GEORGIA DOT RESEARCH PROJECT 14-03 FINAL REPORT ALTERNATIVE STORMWATER COMPLIANCE APPROACHES ON GDOT RIGHT-OF-WAY OFFICE OF PERFORMANCE-BASED MANAGEMENT AND RESEARCH 600 WEST PEACHTREE STREET NW ATLANTA, GA 30308 TECHNICAL REPORT DOCUMENTATION PAGE 1. Report No. 2. Government Accession No. 3. Recipient's Catalog No. FHWA-GA-19-1403 N/A N/A 4. Title and Subtitle: ALTERNATIVE STORMWATER COMPLIANCE APPROACHES ON GDOT RIGHT-OF-WAY 5. Report Date June 2019 6. Performing Organization Code N/A 7. Author(s): Susan E. Burns, Ph.D., P.E. (https://orcid.org/0000-0002-3029-7314) 9. Performing Organization Name and Address Georgia Institute of Technology School of Civil and Environmental Engineering 790 Atlantic Drive Atlanta GA 30332-0355 Phone: (404)894-2285 Email: susan.burns@ce.gatech.edu 12. Sponsoring Agency Name and Address: Georgia Department of Transportation Office of Performance-based Management and Research 600 West Peachtree St. NW Atlanta, GA 30308 8. Performing Organization Report No. 14-03 10. Work Unit No. N/A 11. Contract or Grant No. PI#0013113 13. Type of Report and Period Covered Final: December 2014-June 2019 14. Sponsoring Agency Code N/A 15. Supplementary Notes Prepared in cooperation with the U.S. Department of Transportation, Federal Highway Administration. 16. Abstract The objectives of this study were to evaluate the field performance of stormwater best management practices (BMPs) and perform a statistical analysis on the impact of dimensions of sand filter BMPs on contaminant treatment. Some of the pertinent conclusions from this study are: (1) the combination dry swale/sand filter successfully treated turbidity, total suspended solids, total dissolved solids, some nutrients, and some heavy metals (zinc); (2) the bioretention basin showed removal percentages of > ~60% for solids, nutrients, and metals; (3) statistical analysis of the performance of sand filters demonstrated that sand filters are highly effective at solids, turbidity, some metal and total phosphorus removal, but less effective at removing dissolved constituents; and (4) it is recommended that the GDOT BMP selection flowchart include BMP longevity as a factor, as well as differentiate between short/intermediate term and long term maintenance burden. 17. Key Words Bioretention basin; BMP; Dissolved solids; Dry swale; Sand filter; Nutrient removal; Suspended solids 18. Distribution Statement No restrictions. This document is available through the National Technical Information Service, Springfield, VA 22161. 19. Security Classification 20. Security Classification 21. No. of Pages 22. Price (of this report): (of this page) Unclassified Unclassified 122 Free Form DOT 1700.7 (8-72) Reproduction of completed page authorized GDOT Research Project 14-03 Final Report ALTERNATIVE STORMWATER COMPLIANCE APPROACHES ON GDOT RIGHT-OF-WAY By Susan E. Burns Professor Georgia Tech Research Corporation Contract with Georgia Department of Transportation In cooperation with U.S. Department of Transportation Federal Highway Administration June 2019 The contents of this report reflect the views of the authors, who are responsible for the facts and accuracy of the data presented herein. The contents do not necessarily reflect the official views of the Georgia Department of Transportation or the Federal Highway Administration. This report does not constitute a standard, specification, or regulation. ii Symbol in ft yd mi in2 ft2 yd2 ac mi2 fl oz gal ft3 yd3 oz lb T oF fc fl lbf lbf/in2 SI* (MODERN METRIC) CONVERSION FACTORS APPROXIMATE CONVERSIONS TO SI UNITS When You Know Multiply By To Find inches feet yards miles square inches square feet square yard acres square miles fluid ounces gallons cubic feet cubic yards LENGTH 25.4 millimeters 0.305 meters 0.914 meters 1.61 kilometers AREA 645.2 square millimeters 0.093 square meters 0.836 square meters 0.405 hectares 2.59 square kilometers VOLUME 29.57 milliliters 3.785 liters 0.028 cubic meters 0.765 cubic meters NOTE: volumes greater than 1000 L shall be shown in m3 ounces pounds short tons (2000 lb) Fahrenheit MASS 28.35 grams 0.454 kilograms 0.907 megagrams (or "metric ton") TEMPERATURE (exact degrees) 5 (F-32)/9 Celsius or (F-32)/1.8 ILLUMINATION foot-candles foot-Lamberts 10.76 3.426 lux candela/m2 FORCE and PRESSURE or STRESS poundforce 4.45 newtons poundforce per square inch 6.89 kilopascals Symbol mm m m km mm2 m2 m2 ha km2 mL L m3 m3 g kg Mg (or "t") oC lx cd/m2 N kPa Symbol mm m m km mm2 m2 m2 ha km2 mL L m3 m3 g kg Mg (or "t") oC lx cd/m2 N kPa APPROXIMATE CONVERSIONS FROM SI UNITS When You Know Multiply By To Find millimeters meters meters kilometers LENGTH 0.039 3.28 1.09 0.621 inches feet yards miles square millimeters square meters square meters hectares square kilometers AREA 0.0016 10.764 1.195 2.47 0.386 square inches square feet square yards acres square miles milliliters liters cubic meters cubic meters VOLUME 0.034 0.264 35.314 1.307 fluid ounces gallons cubic feet cubic yards MASS grams 0.035 ounces kilograms 2.202 pounds megagrams (or "metric ton") 1.103 short tons (2000 lb) TEMPERATURE (exact degrees) Celsius 1.8C+32 Fahrenheit ILLUMINATION lux candela/m2 0.0929 0.2919 foot-candles foot-Lamberts newtons kilopascals FORCE and PRESSURE or STRESS 0.225 poundforce 0.145 poundforce per square inch Symbol in ft yd mi in2 ft2 yd2 ac mi2 fl oz gal ft3 yd3 oz lb T oF fc fl lbf lbf/in2 *SI is the symbol for the International System of Units. Appropriate rounding should be made to comply with Section 4 of ASTM E380. (Revise*d SMIaricshth20e0s3y)mbol for the International System of Units. Appropriate rounding should be made to comply with Section 4 of ASTM E380. (Revised March 2003) iii TABLE OF CONTENTS EXECUTIVE SUMMARY .............................................................................................. 1 CHAPTER 1. INTRODUCTION .................................................................................... 3 CHAPTER 2. LITERATURE REVIEW ........................................................................ 6 COMMON BMPS IN USE BY GEORGIA DOT .................................................... 7 Bioretention Basins ............................................................................................... 7 Enhanced Swales................................................................................................... 9 Filter Strips.......................................................................................................... 10 VEGETATED FILTER STRIP PERFORMANCE AS A FUNCTION OF SLOPE ....................................................................................................................... 11 MAINTENANCE OF STORMWATER BMPS .................................................... 13 CHAPTER 3. FIELD PERFORMANCE OF THREE GDOT BEST MANAGEMENT PRACTICE (BMP) SITES.............................................................. 17 BMP FIELD SAMPLE AND TESTING, MATERIALS, AND METHODS ...... 17 BMP: DRY SWALE WITH SAND FILTER IN FORSYTH COUNTY, GA ..... 18 Experimental Results.......................................................................................... 23 Summary Removal Efficiency: Dry swale/sand filter Forsyth County, GA.. 39 BMP: BIORETENTION BASIN IN BARTOW COUTNY.................................. 41 Experimental Results.......................................................................................... 42 Summary Removal Efficiency: Bioretention basin, Bartow County, GA ..... 58 BMP IMPACT ON SURFACE WATER AND GROUND WATER SAMPLES IN THE COASTAL PLAIN ................................................................. 59 Alternative Filter Strips in Coastal Plain BMPs .............................................. 61 SUMMARY: BMP FIELD TESTS ......................................................................... 66 CHAPTER 4. STATISTICAL ANALYSIS OF THE EFFICIENCY OF SAND FILTERS FOR TYPICAL POLLUTANTS ................................................................. 68 FIELD STUDY SITE AND DATA DESCRIPTION ............................................. 70 TYPICAL HIGHWAY RUNOFF CHARACTERISTICS ................................... 71 SAND FILTER STATISTICAL DESIGN PARAMETERS ................................ 88 CHAPTER 5. RECOMMENDATION FOR GUIDANCE ON SELECTION OF STORMWATER BMPS................................................................................................. 93 CHAPTER 6. CONCLUSIONS AND RECOMMENDATIONS ............................... 96 iv APPENDIX A: DETAILED SAMPLE CONCENTRATION DATA FOR TESTED STORMWATER BMPS................................................................................................. 97 APPENDIX B: HYDRAULIC CONDUCTIVITY OF GDOT TOPSOIL MIX ..... 102 DETERMINATION OF HYDRAULIC CONDIUCTIVITY OF GDOT TOPSIOL MIX........................................................................................................ 103 Materials and Methods..................................................................................... 104 Results ................................................................................................................ 105 ACKNOWLEDGEMENTS ......................................................................................... 109 REFERENCES.............................................................................................................. 110 v LIST OF FIGURES Figure 1. Diagram. Mass balance for flow, infiltration, and deposition in a vegetated filter strip. (Figure from Munoz and Carpena, abe.ufl.edu/carpena/vfsmod/FOCUS/VFSMOD_Pmassbalance1.ppt). .......................... 12 Figure 2. Map. Topographic map of the Cumming GA / BMP location (USGS)................. 19 Figure 3. Diagram. Inlet and outlets labeled for the dry swale pond with sand filter. .......... 21 Figure 4. Diagram. Dry swale pond with sand filter showing the rock filter dam (circled in green). The filter ring (circled in red) was removed after construction was completed. 21 Figure 5. Diagram. Cross-section of the dry swale pond with sand filter. ............................ 22 Figure 6. Photos. Test site (dry swale with a sand filter) at Chattahoochee River................ 22 Figure 7. Photos. Test set-up for stormwater monitoring at BMP inflow and outflow......... 23 Figure 8. Graph. First flush turbidity at dry swale with a sand filter. ................................... 24 Figure 9. Graph. First flush total suspended solids at dry swale with a sand filter. .............. 25 Figure 10. Graph. First flush total dissolved solids at dry swale with a sand filter. ............. 25 Figure 11. Graph. First flush total copper at dry swale with a sand filter. ............................ 26 Figure 12. Graph. First flush total lead at dry swale with a sand filter. ................................ 26 Figure 13. Graph. First flush total zinc at dry swale with a sand filter. ................................ 27 Figure 14. Graph. First flush nitrite at dry swale with a sand filter. ..................................... 27 Figure 15. Graph. First flush nitrate at dry swale with a sand filter...................................... 28 Figure 16. Graph. First flush total nitrogen at dry swale with a sand filter. ......................... 28 Figure 17. Graph. First flush total phosphorus at dry swale with a sand filter. .................... 29 Figure 18. Graph. In-situ pH values during the stormwater runoff (dry swale/sand filter). . 30 Figure 19. Graph. In-situ temperature values during the stormwater runoff (dry swale/sand filter). ............................................................................................................................... 30 Figure 20. Graph. In-situ conductivity values during the stormwater runoff (dry swale/sand filter). ............................................................................................................................... 31 Figure 21. Graph. Turbidity test results for inlet samples and outlet samples (dry swale/sand filter). ............................................................................................................................... 32 Figure 22. Graph. Total suspended solid test results for inlet samples and outlet samples (dry swale/sand filter). ............................................................................................................ 32 Figure 23. Graph. Total dissolved solid test results for inlet samples and outlet samples (dry swale/sand filter). ............................................................................................................ 33 Figure 24. Photos. Microfiber filter results for total suspended solid tests (dry swale/sand filter). ............................................................................................................................... 33 Figure 25. Graph. Total copper at inlet samples and outlet samples (dry swale/sand filter). 34 Figure 26. Graph. Total lead at inlet samples and outlet samples (dry swale/sand filter)..... 35 Figure 27. Graph. Total zinc at inlet samples and outlet samples (dry swale/sand filter)..... 35 Figure 28. Graph. Nitrite concentration at inlet samples and outlet samples........................ 36 Figure 29. Graph. Nitrate concentration at inlet samples and outlet samples. ...................... 37 Figure 30. Graph. Total nitrogen concentration at inlet samples and outlet samples. .......... 37 Figure 31. Graph. Total phosphorus concentration at inlet samples and outlet samples. ..... 38 Figure 32. Photo. Overtopping of stormwater at the BMP outlet drainage control structure. ......................................................................................................................................... 39 Figure 33. Map. Topographic Map of the Cartersville BMP area (USGS)........................... 41 vi Figure 34. Diagram. Sample Locations at the Raccoon Creek Bioretention Basin. ............. 43 Figure 35. Diagram. Plan view of drop inlet with dimensions.............................................. 43 Figure 36. Photos. Bioretention basin test site at SR 113 and Racoon Creek, Bartow County . ........................................................................................................................................ 44 Figure 37. Graph. First flush turbidity concentration for the bioretention basin, Bartow County GA. ..................................................................................................................... 45 Figure 38. Graph. First flush total suspended solids concentration for the bioretention basin, Bartow County GA.......................................................................................................... 46 Figure 39. Graph. First flush total dissolved solids concentration for the bioretention basin, Bartow County GA.......................................................................................................... 46 Figure 40. Graph. First flush total copper concentration for the bioretention basin, Bartow County GA. ..................................................................................................................... 47 Figure 41. Graph. First flush total lead concentration for the bioretention basin, Bartow County GA. ..................................................................................................................... 47 Figure 42. Graph. First flush total zinc concentration for the bioretention basin, Bartow County GA. ..................................................................................................................... 48 Figure 43. Graph. First flush nitrite concentration for the bioretention basin, Bartow County GA. .................................................................................................................................. 48 Figure 44. Graph. First flush nitrate concentration for the bioretention basin, Bartow County GA. .................................................................................................................................. 49 Figure 45. Graph. First flush total nitrogen concentration for the bioretention basin, Bartow County GA. ..................................................................................................................... 49 Figure 46. Graph. First flush total phosphorus concentration for the bioretention basin, Bartow County GA.......................................................................................................... 50 Figure 47. Graph. In-situ pH values during the stormwater runoff for the bioretention basin, Bartow County, GA......................................................................................................... 50 Figure 48. Graph. In-situ temperature values during the stormwater runoff for the bioretention basin, Bartow County, GA. ......................................................................... 51 Figure 49. Graph. In-situ conductivity values during the stormwater runoff for the bioretention basin, Bartow County, GA. ......................................................................... 51 Figure 50. Graph. Turbidity test results for inlet samples and outlet samples for bioretention basin, Bartow County, GA. ............................................................................................. 52 Figure 51. Graph. Total suspended solid test results for inlet samples and outlet samples for bioretention basin, Bartow County, GA. ......................................................................... 52 Figure 52. Graph. Total dissolved solid test results for inlet samples and outlet samples for bioretention basin, Bartow County, GA. ......................................................................... 53 Figure 53. Graph. Total copper at inlet samples and outlet samples for the bioretention basin, Bartow County, GA. ............................................................................................. 54 Figure 54. Graph. Total lead at inlet samples and outlet samples for the bioretention basin, Bartow County, GA......................................................................................................... 54 Figure 55. Graph. Total zinc at inlet samples and outlet samples for the bioretention basin, Bartow County, GA......................................................................................................... 55 Figure 56. Graph. Nitrite concentration at inlet samples and outlet samples for bioretention basin, Bartow County, GA. ............................................................................................. 56 Figure 57. Graph. Nitrate concentration at inlet samples and outlet samples for bioretention basin, Bartow County, GA. ............................................................................................. 56 vii Figure 58. Graph. Total nitrogen concentration at inlet samples and outlet samples for bioretention basin, Bartow County, GA. ......................................................................... 57 Figure 59. Graph. Total phosphorus concentration at inlet samples and outlet samples for bioretention basin, Bartow County, GA. ......................................................................... 57 Figure 60. Map. Location of surface water samples taken throughout Skidaway Island, and location of wells for groundwater sampling (at SERF site). ........................................... 60 Figure 61. Photos. Sites that were surveyed and the results obtained from surveying: A) MW-12, B) MW-13, and C) PL-09. ................................................................................ 61 Figure 62. Map-Charts. Values of TIN shown in white, representing percentage of NO3(orange), NH4+ (blue), and NO2- (green). ..................................................................... 62 Figure 63. Map-Charts. Nitrate concentrations in surface water around Skidaway Island. .. 63 Figure 64. Map. Ammonium concentrations in surface water around Skidaway Island. ..... 64 Figure 65. Chart. TIN concentrations found in the groundwater in relation to well location. Distance (0,0) is well B7.5, the well located most closely to the saltwater marshes. ..... 65 Figure 66. Chart. Nitrate concentrations found in the groundwater in relation to well location. Distance (0,0) is well B7.5, the well located most closely to the saltwater marshes. ........................................................................................................................... 65 Figure 67. Chart. Ammonium concentrations found in the groundwater in relation to well location. Distance (0,0) is well B7.5, the well located most closely to the saltwater marshes. ........................................................................................................................... 66 Figure 68. Graph. TIN concentrations in surface water determined as a function of slope angle. ............................................................................................................................... 67 Figure 69. Charts. Distribution of total suspended and dissolved solids observed at inlet of sand filters (International Stormwater Database). ........................................................... 72 Figure 70. Charts. Distribution of pH, temperature and turbidity observed at inlet of sand filters (International Stormwater Database). ................................................................... 73 Figure 71. Charts. Distribution of total nitrogen, NOX, copper, lead, phosphorus and zinc observed at inlet of sand filter (International Stormwater Database).............................. 74 Figure 72. Charts. Distribution of dissolved copper, lead, phosphorus and zinc observed at inlet of sand filter. ........................................................................................................... 75 Figure 73. Charts. Scatter plot of influent v/s effluent concentration of total suspended and dissolved solids in stormwater runoff. ............................................................................ 77 Figure 74. Charts. Scatter plot of influent vs effluent concentration of total pH, temperature and turbidity in stormwater runoff. ................................................................................. 77 Figure 75. Charts. Scatter plot of influent vs effluent concentration of total and dissolved phosphorus, total NOx and Nitrogen in stormwater runoff............................................. 78 Figure 76. Charts. Scatter plot of influent vs effluent concentration of total and dissolved metals in stormwater runoff. ........................................................................................... 79 Figure 77. Diagram. Box plot of influent and effluent concentration of total and dissolved metals in stormwater runoff. ........................................................................................... 82 Figure 78. Diagram. Box plot of influent and effluent concentration of total pH, temperature and turbidity in stormwater runoff. ................................................................................. 82 Figure 79. Diagram. Box plot of influent and effluent concentration of total and dissolved metals in stormwater runoff. ........................................................................................... 83 Figure 80. Diagram. Box plot of influent and effluent concentration of total and dissolved phosphorus, total NOX and nitrogen in stormwater runoff.............................................. 84 viii Figure 81. Graph. Probability plot of influent and effluent concentration of total and dissolved solids in stormwater runoff. ............................................................................ 86 Figure 82. Graph. Probability plot of influent and effluent concentration of total pH, temperature and turbidity in stormwater runoff. ............................................................. 86 Figure 83. Graph. Probability plot of influent and effluent concentration of total and dissolved metals in stormwater runoff. ........................................................................... 87 Figure 84. Graph. Probability plot of influent and effluent concentration of total and dissolved phosphorus and total NOX and nitrogen in stormwater runoff........................ 88 Figure 85. Charts. Clusters of cleaning efficiency of sand filters for total metals and design parameters of the sand filters for each cluster. ................................................................ 89 Figure 86. Charts. Cluster analysis and design parameters of sand filters with respect to total solids................................................................................................................................ 91 Figure 87. Diagram. BMP selection process flowchart (Figure 10.5-1 from GDOT, 2018). 94 Figure 88. Graph. The grain size distribution curve for the inorganic portion of the topsoil. ....................................................................................................................................... 104 Figure 89. Graph. B value measured for specified topsoil hydraulic conductivity sample. 107 Figure 90. Graph. Hydraulic conductivity of specified topsoil with only inorganic minerals (no organic matter included). ........................................................................................ 107 Figure 91. Graph. Hydraulic conductivity of specified topsoil with both inorganic minerals and organic carbon. ....................................................................................................... 108 ix LIST OF TABLES Table 1. Stormwater BMP Routine Maintenance Burden (GDOT Drainage Manual). ........ 13 Table 2. Typical BMP Maintenance Activities and Cost (Adapted from National Academies of Sciences, Engineering, and Medicine, 2014). ............................................................. 15 Table 3. BMP Expected Life Span (Table from National Academies of Sciences, Engineering, and Medicine, 2014). ................................................................................. 16 Table 4. Forsyth County, GA Dry Swale/Sand Filter BMP Data. ........................................ 20 Table 5. Summary of Monitored Tests for Dry Swale/Sand Filter, Forsyth County ............ 23 Table 6. Summary of Rainfall Events for Dry Swale/Sand Filter, Forsyth County. ............. 39 Table 7. Summary of Removal Efficiencies for Dry Swale/Sand Filter, Forsyth County. ... 40 Table 8. Bartow County, GA Bioretention BMP Data.......................................................... 42 Table 9. Summary of Monitored Tests and Rainfall for Bioretention Basin, Bartow County. ......................................................................................................................................... 44 Table 10. Summary of Removal Efficiencies for Bioretention Basin, Bartow County. ....... 58 Table 11. Design Details of Sand Filters Used in the Analysis (International Stormwater Database). ........................................................................................................................ 69 Table 12. P-values of the Kruskal-Wallis Test for Influent Versus Effluent Concentrations. ......................................................................................................................................... 81 Table 13. Median Cleaning Efficiency of Clusters (Metals)................................................. 90 Table 14. Median of Design Parameters for Different Clusters. ........................................... 90 Table 15. Median Cleaning Efficiency of Clusters (Solids).................................................. 91 Table 16. Median of Design Parameters for Different Clusters. ........................................... 92 Table 17. BMP Longevity and Maintenance Burden. ........................................................... 95 Table 18. Summary of Stormwater Event 1 Results (Dry swale with a sand filter). ............ 98 Table 19. Summary of Stormwater Event 2 Results (Dry swale with a sand filter). ........... 99 Table 20. Summary of Stormwater Event 3 Results (Dry swale with a sand filter). .......... 100 Table 21. Summary of Stormwater Event 4 Results (Bioretention basin). ......................... 101 Table 22. GDOT Grain Size for Topsoil (Inorganic). ......................................................... 103 Table 23. Organic Carbon Content Commercial Cow Manure. .......................................... 105 Table 24. Moisture Content Commercial Cow Manure. ..................................................... 106 x EXECUTIVE SUMMARY A variety of stormwater Best Management Practices (BMPs) are being used throughout the United States to naturally attenuate contaminated stormwater runoff. Because each BMP has its own specific characteristics and application, any one BMP may not be applicable to all locations and conditions, which makes selecting the optimum BMP for a given site and suite of stormwater contaminants somewhat challenging. The most common complication in selection is optimizing pollutant removal while minimizing right-of-way acquisition. The Department's MS4 permit requires that a "stormwater management system shall be designed to retain up to the first 2.54 cm (1.0 in) of rainfall on the site, to the maximum extent practicable" and if that is not feasible, the remaining runoff must be treated to 80% total suspended solids (TSS) removal, which is taken as a surrogate for other contaminants that are found in stormwater, such as nutrients and heavy metals. The work performed in this study investigated the field performance of two functioning BMPs: a dry swale/sand filter in Forsyth County, GA and a bioretention basin in Bartow County, GA, as well as a series of steep slopes in the coastal plain region of Georgia (Skidaway Island). During stormwater runoff events, both BMPs demonstrated pronounced first flush peak concentrations, with reduction in the inflow contaminant levels within one hour of stormwater flow. The combination dry swale / sand filter successfully treated turbidity, total suspended solids, total dissolved solids, some nutrients, and some heavy metals. The bioretention basin also successfully treated contaminants and showed better performance in the removal of nutrients when compared 1 to the sand filter. The impact of high slope angles on the concentrations of nutrients was monitored in the coastal plain hydrogeology, and analysis of surface water samples for total inorganic nitrogen (TIN) revealed that the concentration of TIN measured in surface water samples taken at the base of the slopes increased as slope angle was increased, indicating that removal percentages for nutrients may be negatively impacted by steeper slope angles. An in-depth statistical analysis of the performance of sand filters, based on an extensive database collected at the GDOT sand filter in Canton, GA and data drawn from the International Stormwater Database demonstrated that optimization of BMP dimensions for contaminant removal is a complex, multi-constrained problem. Dimensions that were optimized for metal removal did not perform as well for solids removal, and vice-versa; consequently, selection criteria based on optimized BMP dimensions is not feasible at this point in time, due to the complexity of other factors such as regional hydrogeology and precipitation events. Based on the concentrations of heavy metals, zinc (Zn) was the most abundant in stormwater runoff (mean 126.8 g/L), followed by copper (Cu), and lead (Pb). The ratio of mean dissolved to mean total concentration of the pollutants suggests that mitigating the pollution caused by different contaminants would require treatment of both suspended and dissolved solids. In sand filters, mitigation of zinc and phosphorus requires treatment of dissolved pollutants, while mitigation of copper and lead could be treated by removing suspended solids. Finally, it is recommended that the selection criteria flowchart for BMPs on GDOT right-of-way be expanded to include criteria for longevity, as well as distinction between short/intermediate term maintenance versus long term maintenance burden. 2 CHAPTER 1. INTRODUCTION Within recent decades, the Georgia Department of Transportation (GDOT) has been using a variety of permanent stormwater retention/treatment structures, known as high-performance best management practices (BMPs) on right-of-way throughout the State. Typically, the structures are designed for both hydraulic control and for contaminant removal. However, the dimensions of the most commonly implemented BMPs are specified by existing standards, leaving the engineer with limited alternatives in the design and construction of these devices. For example, the Georgia Stormwater Management Manual (i.e., the Blue Book) specifies a maximum slope of 6% on swaletype BMPs constructed for solids removal. This mild slope specification results in strict requirements on the associated right-of-way required along the roadway. Regulation of post-construction pollutants and their mitigation is controlled by the Department's municipal separate storm sewer system (MS4) permit, which specifies requirements for removal of total suspended solids (TSS), for detention of runoff volume, and for mitigation of overland flow. While the current designs for BMPs specified on GDOT right-of-way are functioning well, this research proposes to monitor if the design of the most currently implemented BMPs could be optimized to reduce the cost of rightof-way acquisition, while still maintaining the required environmental protection. According to the National Water Quality Inventory Report (US EPA, 2009), an assessment of 5.7 million km of rivers and streams [representing 16% of the total in the US] revealed that 44% were found to be impaired, i.e., not able to support one or more of its designated uses. The most common sources of impairment include runoff from agricultural activities, hydro-modification, habitat alteration, unspecified non-point 3 sources, atmospheric deposition, and urban runoff from stormwater (US EPA, 2009). According to the water quality assessment report for Georgia (US EPA, 2010), for the 19% of the total rivers and streams [112,896 km] that were assessed, 58% were found to be impaired. In all the impaired rivers and streams, the pollutant contribution from nonpoint sources was highest at 68%, while urban stormwater related runoff contributions to the impairment was second highest, at 25.3%. For GDOT to maintain runoff waterquality by limiting contaminant discharge to receiving waters, understanding the components of runoff originating from highway surfaces in Georgia and the performance of stormwater BMPs to date is important. Therefore, designing and building physically and economically effective solutions to treat pollutants in the highway runoff before they discharge into receiving waters is paramount. Two of the major questions required to assess the efficiency of any BMP in attaining water quality goals (US EPA 2002) are: (1) How varied is the degree of pollution control performance, i.e., effluent quality, provided by the BMP from pollutant to pollutant? (2) How is stormwater volume mitigated? Hydraulic control is relatively straightforward; however, for contaminants, stormwater runoff contains a variety of pollutants that can impact the quality of receiving waters and some parameters may even be site specific (US EPA 2002). Pollutants may be divided into three basic categories which are useful to assess the efficacy of BMP structures: (1) physical characteristics like temperature, pH, conductivity, etc.; (2) concentration of heavy metals (e.g., lead, copper, etc.); and (3) nutrient loadings (e.g., nitrates, nitrites, phosphates, etc., which impact aquatic life quality). 4 This project will provide data and a summary of knowledge that allows for modification of the GDOT drainage manual to serve two critical Department needs: (1) selection criteria specified to implement low maintenance BMPs in order to reduce the long term burden on upkeep, and (2) design parameters optimized for design and construction in transportation right-of-way (as opposed to parameters that were optimized for applications with site development criteria). This project work plan will include: (1) A comprehensive literature review to examine the factors that control contaminant removal in a variety of optimized stormwater structures. This will include all factors that act to increase, or decrease, contaminant removal in the BMPs that are commonly specified on GDOT right-of-way. (2) Performance at three field sites to assess contaminant removal under optimized BMP dimensions with statistical analysis of performance data and comparison to data from the International Stormwater BMP Database. (3) Statewide guidance for design conditions to optimize contaminant removal, while minimizing right-of-way acquisition for construction of stormwater BMPs, with specific emphasis on refining GDOT specific stormwater parameters for design. 5 CHAPTER 2. LITERATURE REVIEW A variety of stormwater BMPs are being used throughout the United States to naturally attenuate contaminated stormwater runoff. Because each BMP has its own specific characteristics and application, any one BMP may not be applicable to all locations and conditions. This tends to make selecting the optimum BMP for a given site and suite of stormwater contaminants somewhat challenging. The current practice is to use selection matrices published in various state DOT manuals to facilitate the selection of an adequate BMP for a particular application. The most common complication is having the desire to optimize pollutant removal while minimizing right-of-way acquisition (Wang et al., 2009). Methods of comparing and balancing these variables have been investigated (Bhatt, 2016), which laid the foundation for implementation of the Analytical Hierarchy Process (AHP), developed by Saaty et al. (1980) and Young (2010). AHP is a hierarchical technique for organizing and analyzing complex decisions, which can be applied to the complex decisions such as development and placement of BMPs. The Department's MS4 permit (2017) requires that a "stormwater management system shall be designed to retain up to the first 2.54 cm (1.0 in) of rainfall on the site, to the maximum extent practicable" and if that is not feasible, the remaining runoff must be treated to 80% total suspended solids (TSS) removal. Typically, TSS is taken as a surrogate for other contaminants that are found in stormwater, such as nutrients and heavy metals, and treatment for TSS is assumed to reduce those concentrations as well. Previous research (Bhatt, 2016) on specific contaminant removals resulted in the following conclusions regarding the removal performance for highway runoff draining into sand filters (for runoff from GDOT right-of-way into the Canton sand filter): the distribution of most incoming 6 pollutants followed a log normal distribution, implying that the occurrence of extreme contaminant loadings was low, and the historical sand filter data revealed that among nutrients, a majority of the total phosphorus was mitigated by the sand filter whereas neither dissolved phosphorus, total NOX, nor nitrogen was mitigated. For metals, zinc was mitigated by the sand filter, but copper and lead were not. However, investigation into the mechanisms that contribute to these results is still ongoing. COMMON BMPS IN USE BY GEORGIA DOT Bioretention Basins Bioretention basins are shallow depressions or designed basins used to pool and reduce the velocity of stormwater. Stormwater runoff is drained into the basin, where it is treated through a combination of processes, including infiltration, physical separation, and biological uptake and/or degradation. Overflow runoff from bioretention basins is designed to drain to an additional BMP or into receiving waters. In most instances, pretreatment of stormwater through physical separation is necessary before flow into the bioretention basin, in order to prevent clogging of the drainage media. The basins are most commonly implemented to treat small drainage areas of less than five acres, with ponding depths limited to 0.3 m (1 ft) or less, in order to drain the structure within 12 hours. The GDOT Drainage Manual (section 10.6.7) specifies the maximum ponding volume drain time for GDOT bioretention basins at 24 hours, with the maximum drain time for the entire structure equal to 72 hours. Bioretention basins are effective at contaminant removal, do not have a large footprint, and can be landscaped for integration into the surrounding areas. Additionally, 7 due to the design of the structure, the basins can be constructed in soils with low hydraulic conductivity, or in areas with impervious cover, such as parking lots. While bioretention basins are effective for contaminant treatment, limitations include their small size, which makes them ineffective for discharge attenuation, as well as their high cost and high level of maintenance. Recent studies on the performance of bioretention basins have focused on the pollutant removal capacity of bioretention basins in a range of climatic conditions. Removal efficiencies are a function of storm size and antecedent dry days, with high rainfall events resulting in lower pollutant removal in the basin and an increasing number of antecedent dry days increasing nitrate concentrations in the basin (Manganka, et al., 2015). Lucke and Nichols (2015) performed controlled tests in five bioretention basins that had been in operation for ten years. Four synthetic stormwater samples were prepared, with concentrations ranging from no pollutants to five times the expected load. The tested basins attenuated flow and phosphorus, and reduced total nitrogen in the effluent for all cases except the clean water sample. In the case of zero nitrogen in the influent, there was leaching from the basin into the effluent. Lucke and Nichols (2015) also tested the metal and hydrocarbon concentrations accumulated in the filter, and demonstrated the concentrations were below regulatory limits after ten years of operation. Climate and rainfall intensity also impact the removal efficiency of BMPs, because more frequent smaller storms produce a lower event mean concentration and lower concentration in the first flush (Wang et al., 2017). Ninety-six storms were monitored in a tropical climate in Singapore and demonstrated that the limited storage capacity of the tested bioretention basins resulted in overflow and reduced pollutant removal capacity. Basins designed for increased water quality volume and depth would provide better 8 removal performance due to reduction in overflow events (Wang et al., 2017). Hunt et al. (2008) measured performance of an urban bioretention cell in Charlotte, North Carolina from 2004 2006. Flow weighted composite samples (23 storms) were used to quantify removal of nutrients, suspended solids, and heavy metals, while grab samples (19 storms) were collected to test for the presence of Escherichia coli (E. coli). There were measurable decreases in the concentrations of total nitrogen (TN), total Kjeldahl nitrogen (TKN), ammonium (NH4-N), biochemical oxygen demand (BOD-5), fecal coliform, E. Coli, TSS, copper, zinc, and lead, but an increase in the concentration of dissolved iron (330%). Their data also demonstrated that the bioretention basin was effective at reducing peak runoff for small to medium sized storm events (Hunt et al., 2008). Enhanced Swales Enhanced swales are vegetated open channels, dry or wet, that are designed with interior cells that control flow and reduce channelized flow velocity with check dams or other similar structures. Swales are constructed on shallow slopes, usually with longitudinal slopes of 4% or less, and bottom widths of 0.6 2.4 m (2 to 8 ft), typically. Both dry and wet swales provide runoff reduction and water quality benefits through the settlement of suspended solids and removal of nutrients and metals; however, dry swales demonstrate better removal of contaminants. Yu et al. (2001) tested the field performance of grass swales in northern Virginia and Taiwan to measure pollutant removal efficiency. The efficiency of removal of solids, chemical oxygen demand, total nitrogen, and total phosphorus varied significantly, from 14% removal to 99% removal, and the authors recommend a minimum length of swale of 68.6 m (225 ft) with a 3% longitudinal slope. 9 Fletcher et al. (2012) performed controlled experiments on grassed swales in Brisbane Australia, with synthetic stormwater as influent. The swales were able to reduce total suspended solids, total nitrogen, and total phosphorus; however, removal efficiency for total suspended solids deceased as the flowrate increased. While grassed swales can remove metals and nutrients, they are most effective at reducing the concentration of suspended solids through sedimentation (Li et al., 2016). Filter Strips Filter strips are uniformly sloped, vegetated structures that are designed to treat sheet stormwater flow through filtration of solid particles, infiltration of water into the soil, and by slowing of the flow of runoff. Filter strips have multiple advantages over other BMPs because they are relatively easy to construct, inexpensive, and are easily blended into the surrounding landscape. In addition, they are suitable for a range of site conditions and can be combined with other BMPs to create a treatment train. However, they treat small drainage areas and they do require a large land area, which is especially disadvantageous for the Department as even small additions to right of way can have a large impact on initial cost and maintenance. Filter strips generally provide little control of volume runoff, especially relative to other BMPs, and if not properly maintained, they can erode and concentrate flow. The Department of Transportation and the Blue Book guidance have recommendations for slopes between 2% and 6%, with the top and the toe of the slope as flat as is feasible. A typical filter strip will be 7.6 to 15.2 m (25 ft to 50 feet) perpendicular to the highway, with the Blue Book recommending a maximum strip of 10 30.5 m (100 feet) along the treatment direction, i.e. with the length of water flow. The grasses in the filter strip must be able to withstand flows of 1.2 m/second (4 feet per second). VEGETATED FILTER STRIP PERFORMANCE AS A FUNCTION OF SLOPE Recent studies of the performance of vegetated filter strips have demonstrated their effectiveness in the reduction of suspended solids and nutrients from the stormwater flow. Robinson et al. (1996) measured the sediment concentration in runoff flowing through a vegetated filter strip as a function of distance, slope, rainfall quantity, and runoff quantity, and showed that filter strips reduce the concentration of suspended solids, with 70% removal occurring in the first 3 m (10 feet) of flow and 85% removal within the first 9 m (30 feet), with little observed reduction in contaminant concentration at lengths greater than 9 m (30 feet). The slope of the filter strip did impact soil loss (or sediment load), with the slope of 12% having increased runoff and soil loss when compared to the slope of 7%. Controlled field tests performed on a grass filter strip in Aberdeen, Scotland and a grass swale in Brisbane, Australia demonstrated that solids content was reduced by approximately 70% at higher slopes, while the Australian tests demonstrated a decrease in nitrogen and phosphorus of approximately 50% (Deletic and Fletcher, 2006). Additional work on the removal of pesticides, nitrogen, and phosphorus from runoff through grass filter strips with lengths ranging from 6 m (20 feet) to 12 m (60 feet) demonstrated 90 100% reduction in suspended solids, 40 100% reduction in runoff volume, 50 100% reduction in nitrate, and 20 90% reduction in soluble phosphorus (Patty et al., 1997). 11 Modeling of the removal of solids and contaminants in a vegetated filter strip has been an area of active research. One of the most sophisticated models, known as VFSMOD-W: Vegetative Filter Strip Modeling System, was developed by Munoz and Carpena (1998) at North Carolina State University and is now maintained at the University of Florida (http://abe.ufl.edu/carpena/vfsmod/). VFSMOD is a finite element solution for overland flow, infiltration, sediment deposition, phosphorus removal (particulate and dissolved), and pesticide removal. The mass balance for removal incorporates runoff, removal of coarse and fine sediments, and infiltration (Figure 1). Field experiments were performed in the Piedmont geology in North Carolina and gave good agreement with model predictions as long as sheet flow was maintained in the filter. Figure 1. Diagram. Mass balance for flow, infiltration, and deposition in a vegetated filter strip. (Figure from Munoz and Carpena, abe.ufl.edu/carpena/vfsmod/FOCUS/VFSMOD_Pmassbalance1.ppt). 12 MAINTENANCE OF STORMWATER BMPS Maintenance of stormwater BMPs represents a significant long-term cost to GDOT. In order to ensure that a BMP is operating as designed, there are several categories of maintenance that should be performed at regular intervals. The primary types of maintenance include structural, routine, runoff pretreatment, maintenance of conveyance channels, and maintenance of slopes. In terms of structural issues, maintenance includes repairing clogged or broken pipes, repairing missing or broken parts (e.g., valves, seals, manholes), repairing cracked concrete, repairing erosion at outfall or on banks, regrading or dredging, and some landscaping refurbishment. In contrast, routine maintenance is primarily focused on vegetation management such as mowing, removal of vegetative overgrowth and woody plants, replacing dead or diseased landscaping, control of invasive plants, and removal of trash, yard debris, or small amounts of sediment buildup. Runoff pretreatment refers to maintenance of the pretreatment bay by controlling vegetation, sediment, and debris, with similar activities performed for maintenance of conveyance channels and slope maintenance. The GDOT Drainage Manual categorizes the routine maintenance burden for commonly used BMPs as low, medium, or high (Table 1). Table 1. Stormwater BMP Routine Maintenance Burden (GDOT Drainage Manual). BMP Filter Strips Grass Channels Enhanced Dry Swale Enhanced Wet Swale Infiltration Trench Bioslope Sand Filter Bioretention Basin Dry Detention Basin Wet Detention Pond Stormwater Wetlands Level 2 Stormwater Wetlands Level 1 Open-Graded Friction Course Maintenance Burden Low Low Medium Low High Medium High Medium Low Low Medium Medium Low Maintenance Tasks Remove sediments Maintain vegetation Inflow/outflow unobstructed Mow grass 13 Maintenance for slopes is particularly important to ensure contaminant removal efficiency and to reduce erosion. A survey of maximum longitudinal slopes for BMPs in the US demonstrated that the maximum recommended channel slope was generally 4% 5% but ranged as high as 10% in limited applications in New Jersey (New Jersey Department of Environmental Protection, 2020). Natural, forested buffers are appealing alternatives for implementation as low cost stormwater BMPs. In general, the primary considerations for good performance of any BMP will include flow path, slope, soil/infiltration, vegetation, and the presence of an organic absorptive zone. The literature on contaminant removal efficiency for forested buffers is scant, but in general, they are limited to low flow applications with gravity drainage in sheet flow but have shown some potential for reduction of nutrients. For applications that rely on grassed buffers or accumulate sediments, similar maintenance burdens will be required. A comprehensive study of the life cycle costs associated with maintenance of stormwater BMPs was performed by the National Academies (2014) to quantify the economic impact and lifespan of the devices that are currently in place (Table 2 Table 3). Maintenance activities ranged simple, low cost but high frequency vegetation management to more complex, high cost but low frequency long term maintenance such as sand filter media replacement (Table 2). The limitation on BMP lifespan was sediment accumulation for vegetated strips and swales, and friction courses, and pipe and concrete longevity for detention basins and filters (Table 3). Aside from friction course, the most commonly used BMPs are predicted to function for decades. Guidance on the frequency of BMP maintenance activities was also reported (Table 2). 14 Table 2. Typical BMP Maintenance Activities and Cost (Adapted from National Academies of Sciences, Engineering, and Medicine, 2014). Category Activity Vegetation Management Interim Maintenance Long Term Maintenance Aesthetic repair Trash and debris removal Mulch management Sediment management and removal Vegetation repair Erosion/rutting Slope inspection Standing water Sediment management Underdrain repair Sand replacement media Compliance Inspection/reporting Frequency 3X per year 2X per year 1X per year Every 2 years Every 5 years Every 10 years Every 20 years Every 30 years Every 50 years Every 4 years Every 8 years Every 12 years Every 3 years Every 5 years Every 10 years 1X per year Hours / Staff 4 hours 2 people 8 hours 2 people 8 hours 4 people 24 hours 4 people 8 hours 4 people 1 hour 2 people Equipment demand Low Medium *disposal High *disposal High High *disposal Low Estimated Cost $640/event $1,280/event $3,800/event $3,800/event $75,600/event - $3,700/event $130/year 15 Table 3. BMP Expected Life Span (Table from National Academies of Sciences, Engineering, and Medicine, 2014). BMP Type Life Span Limiting Factor Vegetated strips 860 years (depending on ecoregion) Sediment accumulation Vegetated swales 1050 years (depending on ecoregion) Sediment accumulation Dry detention basin 80 years Pipe material longevity Bioretention 80 years Pipe material longevity Retention pond 80 years Pipe material longevity Sand filter 75 years Concrete longevity Permeable friction course 14 years Sediment accumulation 16 CHAPTER 3. FIELD PERFORMANCE OF THREE GDOT BEST MANAGEMENT PRACTICE (BMP) SITES Several BMPs were chosen for study with consideration of the type of BMP, location of BMP, and hydrogeology. In the greater metro Atlanta region, the performance of two BMPS were studied: 1) combination dry swale/sand filter and 2) bioretention basin, and in the Coastal Plain, the performance of a series of filter strips were monitored for nutrient impact on surface and groundwater. BMP FIELD SAMPLING AND TESTING, MATERIALS AND METHODS For the stormwater monitoring program, three automatic samplers (Sigma 900 MAX PS1 Portable Automatic Sampler) were used. Each automatic sampler was equipped with four one-gallon polyethylene bottles for sample collection. Flow was measured with an integral HACH Sigma Area-Velocity flow meter (#4041) using a pressure transducer for depth of flow measurement and a pair of ultrasonic transducers for velocity measurement. The area-velocity sensors were installed and secured at the base of the tested pipes. Insitu parameters pH, specific conductance (SC), and temperature (T) were measured with an integral pH- temperature probe (Hach, #8793), and integral conductivity probe (Hach, #3227). The three sensors were also securely placed at the base of the pipes to continuously record the three parameters. Rainfall depths at the site were measured with a tipping bucket rain logger (Sigma). In-situ parameters (temperature, conductivity, pH, flow depth and rainfall) were recorded at an interval of 1 minute throughout the duration of an event. The recorded data were transferred to a personal computer using Hach Insight software. Stormwater samples were collected for each sampler using three bottles to capture the first flush for the first 30-45 minutes of the storm. In the fourth bottle, 200 ml grab samples were collected at an interval of 15 minutes for the whole event, or until 17 capacity was reached. Sample collection was automated, and the triggering condition for the initiation of the sample collection was set as 2.5 cm of flow depth. This was selected to ensure that the intake pipe was sufficiently submerged to collect an accurate volume of the sample. The mouth of the intake pipe had a strainer to prevent clogging of the intake pipe, and the samples were collected by a peristaltic pump. The sampler controller was programmed to rinse the intake pipe once before the collection of a sample. In addition to the in-situ parameters that were recorded continuously, the automatically collected samples were brought from the site to the Geoenvironmental laboratory at the Georgia Institute of Technology within 24 hours (usually within 12 hours) after the completion of sampling program to avoid sample deterioration. The samples were preserved for testing per procedures for different water quality parameters (CFR, 2009). An adequate volume of a sample was passed through a 45m filter paper and preserved separately to test for dissolved parameter concentrations. The runoff samples were tested for total suspended solids, turbidity, specific conductivity, pH, total dissolved solids, nutrients (total nitrogen, nitrate and nitrite and total phosphorus) and metals (total copper, total lead, total zinc, dissolved copper, dissolved lead, and dissolved zinc). Samples were analyzed for metals using Perkin Elmer Optima 7300 DV Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES). Nutrients were measured using a Shimadzu UV-1800-Vis Spectrophotometer. All lab ware, sample bottles and intake tubing utilized for testing or collection of samples were rinsed with 1% nitric acid (HNO3) and de-ionized water (Barnstead, E-pure). BMP: DRY SWALE WITH SAND FILTER IN FORSYTH COUNTY, GA Site 1 is located in Forsyth County (at the Gwinnett County border), Georgia on State Road 20 (SR 20), with outfall to the Chattahoochee River (latitude of 34.127489 and longitude -84.094186). The site contains three different dry swale pond BMPs, which drain the bridge and some watershed surrounding SR 20 at the Chattahoochee River. The 18 dry swale that was monitored was located in the north west corner of the SR 20 / Chattahoochee River crossing. Therefore, the purpose of this BMP was to treat the roadway surface runoff before it enters the Chattahoochee River (Figure 2, Table 4). Figure 2. Map. Topographic map of the Cumming GA / BMP location (USGS). 19 Table 4. Forsyth County, GA Dry Swale/Sand Filter BMP Data. General Test Site Information BMP Test Site Name Location Elevation at top of bioretention pond Structural BMP Information Structural BMP Name BMP Description Treatment Category Number of Inlets Inlet Description Number of Outlets Outlet Descriptions Catchment Area Watershed Stations Regional Watershed Name Station Upstream BMP Downstream BMP SR 20 Dry Swale with Sand Filter SR 20, Cumming, GA 281.6 m (924 ft) Dry Swale Pond with Sand Filter Substantial residence time and storage volume Sedimentation, Filtration 1 Concrete spillway w/ rip rap rock filter 2 0.6 m (24 in) storm drain pipe with outlet control structure and an emergency spillway 2.4 hectare (6.01 acres) Upper Chattahoochee Monitoring stations immediately u/s and d/s of pond None, inflow received directly from SR 113 concrete channel None, effluent discharged to Raccoon Creek A plan view of this BMP shows the inlet, outlet, and emergency spillway (Figure 3); however, it is important to note that the pre-construction designs went through many variations during construction, and the as built configuration had some variation from the initial design. In the as built configuration, the dry swale pond and sand filter also included a rock filter dam (Figure 4), with the cross-section view of a typical dry swale pond with sand filter (Figure 5). The outlet structure consists of a 15.2 cm (6 in) pipe that brings water into a control structure for gradual release, followed by a 61 cm (24 in) drainpipe that discharges water through a rip rap rock filter to the Chattahoochee River. Monitoring occurred at the inlet concrete spillway (Figure 6) and at the outlet control structure prior to water entering the drain pipe, with sensors positioned to measure water quality during the stormwater runoff period (Figure 7). 20 Figure 3. Diagram. Inlet and outlets labeled for the dry swale pond with sand filter. Figure 4. Diagram. Dry swale pond with sand filter showing the rock filter dam (circled in green). The filter ring (circled in red) was removed after construction was completed. 21 Figure 5. Diagram. Cross-section of the dry swale pond with sand filter. Figure 6. Photos. Test site (dry swale with a sand filter) at Chattahoochee River. 22 Figure 7. Photos. Test set-up for stormwater monitoring at BMP inflow and outflow. Experimental Results A total of three storm events were monitored (Table 5) with a range of precipitation from 0.3 cm (0.12 in) to 2.9 cm (1.14 in): Table 5. Summary of Monitored Tests for Dry Swale/Sand Filter, Forsyth County Test No. Date Total Precipitation (cm) Precipitation Duration (hrs) Dry Period (Days) Samples 1 06/22/2018 0.3 (0.12 in) 4 1 Inlet : 2 Outlet: 4 2 02/11/2019 0.7 (0.28 in) 6.5 5 Inlet : 4 Outlet: 4 3 02/19/2019 2.9 (1.14 in) 19 2 Inlet : 4 Outlet: 4 23 The stormwater samples were collected during the initial phase of runoff to assess the anticipated highest contaminant concentrations flowing into the BMP. For the first test, grab samples were taken at 30 minute intervals after initial flow was detected at the BMP inlet. For all other tests, samples were taken at 15 minute intervals, along with one composite sample that was collected over for the first 4 hours of the storm. Sample results indicated that contaminant levels were the highest in the first flush, and reduced within the first 15 minutes of storm duration (Figure 8 - Figure 17). In almost all tests, the BMP demonstrated characteristic first flush decrease over the 45 minute testing program, with reductions in turbidity, dissolved and suspended solids, copper, lead, zinc, and nutrients. Turbidity [NTU] 100 90 80 70 60 50 40 30 20 10 0 0 Test1_Inlet Test2_Inlet Test3_Inlet 15 30 45 60 Time [min] Figure 8. Graph. First flush turbidity at dry swale with a sand filter. 24 60 Test1_Inlet 50 Test2_Inlet 40 Test3_Inlet TSS [mg/L] 30 20 10 0 0 15 30 45 60 Time [min] Figure 9. Graph. First flush total suspended solids at dry swale with a sand filter. 200 Test1_Inlet 150 Test2_Inlet Test3_Inlet 100 TDS [mg/L] 50 0 0 15 30 45 60 Time [min] Figure 10. Graph. First flush total dissolved solids at dry swale with a sand filter. 25 0.2 0.15 0.1 Test1_Inlet Test2_Inlet Test3_Inlet Copper [mg/L] 0.05 0 0 15 30 45 60 Time [min] Figure 11. Graph. First flush total copper at dry swale with a sand filter. 0.25 Test1_Inlet 0.2 Test2_Inlet 0.15 Test3_Inlet Lead [mg/L] 0.1 0.05 0 0 15 30 45 60 Time [min] Figure 12. Graph. First flush total lead at dry swale with a sand filter. 26 0.2 0.15 0.1 Test1_Inlet Test2_Inlet Test3_Inlet Zinc [mg/L] 0.05 0 0 15 30 45 60 Time [min] Figure 13. Graph. First flush total zinc at dry swale with a sand filter. 1 Test1_Inlet 0.8 Test2_Inlet Test3_Inlet 0.6 Nitrite [mg/L] 0.4 0.2 0 0 15 30 45 60 Time [min] Figure 14. Graph. First flush nitrite at dry swale with a sand filter. 27 2.5 Test1_Inlet 2 Test2_Inlet Test3_Inlet 1.5 Nitrate [mg/L] 1 0.5 0 0 15 30 45 60 Time [min] Figure 15. Graph. First flush nitrate at dry swale with a sand filter. 10 Test1_Inlet 8 Test2_Inlet Test3_Inlet 6 Total Nitrogen [mg/L] 4 2 0 0 15 30 45 60 Time [min] Figure 16. Graph. First flush total nitrogen at dry swale with a sand filter. 28 Total Phosphorus [mg/L] 0.2 0.15 0.1 Test2_Inlet Test3_Inlet 0.05 0 0 15 30 45 60 Time [min] Figure 17. Graph. First flush total phosphorus at dry swale with a sand filter. In-situ parameters including pH, temperature, and conductivity were measured at intervals of 1 minute during the storm event (Figure 18 - Figure 20). Measured pH values in stormwater runoff were within the state standard (6.0 -8.5), decreasing as the storm progressed. Measured temperature was also within the state standard of < 32 C (90 F). Conductivity demonstrated an initial peak due to inflow runoff but decreased and stabilized after the initial 15 minutes of inflow. 29 pH 10 9 8 7 6 5 4 19:25 19:55 20:25 20:55 21:25 Time 21:55 22:25 22:55 Figure 18. Graph. In-situ pH values during the stormwater runoff (dry swale/sand filter). 20 15 10 5 0 19:25 19:55 20:25 20:55 21:25 Time 21:55 22:25 22:55 Figure 19. Graph. In-situ temperature values during the stormwater runoff (dry swale/sand filter). Temperature [oC] 30 Conductivity [S/cm] 200 180 160 140 120 100 80 60 40 20 0 19:25 19:55 20:25 20:55 21:25 Time 21:55 22:25 22:55 Figure 20. Graph. In-situ conductivity values during the stormwater runoff (dry swale/sand filter). Conventional parameters including turbidity, total suspended solids, and total dissolved solids were measured for the inlet samples and outlet samples in order to determine percent removals in the BMP. Turbidity was measured using a TB200 portable turbidimeter (Orbeco). Analysis of total suspended solids and dissolved solids were determined using the methodology outlined in EPA 160.2. In all cases, the highest values were measured in the first sample taken at the inlet and concentrations significantly decreased between the inlet and outlet (Figure 21 - Figure 23). Also, suspended solids removal was visually apparent through color change on the microfiber filters used for filtration in the total suspended solids test (Figure 24). 31 Turbidity [NTU] 100 90 80 70 60 50 40 30 20 10 0 0 Test1_Inlet Test2_Inlet Test3_Inlet Test1_Outlet Test2_Outlet Test3_Outlet 15 30 45 60 Time [min] Figure 21. Graph. Turbidity test results for inlet samples and outlet samples (dry swale/sand filter). 60 Test1_Inlet Test2_Inlet 50 Test3_Inlet Test1_Outlet 40 Test2_Outlet Test3_Outlet 30 TSS [mg/L] 20 10 0 0 15 30 45 60 Time [min] Figure 22. Graph. Total suspended solid test results for inlet samples and outlet samples (dry swale/sand filter). 32 200 150 100 Test1_Inlet Test2_Inlet Test3_Inlet Test1_Outlet Test2_Outlet Test3_Outlet TDS [mg/L] 50 0 0 15 30 45 60 Time [min] Figure 23. Graph. Total dissolved solid test results for inlet samples and outlet samples (dry swale/sand filter). Inlet Samples Inlet Samples Outlet Samples Outlet Samples Figure 24. Photos. Microfiber filter results for total suspended solid tests (dry swale/sand filter). 33 Heavy metals concentrations (copper, lead, and zinc) were measured for the inlet samples and outlet samples in order to determine percent removal using inductively coupled plasma optical emission spectroscopy (ICP-OES). For almost all cases, heavy metal concentration was reduced between measured inlet and outlet concentrations (Figure 25 - Figure 27). 0.2 0.15 0.1 Test1_Inlet Test2_Inlet Test3_Inlet Test1_Outlet Test2_Outlet Test3_Outlet Copper [mg/L] 0.05 0 0 15 30 45 60 Time [min] Figure 25. Graph. Total copper at inlet samples and outlet samples (dry swale/sand filter). 34 0.25 0.2 0.15 Test1_Inlet Test2_Inlet Test3_Inlet Test1_Outlet Test2_Outlet Test3_Outlet Lead [mg/L] 0.1 0.05 0 0 15 30 45 60 Time [min] Figure 26. Graph. Total lead at inlet samples and outlet samples (dry swale/sand filter). 0.2 0.15 Test1_Inlet Test2_Inlet Test3_Inlet Test1_Outlet Test2_Outlet Test3_Outlet 0.1 Zinc [mg/L] 0.05 0 0 15 30 45 60 Time [min] Figure 27. Graph. Total zinc at inlet samples and outlet samples (dry swale/sand filter). 35 The concentrations of nutrients, including nitrite (NO2-), nitrate (NO3-), total nitrogen, and total phosphorus, were measured for the inlet and outlet samples (total phosphorus was not measured in the first test). For the analysis of nutrients, EPA 352.1 and EPA 365.2 method were followed, using a UV-spectrophotometer for determination of nutrient concentration. In almost all cases, the highest nutrient concentrations were observed in the first flush, with nutrient concentrations at the outlet significantly reduced when compared to the inlet concentration (Figure 28 - Figure 31). 1 Test1_Inlet Test2_Inlet 0.8 Test3_Inlet Test1_Outlet Test2_Outlet 0.6 Test3_Outlet Nitrite [mg/L] 0.4 0.2 0 0 15 30 45 60 Time [min] Figure 28. Graph. Nitrite concentration at inlet samples and outlet samples. 36 2.5 Test1_Inlet Test2_Inlet Test3_Inlet 2 Test1_Outlet Test2_Outlet Test3_Outlet 1.5 Nitrate [mg/L] 1 0.5 0 0 15 30 45 60 Time [min] Figure 29. Graph. Nitrate concentration at inlet samples and outlet samples. 10 Test1_Inlet Test2_Inlet Test3_Inlet 8 Test1_Outlet Test2_Outlet Test3_Outlet 6 Total Nitrogen [mg/L] 4 2 0 0 15 30 45 60 Time [min] Figure 30. Graph. Total nitrogen concentration at inlet samples and outlet samples. 37 Total Phosphorus [mg/L] 0.2 0.15 0.1 Test2_Inlet Test3_Inlet Test2_Outlet Test3_Outlet 0.05 0 0 15 30 45 60 Time [min] Figure 31. Graph. Total phosphorus concentration at inlet samples and outlet samples. For the monitored stormwater events, the number of antecedent dry days ranged from 1 to 5 (Table 6), which is evident in the higher concentration measured for inflow for the second monitored storm event. Rainfall data for the second storm event show heavy precipitation with high rainfall intensity that resulted in high inflow and outflow concentrations at the outlet samples. This storm event overtopped the BMP outlet drainage (Figure 32), which allowed stormwater to bypass the sand filter, resulting in higher discharge concentrations from the BMP. 38 Table 6. Summary of Rainfall Events for Dry Swale/Sand Filter, Forsyth County. Test No. 1 2 3 Total Precipitation (cm) 0.3 (0.12 in) 0.7 (0.28 in) 2.9 (1.14 in) Precipitation Duration (hrs) 4 6.5 19 Dry Period (Days) 1 5 2 Figure 32. Photo. Overtopping of stormwater at the BMP outlet drainage control structure. Summary Removal Efficiency: Dry swale/sand filter Forsyth County, GA In summary, the dry swale/sand filter effectively functioned to remove solids, metals, and nutrients from the stormwater runoff (Table 7). Solids and turbidity reduction 39 ranged between 60 90% removal, nutrient removals were ~40 90%, and heavy metal concentrations were reduced between 70 90%. Table 7. Summary of Removal Efficiencies for Dry Swale/Sand Filter, Forsyth County. Parameter Removal Efficiency (%) Test 1 Removal Efficiency (%) Test 2 Removal Efficiency (%) Test 3 Turbidity [%] 63 91 83 TSS [%] 92 66 89 TDS [%] 45 63 61 Nitrite [%] 87 89 70 Nitrate [%] 33 57 42 Total Nitrogen [%] 16 17 20 Total Phosphorus [%] - 87 80 Total Copper [%] 71 73 71 Total Lead [%] 88 89 91 Total Zinc [%] 83 77 78 40 BMP: BIORETENTION BASIN IN BARTOW COUNTY Site 2 is located in Bartow County, Georgia (City of Cartersville) along State Road 113 (SR 113) at latitude 34.114389 and longitude -84.890556. This site consists of two bioretention ponds that are located on opposite sides of the SR 113 bridge at Raccoon Creek, which is a 33.8 km (21.0 mile) creek originating in Paulding County, Georgia, joining the Etowah River approximately 2.0 km (1.3 miles) downstream from the observed site. The two bioretention ponds are used to treat the stormwater surface runoff from SR 113, before discharging to Raccoon Creek. The monitored bioretention pond is located on the west side of Raccoon Creek (known as Bioretention Pond A) with approximate dimensions of 6.0 m (20.0 ft) wide and 12.1 m (40.0 ft) long, with length being parallel to the road (Figure 33, Table 8). Figure 33. Map. Topographic Map of the Cartersville BMP area (USGS). 41 Table 8. Bartow County, GA Bioretention BMP Data. General Test Site Information BMP Test Site Name SR 113 Bartow County Bioretention Pond A Location SR 113, Cartersville, GA Elevation at top of bioretention pond 207 m (679 ft) Structural BMP Information Structural BMP Name Bioretention Basin BMP Description Substantial residence time and storage volume Treatment Category Sedimentation, Filtration Number of Inlets 2 Inlet Description 0.5 m (18 in) storm drain pipe and a concrete spillway Number of Outlets 1 Outlet Descriptions Drop inlet Watershed Stations Regional Watershed Name Etowah Station Monitoring stations immediately u/s and d/s of pond Upstream BMP None, inflow received directly from SR 113 Downstream BMP None, effluent discharged to Raccoon Creek Experimental Results The bioretention pond was monitored at the inlet and outlets for the BMP (Figure 34- Figure 35). The concrete spillway discharges runoff directly from SR 113, while the 0.5 m (18 in) storm drain pipe inlets runoff water from the median between SR 113 east and west. The outlet consists of a 0.8 m (2.75 ft) drop inlet with a 0.4 m (7 in) weir that flows into a 0.5 m (18 in) storm drain that drains through a riprap rock filter into Raccoon Creek (Figure 36). Samples were taken at the concrete spillway inlet and also at drop inlet (outlet) on February 28, 2019 (Table 9). 42 Figure 34. Diagram. Sample Locations at the Raccoon Creek Bioretention Basin. Figure 35. Diagram. Plan view of drop inlet with dimensions. 43 Figure 36. Photos. Bioretention basin test site at SR 113 and Racoon Creek, Bartow County . Table 9. Summary of Monitored Tests and Rainfall for Bioretention Basin, Bartow County. Location Date Lat. 34.114389, Long. 84.890556 02/28/2019 Total Precipitation (cm) 0.2 (0.07 in) Precipitation Duration (hrs) Dry Period (Days) 6 4 Samples Inlet : 4 Outlet: 4 First flush grab samples were collected at the inlet and outlet to the BMP at intervals of 15 minutes, and a composite sample was also collected during the first 4 hours of the storm to monitor changes in contaminant levels over the longer storm duration. As 44 anticipated, the results of the first flush monitoring showed the highest concentration of contaminants in the first sample taken, followed by significant decrease after 15 minutes of flow (Figure 37 - Figure 46). For samples taken at 30 minutes and 45 minutes after initiation of flow, the contaminant concentrations were almost invariable with time. 50 Test4_Inlet 40 Turbidity [NTU] 30 20 10 0 0 15 30 45 60 Time [min] Figure 37. Graph. First flush turbidity concentration for the bioretention basin, Bartow County GA. 45 60 Test4_Inlet 50 40 TSS [mg/L] 30 20 10 0 0 15 30 45 60 Time [min] Figure 38. Graph. First flush total suspended solids concentration for the bioretention basin, Bartow County GA. 200 Test4_Inlet 150 TDS [mg/L] 100 50 0 0 15 30 45 60 Time [min] Figure 39. Graph. First flush total dissolved solids concentration for the bioretention basin, Bartow County GA. 46 0.2 0.15 Test4_Inlet Copper [mg/L] 0.1 0.05 0 0 15 30 45 60 Time [min] Figure 40. Graph. First flush total copper concentration for the bioretention basin, Bartow County GA. 0.25 0.2 Test4_Inlet Lead [mg/L] 0.15 0.1 0.05 0 0 15 30 45 60 Time [min] Figure 41. Graph. First flush total lead concentration for the bioretention basin, Bartow County GA. 47 0.2 0.15 Test4_Inlet Zinc [mg/L] 0.1 0.05 0 0 15 30 45 60 Time [min] Figure 42. Graph. First flush total zinc concentration for the bioretention basin, Bartow County GA. 1 Test4_Inlet 0.8 Nitrite [mg/L] 0.6 0.4 0.2 0 0 15 30 45 60 Time [min] Figure 43. Graph. First flush nitrite concentration for the bioretention basin, Bartow County GA. 48 2.5 Test4_Inlet 2 Nitrate [mg/L] 1.5 1 0.5 0 0 15 30 45 60 Time [min] Figure 44. Graph. First flush nitrate concentration for the bioretention basin, Bartow County GA. 10 Test4_Inlet 8 Total Nitrogen [mg/L] 6 4 2 0 0 15 30 45 60 Time [min] Figure 45. Graph. First flush total nitrogen concentration for the bioretention basin, Bartow County GA. 49 0.2 0.15 Test4_Inlet pH Total Phosphorus [mg/L] 0.1 0.05 0 0 15 30 45 60 Time [min] Figure 46. Graph. First flush total phosphorus concentration for the bioretention basin, Bartow County GA. Measurement of in-situ parameters showed that inflow pH was relatively constant (~7.5 8), which was within state standards (Figure 47). Inflow temperature was approximately 15.5 C (60 F), which is also within state standards (Figure 48). Similarly to the dry swale/sand filter, the conductivity spiked within the first 15 minutes of runoff, when the initial stormwater runoff into the basin began, and then decreased for the duration of the storm (Figure 49). 10 9 8 7 6 5 4 7:40 8:10 8:40 9:10 9:40 Time 10:10 10:40 11:10 Figure 47. Graph. In-situ pH values during the stormwater runoff for the bioretention basin, Bartow County, GA. 50 Temperature [oC] 20 15 10 5 0 7:40 8:10 8:40 9:10 9:40 Time 10:10 10:40 11:10 Figure 48. Graph. In-situ temperature values during the stormwater runoff for the bioretention basin, Bartow County, GA. 200 180 160 140 120 100 80 60 40 20 0 7:40 8:10 8:40 9:10 9:40 Time 10:10 10:40 11:10 Figure 49. Graph. In-situ conductivity values during the stormwater runoff for the bioretention basin, Bartow County, GA. Conductivity [S/cm] Substantial reduction in turbidity, total suspended solids, and total dissolved solids were observed within the bioretention basin (Figure 50 - Figure 52). Initial inflow concentrations were low for the storm monitored at this BMP, with initial turbidity at ~40 NTU and TSS at ~40 mg/L. 51 50 Test4_Inlet Test4_Outlet 40 Turbidity [NTU] 30 20 10 0 0 15 30 45 60 Time [min] Figure 50. Graph. Turbidity test results for inlet samples and outlet samples for bioretention basin, Bartow County, GA. 60 Test4_Inlet Test4_Outlet 50 40 TSS [mg/L] 30 20 10 0 0 15 30 45 60 Time [min] Figure 51. Graph. Total suspended solid test results for inlet samples and outlet samples for bioretention basin, Bartow County, GA. 52 200 150 Test4_Inlet Test4_Outlet TDS [mg/L] 100 50 0 0 15 30 45 60 Time [min] Figure 52. Graph. Total dissolved solid test results for inlet samples and outlet samples for bioretention basin, Bartow County, GA. Heavy metal concentrations (copper, lead, and zinc) were low, both at the inflow and outflow, with concentrations < 0.2 ppm (Figure 53 - Figure 55). Nonetheless, heavy metal concentration was reduced within the BMP, and when compared to the initial sample concentration at the inlet, concentration of heavy metals in the outlet samples decreased between 60% and 80% as they were filtered by the bioretention basin. 53 0.2 0.15 Test4_Inlet Test4_Outlet Copper [mg/L] 0.1 0.05 0 0 15 30 45 60 Time [min] Figure 53. Graph. Total copper at inlet samples and outlet samples for the bioretention basin, Bartow County, GA. 0.25 0.2 Test4_Inlet Test4_Outlet Lead [mg/L] 0.15 0.1 0.05 0 0 15 30 45 60 Time [min] Figure 54. Graph. Total lead at inlet samples and outlet samples for the bioretention basin, Bartow County, GA. 54 0.2 0.15 Test4_Inlet Test4_Outlet Zinc [mg/L] 0.1 0.05 0 0 15 30 45 60 Time [min] Figure 55. Graph. Total zinc at inlet samples and outlet samples for the bioretention basin, Bartow County, GA. Consistent reduction of nutrient concentrations occurred within the bioretention basin (Figure 56 - Figure 59). In the case of nitrite and total phosphorus, the reduction was approximately 60%, when comparing inlet and outlet concentrations of stormwater runoff. 55 1 Test4_Inlet Test4_Outlet 0.8 Nitrite [mg/L] 0.6 0.4 0.2 0 0 15 30 45 60 Time [min] Figure 56. Graph. Nitrite concentration at inlet samples and outlet samples for bioretention basin, Bartow County, GA. 2.5 Test4_Inlet Test4_Outlet 2 Nitrate [mg/L] 1.5 1 0.5 0 0 15 30 45 60 Time [min] Figure 57. Graph. Nitrate concentration at inlet samples and outlet samples for bioretention basin, Bartow County, GA. 56 10 Test4_Inlet Test4_Outlet 8 Total Nitrogen [mg/L] 6 4 2 0 0 15 30 45 60 Time [min] Figure 58. Graph. Total nitrogen concentration at inlet samples and outlet samples for bioretention basin, Bartow County, GA. 0.2 Test4_Inlet 0.15 Test4_Outlet Total Phosphorus [mg/L] 0.1 0.05 0 0 15 30 45 60 Time [min] Figure 59. Graph. Total phosphorus concentration at inlet samples and outlet samples for bioretention basin, Bartow County, GA. 57 Summary Removal Efficiency: Bioretention basin, Bartow County, GA The bioretention basin monitored in Bartow County was found to be performing well, with removal efficiencies of ~ 60 80% for the contaminants that were tested (Table 10). Table 10. Summary of Removal Efficiencies for Bioretention Basin, Bartow County. Parameter Removal Efficiency (%) Test 4 Turbidity [%] 77 TSS [%] 72 TDS [%] 42 Nitrite [%] 63 Nitrate [%] 24 Total Nitrogen [%] 12 Total Phosphorus [%] 66 Total Copper [%] 61 Total Lead [%] 82 Total Zinc [%] 68 58 BMP IMPACT ON SURFACE WATER AND GROUND WATER SAMPLES IN THE COASTAL PLAIN A series of field tests were performed on BMPs in the coastal plain of Georgia on Skidaway Island, which is a Pleistocene barrier island on the coast of Georgia, lying just southeast of Savannah, Georgia. The island is bordered by saltwater marshes and rivers that drain into the Atlantic Ocean (Figure 60). The island has yearly temperatures that average 10 C (50 F) in the winter and 28 C (82 F) in the summer, with approximately 1.2 meters (3.9 ft) of rain per year. Skidaway Island is a particularly interesting area to test for nutrient concentrations, because it has a population of approximately 8,500 people with six 18-hole golf courses located throughout the island, which results in high application rates of nutrients through grass fertilization. The study monitored surface and groundwater concentrations of nitrogen at multiple locations at the base of filter strips that were located throughout the island. 59 Figure 60. Map. Location of surface water samples taken throughout Skidaway Island, and location of wells for groundwater sampling (at SERF site). Three of the surface water sample locations were taken at the base of slopes, chosen due to similarity to a stormwater filter strip BMP. However, the slopes were chosen in the field tested sites because they were larger than the 6% limit on filter strips (Figure 61). The length of the filter strip (Lf) and slopes angles were measured in the field, and travel time (Tt) was then determined according to the following equation: = 1.250.6250.5 0.338 (eq. 1) where P is the runoff stormwater (3 cm or 1.2 inches, GSWMM, 2014), S is the slope given in percent (V/H), and n is Manning's roughness coefficient, depending on the density of the grass. Two different n coefficients were used to obtain travel time, because grass density changes throughout the seasons and is maintenance dependent. With this 60 information, a model from the Georgia Storm Water Management Manual (GSWMM, 2014) was used to calculate travel time within the BMP. Figure 61. Photos. Sites that were surveyed and the results obtained from surveying: A) MW-12, B) MW-13, and C) PL-09. Alternative Filter Strips in Coastal Plain BMPs Because the slopes exceeded the 6% maximum in the GSWMM (2014) criteria, the current standards were not met; however, the excess slope allows for determination of the impact of a grassed slope on the surrounding surface and ground water. The locations that were tested were chosen for slopes the drained directly into saltwater marshes or 61 saltwater marsh creeks, where concentrations of nitrogen via transport through stormwater runoff was likely. With slopes all greater than the maximum 6% standard, and all but one travel time less than the standard's minimum 5 minutes, these filter strips would be deemed too short according to Georgia stormwater standards, but they were tested in order to quantify nutrient levels in the runoff. Nitrogen was found in both surface and groundwater at the sites. Total inorganic nitrogen (TIN) levels found in surface water ranged from 8 M to 263 M with an average of 74.0 + 3.8 M (Figure 62), with measurable concentrations of nitrate, nitrite, and ammonium detected (Figure 63, Figure 64, nitrite was not mapped as it was only detected in one sample in low concentration (PA-08)). Figure 62. Map-Charts. Values of TIN shown in white, representing percentage of NO3(orange), NH4+ (blue), and NO2- (green). 62 Figure 63. Map-Charts. Nitrate concentrations in surface water around Skidaway Island. 63 Figure 64. Map. Ammonium concentrations in surface water around Skidaway Island. For the tested groundwater samples, TIN concentrations ranged from 11.7 M to 223.9 M with an average of 84.7 + 6.2 M, with no detectable trend (Figure 65). Similarly, nitrate concentrations did not follow a clear trend; however, its concentration (Figure 66) was very similar to that of TIN, indicating that nitrate is the dominant form of nitrogen in the groundwater wells. Ammonium concentrations increased as the wells approached the saltwater marshes (Figure 67). 64 Figure 65. Chart. TIN concentrations found in the groundwater in relation to well location. Distance (0,0) is well B7.5, the well located most closely to the saltwater marshes. Figure 66. Chart. Nitrate concentrations found in the groundwater in relation to well location. Distance (0,0) is well B7.5, the well located most closely to the saltwater marshes. 65 Figure 67. Chart. Ammonium concentrations found in the groundwater in relation to well location. Distance (0,0) is well B7.5, the well located most closely to the saltwater marshes. SUMMARY: BMP FIELD TESTS For the dry swale/sand filter BMP and the bioretention BMP that were tested, the measured contaminant levels at inlet and outlet samples showed that the first flush concentrations for the measured contaminants were notably higher at BMP inflow compared to the BMP outflow at the initiation of the storm event. For the case of the sand filter, contaminant concentrations (including turbidity, total suspended solids, total dissolved solids, nutrients, and heavy metals) were significantly decreased as the stormwater was treated by the sand filter. Based on the measured contaminant concentration in the first flush, removal efficiencies of each BMPs (dry swale with a sand filter and bioretention basin) were determined (Table 7 and Table 10). In addition, stormwater conditions such as antecedent dry condition, precipitation amount, and rainfall intensity influenced the contaminant levels measured at the inlet and outlet of BMPs. Based on the monitored storm events, both types of BMPs have been performing effectively under 66 current field conditions. Detailed results for individual samples taken at each BMP can be found in Appendix A. For the coastal plain hydrogeology, tests were performed on slopes that were selected to mimic filter strips, but with higher slope angles than are constructed for GDOT under Georgia Stormwater rules. Analysis of surface water samples for total inorganic nitrogen revealed that the concentration of TIN measured in surface water samples taken at the base of the slopes increased as slope angle was increased (Figure 68), indicating that removal percentages for nutrients may be negatively impacted by steeper slope angles. As part of this scope of this project, GDOT requested measurement of the hydraulic conductivity of the GDOT specified mix for topsoil for use in BMPs such as bioretention basins; those data are included in Appendix B. Total Inorganic Nitrogen (M) 300 250 200 150 100 50 0 0 5 10 15 20 25 Slope (%) Figure 68. Graph. TIN concentrations in surface water determined as a function of slope angle. 67 CHAPTER 4. STATISTICAL ANALYSIS OF THE EFFICIENCY OF SAND FILTERS FOR TYPICAL POLLUTANTS In order to assess the impact of dimensions of a BMP on field performance, data for evaluating the removal performance and dimensions of sand filters was obtained from the International Stormwater BMP Database (International Stormwater Database). The original objective of the database was to enable long-term scientific research regarding the factors affecting BMP performance. It was developed using a combination of literature review, of studies conducted prior to 1999, along with ongoing data entry from various agencies and independent researchers. The influent and effluent pollutant data specific to sand filters (Table 11) was extracted from the database for the purpose of this analysis. The pollutants selected to assess the performance of the sand filter were pH, turbidity, temperature, total suspended solids (TSS), total dissolved solids (TDS), dissolved and total heavy metals namely, lead (Pb), zinc (Zn), copper (Cu), dissolved and total phosphorus (P), nitrogen and oxides of nitrogen. 68 Table 11. Design Details of Sand Filters Used in the Analysis (International Stormwater Database). BMP Name Permanent Pool Volume Upstream of Filter Media Permanent Pool's Surface Area Permanent Pool's Length Media Filter's Surface Area Type and Depth (or Thickness) of Each Filter Media Layer Appleyard Drive Delaware Sand Filter 13.5 m Foothill SF 216.6 m La Costa PR 285.7 m Eastern SF 115.5327 m Delaware Sand Filter 3.7 m3 7/8 Shasta Maintenance Station Full Sedimentation Austin Sand Filter 105.6218 m 370 m Mountain Gate Sand Filter - Termination 222.3 m SE Landfill Sand Filter 118.8 m 0.0022 ha 0.0102 ha 0.0179 ha 0.0053 ha 71 ft2 0.0056 ha 518 m2 0.011 ha 0.144 ha 24.38 m 12.49 m 14.93 m 8.99 m 7 ft 7.92 m 37 m 11.9 m 0.8 m Lakewood Sand Filter (95) 9.2 m 0.001 ha 8.2 m Escondido 12.2 m Megginis Ck. Sand Filter 163001.5 m 0.002 ha 8.150 ha 24.9 m 1 m Airpark Sand Filter 11.8 m 0.0024 ha 28.8 m 0.0022 ha 20" layer of sand in concrete box 0.0039 ha 0.0072 ha 0.0026 ha 71 ha 0.0031 ha 18 in. sand; geotextile layer; 6 in. gravel 18 in. sand; geotextile layer; 6 in. gravel 18 in. sand; geotextile layer; 6 in. gravel 2" DE #57 stone; 1.5' sand ASTM C-33; Geotextile Fabric 18 in. sand; geotextile layer; 6 in. gravel 280 m2 450 mm sand 108 m2 460 mm sand 0.006ha 0.144 ha 0.0015 ha 0.0027ha 1.7993 ha 0.0022 ha 18 in. sand; geotextile layer; 6 in. gravel The sand filter is constructed in a basin and consist of 0.6 meter of sand over a 0.225 meter bed of graded # 8910 stone. 12 inches of ASSHTO C-33 type sand (d50-0.85mm) underlain by a 12 inch deep fine gravel layer that is drained by a perforated pipe. The coefficient of uniformity is 5.0. 12 in. sand; geotextile layer; 6 in. Gravel 1. Graded sand to a depth of 0.76 m; 2. Filter fabric; 3. 0.91 m dolomite limestone under drain. 17.4 in,The filter media was sand, specified to meet the requirements of ASTM C-33 concrete sand. Sieve analyses in the supply sand yielded the following results: - Effective Size: 0.125 mm - Uniformity Coefficient: 7.8 Parkrose SF - - - 0.0014 ha Sand - 2.8 ft deep 69 FIELD STUDY SITE AND DATA DESCRIPTION The GDOT field site selected for monitoring stormwater runoff in a sand filter BMP was located in the City of Canton, Cherokee County, Georgia on I-575 at SR-20. I-575 is a 50 km (31 mile) long interstate spur located in north Georgia, which connects the Atlanta metropolitan area with the north Georgia mountains. Motivation for the construction of the Canton sand filter was to limit roadway runoff to the habitat of the Cherokee darter fish, which is a threatened species endemic to the Etowah river system in North Georgia. The sand filter was constructed under an agreement between GDOT and the U.S. Fish and Wildlife Service. Detailed BMP description and stormwater monitoring program for this site has been previously described (Bhatt, 2016). The efficiency of typical sand filters for treating individual pollutants was assessed using the pre-existing data (field and database) in 3 ways. The efficiency of the various sand filters was evaluated for each rainfall event for each pollutant using scatter plots. A statistical analysis for the performance of sand filters for mitigating the different pollutants was conducted by computing the box-plots and probability plots of the influent and effluent concentrations. The efficiency of sand filters to mitigate the various categories of pollutants (metals and total solids) was simultaneously evaluated using the k-means (Loyd, 1982) clustering algorithm. The k-means clustering algorithm partitions a dataset [X1, X2, X3,..... Xp] with each `X' being n dimensional into a pre-specified number k 7). The mean temperature of stormwater runoff is 9.2 C (48.6 F) while the lognormally distributed turbidity values have a mean of 47.8 NTU. Figure 69. Charts. Distribution of total suspended and dissolved solids observed at inlet of sand filters (International Stormwater Database). 72 Figure 70. Charts. Distribution of pH, temperature and turbidity observed at inlet of sand filters (International Stormwater Database). The total concentrations (suspended + dissolved) of nutrients as well as heavy metals and the dissolved concentrations of the pollutants are provided in Figure 71 and Figure 72, respectively. The log normal form of distribution for the pollutants implies that the frequency of events where the concentration of pollutants is extremely high is less than those for low concentrations. Based on the concentrations of heavy metals, zinc (Zn) is the most abundant (mean 126.8 g/L), followed by copper (Cu), and lead (Pb). The ratio (R) of mean dissolved to mean total concentration of the pollutants (Zn = 0.61, Cu = 0.38, Pb = 0.12, P = 0.56) suggests that mitigating the pollution caused by different contaminants would require treatment of both suspended and dissolved solids. Mitigation 73 of Zn and P (R > 0.5) would require treatment of dissolved pollutants, whereas most of Cu and Pb could be treated by treating for suspended solids. Figure 71. Charts. Distribution of total nitrogen, NOX, copper, lead, phosphorus and zinc observed at inlet of sand filter (International Stormwater Database). 74 Figure 72. Charts. Distribution of dissolved copper, lead, phosphorus and zinc observed at inlet of sand filter. The efficiency of sand filters in treating suspended and dissolved pollutants was chosen for statistical analysis to quantify the impact of BMP dimensions on field performance and is described using three types of plots. The first type is scatter plots, which describe the respective effluent concentrations of different pollutants given the influent concentrations for different storm events. The second type is box plots, which describe the distribution of influent and effluent concentrations of different pollutants and the third type are the probability plots of the influent and effluent concentrations, which statistically explain the efficiency of sand filters in treating pollutants based on different influent concentrations. The scatter plots delineate the effluent concentrations corresponding to influent concentrations for different storm events. The scatter plot for the total suspended and 75 dissolved solids (Figure 73) reveals the high removal efficiency of sand filters (data points below the 1:1 line) in treating suspended solids, but not dissolved solids. This follows from the typical understanding of sand filter treating solids through the process of filtration and removal of solid particles. The absence of a correlation between the influent and effluent concentration of suspended solids signifies that regardless of the input suspended solid concentration, the sand filter performs well for removal. The pH of the effluent is typically a little higher (Figure 74) than the influent concentration and is highly correlated, implying that the effect on pH treatment by the sand filter is dependent on the input pH. The temperature of the effluent concentration of the stormwater is similar to the influent temperature except for when the temperature of the influent concentrations is relatively high (~20C). This could potentially be attributed to detention basins which allow time for the cooling of incoming water or thermal diffusion in the filter media indicating filter media acts as a heat sink. Similar to results for suspended solids, turbidity of effluent is also not correlated to influent concentrations and gets treated relatively well. This occurs because most of the turbidity is caused as a result of suspended solids. The scatter plots for the nutrient concentrations reveal that a majority of the total phosphorus is mitigated by the sand filter whereas neither dissolved phosphorus, total NOX or nitrogen is mitigated. On the contrary, the amounts of NOX compounds tended to increase slightly in the effluent. NOX represents the quantity of nitrite and nitrate together. The increase in effluent concentration of NOX could potentially be due to oxidation of nitrogen in the stormwater runoff before reaching the BMP outlet. The results for metals show that total and dissolved copper is not mitigated well by the sand filter whereas zinc and lead (total and dissolved) is mitigated relatively well. 76 Figure 73. Charts. Scatter plot of influent v/s effluent concentration of total suspended and dissolved solids in stormwater runoff. Figure 74. Charts. Scatter plot of influent vs effluent concentration of total pH, temperature and turbidity in stormwater runoff. 77 x Figure 75. Charts. Scatter plot of influent vs effluent concentration of total and dissolved phosphorus, total NOx and Nitrogen in stormwater runoff. 78 Figure 76. Charts. Scatter plot of influent vs effluent concentration of total and dissolved metals in stormwater runoff. Kruskal-Wallis test (or the non-parametric equivalent of ANOVA) was employed to assess a statistically significant difference in median values of influent and effluent concentrations of the various pollutants. The p-values for the test are provided in 79 Table 12 while the corresponding box plots are provided in Figure 77 - Figure 80. It was found that a statistically significant difference (p-values < 0.05) in median concentrations was found in all parameters except temperature, total nitrogen, dissolved phosphorus, and dissolved copper. While the median concentration of total suspended solids decreased, the median concentration of total dissolved solids increased slightly in the effluent (Figure 77). This implies the effectiveness of the sand filters in treating suspended solids but not dissolved solids. The increase in effluent concentration of dissolved solids could potentially occur because of the dissolution of some solids trapped either within the sand filter or detention basins during the transit of the influent stormwater to the BMP outlet. The pH values increased slightly in the effluent, but the turbidity decreased. The decrease in turbidity follows the effective removal of suspended solids by the BMP. The increase of pH on the other hand implies some reduction occurring in the system either during the transit between inlet and outlet or in the detention basin. 80 Table 12. P-values of the Kruskal-Wallis Test for Influent Versus Effluent Concentrations. Parameter TSS TDS TpH Temperature Turbidity Total Nitrogen Total NOX Total Phosphorus Dissolved PThotoaslpChooprpuesr Dissolved Copper Total Zinc Dissolved Zinc DToi tal Lead Dissolved Lead p-value <0.001 <0.001 0.007 0.917 <0.001 0.156 <0.001 <0.001 0.633 <0.001 0.113 <0.001 <0.001 <0.001 0.0153 The total copper concentration decreased, but there was no significant difference between the dissolved and total copper concentration. The concentration of the other two metals decreased significantly in terms of both total metals, as well as dissolved metal. Total phosphorus and nitrogen decreased slightly in the effluent, but there were no statistically significant differences between the NOX and dissolved phosphorus concentrations. 81 Figure 77. Diagram. Box plot of influent and effluent concentration of total and dissolved metals in stormwater runoff. Figure 78. Diagram. Box plot of influent and effluent concentration of total pH, temperature and turbidity in stormwater runoff. 82 Figure 79. Diagram. Box plot of influent and effluent concentration of total and dissolved metals in stormwater runoff. 83 Figure 80. Diagram. Box plot of influent and effluent concentration of total and dissolved phosphorus, total NOX and nitrogen in stormwater runoff. The probability plots (Figure 81 - Figure 84) describe the fraction (y-axis) of times, a value less than the corresponding concentration (x-axis) was observed. Distinct influent and effluent probability curves reflect differences between the influent and effluent concentrations, whereas overlapping curves reflect no differences. For example, if the 0.95 (y-axis) values correspond to 100 mg/L on the influent probability curve, it implies that 95% of the time, the influent concentrations remain less than 100 mg/L. If the effluent curve lies to the left of the influent curve, it represents effective mitigation. Figure 81 shows that the total suspended solids are effectively mitigated (effluent curve to the left of the influent probability curve), but the total dissolved solids are not. In fact statistically, the 84 effluent concentration of the total dissolved solids remains higher than the influent (curve lying to the right of the influent curve). Total turbidity is also effectively mitigated (Figure 82), while the probability curves for temperature are almost identical overlapping, implies that there is no effect of sand filters on the thermal pollution. The sand filter does little to alter the pH of the stormwater runoff. It was also observed that the pH of the influent is greater than ~7 over 25% of the times whereas the pH is higher than 7 in the effluent for more events than it is the influent. This implies that the influent is more acidic than the effluent. The total concentration of metals (Figure 83) reflects some level of mitigation of metals in stormwater runoff through separation between the probability curves for the influent and effluent. The dissolved concentration of zinc also shows distinct mitigation as a result of the sand filter. However, there is no change in the concentration probability curves of dissolved copper and lead, implying the ineffectiveness of the sand filter for the two metals statistically. The sand filter performs variably with respect to nutrients (Figure 84). The total concentration of phosphorus is effectively mitigated through the sand filter whereas the dissolved phosphorus is not. On the other hand, the concentration of NOX is statistically higher in the effluent than in the influent. The concentration of total nitrogen as N on the other hand is unaffected by the presence of the sand filter. 85 Figure 81. Graph. Probability plot of influent and effluent concentration of total and dissolved solids in stormwater runoff. Figure 82. Graph. Probability plot of influent and effluent concentration of total pH, temperature and turbidity in stormwater runoff. 86 Figure 83. Graph. Probability plot of influent and effluent concentration of total and dissolved metals in stormwater runoff. 87 x Figure 84. Graph. Probability plot of influent and effluent concentration of total and dissolved phosphorus and total NOX and nitrogen in stormwater runoff. SAND FILTER STATISTICAL DESIGN PARAMETERS The previous analysis focuses on analysing the impact of sand filters on each pollutant individually. In order to evaluate the efficiency of the sand filter to simultaneously mitigate various pollutants, a k-means cluster analysis was set up for the cleaning efficiency of sand filter for total metals (copper, lead and zinc). The clusters were iteratively computed until a cluster was formed for sand filters effectively mitigating all three metals. The design parameters of the sand filters comprising of this cluster were then 88 evaluated to provide an appropriate design criterion for sand filters that conforms to the cleaning efficiency represented by the cluster. The design parameters of the sand filters as well as the cluster locations are shown in Figure 85. Figure 85. Charts. Clusters of cleaning efficiency of sand filters for total metals and design parameters of the sand filters for each cluster. The median of the removal efficiency in different clusters is provided in Table 13. Cluster 2 delineates the sand filters that perform relatively better than the others in mitigating metal pollutants. 89 Table 13. Median Cleaning Efficiency of Clusters (Metals). Metal Copper Lead Zinc Cluster 1 0.47 0.81 0.2 Cluster 2 0.62 0.87 0.91 Cluster 3 -0.02 0.84 0.87 Cluster 4 0.59 0 0.90 The median values of the design dimensions of the sand filters comprising different clusters are provided in Table 14. The results indicate that a single design parameter cannot be outlined as controlling the removal efficiency of the sand filter. For example, cluster 2 and 4 have similar pool area and depth, yet because of differences in the pool area and depth, the cleaning efficiency of the 2 clusters is different. The best combination of design parameters based on the data analysed is highlighted in Table 14. Table 14. Median of Design Parameters for Different Clusters. Design parameter Cluster 1 Cluster 2 Cluster 3 Cluster 4 Filter Area (m2) 31.99 280 56.99 280 Filter depth (cm) 45.72 45.72 45.72 45.72 Pool area (m2) 56.02 102.01 102.01 518 Pool depth (m) 11.87 7.92 11.88 2.13 A similar cluster analysis (Figure 86) was computed for the total suspended and dissolved solids. 90 Figure 86. Charts. Cluster analysis and design parameters of sand filters with respect to total solids. The median values of the removal efficiency and the design parameters of the corresponding clusters are provided in Table 15 and Table 16, respectively. It is seen that cluster 2 performs the best when assessing the cleaning efficiency of sand filters for total suspended and total dissolved solids simultaneously. Table 15. Median Cleaning Efficiency of Clusters (Solids). Metal Cluster 1 Cluster 2 Cluster 3 Cluster 4 TSS 0.19 0.88 0.91 0.92 TDS -0.41 0.26 -0.85 -0.24 91 Table 16. Median of Design Parameters for Different Clusters. Design parameter Filter Area (m2) Filter depth (cm) Pool area (m2) Pool depth (m) Cluster 1 280 45.72 27.03 2.13 Cluster 2 280 45 518 37 Cluster 3 31.99 45.72 56.02 2.13 Cluster 4 280 45.72 102.01 7.92 In terms of median values of design parameters, the values of cluster 2 for total solids and cluster 4 for total metals are similar and vice-versa. Therefore, by adopting the median values of cluster 2 (for total metals that corresponds to cluster 4 for total solids) as the design criteria for the sand filter, it may provide efficient cleaning of metals and high efficiency for removal of total suspended solids but not for total dissolved solids. In contrast, if design parameters corresponding to cluster 2 for total solids are used to design BMPs, an overall cleaning efficiency for total solids may be obtained but the sand filter will perform poorly for total metal concentrations in the stormwater runoff. 92 CHAPTER 5. RECOMMENDATION FOR GUIDANCE ON SELECTION OF STORMWATER BMPS Specification of a BMP is a complex balance of site hydrogeology, rainfall intensity, contaminant loading, right-of-way restrictions, BMP longevity, and maintenance requirements. Monitoring of the field performance of currently operating GDOT BMPs has demonstrated that the devices are functioning well, with acceptable contaminant removal percentages. Statistical analysis of the performance of sand filters, including the Canton sand filter in addition to BMPs from the International Stormwater Database, demonstrated that the dimensions of a sand filter do impact the contaminant removal percentages; however, the optimum dimensions were different for different contaminants (i.e., heavy metals versus suspended solids). Statistically, the data clearly demonstrated that sand filters are highly effective at solids, turbidity, some metal and total phosphorus removal, but less effective at removing dissolved constituents (solids, nutrients, and metals). Currently, GDOT specification criteria follow a hierarchy of 1) stormwater requirements, 2) site and soil constraints, and 3) BMP feasibility: 93 Figure 87. Diagram. BMP selection process flowchart (Figure 10.5-1 from GDOT, 2018). Based on the review of maintenance and statistical performance, it is recommended that explicit criteria of BMP longevity, and short/intermediate versus long term maintenance burden be added to the BMP selection process flow chart 94 (Figure 10.5-1) (Table 17). For the BMPs that are commonly implemented on GDOT right-of-way, service life is an important criterion, with the optimal BMP having low maintenance burden coupled with a service life of many decades. Additionally, some BMPS, like sand filters and infiltration trenches have relatively minimal maintenance in the short or intermediate terms, but substantial maintenance in the long term. Distinguishing these periods of low versus high levels of required maintenance is important to developing an accurate assessment of the lifecycle cost of operation. Table 17. BMP Longevity and Maintenance Burden. BMP Filter Strips Grass Channels Bioslopes Enhanced Dry Swales Bioretention Basins Enhanced Wet Swales Infiltration Trenches Sand Filters Dry Detention Basins Wet Detention Ponds Stormwater Wetlands Level 2 Stormwater Wetlands Level 1 OGFC Longevity (years) High High Medium High Medium High High High High High High High Low Short-term / Intermediate term Maintenance Burden Low Low Medium Medium Medium Low Low Low Low Low Medium Medium Low Long Term Maintenance Burden Low Low Medium Medium Medium Low High High Low Low Medium Medium Low 95 CHAPTER 6. CONCLUSIONS AND RECOMMENDATIONS The work performed in this study demonstrated that selected monitored BMPs, a dry swale with sand filter in Forsyth County and bioretention basin in Bartow County were performing within their anticipated level of contaminant removal and were operating with State of Georgia standards. During a stormwater event, both BMPs demonstrated pronounced first flush peak concentrations, with reduction in the inflow contaminant levels within one hour of stormwater flow. The combination dry swale / sand filter successfully treated turbidity, total suspended solids, total dissolved solids, some nutrients, and some heavy metals. The bioretention basin also successfully treated contaminants and showed better performance in the removal of nutrients when compared to the sand filter. The impact of high slope angles on the concentrations of nutrients was monitored in the coastal plain hydrogeology, and analysis of surface water samples for total inorganic nitrogen revealed that the concentration of TIN measured in surface water samples taken at the base of the slopes increased as slope angle was increased indicating that removal percentages for nutrients may be negatively impacted by steeper slope angles. An in-depth statistical analysis of the performance of sand filters, based on an extensive database collected at the GDOT sand filter in Canton, GA and data drawn from the International Stormwater Database demonstrated that optimization of BMP dimensions for contaminant removal is a complex, multi-constrained problem. Dimensions that were optimized for metal removal did not perform as well for solids removal, and vice-versa; consequently, selection criteria based on optimized BMP dimensions is not feasible at this point in time, due to the complexity of other factors such as regional hydrogeology and precipitation events. Finally, it is recommended that the selection criteria flowchart for BMPs on GDOT right-of-way be expanded to include criteria for longevity, as well as distinction between short/intermediate term maintenance versus long term maintenance burden. 96 APPENDIX A: DETAILED SAMPLE CONCENTRATION DATA FOR TESTED STORMWATER BMPS 97 Table 18. Summary of Stormwater Event 1 Results (Dry swale with a sand filter). Inlet Outlet Parameter Sample 1 (0 min) Sample 3 (30 min) Sample 5 (0 min) Sample 6 (15 min) Sample 7 (30 min) Sample 8 (45 min) Turbidity [NTU] 34.08 32.76 12.92 13.18 12.24 11.92 TSS [mg/L] 43.2 11.3 3.6 3.9 3.7 2.0 TDS [mg/L] 108.14 71.40 61.48 59.29 55.29 59.78 Nitrite [mg/L] 0.255 0.205 0.037 0.033 0.033 0.032 Nitrate [mg/L] 0.829 0.641 0.611 0.558 0.513 0.529 Total Nitrogen [mg/L] 4.220 3.714 3.583 3.504 3.532 3.520 Total Copper [mg/L] 0.0621 0.0196 0.0149 0.0161 0.0135 0.0126 Total Lead [mg/L] 0.1819 0.0379 0.0246 0.0173 0.0171 0.0080 Total Zinc [mg/L] 0.0465 0.0183 0.0101 0.0031 0.0030 0.0032 98 Table 19. Summary of Stormwater Event 2 Results (Dry swale with a sand filter). Inlet Outlet Parameter Sample 1 (0 min) Sample 2 (15 min) Sample 3 (30 min) Sample 4 (45min) Sample 5 (0 min) Sample 6 (15 min) Sample 7 (15 min) Sample 8 (45min) Turbidity [NTU] 97.5 81.26 85.45 77.9 7.08 5.95 12.46 7.88 TSS [mg/L] 40.50 33.25 36.00 25.50 1.00 4.00 1.75 0.50 TDS [mg/L] 165.43 193.64 177.88 118.18 69.42 49.85 58.90 64.11 Nitrite [mg/L] 0.87 0.78 0.61 0.71 0.14 0.12 0.06 0.06 Nitrate [mg/L] 2.32 1.82 1.64 1.65 1.09 0.98 0.95 0.96 Total Nitrogen 8.03 7.11 6.85 6.93 6.7 6.61 6.57 6.63 [mg/L] Total Phosphorus [mg/L] 0.160 0.155 0.134 0.145 0.020 0.030 0.013 0.017 Total Copper [mg/L] 0.101 0.048 0.044 0.038 0.033 0.025 0.027 0.025 Total Lead [mg/L] Total Zinc [mg/L] 0.225 0.060 0.047 0.035 0.03 0.025 0.020 0.027 0.090 0.040 0.035 0.031 0.021 0.023 0.017 0.021 99 Table 20. Summary of Stormwater Event 3 Results (Dry swale with a sand filter). Inlet Outlet Parameter Sample 1 (0 min) Sample 2 (15 min) Sample 3 (30 min) Sample 4 (composite)1 Sample 5 (0 min) Sample 6 (15 min) Sample 7 (15 min) Sample 8 (composite)1 Turbidity [NTU] 50.38 47.62 36.03 15.93 8.74 7.24 9.42 12.15 TSS [mg/L] 54.75 52.00 42.75 27.50 8.67 3.33 5.33 7.00 TDS [mg/L] 85.26 83.97 81.21 51.56 34.67 26.67 38.00 45.09 Nitrite [mg/L] 0.37 0.30 0.34 0.22 0.12 0.09 0.12 0.83 Nitrate [mg/L] 1.35 1.11 1.05 1.07 0.78 0.82 0.73 1.53 Total Nitrogen 6.13 5.25 5.18 5.31 4.92 4.85 4.91 6.62 [mg/L] Total Phosphorus [mg/L] 0.052 0.044 0.062 0.023 0.012 0.012 0.009 0.015 Total Copper [mg/L] 0.070 0.028 0.025 0.027 0.020 0.021 0.019 0.035 Total Lead [mg/L] 0.195 0.035 0.030 0.027 0.020 0.015 0.017 0.030 Total Zinc [mg/L] 0.053 0.015 0.017 0.016 0.010 0.012 0.013 0.020 1 Composite sample (45min 240min) 100 Table 21. Summary of Stormwater Event 4 Results (Bioretention basin). Inlet Outlet Parameter Sample 1 (0 min) Sample 2 (15 min) Sample 3 (30 min) Sample 4 (composite)1 Sample 5 (0 min) Sample 6 (15 min) Sample 7 (15 min) Sample 8 (composite)1 Turbidity [NTU] 38.41 23.53 23.66 26.11 8.80 8.63 8.87 9.14 TSS [mg/L] 37.50 18.75 12.75 19.25 12.25 7.50 10.25 11.50 TDS [mg/L] 147.42 82.72 65.96 38.79 74.32 91.08 90.29 84.71 Nitrite [mg/L] 0.229 0.158 0.168 0.122 0.073 0.083 0.069 0.119 Nitrate [mg/L] 0.758 0.635 0.623 0.630 0.593 0.578 0.585 0.562 Total Nitrogen [mg/L] 4.032 3.753 3.802 3.766 3.598 3.577 3.483 3.506 Total Phosphorus [mg/L] 0.044 0.033 0.032 0.030 0.009 0.019 0.016 0.016 Total Copper [mg/L] 0.053 0.028 0.025 0.026 0.021 0.023 0.018 0.020 Total Lead [mg/L] 0.158 0.048 0.038 0.042 0.030 0.032 0.028 0.026 Total Zinc [mg/L] 0.038 0.021 0.019 0.017 0.013 0.015 0.010 0.011 1 Composite sample (45min 240min) 101 APPENDIX B: HYDRAULIC CONDUCTIVITY OF GDOT TOPSOIL MIX 102 DETERMINATION OF HYDRAULIC CONDUCTIVITY OF GDOT TOPSOIL MIX GDOT requested testing of the specified topsoil mix for development of the Supplemental Specification Section 893-Miscellaneous Planting Materials. A laboratory based investigation was performed with a soil created to meet the grain size distributions of the specified non-organic portion of the topsoil (Table 22 and Figure 88). For the finegrained soil, 10% of the material was kaolinite (clay) and 90% of the material was Piedmont soil (silt). Table 22. GDOT Grain Size for Topsoil (Inorganic). Sieve Size Percent Passing by Weight Passing 2 inch (50mm) 100 Passing 1- inch (37.5 mm) 100 Passing No.10 (2mm) sieve 83 Passing No.40 (425 um) sieve 60 Passing No. 60 (250 um) sieve 45 Passing No. 200 (75 um) sieve 18 Clay Size (<2 um) 0 103 Figure 88. Graph. The grain size distribution curve for the inorganic portion of the topsoil. Materials and Methods The inorganic materials used for the test included gravel, Ottawa 20-30 sand, FS 50-70 sand, F 75 sand, Piedmont soil, and kaolinite. These soils were mixed together in specified proportion and sieved to create a specimen for testing that met the Department's grain size requirements. In addition to the soil minerals, organic material was added to the topsoil mix (7.5% by mass). Commercial cow manure was purchased and used to add organic carbon to the controlled soil mixture. Organic carbon content was determined using a total organic carbon analyzer (Shimadzu). A series of hydraulic conductivity tests were performed in a falling head hydraulic conductivity test, with flexible membrane (ASTM D5084). All test specimens were back pressure saturated to dissolve discrete air bubbles into the water phase. The B coefficient was monitored until it reached a level of 0.95, while adjusting target effective stress during back pressure saturation. All tests were performed at effective stress levels of 5 psi, 10 psi, and 15 psi. 104 Results The cow manure was sampled eight times to obtain a representative measure of its organic carbon content. Organic carbon content ranged from 10.7% - 20.3%, with an average value of 15.8%, which was used to determine the mass of organic included in the topsoil for hydraulic conductivity testing (Table 23). Six samples of cow manure were also tested to determine moisture content in the as received condition, resulting in an average moisture content of 44% by mass (Table 24). Table 23. Organic Carbon Content Commercial Cow Manure. Sample ID 1 2 3 4 5 6 7 8 Average Median Standard Deviation Maximum Minimum Organic Carbon Content (Current) 16.9% 19.4% 12.4% 15.7% 10.7% 14.7% 16.6% 20.3% 15.8% 16.2% 3.01% 20.3% 10.7% 105 Table 24. Moisture Content Commercial Cow Manure. Sample ID 1 2 3 4 5 6 Average Moisture Content (%) 42.7% 45.8% 46.5% 42.8% 43.1% 43.1% 44.0% A series of samples were prepared for hydraulic conductivity testing of the finegrained soils. In these soils, it is especially critical to ensure saturation of the samples. If the sample is not saturated, unconservative values of hydraulic conductivity may be reported. Tests were performed to ensure the samples reached a B value equal to 0.95 (Figure 89). Hydraulic conductivity for the soil with only inorganic minerals was approximately 1 2 x 10-5 cm/sec. Inclusion of the organic carbon phase increased the hydraulic conductivity of the specified topsoil by approximately one order of magnitude to 2- 5 x 10-4 cm/sec (Figure 90 - Figure 91). 106 B-Coefficient (-) 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 50 100 150 200 250 Total Time Elapsed (min) Figure 89. Graph. B value measured for specified topsoil hydraulic conductivity sample. 1.E-03 1.E-04 Sample 1 Sample 2 Sample 3 Hydraulic Conductivity (cm/s) 1.E-05 0 5 10 15 Confining Pressure (psi) Figure 90. Graph. Hydraulic conductivity of specified topsoil with only inorganic minerals (no organic matter included). 107 1.E-03 Hydraulic Conductivity (cm/s) 1.E-04 1.E-05 0 Sample 4 Sample 5 Sample 6 5 10 15 Confining Pressure (psi) Figure 91. Graph. Hydraulic conductivity of specified topsoil with both inorganic minerals and organic carbon. 108 ACKNOWLEDGEMENTS The author is pleased to acknowledge the contributions of Dr. Aditya Bhatt, Mr. Young Suk Jung, Ms. Lynnae Luettich, and Ms. Hejintao Hunag who worked as graduate research assistants on this project. Thanks to Mr. Jon D. Griffith, P.G., P.E. (deceased), Mr. Binh Bui, Mr. Brad McManus, P.E., and Mr. Dylan Eagleton for their ongoing dedication and significant contributions in the formulation of this project and throughout the course of this work. 109 REFERENCES CFR, U., Code of Federal Regulations Title 40: Protection of Environment, Part 136 Guidelines establishing test procedures for the analyses of pollutants, Appendix B to Part 136definition and procedure for the determination of Method Detection Limit rev. 1.11, 2009. Deletic, Ana & Fletcher, Tim. (2006). Performance of Grass Filters Used for Stormwater Treatment -- A Field and Modelling Study. Journal of Hydrology. 317. 261-275. 10.1016/j.jhydrol.2005.05.021. Fletcher, T.D., Peljo, L., Fielding, J., Wong, T.H.F., and Weber, T. (2012). The Performance of Vegetated Swales for Urban Stormwater Pollution Control, Ninth International Conference on Urban Drainage (9ICUD), Global Solutions for Urban Drainage, ASCE. Georgia Department of Transportation (2018). Drainage Design for Highways, Revision 3.4, 8/18/2018. Hunt, W.F., Smith, J.T., Jadlocki, S.J., Hathaway, J.M., and Eubanks, P.R. (2008). Pollutant Removal and Peak Flow Mitigation by a Bioretention Cell in Urban Charlotte, N.C., Journal of Environmental Engineering, ASCE 2008, 134(5): 403408. International Stormwater Database, Developed by Wright Water Engineers, the American Society of Civil Engineers (ASCE)/Environmental and Water Resources Institute (EWRI), the American Public Works Association (APWA), the Federal Highway 110 Administration (FHWA), and U.S. Environmental Protection Agency (EPA). Editor. Li, H., Li, K., and Zhang, X. (2016). Performance evaluation of Grassed swales for stormwater pollution control, 12th International Conference on Hydroinformatics, HIC 2016, doi: 10.1016/j.proeng.2016.07.481. Lloyd, S.P., Least squares quantization in PCM. Information Theory, IEEE Transactions on, 1982. 28(2): p. 129-137. Lucke, T. and Nichols, P.W.B. (2015). The pollution removal and stormwater reduction performance of street-side bioretention basins after ten years in operation, Science of The Total Environment, Volume 536, Pages 784-792, doi.org/10.1016/j.scitotenv.2015.07.142. Mangangka, I.R., Liu, A., Egodawatta, P., and Goonetilleke, A. (2015). Performance characterisation of a stormwater treatment bioretention basin, Journal of Environmental Management, Volume 150, Pages 173-178, doi.org/10.1016/j.jenvman.2014.11.007. Munoz-Carpena, R., Weblink abe.ufl.edu/carpena/vfsmod/FOCUS/VFSMOD_Pmassbalance1.ppt, Accessed on 1/26/2018. National Academies of Sciences, Engineering, and Medicine. 2014. Long-Term Performance and Life-Cycle Costs of Stormwater Best Management Practices. Washington, DC: The National Academies Press. https://doi.org/10.17226/22275. New Jersey Department of Environmental Protection. (2020). New Jersey Stormwater Best Management Practices Manual, obtained from: 111 http://www.njstormwater.org/bmp_manual/NJ_SWBMP_9.12.pdf, last accessed August 30, 2020. Patty, L., Real, B., and Gril, J.J. (1997). The Use of Grassed Buffer Strips to Remove Pesticides, Nitrate and Soluble Phosphorus Compounds from Runoff Water, Pest Management Science, Volume 49, Issue 3, Pages 243251. Robinson, C.A., Ghaffarzadeh, M. and Cruse, R.M. (1996). Vegetative filter strip effects on sediment concentration in cropland runoff, Journal of Soil and Water Conservation May/June 1996, vol. 51, no. 3, 227-230. Wang, J., Chua, L.H.C., and Shanahan, P. (2017). Evaluation of pollutant removal efficiency of a bioretention basin and implications for stormwater management in tropical cities, Vol. 3, 78-91. Yu, S.L., Kuo, J.-T., Fassman, E.A., and Pan, H. (2001). Field test of grassed-swale performance in removing runoff pollution, Journal of Water Resources Planning Management, 2001, 127(3): 168-171. 112