will result in meeting the water quality objectives.

The allowable amount of pollution in a stream varies depending on several conditions. TMDLs are set to meet water quality standards at the defined critical flow condition. If there is more than one critical flow condition, the TMDL must be constructed to meet water quality criteria at all flow conditions. This is the case in the Mill Run Watershed. For this reason, the long-term average daily load values that will meet water quality criteria 99 percent of the time were used as the basis for the TMDL.

The applicable water quality criteria for metals are a daily average of 1.5 mg/l total iron and a one hour maximum of 0.75 mg/l total aluminum.

How will the loading limits be met? Best Management Practices (BMPs) will be encouraged throughout the watershed to achieve the necessary load reductions.

How can I get more information on the TMDL? To request a copy of the full report, contact Bill Brown at (717) 783-2951 between 8:00 a.m. and 3:00 p.m., Monday through Friday. Mr. Brown also can be reached by mail at the Office of Water Management, PADEP, Rachel Carson State Office Building, 400 Market Street, Harrisburg, PA 17105 or by e-mail at [email protected].

How can I comment on the proposal? You may provide e-mail or written comments postmarked no later than March 9, 2007, to the above address.

REFERENCES

Commonwealth of Pennsylvania. 2001. Pennsylvania Code. Title 25 Environmental Protection.

Department of Environmental Protection. Chapter 93. Water Quality Standards. Harrisburg, Pa.

Donoughe, Michael. 2006. Personal Conversation about the Mill Run Watershed.

Watershed Restoration Action Strategy (WRAS). 2001. Pennsylvania Department of Environmental Protection. State Water Plan Subbasin 11A Little Juniata River and Frankstown Branch Watersheds.

Attachment A

Excerpts Justifying Changes Between the 1996, 1998, 2002 Section 303(d) Lists and 2004 and 2006 Integrated Lists

The following are excerpts from the Pennsylvania DEP 303(d) narratives that justify changes in listings between the 1996, 1998, and 2002 and 2004 and 2006 Intregrated lists. The 303(d) listing process has undergone an evolution in Pennsylvania since the development of the 1996 list.

In the 1996 303(d) narrative, strategies were outlined for changes to the listing process. Suggestions included, but were not limited to, a migration to a Global Information System (GIS), improved monitoring and assessment, and greater public input.

The migration to a GIS was implemented prior to the development of the 1998 303(d) list. As a result of additional sampling and the migration to the GIS some of the information appearing on the 1996 list differed from the 1998 list. Most common changes included:

1. Mileage differences due to recalculation of segment length by the GIS; 2. Slight changes in source(s)/cause(s) due to new USEPA codes; 3. Changes to source(s)/cause(s), and/or miles due to revised assessments; 4. Corrections of misnamed streams or streams placed in inappropriate SWP subbasins; and 5. Unnamed tributaries no longer identified as such and placed under the named watershed listing.

Prior to 1998, segment lengths were computed using a map wheel and calculator. The segment lengths listed on the 1998 303(d) list were calculated automatically by the GIS (ArcInfo) using a constant projection and map units (meters) for each watershed. Segment lengths originally calculated by using a map wheel and those calculated by the GIS did not always match closely. This was the case even when physical identifiers (e.g., tributary confluence and road crossings) matching the original segment descriptions were used to define segments on digital quad maps. This occurred to some extent with all segments, but was most noticeable in segments with the greatest potential for human errors using a map wheel for calculating the original segment lengths (e.g., long stream segments or entire basins). The 2002 Pa. Section 303(d) list was written in a manner similar to the 1998 Section 303(d) list.

In 2004, Pennsylvania developed the Integrated List of All Waters. The water quality status of Pennsylvania’s waters is summarized using a five-part categorization of waters according to their water quality standard (WQS) attainment status. The categories represent varying levels of WQS attainment, ranging from Category 1, where all designated water uses are met, to Category 5, where impairment by pollutants requires a TMDL to correct. These category determinations are based on consideration of data and information consistent with the methods outlined by the Statewide Surface Water Assessment Program. Each PADEP five-digit waterbody segment is placed in one of the WQS attainment categories. Different segments of the same stream may appear on more than one list if the attainment status changes as the water flows downstream. The listing categories are as follows:

Category 1: Waters attaining all designated uses. Category 2: Waters where some, but not all, designated uses are met. Attainment status of the remaining designated uses is unknown because data are insufficient to categorize a water consistent with the state’s listing methodology.

Category 3: Waters for which there are insufficient or no data and information to determine, consistent with the state’s listing methodology, if designated uses are met. Category 4: Waters impaired for one or more designated use but not needing a TMDL. States may place these waters in one of the following three subcategories: • TMDL has been completed. • Expected to meet all designated uses within a reasonable timeframe. • Not impaired by a pollutant. Category 5: Waters impaired for one or more designated uses by any pollutant. Category 5 includes waters shown to be impaired as the result of biological assessments used to evaluate aquatic life use even if the specific pollutant is not known unless the state can demonstrate that nonpollutant stressors cause the impairment or that no pollutant(s) causes or contribute to the impairment. Category 5 constitutes the Section 303(d) list that USEPA will approve or disapprove under the Clean Water Act. Where more than one pollutant is causing the impairment, the water remains in Category 5 until all pollutants are addressed in a completed USEPA-approved TMDL or one of the delisting factors is satisfied.

The 2006 Integrated List was written in a manner similar to the 2004 Integrated List.

Attachment B TMDLs By Segment

Mill Run

The TMDL for the Mill Run Watershed consists of load allocations for four sampling sites along the mainstem. Mill Run is listed as impaired on the Section 303(d) list by unknown causes and siltation from urban runoff/storm sewers and combined sewer overflows, and small residential runoff, respectively. The siltation impairment is addressed in the Mill Run Segment Siltation TMDL Section of this report. Water quality samples taken during 2006 isolated iron and aluminum as the causes of impairment from urban runoff.

An allowable long-term average instream concentration for iron and aluminum was determined at each sample point. The analysis is designed to produce a long-term average value that, when met, will be protective of the water quality criterion for that parameter 99 percent of the time. An analysis was performed using Monte Carlo simulation to determine the necessary long-term average concentration needed to attain water quality criteria 99 percent of the time. The simulation was run assuming the data set was lognormally distributed. Using the mean and the standard deviation of the data set, 5,000 iterations of sampling were completed and compared against the water quality criterion for that parameter. For each sampling event a percent reduction was calculated, if necessary, to meet water quality criteria. A second simulation that multiplied the percent reduction times the sampled value was run to insure that criteria were met 99 percent of the time. The mean value from this data set represents that long-term daily average concentration that needs to be met to achieve water quality standards.

Instream flow measurements were not available for the sample sites. However, there is a USGS gaging station located on Beaverdam Branch with flow data beginning in 1992 and continuing through early 2006. The average flow at the gage is 197.92 cfs and the contributing area is 72.29 square miles. Based on these numbers, one (1) square mile contributes 2.74 cfs of flow to the stream. Contributing area was calculated in square miles for each sample point (Table B1) and then multiplied by 2.74 cfs to determine an estimated flow at each point.

Table B1. Mill Run Sample Point Areas Sample Point Area (mi2) Flow (MGD) BVDM5.0 6.51 11.53 BVDM6.0 9.88 17.50 BVDM7.0 10.15 17.97 BVDM8.0 13.19 23.36

Mill Run above BVDM5.0

Mill Run above point BVDM5.0 has been determined to be meeting its designated use from its headwaters to the Allegheny Reservoir. Mill Run is then listed as impaired due to sediment from small residential runoff from the Allegheny Reservoir to point BVDM5.0. The majority of headwaters of Mill Run is forested, with small residential areas closer to the city limits.

The TMDL for this section of Mill Run consists of a load allocation to all of the watershed area above point BVDM5.0. Addressing the small residential impacts above this point addresses the

impairment for the segment. The estimated flow at BVDM5.0 is 11.53 MGD. The load allocations made at point BVDM5.0 for this stream segment are presented in Table B2.

Table B2. Reductions forMill Run Above BVDM5.0 Measured Sample Reduction Data Allowable Identified

Conc. Load LTA Conc. Load (mg/l) (lb/day) (mg/l) (lb/day) Percent Fe 1.54 148.1 0.33 31.7 79 Al 1.05 101.0 0.13 12.5 88 All values shown in this table are long-term average daily values.

The TMDL for Mill Run at point BVDM5.0 requires that a load allocation be made for all areas above BVDM5.0 for iron and aluminum.

Mill Run between BVDM5.0 and BVDM6.0

Mill Run between BVDM5.0 and BVDM6.0 represents Mill Run between points BVDM5.0 and BVDM6.0. This segment of Mill Run has been determined to be impaired due to urban and small residential runoff. Mill Run has entered the urban setting and receives stormwater from pipes and is channelized in sections.

The TMDL for this section of Mill Run consists of a load allocation to all of the watershed area between points BVDM5.0 and BVDM6.0. Addressing the urban and small residential runoff runoff impacts above this point addresses the impairment for the segment. The estimated flow at BVDM6.0 is 17.50 MGD. The load allocations made at point BVDM6.0 for this stream segment are presented in Table B3.

B3. Long Term Average (LTA) for Mill Run Between Points BVDM5.0 and BVDM6.0 Measured Sample Data Allowable Conc. (mg/l) Load (lb/day) LTA Conc. (mg/l) Load (lb/day) Fe 2.98 434.9 0.47 68.6 Al 1.78 259.8 0.18 26.3 All values shown in this table are long-term average daily values.

The loading reduction for point BVDM5.0 was used to show the total load that was removed from upstream sources. For each parameter, the total load that was removed upstream was subtracted from the existing load at point BVDM6.0. This value was compared to the allowable load at point BVDM6.0. Reductions at point BVDM6.0 are necessary for any parameter that exceeds the allowable load at this point. Necessary reductions at point BVDM6.0 are shown in Table B4.

Table B4. Reductions Necessary at Point BVDM6.0 Iron (lb/day) Aluminum (lb/day) Existing Load at BVDM6.0 434.9 259.8 Existing load from upstream points (BVDM5.0) 148.1 101.0 Difference of existing load and upstream existing load 286.8 158.8 Allowable loads from upstream points 31.7 12.5 Total load at BVDM6.0 318.5 171.3 Allowable load at BVDM6.0 68.6 26.3 Load Reduction at BVDM6.0 (Total load at 249.9 145.0 BVDM6.0 – Allowable load at BVDM6.0) Percent Reduction required at BVDM6.0 78 85

The TMDL for Mill Run at point BVDM6.0 requires that a load allocation be made for all areas above BVDM6.0 for iron and aluminum.

Mill Run Between BVDM6.0 and BVDM7.0

Mill Run between BVDM6.0 and BVDM7.0 represents Mill Run between points BVDM6.0 and BVDM7.0. This segment of Mill Run has been determined to be impaired due to urban runoff. Point BVDM7.0 is located in the Altoona City limits and receives runoff from stormwater pipes and overland flow.

The TMDL for this section of Mill Run consists of a load allocation to all of the watershed area above point BVDM7.0. Addressing the urban runoff impacts above this point addresses the impairment for the segment. The estimated flow at BVDM7.0 is 17.97 MGD. The load allocations made at point BVDM7.0 for this stream segment are presented in Table B5.

B5. Long Term Average (LTA) for Mill Run Between BVDM6.0 and BVDM7.0 Measured Sample Data Allowable Conc. (mg/l) Load (lb/day) LTA Conc. (mg/l) Load (lb/day) Fe 3.14 470.6 0.57 85.4 Al 1.61 241.3 0.27 40.5 All values shown in this table are long-term average daily values.

The loading reduction for point BVDM6.0 was used to show the total load that was removed from upstream sources. For each parameter, the total load that was removed upstream was subtracted from the existing load at point BVDM7.0. This value was compared to the allowable load at point BVDM7.0. Reductions at point BVDM7.0 are necessary for any parameter that exceeds the allowable load at this point. Necessary reductions at point BVDM7.0 are shown in Table B6.

Table B6. Reductions Necessary at Point BVDM7.0 Iron (lb/day) Aluminum (lb/day) Existing Load at BVDM7.0 470.6 241.3 Existing load from upstream points (BVDM6.0) 434.9 259.8 Difference of existing load and upstream existing load 35.7 -18.5 Percent load loss due to instream process 0 7 Allowable loads from upstream points 68.6 26.3 Percent load remaining at BVDM7.0 100 93 Total load at BVDM7.0 104.3 24.5 Allowable load at BVDM7.0 85.4 40.5 Load Reduction at BVDM7.0 (Total load at BVDM7.0 18.9 0 – Allowable load at BVDM7.0) Percent Reduction required at BVDM7.0 18 0

The TMDL for Mill Run at point BVDM7.0 requires that a load allocation be made for all areas above BVDM7.0 for iron.

Mill Run Between BVDM7.0 and BVDM8.0

Mill Run between BVDM7.0 and BVDM8.0 represents Mill Run between points BVDM7.0 and BVDM8.0. This segment of Mill Run has been determined to be impaired due to urban runoff. Point BVDM8.0 is located at the mouth of Mill Run before it enters Beaverdam Branch. Numerous stormwater pipes empty into the stream channel during rain events.

The TMDL for this section of Mill Run consists of a load allocation to all of the watershed area above point BVDM8.0. Addressing the small residential impacts above this point addresses the impairment for the segment. The estimated flow at BVDM8.0 is 23.36 MGD. The load allocations made at point BVDM8.0 for this stream segment are presented in Table B7.

B7. Long Term Average (LTA) for Mill Run Between BVDM7.0 and BVDM8.0 Measured Sample Data Allowable Conc. (mg/l) Load (lb/day) LTA Conc. (mg/l) Load (lb/day) Fe 3.18 619.5 0.22 42.9 Al 1.38 268.9 0.12 23.4 All values shown in this table are long-term average daily values.

The loading reduction for point BVDM7.0 was used to show the total load that was removed from upstream sources. For each parameter, the total load that was removed upstream was subtracted from the existing load at point BVDM8.0. This value was compared to the allowable load at point BVDM8.0. Reductions at point BVDM8.0 are necessary for any parameter that exceeds the allowable load at this point. Necessary reductions at point BVDM8.0 are shown in Table B8.

Table B8. Reductions Necessary at Point BVDM8.0 Iron (lb/day) Aluminum (lb/day) Existing Load at BVDM8.0 619.5 268.9 Existing load from upstream points 470.6 241.3 Difference of existing load and upstream existing load 148.9 27.6 Allowable loads from upstream points 85.4 40.5 Total load at BVDM8.0 234.3 68.1 Allowable load at BVDM8.0 42.9 23.4 Load Reduction at BVDM8.0 (Total load at BVDM8.0 191.4 44.7 – Allowable load at BVDM8.0) Percent Reduction required at BVDM8.0 82 66

The TMDL for Mill Run at point BVDM8.0 requires that a load allocation be made for all areas above BVDM8.0 for iron and aluminum.

Margin of Safety (MOS)

For each TMDL calculated in this study the MOS is applied implicitly. A MOS is built in because the allowable concentrations and loadings were simulated using Monte Carlo techniques and by employing the @Risk software. Other margins of safety used for this TMDL analysis include the following:

• Effluent variability plays a major role in determining the average value that will meet water-quality criteria over the long term. The value that provides this variability in our analysis is the standard deviation of the dataset. The simulation results are based on this variability and the existing stream conditions (an uncontrolled system). The general assumption can be made that a controlled system (one that is controlling and stabilizing the pollution load) would be less variable than an uncontrolled system. This implicitly builds in a margin of safety. • A MOS is also the fact that the calculations were performed with a daily iron average instead of the 30 day average.

Seasonal Variation

Seasonal variation is implicitly accounted for in each TMDL because the data used represent all seasons.

Critical Conditions

The reductions specified in each TMDL apply at all flow conditions. A critical flow condition could not be identified from the data used for this analysis.

Attachment C Water Quality Data Used In TMDL Calculations

TMDL Study Point Company Date Flow Fe Al pH Site (gpm) (mg/l) (mg/l) BVDM5.0 SRBC 6/19/2006 2.32 1.294 8.0 BVDM5.0 SRBC 8/2/2006 0.243 0.0812 8.0 BVDM5.0 BVDM5.0 SRBC 9/2/2006 1.33 0.898 7.9 BVDM5.0 SRBC 10/17/2006 3.776 2.946 7.8 BVDM5.0 SRBC 12/5/2006 0.047 0.0217 7.8

Average= 1.54 1.05 7.9 St Dev= 1.55 1.19 0.1

BVDM6.0 SRBC 6/14/2006 0.041 0.0567 8.1 BVDM6.0 SRBC 6/19/2006 5.16 2.319 7.8 BVDM6.0 BVDM6.0 SRBC 9/2/2006 1.21 1.061 8.0 BVDM6.0 SRBC 10/17/2006 5.512 3.696 7.7

Average= 2.98 1.78 7.9 St Dev= 2.76 1.58 0.18

42 BVDM7.0 SRBC 6/14/2006 3.17 1.779 7.3 BVDM7.0 SRBC 6/19/2006 4.40 2.09 7.7 BVDM7.0 BVDM7.0 SRBC 9/2/2006 0.854 0.374 7.8 BVDM7.0 SRBC 10/17/2006 4.127 2.193 7.8

Average= 3.14 1.61 7.65 St Dev= 1.61 0.84 0.24

BVDM8.0 SRBC 6/19/2006 10.7 3.79 7.7 BVDM8.0 SRBC 8/2/2006 0.112 0.073 8.2 BVDM8.0 BVDM8.0 SRBC 9/2/2006 1.009 0.416 7.9 BVDM8.0 SRBC 10/17/2006 4.003 2.598 7.8 BVDM8.0 SRBC 12/5/2006 0.062 0.0205 8.1

Average= 3.18 1.38 7.94 StDev= 4.50 1.72 0.21

Attachment D YSI Continuous Sonde Data

The majority of the rain during this storm event fell on November 16, 2006. The following charts show that the dissolved oxygen (DO) concentration remained above the water quality standard of 5.0 mg/l below the CSOs.

Mill Run downstream of Altoona West CSO

Dissolved Oxygen 13:00 11/14/2006 to 11:30 11/21/2006 14

8 L

mg/ 6

0 0 0 0 0 0 0 0 0 0 0 :00 :00 0 :00 :0 :00 :00 :0 :00 0 0 :00 :0 :0 :0 :00 :0 :00 :00 :0 8 0 6:00 0: 6 8 0 6 2 8 0: 6: 2 8 0 6:00 8 0:00 6 2 8 0 6:00 1 12:0018:00 12:001 1 1 1 1 12 1 1 1 11/15/2006 11/16/2006 11/17/2006 11/18/2006 11/19/2006 11/20/200611/21/2006

Attachment E Comment and Response

No formal comments were received for the Mill Run Watershed TMDL.