EFFECTIVENESS OF WASTEWATER LAND APPLICATION : MONITORING AND MODELING By Younsuk Dong A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Biosystems Engineering Doctor of Ph ilosophy 2018 ABSTRACT EFFECTIVENESS OF WASTEATER LAND APPLICATION : MONITORING AND MODELING By Younsuk Dong Wastewater land application has been used for decades because of its low cost, energy use, and maintenance requirements, compared to a conventio nal wastewater treatment system. The performance of treatment depends on the hydraulic and organic wastewater loadings, soil characteristics, and soil conditions. Understanding the complexity of soil is important. The aerobic or anaerobic condition of the soil may result in nitrate leaching and metal mobilization into groundwater, respectively. Currently, design criteria are generally based on empirical relationships, which do not adequately consider site and waste - specific conditions. Because organic and h ydraulic loadings are generally fixed based on production, dosing is the only operational parameter that can be adjusted to enhance treatment for site - specific conditions. In this study, an evaluation of domestic and food processing wastewater s land applic ation systems w ere performed including examining their benefits, effectiveness , and techniques for modeling. Monitoring strategies at the demonstration site showed the viability of using land application to treat food processing wastewater and helps in mak ing an operation decision. The HYDRUS C onstructed W etland 2D (CW2D) model was successfully calibrated and validated using data from laboratory experiments. The modeling results showed that m ost of the COD removal in a domestic wastewater land application s ystem occurs within a 30.5 cm ( 1 ft ) depth for a sandy loam soil. Increasing the dosing frequency was effective in slightly reducing the COD effluent concentration. A n increase in nitrate removal by changing dosing frequency while providing sufficient carb on was found to be possible . Copyright by YOUNSUK DONG 201 8 iv ACKNOWLEDGMENTS I would like to thank everyone for their support and I would not be able to start and complete this study without their help. I truly appreciate to my advisor , Dr. Steven Safferman who gave m e the best guidance through the study . Without his encouragement, I probably would not have achieve d this stage. He always kindly provide his endless supports. Thanks also to my committee, Dr. Amirpouyan Nejadhashemi, Dr. Dawn Reinhold, and Dr. Wei Zhang. They have kindly given their time and expertise to help me to complete this research. Thanks to Steve Marquie for helping on the sensor troubleshooting. And I would like to thank to John Hruby, Dan Wurm, and Dave Bratt f or their kindly support for this study. In addition, I would like to thank to my research team, Mr. Steve Miller, Louis Faivor, Jason Smith, Ethan Nussdorfer, Emily Campbell, Shuai Zhang, Daniel Bhur, Davis Roeser, Caitlin Knedgen, Jessica Hauda, Anna Ras chke, and Kiran Lantrip. study would not have been completed. Lastly, t hanks to my family for their endle ss love and support, my parents, parents in law, my sister, my wife Yeoreum Lee, my son Nathaniel Dong, and my dau ghter Elli Dong. v TABLE OF CONTENT S LIST OF TABLES ................................ ................................ ................................ ................................ .... vii LIST OF FIGURES ................................ ................................ ................................ ................................ ... ix KEY TO ABBREVIATIONS ................................ ................................ ................................ ................. xiv Chapter 1. Introduction ................................ ................................ ................................ ....................... 1 1.1. Hypothesis/research question ................................ ................................ ......................... 4 1.2. Objective ................................ ................................ ................................ ................................ ... 4 1.3. Dissertation framework ................................ ................................ ................................ ..... 6 Chapter 2. Literature review ................................ ................................ ................................ ............. 8 2.1. Background of wastewater land application ................................ ............................... 8 2.2. Wastewater treatment technologies ................................ ................................ .............. 9 2.3. Environmental impacts from wastewater land application ................................ 11 2.4. Nitrogen processes in wastewater land application ................................ .............. 12 2.5. Impact o f loadings for wastewater land application treatment ......................... 14 2.6. Current design criteria ................................ ................................ ................................ ...... 16 2.7. Overview of subsurface flow soil mode ling ................................ ............................... 19 Chapter 3. Domestic wastewater land application ................................ ................................ . 25 3.1. Introduction ................................ ................................ ................................ .......................... 25 3.2. Materials and methods ................................ ................................ ................................ ...... 27 3.2.1. HYDRUS CW2D modeling ................................ ................................ ................................ ..... 28 3.2.1.1. Governing equation ................................ ................................ ................................ ................................ ................ 28 3.2.1.2. HYDRUS CW2D component and processes ................................ ................................ ................................ .. 29 3.2.2.3. Limitations of HYDRUS CW2D ................................ ................................ ................................ ........................... 32 3.2.2 .4. Model calibration and validation ................................ ................................ ................................ ...................... 32 3.2.2.4.1. Laboratory experiment ................................ ................................ ................................ ................................ . 33 3.2.2.4.2. Goodness of fit ................................ ................................ ................................ ................................ ................... 35 3.2.2.4.3. Calibration and validation procedure ................................ ................................ ................................ .... 37 3.2.2.5. Scenarios ................................ ................................ ................................ ................................ ................................ ..... 43 3.3. Result and di scussion ................................ ................................ ................................ ........ 46 3.3.1. Monitoring of domestic wastewater land application ................................ ............... 46 3.3.2. HYDRUS CW2D modeling ................................ ................................ ................................ ..... 47 3.3.2.1. Model calibration and validation ................................ ................................ ................................ ...................... 47 3.3.2.2. Scenario capacity of wastewater land application ................................ ................................ ................ 53 3.3.2.3. Scenario treatment performance enhancement ................................ ................................ .................... 55 3.4. Conclusion ................................ ................................ ................................ ............................. 62 Chapter 4. Food processing wastewater l and application ................................ .................... 64 4.1. Introduction ................................ ................................ ................................ .......................... 64 4.2. Materials and methods ................................ ................................ ................................ ...... 69 4.2.1. Monitoring of food processing wastewater land application ................................ . 69 4.2.1.1. Background of demonstration site ................................ ................................ ................................ .................. 70 4.2.1.2. Monit oring strategies ................................ ................................ ................................ ................................ ............ 72 vi 4.2.1.2.1. Hydraulic and organic loadings ................................ ................................ ................................ ................. 72 4.2.1.2.2. Soil sensor cluster ................................ ................................ ................................ ................................ ........... 73 4.3.1.2.3. Groundwater monitoring ................................ ................................ ................................ ............................. 77 4.2.1.3. Site evaluation ................................ ................................ ................................ ................................ ........................... 78 4.2.1.3.1. Visual observation ................................ ................................ ................................ ................................ ........... 7 8 4.2.1.3.2. Soil texture ................................ ................................ ................................ ................................ .......................... 79 4.2.1.3.3. Soil compaction ................................ ................................ ................................ ................................ ................ 79 4.2.1.3.4. Infiltration ................................ ................................ ................................ ................................ ........................... 80 4.2.1.3.5. Uniformity ................................ ................................ ................................ ................................ ........................... 82 4.2.1.3.6. Localized high water table condition ................................ ................................ ................................ ...... 82 4.2.3. HYDRUS CW2D modeling ................................ ................................ ................................ ..... 83 4.2.3.1. Scenarios ................................ ................................ ................................ ................................ ................................ ..... 84 4.3. Result and discussion ................................ ................................ ................................ ........ 85 4.3.1. Monitoring of food processing wastewater land application ................................ . 85 4.3.1.1. Monitoring strategies ................................ ................................ ................................ ................................ ............ 85 4.3.1.1.1. Hydraulic and organic loading ................................ ................................ ................................ ................... 85 4.3.1.1.2. Soil sensor cluster ................................ ................................ ................................ ................................ ........... 86 4.3.1.1.3. Groundwater monitoring ................................ ................................ ................................ ............................. 96 4.3.1.2. Site evaluation ................................ ................................ ................................ ................................ ........................ 106 4.3.1.2.3. Visual observation ................................ ................................ ................................ ................................ ........ 107 4.3.1 .2.2. Soil texture ................................ ................................ ................................ ................................ ....................... 109 4.3.1.2.3. Soil compaction ................................ ................................ ................................ ................................ ............. 110 4.3.1.2.4. Infiltration ................................ ................................ ................................ ................................ ........................ 114 4.2.1.2.5. Uniformity ................................ ................................ ................................ ................................ ........................ 115 4.3.1.2.6. Localized high water condition ................................ ................................ ................................ .............. 117 4.3.3. HYDRUS CW2D modeling ................................ ................................ ................................ ... 120 4.3.3.1. Model calibration and validation ................................ ................................ ................................ ................... 120 4.3.3.2. Scenario Treatment performance enhancement ................................ ................................ ................ 124 4.4. Conclusion ................................ ................................ ................................ .......................... 131 Chapter 5. Conclusion and recommendations ................................ ................................ ....... 135 5.1. Summary ................................ ................................ ................................ ............................. 135 5.2. Recommendation ................................ ................................ ................................ ............. 138 APPENDICES ................................ ................................ ................................ ................................ ....... 142 APPENDIX A: HYDRUS CW2D parameters ................................ ................................ .......... 143 APPENDIX B: R Code (Goodness of fit) ................................ ................................ .............. 155 REFERENCES ................................ ................................ ................................ ................................ ....... 157 vii LIST OF TABLES Table 1. Design criteria for onsite wastewater land application system ................................ ....... 17 Table 2 . Summary of HYDRUS CW2D, LEACHM, SWAP, VS2DT, and DRAINMO D - NII ... 24 .......................... 30 ............................... 31 Table 5. Wastewater characteristic input parameters for HYDRUS CW2D calibration and validation ................................ ................................ ................................ ............................... 41 Table 6. Time variable boundary condition for calibration and validation ................................ .. 43 Table 7. Time variable boundary condition for dosing frequency ................................ ................ 45 Table 8. Typical domestic wastewater land app lication treatment performance .......................... 47 Table 9. Goodness of fit result of calibration and validation ................................ ........................ 48 Table 10. Adjusted HYDRUS CW2D parameters for domestic wastewater land application modeling ................................ ................................ ................................ ............................... 49 Table 11. Simulated effluent concentrations before and after the fitting process for calibration . 50 Table 12. Relative difference between measured and simulated values for calibration ............... 51 Table 13. Simulated effluent concentrations before and after the fitting process for validation .. 52 Table 14. Relative difference between measured and simulated values for validation ................ 53 Table 15. Relative difference from control for COD simulations (3 dosing frequency with 102 mg/L of COD) ................................ ................................ ................................ ....................... 58 Table 16. Relative difference from control for nitrate simulation (3 dosing frequency with 102 mg/L of COD) ................................ ................................ ................................ ....................... 61 Table 17. C haracteristic of food processing wastewater ................................ .............................. 65 Table 18. Characteristics of wastewater at the demonstration site ................................ ............... 73 viii Table 19. Time variable boundary condition for multiple loading strengths ............................... 84 Table 20. Texture for non - optimal and optimal areas ................................ ................................ . 109 T able 21. Soil compaction result using soil compaction meter in Field 1 ................................ .. 111 Table 22. Soil com pact result using soil compaction meter in Field 2 ................................ ....... 112 Table 23. Soil compact result using soil compaction meter in Field 3 ................................ ....... 113 Table 24. Infiltration analysis in Field 1. ................................ ................................ .................... 114 Table 25. High localized water condition analysis for Fields 2 and 3 ................................ ........ 119 Table 26. Goodnes s of fit result for calibration and validation ................................ .................. 122 Table 27. Adjusted HYDRUS CW2D parameters for food processing wastewater land application modeling ................................ ................................ ................................ ........... 122 Table 28. Simulated effluent concentrations before and after the fitting process for calibration 123 Table 29. Relative difference between measured and simulated values for calibration ............. 123 Table 30. Simulated effluen t concentrations before and after the fitting process for validation 124 Table 31. Relative difference between measured and simulated values for validation .............. 124 Table 32. Relative difference from control for COD si mulation (1 dosing frequency with 619 mg/L of COD) ................................ ................................ ................................ ..................... 127 Table 33. Relative difference from control for nitrate simulation (1 dosing frequency with 619 mg/L of COD) ................................ ................................ ................................ ..................... 130 Table 34. Literature review on PPCP found in a domestic wastewater land application system 141 ix LIST OF FIGURES Figure 1. Nitrogen processes in soil ................................ ................................ .............................. 13 Figure 2. Photographs of soil trench es (Dong et al. 2017) ................................ ........................... 34 Figure 3. Inverse modeling data input in HYDRUS CW2D; X - time, Y - measured data, Type - pressure head (1), volumetric water content (2), solute concentration (4), Position - observation node number corresponding to where the volumetric water content is measured, Weight - weight associated with a particular data point ................................ ....................... 39 Figure 4. Fitted HYDRUS to measured value using volumetric water content from Domestic WW ................................ ................................ ................................ ................................ ....... 48 Figure 5. Simulated COD effluent concentrat ion at 15.24, 30.48, 60.96, 91.44, and 121.9 cm (0.5, 1, 2, 3, and 4 ft) depths as increasing loading strength (1x - 5x) on 150 th days of operation ................................ ................................ ................................ .............................. 55 Figure 6. Simulated nitrate effluent conce ntrations with multiple influent COD strength ........... 56 Figure 7. Steady state condition of heterotrophic microorganism (XH) growth .......................... 56 Figure 8. Impact of dosing frequency and influent COD concentration on COD effluent concentration on the 150 th day at 60.96 cm (2 ft) depth of sandy loam soil ......................... 58 Figure 9. Impact of dosing fre quency and COD concentration on nitrate effluent level estimation at 60.96 cm (2 ft) depth of sandy loam soil ................................ ................................ .......... 61 Figure 10. Simulation of the impact of volumetric water content in soil by diff erent dosing frequencies at 150 th day; 3 dosing (left), 6 dosing (middle), and 10 dosing (right) .............. 62 Figure 11. Overview map of demonstration site ................................ ................................ ........... 71 Figure 12. Wastewater flow of the demonstration site ................................ ................................ . 71 Figure 13. Hickenbottom pipe for inducing and collecting run off ................................ .............. 72 Figure 14. Overview of sensor cluster at the demonstration site ................................ .................. 74 Figure 15. Composition of soil sensor cluster ................................ ................................ .............. 75 x Figure 16. Sensors installed at depths ................................ ................................ ........................... 76 Figure 17. MW 103 at the demonstration site ................................ ................................ ............... 77 Figure 18. Juno 3B GPS Handheld ................................ ................................ ............................... 78 Figure 19. Optimal and non - optimal areas ................................ ................................ ................... 79 Figure 20. Soil compaction meter by AgraTronix ................................ ................................ ........ 80 Figure 21. Infiltrometer by Turf - tec International (Tallashassee, FL) ................................ .......... 81 Figure 22. Geoprobe ................................ ................................ ................................ ..................... 83 F igure 23. Hydraulic and organic loading at the demonstration site ................................ ............ 86 Figure 24. Daily volumetric water content at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 1 ................................ ................................ ................................ ................................ 87 Figure 25. Daily oxygen level at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 1 87 Figure 26. Daily temperature at 30.48, 61 .96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 1 . 88 Figure 27. Daily volumetric water content at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 2 ................................ ................................ ................................ ................................ 89 Figure 28. Daily oxygen level at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 2 89 Figure 29. Daily temperature at 30.48, 61.96, and 91. 44 cm (1, 2, and 3 ft) depth on Cluster 2 . 89 Figure 30. Daily volumetric water content at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 3A ................................ ................................ ................................ ............................. 91 Figure 31. Daily oxygen level at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 3A ................................ ................................ ................................ ................................ ............... 91 Figure 32. Daily temperature at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 3A ................................ ................................ ................................ ................................ ............... 91 Figure 33. Daily volumetric water content at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 3B ................................ ................................ ................................ ............................. 93 xi Figure 34. Daily oxygen level at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 3B ................................ ................................ ................................ ................................ ............... 93 Figure 35. Daily temperature at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 3B 93 Figure 36. Daily volumetric water content at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 4 ................................ ................................ ................................ ................................ 95 Figure 37. Daily oxygen level at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth on Cluster 4 95 Figure 38. Daily temperature at 30.48, 61.96, and 91.44 cm (1, 2, and 3 ft) depth o n Cluster 4 . 95 Figure 39. Estimated groundwater flow at the demonstration site ................................ ............... 96 Figure 40. MW locations for Field 1 ................................ ................................ ............................ 97 Figure 41. Groundwater quality from MW 7, upstream of Field 1 ................................ ............... 98 Figure 42. Groundwater quality from MW 101, within Field 1 ................................ ................... 98 Figure 43. Groundwater quality from MW 15, downstream of Field 1 ................................ ........ 98 Figure 44. MW locations for Field 2 ................................ ................................ ............................ 99 Figure 45. Groundwater quality from MW 24 (upstream of Field 2) ................................ ......... 100 Figure 46. Groundwater quality from MW 102 (within Field 2) ................................ ................ 1 00 Figure 47. Groundwater quality from MW 12 (downstream of Field 2) ................................ .... 100 Figure 48. MW location in Field 3 ................................ ................................ .............................. 101 Figure 49. Groundwater quality in MW 28 (upstream of Field 3) ................................ .............. 102 Figure 50. Groundwater quality in MW 20R (within of Field 3) ................................ ............... 102 Figure 51. Groundwater quality in MW 103 (within of Field 3) ................................ ................ 102 Figure 52. Groundwater quality in MW 23 (downstream of Field 3) ................................ ......... 103 Figure 53. Area, owned by neighbor ................................ ................................ .......................... 103 xii Figure 54. MW locations for Field 4 ................................ ................................ .......................... 104 Figure 55. Groundwater quality from MW 11 (upstream of Field 4) ................................ ......... 105 Figure 56. Groundwater quality from MW 16 (downstream of Field 4) ................................ .... 105 Figure 57. D elineated (Red) areas on Field 1 ................................ ................................ ............. 107 Figure 58. Delineated (Red) areas on Field 2 ................................ ................................ ............. 107 Figure 59. Delineated (Red) areas on Field 3 ................................ ................................ ............. 108 Figure 60. Non - optimal areas ................................ ................................ ................................ ..... 108 Figure 61. Locations for soil compaction analysis in Field 1 ................................ ..................... 111 Figure 62. Locations for soil compaction analysis in Field 2 ................................ ..................... 112 Figure 63. Locations for soil compaction analysis in Field 3 ................................ ..................... 113 Figure 64. Catch Can Volume for Field 1 ................................ ................................ .................. 116 Figure 65. Catch Can Volume for Field 3 ................................ ................................ .................. 116 Figure 66. Locations for high localized water condition analysis in Field 2 .............................. 118 Figure 67. Locations for high localized water condition analysis in Field 3 .............................. 118 Figure 68. High localized water condition found in location 1 on Field 3 ................................ . 119 Figure 69. Fitted HYDRUS to measured value using volumetric water content from Domestic/Food WW ................................ ................................ ................................ ........... 121 Figure 70. Simulated COD effluent concentrations with multiple strength of loadings ............ 125 Figure 71. Steady state condition of heterotrophic microorganism (XH) growth ...................... 126 Figure 72. Impact of dosing frequency, and hydraulic and organic loadings on COD effluent concentration on the 150 th day at 60.96 cm (2 ft) depth of sandy loam soil ....................... 127 Figure 73. Effect of dosing frequency and the strength of hydraulic and organic loadings on nitrate effluent concentration ................................ ................................ .............................. 129 xiii Figure 74. Simulation of the impact of volumetric water content in soil by different dosing frequencies at 150 th day; 3 dosing (left), 6 dosing (middle), and 10 dosing (right) ............ 130 Figure 75. Domain t ype and u nits ................................ ................................ ............................... 143 Figure 76. Rectangular d omain d efinition ................................ ................................ .................. 144 Figure 77. Main pro cesses and add - on m odules ................................ ................................ ......... 144 Figure 78. Time information ................................ ................................ ................................ ....... 145 Figure 79. Output information ................................ ................................ ................................ .... 145 Figure 80. Water flow parameters ................................ ................................ .............................. 146 Figure 81. Solute t ransport ................................ ................................ ................................ .......... 146 Figure 82. Solute t ransport parameters ................................ ................................ ....................... 147 Figure 83. Solute transport c onstructed w etland m odel p arameter I (Default) ........................ 148 Figure 84. Solute transport constructed wetland model (CW2D) p arameters II (Default) ...... 149 Figure 85. Solute transport constructed wetland model p arameter I (Adjusted) ..................... 150 Figure 86. Solute transport constructed wetland model p arameter II (Adjusted) .................... 151 Figure 87. Data for inverse solution ................................ ................................ ........................... 152 Figure 88. Rectangular d omain d iscretization ................................ ................................ ............ 152 Figure 89. Water boundary condition ................................ ................................ ......................... 153 Figure 90. Graphic output from HYDRUS CW2D ................................ ................................ .... 154 xiv KEY TO ABBREVIATIONS Abbreviation Full Word Avg Average BOD Biological Oxygen Demand COD Chemical Oxygen Demand E Model Efficiency GHG Greenhouse gas HYDRUS CW2D HY DRUS Constructed Wetland 2D IA Index of Agreement MW Monitoring Wells RMSE Root Mean Squared Error Std Standard Deviation TN Total Nitrogen TP Total Phosphorus WWTP Wastewater Treatment Plant 1 Chapter 1. Introduction Wastewater la nd application has been pertinent for years due to its low cost, energy use, and maintenance requirements. In a conventional activated sludge wastewater treatment system , aeration accounts for the majority of energy usage, requiring 50 - 70% of the facility energy needs (Environmental Dynamics International 2012) . New York State Energy Research and Devel opment Authority ( NYSERDA) stated that the electricity requirement to remove a kg of BOD 5 ranges from 2.87 to 9.04 kWh, depending on plant size (NYSERDA 2007) . L and application treatment system s reduce the energy/electricity cost, which further reduces greenhouse gas (GHG) emissions associated with the energy savings. L and application has been used for various wastewater types such as domestic and food processing . More than 60 million people in the United States depend on individual onsite or sma ll community cluster system s to treat their wastewater (USEPA 2017a) . The density of septic systems varies nationwide , but in general it is higher in the e astern states. The largest density is in Vermont , with 55% of households r elying on septic systems , whereas the lowest is California, with 10% of households depending on septic system s . S eptic systems are used in about 33% of new development throughout the nation and continued growth is expected. Food processing wastewater also has been land - applied for treatment since 1947 (Dennis 1953) , mainly in Minnesota, Michigan, Pennsylvania, Wisconsin, and Washington ( CVRWQCB 2006; Dennis 1953) . In California, approximately 70 % of food processing wastewater is land - applied (Beggs et al. 2007) . Specifically , over 640 food processing plants are in operation in Central Valley, California resulting in the application of a pproximately, 70 % o f wastewater annually 2 (CVRWQCB 2006 ) . In summary, l and application sy stem s are commonly used to treat domestic wastewater and food processing wastewater. F actors including hydraulic and organic loading s , frequency of loading, soil type, soil depth, temp erature, and soil microbial communities play a significant role i n the performance of land application treatment. Understanding the complexity of soil is important. The nitrification process converts ammonia to nitrites and then nitrate under aerobic condi tion s. Denitrification converts nitrate to nitrogen gas under anaerobic condition if an organic carbon source is available . Complete nitrification usually occurs within the first 30 cm (12 in) of the soil depth (Beach 2001; Fischer 1999) . However, complete denitrification typically does not occur in land application systems for domestic and food processing wastewaters (Heatwole and McCray 2007; Redding 2012) . Therefore, nitrate is a concern since it is highly mobile and can flow into groundwater. H igh levels of nitrate in groundwater causes methemoglobinemia, also known as blue baby syndr ome (DEQ 2015) . A case study found that a potato pro cessing facility in Grant County, Washington applying approximately 5.3 million liters of wastewater per day ( 1.42 million gallons per day ) year - round on 9.3 km 2 ( 2,301 acre ) resulted in nitrate contamination in groundwater. The level increased from 1 to 20 mg/L - N in 1986 (Redding 2012) . The United States Environmental Protection Agency (USEPA) set the maximum contaminant level for nitrate at 10 mg/L - N in groundwater, and hav e been strictly enforc ing it. In 2017, a winery in California received a fine of $635,000 for land - applying wastewater that resulted in high levels of nitrates into groundwater (Cuff 2017) . USEPA estimated that a small percentage of most at level above 5 mg/L - N. In Delaware, it is estimated that 53% of t he groundwater has nitrate concentration s above 5 mg/L - N (USEPA 2017b) . 3 Hydraulic loading, organic loading, dosing frequency, soil type , soil depth, and temperature determine the treatment effectiveness. Design procedures are generally based on empirical relationships that prevent water surfacing, which do es not adequately account for site and waste specific conditions (Conn and Siegrist 2009; Leverenz et al. 2009; Siegrist 2007) . Because o rganic and hydraulic loadings are generally fixed based on p roduction, dosing is the only operational parameter that can be practically adjusted to enhance treatment for site - specific conditions . However, research in this area is lacking. To determine and optimize the dosing frequency based on the treatment perform ance, the complexity of soil treatment must be understood. In this study, a modeling effort was conducted using the finite element software, HYDRUS Constructed Wetland 2D ( HYDRUS CW2D), to examine the impact of dosing frequency on treatment performance. 4 1.1. Hypothesis/ r esearch question This research first verified that land application of wastewater is effective for site - specific condition. With this verification, the following hypotheses were researched . Wastewater land application, compared to conventiona l wastewater treatment, can save cost and energy usage, consequently, reducing GHG emissions; and provide resources, such as water and nutrients, for crop production while minimizing environmental pollution. M odeling can effectively simulate the wastewater land application treatment system to enable estimations of treatment performance. I ncreasing dosing frequency in wastewater land application system can maximize the denitrification process . 1.2. Objective The above hypotheses and research question s lead to th e following project objectives. De monstrate the effectiveness of the land application of domestic wastewater by examination of literature. Evaluate the effectiveness of food processing wastewater land application by comprehensively monitoring an actual ins tallation. The monitored parameters include hydraulic and organic loadings , soil conditions ( including its physical characteristics, temperature, moisture content, oxygen concentration , irrigation uniformity, frequency of standing water, and crop growth ), and local subsurface water quality. 5 Compare the benefits of wastewater land application to conventional wastewater treatment system s in terms of energy saving , GHG reduction associat ed with the energy saving , and freshwater reduction and nutrient reuse for crop production. Develop a simulation approach for the wastewater land treatment system using the HYDRUS Constructed Wetland 2D model and c alibrate and validate using laboratory experimental data. Analyze multiple scenarios using the above calibrated mode l to correlate operational parameter to treatment performance including carbon degradation, nitrification , and denitrification. 6 1.3. D issertation framework T he chapters i n this dissertation are , in order , introduction, literature review, general methodology, domestic wastewater land application, food processing wastewater land application, and conclusion. Each are summarized in the subsequent paragraphs. Chapter 2 is a literature review on the following concepts. W astewater land application W astewater treat ment technologies E nvironmental impacts from wastewater land application N itrogen process in wastewater land application I mpact of loadings for wastewater land application treatment C urrent design criteria O verview o f subsurface flow soil modeling Chapte r 3 focuses on domestic wastewater land applications , addressin g all the hypothes es and objectives except for the 2 nd one . This chapter contains, in order, the introduction, methods, results and discussion, and conclusion. First, the performance of domesti c wastewater land application systems was examined and the benefits were estimat ed, including energy conservation and GHG reduction associated with the energy saving s . HYDRUS CW2D modeling of domestic wastewater land application was then discussed. The mod el ing approach was developed , and then calibrated and validated using laboratory experimental data . Using the model, multiple scenarios were examined to observe the capacity of wastewater land application systems and the enhancement of treatment performanc e by changing operation parameters. 7 Included is the assessment of soil depth requirements based on hydraulic and organic loadings and the impact of dosing frequency on the denitrification process. Chapter 4 focuses on food processing wastewater land appl ication, and addresses all the hypotheses and objectives except the 1 st one . This chapter contains, in order, an introduction, methods, result s and discussion, and conclusion s . A comprehensive monitoring strategy for a long - term food processing wastewater land application sites is first discussed . M onitoring include d tracking hydraulic and organic loading s , observing soil condition in real time by soil sensor clusters, and analyzing groundwater quality. Evaluation of non - optimal and optimal areas at the dem onstration site w ere performed by visual observation, soil analysis, and uniformity of irrigation pivot s . This monitoring strategy helps to safely operate the wastewater land application system . Monitoring the h ydraulic and organic loading s and soil condit ion using s oil sensor clusters also help ed in determining the operational strategies . Next, HYDRUS CW2D modeling of food processing wastewater is discussed. Chapter 5 concludes the dissertation by summarizing the effectiveness of domestic and food process ing wastewater land applications. Thereafter, insights and recommendations for further research are provided. 8 Chapter 2 . Literature review This chapter contains background information on wastewater land application , wastewater treatment technolog ies, wa stewater characteristic s , nitrogen processes in wastewater land application, impact s of loadings for wastewater land application, environmental impacts by wastewater land application, current design criteria, and an overview of subsurface flow soil modelin g. 2.1. Background of wastewater land application Land treatment systems are commonly used to treat domestic and food processing wastewater. In 1980, approximately 25% of all housing units (18 million people ) in the United States, disposed of wastewater using an onsite wastewater treatment. Septic tanks with a drain field were the most common (U.S. Census Bureau 2006; USEPA 1980) . In 2017 , approximately 60 million people depended on onsite w astewater treatment systems (USEPA 2017a) . Use of onsite wastewater treatment system is expected to increase to an estimated one - third of all new housing development (USEPA 2017a) . In addition, land a pplication treatment systems have been utilized for many years to treat food processing wastewater, which is highly variable in volume and composition. The first sprinkler irrigation system in the United States with food processing wastewater was demonstra ted in 1947 (Dennis 1953) . A 1964 national survey identified 844 operating land application systems applying food processing wastewater and it is estimated that over 70% of the wastewater p roduced by California food process ors is applied to the land for the treatment 9 (Beggs et al. 2007) . In summary, the users for domestic and food processing wastewater land application system will increase in the foreseeable future . 2.2. Wastewater treatment technologies In general, wastewater treatment is divided into conventional treatment and land application system s . Each technology has advantages and disadvantages . Conventional wastewater treatment system s are complex mechanical system s that includ e activated sludge, ae robic lagoon, membrane treatment system , trickling filter, coagulation and flocculation, clarifier, and biological treatment (Tchbanoglous et al. 2003) . These systems effectively treat the wastewater but have high capital and operation cost s . Factors affecting operation costs include the size and loading of the plant, topography and geography of the site, wastewater characteris tics, technologies associated with the treatment process, type of biosolids treatment, energy supply automation, and organization of the plant and management (Wendland 2005) . If an activated sludge system is emplo yed, the aeration tanks uses 50 73% of the total energy required for a typical wastewater treatment system (Bohn 1993; Environmental Dynamics International 2011) . Approximately operation cost require $0.35 to treat a liter of wastewater (Balmer and Mattsson 1994; Big Fish Environmental 2010) . For example, o peration cost for a 1.89 million liter/day (500,000 gallon/day) wastewater treatment plant is estimated $6 7 2,000 /year (Big Fish Environmental 2010) . In addition, t ypical conventional wastewater treatment plant need to handle their biosolids . Approximately 0.94 kg (1.95 lbs) of dry solids per 3,785 liter (1 , 000 gal lon ) are produced from the primary and secondary processes (Tchbanoglo us et al. 2003) . A case study in New Hampshire found that 40%, 27%, 23%, and 16% of their 10 biosolids were disposed of by land application (class A and B), landfilling, incineration (city of Manchester only), and out of state landfilling, respectively. The cost for biosolid s disposal was estimated at $75/wet ton, $40/wet ton, $71/wet ton, and $77/wet ton for land application (class A and B), landfilling, incineration (city of Manchester only), and out of state landfilling, respectively (Wheeler et al. 2008) . This energy requirement and other operational cost s result in high reoccurring annual expenses for conventional wastewater treatment. W astewater land application treatment costs le ss, uses less energy and chemical s , and requires less maintenance, in comparison to traditional wastewater treatment. Specifically, l and application typically costs 30 50% less to operate than a typical conventional wastewater treatment system (Charmley et al. 2006; Uhlman and Burgard 2001) . Food processing wastewater is often irrigated on crop land to grow c orn and alfalfa for animal feed , redu c ing the use of fresh water and nutrients. Water scarcity is a global issue and agriculture is the primary source of freshwater depletion in the United State s (USDA 2016) . In 2010, total irrigation water withdrawals were 435 ,275 million liter /day (115,000 million gallons/day), which wa s 38 % of total freshwater withdrawals in the United States (Maupin et al. 2014) . Therefore, wastewater land application system reduce t he use of freshwater for crop production . The land application of wastewater requires acceptable site conditions such as area of land availability, soil type, depth to the groundwater, and topography. I mproper operation of land application systems can resu lt in groundwater contamination. S oil s that are either aerobic or anaerobic may result in nitrate leaching and metal mobilization into groundwater, respectively (Dong et al. 2017a; Julien and Safferman 2015) . A balance is essential. 11 Selecting the best wastewater treatment technology is a complex process that requires accounting for site and waste - specific conditions . Included are parameters such as the volume and composition of wastewater, location, type of pr ocessing plant, availability of municipal treatment facility, soil type, cost, and state and local legislation (Harper et al. 1972) . Both conventional wastewater treatment system s and wastewater land application system s can be effective in treating wastewater. However, c onventional wastewater treatment system s are more suitable for urban area and land application system for rural areas. In regard to cost s , wastewater land application system are generally more economical . 2.3. E nvironmental impacts from wastewater land application Wastewater land application can damage the environment by leaching c ontaminants in to the groundwater and /or cause run - off. Domestic wastewater land application systems are generally used in rural areas, representing one of the largest volumetric sources of effluent to groundwater (Koren and Bisesi 2003) . If not proper ly designed and constructed, shallow, unconfined aquifers can become contaminated by nitrate , resulting in a significant public health risk (Robertson et al. 1991; Wilhelm et al. 1994) . In fact, nitrate contamination (concentration in groundwater >10 mg/L) often occurs even in well - constructed and properly functioning domestic wastewater land application systems (Wilhelm et al. 1994) . Nitrate contamination of groundwater has been found under drain fields in the valley soils of the northwest ern U nited S tates (Ver Hey 1987 ) . Similarly, nitrate contamination of groundwater has been documented in the South Valley of Albuquerque, New Mexico (Keleher 2008 ) . If the nitrate is not denitrified, h igh levels enter in groundwater and can cause methemoglobinemia, also more commonly 12 known as blue baby syndrome (DEQ 2015) . A 1950 report listed 144 cases of infant methemoglobinemia with 14 deaths in Minnesota (Rosenfield and Huston 1950) . When the wastewater is applied to soil, soil microorganisms use the organic materials as a food source. During the process of oxidation and decomposition of organic materials, electrons are release. Oxygen is the most favorable electron acceptor (Ta rradellas, Bitton, and Rossel 1997) . When the oxygen is depleted , lower energy electron acceptors such as nitrate, manganese, iron, and sulfate are utilized (Haggblom and Milligan 2000; Matocha et al. 2005; Mokma 2006a) . The low redox po tential condition in soil may reduce metal species to be in a more mobile form (Safferman et al. 2011) . Therefo re, m etal mobilization into groundwater is also a concern where nitrification is limited and nitrate is much less prevalent (McQuilan 2004) . 2.4. Nitrogen processes in wastewater land application A nother concern regarding w astewater land applicati on is nitrate leaching into groundwater (Cuff 2017; Redding 2012) . Adriano et al. (1975) showed that 76% of tot al nitrogen from fruit and vegetable processing wastewater applied on the sandy loam soil leached into subsurface water (Adriano et al. 1975) . Therefore, understanding the nitrogen processes in soil , as shown in F igure 1, is important to protecting the environment. 13 Figure 1 . Nitrogen process es in soil In soil, nitrogen is transformed by nitrogen fixation, ammonification, nitrification, denitrification, and anammox. Nitrogen fixation is the conversion of nitrogen gas to ammonium by microo rganisms. Ammonification is the conversion of organic nitrogen to ammonium resulting from the de composition of dead plant residual, animal tissue, and microbial biomass . Nitrification is the oxidation of ammonia to nitrite, and then nitrate , which is highl y mobile. This is carried out by nitrifying bacteria under aerobic conditions. Nitrifying bacteria includes ammonia - oxidizing bacteria ( Nitrosomonas , Nitrosococcus , and Nitrosospira ) and nitrite - oxidizing bacteria ( Nitrobacter , Nitrospina , and Nitrococcus ) (Watson et al., 1981) . Denitrification converts nitrate to nitrogen gas and is carried out by denitrifying bacteria such as Heterotrophic bacteria , Thiobacillus denitrificans , micrococcus denitrificans , Pseudomonas , and Achromobacter , under anaerobic co ndition (Carlson and Ingraham, 1983) . Denitrification occurs 14 when a carbon source is available for denitrification microorganisms, the soil is under anaerobic conditions , and temperature s are within an acceptable range . At greater soil depths, lower level s of oxygen are likely , which can promote denitrification. On the other hand, carbon is needed for denitrification. Typically, carbo n is oxidized in the upper level s of soil that are often aerobic and , consequen tly , denitrification may not occur resulting in the nitrate leaching into groundwater. Anammox and anaerobic ammonia oxidation converts ammonia to nitrogen gas under anaerobic c ondition s . This process is driven by microorganisms such as Candidatus Anammoxoglobu propionicus and Candidatus Brocadia (Kartal et al. 2007) . 2.5. Impact of loadings for wastewater land application treatment H ydraulic and organic l oading s ha ve an important role when evaluating and designing wastewater land application systems . Typically, these loadings are not controllable at wastewater land application site. Related is the dosing frequency, which may be a crit ical , practical operational parameter as it can be altered without impacting the loadings. The impact s of loadings are discussed in below. Hydraulic and organic loadings are the principal parameters in designing wastewater land application systems. Increas ultimate resulting in its porosity being potentially completely occupied by water (i.e., saturation). When the soil is saturated, oxygen cannot diffuse into its porosity resulting in anaerobi c conditions (Erickson and Tyler 2000) . In addition, w hen wastewater is applied to land, the particulate solids of the wastewater can remain near the surface, limiting oxygen transport to the soil and promoting anaerobic condition s (Beggs et al. 2007; Crites and Tchobanoglous 1998) . 15 Ex cessive organic loading or long - hydraulic conductivity (McDaniel 2006) . High organic loading may also enhance microbial activity because organic carbon is a substrate or food for microorganisms. Excess of microorga nism s can clog the pore space in soil, which may reduce its hydraulic conductivity resulting in a lower redox potential (Hillel 2008) . Fur thermore, i ncreasing hydraulic loading in a well - drained soil decreases retention time of the wastewater, which reduces the efficiency of treatment (Converse and Tyler 1998; Siegris t and Van Cuyk 2001) . Converse and Tyler (1998) studied the treatment of fecal coliform concentrations in a well - drained soil with different hydraulic loadings at 40.75, 122.24, and 244.48 liters per day/m 2 (1, 3, 6 gallons per day/ft 2 ). Higher fecal coli form concentrations were found in effluent wastewater when the soil received 122.24, and 244.48 liters per day/m 2 (3 and 6 gallons per day/ft 2 ) instead of 40.75 liters per day/m 2 (1 gallon per day/ft 2 ) (Converse and Tyler 1998) . Dosing frequency is an operational parameter which may impact on the performance of treatment. A previous study discus se d that a hydraulic resting period of 12 hours provide d the soil immediately after the addition of water. Doses given at a higher frequency, with less resting ti me, were shown to lead to anoxic conditions (Julien and Safferman 2015) . Therefore, increasing dosing frequency may impact on soil reduction condition, which may promote denitrification process. As the frequency of dosing increase, the retention time may increase, which may result in better treatment. In a sand filter treatment system, increasing the dosing frequency was found to impr ove the performance of treatment, but continuous heterotrophic bacterial growth was observed at the surface , which may result in clogging or premature of life of sand filter treatment 16 system (Furman et al. 1955; Grantham et al. 1949; Leverenz et al. 2009) . The optimal dosing frequency should be determined while considering both hydraulic and organic loading s to minimize clogging or premature of life. 2.6. Current design c r iteria Currently, design criteria for wastewater land application systems differ by states. Siegrist (2007) discussed that hydraulic loadings for domestic wastewater land application system s are based on limited empirical evidence and vary widely from stat e to state (Siegrist 2007) . However, most state regulations focus on a few specific wastewater disposal characteristics, the most important of which are hydraulic loading, organic loading, soil depth, and soil type. Table 1 shows the diverse design criteria for domestic wastewater land application system (Arkansas State board of health 2007; Colorad o Department of public health and envrionment 2013; Michigan Department of Environmetnal Quality 2013; Nebraska Department of Environmental Quality 2007; New York State Department of Health 2016; Olivieri and Roche 1979; Oregon Department of Environmental Quality 2017; State of Kansas Department of Helath and Environment 1997; Tennessee State Government 2016) . M any states do not have gu ideline for organic loading. States recommend between 45.72 cm ( 18 in ) to 121.92 cm (4 ft) vertical separation between th e bottom of the drain field and the water table. Regulations on wastewater hydraulic loading are even less uniform, relying on a combination of factors including hydraulic loading, soil type, organic loading, and treatment system size. Many rely on a flow rate per bedroom. Further stipulations are often imposed based on wastewater strength and soil profile. 17 Table 1 . Design criteria for onsite wastewater land application system State Soil depth required for the drain field Soil depth required between drain field and water table Hydraulic loading Organic loading Reference Arkansas 0.46 m (18 in) 0.6 m (24 in) loamy soil, 0.9 m (36 in) s andy soil 1 5 - 30 L/m 2 /day (0.37 - 0.75 gal/ft 2 /day) depends on percolation rate N/A (Arkansas State board of health 2007) California 0.3 m (12 in) 0.91 m (3 6 in ) greater than 5 min/in , 6 m (2 4 0 in ) between 1 and 5 min/in , prohibited less than 1 min/in 9 - 64 L/m 2 /day (0.22 - 1.58 gal/ft 2 /day) depends on percolation rate N/A (Olivieri and Roche 1979) Colorado N/A 1.2 m (48 in) N/A N/A (Colorado Department of public health and envrionment 2013) Kansas N/A 1.2 - 1.8 m (48 - 72 in ) 10 L/m 2 /day (0.25 gal/ft 2 /day) sandy clay loa m 16 L/m 2 /day (0.4 gal/ft 2 /day) sandy loa m 24 L/m 2 /day (0.6 gal/ft 2 /day) l oamy sand 37 L/m 2 /day (0.9 gal/ft 2 /day) medium sand 45 L/m 2 /day (1.1 gal/ft 2 /day) course sand N/A (S tate of Kansas Department of Helath and Environment 1997) Maryland 0.15 to 0.3 m (6 to 12 in) 1.2 m (48 in) T able is provided N/A (Maryland Deaprtment of the Environmemt 2010) 18 Table 1 . Design criteria for onsite wastewater land application system ( ) State Soil depth required for the drain field Soil depth required between drain field and water table Hydraulic loading rate Organic loading Reference Michigan 0.3 to 0.6 m (12 to 24 in) 0.46 m (18 in) 12 L/m2/day (0.3 gal/ft2/day) - sandy cla y 24 L/m2/day (0.6 gal/ft2/day) - loam, sandy loam 40 L/m2/day (1.0 gal/ft2/day) - l oamy sand 48 L/m2/day (1.2 gal/ft2/day) - fine sand 65 L/m2/day (1.6 gal/ft2/day) - coarse sand (DEQ 2013) Nebraska N/A 1.2 m (48 in) N/A N/A (Nebraska Department of Environmental Quality 2007) New York N/A 1.2 m (48 in) 416 - 568 L/day (110 - 150 gal/day) N/A (New York State Department of Health 2016) Tennessee N/A 1.2 m (48 in) 12 L/m2/day (0.3 gal/ft2/day) - sandy clay 24 L/m2/day (0.6 gal/ft2/day) - s ilt loam, loam 28 L/m2/day (0.7 gal/ft2/day) - sandy loam 40 L/m2/day (1.0 gal/ft2/day) fine sand > 150 mg/L BOD; 3 g/m2/day (27 l b BOD/acre/day) for c lays , 4.6 g/m2/day (41 lb BOD/acre/day) for loams, 6.2 g/m2/day (55 lb BOD/acre/day) for sandy (Tennessee State Government 2016) 19 The lack of design consistency is also observed for the land application of food processing wastewater. According to a literature review by ( 2006) , a wide range of hydraulic loading s are observed. Specifically, h ydraulic loadings from 21 food processing facilities range d from 1.96 to 140 liter/m 2 /day ( 2,100 to 150,000 gal/acre/day ) (Carawan et al. 1979) . Hydraulic loading is limited by the organic loading. High organic loading can cause microorganisms to grow extensively , which can clog soil s . When the soil is clogged, surface ponding or run - off may occur. Organic loading also has been roughly estimated based on empirical relationship, which result in a wide range of observations . In the state of New York, t he organic loading was recommended at 56 g of BOD /m 2 /day ( 500 lb of BOD /acre/da y ) (Crites et al. 2000) . Spyridakis and Welch (1976 ) stated that organic loading from two food processing plants were 52 and 84 g o f BOD /m 2 /day ( 460 and 750 lb of BOD /acre/day ) (Spyridakis and Welch 1976) . Crarawa n et al (1979) recommended the maximum organic loading of 22 g of BOD /m 2 /day ( 200 lb/acre/day ) (Carawan et al. 1979) . For Michigan sandy soils, the rough limits, which have been observed by current wastewater application, are between 0.0056 and 0.0224 kg of BOD /m 2 /day (50 and 200 lb/acre/day) with a hydraulic loading less than 3.74 liters/m 2 /day (4000 gal/acre/day) ( Mokma 2006) . This wide range of hydraulic and organic loadings indicates that more research is needed and modeling can be beneficial to determine optimal hydraulic and organic loading s while considering site and waste specific c ondition. 2.7. Overview of subsur face flow soil modeling Many models have been developed to quantify water flow and pollutant movement in soils. These models have been widely used in agriculture, constructed wetland, and septic soil 20 treatment systems. The mode ling approaches can be a simple analytical approach to a complex nonlinear process. Available models include HYDRUS , LEACHM, SWAP, VS2DT, and DRAINMOD. Details about these models are discussed below. Quantification and visualization of pollutant flow patt erns can be modeled using HYDRUS Constructed Wetland 2D (CW2D) software. HYDRUS CW2D simulates the complexity of water flow in unsaturated, partially saturated, and fully saturated soil by numerically solving the Richard equation and the convection dispers ion equation 1999) . This model considers chemical and physical processes of pollutants, soil properties, rainfall, and evapotranspiration, including the aerobic and anoxic transformation and degradation process for organic matter, nitrogen, and phospho rus . This model has been widely used to simulate and understand the transport of pesticides, nitrate, phosphorus, and heavy metals in soil (Anwar and Thien 2015; Crevoisier et al. 2008; Dao et al. 2014; Freiberger et al. 2014; Honegger 2015; Mailhol et al. 2007; Nakamura et al. 2004; Naseri et al. 2011; Nohra Twarakavi et al. 2008; Vilim et al. 2013; Wang et al. 2016) . LEACHM (Leaching Estimation and Chemistry Model) is a one - dimensional finite difference mode l. The model can predict water and solute movement, transformation, plant uptake, and chemical reactions in an unsaturated soil by using the various subroutines . LEACHW describes water movement, LEACHP models pesticides, LEACHN models nitrogen and phospho rus, and LEACHC models salinity in calcareous soils. The model uses the Freundlich - Langmuir isotherm for sorption and desorption (Hutson 2000) . The input of soil parameters , including soil physical properties ( bulk density, particle size distribution, and water 21 retention characteristic s) , are required. Previous studies have used the LEACHM model to predict pesticide, herbicide, and heavy metal tr ansport through soil, as well as soil dynamics of nitrogen and nitrate ( Hutson, 1991; Jemison et al., 1994; Khakural, 1993; Wagenet, 1989; Webb and Lilburne, 2000) . SWAP (Soil - Water - Atmosphere - Plant) is a one - dimensio nal model that solves multiple governing equations using finite difference numerical analysis. This model is used to simulate water flow, solute movement, heat flow, macropore flow, and crop growth in soils. It is designed to simulate water and solute move ment processes at a field - scale with applications during both growing seasons and long - term time series. This model has been used for field - scale water and salinity management, irrigation scheduling, modeling transient drainage conditions, plant growth imp acts from water and salinity, pesticide leaching into water sources, regional drainage from topsoil to different surface water systems, optimization of surface water management, and effects of soil heterogeneity (Van Dam et al. 2008; Kroes et al. 2017) . In addition, the SWAP model can predict preferential flow, adsorption, and decomposition of nutrients and pesticides (Van Dam et al. 1997) . VS2DT (Variably Saturated 2D Flow and Transport) uses the fi nite difference technique to approximate the flow equation, developed using a combination of the law of conservation of fluid mass with a non - nutrient transportation in variably saturated soil conditions. The model can simulate in 1 - dimension and 2 - dimensions with planar or cylindrical geometries. There are multiple options for boundary conditions for flow in unsaturated soil, including infiltration with ponding, evaporation, plant transpi ration, and seepage faces. Options for solute transport include first - 22 order decay, adsorption, and ion exchange. Previous studies used this model to predict pollutant transport to tile drainage, evaluate hydraulic properties of soils for irrigation strateg ies, and to evaluate groundwater transport of tracers (Constantz et al. 2003; Munster et al. 1994) . DRAINMOD is a hydrological model for simulating t he performance of agricultural drainage and related water management systems. The model is effective for simulating the hydrology of poorly - drained, high water table soils on both short and long - term timescales . It predicts the effects of drainage and asso ciated water management practices on water table depths, the soil water regime , and crop yields. Infiltration, subsurface drainage, surface runoff, evapotranspiration, vertical and lateral seepage, water table depth, and water - free pore space in the soil p rofile are considered (Skaggs et al. 2012) . The current version of DRAINMOD simulate solely in 1 - d imension flow. DRAINMOD has several modules, including DRAINMOD - S (salinity), DRAINMOD - NII (nitrogen), DRAIN MOD - DUFLOW (linked to DUFLOW model), and DRAINMOD - W (watershed scale) In the past, this model has mainly been used for nitrogen transport (Salazar et al. 2009; Wang et al. 2005; Youssef et al. 2005) but a recent study used it for phosphorus (Askar et al. 2016) . C omparison of HYDRUS CW2D, LEACHM, SWAP, VS2DT, and DRAINMOD - NII are shown in Table 2 . A ll models can simulate water and solute flow in the soil and account for precipitation, ev apotranspiration, plant uptake, and surface runoff. Only HYDRUS CW2D and SWAP consider macropore in the model . SWAP and DRAINMOD can provide estimated crop yield. HYDRUS C W 2D was selected for this study, be cause it is one of the most comprehensive tool s fo r modeling water and solute flow in soil. HYDRUS CW2D specializes in nutrient flow and it entails both aerobic and anoxic transformation and degradation processes for organic 23 matter, nitrogen, and phosphorus. Nitrate is especially an issue in wastewater la nd application and HYDRUS CW2D has demonstrated capabilities to predict its fate . As a focus of this research is the impact of dosing frequency on treatment performance, it is important to note that several studies successfully used HYDRUS to observe its i mpact on the growth of heterotrophic bacteria, fecal coliform, and moisture content (Leverenz, Tchobanoglous, and Darby 2009; Radcliffe and West 2009; Hassan et al. 2005) . Once the HYDRUS CW2D was calibrated and validated using labora tory experimental data, multiple scenarios were run using different dosing frequencies in order to maximize the treatment performance while protecting environment. This modeling approach may allow for the determination if the operation parameter of dosing frequency can be set to achieve both the degradation of carbon and conversion of nitrate to nitrogen gas. 24 Table 2 . Summary of HYDRUS CW2D, LEACHM, SWAP, VS2DT, and DRAINMOD - NII Variable Model HYDRUS CW2D LEACHM SWAP VS2DT DRAI NMOD - NII Dimension 2D 1D 1D 2D 1D Saturated/ U nsaturated flow Yes Yes Yes Yes Yes Solute flow Yes Yes Yes Yes Yes Hydraulic model van Genuchten Campbell van Genuchten van Genuchten van Genuchten Evapotranspiration Yes Yes Yes Yes Yes Surface runoff Y es Yes Yes Yes Yes Macropore flow Yes No Yes No No Plant uptake Yes Yes Yes Yes Yes Crop yield No No Yes No Yes Application Irrigation management Tile drainage design Drip irrigation design Wastewate r land application Constructed wetland Surface run off Nutrient transport Seasonal simulation Pesticide transport Irrigation water management Nutrient transport Pesticide transport Surface runoff Seasonal simulation Irrigation water management Nutrient transport Crop yield estimation Surface runoff Sea sonal simulation Snow Freezing and thawing Irrigation water management Drip irrigation design Nutrient transport Surface runoff Seasonal simulation Irrigation water management Surface runoff Tile drainage design Manure land application Crop yield estim ation Nitrogen transport Freezing and thawing Seasonal simulation 25 Chapter 3 . Domestic wastewater land application This chapter discusses provides an introduction to domestic wastewater land application, including background, problem statement, and benefits. Then, the treatment performance of domestic wastewater land application system s are evaluated using the literature. HYDRUS CW2D modeling of domestic wastewater land application i s then discussed. 3.1. Introduction W astewater land application has b een used for many years to treat the domestic wastewater . The performance has been studied (Dong et al. 2017b; Gross 2004; Hammerlund and Glotfelty 2016; National Environmental Services Center 2013; Ronayne et al. 1982) . The typical, least expensive configuration includes a 1892.7 3785.4 liter ( 500 1 ,000 gallon) septic tank and a subsurface soil distribution network. This network is referred to as drain field (USEPA 2017c) , leach field (USEPA 2017c) , septic soil tr eatment system (Dong et al. 2017b) , septic field (USEPA 2017c) , soil treatment unit (Wunsch et al. 2009) , and soil absorption field (Lesikar 2008) . In this study, domestic wastewater land application systems refer to a septic tank and drain field in series . In comparison to a conventional wastewater tr eat ment system , land application can save energy , consequently reducing GHG emissions. These benefits were estimated based on standard data on the production and characteristics of wastewater, population, and treatment requirements . A pproximately 497 milli on kg (1,095 million lb) of BOD 5 per year from dome stic wastewater is 26 treated by wastewater land application system s in the United State. To remove 0.45 kg (1 lb) of BOD 5 at a traditional wastewater treatment facility where receiving less than 3.78 million liters/day (1 million gallon/day) requires 4.1 kWh of energy (NYSERDA 2007) . Therefore, approximately 2,037,700 M Wh electricity is saved annually . The cost for electricity was estimated at $326 million/year with an assumption 7.27 Cent/kWh. Thi s result in a GHG reduction of 3.5 million metric tons /day , which is equivalent to GHG emission from 715,432 passenger vehicles driven 11,443 miles/year and a mileage of 22 miles/gallon and (USEPA 2016) . GHG emission from phosphorus treatment in a typical wastewater treatment plant was also estimated. When 13,627 million liters (3,600 million gallons) of dom estic wastewater/day is treated by onsite wastewater treatment system in the United States, approximately 29.8 million kg of phosphorus are treated annually. When 29.8 million kg of phosphorus are treated by wastewater treatment plant by a physical / chemica l process , 476,800 metric tons of CO 2 and 646 metric tons of NO x will be produced . Consequently, this amount of gases can be conserved by an onsite wastewater tre atment system (Coats et al. 2011) . Although domestic wastewater land application system s ha ve been widely used, design criteria are not fully develo ped and vary by state. The depth require d for a soil adsorption field varies from 0.15 to 0.6 m (6 to 24 in). In addition, the depth required from the bottom of an adsorption field to the water table range from 0.46 to 6 m (18 to 240 in). Hydraulic loading s were mainly determined by soil type and many state do not have guideline for organic loading s . Siegrist (2007) discussed that allowable hydraulic loading for domestic wastewater land application systems are based on limited empirical evidence. The need f or computer modeling efforts to design the treatment system is emphasized (Siegrist 2007) . 27 For this research, HYDRUS Constructed Wetland 2D ( HYDRUS CW2D), a finite element model, was selected for simulating the movement of water and multiple sol utes in soil. This model was originally designed to simulate wastewater treatment in wetlands, but was also used in this research for wastewater land application. Previous studies have also use d HYDRUS to simulate nutrient movement in soil (Crevoisier et al. 2008; Dao et al. 2014; Mailhol et al. 2007; Shekofteh et al. 2013; Vilim et al. 2013) . This modeling approach may provide the minimum depth requirements for carbon degradation a nd allow for the understanding of how dosing frequency effects the conversion of nitrate to nitrogen gas. Laboratory column operation and chemical wastewater analyses were used to calibrate and validate the HYDRUS CW2D model. This chapter discusses the ef fectiveness of domestic wastewater by an examination of the literature, develop s a HYDRUS CW2D model, calibrate s and validate s the model using laboratory experiment data, and analyze s multiple scenario to observe the treatment performance including carbon degradation, nitrification, and denitrification. 3.2. Materials and m ethods The m ethod to achieve each objective for domestic wastewater land application are provided. Included is evaluation of domestic wastewater land application , the calibration and va lidation of HYDRUS CW2D modeling using laboratory data, and analysis of multiple scenario to provide a more accurate design approach and observe carbon degradation, nitrification, and denitrification. 28 3.2.1. HYDRUS CW2D m od el ing HYDRUS CW2D is a finite element m odel for simulating two - dimensional water and solutes movement in soil. Included is the visualization of the transmission and degradation processes for organic matter and nitrogen under aerobic and anoxic conditions. In this section, the g overn ing equation , HYDRUS CW2D component and processes , limitation of HYDRUS CW2D , laboratory experimental, calibration and validation procedures , input parameters , goodness of fit, and scenarios are discussed. 3.2.1.1. Governing equation HYDRUS CW2D simulates the water and solut e movement in two dimensions using the Richard and advection - convection dispersion equations . The HYDRUS ion for water flow in unsaturated, partially saturated, and fully saturated soil . The assumption was made that the air phase plays an insignificant role in the liquid flow process. The modified . ( 1 ) where is the volumetric water content [L 3 L - 3 ], h is the pressure head [L], x i (i=1, 2) are the spatial coordinates [L], t is the time [T], K ij A are components of a dimensionless anisotropy 29 ten sor K A , K is the unsaturated hydraulic conductivity function [LT - 1 ], and S is a sink/source term [T - 1 ], which is considered here as the amount of water removed by plant roots. The HYDRUS model solves the advection - dispersion equation for modeling transport of solute in a soil - air - water system. Equation 2 is the governing equation (Warrick 2002) ( 2 ) w here , c is solution concentr ation [ML - 3 ], s is adsorbed concentration [MM - 1 ], is water content [L 3 L - 3 - 3 ], D is dispersion coefficient [L 2 T - 1 ], q is volumetric flux [LT - 1 ], and is the rate constant representing reaction [ML - 3 T - 1 ] 3.2.1.2. HYDRUS CW2D componen t and processes HYDRUS CW2D entails both aerobic and anoxic transformation and degradation processes for organic matter, nitrogen, and phosphorus . There are 12 components (Table 3 ) and 9 processes (Table 4 ). 30 Table 3 . Components in HYDRUS CW2D Symbol Description O 2 Dissolved oxygen (mg/L) CR Readily biodegradable COD (mg/L) CS Slowly biodegradable COD (mg/L) CI Inert COD (mg/L) NH 4 N Ammonium - nitrogen (mg/L) NO 2 - N Nitrite - nitrogen (mg/L) NO 3 - N Nitrate - nitro gen (mg/L) N 2 - N Dinitrogen gas (mg/L) IP Inorganic phosphorus (mg/L) XH Heterotrophic microorganisms (mg/L) XANs Nitrosomonas - autotrophic microorganisms (mg/L) XANb Nitrobacter - autotrophic microorganisms (mg/L) 31 Table 4 . Processes in HYDRUS CW2D Processes Description Hydrolysis C onvert s CS to CR, and small fraction being converted in to CI. Aerobic growth of heterotrophic bacteria C onsumes O 2 and CR. Anoxic bacteria growth using nitrite C onsumes O 2 , CR, ammonium (NH 4 - N ), and IP, and produce N 2 due to denitrification on nitrite. Anoxic bacteria growth using nitrate C onsumes O 2 , CR, ammonium (NH 4 - N ), and IP, and produce N 2 due to denitrification on nitrate. Lysis of heterotrophic organisms P roduces CR, C S, CI, am monium (NH 4 - N ), and IP . Aerobic growth of nitrosomonas C onsumes O 2 and ammonium (NH 4 - N ), and produce nitrite (NO 2 - N ). Aerobic growth of nitrobacter C onsumes nitrite (NO 2 - N ) and nitrate (NO 3 - N ). Lysis of nitrosomonas (XANs) P roduces CR , CS, CI, ammonium (NH 4 - N ), and IP. Lysis of nitrobacter (XANb) P roduces CR, CS, CI, ammonium (NH 4 - N ), and IP. 32 The following assumptions are made in HYDRUS CW2D 2006) . O rganic matter is present only in the aqueous phase and all reactions occur only in the aqueous phase. Adsorption is assumed to be a kinetic process and considered for ammonium , nitrogen , and inorganic phosphorus. All microorganisms are assumed to be immobile. Lysis in HYDRUS CW2D repre sent all decay and loss processes of all microorganism involved, and the rate of lysis doe s not represent the impact of environmental conditions. Heterotrophic bacteria of HYDRUS CW2D include all bacteria responsible for hydrolysis, min eralization of organ ic matter (a erobic growth) , and denitrification (anoxic growth). 3.2.1.3. Limitations of HYDRUS CW2D The limitation of HYDRUS CW2D include the following (Langergraber et al. 2003; . Clogging can occur from particulate matters in the influent wastewater settling and excessive growth of bacteria (biofilm). The resulting pore size reduction is not considered in the model. Impact of environmental condition on pH are not considered in the model. Limited to a temperature range between 10 and 25 °C. 3.2.1.4. Model c a libration and validation This section focuses on the procedure to calibra te and validate HYDRUS CW2D. Included are a description of laboratory experiments, calibration and validation procedure , and goodness of fit. 33 3.2.1.4.1. Laborator y experiment The original purpose of the laboratory experiment was to observe the impact of en zyme pretreated fast - food restaurant wastewater on the performance and life of a drain field. This laboratory experiment is similar to the current study and is suitable for model calibration and validation because it simulated wastewater land application s ystem with multiple strength s of wastewater. Substantial details can be found in Dong et al. ( 2017) . In order to calibrate and validate HYDRUS , data from laboratory experiment was used. This laboratory study measured the required parameters for calibration and validation for HYDRUS such as soil moisture content, chemical oxygen demand, ammonia, and nitrate. Bench - scale drain fields (trenches), including soil moisture sensors embedded within the soil, were designed and operated. All dimensions of the trenches were based on the Michigan Criteria for Subsurface Sewage Disposal (Michigan Department of Public Health 1994) . Figure 2 is a photograph of the soil trenches u s ed for this research . The feedstock flowed by gravity into each trench. At the bottom of the trench, the treated water exite d through a water trap that did not allow air flow into the trenc h. The width of the trench was 60.96 c m (2 ft), selected to accommodate one inlet pipe. A typical septic soil treatment system had multiple inlet pipes with a maximum separation of 91 cm ( 3 f t ) bet ween the pipes. Its length was 121.9 c m (4 ft). The first layer of soil, before wastewater entered, contained 22.9 cm (9 in) of top soil. Wastewater was distributed in the next layer, having a 7.6 cm (3 in) depth of gravel followed by the inlet pipes and then 15.2 cm (6 in) of gravel. The depth of the sandy loam that served as the treatment media was 60.9 cm (2 ft). The loading required for the soil used in this research, sandy loam, is 10.2 L/day/m 2 (0.25 gal/day/ft 2 ) (Michigan Department of Public Health 1994) resulting in a flow rate of 7.57 L/day (2 gal/day) and an empty bed contact time (EBCT) of 60 days. Six CS616 soil moisture sensors , manufactured by Campbell 34 Scie ntific , were placed at two depths at 3 locations along the length of the trench. All soil moisture sensors were connected to a CR1000 data logger, manufactured by Campbell Scientific. Readings from the soil moisture sensors were monitored automatically usi ng the CR1000 data logger. Influent was fed three times every day to simulate the cleaning schedule at a typical fast - food restaurant. The influent and effluent were collected weekly and analyzed for COD, BOD 5 , Total Phosphorus (TP), Total Nitrogen (TN), a mmonia, and n itrate. Figure 2 . Photographs of soil trench es ( Dong et al. 2017) 35 Each trench received only one of the following feedstocks: 1. Domestic wastewater (Domestic WW) control that does not cause premature aging of the septic soil treatment system. 2. Domestic wastewater mixed with food wastewater treated with enzymatic pretreatment ( Domestic/Food WW ) typical test condition. 3. Food wastewater treated with enzymatic pretreatme nt ( Food WW ) high loading test condition. 3.2.1.4.2. Goodness of fit The most common method to evaluate the performance of HYDRUS CW2D model are model efficiency (E), index of agreement (IA), and root mean squared error (RMSE) (Anlauf and Rehrmann 2013; Wallach 2006; Wegehenkel, M. Beyrich 2014) . Model efficiency (E), original ly developed by Nash and Sutcliffe, is defined in Equation 3 (Nash and Sutcliffe 1970) . ( 3 ) Index of agreement (IA) was proposed by Will mott ( 1981) , a s defined in Equation 4. 36 ( 4 ) Root mean squared error (RMSE), is defined in Equation 5 (Anlauf and Rehrmann 2013) . ( 5 ) Where M is measured value, P is predicted value, and N is the number of observations. A range of E lies between - etween 0 and 1 represents an acceptable level of performance. An E value below 0 is considered an unacceptable level of performance (Moriasi et al. 2007) . An E va lue of 1 indicates that simulated and predicted value are equal to observed value. Phogat et al. (2016) suggested E > 0.12 and Qiao (2014) recommended E > 0 for evaluating the performance of their HYDRUS model. Both Analuf and Rehrmann (2013) and Arora et al. (2011) discussed that the acceptable quality should have E > 0.5 (Anlauf and Rehrmann 2013; Arora et al. 2011; Qiao 2014) . A range of IA lies between 0 and 1, and a value of 0 indicates no agreement between measured and simulated values. A value of 1 indi cates a perfect fit of observed to simulated values. The higher value of IA indicates better agreement between observed and simulated values. Phogat et al. (2016) reported acceptable quality of HYDRUS model is IA > 0.8 (Phogat et al. 2016) . RMSE measures the difference between measured and predicted values. Arora et al. (2011) discussed that generally lower RMSE indicates better agreement between measured and 37 predicted v alues. Shekofteh et al. (2013) reported a RMSE of 0.0135, Wang et al. (2016) reported a RMSE of 0.12, and Ramos et al. ( 2012) reported a RMSE of 0.030 for their satisfact ory model performance. (Ramos et al. 2012; Shekofteh et al. 2013; Wang et al. 2016) . T he criteria to evaluate satis factory model performance should include both relative error indices, such as E or IA, and absolute error measured, such as RMSE (Legates and McCabe 1999; Wegehenkel, M. Beyrich 201 4) . Therefore, this study evaluated the quality of the model using the following criteria; E > 0.5 , IA > 0.8, and RMSE < 0.01 4. The calculations for E, IA, and RMSE were performed using Rstudio software (Boston, MA). Details of the code are provided in Ap pendix B. 3.2.1.4.3. Calibration and validation procedure Calibration is described as the process of tuning by adjusting parameters and boundary conditions until the model result agrees with the experimental data. Validation is a process of quantifying the accuracy and credibility of the model . Calibration and validation procedures are describe d below. 1. Calibrate the water flow of the model using measured volumetric water content data of the first two dosing periods of a day. 2. Validate the water flow of the model using the measured volumetric water content data of the last dosing period of the same day used in calibration. 3. Calibrate the solute flow of the model using measured COD, ammonia , and nitrate data in Domestic WW and Domestic/Food WW conditions from 108 days to 170 days (62 days). Initial ly, ammonia was not measured in the laboratory study. After 108 th days, ammonia concentration was measured , which is needed for model calibration. 38 4. Validate the solute flow of the model using measured COD, ammonia , and nitrate data i n Domestic WW and Domestic/Food WW conditions from 171 days to 225 days (54 days). The soil saturated hydraulic and unsaturated hydraulic conductivity function are the most (Radcl iffe and Simunek 2010) . In order to calibrate the water flow, the soil hydraulic parameters need to be optimal . Direct measurement s of all the soil parameters are not always possible. An alternative indirect optimization using inverse modeling, a s commonl y used in hydrology modeling (Gupta et al. 2003) . Inverse modeli ng in HYDRUS uses the initial estimate of the parameters to perform the simulation and compares the simulation results to the observed experimental data. The model is then re - run with modified set of parameter . The process is repeated until the model ed dat a closely match the observed experimental data (Rassam et al. 2003) . Data from a total of 144 me asured volumetric water content s from a laboratory experiment were used for calibration. Figure 3 shows a screenshot of the inverse modeling routine in HYDRUS CW2D. The i nverse solution function optimizes the following soil parame ters : Ks (Saturated hydraulic conductivity), Alpha (Parameter in the soil water retention function), n (Parameter n in the soil water retention function), and I (Tortuosity parameter in the conductivity function). 39 Figure 3 . Inve rse modeling data input in HYDRUS CW2D ; X - time , Y - measured data ; Type - pressure head (1) , volumetric water conte nt (2), solute concentration (4) ; P osition - observation node number corresponding to where the volume tric water content is measured ; W eigh t - weight associated with a particular da ta point Unlike water flow calibration, HYDRUS CW2D solute flow do es not provide inverse modeling. The HYDRUS CW2D solute process parameters need to be manually calibrated by using a trial and error approach (Dittmer et al. 2005; Langergraber et al. 2007; Palfy et al. 2016; Palfy and Langergraber 2014; Pucher 2015) . The first step to calibrate the solute flow of the model is to determine the characteristic s of the wastewater. COD in HYDRUS CW2D model is divided into three fractions including readily CR, CS, and CI. Several approaches to fractionize COD into CR, CS, and CI are discussed. Palfy et al. (2016) reported CR:CS:CI ratio at 60:20:20, 40:40:20, and 30:60:10 for their study (Palfy et al. 2016) . Dalahmeh et al. (2012) conducted a study assuming that CR is being measured as the influent BOD 5 concentration and CI is 0 % of the observed effluent COD 40 level. The remaining COD was set to CS (Dalahmeh et al. 2012) . Henrichs et al. (2007) assumed the percentage of COD for CR, CS, and CI were 5 - 20 %, 60 90 %, 5 19 % of influent COD concentration, respectively. The CI was also considered to be 80 90 % of the observed COD effluent concentration (Henrichs et al. 2007) . Other studies set the CI value to 85% of the measured COD effluent concentration. The CR to CS ratio was then estimated to be approximately 2:1 (Dittmer et al. 2005; Henrichs et al. 2007; Toscano et al. 2009) . A preliminary test to fractionate CR, CS, and CI was conducted and the best estimate for the CI is 85 % of the measured COD effluent and CR to CS ratio being 2:1 of remaining COD. Wastewater composition such as COD (CR, CS, CI), ammonia - nitrogen, and nitrate - nitrogen were inputted in time variable boundary condition as a concentration (mg/L). Table 5 shows the values used to calibrate an d validate HYDRUS CW2D. The calibration and validation values in Table 5 are average concentrations from 108 days to 170 days (62 days) and from 171 days to 225 days (54 days), respectively. Since the COD concentra tion in domestic wastewater is from 99 to 445 mg/L, this model was calibrated for two different COD values of wastewater s using laboratory experiments : Domestic WW (102.1 mg/ L of COD) and Domestic/Food WW (519.1 mg/L of COD) (Brown et al. 1997; Dong et al. 2017; Hammerlund and Glotfelty 2016; Hossain 2008; Ronayne et al. 1982) . 41 Table 5 . Wastewater characteristic input parameter s for HYDRUS CW2D calibration and valida tion Parameter HYDRUS CW2D Symbol Calibration Validation Domestic WW Domestic/Food WW Domestic WW Domestic/Food WW CR (mg/L) cVal1 - 2 51.5 385 48.9 470 CS (mg/L) cVal1 - 3 25.7 193 24.4 235 CI (mg/L) cVal1 - 4 24.9 41 .0 27.4 41.9 Ammonia (mg/L - N) cVal1 - 8 28.7 28.9 30 .0 28 .0 Nitrate (mg/L - N) cVal1 - 10 5.18 1 .0 0 5.01 1.1 0 Once the characteristics of wastewater were determined, a djustment of kinetic parameters were performed using trial and error. According to previous studies, hy drolysis rate cons tant, lysis rate for microorganisms (XH, XANs/b ), maximum aerobic growth rate of XANs, maximum denitrification rate of XH, and f raction of CI generated in biomass lysis were adjusted using the calibration process. (Fuchs 2009; Heatwole and McCray 2007; Palfy et a l. 2016; Pucher and Langergraber 2018) . The procedure for adjusting the kinetic parameters is described below (Palfy et al. 2016; Pucher and Langergraber 2018) . 1. Run th e model using the standard parameter s of the HYDRUS CW2D biokentic model (Langergraber and Simunek 2005) . 2. Adjust the fraction of CI generated in biomass lysis value when the measured and simulated COD effluent concentration s are different . 42 3. Address the growth of bacterial groups, XH and XANs/XANb. using l ysis rates (b h , b ANs , b ANb ) by adjust ing each until steady state is reached (Palfy et al. 2016; Pucher and Langergraber 2017) . 4. Modify the maximum aerobic growth rate , XANs , when measured and simulated ammonia effluent concentrations are different (Pucher and Langergraber 2017) . 5. Adjust the hydrolysis rate and/or maximum denitrification rate for heterotrophic microorgani sms when measured and s imulated nitrate effluent concentrations are different (Pucher and Langergraber 2017) . By decreasing the hydrolysis rate, less organic matter is degraded in the upper layer of soil and more is available for the denitrification process as an electron donor. Time variable boundary condition was used for modeling domestic wastewater land application. This allowed the user to assign specific dosing times, dosing periods, and flux per day, and could repeat these conditions over the entire length of the simulation . In this study, dosing time followed a specific day in the laboratory experiment and the dosing period was set at 30 seconds. Flux was calculated as the t otal volume of applied wastewater divided by surface area. Tabl e 6 shows the time variable boundary condition input values for calibration and validation. Detail of input parameters in HYDRUS CW2D is provided in Appendix A. 43 Tabl e 6 . T ime variable boundary condition for calibration and validation Time (min) Flux (cm/min) 449.5 0 450 0.708 619.5 0 620 0.708 769.5 0 770 0.708 1440 0 3.2.1.5. Scenario s Because the literature review showed that current design c riteria for wastewater land application systems are generally based on limited empirical relationships, the calibrated and validated model was used to simulate multiple , common application scenarios. Included is the soil depth, hydraulic and organic loadin gs, and dosing frequency. Nitrate contamination of groundwater (concentration in groundwater >10 mg/L - N) occurs even in well - constructed and properly functioning domestic wastewater land application systems (Wilhelm et al. 1994) . Complete nitrification usually occurs within the first 30 cm of the soil depth, while complete denitrification ty pically does not occur in domestic wastewater land application systems (Beach 2001; Fischer 1999; Heatwole and McCray 2007) . Therefore, the first scenario was conduc ted to evaluate operation parameters that may enhance nitrate removal. The denitrification process requires carbon and anaerobic condition for denitrifying bacteria such as Heterotrophic bacteria , Thiobacillus denitrificans , micrococcus denitrificans , Pseu domonas , and Achromobacter (Carlson and Ingraham 1983) . Thus, increasing the dosing frequency may cause periodic saturated conditions that result in anaerobic soil conditions leading t o the 44 promotion of growth of denitrifying bacteria that will enhance nitrate removal. Multiple dosing frequencies including 3x dosing frequency, 6x dosing frequency, 10x dosing frequency, and continuous dosing were tested while maintaining constant hydraul ic and organic loading s . Table 7 shows the time variable boundary condition for multiple dosing frequencies. The flux was divided into respective each dosing frequency . 45 Table 7 . Time variable boundary con dition for dosing frequency Flux (cm/day) Time (day) 3x dosing frequency Time (day) 6x dosing frequency Time (day) 10x dosing frequency Time (day) Continuous 0.31215 0 0.31215 0 0.13889 0 1 1.062 0.3125 0 1019.52 0.3125 0 509.76 0.13924 305.86 0.4302 1 0 0.43021 0 0.20833 0 0.43056 1019.52 0.43056 509.76 0.20868 305.86 0.53438 0 0.53438 0 0.31215 0 0.53472 1019.52 0.53472 509.76 0.3125 0 305.86 1 0 0.61111 0 0.43021 0 0.61146 509.76 0.43056 305.86 0.68056 0 0.53438 0 0.6809 0 509.76 0.53472 305.86 0.69445 0 0.61111 0 0.69479 509.76 0.61146 305.86 1 0 0.68056 0 0.6809 0 305.86 0.75 000 0 0.75035 305.86 0.81944 0 0.81979 3 05.86 0.88889 0 0.88924 305.86 1 0 46 3.3. Result a nd discussion Results and discussion for monitoring, benefits, and modeling of domestic wastewater land application are presented in the following sections. 3.3.1. Monito ring of domestic wastewater land application Domestic wastewater land application has been used for many years and its performance is documented in many previous studies. Hence experimentation to determine effectiveness was not part of this project as the literature was used. Table 8 summarizes several manus cripts. The average removal efficiency of COD, BOD 5 , t otal p hosphorus (TP), and ammonia are over 60%. Typical domestic wastewater land application nitrifies most of the ammonia to nitrate but does not denitrify nitrate to nitrogen gas (Beach 2001; Fischer 1999; Heatwole and McCray 2007) . Since nitrate is highly mobile, nitrate leaching into groundwater is a concern. 47 Table 8 . Typical domestic wastewater land application treatment performance Parameter Influent Effluent Reference COD (mg/L) 99 - 445 29.3 (B rown et al. 1997; Dong et al. 2017; Hammerlund and Glotfelty 2016; Hossain 2008; Ronayne et al. 1982) BOD 5 (mg/L) 32 - 217 0 - 17 (Dong et al. 2017; Hossain 2008; Na tional Environmental Services Center 2013; Ronayne et al. 1982; Tchbanoglous et al. 2003) Total Phosphorus (TP) (mg/L - P) 4.5 - 30 0.01 - 4.9 (Dong et al. 2017; Hammerlund and Glotfelty 2016; Hossain 2008; National Environmental Services Center 2013; Tchbanoglous et al. 2003) Total Nitrogen (TN) (mg/L - N) 25 - 63.4 42.3 - 48 (Dong et al. 2017; Gross 2004; Hammerlund and Glotfelty 2016; National Environmental Services Center 2013; Ronayne et al. 1982) Ammonia (mg/L - N) 20 - 60 0.03 - 0.1 (Cui et al. 2003; Dong et al. 2017; Hossain 2008) Nitrate (mg/L - N) 0 - 10 39.1 - 42 (Brown et al. 1997; Dong et al. 2017; Hossain 2008; Ronayne et al. 1982; Tchbanoglous et al. 2003) 3.3.2. HYDRUS CW2D modeling The result s s to determine the depth requirement for the wastewater land application system s and treatment enhancement are discussed below . 3.3.2.1. Model c alibration and validation Model calibration was conducted by inverse modeling using volumetric water content measurement data. Figure 4 shows the comparison of measured and fitted for the HYDRUS 48 volumetric water content. The volumetric water content data measured in the first two dos es were used for the calibration and the remaining data was used for model validation. Figure 4 . Fitted HYDRUS to measured value u sing volumetric water content from Domestic WW Table 9 . Goodness of fit result of calibration and validation Model E IA RMSE Calibration 0.65 0.87 0.004 Validation 0.6 0.82 0.005 The calibrated and validated water flow model was evaluated by E, IA, and RMSE. As shown in Table 9 , all of the values were met the quality of modeling criteria, E > 0.5, IA > 0.8, and RMSE <0.014. 49 Table 10 shows the adjusted HYDRUS CW2D parameters for domestic wastewater land application modeling. Based on literature review, six parameters were considered in the calibration process for the solute flow. After trial and error, a maximum aerobic growth rate of XANs, maximum denitrification rate of XH, and fraction of CI in biomas s lysis were adjusted. Table 10 . Adjusted HYDRUS CW2D parameters for domestic wastewater land application modeling Parameter Description Unit Standard Adjusted k h Hydrolysis rate 1/d 3 - b h Lysis rate for XH 1/d 0.4 - b ANs/b Lys is rate for XANs/b 1/d 0.15 - µ ANs Maximum aerobic growth rate of XANs 1/d 0.9 0.45 µ dn Maximum denitrification rate of XH 1/d 4.8 3 .0 f BM,CI Fraction of CI in biomass lysis 1/d 0.02 0.01 50 T able 11 shows the simulated efflue nt concentrations before and after the fitting process. The table contains average s and standard deviation s of measured i nfluent, effluent, and simulated value s using standard and adjusted parameters from the calibration process . The measured values from d ay 108 to 170 (62 days) in Domestic WW and Domestic/Food WW were used. T able 11 . Simulated effluent concentrations before and after the fitting process for calibration Condition Inf/Eff Type of value HYDRUS CW2D Parameters COD (mg/ L) Ammonia (mg/L) Nitrate (mg/L) Domestic WW Influent Measured avg. (Std.) 102.3 (14.1) 28.7 (3.2) 5.18 (3.5) Effluen t Measured Avg. (Std.) 29.4 (5.4) 0.2 (0.1) 38.4 (3.3) Simulated Standard 28.7 0.1 35.6 Simulated Adjusted 26.8 0.2 35.5 D omestic/Foo d WW Influent Measured Avg. (Std.) 619.0 (142.3) 28.9 (3.2) 1.0 (0.6) Effluen t Measured Avg. (Std.) 48.1 (7.2) 0.5 (0.2) 29.4 (4.6) Simulated Standard 59.5 0.1 21.2 Simulated Adjusted 50.3 0.25 29.6 51 Table 12 shows the relative differences between the measured and simulated values in calibration process. The stimulated COD, ammonia, and nitrate values using standard parameters in Domestic WW conditions are not significantly different with the measured valu es. However, the stimulated COD, ammonia, and nitrate values using standard parameters in Domestic/Food WW conditions were - 19.2 %, 400 %, and 38.7 % different than measured values, respectively. In order to address these differences, the maximum aerobic g rowth rate of XANs, maximum denitrification rate of XH, and the fraction of CI in biomass lysis were adjusted. After adjustment , the relative difference of COD, ammonia, a nd nitrate in Domestic/Food WW decreased from - 19.2 % to - 4.37 %, from 400 % to 100 % , and from 38.7 % to - 0.68 %, respectively. The relative difference of COD and nitrate in Domestic WW condition was increased after an adjustment was made, but not significantly. This was a compromise to maximize the fit for both Domestic wastewater and Do mestic/Food WW conditions. Although the relative difference value for ammonia was high, the predicted model value was trace level, therefore , the model concentration was well predicted even for the standard parameters. Table 12 . R elative difference between measured and simulated values for calibration Condition HYDRUS CW2D COD Ammonia Nitrate Domestic WW Standard 2.44 % 100 % 7.87 % Adjusted 9.70 % 0 % 8.17 % Domestic/Food Standard - 19.20 % 400 % 38.70 % WW Adjusted - 4.37 % 10 0 % - 0.68 % 52 Table 13 shows the s imulated effluent concentrations before and after the fitting process for validation. The table contains average and standard deviations of measured influent, effluent, and simula ted values using standard and adjusted parameters in the validation process. The measured values from day 171 225 (54 days) in Domestic WW and Domestic/Food WW wastewater were used. Table 13 . Simulated effluent concentrations bef ore and after the fitting process for validation Condition Inf/Eff Type of value COD (mg/L) Ammonia (mg/L) Nitrate (mg/L) Domestic WW Influent Measured Avg. (Std.) 100.8 (6.9) 30.0 (4.4) 5.01 (4.8) Effluent Measured Avg. (Std.) 32.3 (8.3) 0.15 (0.1) 39. 4 (3.3) Simulated 29.2 0.2 36.7 Domestic/Food WW Influent Measured Avg. (Std.) 746.5 (111.1) 28.0 (4.5) 1.1 (0.9) Effluent Measured Avg. (Std.) 49.3 (13.1) 0.1 (0.1) 30.9 (3.3) Simulated 52.3 0.26 28 53 Table 14 show the relative differences between measured and simulated values in the validation process. Except for ammonia, the largest relative difference between the measured and simulated values were 10.6 %. Although the large number of relative difference was shown amon g ammonia, the difference between 0.1 and 0.26 is not significant with respect to field conditions. The model was successfully calibrated and validated for both water and solute flow using the laboratory experimental data. Table 14 . Relative difference between measured and simulated values for validation Condition COD Ammonia Nitrate Domestic WW 10.60 % 15.40 % 4.23 % Domestic/Food WW 6.09 % - 61.50 % 10.40 % 3.3.2.2. Scenario c apacity of wastewater land application With the mode l calibrated and validated, observation s on the effects of multiple hydraulic and organic loadings on COD treatment performance w ere conducted. The COD treatment performance was observed at depths of 15.24 , 30.48, 60.96 , 91.44, and 121.9 cm (0.5, 1, 2 , 3 , and 4 ft) with 1x, 2x, 3x, 4x, and 5x strength of hydraulic and organic loadings. This simulation was conducted with 3 dosing frequencies. A COD of 102.3 mg/L and COD effluent concentration at multiple depths of 150 th days were observed. The 150 th day was selected for the comparison because the COD effluent concentration did not significantly change after 60 days. Figure 5 shows the COD treatment performance at multiple depths with different strengths of hydraulic and organic loadi ngs . The efficiency of COD removal at 15.24 cm (0.5 ft ) and 30.48 54 cm (1 ft ) started to decrease after 2x and 3x strength of hydraulic and organic loadings, respectively. However, COD concentrations below 60.96 cm (3 ft) did not significantly change as load ing strength increased. This result shows that most of the COD treatment was within a 30.48 cm (1 ft) depth of soil, as confirmed by other s (Guilloteau et al. 993; Pan et al. 2017) . Consequently, sandy lo am at a soil depth of 60.96 cm (2 ft ) can treat 3 times higher hydraulic and organic loadings without decreasing the COD treatment performance. The s tate of Maryland r equi re s a minimum soil depth of 15 cm ( 6 in ) for drain field . According to the model result, 15 cm (6 in) can adequately treat the typical domestic wastewater loading . However, for households that produce higher strength wastewater, which can be caused by larg e amount of food waste and more frequent laundry, a 60 cm (24 in) soil depth is recommended to ensure the treatment while minimizing environmental impact. HYDRUS does not consider the growth of biofilm. Biofilm is a combination of microbial cells and an ex tra - cellular polymer matrix (Lazarova and Manem 1995) . T he biofilm is often referred to as the clogging zone (Siegrist and Van Cuyk 2001) , crust development (Magdoff et al. 1974) , biofilm (Dong et al. 2017b; Siegrist and Gujer 1985) , biomat, or biozone (Beach et al. 2005; Siegrist and Van Cuyk 2001) . Biofilm creates a hydraulic barrier, which encourage the distribution of wastewater throughout the field (Beach 2001) . However, e xcessive growth of biofilm can restrict the flow and decrease the hydraulic conductivity. Currently, the growth of biofilm and the change of the hydraulic propertie s by the biofilm are not considered in HYDRUS CW2 D . In order to address this limitation, understanding the thickness, hydraulic conductivity, and development rate of the biofilm is needed. High strength domestic wastewater may stimulate the growth of biofi lm because of its high carbon content. To prevent clogging, organic and hydraulic loading should be reduced. 55 Figure 5 . Simulated COD effluent concentration at 15.24, 30.48, 60.96, 91.44, and 121.9 cm (0.5, 1 , 2 , 3, and 4 ft) dept hs as increasin g loading strength (1x - 5x) on 150 th days of operation 3.3.2.3. Scenario t reatment performance enhancement Incomplete nitrate treatment in domestic wastewater land application system is already known. HYDRUS CW2D was conducted to observe the potent ial impact of dosing frequency along with the strength of the influent COD concentration on nitrate treatment performance. These two parameters may reduce oxygen levels in the soil creating anaerobic zones and increase the available carbon sources for deni trifying bacteria. A nitrate effluent concentration on the 150 th day at 60.96 cm (2 ft ) depth of soil was observed. As Figure 6 shows, the n itrate effluent concentration from the HYDRUS CW2D simulation did not sig nificantly change after 100 days, which was mathematically proven by a slope of - 0.00001. In Figure 7 , the growth of heterotrop hic bacteria supports the above since it did not significantly change after 100 days. 56 Figure 6 . Simulated nitrate effluent concentrations with multiple influent COD strength Figure 7 . Steady state condition of heterotrophic microorganism (XH) growth 57 Comparison of COD and nitrate efflu ent concentrations effected by dosing frequency and influent COD strength s were observed. Figure 8 shows the impact of the dosing frequency and the strength of influent COD concentration on its effluent concentration. Table 15 shows the relative differences from the control, which had a 3 X dosing frequency with 102 mg/L of COD. The removal efficiency in Figure 8 is calculated by subtracting continuous from 3 dos ing frequency to sh ow the effect of dosing frequency on COD removal . A s observed , there is a direct relationship between COD influent and effluent concentration - as the influent COD concentration , increase s the COD effluent concentration also increase s . Increasing the dosin g frequency was effective in reducing the COD effluent concentrations by a maximum of 5. 6 mg/L of COD. However, the removal of COD at an influent COD concentration of 102 mg/L was not impacted by an increase in dosing frequency. The removal efficiency incr ease as COD influent concentration increases. COD in the effluent might be the inert form, which is non - biodegradable. A possible explanation for a lower COD effluent concentration with higher dosing frequency is an increase in retention time for treatment correlated with the higher dosing frequency. Because the removal efficiency was not significant, increasing the dosing frequency to improve COD treatment is not recommended. 58 Figure 8 . Impact of dosing frequency and influent COD concentration on COD effluent concentration on the 150 th day at 60.96 cm ( 2 ft ) depth of sandy loam soil Table 15 . R elative difference from control for COD simulation s (3 dosing frequency with 102 mg/L of COD) 59 Figure 9 Error! Reference source not found. shows the i mpact of dosing frequency and COD concentration on nitrate effluent level estimation at 60.96 cm ( 2 ft ) depth of sandy loam soil . Table 16 shows the relative difference from the contro l (3 dosing frequency with 102 mg/L of COD). The removal efficiency in F igure 9 was calculated by subtracting continuous from 3 dosing frequency. At an i nfluent COD concentration of 102 mg/L , the nitrate was not impacted significantly by an increase in dos ing frequency. This may indicate that the influent COD concentration of 102 mg/L is not provide enough carbon for the denitrification process. However, the nitrate effluent concentration was lower when the influent COD concentrations was above 204 mg/L and the dosing frequency was more frequent . The most nitrate removal was observed with the highest concentrations of COD (612 mg/L) with 10 dosing frequency and continuous loading . There was no significant difference observed in nitrate efflu ent concentration s between a 10 dosing frequency and continuous condition. Higher dosing frequency may increase the moisture content of the soil, leading to a reduc tion in the oxygen content, providing optimal conditions for denitrification. Figure 10 shows higher soil water content was observed in 10 dosing frequency than 3 and 6 dosing frequencies. Additionally, a higher influent COD concentration provides more substrate for microorganisms, which may stimulate the denitrification process. For h ou sehold s with typical domestic wastewater, increasing the dosing frequency does not have a significant impact on nitrate removal . H owever, for h ousehold s an d facilities that produce higher COD strength wastewater, than 102 mg/L COD, 10 dosing frequency or continuous dosing can reduce effluent nitrate levels . Increasing dosing frequency may require upgrading the distribution system including pump ing capacity an economic analysis. Currently , the mod el did not consider factors such as precipitation, seasons, topography, and growth of biofilm. However, the model result shows estimated treatment performance 60 affected by dosing frequency and influent COD concentration. When the model considers the above factors, the predicted value of the model may be different bu t might not be significant. F urther , calibration and validat ion of the model using different soil type , precipitation, seasons, topography, a nd growth of biofilm are needed. Seasons can impact nitrate removal. As the soil temperature increases, m icroorgani sm activity also increases, which can promote nitrification and denitrification processes . As the soil temperature decreases, microbi al activity decreases, which may slow the nitrification and denitrification processes. Therefore, the nitrate removal effic iency will decrease in winter. During long - term operation, the drain field can be clogged by biofilm. When the drain field is clogged, the soil beco mes anaerobic, which is the optimal condition for the denitrification process. However, a clogged drain fie ld cannot handle the design hydraulic loading, potentially resulting in an overflow and unpleasant odors. Therefore, replacement or resting the drain field for a year is recommended. 61 Figure 9 . Impact of dosing frequency and COD concentration on nitrate effluent level estimation at 60.96 cm ( 2 ft ) depth of sandy loam soil Table 16 . R elative difference from control for nitrate simulation (3 dosing frequency with 102 mg/L of COD) Influent COD Concentrati on (mg/L) Relative difference from control (mg/L) 3 Dos es 6 Dos es 10 Dos es Continuous 102 0 - 0.3 - 2.3 - 2.2 204 0.1 - 1.5 - 5.9 - 5.2 306 - 0.8 - 4.2 - 8.3 - 7.2 408 - 1.9 - 6 - 10.5 - 10.7 510 - 3.6 - 7.6 - 12.9 - 12.7 612 - 4.8 - 9.5 - 14.3 - 14.5 62 Figure 10 . Simulation of the impact of volumetric water content in soil by different dosing frequencies at 150 th day; 3 dosing (left), 6 dosing (middle), and 10 dosing (right) 3.4. Conclusion In conclusion, a domestic wastewater land application system is effective in treating COD, BOD 5 , TP, and ammonia but not nitrate. In comparison to a conventional wastewater treat ment system , domestic wastewater land application can save energy for the treatment , consequently reducing GHG emissions . Besides th ese benefits, lower capital and operation cost, less maintenance requirement, and no chemicals requirements are other benefits . HYDRUS CW2D was successfully calibrated and validated using measured volumetric water content from a laboratory experiment. Most of the COD treatment in domestic wastewater land application syst em occurs within 15.24 cm (1 ft ) depth of sandy loam soil. Two feet depth of soil for domestic wastewater land application system is ideal with the consideration of a safety factor. The mode l simulation shows the potential nitrate removal by increasing both dosing frequency and influent COD concentration. 63 Future stud ies are warranted. The mode l should be verified in a long - term study. In addition, the model should be calibrated and validat ed using different types of soil, weather condition, and hydraulic and organic loadings . Once the model is calibrated and validat ed , multiple scenario to determine the best operation strategies should be performed. The result s should be provided as an index, which is an integrated approach. The index can be beneficial for onsite wastewater engi neers/designers and regulators to determine the depth operation strategies. 64 Chapter 4 . Food processing wastewater land application This chapter provides background i nformation on the food processing wastewater irrigation demonstration site. Monitoring strategies to prevent and minimize nitrate and metal leaching into the groundwater are first discussed followed by the evaluation of site condition and the benefits of food processing wastewater land application such as freshwater saving, nitrogen and phosphorus reuse, energy savings, and GHG reduction associated with the energy saving. In addition, the potential use of HYDRUS CW2D is discussed. 4.1. Introduction In the United States, 1 trillion gallons of wastewater are produced annually from food processing industry (Aryal 2015) . Food process ing wastewater characteristics vary depending on the facility, technology, and type of food being processed . Typically, food processing wastewater includes organic carbon, nutrients, suspended solids, descaling chemicals, food additives, salts, and equipme nt cleaners (Safferman et al. 2007) . The characteristics of food processing wastewater are summarized in Table 17. 65 Table 17 . Characteristic of food proc essing wastewater Type of processor COD (mg/L) BOD 5 (mg/L) TP (mg/L - P) TN (mg/L - N) Nitrate (mg/L - N) Reference Milk and dairy products 1,025 4,841 154 663 (Christian 2010) Meat 1,684 863 328 2,744 Slaughterhouse 1,000 - 6,000 1,000 - 4,000 80 - 120 250 - 700 (Tritt and Schuchardt 1992 ) Milk 2,833 1,216 77 70 (Konieczny et al. 2005) Meat 2,392 646 1 3 80 Fish 3,017 914 43 181 Confectionery 530 - 2,620 (Di Berardino et al. 2000) Starch 6,222 (Deng et al. 2003) Poultry 364 - 1,219 2.9 - 13.5 (Pierson and Pavlostathis 2000 66 Table 17 . Characteristics of food processing wastewater ( ) Type of processor COD (mg/L) BOD 5 (mg/L) TP (mg/L - P) TN (mg/L - N) Nitrate (mg/L - N) Reference Fish 326 - 1,432 3,500 117 (Chowdhury et al. 2010) Olive oil 220 - 400 (Wang, Huang, and Yuan 2005) Slaughterhouse 2,870 (Sayed et al. 1988) Potato 5,978 12,489 1,277 308 0.22 (Dornbush, Rollag, and Trygstad 1976) Pe ar 3,050 2,040 (Esvelt 1970) Peach 2,150 1,810 Apple 1,400 - 1,520 950 - 1,230 Wine 300 - 30,000 1 - 225 (CVRWQCB 2005) Apple 9,000 (Van Ginkel et al. 2005) Potato 21,000 Confectionery 600 - 20,000 67 This wastewater typically contains nutrients and water, valuable resources for crop production. Producing a valuable crop commodity has the benefits of reducing the use of fresh water, commercial fertilizers, and energy by eliminating the need for a traditional wastewater treatment facility and , consequently , reduces GHG emission . The calculation was conducted using the data from 2016 at the long - term food processing wastewater land application site as a demonstration to show the potential benefits of food processing wastewater land application. The average BOD 5 concentration of the food processing wastewater and groundwater from the sampling well s were 680 mg/L and 2 mg/L, respectively. The total BOD 5 removal was 459,427 kg (1,012,864 lb). To remove 0.45 kg of BOD at a traditional activated sludge wastewater treatment facility with a flow l ess than 3.78 millio n liters/day (1 million gallons/day), 4.1 kWh of energy is required (NYSERDA 2007) . Therefore, 4,200,000 kWh electricity was saved in 2016. This results in a GHG reduction of 3 , 126 metric tons, which is equivalent to the GHG emission from 669 pa ssenger vehicles driven for one year with assumptions of 22 miles/gallon and 11,443 miles/year driven (USEP A 2016) . Nitrogen and phosphorus are essential fo r crop production. Applied nitrogen loading at the demonstration site was estimated at 41.7 g - N/m 2 ( 371.9 lb - N/acre ) in 2016. The amount of nitrogen required for the crop yield was estimated at 22.4 g - N/m 2 (2 00 lb - N/acre ) . This indicates that the excessive nitrogen is applied on the land. The higher than required nitrogen loading is of concern , especially when the crops are not actively growing. The applied nitrogen was adequate for crop production and the yield was expected. In addition to not having to purchase commer cial fertilizers , savings results from minimiz ing fuel us e for tractors to apply fertilize and tr uck s to transport fertilizer s , which result in GHG reduction. 68 Although there are many benefit s, improper operation can result in nitrate leaching or metal mobilization into groundwater (Dong et al. 2017a; Julien and Safferman 2015; Redding 2012) . The USEPA provides water quality standards. The standards include a maximum con tamination level s , which are regulated, and a secondary maximum contamination level s , which are not regulated but of concern. The main concerns in the study were the concentration s of nitrate, arsenic, manganese, and iron in the groundwater. Maximum allowa ble level of contamination for nitrate is 10 mg/L - N (USEPA 2017d) . Secondary maximum contaminant level for arsenic, iron, and manganese are 0.01 mg/L, 0.30 mg/L, and 0.05 mg/L, respectively (USEPA 2017e) . T he monitoring of the demonstration site has been ongoing for 8 years. M on itoring strategies include three parts; tracking hydraulic and organic loading s , using real - time soil sensor clusters to monitor soil condition s , and groundwater monitoring for verification that impact are not occuring . In addition, visual observation s and selected soil sampling were conducted to qualitatively assess site conditions. In particular, areas that appear ed to be less optimal for irrigating were delineated to evaluate the cause. Soil characteristics that were analyzed included texture, compaction , infiltration, uniformity , and localized water condition. Current design criteria are based on limited evidence and have not been fully developed for food processing wastewater. As Table 17 shows , the characteristic s of food processing wastewater vary dep end ing on the type of plant and process es . The COD concentration ranges from 220 to 20,000 mg/L. Because of the diverse nature of the wastewater, it is challeng ing to develop design criteria. HYDRUS CW2D modeling may be a valuable design tool to simulate m ultiple operation strategies and predict the treatment performance , including carbon degradation, nitrification , and denit rification. The model result can provide operation al strateg ies to maximize the treatment while minimizing environmental impact s . 69 4.2. Mate rials and methods Unlike domestic wastewater, there is not a lot of literature on the application of food processing wastewater and that available is not consistent. A detailed summary of the literature found that the reported acceptable hydraulic loading ranged from 2.53 to 14.97 liter/m 2 /day (2,700 to 16,000 gal/acre/day) and the organic loading ranged from 4.48 to 201.75 g/m 2 /day (40 to 1,800 lb BOD/acre/day) (Mokma 2006) . This study also found that there were no clear scientific basis for any of these values (Mokma 2006) . Consequently, this research first reports on the methodology used to demonstrate the feasibility of a food processing wastewater irrigation system and then examines the b enefits compared to a traditional wastewater tr eatment plant. Design approaches using HYDRUS CW2D modeling w ere also explored. 4.2.1. Monitoring of food proces sing wastewater land application Comprehensive monitoring has been continuously ongoing for over 8 years at the demonstration wastewater irrigation facility. Daily hydraulic and organic loadings were tracked. level, and temperature. Groundwater quality was monitored quarterly. Characteristic s of the soil such as its texture, compaction , infiltration , and localized high water condition were also monitored. 70 4.2.1.1. Bac kground of demonstration site Detail information of the demonstration site is described in below, which was extracted from the Water Environmental Federations Technical, Exhibition and Conference (WEFTEC) proceedings ( Dong et al. 2017) . The overview map of the food processing wastewater land application demonstration site is shown in Figure 11 . The size of the demonstration wastewater land application site is 675,825 m 2 (167 acres) . The type of food processing plant at the demonstration site is canning. Corn and alfalfa are grown on the wastewater irrigation site and used as animal feed. The wastewater produced by the food processing plant varies in quantity and characteristic s depe nding on processing at the production plant , technology, type of food, and facility. Before land application, the wastewater is screened and then flows into an aerated equalization tank ( Fig ure 12 ). In this study, the food processing wastewater land application system refers to the sequential treatment of screening, equalization tank, and land application. Because the wastewater is applied to the surface using a center pivot irrigation system, localized runoff is co llected by a Hicke nbottom pipe (Fairfield, IA) ( Figure 13 ), flows into the storage tank, and is then applied on secondary irrigation sites using solid set distributors. This ensures that no water is transported off site by surface movement. The soil types at this the site vary and includes loamy sand, sandy loam, and sand, depending on the location and depth. Average groundwater depth is 9.1 m (30 ft). 71 Figure 11 . Overview map of demonstration site Fig ure 12 . Wastewater flow of the demonstration site 72 Figure 13 . Hickenbottom pipe for inducing and collecting run off 4.2.1.2. Monitoring st rategies Monitoring strategies included tracking hydraulic and organic lo ading s , observing soil condition s using soil sensor clusters, and testing groundwater quality. Each are discussed in the s ubsequent paragraphs. 4.2.1.2.1. Hydraulic and organic loadings Hydraulic loading is tracked daily at the demonstration site. The influent waste water is characterized biweekly. Organic loading was calculated by multiplying the hydraulic loading by the BOD 5 concentration. Table 18 show the average, maximum, and minimum characteristics of wastewater at the demonstration sit e. 73 Table 18 . Characteristic s of wastewater at the demonstration site Parameters Average (mg/L) Maximum (mg/L) Minimum (mg/L) pH (in S.U.) 7.25 7.7 6.3 COD 1,156 2,900 405 BOD 5 651 1,480 153 Nitrogen, total kjeldahl 40.1 64 .4 17.3 Ammonia - nitrogen 3.21 6.4 1.4 Nitrate - nitrogen <0.1 3.3 0.4 Nitrite - nitrogen 0.4 0.4 0.4 Phosphorus, total (as P) 8.59 23.4 3.28 Sodium, total 59.1 416 28.6 Calcium, total 87 87 87 Iron, total 1.3 2.75 0.3 Magnesium, total 35.7 35.7 35.7 M anganese, total 0.1 0.18 0.06 Potassium, total 534.7 717 33.5 Chloride 342.1 526 196 4.2.1.2.2. Soil sensor cluster A remote monitoring system consisting of five soil sensor clusters were used . The locations are shown in Figure 11 . Figure 14 shows an overview of the soil sensor cluster. Each consist of a yagi antenna, surge protector, solar panel, 12v battery, CR 1000 datalogger, RF 401 radio, 3 soil response thermistor reference oxygen sensors, and 3 volumetric water content sensors. All of the parts were manufactured by and purchased from Campbell Scientific (Logan, UT) except for the soil response thermistor reference oxygen sensors , which are from Apogee instrument (Logan, UT)) . Figure 15 shows the composition of the devices in the weather resistant 74 enclosure. In order to communicate with the data loggers, each soil sensor cluster is fitted with a RF401A, 900MHz radio that transmit s up t o one mile with an omnidirectional antenna or up to 10 miles with a higher gain directional antenna. Regular maintenance was conducted twice a year, including examing sensor wire connections, moisture build up in the antenna connector, antenna direction, a nd solids accumulation on the surface of solar panel. Moisture in the antenna connector was found to interfere with the signal strength resulting in communication disruptions and occasional failures to the soil sensor cluster. The battery in each cluster w as routinely replaced every 2 - 3 years to avoid catastrophic failures. Figure 14 . Overview of sensor cluster at the demonstration site 75 Figure 15 . Composition of soil sensor cluster The sensors were installed at depths of 30.48, 60.96, and 91.44 cm (1 , 2, and 3 ft), as shown in Figure 16 . Measurements were taken every 5 minutes and average daily values recorded. From these measurements, it can be determined if the soil is aerobic or anaerobic, which provides information on the potential for nitrat e and heavy metal mobilization. 76 Figure 16 . Sensors installed at depths SO - 110 so il response thermistor reference oxygen sensor were manufa ctured by Apogee (Logan, UT), designed for continuous gaseous oxygen measurement in air, soil/porous media, sealed chambers, and in - line tubing. The sensor s consists of galvanic cell sensing e lement (electrochemical cell) T eflon membrane , and reference tem perature sensor (Apogee Instruments 20 16) . Th e sensor s measure from 0 to 100 % O 2 , and ha ve a standard response time of 60 seconds. The CS 616 water content reflectometer measures the volumetric water content from 0% to saturation in a soil using two 30 cm (11.8 in) stainless steel rods. The variability between probes is ± 0.5% volumetric moisture content in dry soil and ± 1.5% volumetric moisture content in typical saturated soil (Campbell Scientific 2014) . 77 LoggerNet softwa re, was developed by Campbell Scientific (Logan, UT) and was used for programing, communication, and data retrieval between the data logger and the computer. Each day LoggerNet downloads measurement data from all sensors and saves it in a CSV file format. A CSI Web Server, which was also developed by Campbell scientific (Logan, UT), allows users to view the saved data via the web browser. 4.3.1.2.3. Groundwater monitoring Strategically positioned monitoring wells allow for the quarterly collection of groundwater sa mples for carbon, metal, and nutrient testing. In the specific study area, a total of 16 m onitoring wells (MWs) (Figure 1 7 ) are upstream and down stream of each field . Location s of the MWs are shown in Figure 1 1 . D omestic wells downstream of the site are al so monitored to verify Figure 17 . MW 103 at the demonstration site 78 4.2.1.3. Site evaluation In order to evaluate the long - term site function, visual observation, soil characteristics, uniformity, and localized high water table conditions were conducted. 4.2.1.3.1. Visual observation Q uarterly to biannual visual observations and selected soil sampling were conducted to qualitatively evaluate site conditions. In particular, area s that appeared to be less optimal for irrigating were evaluated to determine the cause. Particular attention was focused on non - optimal areas as to identify potential causes. These areas are defined by standing surface water and low crop growth. Figure 19 shows examples of optimal and non - optimal areas. Delineation of non - optimal areas were condu cted using a Juno 3B GPS Handheld ( Figure 18 ) , manufactured by Trimble Inc. (Sunnyvale, CA). T he Juno 3B GPS Handheld is also a computer that can integrated a GPS and digital camera. Figure 18 . Juno 3B GPS Handheld 79 Figure 19 . Optimal and non - opti mal areas 4.2.1.3.2. Soil texture Soil samples were coll ected at the surface, 30.48, 60.96, and 91.44 cm (1 , 2 , and 3 ft) depth of soil in optimal and non - optimal areas of study Fields 1, 2 , and 3 . Soil texture analysis was conducted via M ichigan S tate U niversity Soil and Plant Nutrient Laboratory (East Lansing , MI). 4.2.1.3.3. Soil compaction Soil compaction analysis was conducted using a soil compaction tester, manufactured by AgraTronix ( Figure 20 ) (Streetsboro, OH). This tester can measure up to 60.96 cm ( 24 in ) depths and provide testing ranges such as green 0 - 1,378 kpa (0 - 200 psi) for good growing condition, yellow 1,378 - 2,068 kpa (200 - 300 psi) for fair growing condition, and red 2,068 kpa (300 psi ) and above for poor growing condition. 80 Figure 20 . Soil compaction meter by AgraTronix 4.2.1.3.4. Infiltration Soil infiltration test was conducted using an ASTM 3385 d ouble r ing i nfiltrometer ( Figure 21 ), manufactured by Turf - tec International (Tallahassee, FL). Th e 15.24 cm ( 6 in ) ring was driven into the ground up to the 7.62 cm (3 in ) mark by using a block of wood and mallet. Water was added up to 2.54 cm (1 in ) , 444 mL, into the outer ring, then the inner ring. Once filled, the 81 ring s w ere covered by plastic wra p. The time to infiltrate 2.54 cm ( 1 in ) of water was recorded . The outer ring helps the water in the inner ring flow down, not disperse to the side. Measurements were conducted on optimal and non - optimal areas for comparison. Statistical analysis was cond ucted using a t - test to compare the infiltration rate in optimal and non - optimal areas . The test was conducted using Rstudio software (Boston, MA) and the code is described in Appendix B. Figure 21 . Infiltrometer by Turf - tec Int ernational (Tallashassee, FL) 82 4.2.1.3.5. Uniformity The central pivot irrigation nozzles can clog from suspended solids in the wastewater and freeze during cold weather. This can restrict the flow leading to non - uniform irrigation. Uniformity is critical as it is h ypothesized that metal mobilization can result and impact ground water from small disjointed locations within the field, especially those that have compacted soil or localized high water table condition. Further, uniform application of wastewater is critic al to plant health. Consequently, irrigation uniformity te sting was conducted on fields 1 and 3 . Uniformity was conducted following the ANSI/ASAE S436.1 standard. The procedure includes placing 946 ml ( 32 oz ) disposable soda cups at a 6.1 m (20 ft) distanc e apart in a straight line outward from the pivot elbow. Time was recorded from the point when water touches the cup from the pivot elbow to the point when the water stops hitting the same cup. The water level in each of the cup is measured and recorded . The s ystem uniformity coefficient is a numeric determination of the overall performance of even distribution in an irrigation system. Typically, the coefficient of 85 or higher is considered well distribution. The system uniformity coefficient of 80 below requires an adjustment to the sprinkler system (Kelley 2014) . 4.2.1.3.6. L ocalized high water table condition In order to observe localized high water table, a geoprobe ( Figure 22 ) and auger were used. This equipment allowed for observations down to approximately 182.88 cm ( 6 ft ) . 83 Figure 22 . Geoprobe 4.2.3. HYDRUS CW2D modeling The procedure for model c alibration and validation were performed following the metho d discussed in chapter 3.2.2.4. Minimum, average and maximum of COD concentration at the demonstration site were 405 mg/L, 1,156 mg/L, and 2,900 mg/L, respectively. Similar to the domestic wastewater modeling calibration process, two strengths were conside red, Domestic/Food WW and Food WW from laboratory experiment. Average COD concentrations in Domestic/Food WW and Food WW were 661.5 mg/L and 2900 mg/L, respectively. Calibration using the Food WW condition was attempted but could not be completed due to th e limited resource . Further research is needed. Therefore, the model was calibrated an d validated using the Domestic/Food WW condition. Once the model wa s c alibrated and validated, . The procedure is described in Chapter 3 .2.2.4.2. 84 4.2.3.1. Scenario s In a food processing wastewater land application system, h ydraulic and organic loadings are generally fixed by the food processing production facility , thus dosing frequency is the only operatio nal parameter to maximize the treatment while protecting the environment. Different dosing frequency , including 3, 6, and 10 dos es, were simulated. Multiple strength of hydraulic and organic loadings, 1, 2, and 3 times of strength , w ere also considered bec ause wastewater composition and volumes from food processor is highly variable. Effluent concentrations of COD, ammonia, and nitrate were observed to understand the impact of dosing frequency and hydraulic and organic loadings on carbon degradation, nitrif ication, and denitrification. Table 19 shows the time variable boundary condit ion s for multiple strength s of hydraulic and organic loadings for scenario simulation. Time variable boundary condition multiple dosing frequency were p reviously described in Table 7. Each loading flux was calculated by multiplying the original flux by the respective intensity of loading. Table 19 . Time variable boundary condition for multiple loading strengths Time (day) 1x loa ding 2x loading 3x loading flux (cm/day) flux (cm/day) flux (cm/day) 0.31215 0 0 0 0.3125 0 1019.52 2039.04 3058.56 0.4302 0 0 0 0 0.43055 1019.52 2039.04 3058.56 0.53437 0 0 0 0.53472 1019.52 2039.04 3058.56 1 0 0 0 85 4.3. Result and discussion Results and discussion for monitoring, benefits, and modeling of food processing wastewater land application are presented in the following sections. 4.3.1. Monit oring of food processing wastewater land application The e valuation of monitoring strategies and site cond itions at the long - term demonstration site are discussed. 4.3.1.1. Monitoring strategies Monitoring strategies include tracking hydraulic and organic loadings, monitoring the soil condition via soil sensor clusters, and testing ground water quality. 4.3.1.1.1. Hydraulic a nd organic loading Figure 23 shows the annual hydraulic loading in million gallons and organic loading in lbs of BOD 5 /acre/day. A past concern at the demonstration site was metal mobilization resulting from the ana erobic conditions associated with excessive microbial growth due to the high concentration of the applied carbon. As a solution to the problem, implemented involvement, hydraulic loading was lowered through a reduction in water use within the food processing plant. The decrease of hydraulic loading decreased metal mobilization at the site. 86 Figure 23 . H ydraulic and o rganic loading at the demonstration site 4.3.1.1.2. Soil sensor cluster To ensure that the soil remains aerobic t o prevent metal mobilization , f ive soil sen sor clusters were installed and content, and groundwater quality for over 8 years. Figure 24 shows daily v olumetric water content from soil sensor Cluster 1 . Daily volumetric water shows that soils never reached saturation level ( ~ 30%), except briefly in early 2015. This indicates that the upper levels of the soil are generally aerobic, preventing the growth o f metal reducing microorganisms. Figure 25 shows the daily oxygen content f or soil sensor Cluster 1 . Daily oxygen level in all depths confirms that aerobic condition is maintained throughout the years. The generall y aerobic upper level of soil, indicated by the dissolved oxygen concentration, prevents metal leaching but also limits denitrification. T he oxygen sensors showed similar patterns as the moisture sensors and the greater expense and maintenance did not add significant value in having this more direct environmental measurement. 87 Figure 26 shows the temperature s at different depths in Field 1. It can be observed that the soil never freezes, even though typical soil outs ide of the irrigation zone is frozen to de pths greater than 107 cm (42 in ) during the winter. This indicates that the microbial population remains active year round. Figure 24 . Daily volumetric water content at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s f o r Cluster 1 Figure 25 . Daily oxygen level at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s f o r Cluster 1 88 Figure 26 . Daily temperature at 30. 48 , 61. 96 , and 91. 4 4 cm (1, 2, and 3 ft ) depth s f o r Cluster 1 Figure 27 shows daily volumetric water content f or soil sensor Cluster 2 . The d aily volumetric water content was higher than Cluster 1 . Figure 28 shows the daily oxygen content for soil sensor Cluster 2 . The oxygen sensors in this cluster also malfunctioned at September in 2011 and November 2013 . Regardless of depth, oxygen levels fluctuated and were consistently lower in comparison to the daily oxygen level data f or Cluster 1. Soils with a higher volumetric water content and lower soil oxygen level are more likely to cause metal mobilization but less likely to encourage denitrification. Figure 29 shows the temperatures at different depths in Field 2 . Similar to the results obtained in soil sensor Cluster 1 , t he soil temperature never falls below freezing . This indicates that microbial population within Field 2 remains active year round . 89 Figure 27 . Daily volumet ric water content at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s r o r Cluster 2 Figure 28 . Daily oxygen level at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s for Cluster 2 Figure 29 . Daily temperature at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s for Cluster 2 90 Two soil sensor clusters are installed in Field 3. Soil sensor Cluster 3A is located on optimal condition of F ield 3 . Figure 30 , 3 1 , and 3 2 s how the daily volumetric water content, oxygen content, and temperatures, respectively.. Figure 30 shows daily volumetric water content from soil sensor Cluster 3A . Daily volumetric water shows that soils reached s aturation level ( ~ 30%) during a few time periods . This indicates that the upper levels of the soil are generally aerobic, preventing the growth of metal reducing microorganisms. Figure 3 1 shows the daily oxygen content from soil sensor Cluster 3A . Daily oxygen level in 30.48 (1 ft) depth shows aerobic co ndition . Interestingly, at a depth of 91.44 cm (3 ft) , levels fluctuated and did not show the same trend as the volumetric water content. The oxygen sensor at a depth of 91.44 cm ( 3 ft.) depth malfunction ed . The oxygen sensor also measure temperature. Figur e 32 shows temperature at different depths in Field 3 . It can be observed that the soil never freezes. This indicates that the microbial population remains active year round. 91 Figure 30 . Daily volumetric water content at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s for Cluster 3A Figure 31 . Daily oxygen level at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s for Cluster 3A Figure 32 . Daily temperature at 30. 48 , 61 . 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s for Cluster 3A 92 Cluster 3B wa s placed in the non - optimal area of Field 3 . Figure 33 shows the daily volumetric water content. Interestingly, high levels of volumetric water content s were ob served at all depths. One possible cause is the soil type, which was loamy sand at 30.48 and 61.96 cm (12 and 24 in) and sandy loam, at 76.2 cm ( 30 in ) . Because sandy loam has less porosity than loamy sand, the water flow might be restricted. Restriction o f water flow results in a high volumetric water content. High levels of volumetric water indicates overall anaerobic conditions in the upper levels of soil , which could result in metal mobilization. Groundwater analysis in MW 103 (within Field 3 ) confirmed that manganese was present . However, anaerobic conditions are more conducive for denitrification. Figure 34 shows the daily oxygen content from soil Cluster 3B . The oxygen content is generally lower, which indicates possible deni trification, but may promote metal mobilization. Temperatures at different depths f or Cluster 3B are shown in F igure 35 . Consistent temperatures above freezing indicate that the microbial population remains active year - round . 93 Figure 33 . Daily volumetric water content at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s for Cluster 3B Figure 34 . Daily oxygen level at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s f o r Cluster 3B Figure 35 . Daily temperature at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s f o r Cluster 3B 94 Figure 36 shows daily volumetric water content f or soil Cluster 4. The volumetric water content at a depth of 30.48 cm (1 f t) were generally lower than the volumetric water content at depths of 61.96 and 91.44 cm (2 and 3 ft). The oxygen content ( Figure 37 ) at a depth of 30.48 cm (1 ft) remained aerobic throughout the years, however , oxygen levels at 61.96 and 91.44 cm (2 and 3 ft) fluctuated. Fluctuation could have been caused by high localized water conditions. When the soil was excavated for the installation of the sensors , standing water and heavy wet soil w ere observed at a depth of 121.92 cm (4 f t), however, results from Chapter 4 indicate that treatment likely occurred before the water reached that depth. Figure 38 shows the temperatures at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 feet) depth s from Cluster 4. Given that t he soil never freezes, the microbial population remains active year round . In summary, s ensor readings at the demonstration site indicated a substantial range of soil moisture conditions and oxygen levels, indicating environments that were aerobic to anaer obic. The sensors were also very responsive. Moisture sensors proved to be good indicators if the soil was aerobic or anaerobic, based on a comparison to the oxygen sensors. In addition, the moisture sensors are less complex resulting in better reliability , less maintenance, and significantly less cost. Additionally, temperatures at the lower depths did not drop below freezing indicating that the microbiological environment was at least somewhat active all year. However, as temperature decreases, the activi ty of microorganisms decrease. 95 Figure 36 . Daily volumetric water content at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s for Cluster 4 Figure 37 . Daily oxygen level at 30. 48 , 61. 96 , and 91. 4 4 cm (1, 2, and 3 ft ) depth s for Cluster 4 Figure 38 . Daily temperature at 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 ft ) depth s for Cluster 4 96 4.3.1.1.3. Groundwater monitoring Figure 39 shows the est imated groundwater flow at the demonstration site. In general, the groundwater flows from the north to south. Figure 39 . Estimated groundwater flow at the demonstration site 97 The demonstration site is located in an area with heavy agricultural activity, which may be one cause of the relatively high nitrate concentrations in the groundwater throughout the area. Figure 41 , Figure 42 , and Figure 43 sho ws the groundwater quality for MW 7 ( upstream of Field 1 ), MW 101 (within Field 1 ), and MW 15 ( downstream of Field 1 ) , respectively. Monitoring well locations for Field 1 are shown in Figure 40 . The nitrate concentration in MW 15 was generally equal to or less than 10 mg/L - N of nitrate. Interestingly, this indicates a decrease of nitrate as it progresses across the wastewater irrigation site. Metal mobilization into groundwater was not detected in Field 1 . Figure 40 . MW locations for Field 1 98 Figure 41 . Groundwater quality f or MW 7, upstream of Field 1 Figure 42 . Groundwater quality f or MW 101, within Field 1 Figure 43 . Gr oundwater quality f or MW 15, downstream of Field 1 99 For Field 2 , the groundwater quality of MW 24, (up stream of Field 2 ), MW 102 (within Field 2 ), and MW 12 ( downstream of Field 2 ) are shown in F igure 4 5 , Figure 4 6 , and Figure 4 7 , respectively. Monitoring w ell locations for Field 2 are shown in Figure 44 . Relatively high nitrate concentrations were detected in MW 24 and MW 102. The nitrate concentrations in MW 12 generally maintained equal to or less than 10 mg/L - N of nitrate. This again indicates that nitra te levels were reduced across the site. Relatively high manganese concentrations were initially detected in MW 102 (within Field 2 ), but gradually decreased with time. This decrease correlates to the reduction of hydraulic loading. In general, low concentr ations of arsenic, manganese, and iron were found in MW 12 (downstream of Field 2 ). Figure 44 . MW locations for Field 2 100 Figure 45 . Groundwater quality f or MW 24 (upstream of Field 2 ) Figure 46 . Groundwater quality f or MW 102 (within Field 2 ) Figure 47 . Groundwater quality f o r MW 12 (downstream of Field 2 ) 101 In Field 3 , the groundwater quality of MW 28 (up stream of Field 3 ), MW 20R (within Field 3 ) , MW 103 ( within Field 3 ), and MW 23 (downstream of Field 3 ) are shown in Figure 49 , Figure 50 , Figure 51 , and Figure 5 2 , respecti vely. Monitoring well locations for Field 3 are shown in Figure 48 . Relatively low metals and equal to or less than 10 mg/L - N of nitrate were detected in MW 28 . Relatively high nitrate concentration s w ere detected in MW 20R . 54 highlights the area owned by the neighbor. Initially high iron concentration in MW 103 was detected but the concentration decreased over time. MW 103 had an average manganese concentration of 1.5 mg/L. Manganese was not detected in MW 23, located downstream of Field 3 ( Figure 52 ). This MW also ha d a relatively high nitrate concentration. Figure 48 . MW location in Field 3 102 Figure 49 . Groundwater quality for MW 28 (upstream of F i eld 3 ) Figure 50 . Groundwater quality for MW 20R (within of Field 3 ) Figure 51 . Gr oundwater quality for MW 103 (within of Field 3 ) 103 Figure 52 . Groundwater quality for MW 23 (downstream of Field 3) Figure 53 . Area, owned by neighbor 104 In Field 4, the groundwater quality of MW 1 1 (up stream of Field 4) and MW 16 ( downstream of Field 4) are shown in Figure 55 and Figure 56 , respectively. Monitoring well locations for Field 4 are shown in Figure 54 . Relati vely high nitrate concentrations were detected in MW 11. In contrast , nitrate concentration in MW 16 were generally equal to or less than 10 mg/L - N of nitrate. Once again, the concentration of nitrate decreased across the site. The average m anganese concen tration was 0.058 mg/L i n MW 16, which is slightly higher than USEPA secondary drinking water standard. Figure 54 . MW locations for Field 4 105 Figure 55 . Groundwater quality f or MW 11 (upstream of Fi eld 4) Figure 56 . Groundwater quality f or MW 16 (downstream of Field 4) 106 Nitrate monitoring has proven to be difficult to understand because of the complex site hydrogeology . I nteresting to note that, i n 2017, the average nitr ate concentration in upstream wells was 10.4 mg/L, while downstream wells was 8.8 mg/L. Significant nitrate from area farms is clearly impacting the groundwater coming onto the demonstration site. Further study on understanding of nitrogen process at the d emonstration site is needed while considering plant uptake . Pretreatment for nitrogen removal should also be consider ed . A p reliminary study indicates that a substa ntial amount of nitrogen and carbon can be removed by pretreating the wastewater using coagu lation / flocculation (data is not shown). In summary, monitoring hydraulic and organic loading s, groundwater quality , and soil conditions using sensors has proven to be a safe and effective method to minimize environmental impact s from the land application of food processing wastewater. Further, this data has also proved to be important for making operational decisions. 4.3.1.2. Site evaluation Although monitoring can provide a general overview of site conditions, additional evaluation s w ere performed to delineat e non - optimal areas based on the presence of standing water and poor crop growth. To understand the cause s , soil analyses, uniformity, and localized high water condition were monitored at several of these locations. 107 4.3.1.2.3. Visual observat ion Non - optimal areas were delineated for Fields 1, 2 , and 3 and are shown in Figure 57 , Figure 58 , and Figure 59 , respectively , as delineated in red Figure 60 show a photog raph of a non - optimal area. S ome of the non - optimal areas were possibly due to low elevations . Figure 57 . Delineated ( r ed) areas on Field 1 Figure 58 . Delineated ( r ed) areas on Field 2 108 Figure 59 . Delin eated ( r ed) areas on Field 3 Figure 60 . Non - optimal areas 109 4.3.1.2.2. Soil texture Soil sampling was conducted at the surface, 30. 48 , 61. 96 , and 91. 44 cm (1, 2, and 3 f t ) depths. Soil sample could not be collecte d in Field 2 at 91.44 cm ( 3 ft ) depth because of the high level of compact ion . R esult s show that soil textures differ in each field but there was no significant difference between optimal and non - optimal areas. Consequently, soil texture does not explain t he cause of non - optimal areas. Table 20 shows the texture for non - optimal and optimal areas. Table 20 . Texture for non - optimal and optimal areas Depth Field 1 Field 2 Field 3 Non - optimal Optimal Non - optimal Optimal Non - optimal O ptimal Top Loamy Sand Loamy Sand Sandy Loam Sandy Loam Loamy Sand Loamy Sand 30. 48 cm ( 1 ft) Loamy Sand Loamy Sand Sandy Loam Sandy Loam Loamy Sand Loamy Sand 61. 96 cm ( 2 ft) Sand Sand *Sandy Loam **Sandy Loam Loamy Sand Loamy Sand 91. 44 cm ( 3 ft ) Sand Sand N/A N/A *** Sandy Loam Sand * 50.8 cm ( 20 in ) deep ** 71.12 cm ( 28 in ) deep *** 76.2 cm ( 30 in ) deep N/A: Not available 110 4.3.1.2.3. Soil compaction Soil compaction was measured on both optimal and non - optimal areas up to a 61.96 cm (2 ft) depth. Figures 6 1 , Fi gure 6 2 , and Figure 6 3 show the locations of these measurements and Tables 2 1 , 2 2 , and 2 3 show soil compact result for Field 1 , 2 , and 3 , respectively. A, B, and C designations in each area represent location that were randomly chosen. Overall, there was n o clear difference between optimal areas and non - optimal in regard to soil compaction. 111 Figure 61 . Locations o f soil compaction analysis in Field 1 Table 21 . Soil compaction result using soil compactio n meter in Field 1 ( Green = 0 - 1,378 kpa ( 0 - 200 psi ) , Yellow = 1,378 - 2,068 kpa ( 200 - 300 psi ) , Red = > 2,068 kpa ( 300 psi ) ) 112 Figure 62 . Locations o f soil compaction analysis in Field 2 Table 22 . Soil com pact result using soil compaction meter in Field 2 ( Green = 0 - 1,378 kpa ( 0 - 200 psi ) , Yellow = 1,378 - 2,068 kpa ( 200 - 300 psi ) , Red = > 2,068 kpa ( 300 psi )) 113 Figure 63 . Locations o f soil compaction analysis in Field 3 Table 23 . Soil compact result using soil compaction meter in Field 3 ( Green = 0 - 1,378 kpa ( 0 - 200 psi ) , Yellow = 1,378 - 2,068 kpa ( 200 - 300 psi ) , Red = > 2,068 kpa ( 300 psi )) 114 4.3.1.2.4. Infiltration Infiltration analysis wa s conducted in locations random ly selected in optimal and non - optimal areas on Field 1 (Table 2 4 ). Statistically, there was no different between optimal area and non - optimal area (p - value 0.6116, =0.05). Table 24 . Infiltration analysis in Field 1 . L ocation Infiltration rate (in/sec) Optimal area Non - optimal area 1 524 664 2 330 450 3 534 720 4 414 456 5 534 450 6 762 328 Average 516.3 511.3 Standard deviation 133 .1 136.0 115 4.2.1.2.5. Uniformity The water level in each of the cup distributed throughout the field to measure uniformity was measured and recorded. F igures 6 4 and 6 5 show the result s for Field s 1 and 3 . The average volume for Field 1 and 3 were 69 mL and 42 mL , r espectively, indicating great variability. T he impacts of uneven uniformity can be significant, such as poor plant health and metal mobilization from small disjointed locations. The s ystem uniformity coefficient is a numeric measure of the overall performa nce of an irrigation system (Kelley 2014) . T he system uniformity for Fields 1 and 3 were 61 and 54 , respectively. A system uniformity coefficient of 85 is considered a uniform distribution. Below a coefficient of 80 re quires the sprinkler heads to be adjusted . Further study to address uneven distribution of irrigation system is needed. If suspended solids clog the nozzle, pretreatment may be required to reduce the suspended solids. 116 Figure 64 . Catch Can Volume for Field 1 Figure 65 . Catch Can Volume for Field 3 117 4.3.1.2.6. L ocalized high water condition Observation of localized high water con dition was conducted for Fields 2 and 3 ( Figure 66 and Figure 6 7 , respectively). Figure 68 shows localized high water condition found in Location 1 i n Field 3 . Table 25 shows the results of the depth of the high localized water condition and the total depth the geoprobe drove in the soil. The soils at Areas 1, 3, and 5 of Field 2 were compacted, especially in Location 1. The geoprobe could only proceed 50.8 cm (20 in) into the soil f or that location. Localized high water conditions in Fie ld 2 were observed depths of 121.92 cm (48 in), 142.24 cm (56 in), and 157.48 cm (62 in) for location 2, 3, and 4 , respectively. Two locations were selected in the Field 3 for testing . In both locations, free water at the bottom of the bore hole was found . This indicates that the localized high water tables at Areas 3 and 4 were both less than 121.92 cm ( 4 ft ) . Because location s 2, 3, and 4 in Field 2 and Areas 1 and 2 in Field 3 are within the non - optimal areas, localized high water condition may be a pot ential cause of non - optimal conditions . 118 Figure 66 . Location s for high localized water condition analysis in Field 2 Figure 67 . Location s for high localized water condition analysis in Field 3 119 Figure 68 . High localized water condition found in location 1 on Field 3 Table 25 . High localized water condition analysis for F ield s 2 and 3 Field Location High localized water condition depth Total depth 2 1 N/ A 50.8 cm ( 2 0 in) 2 < 121.9 cm (48 in) 121.9 cm (48 in) 3 < 142.2 cm (56 in) 142.2 cm (56 in) 4 < 157.4 cm (62 in) 157.4 cm (62 in) 5 N/A 121.9 cm (48 in) 3 1 < 121.9 cm (48 in) 121.9 cm (48 in) 2 < 121.9 cm (48 in) 121.9 cm (48 in) N/A no t available In summary, n on - optimal locations were identified at the demonstration site and analyses were conducted to understand the cause . S ite characteristic for non - optimal conditions were evaluated , including soil t exture, compaction, infiltrat ion, n on - uniformity irrigation, and localized high water condition. Interestingly, soil texture, comp action, and infiltration rate did not appear to be the primary cause. T he soil at the demonstration site are generally compacted , 120 likely from d riving heavy equip ment on the wet soil. Drying the fields is desirable to prevent compaction but r esting the field is a challenge because of the fixed hydraulic loading coming from the food processing plant . Precipitation may also delay the drying process. L ow areas in the field combined with localized high water conditions may be problematic, although more research is required. Uniformity testing shows that the water is unevenly distributed, which is typical for center pivot irrigation systems, especially under the circumst ances at the demonstration site. The most likely cause was excessive moisture content as many of the non - optimal locations were in lower areas within the field and contained localized high water condition s. Runoff i s kept on site so environment risk s are eliminated but the practice of catching and redistributing the water increases maintenance and expenses . An alternative is the use of variable irrigation , to reduce water loading in non - optimal areas . C over crops can tolerate high water contents . S oil amen dments such as gypsum, and biochar may help on absorbing water and nutrients. 4.3.3. HYDRUS CW2D modeling The result of model calibration, validation, and scenario simulation to observe impact of dosing frequency on the performance of treatment are provided. 4.3.3.1. Mo del calibration and validation Model calibration was conducted by inverse modeling using volumetric water content measurement data. Figure 69 shows the comparison of measured and fitted HYDRUS volumetric water content values. The measured volumetric water content data of Domestic/Food WW 121 condition in the first two dosing were used for calibration and the remaining measured data was used for model validation. Figure 69 . Fitted HYDRUS to measured value us ing volumetric water content from Domestic/Food WW Calibrated and validated water flow model was evaluated by E, IA , and RMSE. As shown in Table 26 , all the values were met according to the quality of model criteria, E > 0.5, IA > 0.8, and RMSE <0.014. 122 Table 26 . Goodness of fit result for calibration and validation Model Goodness of fit E IA RMSE Calibration 0.9 0.96 0.002 Validation 0.61 0.88 0.002 Table 27 shows adjus ted HYDRUS CW2D parameters for food processing wastewater land application modeling. Based on literature review, six parameters were considered for calibration process of the solute flow. After using the trial and error method, the maximum aerobic growth r ate of XANs, Maximum denitrification rate of XH, and Fraction of CI in biomass lysis were adjusted. Table 27 . Adjusted HYDRUS CW2D parameters for food processing wastewater land application modeling Parameter Description Unit Stand ard Adjusted k h Hydrolysis rate 1/d 3 - b h Lysis rate for XH 1/d 0.4 - b ANs/b Lysis rate for XANs/b 1/d 0.15 - µ ANs Maximum aerobic growth rate of XANs 1/d 0.9 0.45 µ dn Maximum denitrification rate of XH 1/d 4.8 4 f BM,CI Fraction of CI in biomass lys is 1/d 0.02 0.01 Table 28 shows the simulated effluent concentrations before and after the fitting process. The table contains averages and standard deviations of measured influent, effluent, and simulated values using standard and adjusted parameters from the calibration process. The measured values from day 108 to 170 (62 days) in Domestic/Food WW were used. 123 Table 28 . Simulated effluent concentrations before and after the fitting process for calibration Condition Inf/Eff Type of value HYDRUS CW2D Parameters COD (mg/L) Ammonia (mg/L) Nitrate (mg/L) Domestic/Food WW Influent Measured Avg . (Std . ) 619.0 (142.3) 28.9 (3.2) 1.0 (0.6) Effluent Measured Avg. (Std . ) 48.1 (7.2) 0.5 (0.2 ) 29.4 (4.6) Simulate d Standard 59 0.6 24.6 Simulated Adjusted 50.3 0.25 29.6 Table 29 shows the relative differences between measured and simulated values in the calibration process. The COD, ammonia, and nitrate simulated values using standard parameters in Domestic/Food WW condition are different than the measured values. Thus, parameters such as maximum aerobic growth rate of XANs, maximum denitrification rate of XH, and fraction of CI in biomass lysis were adjusted. After adjustments were ma de, the relative difference of COD, ammonia, and nitrate in Domestic/Food WW decreased from - 18.47% to - 4.37 %, from 67 % to 0%, and from 19.51% to - 0.68 %, respectively. Table 29 . Relative difference between measured and simulated v alues for calibration Condition HYDRUS CW2D COD Ammonia Nitrate Domestic/ Food WW Standard - 18.47% 67% 19.51% Adjusted - 4.37% 0% - 0.68% Table 30 the s imulated effluent concentrations before and after the fitti ng process for validation. The table contains average and standard deviations of measured influent, effluent, and 124 simulated values using standard and adjusted parameters in the validation process. The measured values from day 171 225 (54 days) in Domesti c/Food WW wastewater were used. Table 30 . Simulated effluent concentrations before and after the fitting process for validation Condition Inf/Eff Type of value COD (mg/L) Ammonia (mg/L) Nitrate (mg/L) Domestic/Food WW Influent Meas ured Avg (Std) 746.5 (111.1) 28.0 (4.5) 1.1 (0.9) Effluent Measured Avg (Std) 49.3 (13.1) 0.1 (0.1) 30.9 (3.3) Simulated 52.1 0.27 27.9 Table 31 show the relative differences between measured and simulated values in the va lidation process. Except ammonia, the largest relative difference between me asured and simulated was 10.75 %. Although the large relative difference is observed for ammonia, the difference between 0.1 and 0.27 is not significant with respect to field condit ions. Table 31 . Relative difference between measured and simulated values for validation Condition HYDRUS CW2D COD Ammonia Nitrate Domestic/Food WW Adjusted - 5.37% - 63 .0 % 10.75% 4.3.3.2. Scenario T reatment performance enhancement Potent ial nitrate leaching into groundwater is a concern for food processing wastewater land application systems. Monitoring results from the demonstrate site showed the difficulty in determining its fate for a complex, large site. HYDRUS CW2D was conducted to observe the fate of nitrate for current operating conditions and the potential impact of dosing frequency. Multiple strength of hydraulic and loadings along with dosing frequency were observed. A nitrate effluent concentration on the 150 th day at 60.96 cm ( 2 ft) depth of soil was observed. As 125 F igure 7 0 shows, the COD effluent concentration did not significantly change after 120 days as indicated by the very low slope of - 0.001. In F igure 7 1 , the growth of heterotrophic bacteria also did not change after 120 days. Figure 70 . Simulated COD effluent concentrations with multiple strength of loadings 126 Figure 71 . Steady state condition of heterotrophic microorganism (XH) growth Figure 72 shows the relative differences from the control which is 1 dosing frequency with 619 mg/L of COD. Table 32 shows the r elative difference from control for COD simulation (1 dosing frequency with 619 mg/L of COD). Although the s trength of hydraulic and organic loadings increase d , the effluent COD concentration at 60.96 cm (2 ft) depth of sandy loam did not change. Increasing dosing frequency was able to reduce COD effluent concentration by a maximum of 5.4 mg/L of COD. A possible explanation for a lower COD effluent concentration with higher dosing frequency is an increase in the retention time. High organic loading rate can cause the bioclogging, caused by excessive growth of microorganisms. This is also known as a biofilm. The g rowth of a biofilm will change of the hydraulic properties but this is not considered in the HYDRUS CW2 D model and future study is needed. If the soil surface is clogged by biofilm, soil ponding or surface runoff will occur. Surface runoff can result in so il erosion or contamination of local surface water . 127 Figure 72 . Impact of dosing frequency , and hydraulic and organic loadings on COD effluent concentration on the 150 th day at 60.96 cm ( 2 ft ) depth of sandy loam soil Table 32 . R elative difference from control for COD simulation ( 1 dosing frequency with 619 mg/L of COD) Hydraulic Loading (in/day) Relative difference from control (mg/L) 1 Dosing 3 Dosing 6 Dosing 10 Dosing 0.42 0 - 1.7 - 3.2 - 5.4 0.84 - 0. 3 - 1.9 - 3.3 - 5.5 1.26 - 0.5 - 2 - 3.7 - 5.3 Figure 73 shows the effect of the dosing frequency and the strength of hydraulic and organic loadings on nitrate effluent concentration. Table 33 shows the relat ive difference from the control ( 1 dosing frequency with 102 mg/L of COD). As doing frequency and hydraulic and organic loadings increases, the relative difference also increases proportionally. The model simulation shows the potential nitrate removal by 128 increasing doing frequency from 36.5 mg/L - NO 3 - N to 23.9 mg/L - NO 3 - N. When both dosing frequency and hydraulic and organic loading increased, the maximum reduction was from 36.5 mg/L - NO 3 - N to 15.7 mg/L - NO 3 - N. Higher dosing frequency may increase the moist ure content of the soil, leading to a reduc tion in the oxygen content, providing more optimal conditions for denitrification. The Michigan Department Environmental Quality accepted a value of 5.6 g - BOD/m 2 /day (50 lb - BOD/acre/day) as a monthly average with monitoring groundwater and soil condition s . Figure 74 show that the nitrate removal efficiency for 6 and 10 dosing frequency with 6.0 g - BOD/m 2 /day (54 lb - BOD/acre/day) are 33% and 35%, respectively. Thus, 6 or 10 dosing frequency are recommended to achieve at least a 33% of nitrate removal. A c enter pivot irrigation system is typically used for wastewater application. More frequent application requires more energy, which le a d s to high er operation cost s . Since the removal efficiency for 6 dosing frequency an d 10 dosing frequency are not significantly different, 6 dosing frequency is recommended for Michigan. This model did not consider factors such as precipitation, weather conditions, plant uptake, gro wth of biofilm, and topography. If considered, the differ en ces m ay be less or more significant . Consequently, further study is required . Plant uptake can be significant in removing nutrients. W h en the crop is actively growing, nutrients in food processing wastewater are used by crops. Corn and alfalfa are grown at the food processing wastewater land application demonstration site. The yields were a combined average of 20 tons/acres and 6.7 tons/acre for corn and alfalfa, respectively, without adding additional commercial fertilizer. The yields were higher than th e average corn yield in Michigan. While the crop is growing, increasing dosing frequency for nitrate removal is not needed. 129 When the temperature decreases, the activity also decreases, which slow s the denitrification process. During the w inter, pretreatment technology and winter cover crop are recommended to minimize nitrate leaching into groundwater. In Michigan, the permit for wastewater land application systems must be renewed every 5 years. This model can be beneficial to regulator and operator/manager at the food processing wastewater land application facility. Once the model is calibrated and validated with different types of soil, weather condition, and hydraulic and organic loadings, multiple scenarios for different operation approa ch es can be conducted and provide the recommendations as an index. The index allows to determine the best operation strategies for sit e and waste - specific condition. Figure 73 . Effect of dosing frequency and the strength of hy draulic and organic loadings on nitrate effluent concentration 130 Figure 74 . Simulation of the impact of volumetric water content in soil by different dosing frequencies at 150 th day; 3 dosing (left), 6 dosing (middle), and 10 dosi ng (right) Table 33 . R elative difference from control for nitrate simulation (1 dosing frequency with 619 mg/L of COD) Hydraulic Loading (in/day) Relative difference from control (mg/L) 1 Dos es 3 Dos es 6 Dos es 10 Dos es 0.42 0 - 6.9 - 12.1 - 12.6 0.84 - 0.7 - 8.2 - 12.7 - 18.6 1.26 - 1 - 8.3 - 12.9 - 20.8 131 4.4. Conclusion In conclusion, the multiple monitoring practices assured the manufacturer an d regulatory agency that environmental risk was minimized. Results also indicate that nitrate needs to be further investigated as the complexity of the site and likelihood of a substantial amount in the groundwater is actually entering the site makes it di fficult to understand its fate. Regardless, th is research illustrated a methodology for determining if the land application of high strength food processing waste is applicable for examining other unique sites and demonstrated benefits. Included are cost s avings ; the reduced use of freshwater ; nitrogen and phosphorus reuse for beneficial crop production; and energy conservation and the resulting GHG reduction associated with the energy saving. Monitoring strategies including tracking hydraulic and organic loadings, measuring real - time soil condition by soil sensor s , and testing groundwater quality showed the viability of using land application to treat food processing wastewater. This data, and in particular the sensor values, also helps make operation deci sion in regard to the best field to irrigate to at any one time . Although the benefits and effectiveness, and means to evaluate for site - specific conditions, for the land application of food processing wastewater has been shown to be viable, the design bas is is still very empirical . T here is not a consensus in the literature on the most strategic operational strategies. Therefore, a finite element modeling approach was explored to determine its potential utility in predicting performance and aiding in desig n. HYDRUS CW2D model was successfully calibrated and validated using measured volumetric water content from the laboratory experiment previously described. The HYDRUS CW2D simulation shows the potential for enhanced nitrate removal by increasing dosing fre quency. This results because 132 levels, which can stimulate denitrification. However, this strategy may result in metal mobilization into groundwater. The optim al dosing frequency needs to be studied to mitigate nitrate leaching and metal mobilization into groundwater. Further, causing intermittent high and low soil moisture contents by manipulating dosing frequencies - one of the few practical controls possible in a land application system may enable both denitrification to occur and prevent metal mobilization. However, the limitations of HYDRUS CW2D were discussed in Chapter 4.2.3.3, including the failure to represent clogging. Clogging can result from suspe nded solids and excessive growth of biofilm resulting from the application of high strength food processing wastewater. Once the soil is clogged, standing water and higher moisture contents can result, which has a negative effect on crop growth. To prevent clogging, hydraulic and organic loadings may need to be reduced as well as the collection of runoff and its re - application onto the secondary irrigation land. Currently, HYDRUS CW2D does not consider causes of clogging. However, both particle transport a nd bacteria growth models are available in the model (Kildsgaard and Engesgaard 2001; Mackie and Bai 1993) . This data can be potential used to (R adcliffe and Simunek 2010) . Understanding the thickness, hydraulic conductivity, and development rate of the biofilm under different conditio ns , is needed to achieve this. Recommen dation for non - optimal areas are discussed below. Novel management approach es may provide optimal conditions for crop growth and maximize soil treatment capacity. 133 When comparing cover crops, the first step is to identify unique site - specific attributes. Good options for cover crops include, but are not limited to, triticale, ann ual ryegrass, and cereal rye. Further details are provided on each. Triticale is thought to be the best option for a wastewater land application site. Triticale is a cross between wheat and rye with the quality and yield of wheat and the disease and winter hardiness of rye. Its fibrous root system prevents erosion and builds the soil structure, reducing compaction. The dense stand helps dry out wet spring soil enabling the planting of the primary crop earlier in the season. Triticale can be used as feed for various classes of livestock including lactating dairy cows. Another option is variable irrigation. Variable rate irrigation uses solenoid valves through the length of the pivot arm to turn off and on predetermined segments as it progresses around the fie by localized field conditions such as the soil characteristics, topography, and localized high water conditions. Although this approach may work well in warm sea sons, it may be difficult to implement in winter as the valves may freeze and not work properly. Field and soil modifications can also be considered. If the non - optimal section is caused by the topography of the field, such as a low location, fill can b e added. However, this may be very expensive, depending on the extent of the depression. Soil amendments that sorb water and increase porosity are also possible. Such materials include gypsum, lime, and biochar. Gypsum is composed of calcium sulfate dehyd rate and is used for improving soil he alth, specifically saline soils (Grubb et al. 2012) . Lime , composed of calcium carbonate , increases the pH of acidic so il, improves soil infiltration rate s , decreases soil compaction, increases rooting depth, and adsorbs phosphorus (Berglund 1996) . B iochar can optimizes soil conditions to increase crop production by reducing the soil bulk density and increasing its nutrient absorption , pH, water holding 134 capacity, infiltration rate, and soil microbial activity (Anderson et al. 2011; Graber et al. 2010; H ussain et al. 2016; Liu et al. 2015; Schne ll et al. 2012; Ventura et al. 2012) . However, this option may also be expensive so the exten t of required treatment must be carefully considered. 1 35 Chapter 5 . Conclusion and recommendations In this section, results are first summarized. Thereafter, insights and recommendations for further research are provided. 5.1. Summary The objectives of this s tudy were to demonstrate the effective ness of domestic wastewater land application system s by examination of the literature, evaluate the effectiveness of food processing wastewater land application, compare the benefits of wastewater land application to c onventiona l wastewater treatment systems, develop a simulation approach for the complex wastewater land application treatment system using HYDRUS CW2D, and analyze multiple scenarios using the calibrated model to correlate operational parameters. A summary of the results follow. Land application treatment system s for domestic wastewater are effective and can achieve an average removal efficiency of more than 60% for COD, BOD 5 , TP, and ammonia. Denitrification to convert nitrate to nitrogen gas does not occu r in domestic wastewater land application treatment systems, even with modified dosing frequencies and high application rates. Food processing wastewater land application was found to be effective and efficient at the demonstration site as found by m onitor ing hydraulic and organic loading s , real - time soil condition s by soil sensor clusters, and groundwater sample analyses. The method ology 136 developed and used at the demonstration site is applicable to other sites and can aid in making important operational de cisions . Monitoring results are summarized below. o S oil sensor clusters located under the center - pivot irrigation system over a long period of operation proved to be reliable and economical. o Volumetric water content sensors demonstrated the same trends as oxygen sensors in regard to indicating if soil conditions were aerobic or anaerobic. The water content sensors require less maintenance and are more economical than oxygen sensor s . o Corn and alfalfa are grown at the food processing wastewater land applicat ion site. The yields were a combined average of 20 tons/acres and 6.7 tons/acre for corn and alfalfa, respectively, without adding additional commercial fertilizer. These yield were higher than the average corn yield in Michigan. o Non - optimal areas at the demonstration site were delineated based on poor crop growth and standi ng surface water. Soil texture and soil compaction were not found to be the primary causes. Instead, low areas in the field combined with localized high water conditions appear to be th e problem, although more research is needed. Wastewater land application system s , when compare d to conventional wastewater treatment system s , provides benefits such as reducing usage of freshwater and energy saving that is required to treat the wastewater, consequently reducing GHG emission. HYDRUS CW2D was successfully calibrated and validated using measured volumetric water content da ta from laboratory experiments. Multiple scenarios analysis using the calibrated and validated model was conducted. Model s imulation results are summarized below. 137 o Most of the COD removal in a domestic wastewater land application system occurs within a 15.24 cm (1 ft ) depth for a sandy loam soil. o Increasing the dosing frequency was effective in slightly reducing the COD effluen t concentration. A possible explanation is the increase in retention time. o At a typical i nfluent COD concentration of domestic wastewater, nitrate - nitrogen removal could not be achieved by increasing the dosing frequency . Consequently, the hypothesis tha t increasing dosing frequency would increase the soil moisture content and/or increasing the amount of carbon that proceeding to a depth where the soil environment was anaerobic was not proven . o At a high COD, such as in food processing wastewater, n itrate removal increased when both dosing frequency and hydraulic and organic loadings increased. The cause was hypothesized as above. 138 5.2. Recommendation The followings are recommendations from this study. Design criteria for domestic wastewater land applic ation system provides several options for soil types, including coarse sand, medium sand, sandy loam, and loamy sand. The soil types of the food processing wastewater land application demonstration site var y , and includes loamy sand, sandy loam, and sand, depending on location and depth. Calibrating the model using different soil types is suggested. M odel parameters such as hydraulic loading , m aximum aerobic growth rate of XANs , m aximum denitrification rate of XH , and f raction of CI in biomass lysis w ere ca librated using laboratory experiment al data. Calibrating the model using field data is recommended. Tile drainage and soil lysimeters are option s to capture waste water and nutrients transport ed through the soil column . Volumetric water content s oil sensor is also recommended for calibrating water flow. Model simulation result shows that more frequent dosing and carbon are needed to promote denitrification for domestic wastewater land treatment systems. These conditions may stimulate the growth of a biofilm that may restrict flow, resulting in the premature life of drain field. The optimal dosing frequency and influent COD concentration to minimize biofilm growth should be studied. Metal mobilization is a concern at the demonstration food processing wastewat er land application site. Modeling the metal mobilization using HYDRUS PHREEQC is recommended. HYDRUS 1D PHREEQC can simulate the transport of multiple components and mixed equilibrium/kinetic biogeochemical reactions, including interactions with minerals, cation exchange reaction, and pH dependent cation exchanges . Previous 139 studies have used HYDRUS to simulate metal mobilization in soil (Anwar & Thien, 2015; Dao et al., 2014; Nakam ura et al., 2004; Wang et al., 2016) . The HYDRUS PHREEQC modeling approach may help to understand the metal mobilization under different site and waste - specific condition. M odeling result s for f ood processing wastewater land application show that increasi ng dosing frequency can stimulate the denitrification process. Higher dosing frequenc ies may increase the moisture content of the soil, leading to a reduc tion in the oxygen content, providing conditions for denitrification . However, reduced oxygen conditio ns in the soil may result in metal mobilization. The optimal dosing frequency needs to be studied to simultaneous ly mitigate nitrate leaching and metal mobilization into groundwater. In addition to domestic wastewater and food processing wastewater, rese arch on other wastes that are commonly land applied such as winery and milking facility wastewater are suggested including monitoring, benefits, and modeling. Impacts of HYDRUS CW2D limitations should be determined. These limitations include the lack of c onsideration of heterogeneous mixture of soil, precipitation, evapotranspiration, topography, soil pH, plant uptake, macropores, and clogging by suspended particle and excessive growth of biofilm. P oplar tree growth in a land treatment system absorbs wast ewater and nutrients. These trees ha ve been used for phytoremediation for pollutants such as nitrate, atrazine, metals, organics, chlorinated solvent, and benzene (USEPA 2000) . Root system of poplar tree can reach depth of 4.6 m (15 ft.). Aryal and Renhold (2015) have found that a poplar tree is effective in reducing the leaching of iron and manganese and can withstand the continuous saturation of soils condi tion, while maintaining high evapotranspiration rates (Aryal and Reinhold 2015) . More 140 pilot studies are suggested on the effectiveness, efficiency, and benefits of employing poplar trees in a wastew ater land application system are recommended . Onsite wastewater treatment system s can be a source of pharmaceutical and personal care products (PPCP) that enters into groundwater. Sulfamethoxazole, carbamazepine, and nicotine were detected underlying septi c field (Godfrey et al. 2007) . Wastewater i n Cape Cod (Barnstable County, Massachusetts) is mainly treated by onsite wastewater technologies and disposed through septic field s. T etrachloroethylene (analgesic), acetaminophen (antibiotic), sulfamethoxazole, caffeine, carbamazepine, dehydronifedipine, diphenhydramine, and p - Xanthine were found in monitoring wells (Zimmerman 2005) . T able 43 shows PPCPs found in before and after treatment using drain field s . More research on PPCP transport in soils is needed. Stud ies on treatment technologies to capture PPCP are also recommended. Previous studies reported a potential use of biochar to treat PPCPs. Yao et al. (2012) study reported that 2 - 14% of sulfamethoxazole was transported through biochar amended soil, whereas 60% of sulfamethoxazole leached through un amended soil (Yao et al. 2012) . Chen et al. (2017) shows cabrbamazepine was effetely removed by biochar (Chen et al. 2017) . Stud ies on the type of biochar and optimal mixture ratio s are needed. A column study is recommended . 141 Table 34 . Literature review on PPCP found in a domestic wastewater land application system Reference PPCP Before drain field (µg/L) After drain field (µg/L) Swarz et al. (2006) Caffeine 17 23 < 1.7 Paraxanthine 55 65 < 1.7 Conn and Siegrist (2009) EDTA (ethylenediaminetetraacetic acid) 2.4 NP1EC (4 - nonylphenolmonoethoxycarboxylate) 7.2 NP (4 - nonylphenol) 4.1 Sulfamethoxazole 0.51 Matamoros et al. (2009) Salicylic acid 16.4 0.66 Ibuprofen 1.95 0.02 OH - ibuprofen 3.45 0.28 CA - ibuprofen 2.45 0.04 Carbamazepine 4.5 Naproxen 0.09 Diclofenac 0.5 Ketoprofen 1.79 Caffeine 31.9 0.16 Methyl - dihydrojamonate 8 0.04 Hydrocinnamic acid 21.1 0.02 Oxybenzone 3.35 Furosemide 4.65 142 APPENDICES 143 APPENDIX A: HYDRUS CW2D parameters Figure 75 . Domain t ype and u nits 144 Figure 76 . Rectangular d omain d efinition Figure 77 . Main processes and add - on m odules 145 Figure 78 . Time information Figure 79 . Output information 146 Figure 80 . Water flow parameters Figure 81 . Solute t ransport 147 Figure 82 . Solute t ransport parameters 148 Figure 83 . Solute t ransport c onstructed w etland m odel p arameter I (Default) 149 Figure 84 . Solute t ransport - c onstructed w etland m odel (CW2D) p arameters II (Default) 150 Figure 85 . Solute t ransport c onstructed w etland m odel p arameter I (Adjusted) 151 Figure 86 . Solute t ransport c onstructed w etland m odel p arameter II (Adjusted) 152 Figure 87 . Data for inverse solution Figure 88 . Rectangular d omain d iscretization 153 Figure 89 . Water boundary condition 154 Figure 90 . Graphic output from HYDR US CW2D 155 APPENDIX B: R Code (Goodness of fit) The following R code addresses opening file and calculate modeling efficiency, index of agreement, and root mean squared error. ##R code for goodness of fit - water flow #Setting the working directory set wd("C:/Users/Dong/Desktop/validation_R/Raw data/Solute calibration data") #Reading files for calibration obs < - read.csv("Calibration_obs.csv", header = TRUE) sim < - read.csv("Calibration_sim.csv", header = TRUE) #Reading files for validation obs < - read .csv("Validation_obs.csv", header = TRUE) sim < - read.csv("Validation_sim.csv", header = TRUE) #Setting each symbols t_obs < - obs[,1] y_obs < - obs[,2] t_sim < - sim[,1] y_sim < - sim[,2] #Step function to organize y for x time scale 156 ef_sim < - stepfun(x = t_sim, y = c(y_sim[1], y_sim), f = 0.5) # Combine all datas data < - cbind(t = t_obs, obs = y_obs, sim = ef_sim(t_obs)) head(data) #Model efficiency E < - 1 - sum((obs - sim)^2)/sum((obs - mean(obs))^2) #Index of agreement IA < - 1 - sum((obs - sim)^2)/ sum(((abs(sim - mean(obs))) + abs(obs - mean(obs)))^2) # R oot mean squared error (RMSE) RMSE < - sqrt(sum((obs - sim)^2)/length(time)) 157 REFERENCES 158 REFERENCES Adriano, D., Novak, L., Erickson, A., - term land disposal by spray irrigation of food - processing wastes on some chemical properties of soil Journal of Environment Quality , 4. Anderson, C., Condron, L., Clough, T., Fiers, M., Stewart, A., Hill, R., and Sherlock. R. (2011). Pedobiologia , 54, 309 320. of Water and Air Distribution in Growing HYDRUS Software Applications to Subsurface Flow and Contaminant Transport Problems , 33. Using HYDRUS - Inter national Journal of Environmental Science and Development , 6(10), 741 745. - Apogee Instruments, Inc. , (Apr. 1, 2018). A ONSITE WASTEWATER SYSTEMS, DESIGNATED REPRESENTATIVES AND State of Arkansas . m Water resources research , 47(4). WASTEWATER USING POPLAR PLANTATION: EVALUATION OF METAL AND Michgian State University. treatment of wastewaters: small - Ecological Engineering , 78, 53 61. Askar, M., Youssef, M., Chescheir, G. , Hesterberg, D., Amoozegar, A., and Skaggs, R. (2016). - pore Flow and Transport component for DRAINMOD, a Step ASABE , 1 6. 159 Balmer, P., and Mattsson, B. (1 Water Science and Technology , 30(4), 7 15. Beach, D. N., McCray, J. E., Lowe, K. S., and Siegrist, R. L. (200 hydraulic conductivity of sand porous media biofilters during wastewater infiltration due to Journal of Hydrology , 311(1 4), 230 243. ice For Land Kennedy/Jenks Consultants. - continuous anaerobic digestion of a Bioresource Technol ogy , 71(3), 261 266. Soil Use and Management , 12, 176 180. Michigan Water Environment Association . Universität Stuttgart . Brown, K. W., Slowey, J. F., and Wolf, H. W. (1997). Accumulation and Passage of Pollutants in Domestic Septic Tank Dispo sal Fields . College Station, TX. - . wastewater managem Ext. Spec. Rep. , AM - 18E. Applied Environmental Microbiology , 45, 1247 1253. Charmley, E., Nelson, D., industry: utilization of potato by - Canadian journal of soil science , 86(4), 621 629. of 160 Science of The Total Environment , Elsevier B.V., 579, 598 605. fish proces sing wastewater Bioresource Technology , 101(2), 439 449. University of Oradea - Faculty of Environmental Protection , 15, 709 714. Coats, E. R., Watkins, - Cycle Analysis for Removing Phosphorus from Wastewater: Biological versus Water Environment Research , 83(8). Colorado Department of public health and envrionmen - ON - State of Colorado . Onsite Wastewater Treatment Systems and Implications for Wate (210), 132. Groundwater , 41(5), 647 656. Treated Domestic On - Site Wastewater Treatment: Proceedings of the Eighth National Symposium On Individual and Small Community Sewage Systems , ASABE, St. Joseph, MI. Crevoisier, D., Popova, Z., Mailhol, J. C., Agricultural Water Management , 95(4), 354 366. Crites, R., and Tchobanoglous, G. (1998). Small and Decentralized Wastewater Managemetn Systems . McGra w - Hill, NewYork, NY. Crites, R. W., Reed, S. C., and Bastian, R. K. (2000). Land Treatment Systems for Municipal and Industrial Wastes . McGraw - Hill, New York, NY. East Bay Times , Walnut Creek, CA, 1. using vertical - Journal of Environmental Sciences , 15(1). 161 Central Valley Regional Water Quality Control Board. CVRWQCB. (2006). Staff Report: Update on the regulation of food processing waste discharges to land . Dalahmeh, S., Pell, M., Vinnerås, B., H ylander, L., Öborn, I., and Jönsson, H. (2012). Water Air Soil Pollut , 223, 3657 3671. Van Dam, J. C., Groenendijk, P., Hendriks, R. F. A., and Kroes, J. Vadose Zone Journal , 7, 640 653. Van Dam, J., Huygen, J., Wesseling, J., Feddes, R., Kabat, P., van Walsum, P., Groenendijk, P., and van Diepen, C. (1997). Theory of SWAP version 2.0 . - 1D Model to Simulate the Transport of some Selected Heavy Metals in Paddy Soil in Thanh Trì Earth and Environmental Sciences , 30(1), 22 30. Deng, S Applied microbiology and biotechnology , 60(5), 588 593. Sewage and Industrial Wastes , 591 595. - Department of Environmental Quality , (Apr. 1, 2018). DEQ. (2015). Nitrate in drinking water . Water Science and Technology , 51(9), 225 232. Dong, Y., S Proceedings of the Water Environment Federation , 3859 3866. Enzyme pretreatment of fats, oil and grease from restaurant waste to prolong septic soil treatment 162 Journal of Environmental Science and Health, Part A , 52(1), 55 63. on of an anaerobic - aerobic Sixth national symposium on food processing wastes . National Onsite Wastewater Recycle Association , Laurel, MD. First national symposium on food processing wastes , Portland, Oregon. Fische modeling of long - term phosphoru ASABE , Montreal, Quebec. Furman, T - multiple Sewage and Industrial Wastes , 27(3), 261 276. processing and do International Journal of Hydrogen Energy , 30(15), 1535 1542. - site sewage Ground Water , 45(3), 263 271. Graber, E. R., Harel, Y. M., Kolton, M., Cytryn, E., Silber, A., David, D. R., Tsechansky, L., Plant and Soil , Spring er , Dordrecht . Sewage Works Journal , 21(6), 1002 1015. Gross, M. (2004). University Curriculum Development for Decentralized Wastewater 163 Management Wastewater Ch aracterization . National Decentralized Water Resources Capacity Development Project , Fayetteville, AR. Phosphorus - Saturated Gypsum on Soil Phosphorus in a Laboratory In Applied and Environmental Soil Science , 2012, 7. Sand Columns Water Science & Technolo gy1 , 28(10), 109 116. Calibration of Watershed Models , Q. Duan, ed., American Geophysical Union, Washington, DC. Haggblom, M. M., and Milligan, P. W. (2000). Anaerobic Biodegradation of Halogenated Pesticides: Influence of Alternative Electron Acceptors . Soil Biochemistry , (J. Bollag and G. Stotzky, eds.), Marcel Dekker Inc., New York, NY. te Sewage Disposal of High Strength . Harper, J., Blaisdell, J. L., and Grosshopf, J. (1972). Dairy food plant wastes and waste treatment practices . United States Environmental Protection Agency, Washington DC. BEHAVIOR WHERE HIGHLY TR EATED EFFLUENT IS APPLIED TO SOIL AT Soil Science , 170(9), 692 706. - zone transport of nitrogen Journal of Contaminant Hydrology , 91, 184 201. Science of the Total Environment , 380, 196 209. V in unsewered areas underlain by coarse sand and gravel substrates Documentation of rsity of 164 Montana. , Amsterdam;Boston; . University of Illinois at Urbanan - Champaign. Florida. Hussain, M., Farooq, M., Nawaz, A., Al - Sadi, A. M., S olaiman, Z. M., Alghamdi, S. S., Ammara, U., Ok, Y. S., and Siddique, K. H. M. (2016). Biochar for crop production: potential benefits and risks . Journal of Soils and Sediments , Journal of Soils and Sediments. Hutson, J. (2000). LEACHM: Model Description a Physics, and Earth Science . Ground Water Modeling Center, Riverside, CA. Agronomy J ournal , 86, 843 851. Journal of Environmental Science and Health - Part A Toxic/Hazardous Substances and Environ mental Engineering , 50(14), 1452 1457. Kartal, B., Rattray, J., van Niftrik, L. A., van de Vossenberg, J., Schmid, M. C., Webb, R. I., Schouten, S., Fuerst, J. A., Damsté, J. S., Jetten, M. S. M., and Strous, M. (2007). Systematic and Applied Microbiology , 30(1), 39 49. Monitored Natural Attenuation an A - Albuquerque. MICHIGAN STATE UNIVERSITY EXTENSION PURDUE EXTENSION . logging in Journal of Contaminant Hydrology , 50, 261 285. 165 Acta Scientiarum Polonoru m Technologia Alimentaria , 1(123 132). Koren, H., and Bisesi, M. (2003). Handbook of Environmental Health, Fourth Edition, Volume II: Pollutant Interactions in Air, Water, and Soil . CRC Press, Boca Raton, FL. Kroes, J. G., van Dam, J. C., Bartholomeus, R. P., Groenendijk, P., Heinen, M., Hendriks, R. F. A., Mulder, H. M., Supit, I., and van Walsum, P. E. V. (2017). SWAP version 4; Theory description and user manual . Wageningen. ubstrate clogging Water Science & Technology , 48(5), 25 34. Soil Science Society of America , 4, 924 938. The multi - component reactive transport module CW2D . Pc - Progress . distribution of microbial biomass in subsurface vertical fl Water Science and Technology , 56(3), 233 240. Water Research , 29(10), 227 2245. Legates, D. R., and McCabe, G. Water resources research1 , 35(1), 233 241. Texas A&M Agrilife Extension , (Apr. 4, 2018). Water Research , 43, 695 705. Liu, Wang, H. increased water holding capacity but accelerated organic carbon leaching from a sloping Environ Sci Pollut Res . 166 e role of particle size distribution on the performance and Water Science & Technology , 27(10), 19 34. - Type Disposal Systems for Septic Tank Effluent : I. Soil - Journal of Environmental Quality , 3(3), 223 228. nitrogen leaching under furrow irrigation: Experimental and modelling ap Agricultural Water Management , 87(3), 275 284. Maryland Deaprtment of the Environmemt. (2010). Maryland department of the environmental guidelines for land application/reuse of treated municipal wasteaters . Matocha, C. J., Scheckel, K. G., and S Chemical Processes in Soils , M. A. Tabatabai, D. L. Sparks, L. Al - Amoodi, and W. A. Dick, eds., Soil Science Society of America, Inc., Madison, Wisconsin. Maupin, M. A., Kenny, J. F., Hutson, S. S., Lovelace, J. K., Barber, N. L., and Linsey, K. S. (2014). Estimated use of water in the United States in 2010 . McDaniel, P. A. (2006). Anaerobic Processes. In: Encyclopedia of Soil Science . (L. Rattan, ed.), Taylor & Francis, New Yor k. - water quality impacts from on - Onsite Wastewater Recycling Association - 13th Annual Conference, Albuquerque, NM, 6 18. or On - site Michigan Department of Public Health. (1994). Michigan Criteria for Subsurface Sewage Disposal . Michigan. Green Project Alte . Moriasi, D. N., Arnold, J. G., Van Liew, M. W., Bingner, R. L., Harmel, R. D., and Veith, T. L. tion guidelines for systematic quantification of accuracy in watershed 167 Transactions of the ASABE , 50(3), 885 900. aldicarb transport in a drained Soc. Agric. , 37, 1187 1824. Vadose Zone Journal , 3(4), 1353. Naseri, Journal of Applied Science , 11(5), 815 823. Journal of Hydrology , 10(3), 282 290. (Aug. 1, 2016). Nebraska Department of Environmental Quality. (2007). TITLE 124 RULES A ND REGULATIONS FOR THE DESIGN, OPERATION AND MAINTENANCE OF ON - SITE WASTEWATER TREATMENT SYSTEMS . STANDARDS - State of New York . Nohra, J., Madramootoo, C., and reconstructed soil columns using HYDRUS - Environmental Earth Science , 65(6), 1751 1758. NYSERDA. (2007). Energy index development for benchmarking water and wastewater utilities . Oliv ieri, A., and Roche, R. (1979). Minimum guidelines for the control of individual wastewater treatment & disposal systems . Oregon Department of Environmental Quality . Palfy, T. G., Molle, P., Langergraber, G., Troesch, S., Gourdon, R., and Meyer, D. (2016). Ecological Engineering , 87, 340 347. Palfy, T., and Langergraber, G. (2014) Ecological Engineering , 68, 168 105 115. Reclaimed Water d uring Long - Term Simulated Soil Aquifer Treatment System under Water , 9(10), 786. numerical model simulating agro hydro - chemical pro cesses in soil under drip fertigated Irrigat Drainage Sys Eng , 5(155), 2. - time monitoring and control of sequencing batch reactors for secondary treatment of a poultry processing wastewa Water Environment Research , 72(5), 585 592. OF DOMESTIC WASTEWATER USING SAND AND ZEOLITE AS FILTER Pucher, B., a J. Hydrol. Hydromech , 66(2), 227 231. using filter media Journal of Hydrology and Hydromechanics , 66(2), 227 231. - Drained Clay Loam Soil Using HYDRUS Vadose Zone Journal , 8(1), 64 74. Radcliffe, D., and Simunek, J. (2010). Soil Physics with Hydrus Modeling and Applicati on . Taylor & Francis Group, LLC, Boca Raton, FL. - dimensional modeling of water and nitrogen fate from sweet sorghum Agricultural Water Management , 11 1, 87 104. Modelling Variably Saturated Flow with HYDRUS - 2D . ND Consult, Brisbane, Australia. 169 Redding, M. (2012). Basic American Foods Land Application Site (Moses Lake) - Groudnwater Review . Olympia, WA. Groundwater , 29(1), 82 92. Ronayne, M. P., Paeth, R. C., and Wilson, S. A. (1982). Oregon On - Site Experimental Syst ems Program . Salem, OR. Bull Univ Minn Med Found , 21, 315 338. Safferman, S. I., Fernandez - Torres, I., Pfiffner, S. M., Larson, R. A., and Mokma, D. L. (2011). Journal of Environmental Engineering - Asce , 137(2), 97 107. Safferman, S., Wright, T., Miller, S., and Higginbot Michigan State University . DRAINMOD - N II model for Agric. Water. Mgmt , 96(2), 267 281. Biological Wastes , 23 (2), 117 142. of Biochar Application to Maintain Energy Crop Productivity: Soil Chemistry, Sorghum Journal of Enviro nment Quality , 41(July 2016), 1044. Shekofteh, H., Afyuni, M., Hajabbasi, M. A., Iversen, B. V., Nezamabadi - Pour, H., Abassi, F., Field using HYDRUS - Communications in Soil Science and Plant Analysis , 44(20), 2917 2931. Water Research , 19(11), 1369 1378. ration in an onsite Eleventh Individual and Small Community Sewage Systems 170 Conference Proceedings , American Society of Agricultural and Biological Engineers, Warwick, Rhode Island. oil Absorptions Systems: The On - Site Wastewater Treatment: Proceedings of the Ninth National Symposium On Individual and Small Community Sewage Systems , ASABE, St.Joseph, MI, 41 51. Transactions of the ASABE , 55(1987), 1261 1274. modeling of contaminant transport using HYDRUS and its Journal of the Indian Institute of Science , 93(2), 265 284. - 2D software package for simulating the two - dimensi onal movement of water, heat, and multiple Solutes in variably - Phosphorus Treatment Capacities of On - Site Wastewater Lateral Journal of Environmental Engineering , 140(2), 04013002. ASABE , 55(4), 1509 1522. nt processes and environmental impacts of Land treatment and disposal of municipal and industrial wastewater , R. L. Sanks and T. Asano, eds., Ann Arbor Science Publishers, Inc., Ann Arbor, MI, 45 83. Srilert, C., Ong, S., sorption properties on Mn 2 + transport under single , binary , and multiple metals through 27. State of Kansas Department of Helath and Enviro State of Kansas Department of Helath and Environment . Tarradellas, J., Bitton, G., and Rossel. D. (1997). Soil Ecotoxicology . CRC Press Lewis Publishers, Boca Rato n, FL. Tchbanoglous, G., Burton, F. L., and Stensel, H. D. (2003). Wastewater Engineering: Treatment 171 and Reuse . McGraw - Hill, New York. Tennessee State Government. (2016). DECENTRALIZED DOMESTIC WASTEWATER TREATMENT SYSTEMS . Toscano, A., Langergraber, G., C in a pilot - scale two - Ecological Engineering , 35, 281 289. iquid and Bioresource Technology , 41(3), 235 245. and Vadose Zone Using the HYDRUS - Vadose Zone Journal , 7(2), 757 768. U.S. Census Bureau. (2006). American Housing Survey for the U nited States: 2005 . Washington DC. ENVIRONMENTAL PROTECTION - WACO , 8, 28 30. USDA Economic Research Service , . USEPA. (1980). Design Manual - Onsite Wastewater Treatment and Disposal Systems . Washington DC. Environmental Protection Agency. Nationa l Risk Management Research Laboratory Office of Research and Development . USEPA , (Jun. 1, 2017). United States Environmental Protection Agency . United States Environmental Protection Agency , . United States Environmental Protection Agency , (Apr. 4, 2018). 172 United States Environmental Protection Agency , (Apr. 17, 2018). United States Environmental Protection Agency , (Apr. 17, 2018). iochar reduces short - J. Environ. Qual. , 42(1), 76 82. Nitrate Dynamics in Zero Tension Plate Lysimeters Using HYDRUS - 8. Working with dynamic crop models , D. Wallach, J. M akowsk, and W. Jones, eds., Elesevier, 11 53. Soil and Water Research , 11(1), 1 10. Wang, X., Youssef, M., Skaggs, R. analyses of the nitrogen simulation model, DRAINMOD - ASAE , 48(6), 2205 2212. processing wastewater by an anox Process Biochemistry , 40(5), 1733 1739. Watson, S., Valos, F., and Waterbury, J. (1981). The Family Nitrobacteraceae . (S. et Al., ed.), Springe r - Verlag, Berlin. Journal of Hydrology , 38(2), 271 288. g hourly evapotranspiration and soil water content at the grass - covered boundary - Hydrological Sciences Journal , 59(2), 376 394. 173 Efficient Manageme nt of Wastewater , (Apr. 1, 2018). Wheeler, D., Williams, B., Merrick, S., Fuller, M., Hannon, P., Seigle, T., Flesher, K., M alley, J., Majewski, C., Manning, J., Northrop, C., and Uncles, R. (2008). Commision to Study Methods and Costs of Sewage, Sludge and Septage Disposal . Concord. Groundwater , 32(6), 905 916. Physical Geography , 2, 184 194. TREATMENT UN ITS WITH HYDRUS - 2D, CONSTRUCTED WETLANDS MODULE NOWRA 18th Annual Technical Education Conference & Exposition , National Onsite Wasteater Recycling Association, Milwaukee, Wisconsin. Yao, Y., Gao, B., Chen, H., Jiang, L., Inyang, M., Zimmerman, A. R., Cao, X., Yang, L., Xue, Journal of Hazardous Materials , Elsevier B.V., 209 210, 408 413. Youssef, M., Skaggs, R., Chescheir, G., and Gillia DRAINMOD - ASAE , 48(2), 611 626. , and Personal Care Products in Selected Water Supplies , Cape Cod , Massachusetts , June