Wabash River Watershed Water Quality Trading Feasibility Study. Final Report

Transcription

1 Wabash River Watershed Water Quality Trading Feasibility Study Final Report September 2011 Prepared for U.S. Environmental Protection Agency Targeted Watershed Grant WS-00E Prepared by Conservation Technology Information Center With support from Tetra Tech, Inc. Kieser & Associates, LLC

2 CTIC would like to thank our project partners for their support: Agri Drain Corporation Duke Energy Indiana Association of Soil and Water Conservation Districts Indiana Farm Bureau Indiana Soybean Alliance Purdue University Extension

3 Contents 1. Introduction What is a WQT Market Feasibility Analysis? What is the Purpose of the Wabash River Watershed WQT Market Feasibility Analysis? What Does This Report Contain? Understanding the Wabash River Watershed Relationship of the Wabash River Watershed to the Gulf of Mexico Overview of Nutrient Sources and Loadings in the Wabash River Watershed Feasibility Analysis Summary Drivers and Incentives for Trading Suitable Pollutants for Trading Watershed Considerations Timing Geographic Scope for Trading Analysis Potential Credit Buyers and Sellers Potential Credit Demand Potential Credit Supply Potential stakeholder participation Putting It All Together Market Analysis and Trading Considerations Pollutant Loads Regulatory Drivers Trade Ratios Baselines Supply Side Credit Generation Differences in Control Costs Other Trading Considerations Next Steps for Water Quality Trading in the Wabash River Watershed Outreach and Education Prioritized Subwatershed for Future Analysis Trading Program Frameworks Conclusion Appendix A: Letter to Illinois EPA from USEPA Appendix B: Wabash River Watershed TMDL Reduction Summaries and Wasteload Allocations (WLAs) Appendix C: Characterization of Wabash River Nutrient Loads Appendix D: Compilation of Nonpoint Source Analysis Technical Memos Appendix E: Point Source Survey Results Cited References Final Report September 2011 Page i

4 Tables Table 1. Nutrient Breakpoints by Ecoregion Under Consideration by IDEM (from Selvaratnam, S. and J. Frey 2011)... 6 Table 2. Summary of Nutrient Criteria Development Progress for U.S. EPA, Ohio, Minnesota, and Wisconsin... 7 Table 3. Eight-digit HUCs in the Wabash River Watershed across Indiana and Illinois Table 4. Number of facilities with NPDES permits in each 8 digit HUC of the Wabash River watershed Table 5. Tippecanoe County s 2007 Conservation Tillage Data Table 6. Estimated existing nutrient loads from permitted NPDES facilities in the Wabash River watershed Table 7. Changes in pollutant loads and resulting credit demand under different TN permit effluent scenarios Table 8. Changes in pollutant loads and resulting credit demand under different TP permit effluent scenarios Table 9. Summary of CWNS information for all facilities in the Wabash River watershed Table 10. Indicators associated with advanced treatment facilities in the Wabash River watershed Table 11. Summary of treatment type by HUC Table 12. Summary of 2008 CWNS permit information for facilities in the Tippecanoe and Driftwood watersheds Table 13. Permit limit summaries for facilities in the Driftwood and Tippecanoe watersheds Table 14. Summary of facility type and flows for WWTPs included in Driftwood and Tippecanoe nutrient removal analysis Table 15. Summary of ENR treatment levels and assumptions for WWTP upgrade simulations Table 16. Color coding of ENR treatment levels for Tables Table MGD activated sludge ENR upgrade options and costs Table MGD lagoon ENR upgrade options and costs Table MGD activated sludge ENR upgrade options and costs Table MGD lagoon ENR upgrade options and costs Table MGD activated sludge ENR upgrade options and costs Table MGD trickling filter ENR upgrade options and costs Table MGD activated sludge ENR upgrade options and costs Table 24. Total Nitrogen Exported at the Mouth of 8-Digit HUC Watersheds, Independent of Upstream Watershed Loading (USGS, 1997) Table 25. Total Phosphorus Exported at the Mouth of 8-Digit HUC Watersheds, Independent of Upstream Watershed Loading (USGS, 1997) Table 26. SPARROW model estimates of agricultural NPS loading Table 27. General Guidelines for Interpreting NO3-N Concentrations in Tile Drainage Water 1. (Purdue University, 2005) Table 28. Filterstrip Treatment Efficiency Results at the Subwatershed Scale, in Percent, Driftwood Watershed Table 29. Filterstrip Treatment Efficiency Results at the Subwatershed Scale, in Percent, Tippecanoe Watershed Table 30. Filterstrip Treatment Efficiency Results in Percent, Field Scale Results, Driftwood Watershed Final Report September 2011 Page ii

5 Table 31. Cover Crop Treatment Efficiency Results in Percent at the Subwatershed level, Driftwood Watershed Table 32. Cover Crop Treatment Efficiency Results in Percent at the Subwatershed level, Tippecanoe Watershed Table 33. Cover Crop Treatment Efficiency Results in Percent, Driftwood Subwatersheds Table 34. Cover Crop Treatment Efficiency Results in Percent, Tippecanoe Subwatersheds Table 35. No-till Residue Management Treatment Efficiency Percents at the Subwatershed level, Driftwood Watershed Table 36. No-till Residue Management Treatment Efficiency Percents at the Subwatershed level, Tippecanoe Watershed Table 37. No-till Residue Management Treatment Efficiency Results in Percent, Driftwood Watershed Table 38. Annual Nutrient Load Reduction Potential, Wabash River Watershed 8-digit HUC Subwatersheds. (Assuming a 10 and 25 Percent Participation of Agricultural Row Cropped Acres, 20 Percent Reductions, and 40 lbs TN/acre and 3 lbs TP/acre Loading Rates.) Table 39. Estimates of phosphorus bioavialability fractions for specific source categories Table 40. Current and Potential Future Point and Nonpoint Source Baselines in the Wabash River watershed Table Indiana EQIP General Eligible Practices Table 42. Summary of Credit Production per Acre of BMP Table 43. Summary of BMP Credit Production and Annualized Life Cycle Cost per Acre Table 44. Summary of Annualized Life Cycle Cost per Credit: Filter Strips (Prime Land) Table 45. Summary of Annualized Life Cycle Cost per Credit: Filter Strips (Marginal Land) Table 46. Summary of Annualized Life Cycle Cost per Credit: Cover Crops Table 47. Summary of Annualized Life Cycle Cost per Credit: Residue Management Table 48. Comparison of Potential Estimated Upgrade Costs and NPS BMP Costs for Small (<.3 MGD) Facilities by Treatment Level Table 49. Comparison of Potential Estimated Upgrade Costs and NPS BMP Costs for Medium (.3 MGD 5 MGD) Facilities by Treatment Level Table 50. Comparison of Potential Estimated Upgrade Costs and NPS BMP Costs for Large (>5 MGD) Facilities by Treatment Level Table 51. Resulting demand and supply factoring in estimated cost margins for each TP permit effluent scenario Table 52. Resulting demand and supply factoring in cost margins for each TN permit effluent scenario Figures Figure 1. Location of the Wabash River watershed in relation to the MARB and the hypoxic zone in the Gulf of Mexico Figure 2. Location and size of major reservoirs located in the Wabash River watershed Figure 3. Location of karst features in the Wabash River watershed Figure 4. NPDES permitted facilities by size and subwatershed in the Wabash River watershed Figure Landuse Map of the Wabash Watershed (MRLC, 2009) Figure 6. Number of Beef Cattle per Subwatershed within the Wabash-Patoka Watershed Figure 7. Number of Dairy Animals per Subwatershed within the Wabash-Patoka watershed Final Report September 2011 Page iii

6 Final Report September 2011 Page iv

7 1. Introduction In 2008, the U.S. Environmental Protection Agency (EPA) awarded a Targeted Watershed Grant to the Conservation Technology Information Center (CTIC) to conduct a water quality trading (WQT) market feasibility analysis for the Wabash River watershed. The 2008 Targeted Watershed Program funded ten projects focusing on water quality trading or other market-based water quality projects to reduce nitrogen, phosphorus, sediment, or other pollutant loadings that cause hypoxia in the Gulf of Mexico. The projects are located in the three Mississippi River sub-basins with the highest nutrient loads contributing to hypoxia in the Northern Gulf of Mexico: the Ohio River, the Upper Mississippi River, and the Lower Mississippi River. The Wabash River watershed is a major tributary to the Ohio River. CTIC partnered with Kieser & Associates, LLC and Tetra Tech, Inc. (Tt) to conduct a market feasibility analysis to determine if the necessary conditions exist in the Wabash River watershed to support the development and implementation of a viable, sustainable water quality trading program involving agricultural nonpoint sources and permitted point sources. This report summarizes the approach and the findings of the Wabash River Watershed WQT market feasibility analysis. 1.1 What is a WQT Market Feasibility Analysis? Although there might be an interest to conduct WQT in a particular watershed, certain factors need to be present to make it a viable, sustainable program. A WQT feasibility analysis is a process for collecting and analyzing the data and information needed to determine if the technical and economic factors exist to support trading between potential sources. The very basic factors needed to support WQT are as follows: Well-defined sources and amounts of pollution. WQT requires an understanding of pollutant sources. In the case of the Wabash River watershed, nitrogen and phosphorus are the pollutants of concern. Sources generating nutrient loads are potential buyers and sellers in a water quality trading approach. Collecting information to characterize the type of sources and the associated nutrient loads from each source help to determine if there will be an adequate supply and demand for tradable credits. Regulatory drivers and incentives. Without regulatory drivers or some type of incentive, sources wouldn t feel compelled to consider and, ultimately, participate in WQT. The most compelling drivers for WQT are those related to regulatory requirements. In most cases, this is a more stringent permit effluent limit based on a more stringent water quality standard. In other cases, it could be a watershed pollutant reduction goal that might not have a regulatory component, but provides other type of incentives to meet this goal (e.g., avoidance of a total maximum daily load). Difference in control costs among sources. Sources with high pollutant control costs will have an economic motivation to seek out tradable credits from other sources that are able to control pollutants to meet requirements at a lower cost. It is this difference in control costs among sources that will determine which sources might participate as buyers and which sources might have the ability to participate as sellers. WQT feasibility is largely driven by economics, both actual and perceived costs (e.g., transaction costs and risk factors). Final Report September 2011 Page 1

8 A WQT market feasibility analysis has two components: 1) a pollutant suitability analysis and 2) an economic suitability analysis. The pollutant suitability analysis includes information on pollutant type and form, geographic scope, potential buyers and sellers, potential water quality trading credit supply and demand, potential trade ratios to account for pollutant fate and transport as well as uncertainty, issues related to avoiding localized areas of excessive pollutant loading (i.e., hotspots), and duration of water quality trading credits. The economic suitability analysis includes information on potential buyers willingness-to-pay for water quality credits, potential sellers price for generating water quality credits, effect of trade ratios on the cost of water quality trading credits, and the potential costs of involving stakeholders in designing and implementing a water quality trading program. Information from each of these components provides insight as to where WQT might encounter barriers in a particular watershed and what type of trading framework might be most appropriate based on the sources with the greatest potential for participation. A WQT market feasibility analysis is not intended to provide definitive answers about how WQT should work in a particular watershed, only if the conditions are ripe to support such an effort. WQT program design and implementation requires coordination and facilitation with watershed stakeholders to ensure the program integrates well with other efforts. What the product of a WQT market feasibility analysis can do, however, is give watershed stakeholders a starting place and a foundation when moving into the design phase. The analysis can also identify where watershed stakeholders will potentially have to do additional research to obtain detailed, watershed-specific information that could affect WQT success. This might mean holding focus groups with point sources and nonpoint sources to better understand attitudes, perceptions, and concerns. It might also mean public meetings with watershed residents and organizations that have perceptions and opinions about how to meet water quality goals. 1.2 What is the Purpose of the Wabash River Watershed WQT Market Feasibility Analysis? The purpose of the WQT market feasibility analysis for the Wabash River watershed is to conduct a preliminary assessment for the potential of viable, sustainable trading to meet water quality goals. This analysis is an initial assessment focused on using mostly existing data examined through the lens of pollutant suitability and economic suitability. The purpose of this analysis was not to collect new data, but to identify where additional data and information might be needed to support further WQT feasibility assessment activities and future WQT program development. The goal is to create a foundation for future work that, over time, watershed stakeholders contribute to more and more. Ultimately, this WQT market feasibility analysis is intended to characterize the watershed for purposes of trading, identify existing data gaps, and make recommendations about WQT feasibility where the data can support these types of recommendations. Where data are not available, the Project Team has identified next steps and additional data needs to move water quality trading in the Wabash River watershed forward. Through the Wabash River watershed WQT market feasibility analysis, the Project Team led by CTIC reviewed existing watershed data and information available through ongoing Wabash River watershed and Ohio River basin projects, such as the Wabash River Total Maximum Daily Load (TMDL) and TMDLs for subwatersheds in the Wabash River, such as Limberlost Creek, the Little Wabash River, and the Final Report September 2011 Page 2

9 South Fork Wildcat Creek. Information and data on point sources in the watershed were also obtained from the Indiana Department of Environmental Management (IDEM) and the Illinois Environmental Protection Agency (IEPA), and information from nonpoint source loading estimates were obtained from the U.S. Geological Survey (USGS). 1.3 What Does This Report Contain? The Wabash River watershed WQT market feasibility analysis report contains the following: Section 1: Understanding the Wabash River Watershed. An overview of the Wabash River watershed as it relates to the Gulf of Mexico hypoxia issue. Section 2: Feasibility Analysis Summary. This section provides a discussion of the information used in the pollutant suitability and economic suitability analysis the two components of the overall WQT market feasibility analysis. Section 3: Putting It All Together: Market Analysis and Trading Considerations. This section synthesizes the information provided in Section Two to provide an analysis of the overall market potential for water quality trading in the Wabash River watershed. This section also addresses other trading considerations that will affect the market. Section 4: Next Steps for Water Quality Trading in the Wabash River Watershed. This section identifies data needs and additional analyses to move the concept of water quality trading forward in the Wabash River watershed. Appendices. The appendices to the report include the Wabash River TMDL and detailed technical memos generated by the Project Team to inform different components of the WQT market feasibility analysis process. 2. Understanding the Wabash River Watershed Characterizing the physical, chemical, and biological attributes of a watershed is an important first step in assessing the feasibility of WQT. This section provides a brief overview of the Wabash River watershed as it relates to the Gulf of Mexico hypoxia issue. 2.1 Relationship of the Wabash River Watershed to the Gulf of Mexico The area that drains to the Gulf of Mexico is referred to as the Mississippi-Atchafalaya River Basin (MARB). This basin drains 1,245,000 square miles across 31 states and is the focus of the efforts to manage nutrients causing hypoxia in the Gulf of Mexico. Within the MARB, the Wabash River watershed covers approximately 33,000 square miles, draining portions of Indiana and Illinois. Figure 1 shows the location of the Wabash River watershed in relation to the MARB that drains to the Gulf of Mexico. Final Report September 2011 Page 3

10 Figure 1. Location of the Wabash River watershed in relation to the MARB and the hypoxic zone in the Gulf of Mexico. While the Wabash River watershed is only 2 percent of the total area of the MARB, it delivers a relatively significant nutrient load to the Gulf of Mexico via the Ohio River one of the three sub-basins that deliver the highest nutrient loading to the Gulf. According to USGS SPARROW modeling, the Wabash River watershed contributes approximately 9,994 tons of total phosphorus and 139,278 tons of total nitrogen to the Gulf of Mexico each year. Reducing the nutrient load from the Wabash River watershed will contribute to both local water quality improvements and help reduce the load contributing to the Gulf of Mexico s hypoxic zone. 2.2 Overview of Nutrient Sources and Loadings in the Wabash River Watershed In 2006, the IEPA and IDEM developed the 2006 Wabash River Nutrient and Pathogen TMDL report to address water quality impairments in the Wabash River watershed. Information from Indiana and Illinois 2002, 2004, and 2006 Clean Water Act (CWA) Section 303(d) listings demonstrate that several segments in the Wabash River watershed are impaired for nutrients, among other pollutants. Review of the data and comparison to state nutrient targets led to a determination that nutrient TMDLs should be developed for all segments of the Wabash River from the Indiana/Ohio state line to the confluence of the Wabash and Vermilion Rivers. The TMDL addresses nutrient sources that discharge directly to the mainstem of the Wabash River, including point sources permitted under the National Pollutant Discharge Elimination System (NPDES) Final Report September 2011 Page 4

11 program, subwatersheds, and significant tributaries. Permitted NPDES point sources identified in the TMDL include industrial facilities, power plants, wastewater treatment facilities, municipal separate storm sewer systems (MS4s), combined sewer overflows, and confined animal feeding operations (CAFOs). Although a detailed source analysis was not conducted for the subwatersheds, the TMDL does provide an analysis of land use/land cover in the subwatersheds and tributaries that drain directly to the Wabash River. Agriculture is the most significant land use in the area that drains directly to the Wabash River (82 percent), as well as the tributaries (53 97 percent). The TMDL concluded that nonpoint sources, including agriculture, contribute the largest loads of TP and nitrate to the Wabash, but that point sources can have an important impact during low flow periods. The WWTP loads therefore need to be reduced to meet the in-stream 0.30 mg/l benchmark. In summary, the Wabash River watershed contains several segments that are impaired for nutrients due to a variety of point and nonpoint sources. This basic understanding of water quality issues in the Wabash River watershed provides a basic foundation for a more in-depth water quality trading market feasibility analysis, presented in Section Two. 3. Feasibility Analysis Summary This section summarizes the information compiled through the pollutant suitability analysis and the economic suitability analysis. This information provides the basis for the water quality trading market analysis and considerations presented in Section Three. 3.1 Drivers and Incentives for Trading Sources considering WQT generally do so because of more stringent permit effluent limits resulting from changes to water quality standards or implementation of wasteload allocations (WLAs) in approved TMDLs. These are considered regulatory drivers for WQT. In addition to regulatory drivers, sources might consider WQT as a result of other incentives, such as the desire to avoid more stringent permit limits before the development of a TMDL or to meet a pollutant reduction goal established through a watershed management plan. A brief discussion of regulatory drivers and other incentives in the Wabash River watershed is provided below Water Quality Standards State s water quality standards drive NPDES permit effluent limits and TMDLs; therefore, water quality standards are the ultimate regulatory driver for WQT. If water quality standards result in NPDES permit effluent limits that pose a technical or financial burden for sources, the standards can serve as a driver for WQT. To date, Indiana has not adopted numeric water quality criteria for nutrients to protect aquatic life uses. As stated in the 2006 TMDL, Indiana uses draft nutrient benchmarks: Total phosphorus should not exceed 0.3 mg/l. Nitrate + nitrite should not exceed 10 mg/l. Dissolved oxygen should not be below the water quality standard of 4.0 mg/l and should not consistently be close to the standard (i.e., in the range of 4.0 to 5.0 mg/l). Values should also Final Report September 2011 Page 5

12 not be consistently higher than 12 mg/l and average daily values should be at least 5.0 mg/l per calendar day. No ph values should be less than 6.0 or greater than 9.0. ph should also not be consistently close to the standard (i.e., 8.7 or higher). Algae growth should not be excessive based on field observations by trained staff. IDEM considers a segment to be impaired for nutrients when two or more of these benchmarks are exceeded based on a review of all recent data. It is anticipated that IDEM will adopt numeric nutrient criteria in the near future. IDEM is working with USGS to develop numeric nutrient criteria that take into account the relationship between stressors and the biological community. Table 1 presents possible numeric nutrient criteria presented by IDEM at the U.S. EPA Nutrient TMDL Workshop in February The numeric nutrient criteria development effort is still ongoing and these are preliminary values used during the development of the feasibility study. Nutrient Breakpoints Table 1. Nutrient Breakpoints by Ecoregion Under Consideration by IDEM (from Selvaratnam, S. and J. Frey 2011) Glacial North Ecoregion Central/West Plains Low (oligotrophic) TN < 0.60 mg/l None TP < mg/l None High (eutrophic) TN > 1.2 mg/l TN > 1.7 mg/l TP > 0.14 mg/l TP > 0.13 mg/l Once Indiana adopts numeric nutrient criteria, IDEM will issue NPDES permits with more stringent water quality-based effluent limits to meet the new criteria. In Illinois, the state has adopted the following criteria: 0.05 mg/l TP for lakes and 10 mg/l TN for rivers. In addition, U.S. EPA has strongly encouraged IEPA to develop additional nutrient criteria. Progress in developing nutrient criteria in other U.S. EPA Region 5 states provides context for the direction of future nutrient criteria. Table 2 presents a summary of nutrient criteria development progress for EPA (EPA, 2000a, EPA 2000b and EPA, 2000c), Ohio (Ohio EPA, 2011), Minnesota (MPCA, 2010a, 2010b) and Wisconsin (WI DNR, 2010). Final Report September 2011 Page 6

13 Table 2. Summary of Nutrient Criteria Development Progress for U.S. EPA, Ohio, Minnesota, and Wisconsin Authority Status Phosphorus Criteria Nitrogen Criteria EPA Aggregated Level IV Nutrient Ecoregions Guidance VI ug/l VII 33 ug/l IX ug/l VI Total Nitrogen 2.18 ug/l Ohio EPA Developing TMDL derived Site Specific Water Quality Standards Minnesota Pollution Control Agency Wisconsin Department of Natural Resources Developing Promulgated NR 102 (Phosphorus Rule) South River Nutrient Region 150 ug/l Large Rivers 100 ug/l Central River Nutrient Region 100 ug/l Small Rivers 75 ug/l VII Total Nitrogen 0.54 ug/l IX Total Nitrogen 69 ug/l 4-day chronic toxicity standard for Nitrite + Nitrate of 4.9 mg/l NPDES Permit Effluent Limits As stated under the water quality standards discussion, numeric nutrient criteria directly affects NPDES permit limits for nitrogen and phosphorus. To date, IDEM issues NPDES permits that do not contain water quality-based effluent limits for nutrients. This will change when IDEM adopts numeric nutrient criteria, as discussed above. For now, IDEM requires NPDES permitted facilities to conduct discharge monitoring as a way to generate better data on total phosphorus and total nitrogen concentrations. This information will help IDEM in developing future water quality-based effluent limits for nutrients to meet impending numeric nutrient criteria. A recent review of Illinois NPDES permits conducted by U.S. EPA Region 5 showed that reviewed NPDES permits did not contain nutrient effluent limitations. As a result of the findings of this review, U.S. EPA Region 5 issued a letter in January 2011 that directs Illinois EPA to establish nutrient effluent limitations when it makes the determination that a nutrient discharge will cause an excursion beyond Illinois existing narrative nutrient criteria. Appendix A contains a copy of the January 2011 letter from U.S. EPA Region 5 to Illinois EPA about the matter of developing nutrient effluent limitations for nutrient discharges from permitted point sources Wabash River TMDLs The 2006 Wabash River TMDLs contain WLAs for point sources and load allocations (LAs) for nonpoint sources to address nutrient impairments. The TMDL assigns total phosphorus WLAs to point sources. The TP WLAs have not yet been incorporated into NPDES permits because facilities are first required to conduct monitoring to determine their actual discharge concentrations. For tributaries and subwatersheds draining directly to the Wabash River, the TMDL assigns a 4 percent reduction in phosphorus loads and no reductions in nitrate. The 4 percent phosphorus load reduction under the TMDL might serve as a nonpoint source baseline the amount a nonpoint source seller has to reduce by before becoming eligible to sell credits to buyers in need of credits. A list of other TMDLs and associated WLAs for Wabash River subwatersheds is found in Appendix B. Final Report September 2011 Page 7

14 3.1.4 Watershed Management Planning IDEM s Watershed Management Planning Checklist provides watershed organizations with a framework to develop a Section 319 approvable watershed management plan. Several watershed management plans for subwatersheds of the Wabash River watershed are available through IDEM s watershed management planning website. Many of these plans contain nutrient reduction goals. For example, the watershed management plan for the Upper Tippecanoe River contains a 20 percent nutrient load reduction goal to be achieved by 2010 and the watershed management plan for the Lower Eel River contains a 10 percent nutrient load reduction goal. Not every subwatershed in the Wabash River watershed has an IDEM-approved watershed management plan, but for those that do, the nutrient load reduction goals established for nonpoint sources can play a role in promoting participation in WQT. In addition, the nutrient load reduction goals quantified in watershed management plans might also serve as a nonpoint source baseline for nonpoint sources in specific watersheds that want to participate in trading as credit sellers Gulf Hypoxia Action Plan In 2008, the Mississippi River/Gulf of Mexico Watershed Nutrient Task Force released the 2008 Action Plan a national strategy to control hypoxia in the Gulf of Mexico and improve water quality in the MARB. The 2008 Action Plan calls for a dual 45 percent reduction in the riverine total nitrogen load and in the riverine total phosphorus load. This dual nutrient reduction loading goal is not an enforceable goal, but it does provide an overarching target for sources in the Mississippi River Basin which includes the Wabash River watershed to strive for to help reduce the hypoxic zone in the Gulf of Mexico Summary: Drivers and Incentives for Trading In summary, a few point source regulatory drivers for WQT are emerging based on TMDLs. Many waters do not currently have nutrient drivers within the Wabash River watershed and likely will not until IDEM adopts numeric nutrient criteria and IEPA adopts more stringent and encompassing criteria. It is likely that IDEM will adopt numeric nutrient criteria in the near future, which will trigger the need to update existing NPDES permits with water quality-based effluent limits for nutrients. When permitted facilities are required to comply with new, more stringent nutrient permit limits, there will be a more tangible regulatory driver for WQT in the Wabash River watershed. The WQT market feasibility analysis for the Wabash River watershed makes the assumption that new numeric nutrient criteria will become a reality in Indiana and act as a driver for sources to consider WQT as a potential implementation tool to achieve more stringent permit limits. If numeric nutrient criteria are not adopted in Indiana and NPDES permits are not re-issued with more stringent nutrient effluent limits, point sources in the Wabash River watershed would not have a sufficient regulatory driver for WQT. 3.2 Suitable Pollutants for Trading Nitrogen and phosphorus are considered appropriate pollutants for trading under U.S. EPA s 2003 Water Quality Trading Policy and the U.S. EPA Water Quality Trading Toolkit for Permit Writers. Nutrients are relatively persistent in river environments and the focus of the Gulf of Mexico hypoxia issue. Local eutrophication issues such as dissolved oxygen impairments and nuisance algal blooms require TMDLs and are fueling the consideration of statewide nutrient criteria. Therefore, phosphorus and nitrogen forms are the focus of the Wabash River watershed WQT market feasibility analysis. At a very basic level, this means a focus on total nitrogen and total phosphorus. Difficulties in determining loading of Final Report September 2011 Page 8

15 reactive nutrient forms exist because of the sizeable variability in concentrations of soluble reactive forms across very short time periods. This has resulted in water quality monitoring programs relying heavily on TN and TP when estimating loads. However, bioavailability of the nutrients discharged by each source can be an important aspect of offsetting. For example, if a reticent nutrient form is traded for a source s load which is substantially bioavailable the water quality impacts may not be addressed. In addition, pollutant parameter suitability considerations for WQT include determining at what concentration a nutrient form has acutely toxic properties or quickly manifests other stresses. Lastly, the consideration of persistence is important. A parameter that is quickly attenuated is not a viable offset for impacts further downstream. Discharges of sizeable concentrations of ammonia can create acute toxicity concentrations. Ammonia can also consume high levels of oxygen as the form is converted into NO 2 and the NO 3. Because of these interactions WQT to address ammonia effluent limits is not appropriate. However, nitrogen in the form of ammonia is a tradable nutrient form when it is present in concentration levels characteristic of healthy ecosystems and not causing the described impacts. In these settings ammonia is cycling normally through the DIN or TN succession. Persistence, or the fate and transport of nutrients is an important consideration when setting eligible boundaries for trading transactions. For both of the nutrients adequate provisions can be included in the program boundaries and location factor to address quicker attenuation rates in headwater streams and downstream persistence. A brief discussion on the considerations related to the different forms of nitrogen and phosphorus in the context of WQT in the Wabash River watershed is provided below Nitrogen Considerations The 2006 Wabash River TMDL established loading limits for TN, which eliminates the need to consider other forms of nitrogen (e.g., nitrate, organic nitrogen) in the Wabash River watershed. While TN will allow all sources to trade with each other because it represents a stable pollutant that provides an equivalent trading relationship, there are some considerations to keep in mind about the impact from different forms of nitrogen. In general, there is greater environmental benefit to removing the more bioavailable forms of nitrogen, which include dissolved organic nitrogen and dissolved inorganic nitrogen (nitrate, nitrite, and ammonia). For example, the Gulf of Mexico 2008 Hypoxia Action Plan states that nitrogen composition should be emphasized in nutrient reduction strategies. Nitrate is the most important form fueling the primary production that leads to hypoxia development in the spring (April, May, and June). Between 2001 and 2005, total annual nitrogen loads to the Gulf of Mexico declined, but in the critical spring months, the reduction in total nitrogen load is from nitrogen forms other than nitrate. Research conducted by the USGS (1997) suggests that the relationship between nitrate concentration and flow might be due to nitrate leached from soil and the unsaturated zone during high flow conditions. In addition, agricultural tile drainage might also contribute to increased nitrate levels during high flows. As a result, WQT in the Wabash River watershed should place a priority on reducing nonpoint sources of DIN and to a lesser extent on bioavailable forms of DON Phosphorus Considerations Similar to TN, the 2006 Wabash River TMDL uses TP, a form of phosphorus that eliminates the need to consider other forms of phosphorus (e.g., soluble phosphorus) in the Wabash River watershed. Using TP allows for an equivalent trading relationship among sources with a phosphorus contribution. However, there are differences in the type of phosphorus associated with point and nonpoint sources. As a result, the potential effect of phosphorus from these sources on the Wabash River watershed will vary. Agricultural nonpoint sources discharge primarily the non-soluble, sediment-attached form of Final Report September 2011 Page 9

16 phosphorus. Point sources, wastewater treatment facilities in particular, discharge primarily soluble forms of phosphorus. The bioavailability of the TP from each discharger will be considered in Section 4.3 of this report for both the Wabash River and the Gulf of Mexico. 3.3 Watershed Considerations Nutrient water quality standards are emerging in states across the nation (e.g., Wisconsin and Florida) that have phosphorus criteria levels for selected regions at or below 0.1 mg/l TP. Nitrogen water quality standards, to a much lesser extent in the Midwest, are being considered that would substantially reduce stream concentrations to single digits as compared to the 10 mg/l nitrate drinking water standard. These new criteria could create a setting where it is common for numerous impaired waters to be listed and waters achieving water quality attainment may have limited available capacity for future waste loads. Federal regulations (i.e., 40 CFR (d) and 40 CFR 122.4(i)) pertaining to NPDES permit requirements expressly prevent a permit effluent limit from causing or contributing to water quality violations. These requirements apply to WQT as well. A trade cannot create a local hot spot (area of impairment) in one water body because it is protecting another. However, WQT can allow upstream buyers of credits to purchase downstream generated credits if the stream reaches between the two participants is in compliance with water quality standards. In large streams the buyer s discharge may not significantly alter the stream s nutrient concentration. In these setting where the stream is in compliance a new discharger could purchase downstream generated credits to comply with waste load allocation requirements of the system. If the stream s concentrations are above water quality standards or a permittee s discharge is substantial to the point where a local violation would be caused, then the appropriate WQT framework would require generated credits upstream of the buyer s discharge. This assumes the presence of ample credit supply exists upstream of the buyer and that the credit supply will remain persistent within the stream to the point of the buyer s discharge. Because of these considerations, an understanding of the fate and transport of nutrients in rivers and streams is emerging as a critical issue for water quality standards, estuary protection and effluent limit setting programs. WQT is no different. Fortunately, WQT goals are set by the effluent limit requirements of the permit application and WQT eligibility requirements can be created to address upstream generated nutrient credit attenuation concerns. Tools like the USGS SPARROW model for the entire Mississippi River watershed are being developed at finer resolutions to allow for an understanding of scale requirements that can be used to set boundary conditions for WQT (e.g. a 10-digit HUC evaluation provided replacing the current 8-digit HUC results) Summary: Suitable Pollutants for Trading In summary, both total nitrogen and total phosphorus are suitable pollutants for trading in the Wabash River watershed. However, WQT in the Wabash River watershed should focus on strategies to target sources of nitrate and account for the differences in the soluble and non-soluble forms of phosphorus associated with point sources and nonpoint sources. The fate and transport of nutrients within the Wabash River and downstream to the Gulf of Mexico also need to be taken into account. 3.4 Timing WQT frameworks are required to be contemporaneous with NPDES permit effluent limit requirements. NPS generated credits are both episodic in nature and can have a seasonal variation with more credits being generated in one temperature regime or vegetative growth cycle than another (e.g., spring versus Final Report September 2011 Page 10

17 winter or pre-crop canopy versus full canopy). NPDES permit effluent limits are assigned to respond to critical periods. The WQT framework must also respond with contemporaneous credit generation for these effluent limits. The critical period noted in the 2006 Wabash River TMDLs for nutrients include both high flow periods (such as spring runoff) when nutrient loads are high, as well as low flow summer periods when the assimilative capacity of the river is reduced. The critical period in the Gulf of Mexico for the hypoxia issue is late spring/early summer (e.g., April, May, June). Permit effluent limit setting processes for farfield drivers will have to consider other watershed characteristics in addition to critical time periods. For instance, internal loading or recycling of nutrients and transport time for nutrients become relevant considerations when setting protection limits for water bodies. Where possible, seasonal or annual critical periods for NPDES permit effluent limits should be accompanied by adequate supporting justification. For example, in the Chesapeake Bay trading framework U.S. EPA accepted an annual credit generation window and in Wisconsin the DNR Water Quality Trading Framework allows banking of NPS credits to be used within the year generated. Where the critical period is more constrained (e.g., one month) WQT must be structured to contemporaneously generate credits or not be an eligible option Summary: Timing Spring is a critical time period for both the Wabash River watershed and within the Gulf of Mexico. A Wabash River trading framework will therefore need to focus on reducing the load of nutrients during this period. 3.5 Geographic Scope for Trading Analysis Understanding how the geographic scope of the watershed could affect the viability and sustainability of trading is a key aspect of a WQT market feasibility analysis. Geographic scope of the watershed influences important factors such as pollutant fate and transport, which in turn affects credit supply and demand through the application of trade ratios. For purposes of the Wabash River watershed WQT market feasibility analysis, the Project Team focused on the Indiana and Illinois portions of the Wabash River watershed. Beginning in the State of Ohio, the Wabash River watershed covers 32,950 square miles extending across most of the State of Indiana (24,320 square miles) and with significant parts of the State of Illinois. Through the pollutant suitability analysis portion of the WQT market feasibility analysis, the Project Team considered conditions throughout the entire Wabash River watershed that could affect trading and then focused in on subwatersheds for a more detailed analysis. In considering the entire watershed, the Project Team took into consideration factors such as the location of major lakes and reservoirs that could act as a pollutant sink and affect fate and transport. Figure 2 shows the location and size of reservoirs located throughout the Wabash River watershed. If water quality trading were to take place in the Wabash River watershed, the design of the trading program would have to take these reservoirs into account through delivery and location trade ratios (see Section 4.3). Other physical features to consider are the karst regions of the watershed, which total approximately 263,500 acres as shown in Figure 3. Similar to reservoirs, karst features can affect the fate and transport of pollutants and they can also affect runoff rates. Because karst features are limited to the southeast portion of the Wabash River watershed, restrictions on trading due to karst would likely only affect Final Report September 2011 Page 11

18 credit suppliers located in this area. A water quality trading program design should include guidance on how to identify karst features and tailor eligibility requirements accordingly. Other geographic considerations include land use practices that change hydrology, chemistry, and biology of a watershed. In the Wabash River watershed, coal mining is of particular interest. Surface coal mines are found in the southeastern portion of the Wabash River watershed. Subwatersheds with coal mining activities might have lower ph levels due to mine runoff that could affect the bioavailability of nutrients. This could affect potential trading activities in close proximity to these areas because understanding fate and transport of pollutants would be challenging and difficult to account for in program design (e.g., credit estimation and trade ratios). Final Report September 2011 Page 12

19 Figure 2. Location and size of major reservoirs located in the Wabash River watershed Figure 3. Location of karst features in the Wabash River watershed In addition to looking at the characteristics across the entire Wabash River watershed that could affect water quality trading, the Project Team looked at subwatersheds to help in the more detailed analysis necessary to understand the factors affecting potential credit supply and demand. The 8-digit Hydrologic Unit Codes (HUCs) found in the Wabash River watershed are listed in Table 3. Final Report September 2011 Page 13

20 Table 3. Eight-digit HUCs in the Wabash River Watershed across Indiana and Illinois. BASIN SUBBASIN HUC_8 SQ. MILES Wabash Salamonie Wabash Mississinewa Wabash Eel Wabash Middle Wabash-Deer Wabash Tippecanoe ,960 Wabash Wildcat Wabash Middle Wabash-Little Vermilion ,287 Wabash Vermilion ,439 Wabash Sugar Wabash Middle Wabash-Busseron ,019 Wabash Embarras ,442 Wabash Lower Wabash ,321 Wabash Little Wabash ,148 Wabash Skillet ,049 Patoka-White Upper White ,754 Patoka-White Lower White ,675 Patoka-White Eel ,195 Patoka-White Driftwood ,154 Patoka-White Flatrock-Haw Patoka-White Upper East Fork White Patoka-White Muscatatuck ,143 Patoka-White Lower East Fork White ,025 Patoka-White Patoka Total 32,950 No pre-existing watershed model exists for all of the subwatersheds within the Wabash River watershed. Given the size of the Wabash River watershed, conducting detailed modeling for all subwatersheds was not within the scope of this WQT market feasibility analysis. Instead, the Project Team conducted an analysis using the Soil and Water Assessment Tool (SWAT) of two subwatersheds to facilitate a smaller-scale quantitative assessment. The Project Team evaluated factors related to nonpoint source credit suppliers (e.g., agricultural producers) to identify two subwatersheds that would provide characteristics that could be extrapolated to the rest of the Wabash River watershed. The Project Team examined a variety of information such as land use, location of flow gages, cropland data, animal counts, location within the watershed (i.e., headwater or not), and location of karst features. Based on this analysis, the Project Team selected the Tippecanoe and Driftwood subwatersheds for more targeted assessment in the feasibility analysis (the locations of the Tippecanoe and Driftwood subwatersheds are shown in Figures 2 and 3) Summary: Geographic Scope for Trading Analysis In summary, the entire Wabash River watershed appears appropriate for WQT as long as trading program design takes into account notable features. These features include karst areas, major lakes and reservoirs, and other land uses that change the natural hydrology. These features could affect nutrient fate and transport, but could be addressed in WQT program design. Final Report September 2011 Page 14

21 3.6 Potential Credit Buyers and Sellers In WQT, potential credit buyers are sources that need to reduce a pollutant load to comply with a regulatory requirement and credit sellers are sources that have the means to supply a unit of pollutant load reduction by over-controlling a pollutant load. Depending on the type of water quality trading program, a potential credit buyer is typically a facility covered by a NPDES permit. Potential credit sellers could be either other NPDES permitted point sources that over-control a pollutant discharge or eligible nonpoint sources that generate pollutant load reductions through best management practice implementation. For the purposes of the Wabash River watershed WQT market feasibility analysis, the primary focus is on NPDES permitted facilities as potential credit buyers and agricultural nonpoint sources as potential credit sellers NPDES Permitted Facilities as Potential Credit Buyers According to information provided by IEPA and IDEM, there are 943 facilities with NPDES permits within the Wabash River watershed. Table 4 provides a list of the number of NPDES permitted facilities by 8-digit HUC. Table 4. Number of facilities with NPDES permits in each 8 digit HUC of the Wabash River watershed 8 Digit HUC # of All Permits Large Medium Small Total Final Report September 2011 Page 15

22 According to Table 4, the most NPDES permits are located in HUC , which has a total of 157. This HUC has the most facilities across all size categories. Small facilities are the predominant size category in the Wabash River watershed, with more than 50 percent of facilities falling in this category. Figure 4. NPDES permitted facilities by size and subwatershed in the Wabash River watershed. Final Report September 2011 Page 16

23 Section 3.6 provides an analysis of the potential credit demand that these facilities might generate through more stringent permit effluent limits. Once a facility has met the nutrient reduction requirements it is possible for this facility to provide further reduction and generate credits for sale to other facilities Agricultural Nonpoint Sources as Potential Credit Sellers WQT frameworks are created to fit the local setting. This includes working with the watershed characteristics to achieve and protect beneficial uses and water quality standards as well as advance watershed management goals. The nutrients in NPS runoff can be adequately quantified and adjusted for equivalence and location at the field scale and be transferable as credits to offset permitted wastewater discharges. However, WQT must fit within the socio-political structure. WQT has the opportunity to accelerate implementation for TMDLs, provide a mechanism for future growth in capped watersheds, and provide interim compliance leverage in cases of variance request or compliance schedule relief. To realize some of these benefits, the appropriate policies and perspectives must be in place. See the discussion on Baselines in Section 4.4. Land use in the Wabash River watershed includes subregions dominated by forest, urban and agricultural coverages (Figure 5). Forested land use yields relatively light nutrient loading when compared to urban and agricultural land management (Reckhow and Chapra, 1983 and Dodd et al., 1992). The riverine geomorphology includes stable channels, incising channels and enhanced drainage features like channelized streams and ditches. These factors affect nutrient transport pathways and attenuation rates. Final Report September 2011 Page 17

24 Figure Landuse Map of the Wabash Watershed (MRLC, 2009) Final Report September 2011 Page 18

25 The crop coverage map illustrates how much Indiana and Illinois benefit from abundant agricultural resources. Currently these resources are being managed in economically productive livestock and grain commodity production operations. The prevalence of corn-soybean rotations is typical of Midwest states. In addition, livestock density across the watershed is relatively high, but variable in terms of animal type and concentration. A consequence of this production and the physical features of the watershed is that the Wabash River watershed is the highest nutrient loading watershed in the Ohio River Basin (USGS, 2008). This high level of loading sets the stage for reduction opportunities. Current conservation practice adoption varies across the State of Indiana. For example the 2007 CTIC crop residue transect survey results for counties within the Tippecanoe watershed are provided in Table 5. The county variation ranges from 46 percent in some type of corn conservation tillage practices up to 91 percent. Table 5. Tippecanoe County s 2007 Conservation Tillage Data 1 County (rank) No-Till Mulch Till Reduced Till Conventional Till Corn Data Pulaski (19) Kosciusko (24) Fulton (33) White (67) Average corn Soybean Data White (26) Kosciusko (29) Pulaski (35) Fulton (38) Average soybean Indiana State Department of Agriculture Conservation Tillage Data available at: Based on the variable loading rates and variable adoption of conservation practices, the Wabash River watershed consists of both regions containing more than ample credit supply opportunities and those with reduced opportunity. However, within every subwatershed of the Wabash River there exists a number of individual sites that contain the key characteristics desired to supply credits: 1) implementation of a BMP will substantially reduce current NPS loading 2) the site is located in close proximity to the Wabash River or one of its tributaries 3) a willing land owner. These prerequisites are necessary to supply an adequate volume of credits at an economical price. In low volume regions WQT may be limited to individual permits needing assistance with difficult or costly compliance attainment issues. In the regions with ample ability to supply credits a larger program could be available where many buyers and sellers participate. Section 3.7 describes the methods to preliminarily assess the entire Wabash. A higher resolution of assessment at the local level is advised as part of the WQT framework development process, should WQT programs in the Watershed be pursued. To evaluate and quantify the regional potential at a higher resolution stakeholder input is required. One objective would be to gather farm data regarding operational practices of nutrient and conservation management. This information can be both distinctly individualized and considered confidential by the producer and Farm Bill public programs. Producers may choose to divulge historic practices once Final Report September 2011 Page 19

26 funding opportunities are present, but often remain silent when requested for data until they are comfortable with the program and the individuals running it. Section 3.7 provides an analysis of the potential credit supply from these agricultural nonpoint sources located in the Wabash River watershed. 3.7 Potential Credit Demand Determining the potential credit demand from NPDES permitted facilities in the Wabash River watershed requires an understanding of the estimated pollutant load reductions necessary to meet more stringent permit limits, as well as the type of existing treatment and control technology upgrade options that are available to facilities to meet more stringent permit limits. This section examines both factors that play a role in generating a demand for credits Estimates of Existing Pollutant Loads and Pollutant Load Reductions Under Three Permit Limit Scenarios The Project Team developed a technical memorandum entitled Characterization of Wabash River Nutrient Loads as part of the WQT market feasibility analysis to help estimate potential credit demand. This memorandum is found in Appendix C. This technical memo estimates the existing TN and TP loads under current permit limits and then examines the necessary pollutant load reductions to achieve more stringent permit limits in the future. Table 6 summarizes estimated existing nutrient loads from the 943 NPDES permitted facilities in the Wabash River watershed. The estimated existing loads are categorized by size of facility in the categories of small, medium, large. Small facilities are those permitted to discharge no more than 0.3 million gallons per day (MGD). Medium-sized facilities are those with permitted discharges that range between 0.3 to 5 MGD. Large facilities discharge more than 5 MGD. Table 6. Estimated existing nutrient loads from permitted NPDES facilities in the Wabash River watershed. HUC_ Facility Size (# of facilities) Estimated Existing Loads 1 Total Phosphorus Total Nitrogen Daily Load (lbs/day) Annual Load (Tons) Daily Load (lbs/day) Annual Load (Tons) Large (8) , Medium (15) Small (39) Total (62) , Large (1) Medium (7) Small (11) Total (19) Large (2) , Medium (21) Small (16) Total (39) , Final Report September 2011 Page 20

27 HUC_ Facility Size (# of facilities) Estimated Existing Loads 1 Total Phosphorus Total Nitrogen Daily Load (lbs/day) Annual Load (Tons) Daily Load (lbs/day) Annual Load (Tons) Large (2) Medium (8) Small (15) Total (25) Medium (7) Small (4) Total (11) Large (2) Medium (18) Small (28) Total (48) Large (3) , Medium (10) Small (16) Total (29) , Large (4) 3, ,162 4,957 Medium (11) Small (8) Total (23) 3, ,837 5,080 Large (8) 1, , Medium (13) Small (30) Total (51) 1, ,743 1,048 Medium (2) Small (8) Total (10) Large (9) 2, ,571 3,937 Medium (22) , Small (23) Total (54) 3, ,943 4,187 Large (5) , Medium (20) Small (36) Total (61) , Large (1) Medium (10) Small (17) Total (28) Large (6) 3, ,613 4,127 Medium (12) Small (25) Total (43) 3, ,254 4,244 Large (1) Medium (5) Small (10) Total (16) Final Report September 2011 Page 21

28 HUC_ Facility Size (# of facilities) Estimated Existing Loads 1 Total Phosphorus Total Nitrogen Daily Load (lbs/day) Annual Load (Tons) Daily Load (lbs/day) Annual Load (Tons) Large (16) 3, ,752 5,247 Medium (62) , Small (79) Total (157) 4, ,964 5,834 Large (5) 1, ,522 1,738 Medium (15) Small (22) Total (42) 1, ,576 1,930 Large (2) Medium (13) Small (15) Total (30) Large (4) , Medium (20) Small (30) Total (54) , Large (2) , Medium (7) Small (5) Total (14) , Large (1) Medium (11) Small (11) Total (23) Large (1) Medium (19) Small (15) Total (35) Large (2) , Medium (22) Small (28) Total (52) , Large (1) Medium (7) Small (9) Total (17) Existing loads based on flow and TN data reported in the Integrated Compliance Information System (ICIS). TN concentrations were not reported for the vast majority of facilities and therefore were estimated based on reported BOD values. See Characterization of Wabash River Nutrient Loads in Appendix C for details. Estimating existing loads for the different facility size categories in each 8-digit HUC helps to determine the potential change in pollutant loads when different permit effluent limits are considered. As discussed in Section 3.1, the Project Team has made assumptions about likely future permit effluent limits that would result from numeric nutrient criteria based on trends in other Midwest states. For purposes of the WQT market feasibility analysis, the Project Team developed scenarios using different assumed permit effluent limit values. For TN, the assumed values are 3 mg/l, 5 mg/l, and 8 mg/l. For TP, the assumed values are 0.3 mg/l and 0.5 mg/l. Final Report September 2011 Page 22

29 Table 7 summarizes the change in pollutant loads for TN under each permit effluent limit scenario. In addition, this table shows the associated pollutant load reduction that each facility size category is estimated to need to achieve the more stringent permit effluent limits. This is assumed to be the potential credit demand for each facility size category. Table 7. Changes in pollutant loads and resulting credit demand under different TN permit effluent scenarios HUC_ (Tippecanoe) Facility Size (# of facilities) Existing Loads Annual Load (Tons) Estimated Loads Assuming Discharge Value of 3 mg/l of 5 mg/l of 8 mg/l Annual Load Reduction Large (8) Medium (15) Small (39) Total (62) Large (1) Medium (7) Small (11) Total (19) Large (2) Medium (21) Small (16) Total (39) Large (2) Medium (8) Small (15) Total (25) Medium (7) Small (4) Total (11) Large (2) Medium (18) Small (28) Total (48) Large (3) Medium (10) Small (16) Total (29) Large (4) Medium (11) Small (8) Total (23) Large (8) Medium (13) Small (30) Total (51) Final Report September 2011 Page 23

30 HUC_ (Driftwood) Facility Size (# of facilities) Existing Loads Annual Load (Tons) Estimated Loads Assuming Discharge Value of 3 mg/l of 5 mg/l of 8 mg/l Annual Load Reduction Medium (2) Small (8) Total (10) Large (9) Medium (22) Small (23) Total (54) Large (5) Medium (20) Small (36) Total (61) Large (1) Medium (10) Small (17) Total (28) Large (6) Medium (12) Small (25) Total (43) Large (1) Medium (5) Small (10) Total (16) Large (16) Medium (62) Small (79) Total (157) Large (5) Medium (15) Small (22) Total (42) Large (2) Medium (13) Small (15) Total (30) Large (4) Medium (20) Small (30) Total (54) Large (2) Medium (7) Small (5) Total (14) Final Report September 2011 Page 24

31 HUC_ Facility Size (# of facilities) Existing Loads Annual Load (Tons) Estimated Loads Assuming Discharge Value of 3 mg/l of 5 mg/l of 8 mg/l Annual Load Reduction Large (1) Medium (11) Small (11) Total (23) Large (1) Medium (19) Small (15) Total (35) Large (2) Medium (22) Small (28) Total (52) Large (1) Medium (7) Small (9) Total (17) As shown by Table 7, large facilities in the Wabash River watershed have the most significant credit demand for TN under more stringent permit effluent limitations. However, there are significantly more small facilities compared to medium and large facilities and cumulatively they will also generate a significant credit demand. A portion of small rural facilities might experience difficulties such as being understaffed or not able to generate financial resources necessary to meet new restrictive nutrient requirements within the first permit period (typical of most compliance schedules requirements). Table 8 summarizes the change in pollutant loads for TP under each permit effluent limit scenario. This table also shows the associated pollutant load reduction that each facility size category is estimated to need to achieve the more stringent permit effluent limits. This is assumed to be the potential credit demand for each facility size category for TP. Final Report September 2011 Page 25

32 Table 8. Changes in pollutant loads and resulting credit demand under different TP permit effluent scenarios HUC_ (Tippecanoe) Facility Size Estimated Existing Loads Annual Load (Tons) Loads Assuming Discharge Value of 0.3 mg/l Annual Load (Tons) Loads Assuming Discharge Value of 0.5 mg/l Annual Load (Tons) Annual Load Reductions 0.3 Annual Load Reductions 0.5 Large (8) Medium (15) Small (39) Total (62) Large (1) Medium (7) Small (11) Total (19) Large (2) Medium (21) Small (16) Total (39) Large (2) Medium (8) Small (15) Total (25) Medium (7) Small (4) Total (11) Large (2) Medium (18) Small (28) Total (48) Large (3) Medium (10) Small (16) Total (29) Large (4) Medium (11) Small (8) Total (23) Large (8) Medium (13) Small (30) Total (51) Medium (2) Small (8) Total (10) Large (9) Medium (22) Final Report September 2011 Page 26

33 HUC_ (Driftwood) Facility Size Estimated Existing Loads Annual Load (Tons) Loads Assuming Discharge Value of 0.3 mg/l Annual Load (Tons) Loads Assuming Discharge Value of 0.5 mg/l Annual Load (Tons) Annual Load Reductions 0.3 Annual Load Reductions 0.5 Small (23) Total (54) Large (5) Medium (20) Small (36) Total (61) Large (1) Medium (10) Small (17) Total (28) Large (6) Medium (12) Small (25) Total (43) Large (1) Medium (5) Small (10) Total (16) Large (16) Medium (62) Small (79) Total (157) Large (5) Medium (15) Small (22) Total (42) Large (2) Medium (13) Small (15) Total (30) Large (4) Medium (20) Small (30) Total (54) Large (2) Medium (7) Small (5) Total (14) Large (1) Medium (11) Small (11) Total (23) Final Report September 2011 Page 27

34 HUC_ Facility Size Estimated Existing Loads Annual Load (Tons) Loads Assuming Discharge Value of 0.3 mg/l Annual Load (Tons) Loads Assuming Discharge Value of 0.5 mg/l Annual Load (Tons) Annual Load Reductions 0.3 Annual Load Reductions 0.5 Large (1) Medium (19) Small (15) Total (35) Large (2) Medium (22) Small (28) Total (52) Large (1) Medium (7) Small (9) Total (17) Existing Wastewater Treatment and Estimated Upgrade Costs Estimating potential credit demand goes beyond estimating the necessary pollutant load reductions to meet the TN and TP permit effluent limitation scenarios. It also involves understanding the existing type of treatment and which facilities might choose to upgrade their control technologies to meet the more stringent permit effluent limitations. The Project Team conducted an analysis of the type of wastewater treatment and the cost to upgrade. This analysis is part of the Characterization of Wabash River Nutrient Loads technical memorandum found in Appendix C. The level of new pollutant control measures needed to meet nutrient reductions specified by TMDLs or other regulatory drivers will be dependent upon each treatment plant s current operations and the cost associated with the most likely control measure (e.g., biological phosphorus removal). Information on the type of wastewater treatment used by plants within the Wabash River watershed was obtained from the Clean Watersheds Needs Survey (CWNS) for the entire watershed. For the Driftwood and Tippecanoe subwatersheds, the Project Team supplemented CWNS information with data from a review of actual NPDES permits. Information from the CWNS places facilities under the category of Secondary Wastewater Treatment or the category of Advanced Wastewater Treatment. Secondary treatment typically requires a treatment level that produces an effluent quality of less than 30 mg/l of both BOD5 and total suspended solids (secondary treatment levels required for some lagoon systems may be less stringent). In addition, the secondary treatment must remove 85 percent of BOD5 and total suspended solids from the influent wastewater. A facility is considered to have Advanced Wastewater Treatment if its permit includes one or more of the following: BOD less than 20 mg/l; Nitrogen Removal; Phosphorous Removal; Ammonia Removal; Metal Removal; Synthetic Organic Removal. Table 9 summarizes the CWNS information for all of the facilities in the Wabash River watershed. It indicates that 21 percent of the facilities have advanced treatment, 14 percent are known to have secondary treatment, and no information is available for 65 percent of the facilities. Because the CWNS Final Report September 2011 Page 28

35 focuses on larger facilities, and because smaller facilities are less likely to have advanced treatment, it is probable that the majority of the facilities with no information do not use advanced treatment. Table 9. Summary of CWNS information for all facilities in the Wabash River watershed. Treatment Level Number of Facilities Percent Advanced Treatment % Secondary % No Information % Total Facilities in Watershed % Note: the number of facilities in the CWNS (1021) differs from the number obtained by combining the data from IDEM and IEPA (943). The source of the discrepancy is unknown but may be due to the timing of when the two data sets were created. Table 10 summarizes the indicators of advanced treatment and shows that only a small proportion (10%) of the facilities have limits for phosphorus and nitrogen (not including ammonia). Table 10. Indicators associated with advanced treatment facilities in the Wabash River watershed. Advance Indicators Number of Facilities Percent BOD % Nitrogen 1 0.5% BOD, Nitrogen 1 0.5% BOD, Phosphorus 2 0.9% BOD, Ammonia % BOD, Phosphorus, Ammonia % Total % Table 11 summarizes the type of treatment by HUC and indicates that HUCs , , and have the most facilities with advanced treatment. The cities of Terra Haute, Bloomington, Indianapolis, Anderson, and Muncie are located in these HUCs. Final Report September 2011 Page 29

36 Table 11. Summary of treatment type by HUC. HUC 8 Advanced Treatment Secondary Treatment No CWNS Information (Tippecanoe) (Driftwood) Total A more detailed analysis was performed to determine the type of treatment for facilities in the Tippecanoe and Driftwood watersheds (which are being modeled in SWAT to support the feasibility study). The CWNS information for the 92 facilities in these two watersheds is shown in Table 12. Seventeen of the facilities have advanced treatment, nine have secondary treatment, and treatment type was not reported for 66 facilities. Final Report September 2011 Page 30

37 Table 12. Summary of 2008 CWNS permit information for facilities in the Tippecanoe and Driftwood watersheds. Watershed Tippecanoe ( ) Driftwood ( ) Permit Type WWTP or Other Present Treatment Level Present Advance Indicators Number of Facilities Major Other No Info No Info 2 Major WWTP Advanced Treatment BOD (Biochemical Oxygen Demand) Minor Other No Info No Info 23 Minor Other Secondary No Info 1 Minor WWTP Advanced Treatment Minor WWTP Advanced Treatment Minor WWTP Advanced Treatment Ammonia Removal, BOD (Biochemical Oxygen Demand) BOD (Biochemical Oxygen Demand) BOD (Biochemical Oxygen Demand), Ammonia Removal Minor WWTP No Info No Info 3 Minor WWTP Secondary No Info 4 Major Other No Info No Info 1 Major WWTP Advanced Treatment BOD (Biochemical Oxygen Demand) Major WWTP Secondary No Info 1 Minor Other No Info No Info 34 Minor WWTP Advanced Treatment Minor WWTP Advanced Treatment BOD (Biochemical Oxygen Demand) BOD (Biochemical Oxygen Demand), Ammonia Removal Minor WWTP No Info No Info 3 Minor WWTP Secondary No Info Of the 92 facilities, the Project Team was able to obtain permits for the 31 facilities that are WWTP within the watershed. These permits indicate that the following specific types of treatment methods are used within the watershed (either individually or in combination with one another): trickling filters activated sludge (including extended aeration and oxidation) discharge waste stabilization lagoons Out of these only the activated sludge systems can effectively be retrofit for biological nitrogen removal. Therefore, the trickling filter and lagoon treatment processes will require an expansion to reduce nutrient in their effluent. All 31 facilities are required to report CBOD5 and ammonia values, 21 of the facilities have permitted ammonia limits, and only one facility has a phosphorus limit (Table 13). Based on the lack of TN or TP limits in the available permits, and the types of treatment processes specified in the CWNS and the collected permits, it appears that almost none of the facilities in the Driftwood and Tippecanoe are targeting the treatment of TN or TP. The facilities that are treating for ammonia (via nitrification) likely do not include technology for denitrification or phosphorus removal. Final Report September 2011 Page 31

38 Upgrades would therefore be necessary to provide enhanced nutrient removal (ENR) for TN and TP removal. Table 13. Permit limit summaries for facilities in the Driftwood and Tippecanoe watersheds. Permit Limit No Designation in CWNS Secondary Advanced Design Flow (MGD) to to to 5.13 Monthly Average CBOD5 Summer (mg/l) 12 to to to 25 Monthly Average CBOD5 Winter (mg/l) 12 to to 30 Weekly Average CBOD5 Summer (mg/l) 18 to to to 40 Weekly Average CBOD5 Winter (mg/l) 18 to to 45 Monthly Average Ammonia Summer (mg/l) 1.3 to to to 8.6 Monthly Average Ammonia Winter (mg/l) 2 to to to 11 Weekly Average Ammonia Summer (mg/l) 2 to to to 12.9 Weekly Average Ammonia Winter (mg/l) 3 to to to 16.5 Phosphorus Limit (mg/l) N/A N/A 1 To understand potential credit demand, it is necessary to understand the estimated costs for upgrading technology to provide ENR for TN and TP removal. This section of the report provides estimated costs for upgrading permitted WWTPs in the Driftwood and Tippecanoe watersheds to enhanced nutrient ENR for reducing TN and TP effluent loads. The costs are based only on cited literature and the information available in the CWNS and the permits and are intended solely to inform the feasibility study. Actual upgrade costs vary widely, depending on a great number of factors, including: actual target effluent concentrations for nitrogen and phosphorus; existing facilities suitability for various types of upgrades; various wastewater characteristics, including influent TP and TN concentrations, influent rbcod:tp and BOD:TN ratios, alkalinity levels, actual flow and constituent concentrations and their hourly, daily, and seasonal variations; various operating characteristics, including ambient temperatures, mixed liquor characteristics, and plant configuration and control methods; local labor, material, and operational costs which may vary significantly over time; and financing terms. The cost estimates are intended only as a general guide for order-of-magnitude cost ranges for various upgrade options. More detailed, plant-specific analyses are necessary to determine whether WQT would be a more cost-effective option for reducing nutrient loading. In particular, WQT may be advantageous for: long-term compliance and cost savings; short-term compliance for the useful life of the facility or when other regulatory requirements will be better understood in regards to needed upgrades; variance or compliance schedule justification; small or difficult to upgrade facilities; and future growth in fully capped watersheds. Final Report September 2011 Page 32

39 The analysis of permits showed that WWTPs in the Driftwood and Tippecanoe subwatersheds span a variety of process types for permitted flow rates ranging from about 24,000 gallons per day (gpd) to 5.0 million gallons per day (MGD). For the purposes of this evaluation, the secondary (i.e., biological) treatment process identified in the permits for each of these plants was used to categorize the WWTP as either an activated sludge system or a lagoon/trickling filter system. The reason that this simple categorization was chosen is because activated sludge systems are generally relatively simple to convert to biological nutrient removal (BNR) systems (by adding anaerobic and/or anoxic reactors), while lagoons and trickling filters generally require more extensive secondary treatment system modifications to be upgraded to BNR. A number of specific treatment process were listed in the permits for WWTPs categorized as activated sludge systems, including extended aeration, oxidation ditch, waste stabilization, sequencing batch reactor, and activated sludge or conventional activated sludge. The numbers of facilities in each of these two main categories across a range of flow rates are summarized in Table 14. Based on this summary of the characteristics of WWTPs in the target watersheds, seven generic WWTPs (design flow/system type) were used to evaluate the range of potential costs that may be required to upgrade WWTPs in the Driftwood and Tippecanoe watersheds to nitrogen removal, phosphorus removal, or both. The generic WWTPs simulated in the cost analysis are also identified in Table 14. The design flow of each simulated plant is roughly equal to the average flow rate for the WWTPs in a given flow range. Table 14. Summary of facility type and flows for WWTPs included in Driftwood and Tippecanoe nutrient removal analysis Flow Range (MGD) # Facilities Activated Sludge (AS) Total Flow (MGD) Simulation Plant # Facilities Lagoon/Trickling Filter (TF) Total Flow (MGD) Simulation Plant < MGD AS MGD Lagoon 0.1 to MGD AS MGD Lagoon 0.5 to MGD AS to MGD TF > MGD AS In addition to the variety of process types and flow ranges identified in Table 14, for each simulation plant a number of different nutrient removal upgrade options were evaluated. In general, selection of the specific nutrient removal upgrade options simulated was based on the availability of cost information for those options, as found in the literature. Based on the review of relevant literature and previous nutrient removal experience, the Project Team selected two generic levels of treatment for nitrogen removal and two levels of treatment for phosphorus removal for the cost analysis. These levels are as follows: TN1. The low enhanced nutrient removal (ENR) treatment level for nitrogen (or TN1) can be met by adding anoxic reactors (along with nitrified mixed liquor recycle lines) prior to the existing secondary treatment process ( pre-anoxic ) or adding post-secondary anoxic treatment (typically using filters supplemented with an external carbon source), for denitrification. Land application of effluent was also designated as a potential option for the TN1 treatment level, with nitrogen removal attributed to both nitrification/denitrification processes and vegetative Final Report September 2011 Page 33

40 uptake and sequestration (Crites and Tchobanoglous 1998). These processes have been documented to meet a TN of 10 mg/l reliably, but may be designed to meet lower TN levels. TN2. The high ENR treatment level for nitrogen (or TN2) requires both pre- and postsecondary anoxic reactors and has been demonstrated as capable of achieving effluent TN concentrations below 5 mg/l (typically, 2-3 mg/l). TP1. The low ENR level for phosphorus (TP1), which requires an effluent TP of 1 mg/l or less, can be met using enhanced biological phosphorus removal (EBPR), typically involving the addition of anaerobic selector reactors prior to the secondary treatment unit, or using alum, which is typically dosed between the secondary treatment process and the secondary clarifier (but potentially in other configurations), for precipitating phosphorus. TP2. The high ENR treatment level for phosphorus (TP2) requires either multi-point alum addition or EBPR with single- or multi-point alum addition and enhanced solids removal processes can be used to reach TP levels below 0.5 mg/l, often down to 0.1 mg/l or lower. Land application systems are also well documented to be able to meet TP2 treatment levels. These treatment levels are summarized in Table 15, along with the assumptions for influent and baseline (i.e., effluent levels in the absence of ENR) concentrations. For the purposes of the cost analysis, the Project Team assumed a baseline TN concentration of 25 mg/l and TP concentration of 4 mg/l, the midpoints of the ranges shown in Table 15. In the cost calculations, these represent the assumed average effluent concentrations for the WWTPs prior to implementing an ENR process. Another important implicit assumption in all of the cost calculations is that existing secondary treatment processes are nitrifying or can be made to nitrify that is, they are sufficient to convert the majority of influent organic nitrogen and ammonia to nitrate, such that denitrification retrofits would be effective upgrade options. It is important to note that, even under a WQT approach to meeting TN reductions, participating WWTPs would still need to meet existing or revised water quality based effluent discharge limits for ammonia, a potentially mobile and toxic wastewater constituent. Under a WQT approach, nitrifying WWTPs would continue to discharge TN in the form of nitrate at non-toxic levels. Table 15. Summary of ENR treatment levels and assumptions for WWTP upgrade simulations TN Treatment Level Effluent AS options None (influent) mg/l 4-8 mg/l Baseline (no ENR) Low ENR (1) mg/l 2-6 mg/l 5-10 mg/l Pre- or postanoxic retrofit or land application High ENR (2) 1 anoxic retrofit <5 mg/l Pre-/Post- Lagoon/ TF options Effluent AS options Lagoon/TF options Post-anoxic replacement or land application Post-anoxic replacement TP mg/l EBPR or singlepoint alum retrofit <0.5 mg/l EBPR and/or multi-point alum retrofit or land application 1 Enhanced solids removal also generally required for the high ENR process upgrades, particularly for TP removal EBPR replacement or single-point alum retrofit EBPR replacement and/or multi-point alum retrofit or land application Final Report September 2011 Page 34

41 Table 16 summarizes the characteristics of all applicable permutations of the TN and TP treatment levels as used in the cost analysis and provides a guide to the color coded rows in Table 17 through Table 23, which present the results of the upgrade cost analyses. Table 16. Color coding of ENR treatment levels for Tables Treatment Level Effluent TN Effluent TP Color Coding TN mg/l -- Pink TP mg/l Blue TN2 <5 mg/l -- Tan TP2 -- <0.5 mg/l Olive TN1/TP mg/l mg/l Green TN1/TP mg/l <0.5 mg/l Purple TN2/TP2 <5 mg/l <0.5 mg/l Aqua Note that the calculations for cost per pound (cost/#) removed that are summarized in Table 17 through Table 23 use the actual effluent TN and TP treatment levels indicated in the cost reference cited, while also assuming that the design/permitted flows are the actual plant flows (this assumption has little bearing on capital costs, but does affect O&M cost estimates). To be consistent with the basis used in U.S. EPA s nutrient removal reference document (U.S. EPA 2008), the Project Team converted all costs to annual costs assuming 20-year financing terms and a 6 percent interest rate. Additionally, all costs in Table 17 through Table 23 are presented in 2011 dollars using the latest ENR construction cost index (9011, March 2011) as a basis for adjusting the costs generated from the various sources indicated in the tables. Finally, note that where costs per pound removed are presented in Table 17 through Table 23, total costs are used in all calculations. Accordingly, it is difficult to compare the cost per pound of TP removed for a TN1/TP1 system with the cost per pound of TP removed for a TP1 system, for example, since the former number includes costs necessary for TN removal in addition to TP removal, while the latter number would be only associated with TP removal. Table 17 and Table 18 provide estimated costs for 0.05 MGD activated sludge and lagoon upgrade options, respectively. Note that the cost estimates for the single-point alum addition upgrade option may be biased high, as the capital costs cited by Keplinger, et al. (2003) were significantly higher than those cited in comparable references addressing single-point alum treatment (i.e., as summarized in Table 19 and Table 21). Final Report September 2011 Page 35

42 Upgraded Process Table MGD activated sludge ENR upgrade options and costs Annual Cost ($/yr) Treatment Level Cost TN TP $/# TN $/# TP $/# TN+TP Reference MLE added anoxic zone 36, mg/l 2 mg/l Foess (1998) 1 Single-point alum addition 90, mg/l Keplinger (2003) 2 MLE + denitrification filters 59,636 6 mg/l 1 mg/l Foess (1998) 3 Land app. spray irrigation 152, mg/l 0.1 mg/l Buchanan (2010) 4 Land app. drip irrigation 82, mg/l 0.1 mg/l Buchanan (2010) 4 1 Used present worth costs for 50,000 gpd anoxic tank for MLE upgrade retrofit system (option R1) 2 Used average capital and O&M costs for Iredell (0.25 MGD), Valley Mills (0.81 MGD), and Hico (0.87 MGD) plant single-point alum addition upgrades, normalized as unit costs ($/gpd capacity) 3 Used present worth costs for 50,000 gpd deep bed denitrification filter upgrade retrofit system (option R2) 4 Used model-simulated costs for 50,000 gpd systems installed in loam soils (per NRCS Web Soil Survey data for area) and added capital costs for land acquisition (assuming $5,000/acre) and a 25% allowance for various professional fees, both of which are not included in the model Upgraded Process Table MGD lagoon ENR upgrade options and costs Annual Cost ($/yr) Treatment Level Cost TN TP $/# TN $/# TP $/# TN+TP Reference New MLE system 181, mg/l 2 mg/l Foess (1998) 1 Single-point alum addition 90, mg/l Keplinger (2003) 2 New MLE + denitrification filters Land app. spray irrigation 203,661 6 mg/l 1 mg/l Foess (1998) 3 152, mg/l 0.1 mg/l Buchanan (2010) 4 Land app. drip irrigation 82, mg/l 0.1 mg/l Buchanan (2010) 4 1 Used present worth costs for 50,000 gpd MLE process system (option 1) 2 Used average capital and O&M costs for Iredell (0.025 MGD), Valley Mills (0.081 MGD), and Hico (0.087 MGD) plant single-point alum addition upgrades, normalized as unit costs ($/gpd capacity) 3 Used present worth costs for 50,000 gpd MLE and deep-bed filtration process system (option 6) 4 Used model-simulated costs for 50,000 gpd systems installed in loam soils (per NRCS Web Soil Survey data for area) and added capital costs for land acquisition (assuming $5,000/acre) and a 25% allowance for various professional fees, both of which are not included in the model The results presented in Table 17 and Table 18 indicate that for 0.05 MGD activated sludge systems, secondary treatment upgrades and alum addition would typically be most cost effective, while comparably-sized lagoon systems may be cost-effectively upgraded for ENR via land application. Table 19 and Table 20 provides estimated costs for 0.3 MGD activated sludge and lagoon upgrade options, respectively, with average costs for each treatment level provided where more than one option is presented. As indicated above, the average cost estimates for the single-point alum addition upgrade option may be biased high, as the capital costs cited by Keplinger, et al. (2003) were significantly higher than those cited by CH2M-Hill (2010) and others for single-point alum treatment. Final Report September 2011 Page 36

43 Upgraded Process Table MGD activated sludge ENR upgrade options and costs Annual Cost ($/yr) Treatment Level Cost TN TP $/# TN $/# TP $/# TN+TP Reference Single-point alum addition 214,146 1 mg/l Keplinger (2003) 2 Single-point alum addition 10,552 1 mg/l CH2M Hill (2010) 3 TP1 AVERAGE 112, Not specified 154,196 3 mg/l Colorado (2010) 4 Multi-point alum addition 157, mg/l CH2M Hill (2010) 3 Not specified 269, mg/l Colorado (2010) 4 TP2 AVERAGE 213, Not specified 212, mg/l 1 mg/l U.S. EPA (2007) 5 Not specified 469, mg/l 0.8 mg/l M&E (2008) 6 TN1/TP1 AVERAGE 340, Land app. spray irrigation 910, mg/l 0.1 mg/l Buchanan (2010) 7 Land app. drip irrigation 483, mg/l 0.1 mg/l Buchanan (2010) 7 TN1/TP2 AVERAGE 696, Capital costs only. O&M costs not included. 2 Used average capital and O&M costs for Valley Mills (0.081 MGD), Hico (0.087 MGD), and Clifton (0.328 MGD) plant single-point alum addition upgrades, normalized as unit costs ($/gpd capacity) 3 Used average total life cycle costs for Oakley (0.25 MGD), Coalville (0.35 MGD), and Fairview (0.375 MGD) activated sludge plant upgrades, normalized as unit costs ($/gpd capacity) 4 Used average capital and O&M costs for 0.1 MGD plant upgrades from Table 6 of paper, normalized as unit costs ($/gpd capacity) 5 Used average unit capital costs for BNR upgrades based on MD and CT WWTPs for flow range of > MGD ($6,972,000/mgd capacity) 6 Used capital cost equation for BNR upgrades in PA for WWTPs <10 MGD (cost, $M = MGD) 7 Used model-simulated costs for 300,000 gpd systems installed in loam soils (per NRCS Web Soil Survey data for area) and added capital costs for land acquisition (assuming $5,000/acre) and a 25% allowance for various professional fees, both of which are not included in the model Final Report September 2011 Page 37

44 Upgraded Process Table MGD lagoon ENR upgrade options and costs Annual Cost Treatment Level Cost TN TP $/# TN $/# TP $/# TN+TP New MLE system 352,708 3 mg/l Reference CH2M Hill (2010) 2 Single-point alum 80,433 1 mg/l CH2M Hill (2010) 2 Single-point alum 214,146 1 mg/l Keplinger (2003) 3 TP1 AVERAGE 147, Multi-point alum + filters 307, mg/l CH2M Hill (2010) 2 Not specified 269, mg/l Colorado (2010) 4 TP2 AVERAGE 288, Not specified 212, mg/l 1 mg/l U.S. EPA (2007) 5 Not specified 469,884 6 mg/l 0.8 mg/l M&E (2008) 6 TN1/TP1 AVERAGE 340, Land application - spray 910, mg/l 0.1 mg/l Land application - drip 483, mg/l 0.1 mg/l TN1/TP2 AVERAGE 696, Capital costs only. O&M costs not included. 2 Used average total life cycle costs for 0.55 MGD lagoon retrofits, normalized as unit costs ($/gpd capacity) Buchanan (2010) 7 Buchanan (2010) 7 3 Used average capital and O&M costs for Valley Mills (0.081 MGD), Hico (0.087 MGD), and Clifton (0.328 MGD) plant single-point alum addition upgrades, normalized as unit costs ($/gpd capacity) 4 Used average capital and O&M costs for 0.1 MGD plant upgrades from Table 6 of paper, normalized as unit costs ($/gpd capacity) 5 Used average unit capital costs for BNR upgrades based on MD and CT WWTPs for flow range of > MGD ($6,972,000/mgd capacity) 6 Used capital cost equation for BNR upgrades in PA for WWTPs <10 MGD (cost, $M = MGD) 7 Used model-simulated costs for 300,000 gpd systems installed in loam soils (per NRCS Web Soil Survey data for area) and added capital costs for land acquisition (assuming $5,000/acre) and a 25% allowance for various professional fees, both of which are not included in the model The results presented in Table 19 and Table 20 indicate that at these relatively low design flows, land application may be a cost-effective ENR upgrade option, compared with more traditional secondary treatment upgrades or multi-point alum addition, as needed to achieve comparably low TP levels. Table 21 provides estimated costs for 0.75 MGD activated sludge upgrade options, with average costs for each treatment level provided where more than one option is presented. Because the costs cited by Keplinger (2003) were significantly higher than those cited by CH2M-Hill (2010) and U.S. EPA (2008) for single-point alum treatment, the Keplinger data was not used in the average cost calculation for TP1 treatment level upgrade options. Final Report September 2011 Page 38

45 Upgraded Process Table MGD activated sludge ENR upgrade options and costs Annual Cost ($/yr) Treatment Level Cost TN TP $/# TN $/# TP $/# TN+TP Reference MLE added anoxic zone 72, mg/l CH2M Hill (2010) 3 Single-point alum addition 390,779 1 mg/l Keplinger (2003) 4 EBPR or single-point alum addition 29,232 1 mg/l CH2M Hill (2010) 3 Fermenter addition 28, mg/l U.S. EPA (2008) 5 Single-point alum addition 60, mg/l U.S. EPA (2008) 5 Fermenter and filter addition 60, mg/l U.S. EPA (2008) 5 TP1 AVERAGE 44, Phased Isolation Ditch retrofit 69,016 3 mg/l U.S. EPA (2008) 5 MLE retrofit 111,316 3 mg/l U.S. EPA (2008) 5 Step-feed retrofit 111,316 3 mg/l U.S. EPA (2008) 5 Denitrification filter retrofit 230,053 3 mg/l U.S. EPA (2008) 5 TN2 AVERAGE 130, EBPR + multi-stage alum + filters Fermenter, filter, and alum addition 166, mg/l CH2M Hill (2010) 3 118, mg/l U.S. EPA (2008) 5 Multi-point alum and filter addition 134, mg/l U.S. EPA (2008) 5 TP2 AVERAGE 139, Not specified 530, mg/l 1 mg/l U.S. EPA (2007) 6 Not specified 557,904 6 mg/l 0.8 mg/l M&E (2008) 7 TN1/TP1 AVERAGE 543, Land application - spray 2,275, mg/l 0.1 mg/l Buchanan (2010) 8 Land application - drip 1,169, mg/l 0.1 mg/l Buchanan (2010) 8 TN1/TP2 AVERAGE 1,722, Phased Isolation Ditch retrofit 5-stage act. sludge + alum retrofit 238,958 3 mg/l 0.1 mg/l U.S. EPA (2008) 5 271,611 3 mg/l 0.1 mg/l U.S. EPA (2008) 5 Alum addition + denitrification filter 352,253 3 mg/l 0.1 mg/l U.S. EPA (2008) 5 TN2/TP2 AVERAGE 287, Average does not include Keplinger data 2 Capital costs only. O&M costs not included. 3 Used average total life cycle costs for Fairview (0.375 MGD), Moroni (0.9 MGD), Hyrum City (1.3 MGD), and Tremonton (1.9 MGD) activated sludge plant upgrades, normalized as unit costs ($/gpd capacity) 4 Used average capital and O&M costs for Hico (0.087 MGD), Clifton (0.328 MGD), and Meridian (0.36 MGD) plant single-point alum addition upgrades, normalized as unit costs ($/gpd capacity) 5 Used extrapolated life cycle costs per MG treated for retrofit options, normalized to $/gpd capacity 6 Used average unit capital costs for BNR upgrades based on MD and CT WWTPs for flow range of > MGD ($6,972,000/mgd capacity) 7 Used capital cost equation for BNR upgrades in PA for WWTPs <10 MGD (cost, $M = MGD) 8 Used model-simulated costs for 750,000 gpd systems installed in loam soils (per NRCS Web Soil Survey data for area) and added capital costs for land acquisition (assuming $5,000/acre) and a 25% allowance for various professional fees, both of which are not included in the model Final Report September 2011 Page 39

46 The results presented in Table 21 indicate that at these somewhat higher design flows, traditional secondary treatment upgrades or alum addition are more cost effective than land application for achieving very low TP levels. Table 22 provides estimated costs for 2.5 MGD trickling filter upgrade options, with average costs for each treatment level provided where more than one option is presented. For this higher flow rate, we assumed that land application would no longer be a viable upgrade option. Upgraded Process Table MGD trickling filter ENR upgrade options and costs Annual Cost ($/yr) Treatment Level Cost TN TP $/# TN $/# TP $/# TN+TP New MLE system 163, mg/l New phased isolation ditch system Reference CH2M Hill (2010) 2 376,001 5 mg/l U.S. EPA (2008) 3 New MLE system 1,009,265 5 mg/l U.S. EPA (2008) 3 New SBR system 1,128,002 5 mg/l U.S. EPA (2008) 3 New 4-stage bardenpho system 1,385,265 5 mg/l U.S. EPA (2008) 3 TN1 AVERAGE 812, New A/O system 811,098 1 mg/l U.S. EPA (2008) 3 Single-point alum addition 144,841 1 mg/l CH2M Hill (2010) 2 Single-point alum addition 168, mg/l U.S. EPA (2008) 4 New A/O w/fermenters 841, mg/l U.S. EPA (2008) 3 New A/O w/fermenters + filters New mod UCT w/fermenters + filters 989, mg/l U.S. EPA (2008) 3 1,504, mg/l U.S. EPA (2008) 3 New 5-stage bardenpho w/filters 1,553, mg/l U.S. EPA (2008) 3 TP1 AVERAGE 858, New denitrification filters 662,948 3 mg/l U.S. EPA (2008) 4 New A/O w/fermenters + filters + alum Multi-stage alum addition with filters New A/O with fermenters, filters, alum 1,137, mg/l U.S. EPA (2008) 3 1,307, mg/l CH2M Hill (2010) 2 1,137, mg/l U.S. EPA (2008) 3 Multi-point alum addition with filters 395, mg/l U.S. EPA (2008) 4 TP2 AVERAGE 994, New A/O system 3,860, mg/l 1 mg/l Jiang et al (2004) 5 Not specified 441, mg/l 1 mg/l U.S. EPA (2007) 6 Not specified 1,039,337 6 mg/l 0.8 mg/l M&E (2008) 7 New 3-stage UCT system 1,345,686 5 mg/l 1 mg/l U.S. EPA (2008) 3 Final Report September 2011 Page 40

47 Annual Treatment Level Cost Upgraded Process Cost ($/yr) TN TP $/# TN $/# TP $/# TN+TP Reference New step feed AS system 900,422 5 mg/l 1 mg/l U.S. EPA (2008) 3 New 5-stage Bardenpho 1,434,739 5 mg/l 0.5 mg/l U.S. EPA (2008) 3 TN1/TP1 AVERAGE 1,503, New A/A/O with alum + filters 5,301, mg/l 0.1 mg/l Alum addition with denitrification filters Jiang et al (2004) 5 989,475 3 mg/l 0.1 mg/l U.S. EPA (2008) 4 New phased isolation ditch+alum+filters 722,317 3 mg/l 0.1 mg/l U.S. EPA (2008) 3 New SBR + alum + fiters 1,236,844 3 mg/l 0.1 mg/l U.S. EPA (2008) 3 New 5-stage Bardenpho + alum + filters 1,682,108 3 mg/l 0.1 mg/l U.S. EPA (2008) 3 TN2/TP2 AVERAGE 1,986, Capital costs only. O&M costs not included. 2 Used average total life cycle costs for Tremonton (1.9 MGD), Snyderville (2.4 MGD), and Magna (3.3 MGD) activated sludge plant upgrades, normalized as unit costs ($/gpd capacity) 3 Used interpolated life cycle costs per MG treated for expansion options, normalized to $/gpd capacity 4 Used interpolated life cycle costs per MG treated for retrofit options, normalized to $/gpd capacity 5 Interpolated between 1 MGD and 10 MGD de novo plant options 6 Used average unit capital costs for BNR upgrades based on MD and CT WWTPs for flow range of > MGD ($1,742,000/mgd capacity) 7 Used capital cost equation for BNR upgrades in PA for WWTPs <10 MGD (cost, $M = MGD) The results presented in Table 22 indicate that for replacement systems associated with trickling filter upgrades at these higher design flows, the additional costs associated with higher levels of ENR (e.g., TN1/TP1->TN2/TP2) are relatively modest. Table 23 provides estimated costs for 5 MGD activated sludge upgrade options, with average costs for each treatment level provided where more than one option is presented. As for the 2.5 MGD option, for this higher flow rate, we assumed that land application would no longer be a viable upgrade option. Final Report September 2011 Page 41

48 Upgraded Process Table MGD activated sludge ENR upgrade options and costs Annual Cost Treatment Level Cost TN TP $/# TN $/# TP $/# TN+TP MLE added anoxic zone 457, mg/l Reference CH2M Hill (2010) 2 CH2M Hill Single-point alum addition 271,677 1 mg/l (2010) 2 Fermenter addition 118, mg/l U.S. EPA (2008) 3 Single-point alum addition 237, mg/l U.S. EPA (2008) 3 Fermenter addition with filters 316, mg/l U.S. EPA (2008) 3 TP1 AVERAGE 236, Phased isolation ditch retrofit 336,422 3 mg/l U.S. EPA (2008) 3 MLE retrofit 554,106 3 mg/l U.S. EPA (2008) 3 Step feed retrofit 554,106 3 mg/l U.S. EPA (2008) 3 Denitrification filters 1,048,844 3 mg/l U.S. EPA (2008) 3 TN2 AVERAGE 623, EBPR + multi-point alum addition and filters Fermenter addition with alum and filters 1,721, mg/l CH2M Hill (2010) 2 504, mg/l U.S. EPA (2008) 3 Multi-point alum addition with filters 653, mg/l U.S. EPA (2008) 3 TP2 AVERAGE 959, A/O retrofit + alum addition 518, mg/l 1 mg/l Jiang et al(2005) 4 Not specified 882, mg/l 1 mg/l U.S. EPA (2007) 5 Not specified 1,699,014 6 mg/l 0.8 mg/l M&E (2008) 6 TN1/TP1 AVERAGE 1,033, Jiang et al A/A/O system + alum + filters 2,383, mg/l 0.1 mg/l (2005) 4 PID retrofit 1,068,633 3 mg/l 0.1 mg/l U.S. EPA (2008) 3 5-stage w/chem P 1,286,318 3 mg/l 0.1 mg/l U.S. EPA (2008) 3 Alum addition w/denitrification filters 1,484,213 3 mg/l 0.1 mg/l U.S. EPA (2008) 3 TN2/TP2 AVERAGE 1,555, Capital costs only. O&M costs not included. 2 Used average total life cycle costs for Payson (4.5 MGD), Brigham (6 MGD), and Spanish Fork (6 MGD) activated sludge plant upgrades, normalized as unit costs ($/gpd capacity) 3 Used interpolated life cycle costs per MG treated for retrofit options, normalized to $/gpd capacity 4 Interpolated between 1 MGD and 10 MGD retrofit plant options 5 Used average unit capital costs for BNR upgrades based on MD and CT WWTPs for flow range of > MGD ($1,742,000/mgd capacity) 6 Used capital cost equation for BNR upgrades in PA for WWTPs <10 MGD (cost, $M = MGD) The results presented in Table 23 indicate that for a 5 MGD activated sludge upgrade, the cost increase associated with meeting a TP2 standard versus TP1 are significant, although the potential cost increase associated with meeting high versus low ENR combined TN/TP standards are less pronounced. Final Report September 2011 Page 42

49 3.8 Potential Credit Supply Estimation of potential credit supply from agricultural reductions was assessed using two models: SPARROW and SWAT. The SPARROW loads are based on those published by USGS whereas the Project Team set up and calibrated the SWAT model for the Tippecanoe and Driftwood River watersheds. Evaluation of both models was required to increase the resolution of watershed assessment procedures at scale. Results from the SPARROW model are provided in Table 24 and Table 25 for TN and TP, respectively. SPARROW model estimates of agricultural NPS loading are provided in Table 26. The SPARROW model estimates for the Wabash River watershed indicate a substantial amount of TN and TP are generated within and exported at the mouth of each 8-Digit HUC. Each HUC estimate is independent of the upstream loading passing through the HUC. The estimated loading from all land use categories and all HUCs is 923,700 pounds TN and 87,900 pounds TP per year. The estimated cumulative agricultural loading that was generated by the model within each 8-Digit HUC and exported at the mouth is 762,200 pounds of TN and 79,500 pounds of TP per year. The maximum standard error (combined fertilizer and manure related standard errors) is 5 percent for TN and 26 percent for TP. The average maximum standard error estimate is 2 percent for TN and 9 percent for TP. The use of average results does not provide sufficient resolution or information on spatial variability. WQT program frameworks target optimum sites that generate economical transactions. Using an average pound per acre estimate helps illustrate the limited usefulness of the results. The entire watershed has approximately 21,088,000 acres. Using these figures, the average loading from agricultural land uses in the watershed is 0.04 and pounds per acre for TN and TP, respectively. Assuming an average BMP reduction of 20 percent, it would take approximately 125 acres to generate 1 pound of TN load reduction. A finer resolution can capture the spatial variability. Targeting fields that have higher reduction capability has the potential to provide more economical transactions. In addition, the SPARROW model estimates already include reductions attributable to attenuation dynamics within each 8-Digit HUC. Final Report September 2011 Page 43

50 Table 24. Total Nitrogen Exported at the Mouth of 8-Digit HUC Watersheds, Independent of Upstream Watershed Loading (USGS, 1997). Watershed Subwatershed HUC Sq. Miles Acres 1997 SPARROW Estimated Total Nitrogen Exported (Pounds) Average Pounds/Acre Wabash Upper Wabash 5,120,101 1,589 1,016,960 57, Wabash Salamonie 5,120, ,280 9, Wabash Mississinewa 5,120, ,880 24, Wabash Eel 5,120, ,560 14, Wabash Middle Wabash-Deer 5,120, ,280 30, Wabash Tippecanoe 5,120,106 1,960 1,254,400 46, Wabash Wildcat 5,120, ,320 29, Wabash Middle Wabash-Little Vermilion 5,120,108 2,287 1,463,680 73, Wabash Vermilion 5,120,109 1, ,960 37, Wabash Sugar 5,120, ,120 20, Wabash Middle Wabash-Busseron 5,120,111 2,019 1,292,160 61, Wabash Embarras 5,120,112 2,442 1,562,880 82, Wabash Lower Wabash 5,120,113 1, ,440 65, Wabash Little Wabash 5,120,114 2,148 1,374,720 48, Wabash Skillet 5,120,115 1, ,360 17, Patoka-White Upper White 5,120,201 2,754 1,762,560 66, Patoka-White Lower White 5,120,202 1,675 1,072,000 48, Patoka-White Eel 5,120,203 1, ,800 16, Patoka-White Driftwood 5,120,204 1, ,560 36, Patoka-White Flatrock-Haw 5,120, ,040 23, Patoka-White Upper East Fork White 5,120, ,040 26, Patoka-White Muscatatuck 5,120,207 1, ,520 25, Patoka-White Lower East Fork White 5,120,208 2,025 1,296,000 36, Patoka-White Patoka 5,120, ,760 24, Total 32,950 21,088, , Final Report September 2011 Page 44

51 Table 25. Total Phosphorus Exported at the Mouth of 8-Digit HUC Watersheds, Independent of Upstream Watershed Loading (USGS, 1997). Watershed Subwatershed HUC Sq. Miles Acres 1997 SPARROW Estimated Total Phosphorus Exported (Pounds) Total Phosphorus Average Pounds/Acre Wabash Upper Wabash 5120,101 1,589 1,016,960 5, Wabash Salamonie , Wabash Mississinewa ,880 2, Wabash Eel ,560 1, Wabash Middle Wabash-Deer ,280 3, Wabash Tippecanoe ,960 1,254,400 3, Wabash Wildcat ,320 3, Wabash Middle Wabash-Little Vermilion ,287 1,463,680 5, Wabash Vermilion , ,960 2, Wabash Sugar ,120 2, Wabash Middle Wabash-Busseron ,019 1,292,160 4, Wabash Embarras ,442 1,562,880 6, Wabash Lower Wabash , ,440 5, Wabash Little Wabash ,148 1,374,720 4, Wabash Skillet , ,360 1, Patoka-White Upper White ,754 1,762,560 5, Patoka-White Lower White ,675 1,072,000 4, Patoka-White Eel , ,800 1, Patoka-White Driftwood , ,560 3, Patoka-White Flatrock-Haw ,040 2, Patoka-White Upper East Fork White ,040 2, Patoka-White Muscatatuck , ,520 2, Patoka-White Lower East Fork White ,025 1,296,000 2, Patoka-White Patoka ,760 2, Total 32,950 21,088,000 79, Final Report September 2011 Page 45

52 Watershed Subwatershed HUC Table 26. SPARROW model estimates of agricultural NPS loading Agricultural Total Nitrogen Exported (Pounds) Total Nitrogen Maximum Standard Error Agricultural Total Phosphorus Exported (Pounds) Total Phosphorus Maximum Standard Error Wabash Upper Wabash ,815 1% 5,571 4% Wabash Salamonie ,098 5% % Wabash Mississinewa ,120 2% 2,293 9% Wabash Eel ,503 3% 1,503 14% Wabash Middle Wabash-Deer ,686 1% 3,326 6% Wabash Tippecanoe ,985 1% 3,338 7% Wabash Wildcat ,415 1% 3,413 6% Wabash Middle Wabash-Little Vermilion ,534 1% 5,474 5% Wabash Vermilion ,400 4% 2,711 15% Wabash Sugar ,428 2% 2,261 10% Wabash Middle Wabash-Busseron ,001 2% 4,594 7% Wabash Embarras ,343 1% 6,300 4% Wabash Lower Wabash ,613 2% 5,758 5% Wabash Little Wabash ,586 1% 4,357 5% Wabash Skillet ,542 5% 1,204 23% Patoka-White Upper White ,112 1% 5,130 6% Patoka-White Lower White ,288 1% 4,353 5% Patoka-White Eel ,672 4% 1,507 16% Patoka-White Driftwood ,043 2% 3,070 8% Patoka-White Flatrock-Haw ,024 2% 2,290 10% Patoka-White Upper East Fork White ,596 2% 2,627 8% Patoka-White Muscatatuck ,745 2% 2,606 9% Patoka-White Lower East Fork White ,266 1% 2,669 8% Patoka-White Patoka ,584 2% 2,252 10% Total 762,200 79,452 Final Report September 2011 Page 46

53 The Wabash River watershed has significant variability in characteristics that promote or reduce NPS nutrient loading. For example, the variation in land use and vegetative cover is illustrated in Figure 5. Additionally, variations in Indiana animal livestock density estimated using 2007 Indiana Agricultural receipts are presented for beef cattle in Figure 6 and dairy cattle in Figure 7. Figure 6. Number of Beef Cattle per Subwatershed within the Wabash-Patoka Watershed. Final Report September 2011 Page 47

54 Figure 7. Number of Dairy Animals per Subwatershed within the Wabash-Patoka watershed. Metrological conditions also vary significantly across the watershed, with the range of average annual precipitation for 1961 to 1990 generally increases traveling from north to south by approximately 10 inches per year (Oregon Climate Service, 1995). According to a USGS study the average runoff from increases in a general pattern from north to south ranging from 10 to 18 inches per year (USGS, 2008). The ability for an agricultural field to supply nutrient credits is dependent on these and other physical, chemical and biological characteristics. Final Report September 2011 Page 48

55 To illustrate the benefits of using a smaller scale model for estimating nutrient loads, a limited SWAT 2009 model was developed for the Driftwood ( ) and Tippecanoe ( ) watersheds. The model was set up for agricultural land use, forestry and urban runoff were estimated beginning with default values. Another constraint is that point source data was not readily available. The model was then calibrated and validated using hydrology records from USGS stations across a 13-year period ( ) with a 3-year equilibration period. The calibration was completed using the limited water quality data from STORET and USGS and a weight of evidence approach based on other regional model results. The SWAT model calibration and validation results are provided in Appendix D. The model is limited by lack of sufficient operational practice information (e.g., nutrient application rates and methods), lack of point source data and lack of water quality data for calibration. However, the model and resulting outputs are considered sufficient for this preliminary evaluation. Three scenarios for BMPs were created to test each BMP independently for the ability to reduce TN and TP. The three BMPs that were tested are as follows: 1) no-till residue management, 2) filter strips, and 3) cover crops. These three BMPs are not the only practices that can be used to generate nutrient load reductions. These BMPs were selected based on several factors: 1) the three practices provide a range of nutrient load reduction results, 2) the SWAT 2009 model construct manages operational BMPs easier than structural and/or bank stabilization BMPs, and 3) the input data for gully corrections and bank stabilization was not readily available. The NRCS 1 definition of each BMP is provided below (Indiana state office of the NRCS, electronic Field Office Tech Guide. Accessed April 12, 2011 at Filter strips: A strip or area of herbaceous vegetation that removes contaminants from overland flow. Cover Crops: Grasses, legumes, forbs, or other herbaceous plants established for seasonal cover and other conservation purposes. No-till Residue: Managing the amount, orientation and distribution of crop and other plant residues on the soil surface year-round, while growing crops in narrow slots, or tilled or residue free strips in soil previously untilled by full-width inversion implements. A fourth BMP, nutrient management, was considered. Individual field data was not available to determine the range of nutrients currently applied and the percent of fields within each range. Therefore a nutrient management scenario was not fully completed. Nutrient management is critical to successful systems of BMP implementation as discussed in the Final Report of the Lake Erie Millennium Network Synthesis Team (Lake Erie Millennium Network Synthesis Team, 2010). In this final report the findings indicate: And, There is no agronomic benefit to applying P fertilizer when STP levels reach 60 mg/kg Mehlich 3 P. Considering this benchmark, the occurrence of soil samples exceeding 60 mg/kg Mehlich 3 P was < 20% for 19 counties, 20 to 40% for 28 counties, and > 40% in 4 counties. Across the fifty counties, STP levels that are >60 mg/kg occur 30% of the time. (p. 10) Large additions of fertilizer or manure may change the soil mechanisms controlling P mobility by overwhelming a soil s ability to moderate P solubility resulting in a dominant P mineral phase, 1 Indiana State office of the NRCS, electronic Field Office Technical Guide. Accessed April 12, 2011 at Final Report September 2011 Page 49

56 from the amendment (fertilizer/manure), controlling P solubility. This is an important finding because it suggests that management of soils with a low to moderate STP may need to be considered differently than soils with high STP. It may be misleading to lump them together and attempt to predict runoff P at low to moderate STP levels using models developed where sites with very high STP are included, because the mechanisms controlling P solubility (transport risk) are different. (p. 8) The presence of high phosphorus applications (manure and/or fertilizer) occur 30 percent of the time in Ohio. A similar expectation might be made for Indiana. Nitrogen variability is demonstrated in a Purdue University Extension Agronomy Guide (Purdue University, 2005). Table 27. General Guidelines for Interpreting NO3-N Concentrations in Tile Drainage Water 1. (Purdue University, 2005) NO3 N Concentration (ppm) Interpretation < 5 Native grassland, CRP land, alfalfa, managed pastures 5-10 Row crop production on mineral soil without N fertilizer Row crop production with N applied at 45 lbs/acre below economically optimum N Rate 2 Row crop production with successful winter crop to trap N Row crop production with N applied at optimum N rate Soybeans > 20 Row crop production where: N applied exceeds crop need N applied not synchronized with crop need Environmental conditions limit crop production and N fertilizer use efficiency Environmental conditions favor greater than normal mineralization of soil organic matter 1 General guidelines for interpreting NO3-N concentrations in tile drainage water. The interpretation is derived from numerous studies conducted throughout the cornbelt and highlights land management strategies commonly found in association with a concentration measured in tile as the tile leaves the edge of field. 2 Economically optimum N rate is the rate that maximizes the return on investment in N fertilizer and therefore may be slightly lower than the N rate that maximizes crop yield. However, the rates of manure and fertilizer applications within a subwatershed are not well documented in public records. Information like phosphorus soil testing is not available on a field-by-field basis, this is considered confidential information by many programs including those in the Federal Farm Bill. Pragmatic estimation of the TN and TP reductions from nutrient management without ranges of current practices prevents this practice from being accurately estimated Filterstrip Treatment Efficiency Results Table 28 for the Driftwood subwatershed, and Table 29 for the Tippecanoe subwatershed, provide reduction results for implementing filter strips on the edges of row cropped lands. Final Report September 2011 Page 50

57 Table 28. Filterstrip Treatment Efficiency Results at the Subwatershed Scale, in Percent, Driftwood Watershed. Subwatershed Sediment (in short tons) TP (in lbs) TN (in lbs) 3 Baseline 63, , , Filterstrip 45, , , % change Baseline 125, , ,467, Filterstrip 101, , ,168, % change Baseline 49, , ,325, Filterstrip 39, , ,096, % change Table 29. Filterstrip Treatment Efficiency Results at the Subwatershed Scale, in Percent, Tippecanoe Watershed. Subwatershed Sediment (in short tons) TP (in lbs) TN (in lbs) 15 Baseline 8,413 30, ,300 Filterstrip 5,978 22, ,526 % change Baseline 8,533 39,186 2,220,708 Filterstrip 6,677 30,921 1,849,657 % change Baseline 46,706 83, ,397 Filterstrip 35,049 64, ,124 % change To further refine the analysis the Project Team focused on the range of variability as much as the average results. Using field runoff projections for filter strips a percent reduction available from the BMP was completed in three Driftwood subwatersheds. This watershed was selected to evaluate at the field level due it having larger variability in the subwatershed results. This estimate still includes some use of averaging. The SWAT model groups common parcels that have like soils and land use. These groupings are referred to as hydrologic resource units (HRU). The groupings range from hundreds to thousands of acres each. For agricultural row cropping the HRU categorization considers the soil characteristics, crop rotation make up, cropping tillage implements and timing of passes and the nutrient application rates and methods. The variability in these physical and cultural settings, in addition to climate and other factors described above, introduces a range of uncertainty in credit estimation at the watershed scale. To overcome this, implement passes selected to simulate no-till, mulch till and conventional tillage settings were based on the Indiana State Department of Agriculture conservation tillage data 2 were analyzed. SWAT model scenarios for the corn-soybean rotations assessed BMP treatment efficiencies. In the Driftwood subwatersheds only the highest residue rates (simulated by no-till) or the lowest residue rates (simulated by conventional moldboard plow implement passes) exist. Filter strip treatment efficiencies for nitrogen and phosphorus are listed in Table 30. NPS load reductions are in the range of 18 to 23 percent for nitrogen and 19 to 31 percent reduction of phosphorus. The 2 Indiana State Department of Agriculture (ISDA) Conservation Tillage Data by county- Available at: Final Report September 2011 Page 51

58 filter strip average nitrogen reduction is 20.7 percent with a standard deviation of 1.3 percent. For phosphorus the average reduction is 25.3 percent with a standard deviation of 3 percent. An area weighted mean in reduction is 19.6 and 22.7 percent respectively for nitrogen and phosphorus. Therefore, a conservative treatment efficiency value would be 20 percent for nitrogen and 22 percent for phosphorus. Table 30. Filterstrip Treatment Efficiency Results in Percent, Field Scale Results, Driftwood Watershed. Corn Soybean Tillage Practice Subwatershed HRU (acres) No-till Corn & Drilled Soybeans Conventional Tillage of Corn & Soybeans Conventional Till Corn & No-till Drilled Soybeans Nitrogen Percent Reduction Phosphorus Percent Reduction 1 1 (3,113) (1,637) (1039) (3011) (564) (3,630) (1,864) (1,316) (960) (1,844) (3,423) (15,664) (4,108) Cover Crop Treatment Efficiency Results Fall rye cover crop plantings were simulated across corn-soybean rotation row crops in subwatersheds 3, 8 and 23 in the Driftwood Watershed and 19, 30 and 34 in the Tippecanoe Watershed. In Table 31, the reduction results for the Driftwood subwatersheds are provided. Comparing these to the Tippecanoe subwatershed results, provided in Table 32, indicates a larger range of variability than that found in the filter strip investigation. The substantial variability in treatment efficiency can be partially explained by the similarity of residue management and cover cropping. The cover crop leaves residue in the field over winter periods. In fields with high residue management the reductions are usually lower. But the previous use of high residue tillage practices does not fully explain the variability. This is evident in the variability that exists within Tables 32 and 33. Specifically, the variability found in the conventional tillage corn no-till drill soybean category indicates the influence other factors have on the performance of cover cropping. Soil types, rate and timing of nutrients and year-to-year variability can all add to the variability in performance. Having optimum weather conditions allows the cover crop roots to keep nutrients closer to the surface. However, in different years wet weather or poor timing can lead to the nutrients leaching past the root zone prior to the uptake by the cover crop. Table 10 demonstrates an overall higher treatment efficiency gained by cover crops in the Tippecanoe subwatersheds. However, the variability still exists. Table 33 indicates the range of variability in the Driftwood subwatersheds is from 6 to 47 percent for nitrogen reductions. The average is 25.7 percent with a standard deviation of 13.3 percent. The area weighted mean is 18.7 percent reduction in nitrogen. For phosphorus reductions the range is from 8 to Final Report September 2011 Page 52

59 53 percent with an average of 22.6 percent and a standard deviation of 12.6 percent. The phosphorus area weighted mean is 28 percent. Table 34 indicates the range of variability in the Tippecanoe subwatersheds is from 12 to 53 percent for nitrogen reductions. The average is 32.3 percent with a standard deviation of 11.3 percent. The area weighted mean is 34.1 percent reduction in nitrogen. For phosphorus reductions the range is from 5 to 48 percent with an average of 23.6 percent and a standard deviation of 12.3 percent. The phosphorus area weighted mean is 26.6 percent. Selecting a common value to use across the Wabash River watershed (intended for this project only) requires being conservative when assessing these watershed tables. A conservative treatment efficiency to use for cover crops would be 12 percent for nitrogen and 10 percent for phosphorus in the Driftwood and 21 percent for nitrogen and 10 percent for phosphorus in the Tippecanoe. Therefore, a conservative overall Wabash River watershed treatment efficiency factor for cover crops can be 12 percent for nitrogen and 10 percent for phosphorus. This estimate is for extrapolation purposes only. It is evident that there are settings where the average results greatly exceed these low estimates. As such all 8-digit HUC watersheds can be expected to have a fraction of row cropped acres that will generate load reductions substantially greater than these estimates used for a conservative credit supply calculation. Table 31. Cover Crop Treatment Efficiency Results in Percent at the Subwatershed level, Driftwood Watershed. Subwatershed Sediment (in short tons) TP (in lbs) TN (in lbs) 3 Baseline 47, , , Cover crop 35, , , % change Baseline 53, , , Cover crop 41, , , % change Baseline 170, , ,812, Cover crop 165, , ,788, % change Table 32. Cover Crop Treatment Efficiency Results in Percent at the Subwatershed level, Tippecanoe Watershed. Subwatershed Sediment (in short tons) TP (in lbs) TN (in lbs) 19 Baseline 124, ,469 14,102,017 Covercrop 117, ,727 13,451,474 % change Baseline 13, ,721 2,379,021 Covercrop 10,644 83,631 1,401,823 % change Baseline 46,706 83, ,397 Covercrop 30,093 56, ,834 % change Final Report September 2011 Page 53

60 Table 33. Cover Crop Treatment Efficiency Results in Percent, Driftwood Subwatersheds. Corn Soybean Tillage Practice Subwatershed HRU (acres) No-till Corn & Drilled Soybeans Conventional Tillage of Corn & Soybeans Conventional Till Corn & No-till Drilled Soybeans Nitrogen Percent Reduction Phosphorus Percent Reduction 3 39 (2980) (1,363) (8,031) (607) (438) (643) (345) (3801) (3,439) (3,311) (578) (1,168) Table 34. Cover Crop Treatment Efficiency Results in Percent, Tippecanoe Subwatersheds. Corn Soybean Tillage Practice Subwatershed HRU (acres) No-till Corn & Drilled Soybeans Conventional Tillage of Corn & Soybeans Conventional Till Corn & No-till Drilled Soybeans Nitrogen Percent Reduction Phosphorus Percent Reduction (16,793) (10,574) (6,370) (17,525) (2,105) (4,463) (3,171) (6,546) (11,438) (4,674) (1,896) (2,459) (3,673) (5,850) (8,409) (23,325) (7,983) (4,516) No-till Residue Treatment Efficiency Results The subwatershed treatment efficiency results for no-till residue management in the Driftwood subwatersheds are given in Table 35 and Table 36. Many forms of residue management exist. The no-till, ridge-till and strip-till management systems all leave substantially more residue on the field than mulch till (typically chisel plowing once and minimal disc passes) or conventional moldboard plow. Moldboard plowing, a form of conventional tillage, typically leaves less than 5 percent residue on the field. Final Report September 2011 Page 54

61 Subwatershed Table 35. No-till Residue Management Treatment Efficiency Percents at the Subwatershed level, Driftwood Watershed. Sediment (in short tons) TP (in lbs) TN (in lbs) 3 Baseline 47, , , Residue 44, , , % change Baseline 53, , , Residue 47, , ,086, % change Baseline 170, , ,812, Residue 168, , ,958, % change Subwatershed Table 36. No-till Residue Management Treatment Efficiency Percents at the Subwatershed level, Tippecanoe Watershed. Sediment (in short tons) TP (in lbs) TN (in lbs) 19 Baseline 124, ,469 14,102,017 Covercrop 117, ,727 13,451,474 % change Baseline 13, ,721 2,379,021 Covercrop 10,644 83,631 1,401,823 % change Baseline 46,706 83, ,397 Covercrop 30,093 56, ,834 % change Residue management alters the hydrologic pathway of the precipitation. The residue on the field traps more rainfall and snowmelt on the field slightly increasing the infiltration. An increase in nitrogen loading is observed. High residue management also is shown to increase nitrogen loading in both watersheds. Nitrogen is a soluble parameter and increased infiltration associated with high residue tillage can increase the drainage tile loading of nitrates (Pennsylvania State University, 1996). The results of evaluating the range of phosphorus reductions at the HRU level in the Driftwood subwatersheds are provided in Table 37. The field delivery to streams ranges from 11 to 15 percent. The average reduction is 12.6 percent with a standard deviation of 1.2 percent. The area weighted mean is 10.3 percent. A conservative estimate for phosphorus reductions would be 10 percent. Final Report September 2011 Page 55

62 Table 37. No-till Residue Management Treatment Efficiency Results in Percent, Driftwood Watershed. Corn Soybean Tillage Practice Subwatershed HRU (acres) Conventional Tillage of Corn & Soybeans Conventional Till Corn & No-till Drilled Soybeans Nitrogen Reduction (Percent) Phosphorus Reduction (Percent) (8,031) (607) (438) (643) (345) (3801) (3,439) (3,311) (9,953) (1,845) (578) (1,168) Estimation of the Watershed Potential Reductions in Nitrogen and Phosphorus Agricultural credit supply can be estimated using the nutrient reduction estimates, number of row cropped acres and a typical loading rate. The 2008 National Land Cover Data Set was used to supply estimates of corn and soybean acres within each 8-Digit HUC. This data is provided in Table 38. The number of producers that are willing to participate further reduces the number of acres enrolled. A low end range of estimates is provided based on 10, 25 and 35 percent producer involvement. This limitation is applied for three main reasons: 1) not every producer is willing to participate in WQT programs, 2) not every acre in the watershed is capable of generating the load rates identified from this analysis, and 3) the lower estimate of participation better reflects a mature WQT program selection process targeting sites that provide higher credits per dollar. Using the SWAT model estimates of current per acre loading, confirmed by comparisons with regional studies (USGS, 1997), (Smith, 2008) the TN loading for row crop agriculture can be conservatively estimated at 30 lbs TN /acre (33.6 kg TN /ha) and 3 lbs TP/acre (3.4 kg TP/ha). Estimation of the volume of credits that can now be generated based on the pound per acre reduction potential. The three BMPs evaluated demonstrate that a BMP exists that will provide substantial watershed reduction potential by using only a fraction of the row cropped land within the watershed. The following estimates have not assigned a baseline condition or a trade ratio. Therefore, these are reported to provide the potential to supply agricultural NPS credits. The full trade ratio and baseline discussion will follow in later sections. The BMP evaluation summary of credits (i.e., pounds per acre without considering baselines or trade ratios) is as follows: Filter strips: Approximately 240 TN credits and 26.4 TP credits. A filter strip acre serves 40 acres of row crop (SWAT 2009 default application). The nitrogen reduction credit is based on 30 lbs TN/acre of runoff at 20 percent treatment efficiency serving 40 acres. The phosphorus reduction Final Report September 2011 Page 56

63 credit is based on 3 lbs TP/acre runoff loading treated at a 22 percent efficiency serving 40 acres. Table 38. Annual Nutrient Load Reduction Potential, Wabash River Watershed 8-digit HUC Subwatersheds. (Assuming a 10 and 25 Percent Participation of Agricultural Row Cropped Acres, 20 Percent Reductions, and 40 lbs TN/acre and 3 lbs TP/acre Loading Rates.) Wabash River Watershed 8-digit HUCs Corn 1 Soybeans 1 Total Corn and Bean Acres Assuming 20% Removal and 10% Producer Participation TN Reduction TP Reduction Assuming 20% Removal and 25% Producer Participation TN Reduction TP Reduction Patoka 60,230 48, ,331 64,999 6, ,497 16,250 Eel 157, , , ,623 19, ,559 49,656 Mississinewa 145, , , ,825 19, ,562 49,956 Tippecanoe 507, , , ,745 49,075 1,226, ,686 Middle Wabash-Little Vermilion 495, , , ,918 54,192 1,354, ,480 Sugar 181, , , ,965 20, ,912 52,491 Embarras 507, ,372 1,000, ,405 60,041 1,501, ,101 Upper East Fork White 119, , , ,896 15, ,241 38,224 Lower White 158, , , ,412 19, ,530 49,853 Middle Wabash-Deer 183, , , ,430 17, ,575 44,858 Little Wabash 287, , , ,429 39, ,072 98,607 Upper White 347, , , ,629 44,963 1,124, ,407 Wildcat 197, , , ,548 21, ,870 54,887 Lower East Fork White 85,544 81, , ,310 10, ,776 25,078 Eel 144, , , ,471 17, ,677 43,868 Vermilion 380, , , ,603 41,860 1,046, ,651 Upper Wabash 272, , , ,105 35, ,762 88,776 Muscatatuck 79, , , ,191 11, ,978 28,798 Flatrock-Haw 133, , ,671 64,999 6, ,497 16,250 Middle Wabash- Busseron 333, , , ,623 19, ,559 49,656 Skillet 101, , , ,825 19, ,562 49,956 Lower Wabash 232, , , ,745 49,075 1,226, ,686 Salamonie 95, , , ,918 54,192 1,354, ,480 Driftwood 196, , , ,965 20, ,912 52, National Land Cover Data Set Cover Crops: Approximately 3.6 TN credits and 0.3 TP credits. The nitrogen reduction credit is based on 30 lbs of TN/acre runoff treated at 12 percent efficiency. The phosphorus reduction credit is based on 3 lb TP/acre treated at 10 percent efficiency. No-till Residue: Approximately 0.3 TP credits. One acre of moldboard plow conversion to no-till residue management practices increases nitrogen loading to the watershed. However, in phosphorus limited freshwaters this practice provides similar results to that of implementing Final Report September 2011 Page 57

64 cover cropping. The phosphorus reduction credit is based on 3 lb TP/acre treated at 10 percent efficiency. Potential agricultural credit supply is provided in Table 38. The table is based on availability of BMPs that supply a 20 percent reduction of the 30 lbs TN/acre rate and a 20 percent reduction of the 3 lbs TP/acre. The acreage considered in the estimate of supply is limited to 10 or 25 percent of corn and soybean acres in the watershed. In summary, the amount of potential to reduce agricultural row crop nonpoint source nutrient loading is ample. The forecasted potential for nutrient reductions is based on very conservative estimates. The conservative assumptions include: The ability to generate 20 percent reductions across the watershed is predicated by: 1. The BMPs options available are many times greater than those assessed 2. The variability within each BMP assessed indicates that even the poorer performing BMPs have substantial opportunities to outperform the conservative estimates if placed in the right setting 3. Field scale calculations, based on site-specific information, combined with a targeting framework in the WQT program will prioritize site selection Agricultural land management other than corn and soybean acres can be used to generate credits Estimate of nutrient runoff loading rates agrees well with regional research Low producer participation levels used to generate cumulative numbers The expected availability of load reduction to generate WQT credits is higher than Table 14 indicates. 3.9 Potential stakeholder participation Water Quality Trading Feedback: Regulated Point Sources On September 21, 2010, CTIC presented information on the project to approximately 60 participants at the Indiana Water Environment Seminar. With assistance from the Project Team, CTIC developed a survey to assess those wastewater treatment plant representatives knowledge, perceptions and opinions on water quality trading. On Jun 14, 2011, CTIC sent s to 24 of those present at the seminar, asking each to respond to the survey electronically. On July 11, 2011, CTIC contacted these same 24 waste water treatment plant representatives with a reminder to complete the online survey. Twenty-four percent of those surveyed responded. Appendix E contains the survey questions and responses. Duke Energy submitted the following statement regarding water quality trading: Duke Energy endorses the concept of water quality trading. This tool can be an important option for nutrient standard compliance in the future. Municipalities, utilities, other dischargers, and agriculture could potentially benefit greatly with this program by keeping costs low and improving water quality and the environment. Duke Final Report September 2011 Page 58

65 Energy will continue to be engaged by offering expertise and monitoring the progress of the issues associated with water quality trading Water Quality Trading Feedback: Farmer Focus Group On March 25, 2011, CTIC held a farmer focus group meeting in Greenfield, IN which lies in the Driftwood River watershed. Fifteen farmers, agribusiness, agriculture media and watershed group representatives attended. Jim Klang of Kieser and Associates, LLC presented information on water quality trading and fielded questions from the group. Questions asked at the focus group and the associated answers are presented below. Q: How will water quality be measured if farm ground is split between two watershed boundaries? A: This will be specified by those developing the program. Q: Many farmers have at least some best management practices in place. How can the farmer benefit from those practices previously implemented? A: This will be specified by those developing the program. Q: Would pollution control agencies require/drive a program such as this? What is the Indiana Department of Environmental Management s incentive to support a program like this? A: The federal Environmental Protection encourages states to develop nutrient standards for water quality. Some states have approved nutrient management standards, such as Wisconsin, which has approved nutrient standard for lakes. The Indiana Department of Environmental Management has not developed nutrient management standards, nor has the Illinois Environmental Protection Agency. Q: How will urban pollutants be addressed? At what watershed scale (hydrologic unit code) can these programs work? A: This will be specified by those developing the program. The program should be built to avoid hot spots. Q: If farmers are upstream of the regulated facility, can they sell credits? A: Most likely. This will be specified by those developing the program. Q: Is there a program such as this in place in Indiana? A: Not at this time. Q: What are the regulated facilities expecting with regard to a program like this? Are they ready to talk about a program like this in real terms? A: Many regulatory agencies are watching how existing programs function, because they expect nutrient standards will be imposed in the future. Final Report September 2011 Page 59

66 Questions posed to the audience: Q: Are you comfortable with the environmental protection aspect of a program such as this? A: It depends on the accuracy of the models that measure water quality results. Q: Who are you comfortable with coming onto your land to check best management practice performance? A: Crop consultants or USDA Natural Resources Conservation Service representatives. Q: Would you be comfortable with a requirement that directs management practice inspection? A: If the inspection process is transparent. For example, the farmer needs to know exactly what will be assessed. Farmer costs related to inspections should be known. 4. Putting It All Together Market Analysis and Trading Considerations This section synthesizes the information presented in Section Two and provides a summary of findings and recommendations related to overall WQT market feasibility in the Wabash River watershed. 4.1 Pollutant Loads The timing of upgrades, design capacity of the facility, and treatment technologies selected are important factors that determine the economic and performance capability for point source entities to become credit generators. Therefore, while there is a strong potential for point source to point source trading to be effective in the Wabash River watershed a more detailed assessment of options based on the newly required effluent limits will be necessary. 4.2 Regulatory Drivers As discussed in Section , Indiana s current water quality standards and nutrient permit limits provide a regulatory incentive for trading based only on TMDLs developed to protect the narrative water quality standards. However, the regulatory forecast for numeric nutrient criteria indicates that nutrient criteria will emerge within the next decade. These effluent limits will likely affect permit limits once these new standards are approved. Knowing these regulatory changes are likely to occur within the next 3-5 years, stakeholders do have a strong reason to consider water quality trading now as a potential tool for achieving permit limits in the future. 4.3 Trade Ratios A load reduction from one source, at a remote location, must provide equal or greater environmental protection for the water resource. A watershed understanding aids the WQT managers when developing a program. For example, an understanding of the natural nutrient attenuation that occurs on the land and in the streams, lakes and rivers prior to reaching the protected water resource allows appropriate factors to be considered. A trade ratio refers to an explicit factor that is applied to either or both the Final Report September 2011 Page 60

67 buyer and seller. According to the Water Quality Trading Toolkit for Permit Writers development of trade ratios should consider the following elements when being developed: Location Factors and/or Delivery Ratios: address the differences in attenuation of nutrients when discharges occur at spatially different points. A location factor addresses the attenuation of the nutrients being traded between the buyer s or seller s discharge point and the downstream water resource being protected. A delivery ratio addresses the attenuation that occurs between the buyer s and seller s discharge points when the seller is upstream. Equivalency Factors: address the differences in environmental stress that slight differences in discharged pollutant forms or interaction between multiple stressors have on a water resource. For instance, in nutrient trading a buyer may discharge a higher level of bioavailable phosphorus forms than the agricultural nonpoint source runoff discharge being offered as an offset. Other programs trading to relieve impairments with multiple stressors develop ratios for each stressor based on how the parameters interact within that specific watershed setting. Uncertainty Factors: address the introduced variability, errors and lack of understanding that WQT programs work with on a daily basis. Uncertainty occurs from many sources. A few main components that introduce uncertainty into WQT transactions are: 1) from analytical errors when collecting and testing water quality samples, 2) from stochastic variability in discharger loading, climatic events and nonpoint source settings, 3) credit estimation tools, such as models, that introduce simplifications of the real world, and 4) errors in watershed understandings within the current day best available science. Policy Factors: address the socio-political elements watershed and WQT decisions makers implement to address equity issues, incentivize good behavior, advance watershed goals or provide disincentives for less desirable practices. Addressing these components can be done by assigning one or more factors to the buyer and others to the seller or as a block in one trade ratio. The Ohio EPA Water Quality Trading Rules 3 require point source to point source trading to use a ratio where one pound of pollutant reduction equals one pound of water quality credit for that pollutant. For nonpoint source generated credits for point source discharges the trade ratio must be: 1) When there is not an approved TMDL, be calculated using a trading ratio where two pounds of pollutant reduction equals one pound of water quality credit for that pollutant; or 2) When there is an approved TMDL, be calculated using a trading ratio where three pounds of pollutant reduction equals one pound of water quality credit for that pollutant. The rules also allow the director to consider or impose other alternative trade ratios based on watershed, habitat restoration or other considerations. The draft Water Quality Trading Rules in Minnesota 4 quantified phosphorus based risk trade ratios. The Minnesota Pollution Control Agency (MPCA) defined the risk trade ratio as a factor that addresses the total of all risks associated with trading. The point source to point source trades when dealing with an upstream seller are to use a one credit sold to 1.1 credit purchased ratio. For downstream sellers the 3 Ohio EPA Division of Surface Water OAC Chapter Water Quality Trading; available at 4 Minnesota Pollution Control Agency Draft Water Quality Trading Rules and Statement of Needs and Reasonableness; available at: Final Report September 2011 Page 61

68 ratio increases requiring 1.4 credits to be purchased. For nonpoint source generated credits the risk trade ratio is one credit sold to 2.5 credits purchased. In addition, every trade shall include an environmental factor (a policy factor) where the buyer must purchase an additional ten percent of credits that are not available for use. These states are examples of regulatory authorities that have evaluated their watershed settings and the potential beneficial use of WQT for achieving the water quality protection goals. The state then created the more simplified method of a combined trade ratio to roll out watershed implementation when using WQT Location Factors and Delivery Ratios The location factors and delivery ratios are watershed specific attenuation coefficients. The need to use a location factor and/or delivery ratio also depends on the methods used to define the environmental credit value. Upland delivery ratios may be necessary if the load reduction estimation tools do not predict an edge of field result or the fields being credited are not adjacent to waterbodies. Channel attenuation addressed by delivery ratios (assimilation losses between upstream credit generators and downstream buyers) may not be required if the WQT program addresses losses via a location factor for both the buyer and seller. Because an actual WQT framework is not set the location factor method will be used in the feasibility evaluation. This section explains the method used to determine the WQT location factor for each watershed. The location factor is determined by applying values of nutrient loading predicted by the USGS SPARROW model results (USGS, 2009). The model estimates the fraction of incremental nutrient load delivered to the Gulf from upstream watersheds. Incremental loading is defined as the amount of Nitrogen/Phosphorus generated in an individual watershed that arrives at the Gulf of Mexico. This percentage can be used to determine location factors for each watershed and can assist programs attempting to protect both the Gulf of Mexico and upstream waters. As water travels downstream, interaction with the surroundings causes nutrients to be naturally removed from the stream. Water entering streams near the Gulf has less time to interact with its surroundings than water entering farther upstream. In most cases, the percentage of nutrients that reach the Gulf is higher for water that enters streams near the Gulf compared to water that enters upstream watersheds. However, the SPARROW model sometimes predicts that a downstream watershed has a lower percentage of delivered incremental loads than an upstream watershed. This could be caused by error in the model or individual watershed characteristics, including lakes, wetland impoundments, and poor hydrologic connectivity in the local watershed. Such characteristics allow more nutrients to be assimilated than would otherwise occur in the stream channel. In order to conservatively represent the most restrictive watershed, all of the best-fit lines were adjusted down to include the lowest incremental watershed results. This reduced the risk of creating nutrient hotspots by applying a more conservative location factor that did not overestimate natural attenuation in any watershed. According to Bill Franz 5, U.S. EPA Region V, project manager for the SPARROW modeling, U.S. EPA contracts with USGS a 12-digit HUC output from SPARROW will be available in This will allow finer resolution to be applied using this technique. Figures 1-17 graphically show how the SPARROW data was adjusted to determine the location factor. The percentage values provided by the SPARROW model are fitted with a trend line (y1) that best 5 Bill Franz, personal communication, May 25, 2010, Chicago, IL Final Report September 2011 Page 62

69 represents the linear nature of the data. If a point is below the best fit line, it is being credited with more natural nutrient attenuation than is actually predicted in that watershed. To correct this characteristic for the mainstem of the watershed and its larger tributaries, the trend line is adjusted down to the lowest point to assure an individual watershed s percentage is not over-estimated. The resulting best fit line (y2) provides a conservative estimate of a watershed s nutrient load delivery. Within an 8-digit HUC the attenuation factor in subwatersheds is already addressed by the channel process based estimates of SWAT. Prior to the SPARROW update a linear interpolation of the current SPARROW model results could be used. Or to calculate location factors using a step wise process, factors for the subwatersheds can be determined by running the watershed assessment model with multiple scenarios. The scenarios are designed to evaluate the change in loading at the mouth given a change in loading at each subwatershed (entered one at a time). Review of the SPARROW data indicates the maximum phosphorus attenuation rate between Wabash 8-digit HUC watersheds is 49 percent; total nitrogen maximum is 44 percent. Both of these maximums occur when the Salamonie River loading travels to the confluence of the Wabash River with the Ohio River. More commonly, headwaters to confluence based loading losses are approximately 20 percent. Based on these tables, for the purposes of evaluating feasibility the near-field trading will apply a 10 percent location factor and far-field transactions are assigned a 25 percent factor. This reflects the midrange of possible values for each situation Equivalency Factors Equivalency factors address a generated credit reduction providing the same level of stressor impact relief on the water body being protected. This includes addressing potential differences in nutrient bioavailability. This feasibility study provides an in depth review of the potential for differences between buyer and seller s discharged nutrients and the characteristics affecting nutrient bioavailability in each. Phosphorus: While it is possible for a WQT program to require statistical sampling of discharges to determine their bioavailability this may not be cost effective or necessary. Instead an understanding of the forms of phosphorus discharged and the percentages within the total phosphorus discharged can inform the bioavailable estimate. NPS runoff may be assessed using this same breakdown. Addressing fresh water eutrophication in the Midwest is a concern of many states. The Minnesota legislature pursued a desire to better understand the stressors that lead to eutrophication by commissioning a report entitled Detailed Assessment of Phosphorus Sources to Minnesota Watersheds. The oversight task was assigned to the Minnesota Pollution Control Agency. This study contained a technical memorandum that defines the expected variability of phosphorus bioavailability found in different sources (Barr, 2004). The memorandum results are summarized in Table 1 below. Bioavailability equivalence can be determined by dividing the bioavailability of the two sources to achieve a ratio. The denominator of the ratio is the equivalence fraction of the buyer discharge. This ratio can then be used to provide the WQT program with an equivalence discount factor. Table 1 indicates the most likely estimate of phosphorus bioavailability of a source, as well as, the range of expected variability. Combining the finding in Table 39 creates a list of probable equivalency factors. The equivalency factors for various trades are provided below: Final Report September 2011 Page 63

70 Point source to point source domestic WWTPs trades is 85.5 / 85.5 or 1.0 with the likely range of variability in these transactions is expected to be plus or minus 10 percent Equivalence factor in point point domestic WWTP selling to industrial WWTP trades is 85.5 / 88 or 0.97 with a likely range of variability of approximately 15 percent The equivalence factor in domestic point source and agricultural nonpoint sources (fertilizer based) trades are 58 / 85.5 or 0.68 with a range of expected variability of approximately 20 percent The equivalence factor in industrial point sources and agricultural nonpoint sources (fertile based) is 58 / 88 or 0.66 with a likely range of variability of approximately 25 percent Final Report September 2011 Page 64

71 Table 39. Estimates of phosphorus bioavialability fractions for specific source categories. Final Report September 2011 Page 65