Approaches for Determining and Complying with TMDL Requirements Related to Roadway Stormwater Runoff (2019)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

77 This chapter shows the state DOT how to determine the removal performance of proposed BMPs. It provides state DOT practitioners with an overview of methodologies and references to compare compliance strategies with regard to pollutant removal performance. No single solu- tion is optimal for all sites due to variability in site constraints, feasibility considerations, and design objectives. Therefore, the intent of this chapter is not to provide direct comparisons of BMP effluent concentrations or removal efficiencies but to serve as a guide to applicable data sources, tools, and protocols for evaluating BMP performance. Performance Evaluation Methodology The procedure described herein provides a framework for comparing BMP performance based on pollutant removal (Figure 23). This framework is intended to be flexible to allow practitioners the ability to select appropriate methodologies most applicable to a specific location. Prior to conducting a performance evaluation, applicable compliance strategies should be selected based on identifying applicable treatment processes for the POC (see Chapter 4). The stepwise performance comparison methodology highlights key consider- ations, references, and example approaches to serve as a starting point for planning solutions that meet TMDL performance objectives. Each step in this procedure is described in detail within the sections below. Step 1. Identify Evaluation Metrics Understanding the implementation objectives is critical to planning a compliance strategy. Two types of implementation requirements are applicable to state DOT TMDL compliance: 1. Standard approach: The WLA attributed to the state DOT—or jointly specified as part of an MS4—is related directly to a percent load reduction target relative to a baseline condition. The targets may be translated to specific site design requirements as specified in the MS4 permit. This approach measures performance targets on a load reduction basis, which may be expressed as a percentage or absolute reduction. 2. Presumptive approach: In cases where specific WLAs are not developed due to dominant sources or characteristics of the POC or watershed complexities, progress toward meeting a TMDL is based on degree of BMP implementation, total loading relative to the assimilation capacity of the receiving water, and overall trends in receiving water conditions. Since specific percent reduction targets are not set, this approach measures compliance with concentration- based water quality objectives or criteria for the receiving water. The concentration-based objectives consist of the numeric targets identified by the regulators with regard to a specific pollutant. C H A P T E R 6 Best Management Practice Pollutant Removal Performance

78 Approaches for Determining and Complying with TMDL Requirements Related to Roadway Stormwater Runoff Table 39 compares load reduction and effluent target objectives applicable to the standard and presumptive approaches previously described. BMP performance evaluations should consider evaluation metrics that are most applicable to demonstrating TMDL compliance. Step 2. Determine Scale of Comparison With regard to TMDL compliance, project practitioners may be interested in comparing BMP performance on either the site or watershed scale (Table 40). Site-specific evaluations are used to compare BMP design options for a defined location. Watershed scale evaluations aim to plan BMP implementation levels (e.g., acres treated) or specific locations (e.g., near outfalls) to achieve the greatest benefit. Pollutant load reduction is one metric to consider when prioritizing and assessing feasibility of projects on the watershed scale (see Chapter 5). Step 3. Select Evaluation Approach and Pollutant Removal Algorithm The objective of this step is to identify an approach that can be used to compare BMP per- formance for project planning and BMP selection. The framework described below is used to assess BMP performance on the site scale. However, it can be replicated for watershed scale evaluations. For example, performance could be assessed for a representative 1-acre drainage area and then scaled up to estimate the acres requiring treatment to achieve the water qual- ity objectives. The methodology for evaluating load reduction and effluent target metrics is described based on quantifying flow pathways (Figure 24). The figure shows the equation used to estimate load (L), which is composed of concentration (C) and volume (V). This approach can be followed regardless of BMP type or selected model. Applicable modeling tools to evaluate components of this framework are described in Step 4. 4. Conduct Comparative BMP Performance Assessment 1. Identify Evaluation Metrics 2. Determine Scale of Comparison 3. Select Evaluation Approach and Pollutant Removal Algorithm Compliance Strategies Figure 23. BMP performance comparison methodology flowchart. Metric Applicable Approach Objective Description Example Objective Load Reduction Standard The mass of a pollutant removed via volume reduction or treatment Annual or design storm load reduction for a specified pollutant (e.g., 85% TSS removal) Effluent Target Standard and presumptive A specified concentration or load that is required to be met within the receiving water or for discharge from a catchment to the receiving water Nonexceedance threshold for stormwater discharge concentration (i.e., 120 µg/L zinc) or discharge load (e.g., 0.5 lb/acre or lb/day TP) Degree of BMP Implementation Presumptive Pollutant removal is assumed based on the characteristics of BMP implementation. Percent of watershed or impervious area treated Table 39. Comparison of load reduction and effluent target objective.

Best Management Practice Pollutant Removal Performance 79 Best Management Practice Sizing. With respect to pollutant removal performance, struc- tural BMP sizing can generally be grouped into two categories based on type of BMP: • Storage-based: The BMP is sized to capture a certain runoff volume within the storage capacity of the system before overflow occurs. Examples of storage-based practices include bioretention, detention ponds, and permeable pavement. • Flow-based: The BMP is sized based on the capacity of a BMP to provide treatment up to a certain flow rate. Examples of flow-based BMPs include swales, vegetated filter strips, and media filters. Design manuals that account for local conditions can be consulted to determine BMP sizing requirements and acceptable sizing methods. BMP sizing may be an iterative process, by which sizing parameters are adjusted to achieve a target load reduction. Best Management Practice Capture Efficiency. Capture efficiency refers to the proportion of the inflow volume that is treated by the BMP and does not bypass or overflow. The captured volume is shown in Figure 24 as the influent arrow (Linf). Capture efficiency is dependent on the storage volume and treatment–infiltration rate of the BMP relative to catchment runoff volumes and flow rates. Inflows that are not treated by the BMP bypass the practice and remain untreated. Comparison Site Specific Watershed Scale Objective • Compare BMP options at a specific site with defined constraints, or determine if a given project meets project-specific limits. • Compare a limited number of implementation options on a watershed scale to achieve a specified objective or to maximize cost benefit. Design Phase • Site conceptual design • Watershed planning BMP Location • Determined • Flexible Evaluation Components • Defined drainage area (soils, land use, and size) • Conceptual BMP design (size and drainage configuration) • Pollutant removal algorithm for considered BMPs (Step 3) • Existing conditions, watershed scale hydrologic, and water quality model • BMP prioritization and selection methodology • Simplified pollutant removal algorithm for considered BMPs (Step 3) • Defined alternatives Anticipated Results • Ranking of BMP design options • Selection of priority projects Table 40. Performance evaluation considerations based on scale. Figure 24. Conceptual model for calculation of load reduction based on flow pathways (Taylor et al. 2014A).

80 Approaches for Determining and Complying with TMDL Requirements Related to Roadway Stormwater Runoff For calculation of load reduction, capture efficiency can be determined using long-term con- tinuous hydrologic simulation methods (or relationships derived from continuous simulation methods). Pollutant Removal Algorithms by Flow Pathway. BMP effluent can be grouped into three flow pathways: bypass or overflow volume, volume loss, and treated effluent. Bypass or overflow volume is determined based on the design of the BMP capture efficiency. Bypass occurs for BMPs that are offline. Overflow occurs for BMPs that are online. Volume can be reduced through reuse, infiltration, or evapotranspiration pathways. Estimation of these pathways requires quantifica- tion of proportions relative to inflow volumes, which are dependent on infiltration (underlying soil conditions and retention storage), evapotranspiration (plant selection and climate), and reuse (storage capacity and use demand schedule). Treated effluent consisting of inflows that are captured by the BMP and are not removed by volume loss exit the BMP as treated effluent. Upon completion of a hydrologic evaluation, anticipated loads can be determined for each flow pathway (Figure 25). For bypass volume, the concentration is often assumed to be unchanged relative to the influent concentration (although there may be some removal for BMPs that over- flow rather than bypass). Volume losses result in complete removal of pollutant load for the associated volume. However, fate and transport of dissolved constituents—such as chloride and nitrate—to groundwater and downstream reservoirs should be considered. Inflows that are treated by the BMP are assigned a change in concentration based on a selected pollutant removal algorithm. Depending on the POC and BMP type, different pollutant removal algorithms are applicable (Table 41) (Roesner et al. 2013). Determining Performance Metrics. BMP performance can be quantified either by load reduction or effluent target metrics (Table 39). BMP load reductions are determined based on the difference between the cumulative outflow and inflow mass of a POC within a specified period (typically a year or a defined water quality design storm). Effluent concentrations can be directly compared to any applicable target concentration values. Replicating this analysis for various BMPs and design configurations allows for comparison of achieved performance between implementation options. Several tools are available to assist in the calculation of POC removal, as described in Step 4. Step 4. Conduct Comparative Best Management Practice Performance Assessment This section describes applicable sources of BMP performance data and relevant tools to conduct comparative analysis using one or more of the BMP algorithms previously described. Best Management Practice Performance Data A critical input in comparing applicable compliance strategies is the anticipated pollutant removal effectiveness of BMPs. The BMPDB is a robust repository for information on BMP performance. However, limitations exist in evaluating the performance of certain POCs and BMP types. Reference studies can provide valuable information for certain BMP types and pollutants when information is available (Table 42). International Stormwater Best Management Practices Database. The BMPDB is a com- prehensive international data repository for stormwater BMP monitoring projects. The data- base includes influent and effluent BMP hydrology and water quality data, as well as metadata on watershed and site characteristics (Wright Water Engineers and Geosyntec Consultants 2017B). As of January 2017, the BMPDB contains data sets from nearly 650 BMP perfor- mance studies with more than 300,000 water quality records for more than 430 POCs (Wright Water Engineers and Geosyntec Consultants 2017B). More than 150 of these studies are for

Best Management Practice Pollutant Removal Performance 81 transportation-related land uses representing 10 states. The transferability of BMPDB data to state DOT scenarios and the impact of geography and land use parameters on performance are described in Geosyntec Consultants and Wright Water Engineers, Inc. (2014), and a state DOT- specific BMPDB portal is currently planned. The online BMPDB web platform can be used to conduct queries to download specific per- formance information of interest, including paired influent–effluent data for specific POCs and BMP types. Due to the limited amount of available data for some POC and BMP combinations, the number of data pairs available within the BMPDB should be considered when evaluating the representativeness of results. Aggregated summary statistics for specific BMP types and POCs are published in BMPDB Statistical Summary Reports (Wright Water Engineers and Geosyntec Consultants 2017B). Note: BMPs with horizontal-line reduction rates are not influenced by the infiltration rate. Figure 25. Comparison of annual load reduction performance and infiltration rate for six structural BMPs.

82 Approaches for Determining and Complying with TMDL Requirements Related to Roadway Stormwater Runoff Pollutant Removal Algorithm Description Applicable Pollutants Applicable BMPs Particle Settling Effluent concentrations are determined based on particle-settling theory, associating the proportion of influent particle-size distribution classes removed based on residence time and BMP geometry. • TSS and those associated with TSS (see Pollutant Partitioning and Mass Fraction algorithm) • Detention ponds • Filter strips • Vegetated swales • Wet ponds First-Order Decay with Irreducible Constant Effluent concentrations are estimated based on a first-order decay model based on influent concentration and hydraulic loading. The rate constant (k) and irreducible concentration level (C*) are required model inputs. Rate constants may not be available or confirmed for many constituents and BMPs. • All (dependent on researched rate constants) • Detention ponds • Vegetated swales • Wet ponds Influent– Effluent Regression Linear regression equations are developed relating effluent to influent concentrations based on monitoring data. The BMPDB serves as the largest known repository of monitoring data for which regression equations have been developed. • All (dependent on data availability and developed regression equations) • Bioretention • Detention ponds • Filter strips • Media filters • Permeable pavement • Vegetated swales • Wetlands • Wet ponds (dependent on data availability and developed regression equations) Pollutant Partitioning and Mass Fraction Pollutant removal of particulate-bound pollutants can be estimated using TSS as a surrogate. The fraction of the pollutant that is particulate bound (partitioning) and the mass fraction associated with different particle sizes are used to relate the POC to known TSS removal parameters. • Particulate-bound fractions of metals, nutrients, organic compounds, and bacteria • All Volume Reduction Pollutant removal associated with volume loss (infiltration, evapotranspiration, or reuse) is determined based on hydrologic modeling. • All • Bioretention (elevated or no underdrains) • Cisterns • Infiltration basins • Permeable pavement (elevated or no underdrains) Table 41. Comparison of BMP pollutant removal algorithms (adapted from Roesner et al. 2013). Best Management Practice Performance Data by POC. Practitioners are directed to the selected open source references provided in Table 42 for initial BMP performance evaluation information for specific pollutant categories. Extensive BMP performance research exists and is ongoing. Therefore, practitioners are encouraged to conduct literature reviews of specific POCs and BMPs as a preliminary task in any TMDL compliance planning project. Regional design manuals may also provide valuable information with regard to sizing requirements to achieve performance targets. Nonstructural Best Management Practice Performance Data. Nonstructural BMPs are inherently difficult to characterize with regard to pollutant removal due to system and land- scape variability. Monitoring of nonstructural practices is difficult due to undefined inflow and outflow locations, resulting in limited data availability. In addition, nonstructural BMP

Best Management Practice Pollutant Removal Performance 83 performance is likely to be highly site-specific and may vary over time due to public awareness, materials used in the manufacturing of vehicle parts, and technological advancement. While many studies have evaluated nonstructural BMP performance (Herrera et al. 2006, Taylor et al. 2014A), currently, no data repositories or quantitative performance evaluation planning tools that compare nonstructural urban stormwater BMP types are known to exist. Performance considerations of nonstructural BMPs are shown in Table 43. Available Tools for Best Management Practice Performance Evaluation A wide variety of models could be used for assessment of BMP performance, ranging from simple spreadsheet-based tools to advanced continuous simulation methods. These tools allow for comparison between BMP types based on available performance data (Table 44) and site-specific hydrologic evaluations. The EPA TMDL modeling website provides links Pollutant Category Key Performance Evaluation References Pathogens • Colorado E. Coli Toolbox: A Practical Guide for Colorado MS4s (Wright Water Engineers and Geosyntec Consultants 2016) • Effectiveness Evaluation of Best Management Practices for Stormwater Management in Portland, Oregon (Herrera et al. 2006) • Pathogens in Urban Stormwater Systems (American Society of Civil Engineers Environmental and Water Resources Institute 2014) • Statistics for Stochastic Modeling of Volume Reduction, Hydrograph Extension, and Water-Quality Treatment by Structural Stormwater Runoff Best Management Practices (BMPs) (Granato 2014) Sediment • Effectiveness Evaluation of Best Management Practices for Stormwater Management in Portland, Oregon (Herrera et al. 2006) • Performance Comparison of Structural Stormwater Best Management Practices (Barrett 2005) • Statistics for Stochastic Modeling of Volume Reduction, Hydrograph Extension, and Water-Quality Treatment by Structural Stormwater Runoff Best Management Practices (BMPs) (Granato 2014) • Stormwater Best Management Practices (BMP) Performance Analysis (Tetra Tech 2010) Nutrients • International Stormwater Best Management Practices (BMP) Database Pollutant Category Summary: Nutrients (Wright Water Engineers and Geosyntec Consultants 2010) • Effectiveness Evaluation of Best Management Practices for Stormwater Management in Portland, Oregon (Herrera et al. 2006) • Nutrient (Nitrogen/Phosphorus) Management and Source Control (Leisenring et al. 2014) • Performance Comparison of Structural Stormwater Best Management Practices (Barrett 2005) • Statistics for Stochastic Modeling of Volume Reduction, Hydrograph Extension, and Water-Quality Treatment by Structural Stormwater Runoff Best Management Practices (BMPs) (Granato 2014) Metals • Effectiveness Evaluation of Best Management Practices for Stormwater Management in Portland, Oregon (Herrera et al. 2006) • International Stormwater Best Management Practices (BMP) Pollutant Category Summary: Metals (Wright Water Engineers and Geosyntec Consultants 2011) • Performance Comparison of Structural Stormwater Best Management Practices (Barrett 2005) • Statistics for Stochastic Modeling of Volume Reduction, Hydrograph Extension, and Water-Quality Treatment by Structural Stormwater Runoff Best Management Practices (BMPs) (Granato 2014) Organic Enrichment/ Oxygen Depletion (See Pollutant Category, Nutrients references) Salinity/Dissolved Constituents • Methods for Evaluating Potential Sources of Chloride in Surface Waters and Groundwaters of the Conterminous United States (Granato et al. 2015). • Toxicological Effects of Chloride-Based Deicers in the Natural Environment: Synthesis of Existing Research (Fay et al. 2014) Table 42. Selected performance evaluation references by pollutant category.

84 Approaches for Determining and Complying with TMDL Requirements Related to Roadway Stormwater Runoff Nonstructural BMP Factors Affecting Performance Quantifiable Variables Example Evaluations Street Sweeping Type of technology and equipment used, frequency, timing in relation to rainfall, road condition, location Amount of debris and sediment removed, lane miles swept; TSS and/or metals concentrations in runoff • Street Sweeping for Pollutant Removal (Curtis 2002) • Deriving Reliable Pollutant Removal Rates for Municipal Street Sweeping and Storm Drain Cleanout Programs in the Chesapeake Bay Basin (Center for Watershed Protection 2008) • Evaluation of Street Sweeping as a Stormwater-Quality- Management Tool in Three Residential Basins in Madison, Wisconsin (Selbig and Bannerman 2007) • Quantifying Nutrient Loads Associated with Urban Particulate Matter (PM), and Biogenic/Litter Recovery Through Current MS4 Source Control and Maintenance Practices (Sansalone et al. 2011) Catch-Basin Cleaning Schedule, timing, location, type of technology and techniques used Amount of debris and sediment removed; TSS and/or metals concentrations in runoff • Deriving Reliable Pollutant Removal Rates for Municipal Street Sweeping and Storm Drain Cleanout Programs in the Chesapeake Bay Basin (Center for Watershed Protection 2008) • Nutrient and Carbon Loading in Gross Solids in Urban Catch Basins (Hunt et al. 2015) • Quantifying Nutrient Loads Associated with Urban Particulate Matter (PM), and Biogenic/Litter Recovery Through Current MS4 Source Control and Maintenance Practices (Sansalone et al. 2011) Vegetation Management Irrigation rates, runoff prevention, fertilization and pesticide plans, type of vegetation used, staff education, leaf management Runoff reduction, reduction in material use; nutrient and pesticide concentrations in runoff • Reduced River Phosphorus Following Implementation of a Lawn Fertilizer Ordinance (Lehman et al. 2009) • Evaluation of Leaf Removal as a Means to Reduce Nutrient Concentrations and Loads in Urban Stormwater (Selbig 2016) Erosion Control Channel restoration, BMP installation, frequency and location of inspections, design standards TSS and/or nutrient concentrations in runoff • A Review and Evaluation of Literature Pertaining to the Quantity and Control of Pollution from Highway Runoff and Construction (Barrett et al. 1995B) • Impacts of Compost Blankets on Erosion Control, Revegetation, and Water Quality at Highway Construction Sites in Iowa (Glanville et al. 2003) Table 43. Performance considerations of nonstructural BMPs.

Best Management Practice Pollutant Removal Performance 85 to many relevant models useful for evaluating BMP performance related to TMDLs (EPA 2018C). However, it is important to consider that local precedent may dictate use of a cer- tain approach. Planning-level assessments are likely to be useful to state DOT practitioners interested in accurate but time-sensitive methods for quantifying BMP performance. Three relevant tools for this effort are described as follows and compared in Table 45. • EPA National Stormwater Calculator (EPA 2017B): The stormwater calculator is a user- friendly web application that compares the hydrologic performance of various LID practices. National soil, land cover, slope, and meteorological data sets are dynamically linked to the selection of a project location. Runoff volumes are determined based on a long-term simulation using EPA’s Storm Water Management Model (SWMM) (EPA 2017C). State DOT practitioners can use the stormwater calculator for planning-level esti- mates of stormwater runoff volumes, for sizing LID controls for various scenarios, for Nonstructural BMP Factors Affecting Performance Quantifiable Variables Example Evaluations Traction Control Plans Choice of material, weather, timing, rate of application, application technology, staff training, post-season clean up, use of winter severity index and road weather information systems Reduction in quantity of material used and resulting source load reduction, reduction in TSS and chloride concentrations in runoff • Identifying the Parameters for Effective Implementation of Liquid-Only Plow Routes (Peterson et al. 2010) • Quantifying the Impact that New Capital Projects Will Have on Roadway Snow and Ice Control Operations (Sullivan et al. 2017) • Benefit–Cost of Various Winter Maintenance Strategies (Fay et al. 2015) • MassDOT Snow and Ice Control Program: 2017 Environmental Status and Planning Report (Massachusetts Department of Transportation 2017) Public Education on Littering Long-term budgeting, inspection frequency, availability of trash receptacles, frequency of trash pickup Amount of trash captured, reduction in bacteria concentrations in runoff • Key Findings of Don’t Mess with Texas Survey (Baselice & Associates, Inc. 2017) • California Department of Transportation District 7 Litter Management Pilot Study (California Department of Transportation 2000) • Littering Behavior in America: Results of a National Study (Action Research 2009) Spill Prevention and Response Plans Staff training, availability of resources, timing of response Reduction in number of spills, timing to spill response, reduction in nutrient and/or metal concentrations • NA Elimination of Nonstormwater Inflow Watertight joints and pipe coatings, pipe elevation adjustments, backflow preventers, illicit discharge detection and elimination Reduction in runoff volume and/or volume treated by BMP, reduction in influent nutrient and bacteria loads • Illicit Discharge Detection and Elimination: A Guidance Manual for Program Development and Technical Assessments (Brown et al. 2004) • Illicit Discharge Detection and Elimination: Low Cost Options for Source Identification and Trackdown in Stormwater Systems (Irvine et al. 2011) Note: NA = not available. Table 43. (Continued).

86 Approaches for Determining and Complying with TMDL Requirements Related to Roadway Stormwater Runoff Tool EPA National Stormwater Calculator NCHRP Report 792 SELDM Included BMPs • Bioretention (rain garden and street planter) • Cisterns (rain harvesting) • Green roof • Impervious area disconnection • Infiltration basin • Permeable pavement • Bioretention • Dry detention • Filter strip • PFC pavement • Sand filter • Swale • Wet pond BMPs and associated hydrologic and water quality statistic distributions are user defined. Input statistics for the following BMP types using BMPDB designations are defined in Granato (2014): • Bioretention • Composite • Detention basin • Biofilter (swale) • Infiltration basin • Manufactured device • Media filter • Retention pond • Wetland basin • Wetland channel Hydrologic Calculation Methodology Long-term simulations are embedded in the program using SWMM as the computational engine. Runoff volumes and volume bypassed, treated, and lost are estimated from hydrologic performance curves developed using SWMM long-term simulation and defined in the spreadsheets using performance nomographs. Runoff volumes are determined based on statistical distributions of input variables for a selected location. The impact of a BMP is determined by paired statistical distributions irrespective of BMP sizing. Pollutant Removal Algorithm Volume reduction only; pollutant loads are not estimated. Influent–effluent regressions developed from the BMPDB are embedded in the model for each BMP and the following constituents: • Bacteria: E. coli, fecal coliform • Metals: Cu, Pb, Zn • Nutrients: NO3, TKN, TN, TP, DP • Sediment: TSS Export of pollutants is excluded such that effluent concentrations can never exceed influent concentrations. Statistical distributions of the ratio of influent to effluent concentrations from the BMPDB are used to define BMP performance. Input statistics for 11 BMP types are defined in Granato (2014). Key Features • Soils, slope, land cover, and meteorological data are dynamically linked to national data sets for the user-selected location. • Cost module allows for comparison of BMP construction costs using dynamically updated regional cost factors. • Influent runoff quality is defined based on highway runoff monitoring data. • BMP sizing parameters can be adjusted to investigate impact on performance. • Whole life-cycle costing tool allows for calculation of cost of annual load removal ($/lb). Dilution factors and defined water body flow and water quality parameters can be used to assess the effects of BMPs on storm event hydrographs and downstream water body concentrations. Table 44. Comparison of BMP performance evaluation tools. estimating costs for LID controls, and for quickly accessing and viewing soils and meteo- rological data. • NCHRP Report 792: Long-Term Performance and Life-Cycle Costs of Stormwater Best Manage- ment Practices (Taylor et al. 2014A): Seven spreadsheet-based tools (bioretention, dry detention basins, grassed filter strips, PFC pavement, grassed swales, sand filters, and wet ponds)—one for each included BMP—are available to estimate average annual runoff volume reductions, pollutant load reductions, and whole life-cycle costs of each BMP. Runoff volumes and volume captured and lost (through infiltration and evapotranspiration) are estimated from hydro- logic performance curves developed using SWMM (EPA 2017C). Pollutant concentration

Best Management Practice Pollutant Removal Performance 87 reductions are based on influent–effluent regression equations developed from the BMPDB. Hydrologic performance nomographs representing continuous simulation results are avail- able for 343 rain gages included in each tool. A protocol for comparing BMP load reduction using the NCHRP Report 792 spreadsheets is presented in the upcoming section: Structural Best Management Practice Load Reduction Comparison Protocol Using NCHRP Report 792. • SELDM (Granato 2013): SELDM uses statistical distributions of runoff and water quality parameters to estimate the probability distributions of runoff flow volumes, runoff loads, impact of stormwater BMPs, and receiving water body concentrations. SELDM employs user-supplied information and a database of statistics derived from national data sets to com- pute hydrologic and water quality parameters. Location-specific meteorological data sets are available for 2,610 hourly precipitation stations distributed across the country. SELDM uses a statistical approach for estimating BMP performance based on the ratio of influent and efflu- ent data pairs derived from BMPDB to stochastically estimate volume and load reductions (Granato 2014). The summarized statistics can be used to compare performance of various BMP types or evaluate the probability of improving downstream water quality. Structural Best Management Practice Load Reduction Comparison Protocol Using NCHRP Report 792. The objective of this section is to explain the use of NCHRP Report 792 BMP evaluation tools and provide an example comparison of annual load reduction perfor- mance for the BMP types (Taylor et al. 2014A). The user guide provided in Taylor et al. (2014A) should be consulted in conjunction with the guidance provided in this section. A procedure for comparing BMPs based on annual load reduction and corresponding example results is Design Component BMP Type Bioretention with Underdrain Bioretention without Underdrain/ Infiltration Basina Dry Detention Wet Pond Sand Filter Storage Volume (ft3) 1,900 1,900 1,900 1,900 1,900 Infiltration Present? Yes Yes No (impermeable liner) No No Ponding/Surcharge Depth (ft) 1 1 3 1 1 Permanent Pool Depth (ft) – – – 3 – Media Thickness (ft) 2 – – – 2 Media Filtration Rate (in./hr) 2 – – – 2 Stone Reservoir Thickness (ft) 1 – – – – Underdrain Present? Yes No No No Yes Underdrain Discharge Elevation (ft) 0.5 – – – 0 Approximate Footprint at Top of Freeboard (ft2)b 1,260 2,250 1,660 1,630 1,780 Note: – = not applicable. aEvaluated using bioretention tool. bAll BMPs assumed to have 1 foot of freeboard with 3:1 side slopes above ponding/surcharge depth. Media storage is included in calculation of footprint area to hold storage volume. Table 45. Storage BMP input parameters for NCHRP Report 792 load reduction comparison example analysis (Taylor et al. 2014A).

88 Approaches for Determining and Complying with TMDL Requirements Related to Roadway Stormwater Runoff provided below to allow replication by state DOT practitioners for local conditions. For each step, the input values evaluated for the example analysis are defined. 1. Define structural BMP types evaluated: The NCHRP Report 792 BMP evaluation tools include individual spreadsheets for the evaluation of seven BMPs: bioretention, dry deten- tion basins, grassed filter strip, PFC pavement, sand filter, grassed swale, and wet pond. To compare BMP types, each individual BMP spreadsheet tool must be run independently and results must be extracted. The selection of BMPs depends on available site information, resulting in two possible evaluation scenarios: a. Site-specific evaluation: When a specific site has been selected for BMP implementation, site constraints dictate feasibility of different BMP types. NCHRP Report 792 tools can be used to compare feasible BMPs for given site constraints. b. Representative comparison: If a specific location has not been selected, BMPs can be com- pared for a representative watershed. For this relative comparison, BMPs are sized based on standardized guidance (local standards or tool defaults) to compare load-reduction benefits of different BMPs. The NCHRP Report 792 BMP evaluation tools include two types of BMPs: storage and flow-based practices (Taylor et al. 2014A). Storage BMPs include bioretention, dry detention, sand filter, and wet pond. The bioretention tool can be used to evaluate bioretention with and without underdrains. The bioretention without underdrain option can also be used to estimate the performance of infiltration basins by considering volume reduction as the sole component of load reduction. For flow-based practices (swales, filter strips, and PFC pave- ment), performance is based on flow parameters and practice area. To compare flow-based BMPs to storage BMPs, a site-specific evaluation is required to define feasible implementa- tion area or assumptions on site geometry. Example Analysis: A representative comparison of annual pollutant load reduction metrics is completed for five storage BMPs: bioretention with underdrain, infiltration basin (simulated using bioretention spreadsheet without specifying an underdrain), dry detention, wet pond, and sand filter. 2. Specify project location: The first worksheet on each of the tools is Project Location. The selection of a rain gage location references the tool to precipitation statistics and performance nomographs that have been developed for BMP storage components based on continuous simulation modeling in SWMM (Taylor et al. 2014A). The 85th percentile, 24-hr storm depth, and annual precipitation depth for the project location can be specified to adjust the values of the nearest provided rain gage. Example Analysis: This analysis has been conducted for Central Lower Michigan based on the following inputs: – State: Michigan – Rain gage: [9] South Central Lower–Jackson 3N – 85th percentile storm depth: 0.80 in. – Average annual precipitation: 30.6 in. 3. Define highway runoff concentrations and costs: The second worksheet of each tool is Project Options. This worksheet allows for the specification of highway runoff concentra- tions and cost inputs. Highway runoff concentrations based on median values from the HRDB are provided as default values (Granato and Cazenas 2009). Cost inputs are needed if load reduction based on cost is a desired output. Costs are defined based on whole life- cycle costs that include both construction and maintenance. Values can be localized using RSMeans location adjustment factors. Alternatively, unit costs can be defined for specific BMP components. Example Analysis: Tool default values for highway runoff concentrations and BMP costs were used.

Best Management Practice Pollutant Removal Performance 89 4. Define tributary area attributes: The first section of the Project Design worksheet on each tool is used to define tributary area input parameters. To set up the comparison of BMP types, the same tributary area is defined for each BMP evaluated. If conducting a representative analysis, any tributary area can be defined because BMPs are sized relative to runoff parameters. The runoff coefficient input can be determined based on ratio of impervi- ous to pervious area for the watershed or site-specific location. For simplicity, a 1-acre tribu- tary area can be used. If conducting a site-specific evaluation, impervious area and hydrologic soil group inputs should be defined to represent local conditions. Example Analysis: The following tributary area inputs were used: – Tributary area: 1 acre – Impervious area: 90% – Tributary area hydrologic soil group: Silt loam (B) – Runoff coefficient: 0.66 5. Define BMP design parameters: The second section of the Project Design worksheet on each tool is used to specify BMP design parameters. The inputs for this section differ for each BMP type. When conducting a representative comparison analysis for storage BMP types, an equivalent storage volume should be specified for each BMP. The storage volume should be set equivalent to local guidelines for BMP sizing. The provided 85th percentile runoff volume can be used to determine an appropriate storage volume if a particular local stan- dard is not specified. Default values for other design input parameters provided in the tools are appropriate unless not representative of local standards or restricted by site constraints. The critical component in entering inputs for the various BMP types is creating representa- tive conditions in which the BMPs could be installed within similar areas. For each BMP evaluated, the individual BMP tool runs with inputs entered for each practice. A critical performance component for BMPs that allow infiltration is the underlying soil infiltration rate. This parameter dictates the achieved volume reduction—for those BMPs with volume loss components—and consequently load reduction. Depending on the infiltra- tion rate of the in situ soil, BMP load reduction may favor an infiltration practice for high infiltration rates versus a filtration practice for low infiltration rates. If a specific design infil- tration rate is known for a site, this value may be used to compare BMP types. Alternatively, a range of infiltration rates may be assessed to predict the load reduction achieved under different infiltration regimes. This approach is taken in the example analysis with underlying infiltration rates defined for 0.05, 0.1, 0.25, 0.5, 1, and 2 in./hr conditions. Repeated evalua- tions are conducted for each infiltration rate input and BMP type to develop the load reduc- tion and infiltration rate relationship. Example Analysis: Storage BMPs were compared using the inputs provided in Table 45. An equivalent storage volume of 1,900 ft3—equivalent to the 85th percentile, 24-hr runoff volume for the selected rain gage—was defined for each BMP type. The bioretention tool with no underdrain specified was used to represent both an infiltration basin and bioreten- tion without underdrain BMP. For both practices, pollutant removal is only associated with volume reduction. Unless specified in Table 45, default tool design parameters were used. 6. Extract load reduction results: The Results Summary Report provides performance statistics in the same format for each BMP tool. The annual load reduction as a percentage is provided as the last row of the Summary of Volume and Pollutant Load Performance Table. To compare BMP performance this information is extracted for each BMP type and scenario run by copy- ing the information out of the Results Summary Report to a separate Excel worksheet. For BMPs that do not include infiltration (Table 45), changing the soil infiltration rate does not result in changes to annual load reduction, and only one row requires evaluation. Further- more, if an estimate of the soil infiltration rate is known for a specific site, then all evaluations can be conducted at that infiltration rate. The following pollutants can be evaluated using the NCHRP Report 792 tools: E. coli, fecal coliform, total copper, total lead, total zinc, NO3,

90 Approaches for Determining and Complying with TMDL Requirements Related to Roadway Stormwater Runoff Note: BMPs with horizontal-line reduction rates are not influenced by the infiltration rate. Bioretention with Underdrains Dry Detention Infiltration Basin Sand Filter Wet PondBioretention without Underdrains 12,5002,000 1,500 1,000 500 0 10,000 7,500 5,000 2,500 0 75,000 60,000 45,000 30,000 15,000 0 Figure 26. Comparison of annual load reduction cost and infiltration rate for six structural BMPs.

Best Management Practice Pollutant Removal Performance 91 TKN, TN, dissolved phosphorus, TP, and TSS. If a specific pollutant or subset of pollutants is of concern for TMDL compliance evaluation, only those pollutants need to be evaluated. Example Analysis: Values were extracted from the NCHRP Report 792 tools for the six BMP input configurations shown in Table 45. Annual load reduction performance and cost (Figure 25) are compared across varied infiltration rates for four constituents (TSS, total lead, TN, and TP). BMPs without infiltration are shown with constant load reduction (i.e., load reduction is only a function of concentration reduction). The results of this analysis are specific to the climate region for the rain gage analyzed and not representative of nationwide conditions. The cost analysis assumes default whole life-cycle cost parameters presented in the NCHRP Report 792 tools (Taylor et al. 2014A). The annual load reduction comparison (Figure 25) illustrates that BMP performance rankings are dependent on both site conditions (underlying infiltration rate) and POC. For the four constituents shown, infiltrating BMPs (bioretention and infiltration basins) gener- ally have greater load reduction performance at higher infiltration rates than detention or filtration BMPs (dry detention, sand filter, and wet pond). Load reduction performance of infiltrating BMPs decreases for lower infiltration rates with performance relative to other practices dependent on the POC (Figure 25). The POC has substantial impact on the relative performance of BMP types. For example, sand filters have the greatest TSS load reduction for infiltration rates less than 0.5 in./hr. However, they rank fourth or fifth out of six practices at infiltration rates less than 0.5 in./hr for TN. Figure 26 uses the same load reduction analysis presented in Figure 25, except the load reductions are analyzed against the whole life-cycle cost of BMPs using the NCHRP Report 792 default cost parameters to determine annual cost. This results in substantially different relative performance between BMP types than assessing load reduction alone. Wet ponds are the most expensive with regard to annual load reduction cost for each of the four con- stituents shown. Infiltration basins and bioretention have low costs of pollutant removal for TN and TP; however, they are more expensive than dry detention for TSS and total lead. The variation in both load reduction and cost in the example analysis illustrates the value in conducting a site-specific performance evaluation to plan BMP implementations that can achieve the greatest overall benefit.