"Applied Science, Faculty of"@en . "Civil Engineering, Department of"@en . "DSpace"@en . "UBCV"@en . "Dasika, Raghava Kumar"@en . "2009-03-20T23:10:19Z"@en . "1996"@en . "Doctor of Philosophy - PhD"@en . "University of British Columbia"@en . "Shallow unconfined aquifers are prone to contamination by non-point source\r\nnitrate as a result of agricultural and other land use practises. Management of such\r\nsusceptible groundwater resources requires a means of quantifying the transport and fate\r\nof the nitrate from its source to the water table and its subsequent distribution through\r\nthe groundwater. A proper understanding of the dynamics associated with nitrate loading\r\nand its effect on unconfined groundwaters is currently lacking. It has been suggested by\r\nseveral investigators that in-situ management of groundwater nitrate may be achieved by\r\npromoting the natural denitrification capacity of microbial populations present within the\r\naquifers. However, the feasibility of such an approach has not yet been fully evaluated.\r\nA study involving the detailed monitoring of the changing groundwater chemistry\r\nprofiles beneath the water table of two unconfined aquifers, combined with a laboratory\r\ncolumn-test investigation of enhanced denitrification during saturated flow through sand,\r\nhas been performed.\r\nThe groundwater monitoring was performed using a passive sampling approach\r\nthat has enabled detailed multi-level profiling of the groundwater chemistry at and below\r\nthe water table. Measurements were made monthly, over a period of twelve months, in\r\norder to determine the temporal variation in groundwater chemistry as related to the\r\nsurficial land use and recharge patterns. This research has included the first known\r\ndetailed measurement of the seasonal arrival of leached non-point source nitrate to the\r\nwater table beneath agricultural lands, and has provided definitive evidence of the pulsed nature of such arrival.\r\nAt the study sites, the monitoring has also shown that infiltrating recharge waters\r\nresult in the rapid delivery of leached nitrate to the water table of the unconfined\r\naquifers. Furthermore, the research findings suggest that, as a result of the development\r\nof vertical hydraulic gradients at the water table during recharge, leached nitrate fronts\r\ncan be transported to a large depth below the water table within a short time period. A\r\nconceptual model has been developed to describe the observed distribution of nitrate\r\nbelow the water table. Using this conceptual model, it is also concluded that nitrate that\r\narrives at the water table of unconfined aquifers with seasonal recharge waters will\r\nsubsequently be transported through the aquifer in a pulsed manner in response to a\r\ncorresponding seasonal fluctuation in the groundwater flow velocities.\r\nThe laboratory column testing has found that ethanol may be preferred over\r\nmethanol as a carbon source by the subsurface microbial population during enhanced\r\ndenitrification. The findings from this testing also suggest that the effectiveness of carbon\r\ninjection systems, as well as stationary reactive barriers aimed at providing denitrifying\r\nenvironments, may be compromised by clogging as a result of N2 gas bubble\r\naccumulations.\r\nThe study results suggest that the role for carbon additions within aquifers for\r\npromoting in-situ denitrification, on an aquifer wide basis, may be limited due to the\r\nseasonally dynamic nature of the nitrate loading and subsequent distribution through\r\nunconfined aquifers, as well as uncertainty associated with the efficacy of heterotrophic\r\ndenitrifying microbial populations within the aquifers."@en . "https://circle.library.ubc.ca/rest/handle/2429/6301?expand=metadata"@en . "13016713 bytes"@en . "application/pdf"@en . "INVESTIGATIONS INTO THE DISTRIBUTION OF NON-POINT SOURCE NITRATE IN TWO UNCONFINED AQUIFERS & THE R O L E FOR C A R B O N ADDITION IN THE CONTROL OF NITRATE CONCENTRATIONS IN GROUNDWATER by R A G H A V A DASIKA B.E. (Civil), Bendigo College of Advanced Education, 1981 M.S.(Civi l - Geotechnical), Rutgers University, 1985 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES (Department of Civi l Engineering) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA February 1996 \u00C2\u00AERaghava Kumar Dasika, 1996 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of 6 / 1 / / / . FNC-lU&e~\u00C2\u00A3/kJ6~ The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT Shallow unconfined aquifers are prone to contamination by non-point source nitrate as a result of agricultural and other land use practises. Management of such susceptible groundwater resources requires a means of quantifying the transport and fate of the nitrate from its source to the water table and its subsequent distribution through the groundwater. A proper understanding of the dynamics associated with nitrate loading and its effect on unconfined groundwaters is currently lacking. It has been suggested by several investigators that in-situ management of groundwater nitrate may be achieved by promoting the natural denitrification capacity of microbial populations present within the aquifers. However, the feasibility of such an approach has not yet been fully evaluated. A study involving the detailed monitoring of the changing groundwater chemistry profiles beneath the water table of two unconfined aquifers, combined with a laboratory column-test investigation of enhanced denitrification during saturated flow through sand, has been performed. The groundwater monitoring was performed using a passive sampling approach that has enabled detailed multi-level profiling of the groundwater chemistry at and below the water table. Measurements were made monthly, over a period of twelve months, in order to determine the temporal variation in groundwater chemistry as related to the surficial land use and recharge patterns. This research has included the first known detailed measurement of the seasonal arrival of leached non-point source nitrate to the water table beneath agricultural lands, and has provided definitive evidence of the pulsed i i nature of such arrival. At the study sites, the monitoring has also shown that infiltrating recharge waters result in the rapid delivery of leached nitrate to the water table of the unconfined aquifers. Furthermore, the research findings suggest that, as a result of the development of vertical hydraulic gradients at the water table during recharge, leached nitrate fronts can be transported to a large depth below the water table within a short time period. A conceptual model has been developed to describe the observed distribution of nitrate below the water table. Using this conceptual model, it is also concluded that nitrate that arrives at the water table of unconfined aquifers with seasonal recharge waters will subsequently be transported through the aquifer in a pulsed manner in response to a corresponding seasonal fluctuation in the groundwater flow velocities. The laboratory column testing has found that ethanol may be preferred over methanol as a carbon source by the subsurface microbial population during enhanced denitrification. The findings from this testing also suggest that the effectiveness of carbon injection systems, as well as stationary reactive barriers aimed at providing denitrifying environments, may be compromised by clogging as a result of N 2 gas bubble accumulations. The study results suggest that the role for carbon additions within aquifers for promoting in-situ denitrification, on an aquifer wide basis, may be limited due to the seasonally dynamic nature of the nitrate loading and subsequent distribution through unconfined aquifers, as well as uncertainty associated with the efficacy of heterotrophic denitrifying microbial populations within the aquifers. i i i TABLE OF CONTENTS Abstract i i Table of Contents iv List of Tables vii i List of Figures x Acknowledgements xvi 1.0 INTRODUCTION 1 1.1 Research Objectives & Scope 5 1.2 Thesis Organization 7 2.0 L ITERATURE REVIEW 8 2.1 Nitrate in the Terrestrial Environment 9 2.2 Observed Distributions of Non-Point Source Nitrate in Unconfined Aquifers 12 2.2.1 Vertical Profiles of Groundwater Chemistry - Detailed Case Studies 14 2.2.2 Summary 35 2.3 Fate and Transport of Non-Point Source Nitrate in Unconfined Aquifers 37 2.3.1 Denitrification of non-point source nitrate 39 2.3.1.1 Controls on Denitrification 39 2.3.1.2 In-situ Indicators of Denitrification 47 2.3.2 Transport of non-point source nitrate 54 2.4 Control of Groundwater Nitrate 71 2.5 Conclusions 74 3.0 R E S E A R C H METHOD OLOGY 76 3.1 Field Investigations 76 3.1.1 Monitoring Site Locations 76 3.1.1.1 Abbotsford Study Site 80 3.1.1.2 Brookswood Study Site 81 3.1.2 Materials and Methods 83 3.1.2.1 Well Installation & Soil Sampling 83 iv 3.1.2.2 Groundwater sampling 86 3.2 Laboratory Investigations 90 3.2.1 Background 90 3.2.2 Materials and Methods 91 3.2.2.1 Set 1 95 3.2.2.2 Set 2 97 3.2.2.3 Set 3 97 3.2.2.4 Set 4 98 3.2.2.5 Set 5 98 3.2.2.6 Set 6 99 3.3 Analytical Techniques 100 3.3.1 Water Sample Analyses 101 3.3.1.1 pH 101 3 .3 .1 .2 - Chloride (CI) 101 3.3.1.3 Total and Inorganic Carbon (TC/TIC) 105 3.3.1.4 Ca, Na, M g , K 105 3.3.1.5 NCV, NCV, N H 4 + , S0 4 2 \" 106 3.3.1.6 Dissolved 0 2 (groundwater) 106 3.3.1.7 Dissolved 0 2 & N 2 (laboratory samples) 106 3.3.2 Soil Sample Analyses 107 3.3.2.1 Carbon Content 107 3.3.2.2 Exchangeable Cations 108 4.0 RESULTS OF ABBOTSFORD AQUIFER INVESTIGATIONS 110 4.1 Subsurface Geology 110 4.1.1 Soil Characteristics 110 4.1.2 Depth to Water Table 113 4.2 Land Use at Site during Monitoring Period 116 4.3 Groundwater Chemistry in Study Well 119 4.3.1 Major Ion Chemistry, pH, IC & TOC 119 4.3.1.1 Study Well 119 4.3.1.2 Comparison to B C M O E Well Data 123 4.3.2 Dissolved Oxygen 125 4.3.3 Ammonium and Nitrite 130 4.4 Soil Chemistry at Study Well 130 4.4.1 Carbon Content of Subsurface Soils 130 v 4.4.2 Cation Exchange Characteristics 132 5.0 RESULTS OF BROOKSWOOD AQUIFER INVESTIGATIONS 134 5.1 Subsurface Geology 134 5.1.1 Soil Characteristics 134 5.1.2 Depth to Water Table 135 5.2 Groundwater Chemistry in Study Well 139 5.2.1 Major Ion Chemistry, pH, IC & TOC 139 5.2.1.1 North Study Well 139 5.2.1.2 South Study Well 144 5.2.1.3 Comparison to available regional data 144 5.2.2 Dissolved Oxygen 147 5.2.3 Ammonium and Nitrite 148 5.3 Soil Chemistry at Study Well 148 5.3.1 Carbon Content of Subsurface Soils 148 5.3.2 Cation Exchange Characteristics 150 6.0 FINDINGS OF L A B O R A T O R Y C O L U M N TESTING 152 6.1 Results of Set 1 Testing 152 6.2 Results of Set 2 Testing 160 6.3 Results of Set 3 Testing 161 6.4 Results of Set 4 Testing 161 6.5 Results of Set 5 Testing 162 6.6 Results of Set 6 Testing 162 6.7 Comparison to Related Published Findings 167 6.7.1 Ethanol versus Methanol as Carbon Sources 167 6.7.2 Nitrogen Gas Generation and Accumulation 168 6.8 Conclusions 171 7.0 DISCUSSION OF RESULTS 177 7.1 Groundwater Chemistry as Related to Surficial Land Use 177 7.1.1 Source of Leached Nitrate and Associated Chemistry 182 7.1.2 Relative Distribution of K + and S0 4 2 \" 193 7.2 Temporal Variation in Groundwater Chemistry 195 vi 7.2.1 Role of Recharge Pattern 195 7.2.2 Transport Beneath the Water Table 198 7.3 Fate of Nitrate Below the Water Table 208 7.3.1 NCy-N/CT Ratio at Study Wells 209 7.3.1.1 Abbotsford Study Well 209 7.3.1.2 Brookswood Study Wells 212 7.3.2 IC vs pH at the Abbotsford Study Well 214 7.4 Role for Carbon Additions in Promoting Denitrification 216 7.5 Efficacy of Passive Sampling towards Quantifying Groundwater Chemistry 219 7.5.1 Membrane Degradation 219 7.5.2 Repeatability & Accuracy 222 7.5.3 Comparison to Pumped Well Sampling data 224 8.0 CONCLUSIONS 227 9.0 RECOMMENDATIONS FOR FURTHER STUDY 235 Bibliography 237 Appendix A : Evaluation of a Travel Time Estimation Procedure for Unconfined Groundwater Flow 256 Appendix B: Dialysis Membrane Evaluation 266 Appendix C: Measurement of Dissolved Nitrogen & Oxygen by Direct Injection Using Gas Chromatography 277 Appendix D: Detailed Plots of Abbotsford Study Well Chemistry 286 Appendix E: Detailed Plots of Brookswood North Study Well Chemistry 300 Appendix F: Detailed Plots of Brookswood South Study Well Chemistry 312 Appendix G: Description of Methods and Findings Set 2 & Set 3 Laboratory Column Tests 325 vi i LIST OF TABLES Table 3.1 Summary of column tests performed Table 3.2 Summary of column test conditions - Set 1 Table 3.3 Chemical analyses on water samples from Abbotsford study well Table 3.4 Chemical analyses on water samples from Brookswood North study well Table 3.5 Chemical analyses on water samples from Brookswood South study well Table 4.1 Dissolved oxygen levels in study well - March 1995 Table 4.2 DO measurements from B C M O E wells A , B & C Table 4.3 Ammonium in study well - March 27, 1995 Table 4.4 Carbon content of soils from study well Table 4.5 Cation exchange characteristics of soils from study well Table 5.1 Dissolved oxygen levels in North study well - March 1995 Table 5.2 Carbon content of soils from North study well Table 5.3 C rbon content of soils from South study well vi i i Table 5.4 Cation exchange characteristics of soils from North study well Table 5.5 Cation exchange characteristics of soils from South study well Table 6.1 Summary of attached growth systems reviewed Table 7.1 Comparison of GW chemistry at 12m depth in Brookswood wells Table 7.2 Analysis of leachate runoff from poultry manure stockpiles Table 7.3 Comparison of Abbotsford Study well and B C M O E Wells B and C groundwater chemistry ix LIST OF FIGURES Figure 2.1 Nitrogen cycle in the terrestrial environment Figure 2.2 Examples of groundwater nitrate distributions from regional water supply well monitoring Figure 2.3 Schematic representations of groundwater flow systems including unconfined flow and riparian zones Figure 2.4 Vertical distribution of groundwater nitrate and dissolved oxygen data from Long Island, New York Figure 2.5 Vertical distribution of groundwater nitrate in chalk and sandstone outcrop aquifers in Britain Figure 2.6 Vertical distributions of groundwater nitrate in layered aquifer system in Nebraska, USA Figure 2.7 Vertical profiles of groundwater chemistry obtained from three wells in unconfined aquifer in Denmark Figure 2.8 Vertical summer and spring groundwater nitrate, DO and CI\" profiles from unconfined aquifer in Ontario, Canada Figure 2.9 Detailed vertical groundwater chemistry profiles obtained using dialysis membrane passive sampling below the water table of an unconfined sand and sandstone aquifer in Israel Figure 2.10 Vertical profiles of groundwater chemistry obtained from an unconfined aquifer in the Netherlands Figure 2.11 Vertical profiles of groundwater chemistry at the Fuhrberg aquifer in Germany 28 Figure 2.12 Vertical profiles of groundwater chemistry as related to surficial land use 30 Figure 2.13 Vertical profiles of groundwater nitrate, DO and DOC as measured and modelled by Kinzelbach et a l . , (1991) 32 Figure 2.14a Vertical profiles of groundwater and soil chemistry at the Rodney field site, Ontario 33 Figure 2.14b Vertical profiles of groundwater and soil chemistry at the Alliston field site, Ontario 34 Figure 2.15 Vertical profiles of groundwater to illustrate the occurrence of heterotrophic denitrification within aquifers 36 Figure 2.16 Hydrochemical and isotopic variations along groundwater flow path in a chalk aquifer suggesting the occurrence of denitrification 50 Figure 2.17 Areal distribution of groundwater nitrate levels, and delineation of redox zones in a shallow aquifer underlying agricultural lands in Nebraska, USA 52 Figure 2.18 Longitudinal profiles of hydrochemistry in vicinity of redox front shown on Figure 2.17, used to infer denitrification 53 Figure 2.19 Simplified profile of groundwater flow trajectory in an unconfined aquifer receiving uniform recharge 55 Figure 2.20 Measured and simulated tritium distributions beneath the water table of an unconfined aquifer in Delmarva Peninsula, USA 58 xi Figure 2.21 t vs d/D for various values of P/D calculated using equation 2.13 60 Figure 2.22 Aquifer cross-section simulated by DIVAST 63 Figure 2.23 Cross sectional flow net and water table recharge function used to model data shown on Figure 2.11 64 Figure 2.24 Concentration profile beneath water table receiving unit concentration of leached contaminant as computed for various values of vertical dispersion 65 Figure 2.25 Theoretical 500 year concentration profile history at a well location assuming vertical dispersion = 0 m2/yr 66 Figure 2.26 Theoretical 500 year concentration profile history at a well location assuming vertical dispersion = 1 m2/yr 67 Figure 2.27 Calculated boundaries between denitrification and desulfurication zones using transverse dispersivity of 1) 0m, 2) 0.001m, 3) 0.01m 68 Figure 2.28 Calculated boundaries between denitrification and desulfurication zones using nitrate half life of 1) 1 year, 2) 1.5 years, 3) 2.3 years Figure 3.1 Locations of Brookswood and Abbotsford aquifers within the Lower Fraser Valley region in southwestern British Columbia (Carmichael & Wei, 1994) 77 Figure 3.2 Site map of Abbotsford study well area 78 Figure 3.3 Site map of Brookswood study well area 79 Figure 3.4 Schematic description of study wells including completion details 85 69 xii Figure 3.5 Details of monitoring well passive sampling device 87 Figure 3.6 Schematic description of laboratory column test setup 92 Figure 4.1 Geological fence diagram over region including Abbotsford study site 112 Figure 4.2 Monthly depths to water table in study well during monitoring period, and monthly precipitation at Abbotsford Airport between 1992 and June 1995 114 Figure 4.3 Monthly total precipitation at Abbotsford Airport, and water table fluctuations in monitoring wells in vicinity of study site between 1990 and 1993 115 Figure 4.4 Water table configuration and groundwater flow patterns for Abbotsford aquifer 117 Figure 4.5 Area associated with land use changes during monitoring period in vicinity of Abbotsford study well 118 Figure 4.6 Combined plots of monthly major ion chemistry, pH, IC and TOC profiles obtained for groundwater at Abbotsford study well 120 Figure 4.7 Monthly groundwater nitrate profiles at Abbotsford study well 124 Figure 4.8a temporal variation of groundwater nitrate as measured within B C M O E well A 126 Figure 4.8b temporal variation B C M O E well B of groundwater nitrate as measured within 127 Figure 4.8c temporal variation of groundwater nitrate as measured within B C M O E well C 128 xii i Figure 5.1 Vertical geological profile beneath Brookswood study site 136 Figure 5.2 Depths to water table below ground surface at Brookswood study wells 137 Figure 5.3 Combined plots of monthly major ion chemistry, pH, IC and TOC profiles obtained for groundwater at Brookswood North study well 140 Figure 5.4 Monthly groundwater nitrate profiles at Brookswood North study well 143 Figure 5.5 Combined plots of monthly major ion chemistry, ph, IC and TOC profiles obtained for groundwater at Brookswood South study well 145 Figure 6.1a Effluent characteristics measured during Phases 1 and 2 of Set 1 laboratory testing 153 Figure 6. lb Effluent characteristics measured during Phases 3 and 4 of Set 1 laboratory testing 154 Figure 6.1c Effluent characteristics measured during Phases 5 and 6 of Set 1 laboratory testing 155 Figure 6.2 Effluent characteristics measured during Set 6 laboratory column testing 163 Figure 7.1 Schematic vertical cross section illustrating possible relationship between land use and underlying groundwater chemistry in a recharge zone of an unconfined aquifer 180 Figure 7.2 Average nutrient content for commercial and hatching egg layer, turkey and chicken production flocks 184 Figure 7.3 Influence of rate of poultry manure application on soil nitrate-N to 60cm during the growing season in three experimental plots 185 xiv Figure 7.4 Monthly groundwater nitrate and pH profiles at Abbotsford study well 190 Figure 7.5 Groundwater nitrate vs time over depth below ground surface at Abbotsford study well 197 Figure 7.6 Conceptual model showing temporal changes in groundwater flow trajectories in response to fluctuations in water table 200 Figure 7.7 Schematic description of downward movement and dispersion of leached nitrate front below water table 202 Figure 7.8 Groundwater age vs depth as calculated using eq. 2.13 for Abbotsford and Brookswood study sites 204 Figure 7.9 Soil CEC & TEC profiles including groundwater pH and combined monthly groundwater nitrate profiles at Abbotsford study well 206 Figure 7.10 Comparison of monthly N0 3 \" and CI\" profiles from Abbotsford study well 210 Figure 7.11 Monthly groundwater Nitrate-N/Cl profiles from Abbotsford study well 211 Figure 7.12 Monthly groundwater Nitrate-N/Cl profiles from Brookswood study wells 213 Figure 7.13 Plots of groundwater IC vs P H for several months from Abbotsford study well 215 Figure 7.14 Comparison of groundwater nitrate profiles obtained at study wells over consecutive months during period of no recharge 223 Figure 7.15 Plots of data contained in Table 7.3 226 xv ACKNOWLEDGMENTS This work was performed under the supervision of James Atwater, Department of Civi l Engineering. His support, and the free rein, provided throughout the course of the research evolution is gratefully acknowledged. Funding for this work was provided through Environment Canada's Green Plan. The total and enthusiastic support provided by Hugh Liebscher (Environment Canada) towards securing these funds, as well as throughout the research, is also gratefully acknowledged. The rest of the research supervisory committee, including Drs. Richard Campanella and Ken Hall (UBC Civi l Engineering) and Dr. John Paul (Agriculture Canada), are thanked for their guidance of the research, as well as their timely and constructive review of the thesis drafts. Dr. Robert Gillham (University of Waterloo) served as the external examiner for this thesis, and his constructive comments have helped shape its final form. Several individuals have provided valuable data towards this thesis. They include Alan Kohut, Rodney Zimmerman and Mike Wei (BC Ministry of Environment, Lands & Parks - Groundwater Management Group); Steven Cox (USGS, Water Resources Division, Tacoma, Washington); Dr. Bernie Zebarth (Agriculture Canada); and Kevin Chipperfield (Sustainable Poultry Farmers Group). Their contributions are acknowledged through the thesis. Guy Hirsch (UBC Engineering Workshop) is thanked for his timely assistance in fabricating the well sampling devices, and also the laboratory column setups. The laboratory analyses was performed in part by Jufang Zhou and Paula Parkinson, in the Environmental Engineering Laboratory. Susan Harper ensured the timely supply of materials, while patiently accepting mundane questions. Finally, thanks to Natalie, Rohan and Niran. We did it!!! xvi 1.0 INTRODUCTION l Shallow unconfined aquifers which are recharged by infiltrating surface precipitation are prone to nitrate contamination as a result of surface land use practises. The extensive and widespread use of nitrogen based fertilizers is generally considered to have resulted in the non-point contamination of groundwater resources by nitrate (N03\") at levels frequently exceeding the current health based guideline of 10 mg-N/L for drinking waters. Nitrate contamination of groundwater can also originate from localized, or point, sources such as septic tank drain fields, manure stockpiles, animal feeding and handling areas, uncontrolled landfills, and sludge discharges to land (Hallberg, 1989; Harper et a l . , 1992; Keeney, 1989; Kolpin et a l . , 1994; Robertson et a l . , 1991; Spalding et a l . , 1993). Septic tank drain fields are unique in one aspect in that, though individually they could be considered point sources, on a regional basis, with hundreds or thousands of drain fields present, they are collectively, considered a non-point source1 (Hantzsche & Finnemore, 1992; Perlmutter & Koch, 1972). As a consequence of the combined releases from the various sources, nitrate has been found to be the most ubiquitous contaminant in groundwater (Freeze & Cherry, 1979). Recent studies in Canada (Liebscher et a l . , 1992) and other parts of the world (FAO, 1993) show nitrate levels to be increasing in unconfined aquifers. This represents an increased risk as groundwater is the only source of potable water for over one quarter 1 Defining features of a non-point source, as suggested by Dean & Gaboury (1987) are: 1) pollutant releases from the source are discontinuous in time and are driven by natural hydrological processes, and 2) the source is spatially diffuse and exact entry points into the receiving waters cannot be easily determined. 2 of all Canadians (Government of Canada, 1990), and an even higher portion of populations in many parts of the world. The presence of nitrate in groundwater poses a threat not only to subsurface drinking water supplies, but also to surface water bodies (lakes and rivers) into which the groundwater discharges (Cherkauer et a l . , 1992; Spalding & Exner, 1993). Elevated nitrate levels in rivers and lakes, as well as coastal waters, can contribute to eutrophication and subsequent degradation, if nitrogen is the limiting nutrient in these environments. Point sources of nitrate have the capacity to be readily controlled. In the case of septic tank drain fields, there are several recently developed and promising technologies available, but which have yet to be widely applied. These include new septic tank and drain field configurations, as well as the provision of carbon amended reactive barriers to promote denitrification (Gold et a l . , 1992; Hagedorn & Reneau, 1994; Robertson & Cherry, 1995). Controlled site selection and preparation (such as the provision of impermeable barriers to leakage), as well as some form of pretreatment, offer means by which groundwater contamination from other land discharged waste may also be minimised. Several researchers have investigated the potential for adding carbon to groundwaters to promote and enhance denitrification below the water table (Adelman & Spalding, 1988; Adelman & Dahab, 1991; Childs et a l . , 1988; Duthy, 1993; Hiscock et al . , 1991; Rott & Lambeth, 1992). While theoretical and laboratory based studies appear to show that such enhanced denitrification could be used as an in-situ method for controlling nitrate levels, there has been no field scale demonstration of its effectiveness or widespread applicability. One uncertainty currently associated with in-situ enhanced 3 denitrification is the potential for permeability reductions below the water table as a result of N 2 gas bubble accumulations (Soares et a l . , 1991). Though in-situ nitrate reduction approaches may appear feasible, source management represents a more rational long-term alternative, as groundwater treatment would be avoided. Source management of non-point nitrate contamination from agricultural activity continues to present a complex management challenge. As noted by Bouchard et al. (1992), \"the individual who can have the greatest effect on the protection of groundwater from nitrate contamination - the farmer - is sent conflicting signals by an agricultural policy that encourages high nitrate application for maximizing yields, and by an environmental policy that discourages such practises\". There is ongoing research in this area (Francis, 1992; Hall , 1992; Pan, 1994), but the optimization of fertilizer use is confounded also by site and crop specific requirements, as well as seasonal variability. Over-fertilization is practised, and leaching losses continue to be significant. The findings of Hall (1992), from field trials, show that controlled nutrient source management can minimize the extent of nitrate contamination of groundwaters. Effective monitoring methodologies are required in order to implement source management schemes. Though it is currently not widely practised, routine monitoring of the unsaturated (vadose) zone, between the crop root zone and the underlying water table of unconfined aquifers, can provide a means of detecting nitrate leaching losses, and optimizing fertilizer application (Foster et a l . , 1985; Grantham & Lucas, 1985; Spalding & Kitchen, 1988). However, Staver & Brinsfield (1991) considered that, due to the highly variable nature of flow and water storage within the vadose zone, monitoring at and below the water table wil l provide more reliable data with which to quantify the rates 4 of nitrate leaching below non-point surface sources. Detailed measurement of vertical profiles of groundwater chemistry beneath the water table of unconfined aquifers can provide data with which to accurately quantify surface leaching losses, as well as the fate of the leached contaminants (Smith et a l . , 1991). While it is recognised that nitrate leaching below the root zone of agricultural lands occurs as a seasonally driven cycle (Staver & Brinsfield, 1991), detailed measurement of the seasonal arrival of leached nitrate at the water table has not been performed to date. At the current time, monitoring of groundwater resource quality is more commonly performed by sampling from existing water supply wells on a regional basis. Such sampling may be performed annually, though it is more likely that sampling over a given region is less frequent. Monitoring data from water supply wells can be interpreted using a Geographical Information System (GIS) approach to correlate land use practises to groundwater nitrate occurrences (Cook, 1994; Tredoux & du Plessis, 1992). However, as these water supply wells tend to be widely spaced, as well as have intakes that are variable in length and depth, monitoring data obtained from these wells provides little information with which to quantify the dynamics of nitrate transport and fate within the groundwater environment. In particular, monitoring from existing water supply wells cannot be used to quantitatively delineate the arrival of leached nitrate at the water table. Given the widespread consequences of large losses of nitrate from agricultural lands, Burt et al. (1993) proposed that \"a broader catchment based hydrological view is in order\", and stressed the importance of understanding the \"hydrology of nitrate\", from its movement in groundwater systems, to its delivery to surface water bodies. This perspective was also stressed by Nachtnebel & Kovar (1991), while Spalding & Exner 5 (1993) opined that \"the factors affecting the distribution of N0 3 \" in aquifers are complex and poorly understood. Interdisciplinary studies using discrete depth sampling, hydrogeological indicators, isotopic tracers, and microbiological techniques are necessary to unravel the complex dynamics.\" 1.1 RESEARCH OBJECTIVES & SCOPE This thesis describes the findings of research that was commenced with the aim of investigating in some detail the role of carbon-enhanced denitrification for the in-situ control of non-point source contamination within groundwaters. At the onset, the specific objectives consisted of: i) the further investigation of the generation and fate of N 2 under saturated flow conditions through soils during enhanced denitrification, ii) the investigation of the significance of carbon source in the enhancement of denitrification, and iii) the investigation of the feasibility of applying enhanced denitrification for non-point source nitrate control at an aquifer scale. A scope of work consisting of laboratory-based column microcosm flow-through testing, to be followed by carbon injection within local nitrate contaminated aquifers, was conceptually envisaged at the onset of the research. Laboratory column testing was commenced on this basis. Underlying the initial research conceptualization was the presumption that denitrification was occurring naturally at significant levels within local aquifers, and that: a) this denitrification would be readily measurable; and b) that the denitrification would 6 be located within a narrow \"redox-line\" below the zone of nitrate stratification. It was expected that carbon Would be introduced in the vicinity of the redox-line in order to enhance and monitor the denitrification activity there. The identification of a redox-line parallel to the water table requires detailed multi-level sampling of the groundwater chemistry. Such sampling has not been performed within the local (Lower Fraser Valley region of British Columbia) aquifers, and there are only a few instances of such sampling in other parts of the world. A relatively new passive sampling approach was selected, as this method was previously shown (Ronen et a l . , 1986) - although only over small depths below the water table, to be effective in obtaining highly detailed groundwater chemistry profiles beneath the water table. A new sampling device, for passive sampling over extended depth below the water table, was designed and constructed as part of this research. On initial monitoring over of the selected study sites, it was discovered that denitrification was not occurring at significant levels within the local aquifers, nor was there a distinct redox-line below the water table over the monitored depths. The early monitoring did find, however, that the nitrate contamination displayed a sub-annual temporal variation. As no investigations had previously been performed to measure in detail the temporal variation in non-point source nitrate distribution within unconfined aquifers, a long-term monthly monitoring program was commenced at the study sites using the developed multi-level sampling approach. At this stage, it was decided, due to limited financial resources, that the research effort would be concentrated on the long-term monitoring. The laboratory column testing was terminated after a limited, but 7 conclusive, series of tests had been completed. The additional objective, which evolved as a result of the shift in research focus, was: iv) the detailed measurement of the seasonal arrival of non-point source nitrate at, and subsequent distribution beneath, the water table of unconfined aquifers. 1.2 THESIS O R G A N I Z A T I O N A literature review aimed at identifying the current understanding of the distribution, transport and fate of non-point source nitrate in groundwaters; as well as the in-situ control of groundwater nitrate, is presented in Chapter 2. A description of the research methodology is presented in Chapter 3. The results obtained from the field investigations are presented in Chapters 4 and 5. Chapter 6 presents the results and discussion of the laboratory investigations. Chapter 7 contains a detailed discussion of the field and laboratory based findings by relating them to the transport, fate and control of non-point source nitrate below the water table of unconfined aquifers. Conclusions, and recommendations for further research, are contained in Chapters 8 and 9 respectively. 2.0 LITERATURE REVIEW 8 The amount of available literature on non-point source nitrate leaching below the root zone of agricultural lands is large. Research publications on this subject continue to be produced on a regular basis in many journals throughout the world. The motivation behind much of the early research in this area was the better understanding of nitrogen balances within crop-soil systems. A growing awareness of the potential impacts of leached nitrate on the integrity of underlying groundwater resources provided the impetus for the extension of the research into the unsaturated zone, while concomitantly, ongoing monitoring of groundwater resources for nitrate was commenced. While the previously quoted sentiments of Burt et al. (1993) and Spalding & Exner (1993) reflect an apparent limitation in the current state of understanding of the overall transport and fate of non-point source nitrate in groundwater, there is a growing body of findings on specific aspects such as denitrification in aquifers, as well as an increasing ability to quantitatively evaluate the transport and fate of the nitrate within these environments. The objective of this Chapter is to provide a collective review of various recent and published findings towards the development of a more complete understanding of non-point source nitrate travel and fate below the water table of unconfined aquifers. It is hoped, by this synthesis, that an advancement wil l be made towards better understanding the hydrology of non-point source nitrate beneath the water table of susceptible unconfined aquifers. The aspects to be addressed in the following sections include: o nitrogen in the terrestrial environment; 9 and the o observed distributions, o fate and transport, and o control of non-point source nitrate beneath the water table of unconfined aquifers. 2.1 N I T R O G E N IN T H E T E R R E S T R I A L E N V I R O N M E N T Figure 2.1 presents an overview of the nitrogen cycle in the terrestrial ecosystem, and shows several possible, and dominant, processes impacting the fate of nitrogen below the ground surface. The extent to which each of the transformation processes plays a role depends on the form in which the nitrogen enters the ground surface. As shown on Figure 2.1, nitrogen can enter the subsurface environment directly as nitrate (fertilizers and precipitation), or in other forms. Nitrogen in the anionic nitrate (N03~) form is in it's most oxidized state (+5) and, in the concentrations usually present in the terrestrial environment, is generally considered to be non-reactive under oxidizing conditions (Freeze & Cherry, 1979). On the other hand, nitrogen in the cationic ammonium (NFL,*) form is in it's most reduced state (-3), and is susceptible to not only adsorption to negatively charged surfaces of clay and organic matter present within the soil, but also to oxidizing reactions. Nitrogen originating from manure stockpiles, septic tank effluent, and other wastes, may not necessarily be initially present as nitrate or even ammonium, but may be present in some organic form. However, microbial and oxidative processes within the upper soil surface and underlying vadose (unsaturated) zone lead to the transformation of this nitrogen to nitrate. Decomposition of organically bound nitrogen 10 Precipitation N H 3 N H 3 N 0 3 - NO 3-r i Human and animal wastes Mineral ferti l izer Organic N N H 3 N H 3 NO3-N H , Plant residue compost Organic N proteins N 2 I (Ni t rogen^ fixation J 1 Decomposit ion > Nitrif ication ^ \u00E2\u0080\u00A2' \u00E2\u0080\u00A2' J N H / \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 ' (Nitrif icat ion Proteins - ^ D e c o m p o s i t i o n ^ Denitr i f icat ion) '$%ffi$$&\u00C2\u00A3&\u00C2\u00A3$S.D$nitri1lcMion in reducing zones)-:-:] N2(a (itJ Chemical symbols: N, elemental nitrogen; N03\", nitrate; N2, nitrogen gas; NH 3 , ammonia; N 20, nitrous oxide; NH 4 + , ammonium; N0 2\" nitrite; N2(aq), nitrogen gas; dissolved in water. Figure 2.1 Nitrogen cycle in the terrestrial environment (from Dillon, 1988) 11 (amino acids, polypeptides and proteins) that may originate from waste sources leads to nitrogen mineralization to ammonium (\"ammonification\"). This ammonium is then converted to nitrate (\"nitrification\") by nitrifying bacteria in the presence of oxygen. Nitrate can be reduced to N 2 gas, or even ammonium. In the subsurface environment the associated reactions are also mediated by microbes, which couple the nitrate reduction (electron accepting) process with the oxidation (electron donating) of a reduced compound, while using the electron transfer as a source of energy. Nitrate reduction to N 2 , in this dissimilatory manner, is the \"denitrification\" process identified on Figure 2.1. Microbes capable of denitrification are considered ubiquitous in the subsurface environment (Ghirose & Wilson, 1988; Korom, 1992; Tiedje et a l . , 1982). Denitrification is a process that has been extensively investigated and documented (eg. Payne, 1981). It continues to be investigated not only at a fundamental level, but also as a process utilizable for wastewater treatment and as a fate process controlling nitrogen in the environment (Burt et a l . , 1993). It is commonly believed that an oxygen free environment, combined with an availability of biodegradable organic carbon (as the electron donor, or \"energy\" source), are preferred pre-requisites for microbially mediated denitrification to proceed (Tiedje, 1988). Such \"reducing\" conditions are considered to be present to some extent within the root zone (particularly under flooded conditions), and can also be present, to a lesser and varying degree, in the underlying unsaturated and saturated zones. In the absence of sufficient reducing conditions, nitrate can leach below the root zone and move freely with infiltrating recharge waters to the groundwater table. A more detailed discussion on denitrification in the subsurface environment, as well as alternatives to organic carbon as the energy source, is presented in Section 2.3. 12 In addition to the anthropogenic sources such as fertilizer application, natural sources (plant residue, nitrogen fixation, and precipitation, as shown on Figure 2.1) result in the presence of a large nitrogen pool within the upper soil zone. Though much of this nitrogen remains within this zone in natural environments, with a balance between plant uptake and transfers to the atmosphere, some leaching of nitrate to groundwater can still occur (Krajenbrink et a l . , 1988). This leaching is substantially increased by upsets to the nitrogen balance, such as natural or anthropogenic deforestation, and conversion of grasslands to arable lands (Frind et a l . , 1990; Keeney, 1989; Tredoux & du Plessis, 1992). Natural subsurface sources of nitrate, such as organic rich shales or buried cave deposits, can also exist (Feth, 1966). Despite these sources, it is considered that nitrate levels in unpolluted (from anthropogenic sources) groundwaters rarely reach or exceed the drinking water guideline of 10 mg-N/L. On the basis of an extensive survey of approximately 120,000 water supply wells located throughout the United States, Madison & Brunett (1984) considered that N0 3 \" values less than 0.2 mg-N/L may be \"assumed to represent natural background concentration\"; values between 0.2 mg-N/L and 3.0 mg-N/L \"may or may not represent human influence\"; and, values above 3.0 mg-N/1 \"may indicate elevated concentrations resulting from human activities\". 2.2 OBSERVED DISTRIBUTIONS OF NON-POINT SOURCE NITRATE IN UNCONFINED AQUIFERS Figure 2.2 presents two examples of typical plots of groundwater nitrate data collected on a regional scale by the monitoring of water supply, and other similarly installed, wells in unconfined aquifers. Such distributions have been reported from many regions of the world, and they all show the susceptibility of unconfined aquifers to Figure 2.2 Examples of groundwater nitrate distributions from regional water supply well monitoring - a) USEPA (1992); b) Carmichael & Wei (1994) 14 surface source nitrate contamination (Hallberg, 1989). The usefulness (towards quantifying the transport and fate of nitrate) of such distributions is often limited by the absence of any knowledge of the geologic setting and groundwater flow conditions. For example, Figure 2.2b presents a compilation of data collected from a region underlain by several unconfined and unconnected aquifers, so these data would need to be separated before any aquifer-specific evaluation could be performed. Figure 2.3 provides a simplified overview of a groundwater flow system, and illustrates the variability that can be associated with the groundwater flow directions in relation to surface characteristics, including recharge and discharge zones. 2.2.1 Vertical Profiles of Groundwater Chemistry - Detailed Case Studies An early example of a detailed and site-specific appraisal of groundwater nitrate distribution was described by Perlmutter & Koch (1972) for a region (in Long Island, New York) impacted by a high density of septic tank drain fields as well as irrigated fertilizer use. Figure 2.4 presents a vertical cross-section showing the distribution of groundwater nitrate in the region, as related to groundwater flow directions and pumping well locations. This cross-section shows elevated concentrations of nitrate extending to a large depth (up to 120 m, or 400 ft) below the water table at, and some distance downgradient of, the groundwater divide where there is a large vertical component to the groundwater flow. The natural vertical component of flow near the groundwater divide is reportedly also enhanced by the deep pumping wells (Philips & Gelhar, 1978). At further distances from the divide, and where the groundwater flow is horizontal, it was found that elevated nitrate levels occurred within an upper stratified zone at the 15 SOIL R O O T Z O N E Figure 2.3 Schematic representations of groundwater flow systems including unconfined flow and riparian zones (Lowrance & Pionke, 1989) 16 Brook ville \u00E2\u0080\u00A2ft 0 1 2 3 MILES 1 i i i I i i VERTICAL SCALE IS GREATLY EXAGGERATED Figure 2.4 Vertical distribution of groundwater nitrate (upper value shown) and dissolved oxygen (lower) data from Long Island, New York. Data shown at base of wells (Perlmutter & Koch, 1972) 17 water table. The cross-section also shows that, at that particular site, the elevated nitrate levels to be associated with oxygenated groundwaters, while the underlying waters were oxygen-free. Perlmutter & Koch (1972) considered that the observed nitrate and oxygen distributions reflect oxygenated and nitrate contaminated recharge waters displacing older oxygen-free waters1. The oxygen loss in the older waters was considered to be due to consumption through oxidation of lignite present in the aquifer solids. Perlmutter & Koch further speculated that the movement of the \"nitrate front\" may be \"partly controlled by natural chemical reduction of the nitrate to ammonium ions in the reducing environment near . . . . the line of zero dissolved oxygen content.\" However, data confirming this speculation was not obtained by Perlmutter & Koch. Figure 2.5 presents detailed groundwater chemical profiles obtained from single wells by Foster et al. (1982) within sandstone and chalk aquifer outcrop areas in Britain, where the groundwater flow direction is downward. The method of water sampling adopted by Foster et al. (1982) apparently consisted of extracting porewater from sections of drill core. Nitrate at levels exceeding 10 mg-N/1 was found to be extending at least 70 m below the water table in the sandstone aquifer outcrop area, and groundwater dating using tritium measurements showed that the deeper nitrate may be have originated at the ground surface several decades prior to the time of measurement. On the basis of their findings, Foster et al. concluded that there was \"little evidence of denitrification insitu\", but did not entirely discount its existence. The delivery of nitrate to large depths below the water table, as a result of 1 Italics added for emphasis. This approach will be adopted throughout this Thesis as appropriate to emphasise aspects of the overall discussion. Figure 2.5 Vertical distribution of groundwater nitrate in a) chalk and b) sandstone outcrop aquifers in Britain (Foster et al., 1982) 19 downward flow, has been observed elsewhere. In California, elevated nitrate concentrations have reportedly been measured to depths of 400m below ground surface (Freeze & Cherry, 1979). Spalding & Exner (1980) investigated the areal, vertical and temporal distributions in groundwater chemistry by pumping from well clusters with \"shallow\", \"medium\" and \"deep\" intakes placed at 2 m, 9 m and 18 m below the water table of a sand and gravel aquifer system in Nebraska, USA. Well intakes were situated within an upper unconfined and principal aquifer, as well an underlying secondary aquifer, as shown on Figure 2.6. Intake lengths for each well were approximately 0.6 m. Figure 2.6 also contains some chemical profiles obtained by these investigators, and which show elevated nitrate concentrations near the water table of the principal aquifer. Details on the hydraulic continuity between the principal and secondary aquifers were not obtained by these investigators. Despite the large intervals between monitored depths, Spalding & Exner (1982) were able to generally conclude, on the basis of periodic monitoring, that temporal fluctuations in the groundwater nitrate levels occur near the water table in response to seasonal precipitation and fertilization patterns. Several other investigators have also obtained detailed vertical profiles of groundwater nitrate distribution beneath agricultural lands, and these findings are summarised in the following pages. Andersen et a l . , (1980) obtained chemical profiles beneath the water table of an unconfined outwash sand and gravel aquifer in Denmark. They used a drive point sampling technique which provided a comparatively greater extent of vertical detail as shown on Figure 2.7. The thickness of the aquifer is 25 m to 40 m, while annual 22 or\u00C2\u00A9 e i \u00E2\u0080\u00A2 ON DEPTH BtLOW WATER TABLE (M) 02 Test hole no 1 DYBDAL D.G.U. Well tile no. 76IK3 Ft mg/l 2 \u00E2\u0080\u0094I 1 NOJ-N mg/l 20 ^ 5 20 30 SO* 40 H-mg/l 50 30 CI \" S / l ta SO 100 1RI1IUM I U 200 U)0 WI I. I N H , N < Q O 8 mj'u Di-Figure 2.7 Vertical profiles of groundwater chemistry obtained from three wells in unconfined aquifer in Denmark (Andersen et al., 1980) 22 recharge is 375 mm. On the basis of their findings, these investigators concluded the presence of an upper \"zone of oxidation\" containing nitrate, and a lower \"zone of reduction\" containing no nitrate but reduced species such as F e 2 + . No measurements of dissolved oxygen were presented by Andersen et a l . , but they speculated that oxidation of F e 2 + coupled with the reduction of nitrate may explain the sharp nitrate front observed on the vertical chemical profiles. They represented this reaction in the following unbalanced form: 5 F e 2 + + NO; => \u00E2\u0080\u0094 N2 + Fe(OH)3 (2.1) Egboka (1984) investigated the vertical distribution of nitrate contamination of several unconfined glacial lake sand aquifers in Ontario, Canada, using multi-level piezometers, and one example of his findings is shown on Figure 2.8. The two sets of profiles, obtained during different years, shown on this figure also show the presence of an upper \"oxic\" zone where nitrate, as well as oxygen, is present. The profiles also display a relatively steep chemical gradient to an underlying anaerobic zone. By interpreting the C17N03\" ratio, Egboka used these data to infer denitrification to be controlling nitrate levels at depth. This aspect wil l be discussed later in this Chapter. Ronen et a l . , (1986) described the development and application of a multi-layer sampler that permitted \"passive\" sampling of groundwater dissolved constituents at vertical intervals of 0.03 m by diffusion into dialysis cells. Figure 2.9 shows \"microscale\" profiles that these investigators obtained beneath the water table of an unconfined sand and sandstone aquifer located in Israel, and where the water table was at a depth of 27 m below ground surface at the time of monitoring. Land use over this 23 H 3 4 H 3 4 N O f , D.O., Cl\"x 10\"'(mg/l) N O ~ , O . Q , Crx 10\" (mg/1) 0 10 20 0 10 20 Figure 2.8 Vertical summer and spring groundwater nitrate, DO and CI\" profiles from unconfined aquifer in Ontario, Canada (Egboka, 1984) 24 Figure 2.9 Detailed vertical groundwater chemistry profiles obtained using dialysis membrane passive sampling below the water table (depth to groundwater 27m) of an unconfined sand and sandstone aquifer in Israel (Ronen et al., 1986) 25 aquifer includes long term agricultural irrigation by sewage effluent (Ronen et al., 1987a). Through repeated monitoring, these investigators also observed a temporal variation in these profiles (Ronen et al., 1987b), but the limited monitoring time frame provided no clear pattern to this variation. Ronen et al. (1987b) did not speculate on the fate of the observed nitrate, but used their findings to demonstrate the limitations of using large intake pumping well data to interpret groundwater chemistry which, in reality, may be highly stratified (in \"slugs\" or \"parcels\" of water). Van Beek et al. (1989) obtained detailed vertical profiles of the groundwater chemistry in a shallow unconfined aeolian sand aquifer, also in the Netherlands, in order to evaluate the effects of manure spreading and acid rain deposition on groundwater quality. Annual recharge at this site is 150 mm to 300 mm, and Van Beek et al. estimated the travel time through the unsaturated zone to be 1 to 2 years. \"Miniscreen\" wells, which permitted sampling at 1 m intervals, were used by these investigators to obtain the profiles shown on Figure 2.10. No data on dissolved oxygen was presented by Van Beek et al. The miniscreen wells were located within 1 km of a water supply well field - with wells in this field having their intake interval at 15 m to 30 m below the ground surface. On this basis, Van Beek et al., assumed groundwater flow at the miniscreen wells to be predominantly horizontal, and the encountered shallow acidic waters were considered to have originated from acid rain deposition near the miniscreen wells, while the deeper nitrate layer originated from areas upgradient of the wells. These investigators concluded on the basis of the obtained profiles that nitrate reduction below a depth of about 20 m resulted due to nitrate reduction coupled with pyrite oxidation -as evidenced by increased sulfate, iron and arsenic (associated with pyrite), and a Figure 2.10 Vertical profiles of groundwater chemistry obtained from an unconfined aquifer in the Netherlands (Van Beek et al., 1989) 27 reduction in pH. The sharp sulfate front at a depth of 30 m was considered to be associated with a zone of sulfate reduction (coupled to oxidation of solid organic carbon present in the aquifer sediments), and associated precipitation of iron, arsenic, zinc and nickel, as well as an associated increase in pH. Van Beek et al (1989) represented the nitrate and sulfate reduction processes, inferred at this site, as follows: 2FeS2 +6NO; +\u00C2\u00A3H20 =* 3N2\ +2Fe (OH) 3 { a ) +4SOl~ +2H+ <2-2) 2CHzO(s) + SOl~ \u00E2\u0080\u00A2* HCO; + H2S (2-3) Frind et al. (1990) described chemical profiles beneath the water table of an unconfined \"virtually carbonate free\" quaternary sand aquifer in Germany, and these are shown on Figure 2.11. Annual recharge to this aquifer is spatially variable, and ranges from 100 to 250 mm. The thickness of the aquifer is between 20 m and 40 m. Miniscreen wells were also used by these investigators, but the obtained profiles appear less resolved than those of Van Beek et al. (1989). Though not shown on the profiles reproduced in Figure 2.11, Frind et al. reported that the aquifer is virtually oxygen free and that only \"trace\" levels of oxygen were measured at the water table. The nitrate distribution at the site also displayed stratification near the water table, and as with the findings of Van Beek et al. (1989), Frind et al. also concluded nitrate reduction coupled with pyrite oxidation to be affecting the observed vertical distribution. Furthermore, Frind et al. also concluded the reduction of sulfate over depth, and thus defined the \"zone of denitrification\" and \"zone of desulfurication\" as shown On Figure 2.11. These findings wil l be discussed in further detail later in this Chapter. Frind et al. represented the process of nitrate reduction coupled with pyrite oxidation as follows: Figure 2.11 Vertical profiles of groundwater chemistry at the Fuhrberg aquifer in Germany (Frind et al., 1990) ro oo 29 5 F e S 2 +14NO; +4H+ \u00E2\u0080\u0094 7N2 +105O 4 2 \" + 5 F e 2 + +2tf 2 0 ( 2 \u00C2\u00AB 4 ) This representation is different to that shown in equation (2.2), and the reason for this is not apparent as Frind et al. (1990) cite the same source reference as did Van Beek et al. (1989). While equation (2.4) suggests that the reduction of nitrate in this manner should be accompanied by an increase in pH, the profiles shown on Figure 2.11 are not resolved enough to identify this occurring. Postma et al. (1991) extended the previously described monitoring performed by Andersen et al. (1980) to obtain the vertical cross section shown on Figure 2.12, and which compares the distribution of nitrate beneath arable (agricultural) and forested lands. Depth to the water table is approximately 15 m below ground surface, and Postma et al. reported that the water table experiences a \"small\" annual fluctuation of less than 1 m. The sampling methodology was described as consisting of the collection of pore water at 1 m intervals using a N 2 gas-driven double line approach. The obtained profiles show the stratification of nitrate in an upper oxidized zone, and also the effect of recharging nitrate-free waters (beneath forest and heath land) on reducing the distribution of nitrate originating from upgradient arable lands. Postma et al. use the term \"redoxline\" to describe the sharp nitrate (and oxygen) front that divides the upper aerobic and lower anaerobic zones. These investigators not only concluded that nitrate reduction coupled with pyrite oxidation was the process controlling the fate of nitrate at and below the redoxline, but more significantly that \"the sharp decrease of both nitrate and oxygen concentrations at the redoxline indicates that the kinetics of the reduction processes are fast compared to the downward transport rate.\" By modelling the observed geochemistry at this site, Postma et al. (1991) show that even though the vertical transport rate for the ARABLE + FOREST and HEATH +\u00C2\u00BB ARABLE -H Figure 2.12 Vertical profiles of groundwater chemistry as related to surficial land use (Postma et al., 1991) 31 groundwater has an estimated value of about 0.75 m/year, the position of the redox-line remains relatively constant, with a possible downward extension into the aquifer at a rate of a few millimetres to centimetres per year. Kinzelbach et al., (1991) obtained the vertical profiles shown on Figure 2.13 from monitoring of an alluvial sandy aquifer in Germany, by multilevel sampling using a stationary packer system. Annual recharge to this aquifer is 100 mm to 250 mm, and the thickness of the aquifer is 30 m. Unlike the previously described profiles, the profiles obtained by Kinzelbach et al. do not show the presence of a sharp nitrate front, but instead a gradual decrease in nitrate (and oxygen) concentration with depth. On the basis of numerical modelling, these investigators concluded that the observed profile was probably a result of \"a succession of zones with different nitrate input upgradient of the observation well\", and not due to nitrate reduction by denitrification. Figure 2.14 presents profiles obtained by Starr & Gillham (1993) during their investigations of organic carbon availability and its role in controlling nitrate distributions in two unconfined sand aquifers in Ontario, Canada. These investigators also obtained core samples, and performed bioassays to evaluate the differences in denitrification capacity and limiting factors between the two sites. \"In-situ Microcosm\" (ISM as shown on Figure 2.14) testing was also performed for the in-situ evaluation of denitrification potential and capacity. The authors concluded that shallow unconfined aquifers could provide conditions for biodegradable (\"labile\") organic carbon to leach below the root zone and reach the water table, and thus an opportunity for denitrification. However, carbon depletion within the unsaturated zone of deeper unconfined aquifers is considered to limit the subsequent potential for denitrification below the water table. The -DO (mg/l) 0.00 1.00 2.00 3.00 4.00 3.00 0.00 -3.00 -\u00C2\u00AB.oo -soo si--12.00 \u00E2\u0080\u00A2 \"7 -15.00 -la.oo -21.00 -24.00 -27.00 -30 00 aquifer depth (m) 0.00 0.00 nitrate (mg/l) 100.00 200.00 300.00 400.00 900 00 -9.00 -12.00 -15.00 -18.00 DOC (mg/l) 100 2.00 3 00 4 00 5 00 -30.00 aquifer depth (m) -30.00 aquifer depth (m) + measured values modelled profile Figure 2.13 Vertical profiles of groundwater nitrate, DO and DOC as measured and modelled by Kinzelbach et al., (1991) CO r o L i t ho logy 1 rift rumen- DO NO\", CI\" DOC t a l i o n (mg/L) (ng-tf/L) (mg/L) (mg/L) 3-4-Tan f i n o s and Orange mo 11 T \u00C2\u00AB d Gray I i ne sand Gray c l a y 10 0 8 0 25 0 40 0 10 0 J 1 1 I I I I I I I -I I L_J I l i i . i l - N -2 - H -2 - N-3 o o o o o M a y 79 o n o o o Aug 85 ***** Dec 85 \u00E2\u0080\u00A2-\u00C2\u00BB-\u00C2\u00BB-\u00E2\u0080\u00A2-\u00E2\u0080\u00A2 J u l 11 86 J u l 25 86 SOC (\u00C2\u00AB) 0 o 1 i i i i Eh ( mV) Vertical profiles of groundwater and soil chemistry at the Rodney field site, Ontario, as described by Starr & Gillham (1993) L i t h o l o g y I n a t r u m e n - 0 0 NO\", C I * DOC SOC t a t i o n ( m g / L ) ( m g - N / L ) ( m g / L ) ( m g / L ) ( \u00C2\u00AB ) E h (mV) 10 15-\u00C2\u00AB \u00C2\u00AB a. a a -\u00C2\u00ABi o at a _j i I 12 0 l l l l 40 0 _I_J L_J I 25 0 4 0.0 0 B500 600 l l l l A - dark brown medium l a n d B - b r o w n medium l a n d C - t a n f i ne aand D - brown f i n e sandy s i l t E - t a n medium l a n d F - t a n m e d i u m a n d c o a r s e sand w i t h o c c a s i o n a l t i n e g r a v e l G - t a n m e d i u m s a n d Vertical profiles of groundwater and soil chemistry at the Alliston field site, Ontario, as described by Starr & Gillham (1993) 35 implications of these findings wil l be discussed in further detail later in this Chapter. Figure 2.15 presents a series of profiles reproduced from Appelo & Postma (1993). These profiles were reportedly obtained from an unconfined aquifer in Germany. Appelo & Postma provided no further details on the site characteristics, but used these profiles to demonstrate the occurrence of nitrate reduction coupled with organic matter oxidation. 2.2.2 Summary The various profiles shown on Figures 2.3 to 2.15 collectively show the occurrence of nitrate contamination of unconfined aquifers under agricultural lands. The stratification of this nitrate contamination within the upper region of the groundwater is also shown. The several findings of a sharp concentration gradient, between the upper contaminated waters and the lower groundwater, also demonstrate the necessity and usefulness of single well (or location) multi-level sampling (as also discussed by Smith et a l . , 1991) for proper delineation and evaluation of stratified groundwater nitrate. The several findings of a sharp nitrate front have also been individually interpreted to conclude that nitrate reduction processes are totally, or partially, responsible for the development and maintenance of the steep chemical gradients. Furthermore, measurements of temporal variations in these profiles, though limited to short periods and large frequencies of monitoring, have shown that leaching to the water table does not result in the temporally continuous input of nitrate at the water table. Figure 2.15 Vertical profiles of groundwater chemistry used by Appelo & Postma (1993) to illustrate the occurrence of heterotrophic denitrification within aquifers 37 2.3 F A T E AND TRANSPORT OF NON-POINT SOURCE NITRATE IN UNCONFINED AQUIFERS For agricultural lands, the transport and fate of leached nitrate (as introduced, or formed during infiltration through the surface soils) from the ground surface through the vadose zone, and subsequently to the groundwater, has been, and continues to be, the subject of a great deal of research (for example, Alva, 1992; Bergstrom & Johansen, 1991; Devitt et a l . , 1976; Foster et a l . , 1985; Gerhart, 1986; Herbel & Spalding, 1993; Pratt et a l . , 1972; Verdegem & Baert, 1984, 1985; Walther, 1989). The many findings show that the extent and rate of nitrate leaching out of the root zone, and through the vadose zone, will be variable and site specific. Controlling factors include the soil and crop type, irrigation pattern, fertilizer application, and climate. In addition to many uncertainties associated with the fate processes within the root zone, the simplifying assumptions required to overcome the complexity and spatial variability associated with the leaching process make the accurate quantification of nitrate leaching, and subsequent loading on groundwater, a difficult, if not impossible task. Nonetheless, it has been shown that even simplified models enable qualitative assessments such as the evaluation of sensitivity to land use changes (Geng et a l . , 1989; Hantzsche & Finnemore, 1992; Kelly et a l . , 1991; Mercado, 1976; Rijtema & Kroes, 1991; Rogowski, 1990; Tanji, 1980). Nitrate travel times to the groundwater are also variable.' In permeable formations, and in the presence of preferential flow paths such as root channels, fissures or even sinkholes, nitrate can be flushed rapidly through the vadose zone in response to precipitation events (Iqbal & Krothe, 1995). In less permeable formations, and in the absence of preferential flow paths, the nitrate can reside for possibly decades within the 38 vadose zone, while moving slowly as a front towards the groundwater. Depending on the depth of the vadose zone, several such fronts can be present above the water table at any given time (Pratt, 1972). The vadose zone can act a vast storage area, with a capacity for continued contamination of the underlying groundwater with leached nitrate even after surface source removal (Spalding & Kitchen, 1988). One common feature of nitrate contamination of shallow unconfined groundwaters beneath agricultural lands is the annually pulsed nature of nitrate leaching below the crop root zone. This is explained by the accumulation of nitrate within the root zone during the warmer summer months (also the growing season during which fertilization occurs), followed by the leaching of this nitrate and subsequent migration towards the water table as a result of winter precipitation recharge (Lunn & Mackay, 1994; Van Der Ploeg et al . , 1995). The two dominant fate processes with a potential to control the concentration of nitrate in groundwater are dispersion and denitrification (Lowrance & Pionke, 1989), with both processes resulting in a decrease in the concentration of nitrate along the groundwater flow path. Dispersion is essentially controlled by the physical characteristics of the flow system and geologic materials, and may be considered an integral part of the transport process. Denitrification is the only process that can be considered to irreversibly remove nitrate from the groundwater, but is not always a controlling process in aquifers. Quantification (by the use of predictive models) of the fate and transport of non-point source nitrate beneath the water table of unconfined aquifers requires a means of defining the physical transport of the nitrate with the groundwater, coupled to a 39 controlling fate process. For contaminated groundwater evaluations, numerical site specific transport and fate models offer a powerful means by which contamination development may be visualised (National Research Council, 1990). Such transport and fate models can only be used with some confidence after extensive validation using site specific data. Detailed subsurface characterization, such as closely spaced multilevel sampling of groundwater, reduces the extent of uncertainty associated with predictive modelling of contamination exhibiting steep chemical gradients and stratification, such as non-point source nitrate in unconfined aquifers. The following sections present a review of findings related to denitrification, as well as several transport models that have been developed in recent years for the interpretation and prediction of non-point source nitrate distribution, in unconfined aquifers. 2.3.1 Denitrification of non-point source nitrate in unconfined aquifers 2.3.1.1 Controls on Denitrification That a microbially mediated denitrification potential exists in aquifer environments has been well established by many investigators, and a compilation of these findings is contained in Korom (1992). However, the widespread contamination of groundwaters with nitrate, and its persistence over time (as suggested by the previously described findings) shows that any natural denitrification capacity that may be present in subsurface regions is far exceeded by the amounts of nitrate entering this environment. But the previously described findings and interpretation of steep contamination gradients also suggest that, under certain conditions, denitrification can control the distribution of 40 nitrate below the water table of unconfined aquifers. In order to fully appreciate the significance of these reported findings, it is necessary to understand the denitrification process as it occurs within groundwater environments. Microbes2 in aquifers obtain energy from the mediation of redox (reduction-oxidation) processes in which the oxidation (electron release) of either organic or inorganic compounds is coupled to the reduction of several possible electron acceptors (oxidants). These microbes are therefore chemotrophs (as opposed to phototrophs which obtain energy from the sun). If the electron donor is organic, the microbes are organotrophic, and if the electron donor is inorganic the microbes are lithotrophic. Organotrophs use the organic energy source also as a source for cellular carbon, and are therefore also heterotrophic. Similarly, it has been found that most lithotrophs obtain cellular carbon from (inorganic) carbon dioxide (COj), and thus they are also autotrophic. It is noted that the organotrophic-lithotrophic classification pair is synonymous to heterotrophic-autotrophic, and the latter two are the more commonly used terms (Tortora et al . , 1989). For the oxidation of organic carbon, the preferred electron acceptor (from the viewpoint of energy utilizable from the electron transfer process) is molecular oxygen (0 2), and microbes capable of utilizing oxygen are aerobes. In the absence, or on consumption, of oxygen, the next preferred electron acceptor is nitrate (N03\"). Microbes capable of switching to alternate electron acceptors in the absence, or reduction, of oxygen are facultative anaerobes (microbes unable to make this switch are referred to as obligate aerobes or anaerobes). 2 This summary of the terminology associated with microbial classification has been taken from Korom (1992). 41 Denitrification is the reduction of N0 3 \" to N 2 , by the following electron accepting pathway (Korom, 1992): NO; -* N02~ -* NO -* N20 -+ N2 ( 2 \u00C2\u00AB 5 ) Many microbial species are capable of mediating N0 3 \" reduction all the way to N 2 , however several species are only capable of reducing NCy to N 0 2 (nitrite), or N0 2~ to N 2 . Different enzymes are responsible for these two reduction steps, and nitrite accumulation as a result of a difference in the enzymatic reduction rates can also occur (Robertson & Kuenen, 1992; Wilderer et a l . , 1987). Korom (1992) showed in his compilation that the rates of denitrification on aquifer soils, as measured in the laboratory using slurries or core, or in-situ, vary over a wide range of values, and can also be site specific. He noted also that even though laboratory assays may show the presence and activity of denitrifiers on aquifer solids, the occurrence of denitrification in-situ should not be necessarily presumed, given the typically different environmental conditions imposed in the laboratory. Research performed by Priscu & Downes (1987), using lake sediments, showed denitrification rates measured using intact sediments to be less than 1 % of (or two orders of magnitude less than) values obtained from continuously agitated sediment slurries. Such a comparison cannot be made for aquifers using the data compiled by Korom (1992), as both laboratory and field estimates of denitrification at a single site were not obtained. Most microbes capable of denitrification have been identified to be heterotrophic facultative anaerobes, so organic carbon is the preferred electron donor. Though at least one microbial strain capable of using N0 3 \" as the electron acceptor in the presence of high oxygen levels has been apparently isolated (in a wastewater treatment system), it is 42 accepted that most bacteria in the natural environment only switch to using N0 3 \" at low, or zero, 0 2 levels (Korom, 1992). Calculation of free energy values for various redox couples (Stumm & Morgan, 1970) show that in an environment where biodegradable organic carbon is plentiful, the sequence of preferred electron acceptors for the oxidation of this carbon is 02 > NO; > Mnr > Fer > SOf G - 6 ) The consideration of redox processes in such a sequential manner is particularly appropriate for groundwater flow systems, and where recharge water migration below the water table can be associated with the movement of a front as the younger recharge waters displace older groundwaters. Such an interpretation of redox processes in groundwater flow systems was first proposed by Champ et al. (1979), who discussed groundwater flow systems as being either open oxidant or closed oxidant. A closed oxidant system is one in which recharge waters containing an excess of reduced organic carbon, after initial entry into the system, are closed to further input of oxidants, or oxidized species. An open oxidant system is one in which excess dissolved oxygen is present, and which may react with reduced species such as carbon, HS\", NIL/, etc. Heterotrophic Denitrification Heterotrophic denitrification involves the oxidation of organic carbon coupled to the reduction of nitrate. This process can be represented as follows (Champ et a l . , 1979): Ci7,0 + -NO; + -H* - CO, + \u00E2\u0080\u0094N~ + -H0O (2.7) 2 5 3 5 2 5 2 5 2 With regard to the potential for heterotrophic denitrification in groundwaters, 43 Starr & Gillham (1993) concluded from their investigations in Ontario, Canada, that \"under shallow water table conditions there is sufficient labile (biodegradable) organic carbon present for anaerobic conditions to be generated and denitrification to proceed. Under deep water table conditions, there is insufficient labile organic carbon to provide a substrate for dissolved oxygen and nitrate reduction, so denitrification does not occur.\" Starr & Gillham presumed that the source of the organic carbon is the upper soil zone, and speculated on the potential for increasing dissolved organic carbon (DOC) in recharge waters through crop residue management, and as a means of promoting subsequent denitrification below the water table. The generality of this conclusion by Starr & Gillham (1993) is somewhat surprising, as there are many findings which suggest that the availability of labile organic carbon under shallow water table conditions cannot be universally assumed. Beauchamp & Voroney (1994) and Eghball et al. (1994) show that the amount of carbon originating from agricultural lands can be variable, and dependent on such factors as crop and tillage sequences. Earlier work by Paul & Beauchamp (1989a,b) and Paul et al. (1989) showed the complexity and variability associated with dissolved carbon generation in the upper soil zone. These investigators also showed the significance of low molecular weight carbon compound formation (by fermentation of organic matter), and its subsequent preferential utilization during denitrification. Kerner (1993) found that the extent of such fermentation activity, and subsequent low molecular weight carbon generation, can also be seasonally variable due to temperature and moisture content changes. Dunnivant et al. (1992) investigated the transport of dissolved organic carbon 44 through columns of aquifer soils, and observed that a significant portion of the carbon is adsorbed onto the soil surfaces. Similar observations were made by Ronen et al. (1987a). They obtained high estimated values of leached organic carbon content within the deep (27 m) unsaturated zone of their study site in Israel, but found low levels at and beneath the water table. An early study by Leenheer et al. (1974) based on sampling of 100 wells and springs in 27 states in U.S.A., found median dissolved organic carbon (DOC) concentrations for all aquifer types to range from 0.5 mg/l to 0.7 mg/l. But Malcolm (1993) reported that shallow unconfined groundwaters can have higher DOC concentrations. Krajenbrink et al. (1989) found, during their investigations under pasture land, DOC levels of nearly 60 mg/l within the upper 2 m of the water table in a shallow (1-2 m below ground surface) unconfined aquifer. Krajnebrink et al. postulated that \"the DOC leaching is mainly determined by the average organic matter content of the unsaturated zone, while the final DOC flux to the groundwater is controlled by the average depth of the groundwater table.\" Furthermore, they found a temporal variation in the DOC content at and below the water table. Perhaps most importantly, they considered that the DOC reaching the water table to represent the lower degradable fraction of the organic matter originating from the land surface and unsaturated zone. On this basis, Krajenbrink et al. speculated that DOC degradability may be a rate-limiting factor for the consumption (reduction) of oxygen, and presumably therefore any other heterotrophic activity such as denitrification. Other findings suggesting the low biodegradability of DOC in groundwaters were provided by Gron et al. (1992), Bradford et al. (1994) and Wassenaar et al. (1991), who all found that the majority of groundwater DOC consists of high molecular weight 45 compounds. The availability of degradable carbon alone may not be the only factor controlling the potential for denitrification in subsurface conditions. Hickman et al. (1989), demonstrated through a study comparing the biodegradation of organic contaminants between different subsurface soil conditions, that sites could be characterised as either \"fast\" or \"slow\" in relation to their biodegradation rates. \"Fast\" sites were characterised by a historically \"high flux of water, organic matter and nutrients, and a microbial community with a variety of potentially active metabolic pathways\". Biodegradation rates at such sites could be stimulated by the addition of nitrate as an electron acceptor. \"Slow\" sites were characterised by a historically \"low flux of water and nutrients\", and the addition of nitrate was found to decrease degradation rates. Hickman et al. (1989) recognised the significance of their findings towards interpreting the potential reversed role of organic matter addition for nitrate reduction. \"Slow\" sites would be expected to be less responsive to organic matter addition. Though DOC in groundwater, and in particular shallow unconfined aquifers, may originate from surface sources, subsurface solid organic carbon (SOC) such as coal fragments or buried peat deposits, or even methanogenesis, can contribute to a DOC content (Murphy et al., 1989; Aravena & Wassenaar, 1993). Bradley at al. (1992) and Starr & Gillham (1993) also demonstrated using laboratory and field bioassays that heterotrophic denitrification can occur in the presence of sufficiently high (> 0.5% by weight) SOC levels, and that denitrification rates may be directly proportional to the SOC content. In general, however, the available evidence suggests that the combination of low 46 natural levels, and the low biodegradability of residual organic carbon (SOC and DOC) in aquifers limits the extent of natural heterotrophic denitrification of nitrate contaminated waters. The suggestions (Starr & Gillham, 1993; Lowrance & Pionke, 1989; and others) that controlled delivery of labile carbon below the water table could be used to promote denitrification towards limiting the extent of nitrate contamination wil l be discussed further in Section 2.4 Autotrophic Denitrification Robertson & Kuenen (1990), in their review of denitrification by obligate and facultative autotrophs, noted that of the many inorganic compounds (eg. HS\", M n 2 + , F e 2 + , NFL,\"1\") which support aerobic respiration, only hydrogen3 (HJ and reduced sulfur compounds have been generally found to support autotrophic denitrification. On this basis, it is not surprising that the previously reported (Section 2.2) inferences of autotrophic denitrification in aquifers have all been associated with oxidation of pyrite, a reduced sulfur compound. While perhaps currently only a subject of curiosity, Robertson & Kuenen (1990) also reported their own laboratory based findings that there is at least one microbial species (Thiosphaera pantotropha) capable of oxidising ammonia (NH 4 + ) while denitrifying. The occurrence of hydrogen oxidation coupled to nitrate reduction in groundwater environments has so far not been encountered, though Smith et al. (1994) reported that 3 Robertson & Kuenen (1990) described aerobic respiration and denitrification with H 2 as: 5 H 2 + 2 H + + 2N0 3 \" => N 2 + 6 H 2 0 (denitrification) 2 H 2 + 0 2 => 2 H 2 0 (aerobic respiration) 47 microbes capable of mediating this process are probably common in aquifers. The sparingly soluble nature of H 2 (=1 mg/l) in water may in part explain the limited occurrence of this process in-situ, but it has been utilized for reactor based treatment of drinking waters (Gayle et a l . , 1989). Though sequential redox processes appear to be predictable using free energy criteria (Champ et a l . , 1979), available evidence from groundwater environments suggests that energetic favourability should not be considered the sole indicator of reaction potential. Frind et al. (1990) suggested on the basis of their findings in the Fuhrberg aquifer (Figure 2.11) that autotrophic denitrification (with pyrite oxidation) was preferred over heterotrophic denitrification, despite the presence of organic carbon, due to kinetic favourability. This interpretation was subsequently challenged in a discussion between Korom (1991) and Boettcher et al. (1991), but the issue was left unresolved. On a related note, Leuchs & Obermann (1992) postulated that the presence of mixed microbial populations could possibly result in reactions also proceeding in response to species specific preferences. This, combined with the previously described findings related to organic carbon biodegradability, clearly point to the possibility that, in aquifers measured to contain organic carbon, autotrophic denitrification could be preferred over heterotrophic denitrification if the organic carbon has low biodegradability. 2.3.1.2 In-situ Indicators of Denitrification Sequential redox processes can be elucidated for a multiple species system from monitoring spatial changes in reactants and reaction products. This approach also permits some level of reaction quantification on the basis of product-formation and reactant-48 disappearance rates. For heterotrophic denitrification, the redox equation presented in equation (2.7) shows that a decrease in N0 3\" should be accompanied by a decrease in organic carbon (SOC & DOC), and related increases in pH, inorganic carbon (COj subsequently dissolving to HC03\"), and N 2 . The profiles obtained from the Rodney field site by Starr & Gillham (1993), and as shown on Figure 2.14, have been interpreted in this manner to suggest heterotrophic denitrification. On the basis of an increase in pH, HC0 3\" and the occurrence of N0 2\", the profiles shown on Figure 2.15 were interpreted by Appelo & Postma (1993) to also conclude that heterotrophic denitrification affected the observed decrease in nitrate oyer depth. Similarly, several of the other profiles described in Section 2.2 have been interpreted, on the basis of an increase in Fe 2 + and S042~, and a decrease in pH, to conclude the occurrence of autotrophic denitrification by the oxidation of pyrite. An alternative reactant-product approach to inferring the occurrence of denitrification is to correlate any reductions in N0 3\" to the isotopic composition of the 1 5 N stable isotope of N . It has been found (Marriotti et al., 1988, and references contained therein) that denitrifyers preferentially reduce N0 3\" molecules containing the lighter 1 4 N isotope, resulting in the remaining N0 3\" being \"enriched\" in 1 5 N , and the produced N 2 being enriched in 1 4 N . N isotope composition is usually calculated as: l\"15iV /14N\ - \1SN /UN\ d15N = - sampU sumdard x 100 (2.8) 1 * standard [15N /WN]S A positive and increasing value for 5 1 5N indicates enrichment in 1 5 N (Focht, 1973). A plot of the isotopic composition of the nitrate, versus the nitrate concentrations, as measured over a region of concentration decrease can therefore identify whether the 49 decrease may be due to dilution or due to denitrification. Figure 2.16 shows such an interpretation as described by Mariotti et al. (1988), at a site where a decrease in groundwater N0 3 \" concentrations along the flow path is associated with an increase in 5 1 5 N (as well as a decrease in 0 2 ) . However, this figure also shows the unexpected trend of an increase in the organic C content along the flow path in the region where nitrate reduction has been observed. So, while denitrification has been concluded on the basis of the 5 1 5 N data, its nature (autotrophic or heterotrophic) has not been defined. A simpler approach to determining whether denitrification may be occurring is to calculate the ratio of nitrate to some other less reactive constituent known to have originated with the nitrate. Chloride can be used for this purpose, as it can be commonly associated with anthropogenic pollution (eg. Figures 2.8 and 2.9). A plot of N03\"/C1\" versus distance along the groundwater flow path can provide an indication of denitrification if the ratio decreases over distance. However, such a decrease could also reflect historic changes in the N0 37C1\" ratio within the surficial nitrate source. Hem (1989) noted that dissolution of calcite and gypsum can also release trapped CI\" and so contribute to its increase in the groundwater dissolved chemistry. The measurement of E,, in a redox system enables qualitative evaluation of whether the system is oxidising (as indicated by positive E J or reducing (as indicated by a negative E J , and has been applied by several researchers for assessing redox condition changes along regional groundwater flow systems (Edmunds et a l . , 1984, and references contained therein). However, the use of Eh measurement for quantification of specific redox reactions in groundwater environments is fraught with uncertainty, primarily as it has been found that these reactions are generally not at equilibrium, and also as the 50 Figure 2.16 Hydrochemical and isotopic variations along ' groundwater flow path in a chalk aquifer (unit 4) suggesting the occurrence of denitrification (Mariotti et a l . , 1988) 51 measured potentials may actually represent mixed potentials (Lindberg & Runnells, 1984). Other confounding factors include the heterogeneity of aquifer properties (including microbial population distributions and diversity of genera as well as metabolism rates), and the complexity of transport and dispersion processes at the field scale (Barcelona et a l . , 1989; Hirsch, 1988). The inability of the electrodes to respond to geochemically significant redox couples has also been cited as a limitation against quantitative F4 interpretation in natural waters (Lovley et a l . , 1994). However, as can be noted from the profiles shown on Figures 2.11, and the variation along the flow path on Figure 2.16, En measurements can provide another indicator of the stratification between nitrate contaminated waters and underlying nitrate free waters. The use of E\u00E2\u0080\u009E was adopted by Spalding & Parrot (1994) to characterize the distribution and fate of groundwater nitrate in an agricultural area in Nebraska, USA. Figure 2.17 shows the areal distribution of groundwater nitrate levels as measured in domestic and irrigation wells, as well as the inferred location of a redox-front. Figure 2.18 shows three longitudinal profiles of several parameters (inluding 51 5N) developed and used by Spalding & Parrot (1994) to conclude denitrification. While these investigators inferred that denitrification was responsible for the observed areal distribution of groundwater nitrate, they were not able define the conditions responsible for this denitrification. No details on the hydrogeology associated with the Figure 2.17 map area were provided by Spalding & Parrott (1994), so it is not possible to evaluate their observed spatial distributions on this basis. Using their findings, Spalding & Parrott (1994) estimated that denitrification was responsible for the removal of 46 metric tons N/year Figure 2.17 Areal distribution of groundwater nitrate levels, and delineation of redox zones in a shallow aquifer underlying agricultural lands in Nebraska, USA (Spalding & Parrott, 1994) ro 53 Figure 2.18 Longitudinal profiles of hydrochemistry in vicinity of redox front shown on Figure 2.17, used to infer denitrification (Spalding & Parrott, 1994) 54 from groundwater flowing across the 15 km long redox front at their site. 2.3.2 Transport of non-point source nitrate in unconfined aquifers Figure 2.19 presents a generalised representation of a homogeneous isotropic unconfined aquifer, and the groundwater trajectory below the water table. It is readily apparent from the superposition of a surficial non-point source of leaching nitrate that subsequent contamination of the aquifer will commence, and over time extend below, the water table. Numerical models which describe the transport processes beneath the water table offer a means by which observed nitrate (and associated constituent) distributions may be interpreted, and also a means by which potential impacts of land use changes may be predicted. Using the nomenclature shown on Figure 2.19 and assuming an aquifer having a uniform thickness D, porosity e, and receiving uniform recharge P, Appelo & Postma (1993) showed4 that the horizontal velocity, v, at a distance x from the divide is given by v = i * (2.9) De and that the horizontal distance travelled from point xa over time t is 4 this analysis originated from Hoeks (1981) 5 5 Figure 2.19 Simplified profile of groundwater flow trajectory in an unconfined aquifer receiving uniform recharge. Right boundary is groundwater divide, and lower boundary represents an impermeable base (from Appelo & Postma, 1993) 56 x = x e x p \u00E2\u0080\u0094 . (2.10) Furthermore, by considering the following proportionality relationship5 xo D-d Appelo & Postma noted that d = D(l - e x p [ \u00E2\u0080\u0094 ] ) (2.12) De so that a depth versus age relationship can be determined. Equation (2.12) can be rearranged to give: = -e\n(\-dlD) ( 2 < 1 3 ) PID With regard to groundwater flow beneath the water table of unconfined aquifers, Appelo & Postma (1993) concluded the following on the basis of the simplified analysis described by equations 2.9 to 2.13: i) horizontal groundwater velocity at any given location from the divide is constant over depth, but increases with increasing distance from the divide, and ii) the age of the groundwater is dependent only on the depth below the water table, and independent of the location from the divide. That is, water resides in the aquifer in planes of equal age. The latter finding suggests that a single vertical profile of groundwater chemistry could 5 derived by considering Px=vDe and P(x-xJ=vde 57 provide sufficient information to assess non-point contaminant source history. Geologic heterogeneity, anisotropy, stratification and the presence of confining lenses or permeable pathways can all result in an increased complexity of the simplified flow system depicted on Figure 2.19 (Freeze & Witherspoon, 1967). But the profiles obtained by Postma et al. (1991), and presented in Figure 2.12, show that the nitrate front (or redox-line) at their site is at a relatively constant depth in the direction of groundwater flow. Reilly et al. (1994) also presented some findings which demonstrate the existence of solutes in planes of equal age. Figure 2.20 shows measured and simulated distributions of tritium as obtained by these investigators from an unconfined \"permeable sand and gravel\" (estimated anisotropy of J ^ r K y of 5) aquifer in the Delmarva Peninsula, USA. The tritium originated from atmospheric releases during nuclear weapons testing performed during the 1950s to 1960s. Precipitation waters at the study site were found to contain three tritium concentration peaks between 1950 and 1970, with a dominant peak occurring around 1965. Using this peak as a single event non-point slug (or front) arriving at the water table, Reilly et al. found that it remained as a narrow plane (at a constant depth below the water table) of solute during downward movement below the water table. On the basis of simulation performed by assuming decay, but no dispersion (Figure 2.20A), these investigators were able to reproduce the sharp concentration gradients implied by the measured tritium data. Though the simplified analysis described by equations 2.9 to 2.13 considers groundwater flow only (and not transport fate processes such as dispersion and denitrification), it is useful for a first estimate of the age distribution of measured vertical 0 300 METERS 0 1000 FEET VERTICAL EXAGGERATION IS 50.0 EXPLANATION 16.5 Location of well, tritium concentration * measured Nov. 1990. Contour of simulated tritium S distribution for 1990. contour interval 25 Tritium Units (TU). \u00E2\u0080\u00A2Area above the water table or below the lower confining unit Simulated tritium distribution at the end of 1990: (a) with dispersivity aL = 0.0 m (0.0 feet) and aT = 0.0 m (0.0 feet), and (b) with dispersivity aL = 0.15 m (0.5 feet) and aT = 0.015 m (0.05 feet). Contour interval 25 tritium units (TU). Measured concentrations from samples obtained from wells in November 1990 are given for their location in bold italics. Figure 2.20 Measured and simulated tritium distributions beneath the water table of an unconfined aquifer in Delmarva Peninsula, U S A (Reilly et a l . , 1994) 59 profiles of comparatively non-reactive contaminants such as nitrate, particularly where vertical dispersion can be assumed to be negligible6. Figure 2.21 presents a plot of t versus d/D computed using equation (2.13) for various values of P/D. A l l plots were computed by assuming a value of 0.35 for porosity e. These plots show that the \"age\" of the groundwater is sensitive to the ratio of annual recharge (P) to the total aquifer depth (D). As expected, for a given depth d below the water table, the travel time (age) t increases as P/D decreases. One striking feature is the sensitivity of t to the d/D ratio at values between 0 (water table) and 0.05. This highlights the significance of sub-annual distributions of the annual recharge and the subsequent temporal variations on the distribution of leached contaminants below the water table. This aspect will be discussed further in Chapter 7, in relation to the findings of the current research. Numerical flow and transport models developed specifically for the evaluation of observed groundwater non-point source nitrate distributions have been described by Frind et a l . , (1990), Kinzelbach et a l . , (1991), Postma et a l . , (1990), and Widdowson (1989). Dillon (1988, 1989) described a model of non-point source contaminant transport beneath the water table of unconfined aquifers, but did not verify it with observed nitrate profiles. Engesgaard & Kipp (1992) described a geochemical transport model for redox controlled movement of mineral fronts in groundwater flow systems, and verified it using the findings on nitrate removal by oxidation of pyrite as described by Postma et a l . , (1991). Bouwer & Cobb (1986), Molz et a l . , (1986) and Widdowson et a l . , (1988) 6 this usefulness is demonstrated by performing such analysis using the site data of Postma et al. (1991) and Reilly et a l . , (1994) and comparing the computed ages against their CFC and tritium age estimates. This comparison is presented in Appendix A 60 t (years ) Figure 2.21 t vs d/D for various values of P/D calculated using equation 2.13 61 presented mathematical descriptions of competitive and sequential heterotrophic redox processes, including oxygen and nitrate respiration. These investigators also describe a number of parametric variations - including initial microbial population, substrate (carbon) availability, and reaction rates; and used these to evaluate the advective-dispersive behaviour in theoretical one-dimensional flow through columns of porous media. Their findings, and the extension by Widdowson (1989), show that the developed formulations are able to simulate the sharp concentration gradients observed in-situ. Of the developed flow and transport models, only those of Dillon (1989) and Frind et al. (1990) are capable of describing the vertical distribution in two dimensions, and thus provide a means by which spatial variability in the source of nitrate as well as recharge may be incorporated into the analyses. Both of these models use the advection-dispersion equation as the basis for their model formulations. The model described by Postma et al. (1991) is based on a time-stepped \"mixing-cell\" approach, that is further described in Appelo & Postma (1993). Using this approach, Postma et al. (1991) modelled the vertical column beneath the water table and used the vertical component of the groundwater flow velocity in their time stepping computations. Both Dillon (1989) and Frind et al. (1990) describe denitrification as an exponential decay process (c=c0e~Xt). The advection-dispersion equation incorporating this decay, as well as any retardation, is then written as Z>,*!E + DJ^ - v^\u00C2\u00A3 - R\c = (2.14) dx2 dx dt and where DL is the longitudinal coefficient of dispersion; DT is the transverse coefficient of dispersion; c is solute concentration; R is the retardation factor; and X is the 62 exponential decay constant. For nitrate, Frind et al. (1990) consider R to equal 1.0 (ie. not retarded) Dillon (1989) and Frind et al. (1990) differ in the method which they each adopted for the solution of the advection-dispersion transport equation. The DIV AST (acronym for Diffuse Source Vertical Analytical Solute Transport) model developed by Dillon (1989) contains analytical solutions for longitudinal transport along the groundwater flow lines, with an imposed transverse dispersion perpendicular to the flow. Figure 2.22 shows the general flow field used in the analyses developed by Dillon (1989). Frind et al. (1990) describe a solution method consisting of a finite element based approach in which a 2D \"steady state flow field\" is assumed because the \"time scale of the transport problem is of the order of years or decades\". Figure 2.23 shows the site specific (to the Fuhrberg aquifer) flow field developed by Frind et al. (1990). The modelling approaches adopted by Dillon (1989) and Frind et al. (1990) assume that the water table boundary is fixed in space, and that the flow system is constant, with only the recharge nitrate concentrations varying over space and time. Such an approach ignores the dynamics associated with large seasonal water table fluctuations, and any subsequent impacts that this may have on the flow trajectory immediately beneath the water table, an aspect implied by the previously described interpretation of Figure 2.21. Figures 2.24 to 2.28, compiled from Dillon (1988) and Frind et a l . , (1990) demonstrate that, during the interpretation of observed vertical profiles of non-point source contamination beneath the water table of unconfined aquifers, and where 63 Z \u00E2\u0080\u00A2 Z = D Z = 0 X = 0 Figure 2.22 Aquifer cross-section simulated by DIVAST. Groundwater divide at x=0, water table at z=D. Constant saturated thickness and recharge. Allowance for leakage, q 0 (Dillon, 1989) 64 Figure 2.23 Cross sectional flow net and water table recharge function used by Frind et al. (1990) to model the data shown on Figure 2.11 65 160 r CONCENTRATION Figure 2.24 Concentration profile beneath water table receiving unit concentration of leached contaminant as computed for various values of vertical dispersion (Dillon, 1988) 0 L J 1 1 1 ' ' \u00E2\u0080\u0094 2 0 0 0 2 1 0 0 2 2 0 0 2 3 0 0 2 4 0 0 2 5 0 0 Date Figure 2.25 Theoretical 500 year concentration profile history at a well location assuming vertical dispersion = 0 m2/yr and unit recharge concentration a) eliminated in 1986; b) sustained (Dillon, 1988) 67 (3) G r o u n d surface 2000 2100 2200 2300 2400 2500 Date Figure 2.26 Theoretical 500 year concentration profile history at a well location assuming vertical dispersion = 1 m2/yr and unit recharge concentration a) eliminated in 1986; b) sustained (Dillon, 1988) Figure 2.27 Calculated boundaries between denitrification and desulfurication zones using transverse dispersivity of 1) Om, 2) 0.001m, 3) 0.01m to simulate findings shown on Figure 2.11 (Frind et al., 1990) 69 Figure 2.28 Calculated boundaries between denitrification and desulfurication zones using nitrate half life of 1) 1 year, 2) 1.5 years, 3) 2.3 years to simulate findings shown on Figure 2.11 (Frind et al., 1990) 70 groundwater flow is predominantly horizontal: 1) observed steep concentration gradients can be explained simply by presuming a zero (or very small) value for transverse (vertical) dispersion, and not necessarily by the occurrence of denitrification7 (Figures 2.24, 2.25, 2.27 and 2.28); 2) under continued leaching without source reduction, the nitrate front (or \"redox-line\") will gradually move deeper below the water table in the absence of any denitrification, but that this rate of movement may be quite slow and imperceptible over years, and possibly even decades, of monitoring (Figures 2.25 and 2.26). Both of these conclusions have implications for the interpretation of measured vertical profiles. In particular, the several previously described findings of a steep nitrate concentration gradient between the upper nitrate-contaminated, and the lower nitrate-free zones can be explained to have resulted due to horizontal flow combined with zero transverse dispersion. On this basis, the reduced species encountered in the groundwater beneath the nitrate front may simply have evolved during the longer residence times associated with these older, and often anaerobic, waters. It is of some significance that none of the vertical distributions, from which autotrophic denitrification (by pyrite oxidation) was concluded, contain measurements of nitrogen products (N 2 , or 51 5N) which could confirm the inferred concentration of denitrifying activity at the redox-line. 7 as also inferred from the work of Reilly et al. (1994) who used zero dispersion to simulate the observed steep tritium concentration gradients as shown on Figure 2.20. 71 2.4 C O N T R O L O F G R O U N D W A T E R N I T R A T E Francis (1992) concluded his study on fertilizer nitrogen control policy mechanisms by stating that \"nitrate leaching losses cannot be brought to zero with current technology, however the use of BMPs (Best Management Practises) can ensure the lowest possible nitrate loss\", and further noted that the treatment of groundwater for nitrate removal may become a necessary component of this source of drinking water. The anthropogenic control of groundwater nitrate may be considered to consist of three separate components: o control of nitrate at its source, o control of nitrate within the groundwater flow system, and o control of nitrate in extracted groundwater. The approaches and limitations associated with the control of nitrate at its source have already been briefly discussed in Chapter 1. The control of nitrate in extracted groundwater is a treatment application for which many exsitu approaches have been engineered, including denitrification and ion-exchange processes (Clifford & L iu , 1993; Gayle et a l . , 1989). There have also been several systems developed in Europe and which utilize a portion of the aquifer formation as an in-situ fixed growth bio-reactor, through which carbon amended groundwater is pumped to remove nitrate by denitrification. A review of these systems is contained in Dahab (1993) and Hiscock et al. (1991). While nitrate reduction has been shown to be achievable, the efficiency of these in-situ systems reportedly is significantly limited by a tendency for formation clogging as a result of biogrowth and nitrogen gas (N2) bubble accumulations. The long term effectiveness of these systems appears to be limited on this basis. 72 As previously noted, it has been proposed by others that carbon additions to groundwater environments could offer a means by which the natural denitrification potential, where present, can be enhanced to remove nitrate during groundwater flow. Adelman & Spalding (1988) and Adelman & Dahab (1991) reported on some early theoretical investigations on the possibility of injecting ethanol into aquifers to enhance denitrification. Childs et a l . , (1988) developed a numerical model for carbon substrate injection to enhance denitrification, as did Duthy (1993). Rott & Lambeth (1992) presented the findings of their conceptual investigation into \"subterranean denitrification\". There has been no extension of these various findings to the field scale. The feasibility of such an approach therefore remains uncertain, in particular as all of these evaluations have also been based on an optimistic assumption of rapid mixing of the introduced carbon into the receiving groundwaters. Weier et al. (1994) investigated the potential for ethanol addition to irrigation waters at the end of crop harvest to promote denitrification of the residual soil nitrate. Though laboratory column testing showed that denitrification could be promoted in this manner, the feasibility of extending this approach to the field scale was not evaluated, and therefore also remains unknown. Of the various schemes suggested to promote denitrification at or below the water table, the previously mentioned proposal by Starr & Gillham (1993), of crop residue management aimed at increasing the amount of biodegradable carbon leaching below the root zone, may represent a more feasible alternative, as it could be incorporated into seasonal farming cycles, with no additional requirements such as injection wells and associated capital costs. However, as discussed in Section 2.3, the retardation of carbon 73 through the unsaturated zone, may require that large quantities of residual organic matter be provided to ensure labile carbon delivery to the water table. Much work is required before the feasibility of this approach can be fully evaluated. Two other approaches which offer a means for controlling groundwater nitrate levels at an aquifer-wide scale, are the reliance on natural or formed riparian zones (near discharge areas, or in areas of shallow water table - see also Figure 2.3); and interceptor ditches for capturing shallow nitrate contaminated groundwaters and diversion to natural or engineered wetlands where conditions for denitrification and nitrate uptake could be provided (Olson, 1992). These two approaches also offer a benign long-term means by which groundwater nitrate levels could be controlled. The feasibility of interceptor ditches for the capture of stratified contamination from shallow groundwaters was demonstrated by Zheng et a l . , (1988) and Chambers & Bahr (1992) in theoretical and field tracer studies, but there appears to have been no application of this approach for channelling away stratified groundwater nitrate. Groundwater nitrate reduction on flow beneath riparian zones can occur as a result of direct uptake by vegetation and due to organic carbon leaching from the surface and subsequent denitrification. Research conducted to date has shown that the extent of nitrate reduction beneath riparian zones can be variable and depend on such factors as soil type, vegetation and hydrology (Groffman et a l . , 1992; Haycock & Pinay, 1993; Licht & Schnoor, 1993; Lowrance, 1992; Simmons et a l . , 1992; Schipper et a l . , 1991). Though still in its infancy, the growing body of data on riparian zone nitrate removal can be expected to provide means by which riparian zones may be engineered for this purpose. 74 2.5 CONCLUSIONS Detailed vertical profiling of non-point source nitrate and associated chemistry below the water table of unconfined aquifers is essential towards understanding the origin and fate of this nitrate within the aquifer environment. A proper understanding of the associated groundwater flow regime is also required. Consideration of the seasonal recharge cycle, associated with annual precipitation patterns, shows that arrivals of nitrate to the water table wil l be in an annual pulse form in areas where the travel time through the unsaturated zone is short. This would be expected to result in a temporal variation in the groundwater chemistry profile. However, it has also been shown (equation 2.13) that subsequent travel times beneath the water table are highly dependent upon the surface infiltrating recharge amount and the thickness of the aquifer. The majority of the reviewed findings suggest that natural denitrification, and its impact on non-point source nitrate concentrations, within aquifers may to be concentrated within a narrow redox-front between oxic recharge waters and underlying (or downgradient) anoxic groundwaters. It has been inferred that denitrification under these conditions results in the development of steep chemical gradients at the redox-front. Encountered occurrences of natural denitrification beneath agricultural lands, and the subsequent control on leached nitrate levels, have been primarily associated with autotrophic denitrification by pyrite oxidation. Low natural levels of labile (biodegradable) carbon are considered to limit the extent of natural heterotrophic denitrification activity within groundwater environments. The addition of labile carbon into aquifers, to promote heterotrophic 75 denitrification for aquifer wide control of groundwater nitrate levels, has been investigated at a conceptual level. But how and where this carbon should be added to ensure effective nitrate control is currently not clear. While the potential for aquifer clogging as a result of N 2 gas bubble formation and accumulation has been identified, this phenomenon has not been investigated in detail to determine how it may be minimized during aquifer remediation. 3.0 RESEACH METHODOLOGY 76 This Chapter describes the methodologies adopted for both the field and laboratory based investigations, and is organised as follows: Section 3.1 describes the overall investigation objectives, materials and procedures associated with the field investigations; Section 3.2 similarly describes the laboratory investigations; and Section 3.3 contains a description of the analytical techniques used to quantify the chemical characteristics of the water and soil samples collected during the both the field and laboratory investigations. 3.1 FIELD INVESTIGATIONS The overall approach taken for the field investigations was to obtain detailed measurements of groundwater chemistry profiles at monthly intervals in order to determine the temporal variation in non-point source nitrate, and associated chemistry, loading to the water table of two unconfined aquifers. 3.1.1 Monitoring Site Locations Three locations were selected for the installation of monitoring wells. Two wells were installed in the Brookswood Aquifer, and one well installed in the Abbotsford Aquifer. Both of these aquifers are located within the Lower Fraser Valley region in southwestern British Columbia. Figure 3.1 shows the locations of these aquifers. Both of these aquifers are known to be contaminated with nitrate (Carmichael & Wei, 1994). Detailed site maps, including well locations, are presented in Figures 3.2 and 3.3. Figure 3.1 Locations of Brookswood and Abbotsford aquifers within the Lower Fraser Valley region in southwestern British Columbia (Carmichael & Wei, 1994) 78 N S s ^ ^ groundwater divide 0 BCMOEWefl Groundwater flow direction Abbots ford A1rport \u00E2\u0080\u00A2a s 0 A Mt Lehman / / B o 0 C s E 0 Avenue \u00E2\u0080\u00A2 Huntingdon Read \ Canada \"\"U'SA\"\" Study Well 5 0 0 1000 m Figure 3.2 Site Map of Abbotsford study well area, showing location of B C M O E wells A, B & C, and groundwater flow directions 79 250 500 750 m N oj pj T3 C Groundwater flow direction O water supply pumping well \u00E2\u0080\u00A2 Residential site with drainfleld . A Manure stockpile location CRTC \ Property ^ A North \u00E2\u0080\u00A2 Study Well South Study Well Forest . Residential land with septic drain fields 36th Avenue Figure 3.3 Site Map of Brookswood study well area, also showing location of pumping well and site of Piteau Associates (1994) investigation. Presumed groundwater flow direction shown. Section AA' shown on Figure 5.1. 80 The selection of the monitoring sites was coordinated as part of a larger, and ongoing, regional subsurface investigation program being performed by the Geological Survey of Canada. As a result of this, drilling of the wells was coordinated by GSC staff, who also collected samples of encountered subsurface soils. A brief description of the site conditions is presented below, while a more detailed discussion is presented in Chapter 7. Kohut (1987) has noted that the region is characterized by \"a cool Mediterranean type of climate in which precipitation falls principally as rainfall during the period September to May. The area receives an average of 1513 mm of precipitation annually. The period from June to September is normally dry and may be subject to drought conditions for short periods\". 3.1.1.1 Abbotsford Study Site The study well was installed close to the southern edge of a raspberry field, and near the foot of a roadway (0 Avenue) shoulder. The distance between the well and the edge of the field is approximately 1 m. The topography over the Figure 3.2 map area is flat to gently undulating. Raspberry production (row crop) presently constitutes the primary land use over much of the area surrounding the study site. Most of the map area on Figure 3.2, apart from the airport footprint, is occupied by raspberry fields. Poultry production is also substantial, and is of the barn type. The land adjacent and upgradient to the study well along O Avenue has been in raspberry production for at least 20 years. Poultry production in the area has also increased during this time (Froese, W. , 1995 - personal 81 discussions with the site farmer) Large quantities of poultry manure are presently generated over the aquifer, and almost all of this manure is applied as fertilizer and soil conditioner over the surficial soils. Exact details on the quantity of applied manure are not known, but Liebscher et al. (1992) estimated annual application rates to be in excess of 200 kg-N/ha. In the vicinity of the study well, inter-row applications of manure in layer thicknesses of about 0.2 m has been observed in the past (Davis, 1995). The amount of manure applied has gradually increased over the past three to four decades, with a significant increase during the 1970s as a result of a concentration of poultry production over the aquifer (Chipperfield, K., 1995 - personal communication). In the vicinity of the study site, it is likely that manure application rates may have doubled during the last two decades to reach current values (Froese, W. , 1995 - personal discussions with the site farmer). Poultry manure applications to cropland are typically performed during the spring months (February - April), though year-round application as well as manure stockpiling (occasionally on bare ground) is not uncommon (Sustainable Poultry Farming Group, 1993). As shown on Figure 3.2, the groundwater flow direction at the study site is south-southwesterly. The groundwater chemistry at the study well is therefore presumed to be affected by the land use to the north. Further discussion of the groundwater flow regime is presented in Chapter 4. 3.1.1.2 Brookswood Study Site The site is flat in the vicinity, and upgradient, of both the study wells. Land use 82 over this aquifer is mixed, and consists of residential (low and high density) areas with septic tank drain fields, and rural farm land. The study site is the location of a Canadian Overseas Telecommunications Corporation (CRTC) radio tower complex. The land surrounding the towers is leased for farming. Crop production at the site has primarily consisted carrots and potatoes for at least the past twenty years (Liebscher, H . , 1995 - personal communication). The South study well is located at the southern edge of the CRTC site, and approximately 1 m inside the site fence line. The North study well is located within a cropped field and approximately 100 m from the edge of a wooded area. Poultry manure is imported onto the site for use as a fertilizer and soil conditioner, and is applied typically between February and Apri l . This poultry manure is stockpiled on a continuous basis at a location near the northern site boundary as shown on Figure 3.3. Additionally, a large poultry manure stockpile was temporarily located in the area between the North study well and the wooded land for a period of possibly six months until Apri l 1994. This stockpile was removed prior to the commencement of monitoring at the monitoring wells. Inorganic nitrogen-based fertilizer is also applied. Records of the applied quantities of manure and fertilizer application at this site are not available. Winter cover of rye is occasionally applied at the end of the growing season, though during the monitoring period much of the land upgradient of the North study well was left uncovered after potato harvest in September 1994. The groundwater flow direction beneath the study site has been previously determined (Piteau Associates, 1994) to be north-northwesterly, as shown on Figure 3.3. 83 This aspect is further discussed in Section 5.1.2. The water supply pumping well located approximately 50 m east of the South well is not considered to significantly impact the groundwater flow regime upgradient of both the South and North study wells at this site. The pumping well typically operates intermittently between the months of March and September each year, and it has been observed that the well typically only pumps for a few minutes any given day (Piteau Associates, 1994). No records of the pumping rates or duration have been maintained. Piteau Associates (1994) report that the actual drawdown resulting from the pumping is negligible at a distance of 30 m upgradient and surrounding the well. It has also been inferred that the majority of the water drawn into the pumping well will originate within a narrow zone upgradient (or immediately to the south - south east side) of the well (Piteau Associates, 1996 - personal communication). On the basis of the preceding considerations, the groundwater chemistry at the South study well is therefore presumed to be affected by the residential land use to the south, while the chemistry at the North study well is presumed to be affected by the overlying cropped land, as well as the upgradient wooded and residential lands. 3.1.2 Materials and Methods 3.1.2.1 Well Installation & Soil Sampling A l l three wells were advanced with a cable-tool drill rig 1 , by which 6 inch (0.15 m)2 steel casing was driven into the ground and the encountered materials \"loosened\" 1 Under the local geological conditions, this is the preferred approach for well installation in the region 2 For the purposes of this thesis all well installation and sampling device dimensions wil l be presented primarily in inches and feet, as these were the units used by the driller 84 using a cutting bit. Cutting retrieval was performed using a wireline connected bailer barrel. Such advancement and cutting retrieval was performed in approximately 5ft (1.5 m) intervals. The recovered cuttings were therefore, at best, a composite sample of subsurface materials encountered over each 5 ft (1.5 m) advance interval. The bailer contents were emptied into a bucket, from which a sample (approximately 2 kg) was taken and stored in marked polyethylene bags. Water was used as the drilling fluid during all stages of drilling and well installation. On reaching the desired depth, and the casing cleaned of all cuttings, well installation was performed by lowering 2 inch (0.05 m) ID P V C pipe into the casing. The pipe was continuously screened (0.01 inch slots at 0.25 inch spacing - also referred to locally as \"10 slot\") over a substantial portion of its length, and the well was completed by backfilling the annulus between the pipe and casing using washed medium silica sand, upon which the casing was removed. Bentonite pellets and cement slurry were used to seal the upper portion (above the screened zone) of the well annulus after casing removal. Figure 3.4 presents a schematic description of the well completion details, as well as the individual completion depths of the three study wells. The completion dimensions are presented in feet as the P V C pipe was obtained in 10 ft (3 m) lengths. At the time of well installation it was decided that the top of the screened portion of each monitoring well would be no shallower than a depth of 15 ft (approx 5 m) below ground surface. This choice was made purely on a judgemental basis, in order to minimize the likelihood of surface recharge preferentially flowing towards the top of the screen. This decision to measure progress, and as the majority of the materials were specified in these units 85 cement (3ft) bentonite (3ft min.) slotted 2\" ID FVC (threaded joints at I Oft interval) washed medium silica sand backfill threaded e n d o p Well Location D L Abbotsford 75ft 60ft (23 m) (18m) Brookswood North 55ft 40ft (I7m) (I2m| Brookswood South 55ft 40ft Figure 3 .4 Schematic description of monitoring well completion, and table of completion depths at study wells 86 is of some significance for the case of the Abbotsford well, as it was subsequently found that the high water table mark at that well was shallower than 15 ft (5 m), resulting in no water chemistry data being collected in the vicinity of the water table during the months associated with a shallow water table. This aspect wil l be discussed further in Chapter 4.0. Additionally, although the Abbotsford well was initially completed to a depth of 75 ft (23 m), the bottom 20 ft (6 m) was lost, presumably due to a breakage in the pipe. 3.1.2.2 Groundwater Sampling The method of sampling used to obtain the profiles of groundwater chemistry was developed as part of this research program, and is based on the dialysis membrane technique originally proposed for groundwaters by Ronen et al. (1986). The principle of operation consists of lowering membrane enclosed deionised water below the water table, upon which diffusion of groundwater dissolved constituents across the membrane occurs. After a period of time, the membrane enclosed water wil l have the same concentrations of dissolved constituents as the surrounding groundwater. Retrieval and subsequent analysis of the membrane enclosed waters provide a measurement of the groundwater chemistry. As no pumping is involved in this \"passive\" sampling technique, highly detailed spatial profiles of groundwater chemistry can be obtained by the appropriate design of the sampling device. Figure 3.5 presents the key features of the sampling device developed for the present research. Three sampling devices were constructed in order to provide each monitoring well with a dedicated sampler. 87 GS W T Stainless steel wire rope 2\" ID Slotted PVC Well (0.01\" slots at 0.25\" spacing) detail 12\" Perforated 1.75\" ID PVC (with snap-fit PVC endcaps) o o o o o o o o o o o o 2\" diam. NEOPRENE spacer 7 Figure 3 . 5 Details of monitoring well passive sampling device 88 The sampling device is lowered into the screened well, and consists of a series of perforated P V C canisters that are strung on to stainless steel wire rope. Neoprene discs are used as spacers between each canister and prevent the movement of dissolved constituents up and down the well. Each canister is 12 inches (0.3 m) in length, and is perforated by several drilled holes, 0.5 inch (0.013m) in diameter. The entire sampling device was constructed as a single unit by threading 1/8 inch (0.003m) wire rope through a slightly larger central hole drilled in each endcap. A copper nicopress sleeve was crimped onto the wire rope at intervals of five canisters, in order to prevent the vertical movement of the canisters on initial lowering below the water table. Each sampling event was performed by placing two dialysis bags (\"sausages\" of dialysis membrane tubing filled with de-ionised water) into each canister and lowering the entire device into the well. The dialysis \"sausages\" were prepared in the laboratory by knotting lengths of dialysis membrane tubing (Spectra/Por 1: 32 mm flat width, regenerated cellulose with M W C O 6-8,000) filled with de-ionised water. These bags were stored and transported to the site in a cooler filled with de-ionised water. Canister filling was performed at the site initially by laying the device on a clean tarpaulin spread on the ground adjacent to the well, and subsequently using a rack over which the device was suspended. The prepared device was then gently pushed into the screened well. On retrieval, the dialysis bags were removed from each canister and perforated to collect the enclosed water into individual labelled nalgene bottles. These bottles were transported on ice back to the laboratory for testing. Sampling in this manner provided approximately 150 ml of water from each sampling interval (1 ft, or 0.3 m) The sampling device was scrubbed and rinsed with tap water prior to reuse. 89 A monthly cycle of filling and retrieval was adopted at each site in order to monitor the groundwater chemical profile. Sampling was commenced in July 1994. The wells at the Brookswood aquifer site were undeveloped prior to device placement, whereas the Abbotsford well was fully developed using a submersible pump. At the end of month 1, and after the devices had been left in the wells for the full month, the dialysis bags retrieved from the Brookswood wells were found to be totally degraded, whereas those from Abbotsford were clean and intact. This degradation occurred in conjunction with a high content of fines accumulation within the device, and it is likely that microbes attached to the fines were responsible for the bag degradation. At the beginning of month 2, the Brookswood wells were developed. The devices were then placed and left in the wells for all of month 2. On subsequent retrieval, the bags from all wells showed a slight degradation at the end of the second sampling period, suggesting (at least in the Abbotsford well) that fines-associated microbial accumulation within the wells may have occurred as a result of repeated lowering and raising of the sampling device (and the flushing that this imposes across the well screen). A l l three wells were again developed at the beginning of month 3, and on the basis of the observations made after the first two periods of sampling (during which time the device was left in the well for one month prior to each retrieval), the devices were only lowered into the wells during mid-month and removed at the end of month 3 after only a two week residence in the wells. No bag degradation was observed after the two week residence time. A two week residence time was therefore adopted for the remainder of the monitoring. A series of laboratory tests, aimed at investigating the diffusion of dissolved 90 groundwater constituents into the dialysis membrane sausages, was performed to confirm the basis for the developed passive sampling technique. The methodology and the results are described in Appendix B. 3.2 L A B O R A T O R Y INVESTIGATIONS 3.2.1 Background The enhancement of the natural denitrification potential, or the introduction of nitrate (N03\") as an electron acceptor, within an aquifer wil l both result in the generation and introduction of molecular nitrogen (N2) into the groundwater. This N 2 wil l either remain in solution and travel along with the groundwater while diffusing into areas of lower concentration, or it may enter a separate gas phase and form bubbles within the subsurface formation. Ronen et al. (1989) postulated that such gas bubble evolution occurs from discrete microbial colonies, and mathematically showed that bubble accumulation has the potential to clog the formation and lead to localised impacts on the flow regime. As previously noted, there is a growing interest in denitrification in groundwater environments (Hiscock et a l . , 1991; Korom, 1992; Smith et a l . , 1994), as well as the injection of a carbon source for the reduction of groundwater nitrate levels (Duthy, 1993; Devlin and Barker, 1994). While clogging of subsurface formations near injection wells, or other recharge systems, as a result of biomass accumulations is well recognized (Taylor et a l . , 1990), the potential significance and manageability of nitrogen gas bubble formation and accumulation, during enhanced denitrification, is a question that remains unanswered. Soares et a l . , (1991), presented the findings of laboratory column testing aimed 91 at investigating the phenomenon of gas bubble accumulation during enhanced denitrification (by adding a carbon source to nitrate laden water) under conditions of saturated flow through sand. Sand columns were inoculated (apparently after column packing) with a \"coarsely-filtered suspension of fresh garden soil\". Potassium formate (KCH0 2 ) was used as the carbon source. Soares et al. reported that application of this substrate in a continuous manner led to a significant and localised accumulation of gas near the top (inlet) of the columns, as evidenced by the formation of \"dry patches\" through the sand. Pulse application of the substrate apparently did not result in gas bubble formation, but Soares et al. found similar long-term decreases in permeability under both continuous and pulse application of the formate. A review of other related findings on denitrification and nitrogen gas generation during saturated flow through porous media is presented in Chapter 6. Soares et al . , (1991), did not directly measure or identify the N 2 produced during their testing, and concluded denitrification solely on the basis of a reduction in N0 3 \" as well as the evidence of gas formation (presumed to be N 2). 3.2.2 Materials and Methods Laboratory investigations consisted of several sets of column tests used to investigate the fate of nitrogen gas that evolves during denitrification under saturated flow through sand. The column tests were similar to those used by Soares et a l . , (1991), and the basic column design used for the research was based on the descriptions provided by those investigators. Figure 3.6 presents a schematic description of the column setup used for the current study. Columns were constructed using lengths of clear acrylic tubing to 92 Figure 3.6 Schematic description of laboratory column setup 93 which matching polyethylene funnels were attached (using epoxy adhesive) as the base end. Silicone tubing was used to pump nitrate and carbon amended tap water into and out of the column. Constant flow rate conditions were applied using peristaltic pumps. The applied flow rates were selected to simulate pore water velocities typically encountered within the local aquifers, where the groundwater velocities are typically in the order of 0.1 to 1 m/day (Liebscher et a l . , 1992; Piteau Associates, 1994). The range of flow velocities applied during the column testing was 0.2 to 0.8 m/day. Six (6) sets of column tests was performed during the laboratory testing. With the exception of Set 1, each set consisted of two or more column tests performed simultaneously. Table 3.1 presents a summary of the test conditions during the various sets of column experiments. These sets are numbered in the order in which they were performed, and the six sets resulted in a combined total of 17 individual column tests. The column testing was terminated after the completion of Set 6, and by which time some clear conclusions were reached. This completion also coincided with the concentration of the research effort on the in-situ monitoring of groundwater nitrate distribution. The objectives and methods associated with the test sets are presented in the following sections. For Sets 2, 3, 4 and 5, only a brief summary of the test methods is presented in this Chapter, as the findings from these Sets are considered to be relatively less significant than those from Set 1 and 6. Further descriptions of Set 2 and 3 are presented separately in Appendix G. Table 3.1 Summary of column tests performed SET NO. L (cm) 0 (cm) TEST DURATION (days) FLOW RATE (ml/min) CARBON SOURCE TOPSOIL COLUMN MATERIAL SOURCE NITRATE (mg-NA) NITRATE REMOVAL (%) 1 18 8.25 90 0.67-1.25 methanol yes sand 13 see text 2 1 20 10 155 0.7 methanol yes sand 13-26 < 10 2 20 10 155 0.7 none yes sand 13-26 0 3 40 20 145 5 methanol yes gravel 13 - 26 < 10 4 40 20 145 5 none yes gravel 13-26 0 3 5 20 10 45 0.4 - 0.8 methanol yes sand 13 40 at lower flow 6 20 10 45 0.4 - 0.8 methanol yes sand 13 40 at lower flow 4 7 20 10 53 0.5 ethanol yes sand 13 100 8 20 10 60 0.5 ethanol yes sand 13 100 5 9 20 10 22 0.5 methanol yes sand 13-26 less than 40 10 20 10 22 0.5 ethanol yes sand 13-26 100 11 20 10 22 0.75 ethanol yes sand 13-26 100 12 20 10 22 0.5 ethanol no sand 13 100 6 13 20 10 30 0.5 ethanol no sand 26 see text 14 20 10 30 0.75 ethanol no sand 26 see text 15 20 10 30 1.0 ethanol no sand 26 see text 16 20 10 30 1.25 ethanol no sand 26 see text 95 3.2.2.1 Set 1 The objective of this single column test (Set 1) was to replicate the findings of Soares et al. (1991) and, while doing so, to also obtain a better \"feel\" for the column test methodology. Column preparation consisted of placing pea gravel in the funnel, followed by a layer of fine teflon wire mesh to prevent subsequent sand loss through the gravel. Tap water was then pumped into the base of the column in an upflow mode, using a peristaltic pump, in order to wet and saturate the gravel, and to also partially f i l l the column with water. Pumping was then stopped, and a wetted sand (uniform medium silica sand) and topsoil (amount approximately a tablespoon full) mixture was poured into the column and tamped gently in layers of 1 to 2 cm to the final height (18cm). The general approach taken during the testing was to maintain a constant flow rate through the column of sand and monitor the effluent characteristics while imposing various test conditions. Flow was vertically downward. Periodic visual inspection of the column was also performed. With the exception of a period during which total recycle of the flow was imposed, the source feed (contained in a 1 litre flask) consisted of tap water to which 90 mg-C/1 and 14 mg N0 3 -N/1 was added or removed as summarised in Table 3.2. The carbon source used was methanol (CH 3OH), and the nitrate-nitrogen source was potassium nitrate (KN0 3 ) . At the onset of the recycle in Phase 2, a 250 ml portion of the source feed was transferred into a mixing flask containing a stirring magnet. This flask was then placed on a magnetic stirrer, and the influent and effluent recycled through it. During the recycle phase, all samples were obtained from the effluent discharge, resulting in a gradual decrease in the total volume of water in the system, and also a decrease in the hydraulic residence time. 96 Table 3.2: Summary of column test conditions - Set 1 FLOW RATE FLOW CONDITION PORE VOLUMES INFLUENT CARBON INFLUENT NOj-N Phase 1 1.5 ml/min non-recycle 0 to 60 90mg-C/l introduced after 43 pore volumes and maintained until 60 pore volumes. 14mg-N/l introduced after 14 pore volumes and maintained until 60 pore volumes Phase 2 1.5 ml/min full recycle 60 to 127 a/a n/a Phase 3 1.5 ml/min non-recycle 127 to 154 90mg-C/l introduced after 127 pore volumes 14mg-N/l introduced after 142 pore volumes Phase 4 0.67 ml/min non-recycle 154 to 203 90mg-C/l maintained until 203 pore volumes 14mg-N/l maintained throughout this phase Phase 5 0.67 ml/min non-recycle 203 to 249 90 mg-C/1 reintroduced after 205 pore volumes and maintained 14mg-N/l removed after 212 pore volumes and reintroduced after 215 pore volumes Phase 6 1.25 ml/min non-recycle 249 to ... 90mg-C/l 14mg-N/l 97 During each of the phases of the testing, the addition and removal of C and N 0 3 -N from the source feed was performed while also draining (using a large syringe) the column supernatant to just above the top of the sand surface, and immediately replacing (again using a syringe) it with the new source feed. This was done so as to ensure that the concentrations of C and N0 3 \" -N entering the sand column were at known values at all times. Samples of effluent were collected periodically during all phases for the analysis of TOC, IC, N 0 3 \" and N0 2 \" . During Phases 4 and 5, samples were also collected for the analysis of ammonium (NH 4 + ) . 3.2.2.2 Set 2 This set of testing repeated Set 1, while using control columns to which no carbon was added. Methanol was again used as a carbon source. The behaviour under long term continuous flow without recycle was also monitored. Additionally, columns packed with a medium gravel were also tested. Phosphorus was added during later stages of flow through to assess the likelihood of denitrification being phosphorus limited. Further details of this Set are presented in Appendix G. 3.2.2.3 Set 3 This set of testing was performed to investigate the effect of flow rate reduction on the extent of denitrification across the column length, as well as the effect of using a lower concentration of methanol in the source feed. 98 Further details of this Set are presented in Appendix G. 3.2.2.4 Set 4 The Set 4 column tests were aimed at developing a method by which the dissolved N 2 and 0 2 content of small volumes of water samples could be quantified. A successful method of direct injection of liquid samples into a GC was developed (see Appendix C). Initially, the effluent was sampled at the discharge end of the silicone tubing that drained the base of each column. However, even though significant nitrate removal was being measured through the column, the dissolved gas measurements showed that the effluent was oxygen saturated. It was subsequently discovered that effluent sampled (using a gas tight syringe) closer to the base of the column, and at the entry into the silicone tubing, contained no oxygen. This finding suggested that oxygen diffusion across the wall of the silicone tubing may have led to its observed presence at the discharge end. Though this finding is of some significance, it was not investigated further as part of this research. This Set of tests was also performed using ethanol (CH 3CH 2OH) as a carbon source during flow through. 3.2.2.5 Set 5 This set consisted of a short series of tests aimed at confirming the earlier findings, from Sets 1 to 4, in the difference in the extent of denitrification achieved through the columns between methanol and ethanol amended waters. Additionally, a sand column to which no topsoil was added was also investigated. 99 3.2.2.6 Set 6 This set of column tests was performed to investigate in detail, the role of flow rate variations on denitrification and the generation and fate of N 2 , with ethanol as the carbon source through sand columns unamended with topsoil. Detailed measurements of the influent and effluent dissolved N 2 and 0 2 were obtained during this testing. This set of testing was a culmination of the various findings from Sets 1 to 5, and was performed to obtain a detailed and controlled measurement of denitrification and N 2 generation. Four columns were packed with sand (with no topsoil) as for the Set 1 tests. The prepared columns (each with remnant supernatant) were then covered with aluminium foil and left to stand for one week to allow attachment of any bacteria introduced with the tap water. At the end of one week, the supernatant in each column was gently stirred to resuspend any fines that may have settled onto the top of the packed sand column. The supernatant was slowly removed using a syringe to avoid disturbances to the packed sand. The supernatant was replaced with freshly prepared nitrate amended tap water (2mM KN0 3 ) , and the peristaltic pumps turned on to commence downflow through the columns. A different flow rate was adopted for each of the columns (0.5, 0.75, 1.0 and 1.25 ml/min). These flow rates were checked daily by collecting effluent into a measuring flask for one hour. The supernatant in each column was topped up daily with freshly prepared nitrate solution. At the end of one week (day 7) of continuous downflow with nitrate amended tapwater, the supernatant in each column was again removed, and replaced with 2mM KNO3 tapwater to which denatured alcohol (85 % ethanol and 15 % methanol) was added 100 at a proportion of 0.2 ml/1. Flow was not stopped during this introduction of a carbon source to the columns, as the rate of flow was small enough not to have lowered the water level below the sand surface during supernatant replacement. Continuous downflow with the carbon and nitrate amended tapwater was maintained for 16 days. Due to visible biogrowth development in the effluent tubes, on day 16, these were replaced with new tubing. This procedure led to the free draining of the pore fluids from the columns for a period of a few seconds. On day 22, the supernatants were again lowered, and replaced with 2mM KN0 3 amended tap water, with flow continued for another week. Flow was then stopped and the tests concluded. The sampling strategy during the testing consisted of daily collection, of effluent discharged from the peristaltic pump, as well as sampling directly below the base of the columns using a gas-tight syringe. Samples of the supernatant were also sampled for gas analyses. The water samples were analyzed for N03\", N02\", N H 4 + , IC and TOC. The samples (100 fil) obtained using the gas-tight syringe were analyzed for dissolved N 2 and 0 2. 3 . 3 ANALYTICAL TECHNIQUES All water analyses, with the exception of field DO measurements, were performed at the Environmental Engineering Laboratory (EEL), Department of Civil Engineering, University of British Columbia. The soil CEC/TEC analyses were also performed at the EEL. The soil TC/TOC analyses were performed by Norwest Laboratories, located in Langley, British Columbia. QA/QC procedures including analyzer calibrations using reagents of known 101 concentrations, as well as spike recovery testing, are routinely performed in these laboratories. No replicate testing was performed during the course of this study. The relatively high frequency of sampling during the laboratory testing, and the high frequency and closely spaced multi-level sampling of the groundwater, is considered to reduce sampling and analytical error uncertainty. 3.3.1 Water Sample Analyses Water samples obtained from the laboratory column testing were analyzed for the constituents described in Section 3.2.2. Water samples obtained from the monitoring wells were analyzed for \"major ions\"3, as well as N0 3 \" , N H 4 + , N0 2 \" , K + , pH, and DO (dissolved oxygen). Tables 3.3,3.4 and 3.5 present a summary of the analyses performed on the monitoring well samples. The techniques used for these quantifications are described below. 3.3.1.1 pH pH was measured using an Ag/AgCl electrode and a Bradley-James 10507 pH meter. The detection limit is 0.01 for this unit. 3.3.1.2 Chloride (Ct) CI\" measurements were made by either using a specific ion electrode (Orion 3 C a 2 + , N a + , M g 2 + , CI\", H C 0 3 - , S0 4 2 \" . Freeze & Cherry (1979) note that the total concentration of these six ions normally comprises more than 90% of total dissolved solids in natural waters. Table 3.3: Chemical analyses on water samples from Abbotsford Study Well DATE N O x N02 N H X TOC IC CI pH so42 C a 2 + Na + Mg 2 * K + DO Jul 26 '94 X X X X X X Aug 29 X X X X X Sep 27 X X X X X Oct 25 X X X X X X X X X Nov 28 X X X X X X X X Dec 19 X X X X X X X Jan 23 '95 X X X X X X X X Feb 27 X X X X X X X X X Mar 27 X X X X X X X X X X X Apr 24 X X X X X X X X X May 22 X X X X X X X X X X Jun 26 X X X X X X X X X X Table 3.4: Chemical analyses on water samples from Brookswood North Study Well DATE N O x N 0 2 N H X TOC IC CI pH S0 4 ^ C a 2 + Na + M g 2 + K + DO Sep 7 '94 X X X Sep 27 X X X X X X Oct 24 x X X X X X X X X Nov 28 X X X X X Dec 19 X X X X X X X Jan 23 '95 X X X X X X X X Feb 27 x X X X X X X X X Mar 27 X X X X X X X X X Apr 25 X X X X X X X X X X May 23 X X X X X X X X X X Jun 27 X X X X X X X X X X Table 3.5: Chemical analyses on water samples from Brookswood South Study Well DATE N O x N 0 2 N H X TOC IC CI pH S O ^ C a 2 + Na + M g 2 + K + Jul 8 '94 X X X X X Aug 22 X X X X Sep 12 X X X X Oct 11 X X X Nov 14 X X X X X X X < Dec 19 X X X X X X Jan 23 '95 X X X X X X X X Feb 27 X X X X X X X X X Mar 27 X X X X X X X X X X Apr 25 X X X X X X X X X X May 23 X X X X X X X X X X Jun 27 X X X X X X X X X X 105 Model 96-17B Combination Chloride Electrode, connected to a Perkin Elmer millivolt meter), or by using a colorimetric method in a L A C H A T Instruments QuickChem Automated Ion Analyzer. The detection limit for both of these techniques is 0.05 mg/l. 3.3.1.3 Total and Inorganic Carbon (TC/TIC) Total Carbon (TC) and Inorganic Carbon (IC) determinations were separately made by injecting water samples into a SH IMADZU TOC-500 Total Organic Carbon Analyses, operated using a \"combustion non-dispersive infrared gas analysis method\" (from Users Manual notes). The TC is converted to C 0 2 at 680\u00C2\u00B0C, and IC is converted to C 0 2 at 150\u00C2\u00B0C, and the generated C 0 2 quantified and converted to TC or IC by using calibration standard data. Total Organic Carbon (TOC) is computed as TC - IC. The detection limit for the TOC analyzer is rated at 1 mg/l, however experience within E E L suggests that repeatability of measured values is reduced for TC/IC values less than 5 mg/l. 3.3.1.4 Ca, Na, Mg, K These constituents were quantified using a Thermo Jarrel Ash IL VIDEO 22 AA/AE- Spectrophotometer. Ca, Na and M g were quantified using A A (atomic absorption), while K was determined using A E (atomic emission) spectrophotometry, all with an air/acetylene flame. These determinations were made after the bottle contents had been acidified to approximately pH 2 using three to four drops of concentrated HNO3. 106 The detection limits are 0.01 mg/1 for Ca and K; 0.001 mg/1 for Na; and 0.003 mg/1 for M g . 3.3.1.5 N03, N02, NH4+, SO? These constituents were individually quantified using a L A C H A T Instruments QuickChem Automated Ion Analyser using colorimetric techniques. Samples tested for N0 3 \" and N0 2 \" were preserved with a few drops of a solution of \"NOx Preservative\" (0. lg phenyl mercuric acetate, 20ml acetone, 280ml H 2 0) immediately on collection, and samples tested for N H 4 + were acidified immediately on collection to pH 2 using a few drops of dilute H 2 S0 4 . The detection limits are 0.05 mg-N/1 for N0 3 \" and N 0 2 ' ; and 0.5 mg/1 for N H 4 + -N and S 0 4 2 \ 3.3.1.6 Dissolved 02 (DO) - Groundwater Dissolved oxygen measurements were made on samples of the groundwater using a H A C H Model OX-2P Test Kit, which is based on the Winkler Technique for DO quantification. The rated detection limit for this technique is 0.1 mg/1. 3.3.1.7 Dissolved 02 & N2- Laboratory Column Samples Effluent from the laboratory columns was analyzed for dissolved 0 2 and N 2 using a gas chromatograph with a direct injection technique that was developed as part of the current research. The samples (100 /xl) obtained using the gas-tight syringe were 107 analyzed for dissolved N 2 and 0 2 by direct injection into a gas chromatograph (Hewlett Packard 5750 Research Chromatograph) with a thermal conductivity detector, and a 2 mm stainless steel column packed with Molecular Sieve 5A. Column length was 2 m. The column temperature was set at 30 \u00C2\u00B0C during testing. At the end of each daily set (four effluent and four supernatant samples) of injections, the GC column temperature was raised to 190\u00C2\u00B0C for a period of 8 hr, and then lowered to, and left overnight at, 30 \u00C2\u00B0C in preparation for the following day's set of injections. This technique was found to produce repeatable results with a linear correlation between peak areas and dissolved gas concentrations over the encountered range. Further discussion of the injection technique, and its development, is presented in Appendix C. 3.3.2 Soil Sample Analyses 3.3.2.1 Carbon Content Total Carbon (TC) and Total Organic Carbon (TOC) determinations were made on several samples of soil that was recovered during the drilling and installation of the three monitoring wells. This analysis was performed using a Leco-CR12 Carbon Analyzer. Each sample was air dried and then pulverised to separate the soil grains, prior to being passed through a 2 mm sieve. 0.5 g of the sieved fraction was weighed into a ceramic crucible, and inserted into a combustion tube located within the Leco-CR12 Analyses. Pure oxygen was passed through the combustion tube. For Total Carbon measurements, the combustion tube was heated to a temperature of 1371 \u00C2\u00B0C (2500 \u00C2\u00B0F), and for Total Organic Carbon measurement was heated to 816 \u00C2\u00B0C (1500 \u00C2\u00B0F). In both 108 cases, carbon is oxidised to carbon dioxide, which is then quantified using an infrared detector, and the carbon content computed using calibration standard data. Total Inorganic Carbon (TIC) was computed as the difference between TC and TOC. The detection limits for TC and TIC using this technique are 0.01 % dry weight. 3.3.2.2 Exchangeable Cations 30 g of oven dried (at 104 \u00C2\u00B0C) soil samples were weighed out into individual 125 ml erlemeyer flasks. 125 ml of I N ammonium acetate (NILOAc) solution (pH 7 as prepared) was then added to each flask, which was then stoppered. The flask was shaken by hand and left overnight. The flask was then shaken again, and the contents filtered through No. 541 filter papet into a collection flask. The retained soil was washed further with either 75 ml or 125 ml of I N NILOAc to give 200 ml or 250 ml of filtrate. The filter paper, with the retained soil, was placed aside for subsequent exchange capacity determination. The filtrate was analyzed for Ca, Na, M g and K (in mg/l) by the methods previously described. The Total Exchangeable Cation (TEC) value, in meq/100 g, was computed as TEC = '^-A x _ ? _ X 1 0 0 g (3.1) 1000 ml 30 g 6 where A is the sum total of Ca, Na, M g and K expressed as meq/14, and B is either 200 ml or 250 ml depending on the total amount of filtrate produced. The filter paper containing the ammonium saturated soil was then placed over a 4 meq/1 = milliequivalents/litre, computed as (molarity X valence), with molarity expressed as mmol/1 109 clean collection flask and washed with 200 ml of iso-propanol in approximately 40 ml portions, while allowing each portion to drain completely before adding the next. The filtrate (iso-propanol) was discarded. The soil-filter funnel was replaced over the collection flask, and 250 ml of IN KC1 was leached through the soil in portions, allowing each portion to drain completely. The collected filtrate was analyzed for ammonium. Cation Exchange Capacity (CEC) was computed as CEC = C x 2 5 0 m l X100 g (3.2) 1000 ml 30 g 6 where C = filtrate concentration of ammonium in meq/1. 110 4.0 RESULTS OF ABBOTSFORD AQUIFER INVESTIGATIONS This Chapter presents a compilation of the results of the analyses performed on samples of groundwater and soil obtained from the Abbotsford study well. A brief comparison of these results to other published findings from the Abbotsford aquifer is also presented in this Chapter. Detailed evaluation of these results, and a discussion of the broader implications with respect to the source, transport and fate of non-point source nitrate in unconfined aquifers, are contained in Chapter 7. 4.1 SUBSURFACE GEOLOGY 4.1.1 Soil Characteristics The soil samples obtained during the installation of the study well can be all classified as gravelly medium to coarse grained sand with varying amounts of fines, overlain by a thin (less than lm) top soil layer. The nature of the sampling (bailer cuttings obtained over 1.5 m intervals) preclude any detailed delineation of the stratigraphy of the underlying deposits. However, the sample obtained from a depth of 16.8 m (55 ft - ie. bailer contents obtained between 50 and 55 ft) was described by the driller as \"t i l l \" , and the obtained sample appeared to have a higher fines content than the overlying soils. The thickness of this till zone could not be determined, but the samples obtained from the 18.3 m (60 ft) and 19.8 m (65 ft) depths were similar to the soils in the upper 15 m (50 ft) of the well. From the Soil Map Mosaics and Legend contained in Luttmerding (1980), it is noted that the surficial soil type over the study site is named \"Abbotsford\" (abbreviation I l l AD) , and described as consisting of \"20 to 50 cm of medium textured eolian deposits over gravelly glacial outwash\", and exhibit well to rapid drainage. The soils at the study site also form part of the \"Abbotsford Soil Management Group\" (Luttmerding, 1984), in which the surficial eolian deposits provide some water storage capacity. However, the rapidly draining subsoils and the limited thickness of the aeolian deposits are considered to result in the relatively free and rapid infiltration of water to the water table. A l l major aquifers in the Fraser Lowland are situated in glaciofluvial sand and gravel deposits. The Abbotsford aquifer is composed of a succession of stratified permeable glaciofluvial sands and gravels, interspersed with minor till and clayey silt lenses (Atwater et a l . , 1994). Ricketts & Liebscher (1994) note that, though the aquifer is dominated by sand and gravel, stratigraphic sections as observed in exposures (eg. gravel pits) exhibit substantial differences in grain size and compaction. The base of the aquifer has not been fully explored, but the aquifer has been found to reach a thickness of 70 m locally (Liebscher et a l . , 1992). The glacial till lenses can lead to locally confined conditions for groundwater flow, though for the most part the aquifer is considered to be unconfined (Kohut, 1987). The till considered to have been encountered at a depth of 16.8 m (55 ft) at the study well may be such a localised lens. Other findings (Liebscher et a l . , 1992; Roth, 1994) suggest that in the vicinity of the study site the sand and gravel deposits are relatively uniform with no major till lenses having been encountered. The base of the permeable sands and gravels is present at a relatively uniform depth of 25 to 30 m below existing ground surface. An underlying thick strata of low permeability glaciomarine and marine clay has been encountered in several deep wells located along Mt. Lehmann Road, Hamm Road and 0 Avenue. Figure 4.1 contains Figure 4.1 Geological fence diagram over region including Abbotsford study site (from Liebscher et al., 1992) 113 a geological fence diagram that was prepared by Liebscher et al. (1992), and which shows relatively uniform geology in the vicinity of the study site. 4.1.2 Depth to Water Table Figure 4.2 presents a plot of the depth to water table versus time as measured at the study well during the monitoring period (July 1994 to June 1995). A plot of the monthly precipitation measured at the Abbotsford airport between January 1992 and June 1995 is also presented on Figure 4.2. As shown on Figure 4.2, the water table at the study well exhibited a fluctuation of 2.8 m during the monitoring period. The water table rose rapidly between October 1994 and March 1995, after which it dropped equally rapidly over the remainder of the monitoring period. The water table was shallowest during March 1995, when it was measured to be at a depth of approximately 2.7 m. This meant that during that month, the water table was 1.9 m above the well screen (at a depth of 4.6 m below ground surface). The water table was shallower than 4.6 m during and between the months of December 1994 and May 1995. The water table was deepest during October 1994, when it was at a depth of 5.5 m. Depth to groundwater is variable across the aquifer. Figure 4.3 contains a compilation of plots of water table elevations as measured by Environment Canada between 1990 and 1993 at several individual and clustered monitoring wells that were located to the south and west of the airport (Hii, B., 1995). These plots are presented here for illustrative purposes only, and so well depths and locations are not included. From these plots it is noted that seasonal water table fluctuations range between 2 to 3 m, but that the low and high water table elevations vary from year to year. 114 2 - T Dec 92 Dec 93 Dec 94 Dec 95 Figure 4.2 Monthly depths to water table in Abbotsford study well during monitoring period, and monthly precipitation at Abbotsford Airport (precipitation data from Environment Canada) 115 50 - r Dec 90 Dec 91 Dec 92 Dec 93 Figure 4.3 Monthly total precipitation at Abbotsford Airport, and water table fluctuations in monitoring wells in vicinity of study site between 1990 and 1993 (data from Environment Canada) 116 Furthermore, the months associated with the high and low water table also vary from year to year. Kohut (1987) presented longer term data which shows that long term overall fluctuations of up to 6 m also occur, and can be correlated to departures in precipitation patterns from the average. The plots presented on Figure 4.3 show water table fluctuations typically exhibiting a 2-3 month lag between maximum precipitation (December - February) and highest water table level (March - May). The aquifer is recharged by direct precipitation over much of its area. The topography at, and surrounding, the study site is flat, so surface runoff is not a controlling factor in the hydrologic balance. Kohut (1987) has estimated that 37% to 81 % of the annual precipitation may recharge to the aquifer, with the higher value being more probable in areas not affected by surface runoff and stream drainage. Using these values, and assuming an average annual rainfall of 1500 mm, the recharge amount is then estimated to be between 560 mm and 1200 mm. Figure 4.4 shows the water table configuration and flow patterns as interpreted for the aquifer by Liebscher et al. (1992). In the vicinity of the study site, south of the airport and along 0 Avenue, the groundwater flow direction is southwesterly. A similar plot contained in Roth (1994) suggests that the 50 m water table contour located north-east of the airport as shown on Figure 4.4 may be associated with a groundwater divide. 4.2 LAND USE AT SITE DURING MONITORING PERIOD A change in the land use in the vicinity of the study well was observed during the monitoring period. The area associated with this land use change is shown on Figure 4.5. During July and August 1994, the field upgradient of the well supported mature C U a r b c o o k LEGEND * Mil l Laka B L a . l o n l a k a C Judaon Laaa D Abbofaford Airpoil A b b O t t f o i d * f E E E E 6 o \u00C2\u00BB o \u00C2\u00AB o \u00C2\u00A35 S u m n MOTES: 1. Equlpolanltal llnaa daiUad tfom data point, a . lacoid.d In Match of 1987.88.t SS and aaaaonallf ad|ualad Or IWD-P4V laslon.Vancouvai * USGS Pacific N o i l h \u00C2\u00AB \u00C2\u00AB . i . lacoma. 2. Oafa polnla ara not ahoavn on thla map. LEGEND *Qm \u00C2\u00BB A I I \u00C2\u00AB l a e i t c O N l o u a I fc l t lRES AOOVt StA I C V I L I GaOUNOWAIIN f tUW Lyndon Figure 4.4 Water table configuration and groundwater flow patterns for Abbotsford Aquifer as interpreted by Liebscher et al. (1992) 118 Huntigdon Road s on c c E -C A o o B C o berry canes removed over this region in Sept '94 0 Avenue ra a E E ra I Study Well 250 500m Figure 4.5 Area associated with land use changes during monitoring period in vicinity of Abbotsford study well 119 raspberry canes that had been harvested. During the September site visit it was noted that all of these canes had been removed and the field cleared of support posts. During the October site visit it was noted that the field had been tilled. The field was left untouched until February 1995, when it was tilled again and young raspberry canes were planted. During March and April 1995, manure was applied on to the field. Discussions with the farmer at the site ascertained that such crop removal is performed every eight years or so to replace old crops (Froese, W. , 1995 - personal communication with site farmer). The significance of this land use change in relation to the measured groundwater chemistry at the study well is discussed in further detail in Chapter 7. 4.3 GROUNDWATER CHEMISTRY IN STUDY W E L L 4.3.1 Major Ion Chemistry, pH, IC & TOC 4.3.1.1 Study Well Figure 4.6 presents a series of combined plots of major ion, pH, TOC and IC data obtained between July 1994 and June 1995 at the study well. Detailed plots of the data as obtained during each monitoring month are contained in Appendix D. This data shows the water chemistry as encountered over the screened depth of the well. As the top of the screen is at a depth of 4.6m (15ft) below ground surface, the chemistry at and below the water table to a depth of 4.6 m was not obtained for the months when the water table was shallower than 4.6 m. The plots on Figure 4.6 all show that the concentrations of N0 3 \" and the major ions (except HC0 3 \" , as well as pH) are highest near the water table and decrease over depth, pointing to recharge waters being the principal source of dissolved chemistry to 120 Nitrate (mg-N/l) Sulfate (mg/l) Chloride (mg/l) 0 10 20 30 40 50 0 20 40 60 0 5 10 15 20 25 TOC (mg/l) IC (mg/l) pH 0 10 20 30 40 50 0 10 20 30 40 50 5.8 6.2 6.6 7.0 Figure 4.6 Combined plots of monthly major ion chemistry, pH and TOC profiles obtained for groundwater at Abbotsford study well Calcium (mg/1) 0 20 40 60 4 - j \u00E2\u0080\u0094 i \u00E2\u0080\u0094 I i I i I W 1 8 J \u00C2\u00A7 Sodium (mg/l) 2 O 0 4 8 12 I 4 Magnesium (mg/l) 0 5 10 15 \u00E2\u0080\u00A2 ' ' 1 i 1 1 ' 1 1 1 ' ' 1 1 Potassium (mg/l) 0 5 10 15 20 25 Figure 4 . 6 ( . . .cont) 122 the upper part of the aquifer. Nitrate leaching to the water table has been found to be accompanied by the delivery of C a 2 + ; N a + , M g 2 + , K + , S 0 4 2 and CT. pH and IC (HC0 3 ) both were found to increase with depth. The measured pH values range from 5.8 to 6.8 over the monitoring depth. In comparison to the several detailed vertical groundwater nitrate profiles encountered at other sites in the world, and which were presented in Chapter 2, the profiles obtained at the study well do not display a steady state steep concentration gradient, nor do they display a reduction of nitrate levels to zero, or even small \"natural\" (<3 mg-N/1) levels. The nitrate levels over the monitoring depth remained above 10 mg-N/1 throughout the monitoring period. The profiles shown on Figure 4.6 show that the groundwater chemistry exhibited significant temporal fluctuation during the monitoring period, except below a depth of 13m where the chemistry showed little to no change over the monitoring period. The relatively greater temporal variation associated with the pH values, even below 13 m, are partly considered to be due to the limited accuracy associated with pH measurement. The measured TOC values are less than 5 mg/l over the full monitoring depth, and show no distinct temporal variation. It is possible that this TOC may have originated from the dialysis membrane used as part of the passive sampling approach. The testing performed on the regenerated cellulose membrane material, and as described in Appendix B, showed that TOC contributions from the membrane into the enclosed waters can be expected on prolonged submergence. Given the uniform nature, and relatively low magnitude, of the TOC values with depth, the values are not considered to be representative of the TOC content of the surrounding groundwaters. This conclusion can 123 only be verified by direct sampling of the groundwater, such as through pumping. Figure 4.7 presents individual plots of the nitrate profiles obtained during the monitoring period. These profiles exhibited significant temporal variation at this site over the twelve months of monitoring. A feature of the observed distribution of nitrate (and associated chemistry) beneath the water table at this site is that, unlike many of the cases reviewed in Chapter 2, the vertical chemistry profile here is not as static over time as observed elsewhere. A temporally stable steep chemical gradient has not been encountered at this site over the monitoring depth. While a comparatively steep gradient was observed during March, 1995, this was found to be shortlived. Despite the temporal fluctuations associated with the vertical distributions of chemistry beneath the water table, Figures 4.6 and 4.7 show that nitrate concentrations below the water table at this site did not fall below 10 mg-N/1 during the monitoring period, except in the immediate vicinity of the water table. A detailed discussion of the observed temporal variation is presented in Chapter 7. 4.3.1.2 Comparison to BCMOE Well data The findings obtained from the study well can be compared to those obtained by the BC Ministry of Environment (BCMOE) during monitoring from three well nests A , B and C located in the vicinity of the study site (Figure 3.2). The B C M O E installed these wells in 1987. Each well nest consists of four individual piezometers with screened depths of 7.6 m (25 ft), 10.7 m (35 ft), 16.8 m (55 ft) and 22.9 m (75 ft) below ground surface (Chwojka, 1990). Typically, each piezometer consists of 0.05 m (2 inch) P V C pipe with a 1 m (3 ft) screened interval (for the 10.7, 16.8 and 22.9 m depth wells) and 124 0 2 H 4 6 8 10 -| 12 14 16 -18 -Jul '94 0 10 20 30 40 50 i I i I i I i I i I Nov '94 0 10 20 30 40 50 CD O \u00C2\u00AB CO T3 C 3 2 CD $ _o o . (D Q 2 H 4 6 8 10 12 14 16 -\ 18 \u00E2\u0080\u00A2 I i I i I i I i I 0 2 H 4 6 8 -\ 10 12 H 14 16 18 Mar'95 0 10 20 30 40 50 \u00E2\u0080\u00A2 I i I i I \u00E2\u0080\u00A2 I i I Nitrate (mg-N/l) Aug '94 Sept '94 0 10 20 30 40 50 0 10 20 30 40 50 I i I i I i I i I I i I i I \u00E2\u0080\u00A2 I Dec '94 Jan '95 0 10 20 30 40 50 0 10 20 30 40 50 I i I i I i I i i I i I i I i I i I Apr '95 0 10 20 30 40 50 i I i I i I i I i I May '95 0 10 20 30 40 50 \u00E2\u0080\u00A2 l \u00E2\u0080\u00A2 I \u00E2\u0080\u00A2 I . I \u00E2\u0080\u00A2 I Oct '94 0 10 20 30 40 50 I i I i I i I i I Feb '95 0 10 20 30 40 50 \u00E2\u0080\u00A2 I \u00E2\u0080\u00A2 I \u00E2\u0080\u00A2 \u00C2\u00BB . I \u00E2\u0080\u00A2 I Jun '95 0 10 20 30 40 50 I \u00E2\u0080\u00A2 I \u00E2\u0080\u00A2 I \u00E2\u0080\u00A2 I \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 Figure 4.7 Monthly groundwater nitrate profiles at Abbotsford study well 125 an 2.4 m (8 ft) screened interval (for the 7.6 m depth well). Figure 4.8 shows the temporal variation, between 1988 and 1991, of nitrate-N concentrations within the three B C M O E wells (Kohut et a l . , 1993). These plots were prepared using data collected only every six months. The plots show that nitrate concentrations decrease with depth, and also exhibit temporal variation, with the greatest variations occurring near the water table within any given year or over the long term. The plots shown on Figure 4.8 show no clear trend to the temporal variations, but Kohut et al. (1993), on the basis of their various findings, concluded that \"concentrations with depth oscillate with time, suggesting a pulsating effect with contaminants moving downwards and southwesterly in a downgradient direction\". Generally, the temporal variations over depth at the study well are comparable to the B C M O E findings. A more detailed comparison of the study findings with the findings of the B C M O E , is presented in Chapter 7. 4.3.2 Dissolved Oxygen Table 4.1 presents the dissolved oxygen data obtained using the Hach kit during the March 27, 1995 sampling event. The data show that the dissolved oxygen content decreases with depth, but is still above 4 mg/l at the base of the well. The dissolved oxygen measurements show that the groundwater encountered in the study well is oxic over the monitoring depth (16 m). This finding is consistent with measurements obtained by the U.S. Geological Survey (USGS) from the B C M O E wells A , B and C located in the vicinity of the study well. The USGS data, shown on Table 4.2, forms part of larger database of DO measurements made from wells located 126 SfTE A-25 6.00 5.00 4.00 3.00 -Mill c i T F A O l 1 C n 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I I 1 1 1 1 1 1 1 1 1 SITE A-55 9.00 SITE A-75 5.00 | i i i i i i i i i i i 1988 I I l : I i I i i I I | I I I I I 1 I i I I I | i I I 1 i I I i i I I 1989 1990 1991 Figure 4.8a Temporal variation of groundwater nitrate as measured within BCMOE well A between 1988 and 1991 (Kohut et al., 1993) 127.. 45.00 40.00 35.00 30.00 Nitrate-N SITE B - 3 5 5.00 5.00 0.00 SITE B - 5 5 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ! 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 . 1 1 1 1 M 1 1 1 I ! 1 i i l l 1 1 i 1 l 1 l 1 1 i 1 1 i I 1 1 1 1 1 i i i i i i i i i i i | i i i i i i i i i i i | 1988 1989 1990 1991 Figure 4.8b Temporal variation of groundwater nitrate as measured within BCMOE well B between 1988 and 1991 (Kohut et al.,. 1993) 25.00 20.00 15.00 1 0.00 5.00 Nitrate-N mg/L | -SITE C - 2 5 0.00 | i i I i i i i i i i i I i i i i i i i i i i i I i i i i i i i i i i i I i i i i i i > i i i i 25.00 20.00 15.00 SITE C - 3 5 10.00 \u00E2\u0080\u0094 ) i i i i i i i i i i \u00E2\u0080\u0094 i i i i i i i i i i i I i i i i i i i i i i i \u00E2\u0080\u0094 i M i i M i i i 15.00 10.00 SITE C - 7 5 1990 1991 Figure 4.8c Temporal variation of groundwater nitrate as measured within BCMOE well C between 1988 and 1991 (Kohut et al., 1993) Table 4.1 Dissolved Oxygen Levels in Study Well - March , 1995 Depth DO (m) (mg/l) 4.9 7 6.2 7 8.1 6 9.8 8.5 11.4 7.5 13.0 6.5 14.6 4.5 Table 4.2 D O (mg/l) measurements from B C M O E Wells A , B & C (Cox, 1995) Depth Depth Well Well Well (m) (ft) A B C 6.1 20 8.8 7.6 25 8.1 9.0 10.7 35 7.6 8.5 9.2 16.8 55 6.0 6.4 5.0 22.9 75 1.0 0.2 0.8 130 throughout the Abbotsford aquifer, and which shows the underlying shallow groundwaters to be oxic over most of the aquifer (Cox, 1995). The dissolved oxygen levels in the groundwater at a depth of 22.9 m are at, or less than, 1 mg/l (Table 4.2). 4.3.3 Ammonium and Nitrite Nitrite (N02\") measurements were obtained during July and August 1994. Nitrite was detected at concentrations of 0.673 and 0.655 mg-N/1 in the upper 0.6 m of the water table during July, but was below detection level (0.05 mg-N/1) over the rest of the well depth during that month as well as during August. Ammonium (NH 4 + ) measurements were made during July 1994 and March 1995. Ammonium was detected at a concentration of 0.099 mg-N/1 at the water table during July, but was below detection level (0.05 mg-N/1) over the rest of the well depth. During March 1995, ammonium was at concentrations above the detection level above a depth of 8.1 m below ground surface, and to the top of the screened interval (4.6 m) as summarised in Table 4.3. The significance of the findings of ammonium and nitrite near the water table, at levels above detection limits, is also discussed in Chapter 7. 4.4 SOIL C H E M I S T R Y A T STUDY W E L L 4.4.1 Carbon Content of Subsurface Soils The results of the total organic carbon (TOC), total carbon (TC), and total inorganic carbon (TIC) determinations in soils of the retrieved drill cuttings are presented in Table 4.4. Most of the obtained values for TOC are at or below the detection limit of Table 4.3 Ammonium (mg-N/1) in Study Well March 27, 1995 Depth (m) 8.1 7.8 7.5 7.2 6.9 6.5 N H 4 + 0.063 0.073 0.074 0.295 0.271 0.180 Depth (m) 6.2 5.9 5.6 5.3 4.9 4.6 N H 4 + 0.304 0.312 0.460 0.626 0.469 0.463 Table 4.4 Carbon Content of Soils from Study Wel l Depth (m) Depth (ft) TC (% dry wt.) TOC (% dry wt.) TIC (% dry wt.) 1.5 5 0.056 0.021 0.035 3.0 10 0.064 0.020 0.044 4.6 15 0.179 0.009 0.170 6.1 20 0.176 0.129 0.047 7.6 25 0.013 0.010 0.003 9.1 30 0.010 0.001 0.009 10.7 35 0.043 0.015 0.028 12.2 40 0.023 0.009 0.014 13.7 45 0.020 0.002 0.018 15.2 50 0.022 0.001 0.021 16.8 55 0.068 0.001 0.067 18.3 60 0.032 0.003 0.029 132 0.01 % 1 for the analyzer used for these quantifications, and on the whole are less than 0.02% except for the sample from 6 m which had a TOC value of 0.129%. 4.4.2 Cation Exchange Characteristics of Subsurface Soils The measured values of the total exchangeable cations (TEC) and cation exchange capacity (CEC) are presented in Table 4.5. Base saturation values computed as the ratio of TEC to CEC are also shown on Table 4.5. The CEC values of the soil are all less than 2.5 meq/lOOg, with an average value of 1.6 meq/100 g. The TEC values, and therefore base saturation, have been found to increase with depth, with base saturation is near 100% between 13.7 m and 16.8 m. The low measured values of CEC for the soils encountered at the Abbotsford site are comparable to those measured by Ceazan et al. (1989), Dance & Reardon (1983) and Bjerg & Christensen (1993a) for the similar glaciofluvial soils encountered at their sites. 1 personal communication with Norwest Labs. 133 Table 4.5 Cation Exchange Characteristics of Soils from Study Well Depth (m) Ca exch. M g exch. Na exch. K exch. TEC (meq/lOOg) CEC (meq/lOOg) Base Sat. (%) 1.5 0.29 0.04 0.06 0.09 0.48 1.43 34 3.0 0.32 0.05 0.04 0.11 0.52 0.96 54 4.6 0.55 0.07 0.07 0.14 0.83 1.75 47 6.1 0.36 0 0.05 0.13 0.54 1.60 34 7.6 0.43 0 0.05 0.12 0.60 1.31 46 9.1 0.87 0 0.05 0.10 1.02 2.38 43 10.7 1.25 0.02 0.06 0.12 1.45 2.37 61 12.2 1.27 0.02 0.05 0.10 1.44 1.81 80 13.7 0.9 0.15 0.04 0.05 1.14 1.06 107 15.2 0.98 0.08 0.05 0.06 1.17 1.12 104 16.8 1.22 0.09 0.10 0.05 1.46 1.47 99 18.3 1.27 0.16 0.12 0.03 1.58 2.07 76 134 5.0 RESULTS OF BROOKSWOOD AQUIFER INVESTIGATIONS This Chapter presents a compilation of the results of the analyses performed on samples of groundwater and soil obtained from the Brookswood North and South study wells. A brief comparison of these results to other published findings from the Brookswood aquifer is also presented in this Chapter. Detailed evaluation of the results is presented in Chapter 7. 5.1 SUBSURFACE GEOLOGY 5.1.1 Soil Characteristics The soil samples collected during the installation of the study wells can be described as gravelly sands containing varying amounts of silt and occasional clay. A lump of clay was present in the cuttings retrieved from a depth of 15.2 m (50 ft) at the location of the North study well, but it is not apparent if this represents a continuous layer or a small lens within the gravelly sands. From the Soil Map Mosaics & Legend contained in Luttmerding (1980), the surficial soil type at the study site is named \"Lynden\" (abbreviation LY ) , and described as \"coarse textured glacial outwash deposits\" exhibiting well to rapid drainage. The Brookswood aquifer (also known as the Langley aquifer) is also located within glaciofluivial sands and gravels, and contains occasional interbeds of silt lenses and layers. This aquifer has been less studied in comparison to the Abbotsford aquifer, but available information shows that the subsurface sediments become finer with increasing depth. The thickness of the aquifer ranges from 15 m to 55 m, and in the 135 vicinity of the study site has been estimated to be about 20 m. A lower permeability strata of clayey silt and silty fine sand is considered to represent the base of the aquifer (Piteau Associates, 1994). Figure 5.1 presents a simplified vertical profile of the geology beneath the study site along the cross section A A ' location shown on Figure 3.3. 5.1.2 Depth to Water Table Figure 5.2 presents a plot of the depth to water table versus time as measured at the North and South study wells during the monitoring period. The monthly precipitation data collected at Abbotsford airport is also plotted on Figure 5.2, on the assumption that the precipitation pattern over the Brookswood site is similar to the Abbotsford site. The measured temporal fluctuations are similar at the two study wells. The depth to the water table at the two well sites ranges from 6.5 m to 9.3 m below ground surface. The seasonal water table fluctuation of 2.8m is similar to the fluctuation measured at the Abbotsford study well, as is the occurrence of high water table in March/April, and the low water table in October. The similar fluctuations observed at the South and North study wells suggest also that drawdown associated with the high capacity pumping well that is located to the east of the South well (Figure 3.3) does not extend as far as the South well (as previously discussed in Section 3.1.1.2). Based on a level survey performed after the completion of all monitoring activity, it has been estimated that the ground surface elevation at the North well is approximately 0.4 m lower than at the South well. The plots on Figure 5.1 show that the water table levels at the North well are lower by approximately 0.3 m than the levels at the South North Well South Well R t e a u Associates f 1994) Wells 7^ water table sand & gravel nroundwater flow direction presumed base of sand & gravel Figure 5.1 Vertical geological profile beneath Brookswood study site along Section AA' (Figure 3.3) 137 3 7 CD I-oo Q. Q E c g \"co \u00E2\u0080\u0094^\u00C2\u00BB \"5. \"o cu D_ c o 8 9 -10 300 South Wall - e - North Well 200 ^ 100 0 Dec 92 Dec 93 Dec 94 Dec 95 Figure 5.2 Depths to water table below ground surface at Brookswood North and South study wells, as related to monthly total precipitaion at Abbotsford airport (precipitation data obtained from Environment Canada) 138 well - the exception being for the months of November and December, when there was no difference in depth to water between the two wells. On the basis of this data it is noted that the water table elevation drops by 0.4 to 0.7 m between the South and North wells, over a distance of 350 m. The calculated gradient between the two wells is therefore between 0.001 and 0.002 near the surface, with flow direction northward. However, this may not be indicative of the gradient or the flow direction at depth. Unlike for the Abbotsford aquifer, the groundwater flow regime at the Brookswood aquifer is not very well delineated (Ricketts & Liebscher, 1994). Much of the available information on the recharge and groundwater flow characteristics of the Brookswood aquifer has been recently summarised in a report prepared by Piteau Associates (1994). This report presented the results of an impact assessment of a septic tank drain field that is located across the road (36th Ave), and upgradient of the South study well as shown on Figure 3.3. The findings presented in the Piteau Associates report suggest that groundwater flow at the study site is predominantly horizontal, with a water table gradient of approximately 0.01, and the flow direction being to the north-east. Groundwater velocities within the aquifer have been estimated to be between 250 m/year and 875 m/year. The water table elevations measured at the South and North wells therefore support the reported findings of flow direction to the north. The Brookswood aquifer is recharged primarily by direct precipitation, with some contributions from surrounding creeks. It has been estimated that 25 to 30% of annual precipitation recharges to the aquifer (Piteau Associates, 1994). Assuming an annual precipitation of 1500 mm, this equates to recharge of between 350 mm and 450 mm, which is less than the estimated range for the Abbotsford site. 139 5.2 G R O U N D W A T E R C H E M I S T R Y A T STUDY W E L L S 5.2.1 Major Ion Chemistry, p H , IC & T O C As the water table is located well below the top (at a depth of 4.6 m) of the screened interval at both the North and South study wells, the obtained profiles include the chemistry at and below the water table during all the months of monitoring at each well. 5.2.1.1 North Study Well Figure 5.3 presents a series of combined plots of major ion, pH, TOC and IC data between September 1994 and June 1995 at the North study well. Detailed plots of the profiles for each month are contained in Appendix E . The plots show elevated concentrations of N0 3 \" , CI\", C a 2 + , M g 2 + and N a + at a depth of approximately 3m below the water table at the North study well. The profiles suggest that all of these constituents arrived at the water table with recharge waters. While the concentrations of these various constituents are lower, and remained relatively constant over time below a depth of 3 m below the water table, they also show a slight increase over depth. Unlike at the Abbotsford study well, elevated levels of IC, K + and S0 4 2 \" are not associated with the elevated nitrate levels near the water table at the North well. The concentrations of these constituents remained relatively constant over the monitoring period, but also showed a slight increase with depth. The values for pH exhibited considerable variation over the monitoring period, but most of the profiles of pH decrease sharply near the bottom of the well. The variation is considered to be due 140 Nitrate (mg-N/1) 0 10 20 30 40 50 1 . i \u00E2\u0080\u00A2 i . i \u00E2\u0080\u00A2 i . i TOC (mg/l) 0 10 20 30 40 50 o -2 6 -Sulfate (mg/l) 0 20 40 60 i i I i I IC (mg/l) 0 10 20 30 40 50 1 i I i l i l i I i l Chloride (mg/l) 0 5 10 15 20 25 1 I H I I I 1 I I I I I I I I I I I I I I I I I I pH 5.8 6.2 6.6 7.0 I ' ' ' i ' ' ' i ' \u00E2\u0080\u00A2 i i Figure 5.3 Combined plots of monthly major ion chemistry, pH, IC and TOC profiles obtained for groundwater at Brookswood North study well Figure 5.3 (...cont) 142 to measurement limitations. The pH decrease with depth near the base of the well is discussed in Chapter 7. The TOC profiles are all uniform over depth, with the values being less than 5 mg/l. As with the findings at the Abbotsford well, the TOC measured at the North well may also have originated from the dialysis membrane used in the sampling device. Figure 5.4 presents the monthly groundwater nitrate profiles, and demonstrate clearly that nitrate arrives at the water table during Decmeber 1994. However, unlike the behaviour observed at the Abbotsford study well (Figure 4.6), a striking aspect of the arrival of nitrate at the Brookswood North study well is that the elevated nitrate levels are localised to the region above the low water table level (approximately 9.5 m below ground surface). Figure 5.4 shows that nitrate that has arrived at the water table with recharging waters has not been flushed deeper into the groundwater, but that it remains perched above the older groundwaters. Although data was not obtained for the months of July and August, the profiles obtained during September 1994, and between April and June of 1995, show that the perched body of nitrate contaminated water dissipates with the drop in the water table. These monthly plots also demonstrate the development and persistence of a steep chemical gradient between a depth of 9 m and 10 m below the ground surface. Below this depth the nitrate levels remained low (less than 5 mg-N/1) throughout the monitoring period. While the observed steep gradient appears similar to those displayed in several of the profiles presented in Chapter 2, the temporal variation associated with the overlying body of nitrate contaminated water is not. This aspect wil l be discussed further in Chapter 7. Nitrate (mg-N/1) Sept7 '94 Sept 27'94 Oct'94 0 10 20 30 40 o 10 20 30 40 0 10 20 30 40 I . i . i . i 2 -4 -6 8 10 12 -14 -16 . i i i i i i Nov '94 Dec '94 Jan '95 Feb '95 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 0 1 D CO C zs 2 O o Q. Q 2 -4 -6 8 10 12 -14 -16 -i , i i i i i _ i i i i i I i Mar '95 Apr '95 May '95 Jun '95 0 10 20.30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 2 -4 -6 -8 -10 -12 -14 -16 -I i i i I i I i I i Figure 5.4 Monthly groundwater nitrate profiles at Brookswood North study well 5.2.1.2 South Study Well Figure 5.5 presents a series of combined plots of major ion, pH, TOC and IC data obtained between July 1994 and June 1995 at the South study well. Detailed plots of the profiles for each month are contained in Appendix F. Unlike at the North well, elevated concentrations of any of the measured constituents was not encountered at, or below, the water table at the South well. pH profiles show similar variation as the North well, with a decrease over depth near the bottom of the well. The measured TOC values at the South study well were generally larger than at the North well, and also exhibited more scatter. This well was also associated with a consistently larger accumulation of fines by the end of each monitoring event than at the North well, or the Abbotsford well (which had only slight fines accumulation throughout most of the monitoring period). The higher TOC values at the South well are possibly due to degradation of the membrane by microbes associated with the fines. 5.2.1.3 Comparison to available regional data Monitoring of the groundwater in the Brookswood aquifer has been less extensive in comparison to the Abbotsford aquifer. Findings reported in Piteau Associates (1994) and Gartner Lee (1993) show that no more than one or two dozen wells have been monitored over this aquifer. This monitoring has shown that localised zones of elevated nitrate concentrations are present, and this is also suggested by the mixed land use over the aquifer. Based on estimates of total nitrogen leaching from both residential (septic drain field and lawn fertilizers) as well as rural areas (poultry manure), Piteau Associates 145 Nitrate (mg-N/1) Sulfate (mg/l) Chloride (mg/l) 0 10 20 30 40 50 0 20 40 60 0 5 10 15 20 25 O CO 6 -8 -10 -12 -14 -16 18 J \u00E2\u0080\u00A2 i \u00E2\u0080\u00A2\u00E2\u0080\u00A2 i II TOC (mg/l) 0 10 20 30 40 50 I . I ' i i i i i . 1.111111111111111111111111 IC (mg/l) pH 0 10 20 30 40 50 5.8 6.2 6.6 7.0 i i ' i i \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 i . . . i Figure 5.5 Combined plots of monthly major ion chemistry, pH, and TOC profdes obtained for groundwater at Brookswood South study well 146 Calcium (mg/l) Magnesium (mg/l) 0 20 40 60 0 5 10 15 o D providing direct confirmation of the denitrifying activity within the columns. 6.6 RESULTS O F SET 6 TESTING The results of the Set 6 column testing are presented in plot form on Figure 6.2, which includes the effluent IC, N03\" and N02\" variation with time, as well as the percentage change in dissolved 0 2 and N 2 between the influent and effluent. The values of percentage change in dissolved gas peak areas were calculated as moving averages 20 -r 0 10 20 30 TIME (days) Figure 6.2 Effluent characteristics measured during Set 6 laboratory column testing 164 (forward-back) of values obtained during three consecutive days. The relative scatter exhibited by the effluent IC values for the 0.75, 1.0 and 1.25 ml/min columns is associated with the detection limit of the analyzer. Visual inspection of the columns showed no evidence (such as the \"dry patches\" reported by Soares et a l . , 1991) of gas bubble formation and accumulation within the sand columns. However, some bubbles did form within the larger pore spaces in the gravel at the base of the 0.5 ml/min and 0.75 ml/min flow rate columns - with fewer bubbles noted in the 0.75 ml/min column. The results shown on Figure 6.2 provide direct evidence for denitrification, on introduction of carbon, in all four columns: a reduced level of N0 3 \" , and an increased level of IC and dissolved N 2 in the effluent. The largest reduction in nitrate is associated with the slowest flow rate (and, therefore, longest residence time) column. Complete nitrate removal was not achieved under the imposed test conditions. In contrast to the column tests (Sets 1, 2, 3) in which methanol was used as the carbon source, the columns in Set 6 show that denitrification commences very soon after the introduction of ethanol. N0 2 \" was also measured well above detection limits in the effluent of the 0.5ml/min column. It is noted that the N0 2 \" minima at day 11 is associated with IC and dissolved N 2 maxima, as well as a N0 3 \" minima. The subsequent increase in effluent nitrite is associated with a decrease in effluent IC as well as dissolved N 2 , in accordance with the partial reduction of nitrate to nitrite depicted by equation 2.5. The dissolved N 2 and 0 2 plots exhibit several important features. Firstly, the onset of denitrification is associated with a sharp reduction in the dissolved 0 2 levels. A pronounced increase in the effluent dissolved N 2 levels is observed only after day 9, 165 when the effluent 0 2 levels are zero in all columns, confirming the anoxic nature of the denitrifying population. The second feature of the plots is the reduction in dissolved 0 2 levels through the columns even prior to the introduction of carbon on day 7. This behaviour is noted, at an increased level as expected, even after the removal of the carbon on day 23 and by which time the microbial population in each column has presumably increased substantially. The most important feature, in the context of the study objective, is the change in dissolved N 2 . The percentage change values exhibit peak values at, or above, 50% for the 0.5, 0.75 and 1.0 ml/min columns. It is also noted that the peaks \"breakthrough\" at consecutively later times with increasing flow rate. If it is assumed that the dissolved N 2 level in the influent is approximately 15 mg/l (at room temperature and in equilibrium with dissolved 0 2 ) , then a 50% increase indicates that the breakthrough peak value is 22.5 mg/l. However, the saturation value for N 2 in the absence of 0 2 is approximately 19 mg/l (approximately 30% increase in GC peak area), so the measured breakthrough peaks represent effluent supersaturated with N 2 . For the 0.5 ml/min and 0.75 ml/min columns, the subsequent decrease over time in effluent dissolved N 2 levels may be associated with the onset of bubble formation, and subsequent coalescence, at the column bases. The bubble formation within the gravel pore spaces may have been triggered by a decrease in the interstitial flow velocity across the sand-gravel interface (due to an increase in porosity and pore size). It is noted that for the 0.5 ml/min column, the effluent dissolved N 2 continues to decrease with time between days 10 and 20, a period which also is associated with an increase in nitrite production. For this column, incomplete nitrate dissimilation also explains the decrease in effluent dissolved N 2 after THIS PAGE BLANK 167 the breakthrough peak. The temporally consecutive trend observed for the N 2 breakthrough peaks may have resulted due to a related lag in microbial population growth, and total denitrification capacity, between the columns at different flow rates. 6.7 C O M P A R I S O N TO R E L A T E D PUBL ISHED FINDINGS 6.7.1 Ethanol versus Methanol as Carbon Sources The study findings show that under the imposed test conditions, ethanol more rapidly promotes denitrification during saturated flow through soil, than does methanol. The findings of Set 2 (Appendix G) show that, even after 60 days, little denitrification was promoted by methanol. In contrast, the Set 6 tests show that denitrification commenced within 2 days of ethanol addition under similar test conditions. These findings are consistent with the findings recently reported by Christensson et al. (1994). While the acclimatization period associated with methanol, in comparison to other carbon sources, during denitrification of wastewaters is well known (McCarty et a l . , 1969), Christennson et al. (1994) describe what is possibly the first direct and detailed comparison of methanol and ethanol as carbon sources for denitrification. These investigators found from chemostat studies that under completely mixed conditions up to 20 days were required for significant denitrification to commence with methanol; while, in the ethanol amended chemostat, denitrification commenced immediately and was completed within approximately two weeks. These investigators were unable to explain the reason for their finding of a preferential utilization of ethanol. The buildup and persistence of nitrite during the Set 6 tests with ethanol is also consistent with the findings of Christensson et al. (1994); who observed a buildup of nitrite in their chemostats during denitrification with ethanol, but no nitrite buildup in the methanol amended chemostat. These findings suggest that, when ethanol is provided as the carbon source, there may be an \"acclimatization period\" associated with the activation of the enzymes responsible for the reduction step of N0 2 \" to N 2 along the nitrate reduction pathway (see equation 2.5 and associated discussion in Section 2.3.1.1). 6.7.2 Nitrogen Gas Generation and Accumulation Table 6.1 presents a summary of a literature review of pertinent findings on the generation and accumulation of nitrogen gas during attached growth denitrification under saturated flow conditions. In the context of the column tests, it is of some relevance that all of the systems noted in Table 6.1 (with the exception of the rotating fixed film reactor used by Harremoes et a l . , 1980, and the groundwater environment investigated by Smith et a l . , 1991) were operated under essentially a one dimensional flow regime with a constant substrate (carbon and nitrate) source at an \"inlet\" boundary. The subsequent flow away from this boundary may be longitudinal (as for the columns), or radial (as for the injection wells). In either case, prolonged operation under this condition is expected to lead to a concentration of microbial activity, and the consequent accumulation of biomass, at, and near, the inlet end. The rate and quantity of biomass accumulation wil l depend upon the substrate loading rate as well as the cell yields, assuming that the other growth factors are present in sufficient quantities. As found experimentally by Young and McCarty (1969), and also inferred theoretically by Bouwer and Cobb (1987) and Widdowson et al. (1988), the localization and rate of biomass accumulation (and 169 therefore substrate removal) wil l also depend upon the initial distribution of the microbial population. Young and McCarty (1969) noted a significant difference in behaviour between a column initially receiving a heavy seed of sludge at the inlet end, and one initially receiving a light seed of sludge uniformly supplied throughout the column length. Maximum substrate removal was reached quickly for the inlet seeded column, while the corresponding time required for the light and uniformly seeded column was much longer. In general, the accumulation of biomass may occur either as floes within the pore spaces, as observed by Young and McCarty (1969), or result in the development of microcolonies or biofilms on the porous media surfaces. Related studies performed by Vandevivere and Baveye (1992) show that the form of colonization may also be bacterial strain specific. Bouwer and Cobb (1987) showed that the localization of denitrification activity (and therefore N 2 evolution) wil l be controlled by the rate of denitrification, which too wil l be bacterial strain specific. In regions where the microbial population consists of obligate oligotrophs (preferring low nutrient environments) or where the microbial population is sparsely distributed, Bouwer and Cobb (1987) showed that, under one-dimensional flow, the rate of nitrate reduction is gradual and extends some distance away from the inlet boundary. The localization of substrate utilization, and accumulation of biomass, at the inlet end is a feature common to most of the systems noted in Table 6.1. Where large reductions in nitrate concentration have been found to occur across this inlet area of activity, nitrogen gas accumulation has also been observed, or inferred, immediately downgradient of this area (Hiscock et a l . , 1991; Jepsen et a l . , 1992; Soares et a l . , 1991). On this basis, it could be presumed that the N 2 is being generated so rapidly, and in 170 sufficient quantity, so as to supersaturate the surrounding porewater and enter the gas phase before any reduction in concentration is affected by the inflowing waters. It is also possible that the gas accumulations, where observed or inferred, in the reviewed systems may be associated with biofilm development - as in the case .of the fixed film reactor testing performed by Harremoes et al. (1980) and also the floes observed by Young and McCarty (1969). Since most of these systems were seeded with denitrifyers (in the columns, and during subsurface injection of wastewaters) it is possible that a concentration of biomass at the reactor inlet could have been experienced by other investigators, and which led to the development of biofilm in the respective systems. The column testing performed as part of the current research was associated consistently with the commencement of gas accumulation (inferred from \"dry patch\" formation) near the base of the sand columns. In the columns in which no topsoil was added (Column 12, Set 5), this can be explained by a gradual decrease in oxygen levels through the column, and the subsequent initial development of anaerobic conditions (favoured for denitrification) at the column base. Where topsoil was added and where dry patches were observed (Set 1, and columns 10 and 11 of Set 5) it is again possible that the rate of oxygen reduction was still sufficiently gradual over the column length. The inoculation method used by Soares et al. (1991), in which topsoil was only introduced to the sand surface after column packing, may have resulted in a high microbial density at the surface. Oxygen consumption may have been rapid across the inlet area, and the conditions for denitrification may have initially developed immediately below the sand surface, where Soares et al. observed the commencement of dry patch development. Soares et al. did not report on the quantity of topsoil that was used in their 171 inoculation procedure. In the current research, the amount of topsoil used was less than 10 g/column (with the total column sand weight per column being greater than 1500 g). Though not quantified, the subsequent mixing of this comparatively small amount of topsoil with the sand prior to placement may have resulted in a sparse microbial population distribution through the columns. Harremoes et al. (1980) showed, using biofilm kinetics, that the diffusional resistance across a biofilm has a significant role to play in the fate of the N 2 that evolves as a result of denitrification within the biofilm. The role of the biofilm in trapping and accumulating the generated N 2 , and as clearly demonstrated in the experiments performed by Harremoes et al. (1980), may explain the gas accumulations observed or inferred by Soares et al. (1991) and other investigators. Rittmann (1993) has suggested that diffusion resistance wil l not be significant in groundwater environments devoid of continuous biofilms. In this context, the assumption, as made by Ronen et al. (1989), of gas bubble evolution from the surface of a microcolony (as opposed to bubble development within a biofilm) is an aspect that requires further investigation. It is possible that gas bubble formation under enhanced denitrification conditions will only occur after the microbial population density has increased sufficiently to form biofilms, or under conditions of low flow velocity with high substrate loading, and a high denitrification rate. 6.8 CONCLUSIONS The packed soil column testing is considered to have provided the first known set of data in which denitrification under saturated flow has been conclusively identified by 172 the direct measurement of dissolved N 2 and 0 2 . These measurements have also shown that, under the test conditions, denitrification will only proceed after dissolved oxygen levels decrease to levels at, or near, zero within the flowing porewaters. The column testing has found that ethanol is more immediately utilized than methanol by the denitrifying population present within the packed sand columns. The testing also found that, under similar flow and packed sand conditions, ethanol also results in a more rapid reduction of nitrate through the columns. While the column testing has confirmed the general feasibility of carbon additions to enhance denitrification, it has not, led to a definitive insight into the potential for clogging of aquifer media during in-situ enhanced denitrification. Based on these findings, and the review of the literature on nitrogen gas evolution in wastewater treatment systems, the following speculative conclusions are reached. Nitrogen gas bubble formation and accumulation does not necessarily occur immediately at the onset of N 2 evolution. Supersaturation of the pore fluids with dissolved N 2 can occur. The study findings suggest that, under sufficient flow rates, this supersaturated fluid may be transported and subsequently diluted without any of the N 2 entering a gas phase and forming bubbles. Under conditions which favour the localization of denitrifyer activity (such as injection wells; or stationary reactive barriers such as the systems discussed by Devlin & Barker, 1994), and where biofilm formation occurs, bubble formation and accumulation may be inevitable. These bubbles have the potential to enhance the rate and extent of clogging that may have otherwise occurred due to biomass accumulations alone. In this context, pulse injection of substrates, which has the potential to minimize biomass accumulation (Duthy, 1993; Taylor & Jaffe, 1990c), may 173 also offer a method by which gas bubble accumulations can be avoided. More research needs to be done before general quantitative models of nitrogen gas bubble formation and subsequent clogging of porous media, under enhanced denitrification conditions, can be developed. TABLE 6.1: SUMMARY OF ATTACHED GROWTH SYSTEMS REVIEWED REFERENCE A T T A C H E D GROWTH SYSTEM REMARKS Bengtsson & Annadotter (1989) Hanaki & Polprasert (1989) Harremoes et al. (1980) Hiscock et al. (1991) Jepsen et al. (1992) Intact saturated cores of sandy aquifer material receiving nitrate ammended filter sterilized groundwater in downflow mode. Upflow filter system of 2cm PVC pipe pieces (porosity 0.82) receiving nitrate + methanol solution Transparent fixed film reactor receiving nitrate + methanol solution Unconsolidated aquifer environments receiving nitrate contaminated groundwater amended with carbon Proprietary \"Biocarbone\" system treating acetate amended nitrified wastewater in downflow Measured dissolved N 2 in effluent correlated to N0 3 \" reduction. N 2 bubble formation not found or inferred. A l l N 2 produced found to have evolved out of top of column, as collected and measured. Distinct and extensive bubble formation at rear of biofilm Gas bubbles inferred to have led to clogging that was experienced during injection of wastewaters Gas bubble accumulation inferred to be occurring below clogged surface layer of columns T A B L E 6.1 (cont..) REFERENCE ATTACHED GROWTH SYSTEM REMARKS Lance & Whisler (1972) Oberdorfer & Peterson (1985) Polprasert & Park (1986) Columns of loamy sand intermittently flooded with secondary sewage effluent Aquifer formations of fine to medium carbonate sand, and cemented reef rubble receiving injected secondary effluent Anaerobic filters of crushed stone in upflow mode to treat effluent from waste stabilization basin Bubbling gas collected after 20 days of flooding found to be 98% N 2 . Associated sharp decrease in flow rate through column inferred to have resulted due to gas bubble accumulation Gas bubble formation inferred from -measured increase in dissolved nitrogen gas away from wells, as well as associated steep head loss \"Bubbling of gaseous nitrogen\" observed only at influent N0 3 \" concentration greater than 400 mg/l Savage & Chen (1975) Proprietary \"Denite\" filters (anaerobic) treating clarified secondary effluent in upflow mode \"Vigorous release of nitrogen gas bubbles sufficiently violent to boil the entire surface of filter\" occurred during filter \"bumping\" (backwashing) T A B L E 6.1 (cont..) R E F E R E N C E ATTACHED GROWTH SYSTEM REMARKS Smith et al. (1991) Scares et al. (1991) Sand & gravel aquifer containing an \"Excess\" levels of dissolved N 2 in extensive plume of nitrate contamination groundwater associated with areas of from continuously discharged secondary sewage effluent Sand columns operated in downflow mode receiving nitrate + formate solution . reduced N0 3 \". N 2 bubble formation not found or inferred Visible gas bubbles reported under continuous flow and denitrification Tucker et al. (1974) Yamaguchi et al. (1990) Anaerobic filter of medium sand operated in upflow mode, receiving nitrate + ethanol solution Decomposed granite and sand (< 5mm) columns operated in upflow mode K N 0 3 + methanol solution Expansion of sand bed on denitrification considered to have resulted due to gas accumulation Evolved gas from top of columns reported to have higher N 2 /0 2 ratio than influent, but no other impact of N 2 gas appears to have been encountered Young & McCarty (1969) Anaerobic filter of smooth quartzitic gravel (1\" to 1.5\") operated in upflow mode to treat synthetic wastewater Biomass in filter present as floes within porespaces. Gas bubbles visibly attached to floes, and which rise and release bubbles on contact with overlying stone surface. Bubbles continue to float upward and out of filter 7.0 DISCUSSION OF RESULTS 177 The various results presented in Chapters 4, 5 and 6 are discussed together in this Chapter, while maintaining a focus on, non-point source nitrate, under the following general headings: i) Groundwater chemistry as related to surficial land use ii) Temporal variations in groundwater chemistry iii) Fate of nitrate below the water table iv) Role of carbon additions for promoting denitrification within aquifer environments. Since this research also involved the use of the relatively new and little used dialysis membrane technique for passive sampling of groundwater dissolved constituents, a discussion on the efficacy of this technique is also presented. 7.1 G R O U N D W A T E R C H E M I S T R Y AS R E L A T E D TO S U R F I C I A L L A N D U S E The findings from the Brookswood North and Abbotsford study wells both provide direct evidence of surface infiltration contributions to the groundwater chemistry at the two sites. The findings from the Brookswood South study well, which is located downgradient of a residential area, are less clear. The TDS (total dissolved solids) values associated with the measured groundwater chemistry at the Brookswood South well are less than 100 mg/l over the full depth of the well. Using the classification system described by Freeze & Cherry (1979) for natural 178 groundwaters in glacial deposits, the waters encountered at this well can be classified as Type I (slightly acidic, very fresh waters, TDS < 100 mg/l, soft to very soft). While the waters encountered within the lower half of the Brooksood North well, and near the base of the Abbotsford well also generally fit this classification, it is clear that the chemistry near the water table at these two wells does not. The composite profiles presented on Figure 4.6 show that at the Abbotsford study well recharge waters resulted in a significant change in the groundwater chemistry over a depth of at least 8 m below the water table during the monitoring period. The composite profiles presented on Figure 5.3 for the Brookswood North well show that recharge waters resulted in a significant change in the groundwater chemistry over a depth of approximately 3 m below the water table during the same monitoring period. As the Brookswood South study well is located upgradient of the North well within the same aquifer, a comparison of the chemistry associated with the deeper groundwater encountered within the two wells can be made. Table 7.1 presents a comparison of the groundwater chemistry at a depth of 12 m (below ground surface) within the Brookswood North and South study wells. The 12 m depth is below the zone found to be impacted by annual recharge waters at the North well. The groundwater chemistry at the 12m depth is found to be similar between the Brookswood North and South Wells, with the exception of N a + and S0 4 2~, both of which are present in higher concentrations at the South well. Figure 7.1 presents a schematic composite vertical section showing three well locations (X, Y , Z) in relation to a spatially varying surficial land use over a recharge area of an unconfined aquifer. Wells X and Y can be considered analogous to the Table 7.1 Comparison of GW chemistry at 12 m depth in Brookswood wells Parameter North South (mg/l) Well Well pH 6.1 - 7 6.1 - 6.7 N O 3 - N < 5 < 5 S0 4 2 + < 10 10 - 20 cr 3-9 5 - 12 IC 5 - 10 6 - 12 C a 2 + 5 - 10 6 - 12 M g 2 + 2 - 4 3 - 4 N a + 3-4 7- 10 K + < 3 < 3 ; agricultural with | fertilization j and irrigation forest - X - residential with septic drain fields agricultural with fertilization and Irrigation WELL X WWWwww WELL Y A A A WELL Z wwwwwwwwwwwwwwwiw \u00E2\u0080\u0094 water table nftratg cortarrfnated region groundwater flow Figure 7.1 Schematic vertical cross section illustrating possible relationship between land use and underlying groundwater chemistry in a recharge zone of an unconfined aquifer 181 Brookswood North and South wells respectively, while Well Z is analogous to the Abbotsford study well. The nitrate contaminated regions as shown on Figure 7.1 have been deduced by considering the surface loading geometries and an assumed recharging flow regime for the underlying groundwater. As the upgradient extent of agricultural land use at the Brookswod North study well is only 100 m, the shallow depth of the nitrate contamination below the water table as encountered at this well can therefore be explained on the basis of a limited upgradient extent of a surface nitrate source. At the Brookswood North study well, the surface sources of nitrate include the annually applied manure and fertilizer, as well as any nitrate that may have leached from the manure stockpile temporarily placed upgradient of the well (Figure 3.3). The deeper extent of nitrate contamination encountered at the Abbotsford study well can be considered to have resulted due to the greater upgradient extent of the nitrate source. Although plumes of nitrate are shown (Figure 7.1) in the vicinity of Well Y , no plume has been intercepted by the Brookswood South study well - which was primarily located to intercept a plume emanating from the residential septic tank drain field located to the south of this well (Piteau Associates, 1994). Generally, the groundwater chemistry profiles obtained from the three study wells validate this idealised deduction. The higher N a + and S0 4 2~ values encountered at the Brookswood South well are possibly due to the releases from the drain field, and may be representative of the overall chemistry of the groundwaters flowing beneath the residential lands. While the North well is not directly downgradient of the South well along the assumed groundwater flow path at this site, the concentrations of both N a + and S0 4 2 \" in the groundwater originating beneath the residential area may be decreasing (as a result of dispersion, or exchange 182 processes) along the flow path to the levels measured at the North well. Given the relatively pristine nature of the waters encountered at the Brookswood South well, this well could be considered a \"control\" for further discussion. The following discussion wil l also focus on the Abbotsford study well, as the extent of nitrate contamination is more significant and better defined at this well. 7.1.1 Source of Leached Nitrate and Associated Chemistry The almost total reliance on poultry manure as a fertilizer and soil conditioner in the vicinity of the Brookswood North and Abbotsford study wells, and the accumulation of published evidence showing the extent of nitrate leaching under poultry manure amended soils, point to this manure being the source for the leached nitrate arrivals observed at the study wells. Findings of large losses of nitrate beneath the root zones of crop lands receiving poultry manure have been reported by Adams et al. (1994), Cooper et al. (1984), Liebhardt et al. (1979), Sims (1986) and,Weil et al. (1979). For the Abbotsford aquifer, recent findings by Wassenaar (1994) show relatively conclusively, on the basis of N and O isotope quantification, that the groundwater nitrate in the aquifer originated primarily from an organic source (presumed to be poultry manure), and to a lesser extent from applications of inorganic nitrogen based fertilizers. Over the study region, recent investigations by Chipperfield (1992, 1994), the Sustainable Poultry Farming Group, SPFG, (1994) and Zebarth et al. (1994) have led to the quantification of the nutrient content of poultry manure, the leachate generated from stockpiled manure, and the temporal variations in the N content within the crop (raspberry in particular) root zone in manure amended fields. The findings from these 183 investigations are useful in evaluating the observed temporal variation in the groundwater chemistry at and below the water table at the Abbotsford study well: a) The nitrogen content of poultry manure at the time of application is present mainly in organic and ammonium form. Figure 7.2 shows the nutrient content of manure from various poultry types as reported by the SPFG (1994). b) On manure application, typically between February and Apri l , the ammonium is rapidly (within days to weeks) converted to nitrate, the extent and rate of conversion depending upon the time of manure incorporation into the soil (Chipperfield 1992; SPFG, 1994). c) Soil nitrate-N content increases steadily during the growing season as a result of ongoing mineralization of organic-N (residual and from poultry manure), and due to a comparatively low N requirement by the raspberry crops. Soil nitrate-N levels as high as 350 kg/ha have been measured in late October at the end of the growing season. Between October and February, soil nitrate-N disappears. Figure 7.3 shows plots of the temporal variations in soil nitrate-N, to a depth of 0.6 m, during the growing season and under different rates of manure application as reported by Zebarth et al. (1994). The findings presented on Figure 7.3 have been inferred to indicate a total reduction in the soil nitrate-N content within the upper 0.6 m of the land surface between October and February as a result of leaching (Zebarth, 1995). There is currently no data from the study region on the nitrate-N content, or the general chemistry, of porewaters within the unsaturated zone beneath croplands fertilized by poultry manure. Reports by Paul & Zebarth (1992, 1993) show that denitrification within the upper 0.6 m (root-zone) 184 V. Average Nutrient Content for Commercial and Hatching Egg Layer, Tun<*y, and Chicken Production Flock* f% by weight) 7 Total Ammonium Phosphorus Potassium Calcium Magnesium Nitrogen Nitrogen Nutrient Figure 7.2 Average nutrient content (% by weight) for commercial and hatching egg layer, turkey and chicken production flocks (SPFG, 1994) 185 History of manure use, 1992 Layer manure applied (kg total N/ha) 0 100 200 No history of manure use, 1992 F M A M J J A S O 300 250 200 150 100 50 0 No history of manure use, 1993 1 r \u00E2\u0080\u00A2 \u00E2\u0080\u0094 i 1 i F M A M J J A S O Figure 7.3 Influence of rate of poultry manure application on soil nitrate-N to 60cm during the growing season in three experimental plots (Zebarth et al., 1994) 186 can also partly explain the reduction in soil nitrate-N content. These investigators measured high levels of denitrification during the growing season (March to September) also. The ongoing accumulation of soil nitrate-N as shown on Figure 7.3, as well as the arrival of nitrate at the water table in high concentrations as found in this study, suggest that the organic nitrogen mineralization rate combined with the leaching rates currently exceed the rate of root-zone denitrification in the study region. Zebarth et al. (1994) estimated that, for the Abbotsford aquifer, the leaching of 200 kg-N/ha out of the root zone could result in the equivalent of 20 mg/l nitrate-N in the underlying groundwater! This value appears to have been computed by assuming a value of 1 m for recharge, with the recharge water containing the computed 20 mg/l nitrate-N. While the quantities of manure applied at the study sites are not known, given the long term history of manure use at and upgradient of the Abbotsford study well, it is presumed on the basis of the descriptions and findings shown on Figure 7.3 that the soil nitrate-N level at the commencement (typically October) of fall precipitation at the site is probably at least 150 kg/ha. On this basis, and assuming total recharge of l m , leaching to the water table wil l consist of the arrival of fronts of nitrate at a concentration of at 15 mg-N/1 or more. The observed fluctuation of nitrate-N levels near the water table at the Abbotsford study well (Figure 4.6) show that this is the case. This is discussed further in Section 7.2.1. The findings shown on Figure 7.3 also suggest that, in the vicinity of the Abbotsford study site, the change in land use (cane removal in September) may not have led to a change to the annual recharge and nitrate leaching cycle at that site. The removal 187 of the canes may or may not have effected the levels of residual soil nitrate. As the soils are rapidly draining, the removal of the canes should not have changed the infiltration quantities. A closer look at Figure 4.5 shows that, while the region associated with cane removal is large, the extent of the effected area actually upgradient (of the groundwater flow) from the study well location is small. On the basis of these considerations, the removal of the canes is not considered further here. With regard to the major ion chemistry associated with the leaching nitrate, it is noted from Figure 7.2 that poultry manure contains a high percentage of K + , C a 2 + and M g 2 + . Table 7.2 presents some previously unpublished data that was obtained by the B C M O E during December 1991, using three samples of leachate runoff from manure stockpiles that were located during that time along 0 Avenue and Hamm Road (Figure 3.2). While the data presented on Table 7.2 shows the expected variability (due to variability in the source, as well as possible localised dilution), it can be used to interpret the detailed groundwater chemistry data collected as part of this study. A direct comparison of Table 7.2 and the groundwater chemistry associated with the nitrate near the water table at the Abbotsford study well (Figure 4.6) shows that CI\", S0 4 2 \" , N a + and K + most probably originated from the land applied poultry manure. But the comparison also highlights several differences between the quality of the leachate and groundwater: a) while the leachate has neutral to basic pH, and has high alkalinity, the groundwater at the water table is acidic with almost no alkalinity, b) the C a 2 + and M g 2 + content of the leachate are lower than would have been expected from the manure constituent data shown on Figure 7.2. On the other Table 7.2 Analysis of leachate runoff from poultry manure stockpiles (BCMOE, 1995) Parameter (mg/l) Sample 1 Sample 2 Sample 3 Alkalinity (total) 14900 1690 1150 pH 7.6 7.8 7.7 C a 2 + 24.7 30 193 M g 2 + 4.6 4.9 14.2 Na + 570 60 41 K + 2430 275 190 N H 4 + - N 4160 498 350 N O 3 - N 5 0 0 S042\" 1620 150 113 ci- 900 102 71 189 hand, the C a 2 + and M g 2 + concentrations near the water table are higher than would be expected from their relative concentrations in the leachate, c) the N content in the leachate is mainly ammonium (NH 4 + ) , while little to no ammonium was detected at the water table. It has already been noted that, in manure amended soils, ammonium is rapidly converted to nitrate. The predominance of ammonium as detected in the runoff leachate may simply reflect a slower rate of nitrification during the winter months when the leachate samples were collected. The nitrification process is mediated by chemoautotrophic microbes, which obtain their carbon requirements from inorganic carbon (IC) compounds in the form of carbon dioxide, carbonates or bicarbonates (Benefield & Randall, 1980). The overall nitrification process is a two step process which consists of the oxidation of N H 4 + to N0 2 \" , followed by the oxidation of N0 2 \" to N0 3 \" . These two reaction steps are mediated by different microbes, but the overall conversion of NIL/ to N0 3 \" can be simply written as shown on equation 7.1. M f 4 + + 202 =\u00E2\u0080\u00A2 NO; + 2H+ + H20 (7.1) The nitrification process is therefore associated with the reduction in IC (or alkalinity, in the carbonate or bicarbonate form) as well as a reduction in pH. The low values of IC and pH measured near the water table are therefore consistent with the arrival of recharge waters containing nitrified manure leachate from the surficial soils. Figure 7.4 presents the profiles of nitrate-N and pH for the months that these two parameters were quantified together on samples from the Abbotsford study well. These profiles all show that nitrate maxima in the vicinity of the water table are associated with pH minima. A similar trend is also observed at the Brooksood North study well (Figure E9, Appendix Figure 7.4 Monthly groundwater nitrate and pH profiles at Abbotsford study well Figure 7.4 (....cont) 192 E). As noted by Freeze & Cherry (1979) rain water itself can have a low pH (approximately 5) as a result of C 0 2 dissolution. pH increases within percolating waters can subsequently occur, for example, as a result of neutralisation during travel through the subsurface soils. The low pH values encountered at the water table of the Abbotsford study well cannot therefore solely be considered a result of nitrification. However, the pH values at the Brookswood site were all typically greater than 6 near the water table, even at the North well where nitrate leaching occured. The shallower water table depth, combined with the nitrification process is considered to maintain the low pH condition of recharging water at the Abbotsford site. The N H 4 + encountered at and below the water table during March 1995 at the Abbotsford studty well is considered to represent incomplete nitrification, as well as reflect the rapid flushing through the unsaturated zone during the winter months. The N0 2 \" encountered at the water table in July 1994, suggests incomplete nitrification of the leached N H 4 + during groundwater flow, but there is insufficient temporal data to support this conclusion. The H + production on nitrification of the poultry manure may explain the elevated concentrations of C a 2 + and M g 2 + associated with the leaching nitrate. Cahn et al. (1993) have reported that the production of protons (H + ) and N0 3 \" , in soils amended with ammonium based fertilizers, can result in the co-leaching of \"base cations\", namely C a 2 + , M g 2 + and K + . The H + ions act to displace these cations from exchangeable sites on soil surfaces into the soil solution. Cahn et al.,(1993) also note that the requirement of charge neutrality in leaching solutions can lead to the release of cations into soil 193 porewaters initially containing an anionic charge imbalance. An alternate hypothesis for the arrival of major cations, and in particluar C a 2 + , to the water table is mineral dissolution within the unsaturated zone as a result of infiltrating acidic waters. Freeze & Cherry (1979) describe the potential for C 0 2 generation due to oxidation of residual organic matter within the upper soil zone, and the subsequent infiltration of carbon dioxide-charged water through the soil zone. However, the Brookswood South well encountered no elevated levels of major cations (or anions) near the water table during the monitoring period, even though the surface condsitions near this well are similar to those at the North study well and the Abbotsford study well. The difference in water table chemistry at these wells and the Brookswood South well is therefore concluded to be solely due to the application of poultry manure, which is a source of not only readily teachable major ions (Table 7.2), but also a significant amount of organic matter. On the basis of the preceding analyses, the study findings suggest that the major ions encountered near the water table in the Abbotsford study well, as well as the Brookswood North study well, originated from the surface applied poultry manure, with possibly some leaching from soil surfaces within the unsaturated zone. The exception to this generalization is the difference in the observed levels and distribution of K + and S0 4 2 \" between the Abbotsford and Brookswood North study wells. 7.1.2 Relative Distribution of K + and S042\" The water table chemistry at the Brookswood North study well (Figure 5.2) also shows recharging N0 3 \" to be associated with CI\", C a 2 + , M g 2 + and N a + . However, the K + 194 and S0 4 2 \" content near the water table at this well does not increase along with CI\", C a 2 + , M g 2 + or N a + during recharge. A closer look at the K + and S0 4 2 \" profiles for the Abbotsford study well (Figure 4.6) shows that these two constituents have similar profiles in which their concentrations decrease more rapidly with depth below a sharp peak. This suggests that K + and S0 4 2 \" may be selectively co-retarded during their transport beneath the water table. The retardation of K + by cation exchange processes is well recognised. Bjerg & Christensen (1993) demonstrated, by performing natural gradient tracer tests in a glacial outwash sand aquifer in Denmark, that K + was substantially retarded in comparison to N a + , C a 2 + or M g 2 + . The CEC of the aquifer soils was reported to be approximately 1 meq/100 g. Similar findings of K + retardation have been reported by Dance & Reardon (1983) from their investigations at the Borden landfill site in Ontario, and by Ceazan et al. (1989) from work at the glacial outwash plain aquifer in Cape Cod. The soils at the Cape Cod and Borden sites are also associated with a low CEC (less than 1 meq/100 g at both sites). The CEC data shows slightly higher values at the Brookswood and Abbotsford study wells, so K + retardation is probably occurring at these sites. Preferential adsorption of S 0 4 2 ; by soils is well recognised. Chao (1964) and Chao et al. (1962, 1963, 1964) found that sulfate adsorption can be dependent on many factors including associated cation and anion concentrations in solution, organic content of the soil, presence of oxide (Fe/Al) coatings on the soil, and pH. Of these factors, pH and oxide coatings appear to be the most important, and greatest adsorption is associated with low pH and the presence of oxide coatings. Chao et al. (1963) measured 10% adsorption at a pH of 6 in their study soils, with an increase in adsorption at lower pH values. On 195 the basis of their findings, Chao et al. (1963) speculated that the possible adsorption mechanism is sulfate attachment to positive charges on oxide surfaces, or on crystal edges of clays, at low pH values; The greater depth of the unsaturated zone at the Brookswood North study well, and the relatively higher CEC values associated with these soils, could cause greater retardation potential at this site. This could explain why temporal K + and S0 4 2~ arrivals and concentration changes near the water table do not mimic those of N0 3 \" , C a 2 + , N a + , M g 2 + a n d C l - . 7.2 TEMPORAL VARIATION IN GROUNDWATER CHEMISTRY It has been concluded that infiltrating winter precipitation and subsequent recharge is responsible for the delivery of surface leached chemicals to the underlying water table at the Abbotsford and Brookswood North study wells. This Section discusses in more detail the temporal changes observed in the chemical profiles beneath the water table at the Abbotsford and Brookswood North study wells. As the groundwater at the Brookswood South well exhibited no change in chemistry during the monitoring period, it wil l not be discussed. While the following discussion refers only to the nitrate profiles, it is applicable to all of the other constituents associated with the nitrate, with the exception of pH and IC which will be discussed separately (Section 7.2.3). 7.2.1 Role of Recharge Pattern The precipitation patterns and associated fluctuations in the water table within the study aquifers have been used to conclude that winter precipitation contributes to the 196 recharge at the study sites. The rapidly draining nature of the gravelly sands overlying the water table result in infiltrating precipitation causing a rapid response in water table fluctuation. Figures 4.2 and 5.2 indicate that the water table rises almost immediately after the commencement of winter precipitation in October. Viswanathan (1984) found from detailed studies of the recharge characteristics of an unconfined fine to medium grained sand aquifer in Australia that significant recharge can occur on the first day of rainfall. Figure 7.5 presents a temporal plot of nitrate-N versus depth at the Abbotsford well. Only values at and below a depth of 5.9 m from the ground surface are presented for clarity. While this depth does not represent the top of the water table, it is the closest depth for which continuous data was measured over the twelve months of monitoring. The plots on Figure 7.5 show that nitrate levels commence increasing by November soon after the onset of winter precipitation. The overall increase in nitrate-N levels at the water table between October and March is approximately 15 mg-N/1. The nitrate levels show the greatest fluctuation near the water table, with smaller fluctuations occurring at depths over 6 m below the water table during the recharge cycle (Figure 7.5). While the observed magnitude of the nitrate concentration in the recharge waters is consistent with the presumed levels of residual soil nitrate level at the Abbotsford site, it should be noted that the residual soil nitrate level and recharge pattern are not the only factors controlling the concentration of nitrate reaching the water table. While aiming to develop a rational management approach by which nitrate seepage losses during winter months could be estimated by measurement of residual soil nitrate, Van Der Ploeg et al. (1995) showed that the soil type in the unsaturated zone also plays a controlling role. By 197 50 -y Jul '94 Dec '94 Jun '95 Time (months) Figure 7.5 Groundwater nitrate vs time over depth below the ground surface at the Abbotsford study well 198 using a one-dimensional mixing cell model, Van Der Ploeg et al. showed that for a given residual soil nitrate level and recharge rate, groundwaters overlain by soils having high field capacity1 are likely to be less vulnerable to concentrated arrivals of nitrate with the recharge waters. The gravelly sand deposits that are associated with the Abbotsford and Brookswood have a low field capacity, as well as have higher saturated conductivities, so recharging waters move rapidly through these formations. This results in high nitrate at the onset of recharge and the rapid movement of the nitrate as a front to the water table. 7.2.2 Transport Beneath the Water Table As seen on Figure 4 . 7 , the profiles at the Abbotsford well suggest the arrival of a nitrate front which subsequently moves below the low water table level. In contrast, the profiles at the Brookswood North well (Figure 5 . 3 ) show that the zone of elevated nitrate increases in thickness as the water level rises in response to the arrival of the nitrate laden recharge waters, but this zone does not move below the low water table level. It is intuitive that the water table rise and fall associated with this seasonal recharge pattern will affect a temporal variation in the flow trajectory at and below the water table. A search of available literature shows that this phenomenon has not been previously investigated or quantified, so there is no analytical approach currently available towards interpreting, on this basis, the vertical profiles of groundwater 1 the moisture content that must be exceeded during wetting before free drainage occurs 199 chemistry as measured at the study wells. In order to describe the observed temporal variations in the profiles, a conceptual model describing the temporal variations in the flow trajectory beneath a seasonally fluctuating water table has been developed,\" and is presented in Figures 7.6 and 7.7. These figures also show the presence of a passive sampling well. The passive sampling approach, unlike a pumped approach, does not induce any changes in the groundwater flow paths in the vicinity of the well. Figure 7.6a shows the groundwater flow paths beneath the water table during the low water table period when no recharge is occurring. Flow under this condition wil l be horizontal over the full depth of the aquifer (assumed to be homogeneous and isotropic). Beneath a uniform non-point surface source, temporal variations in the chemistry profiles measured at the well location during this time would be expected to be minimal. This is supported by the profiles measured between August and December 1994 at the Abbotsford well (Figure 4.7), although the disappearance of the nitrate gradient at a depth of 13 m (below ground surface) is noted to have occurred over these months. The rising water table wil l result in the development of vertical hydraulic gradients in the vicinity of the water table, and consequently a vertical component to the groundwater flow as shown on Figure 7.6b. The front of leached nitrate, which arrives at the water table during this time, will therefore continue to move downward to some depth before being transported in the horizontal direction. The profiles shown on Figure 4.7 for the months of February through to May 1995 show the arrival and downward movement of a nitrate front. These profiles also show that a decrease in concentration is associated with the downward movement of the nitrate front. This decrease in Figure 7.6 Conceptual model showing temporal changes in groundwater flow trajectories in response to fluctuations in water table at Abbotsford study site: a) October, b) February, c) June, d) August g 201 concentration can be explained on the basis of longitudinal dispersion in the direction of flow (which is initially downward). Figure 7.7 illustrates the downward movement and dispersion of a single front of leached nitrate on arrival at the water table. Figures 7.6c and 7.6d show the transition back to a totally horizontal flow regime during the period of water table decline! As part of this conceptual model (Figure 7.6), the increase in flow volumes during winter recharge is considered to result in an increase in the velocities along the flow path. On this basis, the groundwater flow velocities would be expected to be at their lowest value just before the onset of recharge (Figure 7.6a). The velocities would be expected to be at their highest value during the months of recharge and subsequent water table decline (Figure 7.6b to 7.6d). On the basis of this reasoning, it can be concluded that nitrate (and other dissolved constituents) that is delivered to the water table by recharge waters wil l subsequently be transported through the groundwater in a pulsed manner in response to temporal variations in groundwater flow velocities. Such a pulsed behaviour in unconfined aquifers was previously speculated by Staver & Brinsfield (1991), though they did not have any data to support their speculation. While actual groundwater velocity measurements were not made during the current research, the conceptual flow model developed in Figures 7.6 and 7.7, and the measured profiles of groundwater chemistry provide further support for the Staver & Brinsfield (1991) speculation of pulsed behaviour. As developed on Figure 7.6, the temporal variation in the flow trajectory beneath the water table is considered to extend over only a portion of the aquifer thickness at the well location. The well shown on Figure 7.6 will encounter, at depth, groundwaters not < Figure 7.7 Schematic description of downward movement and dispersion of leached nitrate front below water table o ro 203 immediately affected by the seasonal variations in chemical loading and distribution in the vicinity of the water table. The almost negligible change in the chemistry profile measured beneath a depth of approximately 13 m at the Abbotsford well provides support for this aspect of the conceptual model. The monthly concentration profiles for the Abbotsford well (Figure 4.7) show tha the downward movement and distribution of the winter leached nitrate occurs rapidly and extends to a depth of nearly 8 m below the water table. This is clearly seen by the disappearance (by October) and subsequent reappearance (by May) of a nitrate gradient at a depth of 13 m below the ground surface (Figure 4.7). From this observation it can be concluded that the groundwater could be as young as 1 year old to a depth of 8 m below the water table at this site. However, the pulsed nature of the groundwater flow confounds this conclusion. The fluctuation in the nitrate gradient at 8 m could be explained by the pulsed arrivals of upgradient waters to the study well location. On the basis of the conceptual model described on Figures 7.6 and 7.7, the profiles at the Brookswood North study well can be readily explained by the arrival of a nitrate front which extends over only a small upgradient (from the well) portion of the water table. Any subsequent movement of the nitrate below the water table may not be detected at the well location if the flow trajectories beneath the nitrate front do not intersect deeper portions of the well. Figure 7.8 presents a plot of groundwater age below the water table calculated using equation 2.13 for the Abbotsford and Brookswood sites using the precipitation and aquifer depth data applicable to these sites. A porosity value of 0.3 has been assumed for these calculations. The groundwater! age at a depth of 8 m below the water table at the Figure 7.8 Groundwater age vs depth as calculated using eq. 2.13 for probable ranges of P and D at Abbotsford and Brookswood study sites. Porosity assumed to be 0.3 g 205 Abbotsford aquifer has an estimated age range of between 3 and 7 years based on the most probable range (0.5 m to 1 m) of recharge at this site (Figure 7.8a). For the Brookswood wells, the estimated age of between 2 and 5 years at a depth of 3 m below the water table (Figure 7.8b). These estimated ages may be considered to be of the same order of magnitude as the 1 year value implied by the profile evolutions on Figures 4.7 and 5.3. However, this finding also shows that groundwater age determinations using the methodology described by equations (2.9) to (2.13), and as illustrated on Figure 2.21, may overestimate the age in cases where the groundwater flow trajectory below the water table exhibits dynamic sub-annual fluctuations. This finding was anticipated on the basis of the plots on Figure 2.21, and was previously noted in Section 2.3.2. For the Abbotsford study well, the division of the groundwater beneath the water table into an upper zone exhibiting dynamic temporal variation in flow conditions as well as distribution of leached contaminants, and an underlying zone exhibiting a relatively steady state (or less temporally dynamic) condition, can also be inferred by considering the CEC and TEC data obtained for the surrounding soils. Figure 7.9 presents a plot of the soil CEC and TEC data in relation to the groundwater pH and nitrate profiles. The CEC and TEC profiles show that below a depth of 13 m base saturation (TEC/CEC expressed as a percentage) is close to 100%. The zone below 13 m is also associated with a constant and near neutral groundwater pH, as well as a stable and constant distribution in groundwater nitrate (and associated chemistry). Above 13 m, the groundwater chemistry fluctuates, and pH values are acidic (< 7). It is possible that TEC and therefore base saturation values within the surrounding soils also fluctuate over time in response to changes in the groundwater chemistry, as cation exchange processes occur 206 Nitrate (mg-N/1) PH 0 10 20 30 40 50 6.0 6.2 6.4 6.6 6.8 7.0 0.0 0.5 1.0 1.5 2.0 2.5 CEC & TEC (meq/100g) Figure 7.9 Soil C E C & T E C profiles including groundwater pH (October 1994) and combined monthly groundwater nitrate profiles at Abbotsford study well 207 at fast rates (Reardon et a l . , 1983). The TEC/CEC quantifications at the study sites were made on one set of samples collected at the time of drilling, and a temporal variation in TEC contents of the subsurface soils can only be determined from repeated sampling of these soils. The conceptual model as developed on Figure 7.6 and 7.7, and the simplified analysis associated with equation 2.13, are both based on the assumption of a homogeneous isotropic aquifer. While these two approaches have been used in the preceding discussions to explain the observed findings, the abrupt nature of the decrease in temporal fluctuations of the groundwater chemistry below a depth of 13 m in the Abbotsford aquifer, also suggests that there may be a physical control on the groundwater flow over these depths. While a detailed stratigraphy at the soils underlying the study sites was not obtained, the reported finding of till at a depth of 17 m at the Abbotsford well points to the possibility that there may be strata based control on the groundwater flow regime at this site. There is insufficient information available at the current time with which to evaluate this aspect further, and this poses a major limitation in reaching a definitive conclusion with regard to the applicability of the conceptual model. Another assumption associated with the conceptual model shown on Figures 7.6 and 7.7 is that of uniform recharge across the land surface. While this may be applicable to the Abbotsford study site, it may not at the Brookswood site, where the surficial land use changes from agricultural, to forested to residential upgradient of the North study well. r. 208 7.3 F A T E O F N I T R A T E B E L O W T H E W A T E R T A B L E The oxic nature of the groundwaters over the monitoring depths at the study sites, and the relatively low dissolved and solid organic carbon content within the aquifers, suggest that denitrification may play little to no role in the fate of nitrate over these depths beneath the water table. The measured TOC values in the groundwater are less than 5 mg/l, and in fact could be less than 1 mg/l if it is assumed that the cellulose bags may have contributed to most of the measured TOC as discussed in Chapters 4 and 5. The SOC values in the recovered soil samples are generally well below 0 .5%, a value suggested by the findings of Bradley et al. (1992) and Frind et al. (1991) to be a minimum that may be required for denitrification to occur at any significant rates. On this basis, and from the preceding analysis of the transport of nitrate below the water table, the decreases in nitrate concentrations with depth below the water table at the Abbotsford well, and with time at the Brookswood North well, can be explained by considering the flow path development beneath the water table and the subsequent longitudinal dispersion of the nitrate along the flow path (using the analyses described in Section 2.3.2). The steep concentration gradients measured at the Brookswood North well clearly demonstrate that such gradients do not result from only chemical transformation processes (as suggested by previous investigators). As noted in Chapter 2, the spatial variation of the N/Cl\" ratio has been used by other investigators to deduce the occurrence of denitrification along groundwater flow systems. Given the detailed chemical characterization associated with this study, this approach can be used to further investigate the role of denitrification over the monitored depths at the study sites. 209 7.3.1 NOy-N/Cl Ratio at Study Wells 7.3.1.1 Abbotsford Study Well Figure 7.10 presents a detailed comparison of the N 0 3 \" and CI\" profiles that were obtained at the Abbotsford study well. Figure 7.11 presents monthly profiles of the nitrate-N/Cl\" (herein referred to as N/Cl\") ratio for this well. Figure 7.10 shows that the CI\" shows a greater relative temporal variation (even below 13 m) than the N0 3 \" . The N/Cl\" profiles show an associated large temporal variation. The measured N/Cl values near the water table range between 1.4 and 3. Based on the poultry manure leachate characteristics shown on Table 7.2, a value of between 4.5 to 5 would be expected for the N/Cl ratio if it is assumed that all of the nitrogen and chloride present within the manure is leached to the groundwater. The lower measured values can be explained to be a result of N uptake by crops, N volatilization and denitrification within the soil zone. The temporal variation in the N/Cl\" profiles may be explained due to a spatial variation in the N/Cl\" ratio at the ground surface. Despite the scatter and temporal variation associated with the N/Cl\" profiles, they display a decrease in the N/Cl\" ratio with depth below the water table, although below 13 m the N/Cl\" ratio remains relatively constant for any given time of monitoring. In the absence of the dissolved oxygen (DO) and TOC/SOC data, the N/Cl\" profiles, combined with the increasing IC over depth, suggest that heterotrophic denitrification may be occurring beneath the water table at the Abbotsford well site. The disparate nature of the DO/TOC-SOC findings and the N/Cl\" findings from the Abbotsford study well confound analysis of the role of denitrification at this site. The finding of nitrate levels remaining well above 10 mg-N/1 throughout the year over the Chloride (mg/l) Nitrate-N (mg/l) 5 10 15 20 25 30 35 40 45 50 1111iii11 i i ii|i i i i | i iii| i i i i | i i i i | \u00E2\u0080\u00A2\u00E2\u0080\u00A2 i i | \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2\u00E2\u0080\u00A2 | Figure 7.10 Comparison of monthly N03\" and CI\" profiles obtained from Abbotsford study well o Figure 7.11 Monthly groundwater Nitrate-N/Cl profiles from Abbotsford study well 212 monitoring depth suggests that any denitrification activity that may be occurring is small in comparison to the rates of nitrate delivery and movement through the groundwater at this site. Wassenaar (1994) reached the same conclusion on the basis of his isotope data. It is possible that denitrification may be occurring in anaerobic micro-environments within the sand and gravel aquifer. Such micro-environments can result at dead-end pores, and in low permeability soil zones (layers or lenses), where porewater velocities are small and thus more opportunity is provided for 0 2 consumption as well as subsequent denitrification of N0 3 \" (Bengtsson & Annadotter, 1989; Lowrance & Pionke, 1989). There is also other evidence to support the occurrence of denitrification within the Abbotsford aquifer. Unpublished data maintained by the B C M O E and the USGS include findings of N0 2 \" and dissolved N 2 0 from the deeper (22.9 m, or 75 ft, below ground surface) screened intervals of the B C M O E wells A , B, and C (BCMOE, 1995; Cox, 1995). The dissolved oxygen levels are less than 1 mg/l (Table 4.2) at depth over this region, so there is a greater likelihood of denitrification occurring within the deeper portions of the aquifer. The observed decrease in N/Cl\" over depth below the water table may also be explained by the increase in groundwater CI\" content along the groundwater flow path. Hem (1989) noted that CI\" trapped in gypsum and calcite can be released into flowing groundwaters. The likelihood of calcite dissolution at the study site is discussed further in Section 7.2.3. 7.3.1.2 Brookswood Study Wells Figure 7.12 presents the nitrate-N/Cl\" profiles for the Brookswood North and 213 Nitrate-N/Chloride 0.4 0.8 1.2 1.6 0.4 0.8 1.2 1.6 Figure 7.12 Monthly groundwater Nitrate-N/Cl profiles from Brookswood study wells 214 South study wells. The South study well profiles show no change in the ratio over depth. At the North study well, the higher N/CT values above 10 m depth can be associated with the arrival of nitrate with the rdcharge waters. The N/Cl\" profiles below 10 m are of a similar magnitude to those at the South well, but exhibit an increase with depth. While the N/Cl\" profiles at the South well point to no denitrification, the North well profiles are somewhat complicated. The measured increase in N/Cl\" over depth could be explained by the production of N0 3 \" . The production of N0 3 \" as a result of mineralization of organic matter embedded in subsurface soils has been encountered elsewhere, as discussed in Chapter 2. The data appears to suggest that mineralization might be occurring at the Brookswood site, as the depth related increase in N0 3 \" is also associated with a decrease in pH (Figure 5.2) - a trend that is associated with the process of nitrification. 7.3.2 IC vs pH at the Abbotsford Study Well Figure 7.13 presents plots of IC vs pH for the Abbotsford well for the several months that these two parameters were measured. In Section 2.3.1 it was noted that the occurrence of heterotrophic denitrification along a groundwater flow path would be associated with an increase in pH and IC. The plots on Figure 7.13 show such a trend, as pH and IC increase with depth below the water table. If the IC measured in the groundwater has been produced as a result of denitrification, a value of 10 mg/l IC (as encountered near the bottom of the well) suggests that 10 mg-N of N0 3 \" must have been denitrified - assuming a 1:1 ratio for N consumed and IC produced (equation 2.7). While the nitrate profiles (Figure 7.10) show nitrate concentrations to be decreasing by at least 215 15 10 5 0 Oct '94 Nov '94 15 \u00E2\u0080\u0094 i 1 1 0 -O 5 -0 -Mar '95 Apr'95 15 -10 \u00E2\u0080\u0094 5 \u00E2\u0080\u0094 0 -May '95 ^ 5.5 6.0 6.5 Jun '95 7.0 5.5 6.0 6.5 PH 7.0 Figure 7.13 Plots of groundwater IC vs pH for several months from Abbotsford study well 216 10 mg-N/1 over the monitoring depth, it has been concluded previously that dilution by dispersion can also explain the observed decrease in concentrations. Despite the fact that the measured increase in IC values may be correlated to the decrease in N0 3 \" values, the very low TOC and SOC levels do not support the conclusion of heterotrophic denitrification being totally responsible for the production of IC at the measured levels. An organic carbon content of at least 10 mg/l would b required to reduce 10 mg-N/1 from the groundwater (see equation 2.7). It has been estimated that the TOC values in the groundwater are probably less than 1 mg/l, and it is likely that this TOC has low biodegradability. In the absence of heterotrophic denitrification such a close correlation between IC and pH can be explained on the basis of calcite dissolution by recharging low pH waters (Heathcote, 1985). The associated reaction can be represented as: CaCOz + H* =\u00E2\u0080\u00A2 Ca 2 + + HCO; (7'2) However, the observed profiles do not show that C a 2 + is being released into solution over depth. This \"missing\" C a 2 + can be accounted for by the relative increase, over depth, in C a 2 + at the soil surface exchange sites, as shown in the TEC data on Table 4 .5 . Nesbitt & Cramer (1993) also deduced such C a 2 + adsorption at a site in Ontario, and where gypsum dissolution resulted in the release of S0 4 2 \" to the groundwater, but not C a 2 + . 7.4 R O L E F O R C A R B O N ADDITIONS IN P R O M O T I N G DENITRIF ICAT ION O F NON-POINT S O U R C E N I T R A T E At the Abbotsford and Brookswood study sites, it was found that non-point source 217 nitrate arrival and distribution beneath the water table can be a temporally and spatially dynamic process. The groundwater was found to be oxic, and no evidence of significant denitrification was found at the well sites over the monitoring depths. Under these conditions, the feasibility of carbon addition as a means of controlling the observed groundwater nitrate distributions appears to be limited. Moreover, the presence of a microbial population capable of heterotrophically denitrifying the leached nitrate cannot be automatically presumed. Given the observed temporally dynamic nature of the nitrate leaching process, a carbon delivery system functioning in sympathy with the dynamics of groundwater nitrate distribution would appear to be preferable. As the temporal and spatial variations associated with nitrate arrivals to the water table are to a large extent controlled by land use cycles and recharge patterns, the management of crop (or other organic) residues appears to offer the most potential as a means of providing carbon to the underlying shallow groundwaters. There is very little available information with which to estimate the amounts of residue required to ensure the delivery of carbon at the required levels. At the study sites, it was found that poultry manure applications typically are in the order of tonnes per hectare, yet the soluble carbon content of the underlying shallow groundwaters is negligible. As the manure is applied at the end of the recharge period, degradation within the root zone during the growing season could destroy the carbon, or it could be bound to the soil surfaces, prior to the next recharge period. This points to the application of organic residue at the end of the growing season just before recharge as a means of ensuring carbon leaching below the root zone. The major problem with such an approach 218 however, is that any increase in organic residue content within the surficial soils could also lead to an increase in soil nitrate levels, as suggested by the findings of Zebarth et al. (1994), and shown on Figure 7.3. So it would seem that the potential solution may infact compound the problem. The use of reactive barriers placed across the groundwater flow path, and the provision of slow leaching carbon within these barriers, for nitrate removal has been discussed by Devlin & Barker (1994), Robertson & Cherry (1995) and Starr & Cherry (1994). The use of these barriers has been promoted by these various investigators in anticipation of low maintenance requirements as well as the long term effectiveness of the systems. The findings of the laboratory column testing performed as part of the current research suggest that clogging of such stationary systems, by the accumulation of N 2 gas bubbles may be a potential problem. This aspect requires further investigation, preferably on a field-scale, so that any constraints associated with the application of such a barrier approach can also be ascertained. The costs associated with the injection of carbon into groundwater to reduce nitrate levels at an aquifer scale may be prohibitively high. The findings of this research have not provided any information to suggest that such carbon injection may be more feasible than other approaches. Furthermore, a consideration of groundwater chemistry alterations as a result of anoxification would suggest that large scale injection of biodegradable carbon to groundwaters has the potential to reduce the quality of the water towards human consumption. On the whole, it would appear that the role of carbon additions to groundwaters towards controlling ongoing non-point source nitrate contamination is limited. 219 7.5 EFFICACY OF PASSIVE SAMPLING TOWARDS QUANTIFYING GROUNDWATER CHEMISTRY Passive sampling of groundwater dissolved constituents using the dialysis membrane technique is currently not commonly practised. A review of available literature shows that all of the related work performed to date has involved the use of the device as described by Ronen et al. (1986). Furthermore, the device developed by Ronen et al. (1986) is capable of providing a detailed vertical profile over a depth of only 2 m below the water table. This research has involved the development and use of a device that differs in several aspects from that of Ronen et al. (1986), not least of which is the extended depth of monitoring permitted by the equipment that was developed. As the study extended the application of this passive sampling technique to new sites, and used to measure temporal variations in groundwater chemistry, an opportunity is provided for further evaluation of the technique for widespread utility towards quantifying groundwater chemistry. The following discussion considers several aspects related to the dialysis membrane approach to passive sampling. 7.5.1 Membrane Degradation As previously noted, degradation of the dialysis membrane was observed during this study. Ronen et al. (1986) and others (Magaritz et al., 1989) discuss the use of nylon and cellulose acetate membranes for groundwater monitoring, but did not provide any discussion on the degradability of these materials. The regenerated cellulose membranes used during this research were selected 220 because they were available in the tubular form that suited the device configuration as well as the dialysis membrane sausage mode chosen to contain the initially deionised water. Cellulose dialysis membrane degradation has been reported by previous researchers who have used this material for monitoring surface water bodies. Vargo et al. (1975) found, in open waters, that regenerated cellulose can be degraded by microbes within 9 days under summer conditions. This study found that regenerated cellulose membranes can be degraded totally after a residence period of one month within a well below the water table. At the Brookswood North study well, the membranes retrieved after the first sampling event attempted in July 1994 were found to have degraded to a slimy mass. The observed membrane degradation was associated with an accumulation of fines within the sampling device, and microorganisms associated with the fines are presumably responsible for the observed degradation. Fines accumulation and membrane degradation were minimised by developing (once) the wells using a submersible pump, and also by limiting the residence time of the device in the well to two weeks. After initial development, both the Brookswood wells continued to accumulate fines, though in small amounts, throughout the monitoring period. In these wells , fines accumulation was observed within the canisters, and as a partial coating on the membrane bags. Tactile examination of the fines showed these to consist of both silt (\"gritty\") and clay (\"slippery\") fractions. No such accumulations was experienced at the Abbotsford study well. The limited and ongoing accumulation of fines during sampling events within the Brookswood wells was also associated with a slight but perceptible degradation of the membrane bags. 221 Ronen et al. (1992) also described the accumulation of fines within their device, and they took the opportunity to characterize in detail the nature of these fines, which they considered to be suspended during groundwater flow. At their study site, the fines were found to consist of detrital calcite, quartz and clay released into the groundwater as colloidal particles. The potential impact of membrane degradation on the chemistry of the retrieved membrane enclosed waters is not readily apparent. However, given the reliance of this passive sampling technique on the diffusion of constituents across the membrane, it is likely that any degradation products (which would presumably include organic C) would not have been concentrated within the enclosed waters. Furthermore, the degradation of the membrane presumably leads to an increase in its permeability, and possibly a decrease in the diffusion resistance (if the membrane wall thickness is reduced), both of which would only enhance the equilibration of the chemistry between the membrane enclosed water and the surrounding groundwater. The comparatively higher levels of TOC measured from the Brookswood wells, and the relative uniform distribution of these values over depth suggests that bag degradation may have contributed to the TOC content of the retrieved waters. This is particularly evident for the Brookswood South well, which consistently showed the greatest relative amount of fines accumulations as well as bag degradation. The laboratory evaluations of the membrane materials (Appendix B) have also shown that TOC contributions from the bag can occur on prolonged submergence. 222 7.5.2 Repeatability & Accuracy The results of the laboratory evaluations of the membrane material, as described in Appendix B, show that diffusion of dissolved constituents across the membrane occurs rapidly. In a stationary bath, it was found that all major ions diffused into the prepared sausages and reached bath concentrations within 48 h. Figure 7.14 presents a plot of the early profiles obtained from the Abbotsford and Brookswood North study wells. This figure compares the nitrate profiles obtained over successive months from July to September 1994 - a period associated with no recharge to the water table, and during which time the chemical profiles showed little overall temporal change. These early months of monitoring were also associated with the reduction in device residence time from one month to two weeks. Figure 7.14 shows that there was little effect on the chemical profiles as a result of reducing the device residence time in the well to two weeks. On this basis it is concluded that any disturbance to the groundwater flow regime in the vicinity of the well as a result of device lowering must have dissipated and the chemistry of the membrane enclosed waters was equilibrated with the surrounding groundwater within the two week period. The nitrate concentrations encountered in Abbotsford and Brookswood North study wells range between 1 mg-N/1 and 50 mg-N/1. The profiles shown on Figure 7.13 show good repeatability over this range - while recognising the possibility of some temporal variation between successive months. The passive sampling technique is therefore shown to be capable of providing not only a detailed and repeatable characterization of the groundwater chemistry, but also an 223 0 ' ' ' I I I I I I I I I I L Nitrate (mg-N/1) 10 15 0 10 20 30 40 50 _ J I i I i I i I i I Brookswood Aquifer North Study Well 6 -E 8 i CD 0 f co 1 0 \u00E2\u0080\u00A2o c 2 CD 1 1 2 CD CL \u00C2\u00A3 14 16 -Sep 7 '94 (one month in well) Sep 27 '94 (two weeks in well) 18 Abbotsford Aquifer Study Well Jul 26 '94 (one month in well) Aug 29 '94 (one month in well) Sep 27 '94 (two weeks in well) Figure 7.14 Comparison of groundwater nitrate profiles obtained at study wells over consecutive months during period of no recharge 224 accurate characterization when combined with the adopted methods of water analyses2. 7.5.3 Comparison to Pumped Well Sampling Data While it has been demonstrated that the passive sampling technique is capable of providing a repeatable and accurate set of chemical data, its effectiveness in providing a representative indication of actual groundwater chemistry can only be confirmed by comparison of the data to measurements obtained through direct sampling of the groundwater, such as pumping. However, this requires the placement of a pumped well in close proximity to the passively sampled well, in order to ensure that any spatial variability associated with the groundwater chemistry does not influence the comparison. At the Abbotsford study site, the groundwater chemistry at the BCMOE wells B and C (Figure 3.2) may be considered to be influenced by the same surficial land use impacting the groundwater encountered at the study well. Furthermore, as the BCMOE completed a round of sampling from these three wells during November, 1994, (BCMOE, 1995), the data obtained during that month can be compared to the findings at the study well. Table 7.3 presents a comparison of groundwater chemistry obtained using the passive sampling technique at the Abbotsford study well on October 25, 1994, and the groundwater chemistry as quantified by. the BCMOE on samples obtained on November 15, 1994, by pumping from their Wells B and C. Figure 7.15 presents a plot of the Table 7.3 data. In general, the values measured at the study well can be considered to 2 the Lachat autoanalyser is capable of accurately measuring nitrate concentrations as low as 0.05 mg-N/1 and at least as high as 100 mg-N/1 2 2 5 be within the orders of magnitude obtained through pumping from wells B and C, and there is no consistent difference between the study well data and the pumped well data. The exceptions to this are the values measured for nitrate and M g , both of which have higher measured values over all depths at the study well. The reason for this is not readily apparent. On the basis of this comparison, it can be concluded that the passive sampling technique provides data that is, at least, comparable to data obtained through well pumping. Ronen et al. (1986) reported that during early investigations with their passive sampling device, they were able to measure groundwater chemistry above detection limits using the passive sampler, but that pumping from a nearby well was unable to detect the same chemistry. This has not been the case at the Abbotsford site. Table 7.3 Comparison of Abbotsford study well (Oct 25' 94) and BCMOE Wells B and C (Nov 15 '94) groundwater chemistry Depth Well N03- CI IC PH Ca Na Mg 6.1 m(20ft) 7.6 m (25 ft) 7.6 m (25 ft) BCMOE B 13.4 11.9 4.3 2.4 6.3 21.2 4.4 3.9 BCMOE C 9.6 19.7 6.7 2.5 6.3 17 5 3.5 Study 28.6 23.1 10.3 2.5 6.3 26.5 6.6 10.4 10.7 m (35 ft) BCMOE B 17.5 10.9 5.9 3.3 6.4 26.6 5.6 5 BCMOE C 13.7 19.3 9.3 3 6.4 26 6.4 4.5 Study 23.4 16.9 10.4 4.7 6.5 25.6 6.1 9.4 16.8 m (55 ft) BCMOE B 6.8 12.3 3.6 11.5 7 21.4 4.2 5.3 BCMOE C 6 14.8 8.6 12.6 7.1 25 4.6 5.4 Study 12.9 12.7 7.3 10 6.7 18.3 4.6 8.9 226 Figure 7.15 Plots of data contained in Table 7.3: a) 6.1/7.6 m (20/25 ft), b) 18.3 m (35 ft) and c) 16.8 m (55 ft) depths 8.0 CONCLUSIONS 227 A study involving the detailed monitoring of the changing groundwater chemistry profiles beneath the water table of two unconfined aquifers, combined with a laboratory column-test investigation of enhanced denitrification during saturated flow through sand has been performed. The completed field and laboratory based investigations have led to a better understanding of non-point source nitrate contamination of unconfined aquifers, and the enhancement of denitrification in-situ by the introduction of carbon. The unconfined aquifers (Abbotsford and Brookswood, within the Lower Fraser Valley region of British Columbia), selected for the study, are within unconsolidated sand and gravel deposits which are recharged by heavy winter precipitation. Groundwater flow velocities within these aquifers have been estimated to range from 0.5 m/day to nearly 3 m/day. The laboratory column tests have been performed using flow velocities of between 0.4 m/day and 0.8 m/day in order to simulate the local conditions. The laboratory based investigations have provided further insight into the fate of N 2 that is generated during enhanced denitrification under saturated flow conditions, and insight into the importance of carbon source in the enhancement of denitrification. An improved sampling device (using the passive sampling concept previously developed by Ronen et a l . , 1986) has been designed and applied to obtain vertical chemistry profiles over extended depth beneath the water table of unconfined aquifers. The application of this technique has not only provided more detailed information on the distribution of nitrate within local aquifers, but also the first known measurement of the arrival of leached non-point source nitrate at the water table. It has also led to the first 228 detailed measurement of the sub-annual temporal variation of groundwater chemistry profiles beneath the water table of unconfined aquifers. A conceptual groundwater flow model has been developed to describe the measured temporal variation in ground water chemistry in response to seasonal fluctuations in recharge. The improved understanding of non-point source nitrate contamination of unconfined aquifers has permitted a more rational evaluation of the role of carbon addition for in-situ control of groundwater nitrate. It has also led to an improved definition of the monitoring requirements towards effective groundwater protection from non-point source contamination. With respect to the study objectives, the study findings suggest the following: 1) Generation and fate of N 2 under saturated flow conditions through soil: The laboratory column tests have found that N 2 , produced on denitrification, can enter the gas phase and form bubbles within sand. This finding has confirmed the previous work of Soares et al. (1991). Additionally, definitive evidence of this tendency was obtained during this study by the direct measurement of N 2 through sampling such a bubble. However, the study has also shown that N 2 does not immediately enter the gas phase on being produced on denitrification. Dissolved N 2 , at concentrations above saturation, has been measured in effluent from packed sand columns. The findings suggest that gas bubble formation may only occur after the microbial population density has increased sufficiently to form biofilms, or under conditions of low flow velocity with high substrate loading and a high denitrification rate. 229 It has been found that under the imposed test conditions, denitrification, and subsequent N 2 generation, will only proceed on the total reduction of dissolved within the pore waters. 2) Significance of carbon source in the enhancement of denitrification: It has been found that ethanol is more immediately utilized than methanol by the denitrifying population present within the packed sand columns used to simulate groundwater environments. While denitrification was achieved with methanol, it was found that this was achieved at low flow rates, and after an acclimatization period. The denitrifying population responded immediately to the introduction of ethanol. However, nitrite accumulation (incomplete denitrification) was associated with ethanol as a carbon source. These findings suggest that any scheme aimed at introducing biodegradable carbon into aquifers should consider the type of carbon being introduced, and any acclimatization period (relative to the groundwater travel times) required by site specific microbes. The potential for nitrite accumulation, as encountered with ethanol, should be considered as drinking water guidelines for nitrite (1 mg-N/1) are less than for nitrate (10 mg-N/1). The amount of dissolved 0 2 present within the groundwaters should also be ascertained, in order that sufficient carbon is provided to reduce both the and NCy within the groundwaters. 3) Feasibility of enhanced denitrification for non-point source nitrate control: The role for carbon addition to groundwater environments for control of non-point source nitrate may be constrained by the generally large scale nature of such contamination, as 230 well as an uncertainty associated with the viability of heterotrophic denitrifying microbial populations toward rapid utilization of introduced carbon within aquifers. The injection of carbon below the water table, or the introduction of carbon at the ground surface, wil l have to carefully consider the groundwater flow regime and, in particular, the extent of mixing affected by dispersion through the aquifer. Immediate and total mixing (as suggested by previous investigators) of introduced carbon within receiving groundwaters should not be presumed. In aquifers where transverse dispersion is limited, carbon injection may need to be performed over the full depth of nitrate contamination to affect any mixing with the groundwaters, and to promote sufficient denitrification to control groundwater nitrate levels. Reactive barriers containing a biodegradable carbon source, and which are placed across the groundwater flow path, may be infeasible in regions where nitrate contamination extends to a large depth below the water table. Furthermore, clogging of reactive barriers as a result of N 2 gas bubble formation and accumulation, as well as due to biogrowth, may be possible as these barriers offer the opportunity for biofilm development and concentration of denitrifyer activity. 4) Seasonal cycle of nitrate loading to unconfined aquifers: At the study sites, where the onset of winter precipitation was associated with a rapid rise in the water table, the onset of the recharge was also associated with the arrival of a nitrate front to the water table. At the study sites, the arrival of a nitrate front to the water table has been found to produce sharp, but temporally unstable, chemical gradients of the form previously 231 reported in the literature. The study findings clearly show that such steep gradients can develop in the absence of denitrification. At a given well location, the distribution and level of nitrate contamination below the water table of unconfined aquifers can show substantial sub-annual temporal variation. It has been found that nitrate concentrations can fluctuate over a range of greater than 10 mg-N/1 (the drinking water guideline). Over a given region, the vertical distribution of nitrate below the water table can show substantial spatial variation. Factors controlling the areal and temporal distribution of nitrate below the water table include the areal source extent of nitrate, recharge pattern and groundwater flow regime. Deep and rapid transport of leached nitrate below the water table may occur as a result of a seasonal variation in. the groundwater flow pattern and associated flow velocities. It has been inferred that, under sufficient recharge rates, the development of vertical hydraulic gradients near the water table results in the vertical transport of leached contaminants below the water table. Furthermore, the corresponding seasonal fluctuations in flow wil l result in the movement of nitrate through the groundwater in a pulsed manner. Other findings from the study include: 5) C h e m i s t r y associated w i t h l eached n o n - p o i n t source n i t ra te : Leaching of nitrate through the unsaturated zone wil l be associated with the leaching of other ionic constituents. These constituents may originate from the source itself, and/or mineral dissolution and leaching from the soil surfaces as a result of charge imbalances in the 232 flowing porewaters. Non-point source nitrate contamination of \"fresh\" groundwaters can lead to significant changes in the major-ion chemistry of these waters. At the study sites, it was found that S042\", and C a 2 + also arrived at high concentrations to the water tabie. Additionally, the recharge waters were found to have low pH and alkalinity. It is concluded that this was due to the nitrification of the NFL/ originating from surface applied poultry manure. At the study sites, it was found that the water table region was oxic over the monitored depths, suggesting the arrival of oxygen saturated recharge waters. 6) Role of sampling from existing water supply wells for aquifer management: Region wide sampling of existing water supply wells can provide a means of assessing the quality of the groundwater in the vicinity of individual wells, while enabling qualitative assessments of the susceptibility of localised regions to surrounding surficial land use. In the absence of knowledge of the underlying groundwater flow regime, as well as a proper site-specific understanding of surficial land use impacts on the underlying groundwater, the study findings suggest that regional sampling of water supply wells cannot be solely used to provide information for the effective management of the integrity of the underlying groundwater resource under current and future land use. Annual, and less frequent, monitoring of water supply wells cannot be used as a means of predicting long-term trends within aquifers where large sub-annual fluctuations of groundwater nitrate concentrations occur. 233 7) Role of multi-level sampling: Detailed multi-level sampling of groundwater beneath the water table of unconfined aquifers will enable the more effective delineation of non-point source nitrate distribution, and also provide more appropriate data for evaluating the transport and fate of the nitrate through the groundwater. Frequent multi-level sampling can be used to distinguish seasonally pulsed arrivals of nitrate to the water table, as well as provide a means by which the distribution of these fronts below the water table can be followed. Detailed multi-level sampling near the water table also enables the site-specific quantification of leaching losses to the water table. 8) Role of passive sampling using the dialysis membrane technique: The study has provided further confirmation of the effectiveness of passive sampling, using the dialysis membrane technique, towards obtaining accurate, repeatable, and highly detailed groundwater chemistry profiles beneath the water table. The study has also shown that with an appropriate design, the technique can be conveniently applied over an extended depth below the water table. The advantages offered by the dialysis membrane technique include: i) non-reliance on pumping (and therefore the associated equipment) as a means of obtaining groundwater samples. ii) rapidity of sampling: the on-site time required for the installation and retrieval of the sampling device and water sample collection is short. Under the conditions of the study, using pre-prepared dialysis bags, device installation was typically completed within 30 mins to 1 hr, while retrieval 234 and sample collection took 1 to 2 hrs. Some difficulties and potential limitations associated with the dialysis membrane passive sampling technique include: iii) microbial degradation of regenerated cellulose dialysis membrane has been found to occur, though the extent of degradation can be controlled by reducing the residence time within the monitoring well. Such degradation due to microbial activity has the potential to effect the interpretation of the dissolved oxygen, organic and inorganic carbon content within the surrounding groundwater. iv) the sampling device, as used in the study, relied on a continuously screened well extending over some depth through the aquifer. The provision of such a preferential pathway may impact the local groundwater flow where high permeability lenses or layers are encountered by the well. While the placement of neoprene spacers through the well has been used to limit any vertical flows through the well, this aspect needs to be evaluated further. v) the quantity of water sample obtained at any given depth, during any given sampling event, is small, and can be limited by the sampler dimensions and the amount of deionised water initially enclosed by the dialysis membrane. 235 9.0 RECOMMENDATIONS FOR FURTHER STUDY Further investigation of the following aspects is recommended: 1) The role of the unsaturated zone in controlling the rate of nitrate delivery to the water table. Detailed and frequent monitoring of the porewater chemistry in the unsaturated zone over the full depth between the root zone and the water table, concomitant to detailed and frequent profiling of the groundwater chemistry beneath the water table, should be performed at the Abbotsford site in order to confirm the inference of complete leaching of nitrate from the root zone during the winter recharge period. 2) The areal distribution of leaching nitrate fronts. At each of the study sites, only a single passive sampling well was used to investigate the nitrate distribution under the various land uses. Monitoring using a grid of passive sampling wells is recommended in order to obtain detailed 3-dimensional profiles of groundwater chemistry, as well as its temporal variation. For example, at the Abbotsford study site, placement of passive sampling wells adjacent to the B C M O E wells A , B and C would enhance the monitoring effort in that region. 3) The temporal variation in groundwater flow patterns in response to seasonal recharge. The development of, a numerical model, to describe the inferred variation in groundwater flow direction and velocity; at and below the water table of unconfined aquifers in response to seasonal recharge cycles, should be pursued. Detailed mapping 236 of groundwater flow velocities through in-situ measurements should be performed to validate the developed model. 4) The efficacy of passive sampling under a variety of hydrogeological conditions. The passive sampling technique was used in this study in relatively homogeneous sand and gravel deposits having high conductivities and flow rates. Further application of this technique should be attempted under less ideal conditions, such as in stratified deposits and in areas of local confining conditions, to assess the wide scale applicability of the technique for monitoring of groundwater chemistry. 5) The generation and fate of N 2 from a microbial colony during saturated flow through soil. Further testing to elucidate the exact conditions required for the onset of bubble formation on denitrification should be performed in order to quantify the likelihood of clogging due to bubble accumulations in subsurface treatment systems such as reactive barriers. Field trials should also be performed to assess actual in-situ impacts of such clogging. 237 BIBLIOGRAPHY Adams, P.L . , Daniel, T .C. , Edwards, D.R., Nichols, D.J., Pote, D .H. & Scott, H.D. (1994) Poultry litter and manure contributions to nitrate leaching through the vadose zone. Journal Soil Science Society of America, 58:1206-1211 Adelman, D. & Dahab, M.F . (1991) Effects of solute transport processes on in-situ denitrification for nitrate control in groundwater. In Proceedings of International Symposium on Groundwater, G.P. Lennon (ed), A S C E Adelman, D.D. & Spalding, R.F. (1988) The economic feasibility of using ethanol to accelerate groundwater denitrification. In \"Planning now for irrigation & drainage in the 21st century\", Proceedings of Conference, A S C E Alva, A . K . (1992) Differential leaching of nutrients from soluble vs controlled-release fertilizers. Environmental Management, 16(6):769-776 Andersen, L.J . , Kelstrup, N. & Kriatiansen, H. (1980) Chemical profiles in the Karup water-table aquifer, Denmark. In \"Nuclear techniques in groundwater pollution research\", International Atomic Energy Agency Appelo, C .A .J . & Postma, D. (1993) Geochemistry, groundwater & pollution. Balkema Aravena, R. & Wassenaar, L.I. (1993) Dissolved organic carbon and methane in a regional confined aquifer, southern Ontario, Canada: carbon isotope evidence for associated subsurface sources. Applied Geochemistry, 8:483-493 Atwater, J.W. et al (1994) Groundwater resources of British Columbia. BC Environment and Environment Canada Barcelona, M.J . , Holm, T.R., Schok, M.R. & George, G.K. (1989) Spatial and temporal gradients in aquifer oxidation-reduction conditions. Water Resources Research, 25(5):991-1003 Benefield, L .D. & Randall, C.W. (1980) Biological process design for wastewater treatment. Prentice Hall Beauchamp, E.G. & Voroney, R.P. (1994) Crop carbon contribution to the soil with different cropping and livestock systems. Journal Soil & Water Conservation, 49(2):205-209 Bengtsson, G. and Annadotter, H. (1989) Nitrate reduction in a groundwater microcosm determined by 1 5 N gas chromatograph-mass spectrometry. Applied & Environmental Microbiology, 55(ll):2861-2870 Bergstrom, L. & Johansson, R. (1991) Leaching of nitrate from monolith lysimeters of 238 different types of agricultural soils. Journal of Environmental Quality, 20:801-807 Bjerg, P.L. & Christensen, T .H . (1993a) A field experiment on cation exchange-affected multicomponent solute transport in a sandy aquifer. Journal of Contaminant Hydrology, 12:269-270 Bjerg, P.L. & Christensen, T .H . (1993b) Model simulations of a field experiment on cation exchange-affected multicomponent solute transport in a sandy aquifer. Journal of Contaminant Hydrology, 12:291-311 Boettcher, J . , Strebel, O., Duynisveld, W . H . M . & Frind, E.O. (1991) Reply to Korom (1991). Water Resources Research, 27(12):3275-3278 Bottcher. J . , Strebel, O. & Kolle, W. (1992) Redox conditions and microbial sulfur reactions in the Fuhrberger feld study aquifer. In \"Progress in Hydrogeochemistry\", G.Matthess et al. (eds), Springer-Verlag Bouchard, D . C , Williams, M .K . & Surampalli, R.Y. (1992) Nitrate contamination of groundwater: Sources and potential health effects. Journal of the American water Works Association, 84(9): 85-90 Bouwer, E.J. & Cobb, G.D. (1987) Modelling of biological processes in the subsurface. Water, Science & Technology, 19:769-779 Bradford, S . M . , Palmer, C.J. & Olson, B.H. (1994) Assimilable organic carbon concentrations in southern California surface and groundwater. Water Research, 28(2):427-435 Bradley, P . M . , Fernandez, M . & Chapelle, F .H . (1992) Carbon limitation of denitrification rates in an anaerobic groundwater system. Environmental Science & Technology, 26(12):2377-2381 Broers, H.P. & Griffieoen, J. (1994) Reactivity of natural sediments in relation to groundwater quality in several hydrogeochemical environments. In \"Transport and reactive processes in aquifers\", T .H . Dracos 8c F. Stauffer (eds). Balkema Burt, T .P. , Heathwaite, A . L . & Trudgill, S.T. (eds) (1993) Nitrate: Processes, patterns and management. John Wiley & Sons Cahn, M . D . , Bouldin, D.R., Cravo, M.S. & Bowen, W.T. (1993) Cation and nitrate leaching in an oxisol of the Brazilian Amazon. Agronomy Journal, 85:334-340 Carmichael, V . E . 8c Wei, M . (1994) A strategy for managing groundwater and community drinking water systems in the Fraser Valley, British Columbia. Proceedings of the 6th National Conference on Drinking Water, Victoria, BC , October 16-18 1994 239 Ceazan, M . L . , Thurman, E . M . & Smith, R.L. (1989) Retardation of ammonium and potassium transport though a contaminated sand and gravel aquifer: the role of cation exchange. Environment, Science & Technology, 23:1402-1408 Chambers, L .W. & Bahr, J . M . (1992) Tracer test evaluation of a drainage ditch capture zone. Ground Water, 30(5): 667-675 Champ, D.R., Gulens, J. & Jackson, R.E. (1979) Oxidation-reduction sequences in groundwater flow systems. Canadian Journal of Earth Sciences, 16:12-23 Chao, T.T., Harward, M.E . & Fang, S.C. (1962) Soil constituents and properties in the adsorption of sulfate ions. Soil Science, 94:276-283 Chao, T.T., Harward, M.E . & Fang, S.C. (1963) Cationic effects on sulfate adsorption by soils. Proceedings Soil Science Society of America, 27:35-38 Chao, T.T. (1964) Anionic effects of sulfate adsorption by soils. Proceedings Soil Science Society of America, 28:581-583 Chao, T.T., Harward, M.E . & Fang, S.C. (1964) Iron and aluminum coatings in relation to sulfate adsorption characteristics of soils. Proceedings Soil Science Society of America, 28:632-635 Cherkauer, D.S., McKereghan, P.F. & Schalch, L .H . (1992) Delivery of nitrate and chloride to the Great Lakes: Case study for the Door Peninsula, Wisconsin. Ground Water, 30(6): 885-894 Childs, S .W. , Lindstrom, F.T., Boersma, L. & Myrold, D.D. (1988) Model of carbon substrate injection to enhance denitrification in aquifers. In \"Agricultural impacts on groundwater\", Proceedings of National Water Well Association Conference Chin, K.J . , Janssen, L.J.J., van Strelen, S.J .D. , Verbunt, J .H.G. and Sluyter, W . M . (1988) Bubble parameters and efficiency of gas bubble evolution for a chlorine-, a hydrogen and an oxygen evolving wire electrode. Electrochimica Acta 33(6):769-779 Chipperfield, K. (1992) Raspberry field soil nitrate survey. Sustainable Poultry Farming Group Chipperfield, K. (1994) An evaluation of leachate and soil effects for stockpiling practises for poultry litter in the Fraser Valley. Sustainable Poultry Farming Group Christensson, M . , Lie, E. and Welander, T. (1994) A comparison between ethanol and methanol as carbon sources for denitrification. Water, Science & Technology, 30(6): 83-90 240 Chwqjka, F.R. (1990) Drilling and construction of seven drill holes for installation of 13 piezometers at Matsqui, British Columbia. Contracts No. 87 and 88. Groundwater Section, BC Ministry of Environment Clifford, D. & L iu , X . (1993) Ion exchange for nitrate removal. Journal of the American Water Works Association, 85:135-143 Cook, K .E . (1994) An evaluation of water quality and land use in the Salmon River watershed, Langley, BC , using GIS techniques. Unpublished Master of Science thesis, Department of Soil Science, The University of British Columbia Cook, P .L . , Solomon, D.K. , Plummer, L . N . , Rosenberg, E. & Schift, S .L . (1995) Chlorofluorocarbons as tracers of groundwater transport processes in a shallow, silty sand aquifer. Water Resources Research, 31(3):425-434 Cooper, J.R., Reneau, R.B. , Kroontje, W. & Jones, G.D. (1984) Distribution of nitrogenous compounds in a rhodic paleudult following heavy manure application. Journal of Environmental Quality, 13(2): 189-193 Cox, S. (1995) personal communication Dahab, M.F . (1993) Comparison and evaluation of in-situ bio-denitrification systems for nitrate reduction in groundwater. Water, Science & Technology, 28(3-5):359-368 Dance, J.T. & Reardon, E.J. (1983) Migration of contaminants in groundwater at a landfill: A case study. 5. cation migration in the dispersion test. Journal of Hydrology, 63:109-130 Davis, J.O. (1995) Tracking nitrate-nitrogen contaminants in the Abbotsford aquifer using hydrogeologic and geophysical methods. Unpublished report of a directed study. Department of Geological Sciences, University of British Columbia Dean, J .D. & Gaboury, D.R. (1987) Engineering aspects of point versus nonpoint sources of groundwater pollution. In \"Detection, control and renovation of contaminated groundwater\", N. Deo et al. (eds). A S C E De Haan, S .B . , Rae, J.E. & parker, A . (1994) Pyrite oxidation in the tertiary sands of the London Basin aquifer. Applied Geochemistry, 9:161-173 Devitt, D., Letey, J . , Lund, L.J. & Blair, J.W. (1976) Nitrate-nitrogen movement through soil as affected by soil profile characteristics. Journal of Environmental Quality, 5(3):283-288 Devlin, J.F. and Barker, J.F. (1994) A semipassive nutrient injection scheme for enhanced in situ bioremediation. Ground Water 32(3):374-380 241 Dillon, P.J. (1988) An evaluation of the source of nitrate in groundwater near Mount Gambier, South Australia. CSIRO Water Resources Series No. 1. CSIRO Australia Dillon, P J . (1989) An analytical model of contaminant transport from diffuse sources in saturated porous media. Water Resources Research, 25(6): 1208-1218 Dunkle, S .A . , Plummer, L . N . , Busenberg, E., Philips, P.J . , Denver, J . M . , Hamilton, P .A. , Michel, R.L. & Coplen, T.B. (1993) Chlorofluorocarbons (CC13F and CC12F2) as dating tools and hydrologic tracers in shallow groundwater of the Delmarva Peninsula, Atlantic Coastal Plain, United States. Water Resources Research, 29(12):3833-3860 Dunnivant, F . M . , Jardine, P . M . , Taylor, D.L. & McCarthy, J.F. (1992) Transport of naturally occurring dissolved organic carbon in laboratory columns containing aquifer material. Journal Soil Science Society of America, 56:437-444 Duthy, M .L . (1993) Modelling bioremediation of nitrate-contaminated waste and groundwater. Journal of Australian Geology & Geophysics, 14(2/3):249-257 Edmunds, W . M . , Miles, D.L. & Cook, J . M . (1984) A comparative study of sequential redox processes in three British aquifers. In \"Hydrochemical balances of freshwater systems\", E.Eriksson (ed), IAHS-AISH Publication No. 150 Egboka, B.C.E. (1984) Nitrate contamination of groundwater in Ontario, Canada. The Science of the Total Environment, 35:53-70 Eghball, B., Mielke, L . N . , McCallister, D.L. & Doran, J.W. (1994) Distribution of organic carbon and inorganic nitrogen in a soil under various tillage and crop sequences. Journal Soil & Water Conservation, 49(2):201-205 Engesgaard, P. & Kipp, K .L . (1992) A geochemical transport model for redox-controlled movement of mineral fronts in groundwater flow systems: A case of nitrate removal by oxidation of pyrite. Water Resources Research, 28(10):2829-2843 F A O (1993) Prevention of water pollution by agriculture and related activities. Water Reports 1. Proceedings of a F A O Expert Consultation, Chile Feth, J .H . (1966) Nitrogen compounds in natural waters - A review. Water Resources Research, 2(l):41-58 Focht, D.D. (1973) Isotope fractionation of 1 5 N and 1 4 N in microbial nitrogen transformations: A theoretical model. Journal of Environmental Quality, 22(2):247-252 Foster, S .S.D, Cripps, A . C . & Smith-Carrington, A . (1982) Nitrate leaching to 242 groundwater. Philosophic Transactions of the Royal Society, London, 296:477-489 Foster, S .S .D. , Geake, A . K . , Lowrence, A .R . & Parker, J . M . (1985) Diffuse groundwater pollution: Lessons of the British experience. In \"Hydrogeology in the service of man\", IAHS Publication No. 154, Vol . 3:168-177 Francis, D.D. (1992) Control mechanisms to reduce fertilizer nitrogen movement into groundwater. Journal of Soil and water Conservation, 47(6):444-448 Freeze, R. A . & Witherspoon, P. A . (1967) Theoretical analysis of regional groundwater flow. 2. effect of water table configuration and subsurface permeability variation. Water Resources Research, 3(2):623-634 Freeze, R.A. & Cherry, J .A. (1979) Groundwater. Prentice Hall Frind, E.O. (1987) Modelling of contaminant transport in groundwater - an overview. In Proceedings of CSCE Centennial Symposium on Management of Waste Contamination of Groundwater, R .N . Yong (ed) Frind, E.O., Duynisveld, W . H . M . , Strebel, O. & Boettcher, J. (1990) Modelling of multicomponent transport with microbial transformation in groundwater: The Fuhrbreg case. Water Resources Research, 26(8): 1707-1719 Gartner Lee Ltd. (1993) Fraser valley groundwater monitoring program: phase I report. Prepared for the BC Ministry of Health Garrels, R . M . & Christ, C .L . (1965) Solutions, minerals and equilibria. Freeman, Cooper & Co. Gayle, B.P. , Boardman, G.D. , Sherrard, J .H . & Benoit, R.E. (1989) Biological denitrification of water. Journal of Environmental Engineering, 115(5):930-943 Gelhar, L .W. & Wilson, J .L. (1974) Ground water quality modelling. Ground Water, 12(6): 399-407 Geng, Q.Z., Girard, G. & Ledoux, E. (1989) Coupled modelling of water cycle and nitrate transport in a hydrological system. In \"Appropriate methodologies for development and management of groundwater resources in developing countries\", C P . Gupta (ed). A . A . Balkema Gerhart, J . M . (1986) Ground-water recharge and its effects on nitrate concentration beneath a manured field site in Pennsylvania. Ground water, 24(4):483-489 Gold, A . J . , Lamb, B.E. , Loomis, G.W., Boyd, J.R., Cabelli, V . J . & McKie l , C F . (1992) Wastewater renovation in buried and recirculating sand filters. Journal of 243 Environmental Quality, 21:720-725 Government of Canada (1990) Canada's Green Plan: Canada's green plan for a healthy environment Ghirose, W . C . & Wilson, J.T. (1988) Microbial ecology of the terrestrial subsurface. Advances in Applied Microbiology, 33:107-172 Glas, J .P. and Westwater, J.W. (1964) Measurements of the growth of electrolytic bubbles. International Journal of Heat & Mass Transfer. 7:1427-1443. Grantham, G. & Lucas, J .L. (1985) Monitoring of the unsaturated zone as an aid in aquifer protection. In \"Hydrogeology in the Service of Man\", IAHS Publication No. 154, Vol . 3:70-83 Groffman, P . M . , Gold, A . J . & Simmons, R.C. (1992) Nitrate dynamics in riparian forests: microbial studies. Journal of Environmental Quality, 21:666-671 Gron, C , Torslov, J . , Albrechtsen, H.J. & Jensen, H . M . (1992) Biodegradability of dissolved organic carbon from an unconfined aquifer. The Science of the Total Environment, 117/118:241-251 Hagedorn, C. & Reneau Jr., R.B. (1994) Evaluation of denitrification in two shallow placed low pressure distribution systems. Water, A i r & Soil Pollution, 73:95-103 Hall , D.W. (1992) Effects of nutrient management on nitrate levels in groundwater near Ephrata, Pennsylvania. Ground Water, 30(5):720-730 Hallberg, G.R. (1989) Nitrate in groundwater in the United States. In \"Nitrogen management and groundwater protection\", R.F. Follett (ed), Developments in Agricultural and Managed-Forest Ecology 21. Elsevier Science Hanaki, K. and Polprasert, C. (1989) Contribution of methanogenesis to denitrification with an upflow filter. Journal of the Water Pollution Control Federation. 61(9): 1604-1611 Hantzsche, N .N . & Finnemore, E.J. (1992) Predicting ground-water nitrate-nitrogen impacts. Ground Water, 30(4):490-499 Harper, C R . , Goetz, W.J . & Willis, C.E. (1992) Groundwater protection in mixed land-use aquifers. Environmental Management, 16(6):777-783 Harremoes, P., la Cour Jansen, J . and Kristensen, G .H . (1980) Practical problems related to nitrogen bubble formation in fixed film reactors. Progress in Water Technology. 12:253-269 244 Haycock, N.E. & Pinay, G. (1993) Groundwater nitrate dynamics in grass and poplar vegetated riparian buffer strips during the winter. Journal of Environmental Quality, 22:273-278 Heathcote, J .A. (1985) Carbonate chemistry of recent chalk groundwater in a part of east Anglia, U.K. Journal of Hydrology, 78:215-227 Herbel, M.J . & Spalding, R.F. (1993) Vadose zone fertilizer derived nitrate and 5 1 5 N extracts. Ground Water, 31(3):376-382 Hickman, G.T., Novak, J.T., Morris, M.S. & Rebhun, M . (1989) Effects of site variations on subsurface biodegradation potential. Journal Water Pollution Control Federation, 61(9): 1564-1575 H i i , B. (1995) personal communication Hirsch, P. (ed) (1988) Groundwater. Microbial Ecology, 16(1). Special Issue Hiscock, K . M . , Lloyd, J.W. & Lerner, D .N . (1991) Review of natural and artificial denitrification of groundwater. Water Research, 25(9): 1099-1 111 Hoeks, J. (1981) Analytical solutions for transport of conservative and nonconservative contaminants in groundwater systems. Water, A i r & Soil Pollution, 16:339-350 Iqbal, M .Z . & Krothe, N.C. (1995) Infiltration mechanisms related to agricultural waste transport through the soil mantle to karst aquifers of southern Indiana, USA. Journal of Hydrology, 164:171-192 Jackson, W. A . , Leonard, R.A. & Wilkinson, S.R. (1975) Land disposal of broiler litter -changes in soil potassium, calcium and magnesium. Journal of Environmental Quality, 4(2):202-206 Jackson, W . A . , Wilkinson, S.R. & Leonard, R.A. (1977) Land disposal of broiler litter: changes in concentration of chloride, nitrate nitrogen, total nitrogen and organic matter in a Cecil sandy loam. Journal of Environmental Quality, 6(1) 58-62 James, R.V. & Rubin, J. (1979) Applicability of local equilibrium assumption to transport through soils affected by cation exchange. In \"Chemical modelling of aqueous systems\", E.A. Jenne (ed). ACS Symposium Series 93, American Chemical Society Jepsen, S .E . , Laursen, K., la Cour Jansen, J. and Harremoes, P. (1992) Denitrification in submerged filters of nitrified wastewater and chemical pre-precipitated wastewater. In \"Chemical water and wastewater II\", R. Klute & H.H. Hahn (eds). Springer-Verlag 245 Keeney, D. (1989) Sources of nitrate to groundwater. CRC Critical Reviews in Environmental Control, 16(3):257-304 Kelley, W.P . (1948) Cation exchange in soils. Reinhold Publishing Co. Kelley, C.C. & Spilsbury, R .H. (1939) Soil survey of the lower Fraser Valley. Publication 650, Technical Bulletin 20, Dominion of Canada - Department of Agriculture Kelly, W . E . , Curtis, B. & Adelman, D. (1991) Nitrate ground-water modelling for agricultural and non-point sources. In \"Nitrate contamination\", I. Bogardi & R.D. Kuzelka (eds). NATO ASI Series Volume G30. Springer-Verlag Kerner, M . (1993) Coupling of microbial fermentation and respiration processes in an intertidal mudflat of the Elbe estuary. Limnology & Oceanography, 38(2): 314-330 Kinzelbach, W. , Schafer, W. & Herza, J . (1991) Numerical modelling of natural and enhanced denitrification in aquifers. Water Resources Research, 27(6): 1123-1135 Kohut, A . P . (1987) Groundwater supply capability, Abbotsford upland. BC Ministry of Environment, Lands & Parks Kolpin, D.W., Burkart, M.R. & Thurman, E . M . (1994) Herbicides and nitrate in near-surface aquifers in the midcontinental United States, 1991. USGS Water Supply Paper 2413 Korom, S.F. (1991) Comment on \"Modelling of multicomponent transport with microbial transformation in groundwater: The Fuhrberg Case\" by E.O. Frind et al. Water Resources Research, 27(12):3271-3274 Korom, S.F. (1992) Natural denitrification in the saturated zone: A review. Water Resources Research, 28(6): 1657-1668 Krajenbrink, G.J.W., Ronen, D., Van Duijvanbooden, W. , Magaritz, M . & Wever, D. (1988) Monitoring of recharge water quality under woodland. Journal of Hydrology, 98:83-102 Krajenbrink, G.J.W., Boumans, L . J .M . & Meinardi, C R . (1989) Hydrochemical processes in the top layer of groundwater under pasture land. In \"Nitrogen in organic wastes applied to soils\", J .AA. Hansen & K. Henriksen (eds), Academic Press Lance, J.C. and Whisler, F.D. (1972) Nitrogen balance in soil columns intermittently flooded with secondary sewage effluent. Journal of Environmental Quality. 1(2): 180-186 246 Leenheer, J .A. , Malcolm, R.L . , McKinley, P.W. & Eccles, L .A . (1974) Occurrence of dissolved organic carbon in groundwater samples in the United States. USGS Journal of Research, 2(3):361-369 Leuchs, W. & Obermann, P. (1992) Anoxic reaction zones in an aquifer influenced by increasing nitrate and sulfate contents. In \"Progress in Hydrogeochemistry\", G. Matthess et al. (eds). Springer-Verlag. Licht, L .A . & Schnoor, J .L. (1993) Tree buffers protect shallow groundwater at contaminated sites. Ground Water Currents, USEPA Newsletter, EPA/542/N-93/011 Liebhardt, W . C , Golt, C. & Tupin, J. (1979) Nitrate and ammonium concentrations of groundwater resulting from poultry manure applications. Journal of Environmental Quality, 8(2):211-215 Liebscher, H . , H i i , B. & McNaughton, D. (1992) Nitrate and pesticides in the Abbotsford aquifer, southwestern British Columbia. Environment Canada Lindberg, R.D. & Runnells, D.D. (1984) Groundwater redox reactions: an analysis of equilibrium state applied to E,, measurements and geochemical modelling. Science, 225:925-927 Lloyd, J.W. & Heatchote, J .A. (1985) Natural inorganic hydrochemistry in relation to groundwater. An introduction. Clarendon Press Lovley, D.R., Chapelle, F .H . & Woodward, J . C (1994) Use of dissolved H 2 concentrations to determine distribution of microbially catalysed reactions in anoxic groundwaters. Environmental Science & Technology, 28(7): 1205-1210 Lowrance, R. (1992) Groundwater nitrate and denitrification in a coastal plain riparian forest. Journal of Environmental Quality, 21:401-405 Lowrance, R.R. & Pionke, H.B. (1989) Transformations and movement of nitrate in aquifer systems. In \"Nitrogen management and groundwater protection\", R.F.Follett (ed), Developments in Agricultural & Managed Forest Ecology, 21. Elsevier Lunn, R.J. & Mackay, R. (1994) An integrated modelling system for nitrogen transport. In \"Transport and reactive processes in aquifers\", T .H . Dracos & F. Stauffer (eds). A . A . Balkema Luttmerding, H.A. (1980) Soils of the Langley-Vancouver map area, Volume 1 - soil map mosaics and legend, Lower Fraser Valley (scale 1:25,000). Report No. 15, BC Soil Survey. BC Ministry of Environment 247 Luttmerding, H.A. (1981a) Soils of the Langley-Vancouver map area, Volume 3 -description of the soils. Report No. 15, BC Soil Survey. BC Ministry of Environment Luttmerding, H.A. (1981b) Soils of the Langley-Vancouver map area, Volume 6 -technical data, soil profile descriptions and analytical data. Report No. 15, BC Soil Survey. BC Ministry of Environment Luttmerding, H.A. (1984) Soils of the Langley-vancouver map area, Volume 5 -agricultural soil management groups. Report No. 15, BC Soil Survey, BC Ministry of Environment Madison, R.J. & Brunett, J.O. (1984) Overview of the occurrence of nitrate in ground-water of the United States. USGS Water Supply Paper 2275, pp93-105 Malcolm, R.L. (1993) Concentration and composition of dissolved organic carbon in soils, streams and groundwaters. In \"Organic substances in soil and water: Natural constituents and their influence on contaminant behaviour\", A.J.Beck et al. (eds). Royal Society of Chemistry Mariotti, A . , Landreau, A . & Simon, B. (1988) 1 5 N isotope biogeochemistry and natural denitrification processes in groundwater: application to the chalk aquifer of northern France. Geochimica et Cosmochimica Acta, 52:1869-1878 Mattson, B. (1995) Examining nitrate distribution at two sites in the Abbotsford aquifer. Unpublished report prepared as part of Hydrology 504, University of Idaho McCarty, P .L . , Beck, L. and St. Amant, P. (1969) Biological denitrification of wastewaters by addition of organic materials. Proceedings of the 24th Industrial Waste Conference. Part Two. 1271-1285 McCarty, P .L . , Rittmann, B.E. & Bouwer, E.J. (1984) Microbial processes affecting chemical transformations in groundwater. In \"Groundwater Pollution Microbiology\", G. Bitten & C P . Gerba (eds). John Wiley & Sons Mercado, A . (1976) Nitrate and chloride pollution of aquifers: A regional study with the aid of a single cell model. Water Resources Research, 12(4):731-747 Molz, F.J. , Widdowson, M .A . & Benefield, L .D. (1986) Simulation of microbial growth dynamics coupled to nutrient and oxygen transport in porous media. Water Resources Research, 22(8): 1207-1216 Murphy, E . M . , Davis, S . N . , Long, A . , Donahue, D. & Jull, A.J.T. (1989) Characterization and isotopic content of organic and inorganic carbon in the Mi lk River aquifer. Water Resources Research, 25(8): 1893-1905 248 Nachtnebel, H.P. & Kovar, K. (eds) (1991) Hydrological basis for ecologically sound management of soil and groundwater. IAHS Publication No. 202 National Research Council (1990) Groundwater models: Scientific and regulatory applications. National Academy Press Nesbitt, H.W. & Cramer, J J . (1993) Genesis & evolution of HCCyrich and S0 4 - r ich groundwaters of quarternary sediments, Pinawu, Canada. Geochimica et Cosmochimica Acta, 57:4933-4946 Nicholson, R.V. , Cherry, J .A. & Reardon, E.J. (1983) Migration of contaminants in groundwater at a landfill: A case study: 6. Hydrogeochemistry, Journal of Hydrology, 63:131-176 Oberdorfer, J .A. and Peterson, F.L. (1985) Waste-water injection: geochemical and biogeochemical clogging processes. Ground Water. 23(6):753-761 Olson, R.K. ed (1992) The role of created and natural wetlands in controlling non-point source pollution. Ecological Engineering, 1(1/2), Special Issue Paul, J.W. & Beauchamp, E.G. (1989a) Effect of carbon constituents in manure on denitrification in soil. Canadian Journal of Soil Science, 69:49-61 Paul, J.W. & Beauchamp, E.G. (1989b) Denitrification and fermentation in plant-residue ammended soil. Biology & Fertility of Soils, 7:303-309 Paul, J .W., Beauchamp, E.G. & Trevors, J.T. (1989) Acetate, propionate, butyrate, glucose and sucrose as carbon sources for denitrifying bacteria in soil. Canadian Journal of Microbiology, 35:754-759 Paul, J.W. & Zebarth, B.J. (1992) Influence of time of manure application on leaching and denitrification - 1991 highlights. Technical Report No. 86, Agriculture Canada Agassiz Research Station Paul, J.W. & Zebarth, B.J. (1993) Denitrification in manured soils in the Fraser Valley: how important is it? In \"Soil management effects on soil and water quality\", E.Wolterson & H. Hamilton (eds). Proceedings of the 16th BC Soil Science Workshop Pan, J. (1994) Comparative effectiveness of discharge and input control for reducing nitrate pollution. Environmental Management. 18(l):33-42 Payne, W.J . (1981) Denitrification. John Wiley & Sons Perlmutter, N . M . & Koch, E. (1972) Preliminary hydrogeologic appraisal of nitrate in groundwater & streams, southern Nassau County, Long Island, New York. USGS 249 Professional Paper 800-B:B225-B235 Phillips, K J . & Gelhar, L .W. (1978) Contaminant transport to deep wells. Journal of the Hydraulic Division, ASCE , 104(HY6):807-819 Piteau Associates (1994) Impact assessment of sewage effluent on groundwater quality in an unconfined aquifer, Lower Fraser Valley, British Columbia. Project 92-620. Prepared in conjunction with Dayton & knight Ltd. , for Conservation & Protection, Environment Canada & BC Ministry of Health. June 1994 Polprasert, C. and Park, H.S. (1986) Effluent denitrification with anaerobic filters. Water Research. 20(8): 1015-1021 Postma, D., Boesen, C , Ktistiansen, H. & Larsen, F. (1991) Nitrate reduction in an unconfined sandy aquifer: water chemistry, reduction processes, and geochemical modelling. Water Resources Research, 27(8):2017-2045 Pratt, P.F. (1972) Nitrate in the unsaturated zone under agricultural lands. USEPA water Pollution Control Research Series: 16060DOE04/72 Pratt, R.F. , Jones, W.W. & Hunsaker, V .E . (1972) Nitrate in deep soil profiles in relation to fertilizer rate and leaching volume. Journal of Environmental Quality, 1(1):97-102 Pratt, P.F. & Jury, W . A . (1984) Pollution of the unsaturated zone with nitrate. In \"Pollutants in porous media: The unsaturated zone between the soil surface and groundwater\", B. Yaron et al. (eds). Ecological Studies 47. Springer-Verlag Priscu, J .C. & Downes, M.T. (1987) Microbial activity in the surficial sediments of an oligotrophic and eutrophic lake, with particular reference to dissimilatory nitrate reduction. Archive fuer Hydrobiologie, 108(3):385-409 Reardon, E.J., Dance, J.T. & Lolcama, J .L. (1983) Field determination of cation exchange properties for calcareous sand. Ground Water, 21(4):421-428 Reilly, T .E. , Plummer, L . N . , Phillips, P.J. & Busenberg, E. (1994) The use of simulation and multiple environmental tracers to quantify groundwater flow in a shallow aquifer. Water Resources Research, 30(2):421-433. Ricketts, B.D. & Liebscher, H. (1994) The geological framework of groundwater in the greater Vancouver area. In \"Geology and geological hazards of the Vancouver region, southwestern British Columbia\", geological Survey of Canada Bulletin 481 Rijtema, P.E. & Kroes, T .G. (1991) Nitrogen modelling on a regional scale. In \"Nitrate contamination\", I. Bogardi & R.D. Kuzelka (eds). NATO ASI Series Volume 250 G30. Springer-Verlag Rittmann, B.E. (1993) The significance of biofilms in porous media. Water Resources Research. 29 (7), 2195-2202 Robertson, W.D . & Cherry, J .A. (1989) Tritium as an indicator of recharge and dispersion in a groundwater system in central Ontario. Water Resources Research, 25(6): 1097-1109 Robertson, W . D . , Cherry, J .A. & Sudicky, E.A. (1991) Ground-water contamination from two small septic systems on sand aquifers. Ground Water, 29(1):82-92 Robertson, W.D . & Cherry, J .A. (1995) Insitu denitrification of septic-system nitrate using reactive porous media barriers: Field trials. Ground Water, 33(1):99-111 Robertson, L .A . & Kuenen, J .G. (1990) Denitrification by obligate and facultative autotrophs. In \"Autotrophic microbiology and one-carbon metabolism\", G.A. Codd et al. (eds). Kluwer Academic Publishers Roberston, L .A . & Kuenen, J .G. (1992) Nitrogen removal from water and waste. In \"Microbial control of pollution\", J .C. Fry et al. (eds). Cambridge University Press Ronen, D., Berkowitz, B. and Magaritz, M . (1989) The development and influence of gas bubbles in phreatic aquifers under natural flow conditions. Transport in Porous Media, 4:295-306 Ronen, D., Magaritz, M . & Levy, I. (1986) A multilayer sampler for the study of detailed hydrochemical profiles in groundwater. Water Research, 20(3):311-315 Ronen, D., Magaritz, M . , Almon, E. & Amiel, A . J . (1987a) Anthropogenic anoxification (\"eutrophication\") of the water table region of a deep phreatic aquifer. Water Resources Research, 23(8): 1554-1560 Ronen, D., Magaritz, M , Gvirtzman, H. & Garner, W. (1987b) Microscale chemical heterogeneity in groundwater, Journal of Hydrology, 92:173-178 Ronen, D., Magaritz, M . , Weber, U. , Amiel, A . J . & Klein, E. (1992) Characterization of suspended particles collected in groundwater under natural gradient flow conditions. Water Resources Research, 28(5): 1279-1291 Rott, U. & Lamberth, B. (1992) Subterranean denitrification for the treatment of drinking water. Water Supply, 10(3): 111-120 Rugowski, A .S . (1990) Estimate of the groundwater pollution potential of an agricultural watershed. Agricultural Water Management, 18:209-230 251 Savage, E.S. and Chen, J J . (1975) Operating experiences with columnar denitrification. Water Research. 9:751-757 Schipper, L .A . , Cooper, A . B . & Dyck, W.J . 91991) Mitigating nonpoint-source nitrate pollution by riparian zone denitrification. In \"Nitrate contamination\", I. Bogardi & R.D. Kuzelka (eds), NATO ISI Series, Vol . G30. Springer-Verlag Shepherd, M .A . (1993) Poultry manure and nitrate leaching. World's Poultry Science Journal, 49(2): 171-172 Simmons, R.C. , Gold, A . J . & Groffman, P . M . (1992) Nitrate dynamics in riparian forests: Groundwater studies. Journal of Environmental Quality, 21:659-665 Sims, J.T. (1986) Nitrogen transformations in a poultry manure amended soil: temperature and moisture effects. Journal of Environmental Quality, 15(l):59-63 Smith, R.L . , Howes, B.L. & Duff, J .H. (1991a) Denitrification in nitrate-contaminated groundwaters: occurrence in steep vertical geochemical gradients. Geochimica et Cosmochimica Acta, 55:1815-1825 Smith, R.L . , Harvey, R.W. & LeBlanc, D.R. (1991) Importance of closely spaced vertical sampling in delineating chemical and microbiological gradients in groundwater studies. Journal of Contaminant Hydrology, 7:285-300 Smith, R .L . , Ceazan, M . L . , Brooks, M . H . (1994) Autotrophic, hydrogen-oxidizing, denitrifying bacteria in groundwater; Potential agents for bioremediation of nitrate contamination. Applied & Environmental Microbiology, 60(6): 1949-1955 Soares, M . I . M . , Braester, C , Belkin, S. and Abelovich, A . (1991) Denitrification in laboratory sand columns: carbon regime, gas accumulation and hydraulic properties. Water Research. 25(3):325-332 Solomon, D.K. , Poreda, R.J., Schiff, S .L . & Cherry, J .A. (1992) Tritium and helium 3 as groundwater age tracers in the Borden aquifer. Water Resources Research, 28(3):741-755 Spalding, R.F. & Exner, M.E . (1980) Areal, vertical & temporal differences in groundwater chemistry: I. inorganic constituents. Journal of Environmental Quality, 9(3):479-483 Spalding, R.F. & Exner, M.F . (1993) Occurrence of nitrate in groundwater - A review. Journal of Environmental Quality, 22:392-402 Spalding, R.F. , Exner, M .E . , Martin, G.E. & Snow, D.D. (1993) Effects of sludge disposal on groundwater nitrate concentrations. Journal of Hydrology, 142:213-228 252 Spalding, R.F. & Kitchen, L .A. (1988) Nitrate in the intermediate vadose zone beneath irrigated cropland. Ground Water Monitoring Review, 8:89-95 Spalding, R.F. & Parrott, J .D. (1994) Shallow groundwater denitrification. The Science of the Total Environment, 141:17-25 Starr, R.C. & Cherry, J .A. (1994) Insitu remediation of contaminated groundwater: the funnel-and-gate system. Ground Water, 32(3):465-476 Starr, R.C. & Gillham, R.W. (1993) Denitrification and organic carbon availability in two aquifers. Ground Water, 31(6):934-947 Staver, K .W. & Brinsfield, R.B. (1991) Monitoring agrochemical transport into shallow unconfined aquifers. In \"Groundwater residue sampling design\", R.G. Nash & A .R . Leslie (eds). ACS Symposium Series 465. American Chemical Society Sustainable Poultry Farming Group (1994) Standardizing measurements of nutrient content and density of poultry manure. Tanji, K .K . (1980) Problems in modelling nonpoint sources of nitrogen in agricultural systems. In \"Environmental impact of nonpoint source pollution\", M.R. Overcash & J . M . Davidson (eds). Ann Arbor Science Taylor, S.W. and Jaffe, P.R. (1990a) Biofilm growth and related changes in physical properties of a porous medium. 1. experimental investigation. Water Resources Research. 26(9):2153-2159 Taylor, S.W. and Jaffe, P.R. (1990b) Biofilm growth and related changes in physical properties of a porous medium. 3. dispersivity and model verification. Water Resources Research. 26(9):2171-2180 Taylor, S.W. and Jaffe, P.R. (1990c) Substrate and biomass transport in a porous medium. Water Resources Research. 26(9):2181-2194 Taylor, S .W. , Mi l ly , P .C.D. and Jaffe, P.R. (1990) Biofilm growth and related changes in physical properties of a porous medium. 2. permeability. Water Resources Research. 26(9):2161-2169 Tiedje, J . M . , Sextone, A . J . , Myrold, D.D. & Robinson, J .A. (1982) Denitrification: Ecological niches, competition and survival. Antonie van Leeuwenhoek, 48:569-583 Tiedje, J . M . (1988) Ecology of denitrification and dissimilatory nitrate reduction to ammonium. In \"Biology of anaerobic organisms\", A J . B . Zehnder (ed). John Wiley & Sons 253 Tortora, G.J. , Funke, B.R. & Case, C .L . (1989) Microbiology - An introduction, 3rd Edition. The Benjamin/Cumings Publishing Company Inc. Tredoux, G. & du Plessis, H . M . (1992) Situation appraisal of nitrate in groundwater in South Africa, water Supply, 10(3):7-16 Tucker, D.O., Randall, C.W. and King, P.H. (1974) Columnar denitrification of a munitions manufacturing wastewater. Proceedings of the 29th Industrial Waste Conference, Purdue University, 165-175 USEPA (1992) Proceedings of the National Rural Clean Water Program Symposium -10 years of controlling agricultural nonpoint source pollution: the RCWP experience. Summary Publication EPA/625/R-92/006 Van Beek, C . G . E . M . , Hettinga, F . A . M . & Straatman, R. (1989) The effects of manure spreading and acid deposition upon groundwater quality at Vierlingsbeek, the Netherlands. In \"Groundwater contamination\", L. Abriola (ed), IAHS Publication No. 185 Van Der Ploeg, R.R., Ringe, H. & Machulla, G. (1995) Late fall site-specific nitrate upper limits for groundwater protection purposes. Journal of Environmental Quality, 24:725-733 Vandevivere, P. and Baveye, P. (1992) Effect of extracellular polymers on the saturated hydraulic conductivity of sand columns. Applied & Environmental Microbiology. 58(5): 1690-1698 Verdegem, L. & Baert, L. (1984) Losses of nitrate nitrogen in sandy and clayey soils. 1. Leaching of N 0 3 - N to the subsoil; hints to chemical N 0 3 - N reduction; N 0 3 - N losses by artificial drainage. Pedologie, 34(3):235-255 Verdegem, L. & Baert, L. (1985) Losses of nitrate nitrogen in sandy and clayey soils. 2. A qualitative and quantitative approach to the chemical NC^-N reduction in reduced subsoils. Pedologie, 35(l):39-54 Vogel, J .C. (1967) Investigation of groundwater flow with radiocarbon. In \"Isotopes in Hydrology\", Proceedings of 1966 Symposium, International Atomic Energy Agency Viswanathan, M . N . (1984) Recharge characteristics of an unconfined aquifer from the rainfall-water table relationship. Journal of Hydrology, 70:233-250 Walther, W. (1989) The nitrate leaching out of soils and their significance for groundwater: Results of long term tests. In \"Nitrogen in organic wastes applied to soils\", J. A A . Hansen & K. Henriksen (eds), Academic Press 254 Ward, C .A . , Tikuisis, P. and Tucker, A .S . (1986) Bubble evolution in solutions with gas concentrations near the saturation value. Journal of Colloid Interface Science. 113(2):388-398 Wassenaar, L.I., Aravena, R., Fritz, P. & Barker, J.F. (1991) Controls on the transport and carbon isotopic composition of dissolved organic carbon in a shallow groundwater system, central Ontario, Canada. Chemical Geology, 87:39-57 Weier, K .L . , Doran, J .W. , Mosier, A . R . , Power, J.F. & Peterson, J .A. (1994) Potential for bioremediation of high nitrate irrigation water via denitrification. Journal of Environmental Quality, 23:105-110 Weil , R.R., Kroontje, W. & Jones, G.D. (1979) Inorganic nitrogen and soluble salts in a Davidson clay loam used for poultry manure disposal. Journal of Environmental Quality, 8(1):86-91 Westerheide, D.E. ad Westwater, J.W. (1961) Isothermal growth of hydrogen bubbles during electrolysis. Journal of the American Institute of Chemical Engineers. 7(3): 357-362 Widdowson, M . A . , Molz, F.J. & Benefield, L .D. (1988) A numerical transport model for oxygen- and nitrate-based respiration linked to substrate and nutrient availability in porous media. Water Resources Research, 24(9) Widdowson, M .A . (1989) Modelling of nitrate transport coupled to denitrification in the saturated zone. In Proceedings of the 1989 ASCE National water Conference, T.A.Austin (ed) Wild, A . (ed) (1988) Russel's soil conditions and plant growth, 11th edition. Longman Scientific & Technical Wilderer, P .A. , Jones, W . L . & Dan, U. (1987) Competition in denitrification systems affecting reduction rates and accumulation of nitrite. Water Research, 21(2):239-245 Yamaguchi, T., Moldrup, P., Teranishi, S. and Rolston, D.E. (1990) Denitrification in porous media during rapid, continuous leaching of synthetic wastewater at saturated water flow. Journal of Environmental Quality. 19:676-683 Young, J.C. and McCarty, P.L. (1969) The anaerobic filter for waste water treatment. Journal of the Water Pollution Control Federation. 41(5), Part 2, R160-R173 Zebarth, B.J., Dean, D . M . , Kowalenko, C .G. , Paul, J.W. & Chipperfield, K. (1994) Improved manure nitrogen management in raspberry production. Agassiz Research Station Technical Report No. 96, Agriculture Canada 255 Zebarth, B J . (1995) personal communication Zheng, C , Bradbury, K.R. & Anderson, M.P . (1988) Role of interceptor ditches in limiting spread of contaminants in groundwater. Ground Water, 26(6):734-742 256 APPENDIX A EVALUATION OF A TRAVEL TIME ESTIMATION PROCEDURE FOR UNCONFINED GROUNDWATER FLOW This Appendix presents an evaluation of the analytical approach described by equation (2.13), in Chapter 2, for estimating the travel time (and, therefore, the age) of recharge waters beneath the water table of unconfined aquifers. The usefulness of this analytical approach for estimating groundwater travel times was briefly discussed by Appelo & Postma (1993). The objective of this Appendix is to consider the site data recently reported in Reilly et al. (1994), as well as the site data of Postma et al. (1991), to evaluate the relative accuracy of equation (2.13). Data of Reilly et al. (1994) Figure A l presents a map of the site area, including well locations and groundwater flow directions, as contained in Reilly et al. (1994). The depth of the surficial sand and gravel aquifer at this site over the investigation area is approximately 20m. Recharge over the region has an estimated range of 0.20m/y to 0.51m/y. The porosity of the underlying aquifer materials has an estimated range of 0.25 to 0.40. The aquifer system is considered to be relatively homogeneous, but anisotropic. Figure A2 presents a model grid used by Reilly et al. to simulate the cross sectional flow over the region shown on Figure A l . Figure A2 also shows the location of several multilevel well \"clusters\". Table A l presents a comparison of groundwater age at various well locations as estimated by Reilly et al. from the analysis of the groundwater chlorofluorocarbon (CFC) 257 -SO\u00E2\u0080\u0094 \u00E2\u0080\u0094 161, 160. 15 51. EXPLANATION GENERAL LOCATION OF SIMULATED CROSS SECTION WATER-TABLE CONTOUR-Shows altitude of water table. Dashed where approximately located. Contour interval 5 feet. 58 and 59-foot contours are dotted. Datum is sea level WELL CLUSTER AND IDENTIFICATION NUMBERS SINGLE WELL AND IDENTIFICATION NUMBER GENERAL DIRECTION OF GROUND-WATER FLOW Figure A l Site map of area investigated by Reilly et a l . (1994) showing location of wells and water table configuration, as well as the general location of the region simulated by the cross section shown on Figure A2 Inactive cells above water table N CO ii B c > o CO 4> t/J U > o X) ca U \u00E2\u0080\u00A2o 3 *-\u00C2\u00BB \u00E2\u0080\u00A2a < 18.3 m (60 ft) 9.1 m (30 ft) Free surface with constant flux boundary -9.1 m (-30 ft) 2440 m (8000 ft) 1220 m (4000 ft) Horizontal Distance, in meters (feet) 0 m (0 ft) Figure A2 Model grid used by Reilly et al. (1994) to simulate the groundwater flow region identified in Figure A l 259 content, and from a refined numerical simulation of the groundwater flow performed using a trial and error approach. The data contained in Table A l shows that Reilly et al. were able, after what appears to have been considerable manipulation of the recharge configurations as well as aquifer properties, to model the CFC age estimates relatively accurately. The values for well depth (below the water table) shown on Table A l have been estimated for the present evaluation by using Figure A2 and counting the number of grid squares above the well location. Data of Postma et al. (1991) Figure A3 presents a map of the site area investigated by Postma et al. (1991). The land use and encountered nitrate distributions are shown on Figure 2.12, in Chapter 2 of this thesis. The site is underlain by an unconfined glaciofluvial sand aquifer with the Water table at a depth of approximately 15m below ground surface. The thickness of the aquifer is variable, ranging from 90m near the groundwater divide, to 20m at the downstream discharge zone. Recharge at the site is 0.375m/year. Postma et al. provided no values for the porosity of the sand deposit at the site. For the purposes of the present evaluation it is assumed that the porosity of the site soils has an average value of 0.33 (based on the data provided for the sand and gravel aquifer investigated by Reilly et a l . , 1994). At the location of well T2, Postma et al. obtained a profile of the tritium concentrations in the groundwater. Atmospheric nuclear testing is considered to have produced a tritium peak in precipitation during 1963. By also computing a date for the groundwater at the water table at the time of tritium measurement (1988), Postma et al. Figure A3 Site map of area investigated by Postma et al. (1991) showing the location of monitoring wells and water table configuration, as well as the location of a groundwater divide ro o 261 estimated the date/age profile to be as shown on Figure A4. Postma et al. estimated the time of travel to the water table at the site to be approximately 3 years. So in 1988, the water at the top of the water table originated at the ground surface in 1985. As the tritium peak is associated with the year 1963, in 1988 the age of this peak relative to the water table was 22 years. Age estimation using Equation 2.13 Equation 2.13 was used to compute the age of the groundwater below the water table at the two sites, and by using the ranges of values for aquifer characteristics as noted above. Plots of the computation results are presented in Figure A5. Figure A5a presents curve plots of the age versus depth below water table (d) for the site of Reilly et a l . , as computed for two conditions; i) porosity (e) of 0.25 and recharge (P) of 0.51m/y, and ii) porosity of 0.40 and recharge of 0.20m/y. Figure A5a also contains a plot of the CFC and numerically modelled age data contained in Table A l . Figure A5b presents curve plots of age versus depth below water table for the site of Postma et al. (1991), as computed for two aquifer thickness (D) conditions; i) 90m, and ii) 20m. Figure A5b includes a plot of the single age value (22 years) for the tritium peak as obtained from Figure A4. Discussion On the whole, Figure A5 shows that equation (2.13) can be used as a reasonably accurate first estimate of the groundwater age beneath the water table of unconfined Figure A4 Tritium concentration profile at the location of well T2, shown on Figure A3, as well as estimated year age based on tritium peak association to 1963, as presented by Postma et al. (1991) Age below Water Table (years) Figure A5 Age versus depth below water table as estimated using equation (2.13) for a) site of Reilly et al. (1994), including age data from Table A l ; b) site of Postma et al. (1991), including tritium peak age ro co 264 aquifers receiving surface recharge. The required input parameters of porosity, aquifer thickness and annual recharge can be estimated for any given site relatively easily and in a comparatively accurate manner. Furthermore, Figure A5a shows that by considering the low and high values of the input parameters it is possible to obtain an envelope of estimated age ranges at any given depth, and from which a \"best\" (average) estimate can be made. 265 Table A l : Age data from Reilly et al. (1994) Well Depth below CFC Age 2nd Simulation Number WT (m) (years) Age (years) 59 3.4 6 6.0 164 10.3 24 23.5 53 0.6 6 0.8 52 4.6 8 8.1 61 9.1 21 17.5 62 4.6 4 3.9 163 9.1 9 8.8 162 17.1 35 36.7 APPENDIX B DIALYSIS MEMBRANE EVALUATION 266 A series of laboratory tests was performed to evaluate the performance of the dialysis membrane (SPECTRUM Spectra/Por 1, M W C O 6,000-8,000 regenerated cellulose, tubular form, flat width 32mm) towards quantifying the dissolved chemistry of the encountered groundwater at the study wells. Testing was performed generally at a temperature of 10\u00C2\u00B0C, a condition which has been presumed to reflect the groundwater environment. The membrane was obtained from the supplier in rolls of 100ft (30m). As supplied, the membrane contains glycerol (HOCH 2 CH.OH.CH 2 .OH) as a humectant. In order to remove the glycerol, the membrane was cut into the required lengths (typically 0.3m for this research) and rinsed in a flow-through tap water bath for 24 hours, in general accordance with the manufacturer's recommendations. On completion of rinsing, each membrane segment was knotted by hand at one end, filled with distilled water, and then knotted at the other end to produce one \"sausage\" shaped membrane bag. The knotting technique used was found to result in no trapped air bubbles within the membrane sausages. The volume of water enclosed in each sausage prepared in this manner is approximately 50ml. The laboratory testing consisted of placing a number of dialysis membrane sausages, containing distilled water, into a bath of water having known chemistry, and monitoring the diffusion characteristics of the various dissolved constituents into the membrane enclosed distilled water. Additionally, testing aimed at evaluating the potential for organic carbon leaching from the membranes into the enclosed waters was also 267 performed. Three series of tests were performed, as described below. Series 1: Diffusion of major ions into membrane enclosed water A water bath was prepared by dissolving 1.93g Mg(N0 3 ) 2 .6H 2 0; 1.52g K N 0 3 ; 1.10g CaCl 2 .2H 2 0; 1.28g N a H C 0 3 ; and 1.30g CaS0 4 .2H 2 0 into 15 litres of distilled water. These masses were selected to produce major ion concentrations at the levels encountered at the Abbotsford study well. The water bath was kept in a cool room maintained at a temperature of 10\u00C2\u00B0C. Twenty (20) dialysis membrane sausages, prepared as described above, were then placed in the water bath. Two sausages were removed at predetermined time intervals, and perforated to collect the enclosed water into nalgene bottles, similar to the procedure adopted in the field. At the same time that the sausages were removed from the bath, a 60ml sample of the bath water was also collected. The collected samples were then stored in a refrigerator maintained at 4\u00C2\u00B0C until the completion of sampling (72hrs) Sample portions to be tested for HCOy were frozen, except for the 48hr and 72hr samples which were stored at 4\u00C2\u00B0C. At the completion of sampling, analyses for major ions (N0 3\", S0 4 2 \" , CI\", HC0 3 \" , C a 2 + , N a + , M g 2 + , and K + ) was performed on the collected samples in accordance with the methodology described in Chapter 3. HC0 3 \" was quantified using a Total Carbon Analyzer, and by assuming that the measured IC was all in the HC0 3 \" form. The TC analyses also provided data on the organic carbon (TOC) content of the collected samples. The results of the analyses are presented on Tables BI and B2. Plots of the obtained data are presented in Figures BI and B2. Table B I . Major anion chemistry and TOC data from Series 1 Testing Bath water Membrane Enclosed Water T IME (hrs) N0 3 \" (mg-N/1) so42-(mg/1) ci-(mg/1) IC (mg/l) TOC (mg/l) N 0 3 -(mg-N/1) so42-(mg/1) ci-(mg/1) IC (mg/l) TOC (mg/l) 0 27.61 49.4 27.84 8 0 0.22 1.23 0.43 0 0 0.5 27.08 47.52 27.62 8 0 6.56 6.62 8.66 0 0 1.0 26.52 47.12 27.32 7 0 11.30 7.75 13.46 2 1 2.0 25.94 46.0 27.10 5 0 16.82 11.40 19.45 2 1 4.0 25.99 46.46 26.87 6 0 22.16 18.20 24.80 3 0 8.0 25.84 45.44 26.68 8 0 25.06 27.07 25.59 5 1 12.0 25.78 45.13 26.73 7 0 25.50 32.65 26.93 3 1 16.0 25.76 44.39 26.90 9 0 25.69 36.70 27.05 4 1 24.0 25.74 44.31 26.74 9 0 25.68 40.36 26.90 9 0 48.0 25.79 44.25 26.70 11 0 25.58 42.86 27.01 11 2 72.0 25.73 43.95 26.71 11 0 25.55 43.10 27.03 11 2 269 Table B2. Major cation chemistry data from Series 1 Testing Bath water Membrane Enclosed Water T IME (hrs) C a 2 + (mg/l) N a + (mg/l) M g 2 + (mg/l) K + (mg/l) C a 2 + (mg/l) N a + (mg/l) M g 2 + (mg/l) K + (mg/l) 0 37.6 24.5 12.4 64.5 0.1 0.1 0.2 0 0.5 37.4 24.3 11.9 64.5 6.8 5.7 2.0 16.7 1.0 38.2 23.9 12.3 63.7 11.5 9.3 3.2 24.8 2.0 34.7 23.9 12.4 63.2 17.9 14.0 5.2 36.0 4.0 37.8 23.4 11.6 63.5 25.9 19.3 8.1 51.4 8.0 36.0 23.2 12.0 66.1 31.9 21.9 10.2 58.0 12.0 37.0 22.9 11.3 65.3 33.3 22.9 11.0 62.6 16.0 34.6 22.9 11.7 62.8 35.6 22.6 11.6 61.5 24.0 35.8 22.8 11.6 64.8 36.4 23.0 12.2 61.5 48.0 35.8 22.8 11.8 63.9 35.0 23.5 12.5 60.9 72.0 36.7 23.0 .11.5 62.8 37.5 23.1 12.2 63.4 270 12 24 36 48 60 72 Time (hours) Figure BI . Major cation chemistry data of bath water (open circles) and membrane enclosed waters (solid circles) from Series 1 tests 271 \u00E2\u0080\u00A2 30 25 20 15 10 5 0 50 40 30 20 10 0 30 25 20 15 10 5 0 12 4 - \u00E2\u0080\u00A2 . g i \u00C2\u00ABj : \" \u00E2\u0080\u00A2 f 12 24 36 48 Time (hours) 60 72 Figure B2 Major anion chemistry data of bath water (open circles) and membrane enclosed waters (solid circles) from Series 1 tests 272 With the exception of HC0 3 \" and S0 4 2 \" , the obtained results show that equilibration of the chemistry between the membrane enclosed water and the surrounding bath water occurs within 24hrs. S0 4 2 \" concentrations within the membrane enclosed water is seen to essentially reach equilibrium within 48hrs. The slight decrease in the concentrations within the bath waters is explained by the fact that the membrane enclosed water resulted in an increase in the total water volume within the bath, and as the dissolved constituents diffused into the bags, the concentration within the surrounding bath water decreased until equilibrium was reached. The reason for the slightly higher measured values for M g and Na in the membrane enclosed waters, in comparison to the bath waters, is not obvious, but this could be a result of analyzer drift. As noted in Table B I , TOC concentrations in the membrane enclosed waters were 0 or lmg/1, except for the sausages left in the bath for 48 and 72hrs and which had TOC concentrations of 2mg/l. The reason for the erratic nature of the measured concentrations of HC0 3 \" in both the bath and membrane enclosed waters sampled during the initial 24hr period is not clear. Freezing of the samples may have had some impact, and in order to evaluate this further a second set of testing was performed as described below. Series 2: Diffusion of HC03\" and organic carbon across membrane 3g of N a H C 0 3 and 3ml of denatured ethyl alcohol (containing equal volumes of ethyl acetate, hydrocarbon solvents and methyl iso-butyl ketone) were dissolved into 11 of distilled water and added to the bath water that remained after Series 1, and after 273 approximately 31 of this water was removed for Series 3 testing. Twenty membrane sausages were then placed in the bath, and sampling performed as per Series 1. the bath and membrane enclosed water samples were stored in a refrigerator, but not frozen, and then analyzed for TOC and IC. The results of this series of testing are presented in Table B3 and Figure B3. On the whole, these results show that HC0 3 \" and the ethyl alcohol diffuse into the dialysis bags within 24 hrs at the concentrations used for the testing. The reason for the fluctuations in, and the discrepancy between, the bath water and bag-enclosed water TOC concentrations is not readily apparent. Series 3: Leaching of organic carbon from membrane As noted earlier, the membrane as obtained contains glycerol. An earlier test which consisted of placing a distilled water sausage made with unrinsed dialysis membrane in a 1000ml graduated flask of distilled water, found that after a few days, the water surrounding the bag had a TOC concentration of about 20mg/l. It is presumed that this TOC is glycerol that had leached into the water. However, this TOC analysis was performed as an afterthought, as the main purpose of the test was to evaluate the relative buoyancy of the prepared sausage. Additionally, the 48hr and 72hr TOC data obtained from the Series 1 testing suggested that long term submergence of the membrane may lead to TOC contributions into the enclosed waters from the membrane itself. The Series 3 tests were therefore performed to evaluate the likelihood of organic carbon contributions from rinsed membrane into the enclosed water on prolonged 274 Table B3. IC and TOC data from Series 2 Testing Bath water Membrane Enclosed Water T IME IC TOC IC TOC (hrs) (mg/l) (mg/l) (mg/l) (mg/l) 0 41.4 160 0 0 0.5 41 158 3.5 29.1 1.0 41 154 8.2 45.5 1.5 41 152 12.3 60.1 2.0 40.6 150 14.8 67.8 4.0 40.2 148 25.1 105.5 7.0 39.9 145 31.9 121.9 11.0 40 145 36.7 131.9 14.0 39.5 144 38 136.2 24.0 40.2 142 39.7 132.7 48.0 39.9 142 40.2 126.8 72.0 40.5 157 40.7 156.1 275 175 - r 0 12 24 36 48 60 72 Time (hours) Figure B 3 TOC and IC data for bath water (open circles) and membrane enclosed waters (solid circles) from Series 2 tests 276 storage. Approximately 3 litres of bath water that remained after Series 1 testing was placed in a nalgene jug. Three membrane sausages were placed in the jug, which was then left at room temperature for one week. Simultaneously, three membrane sausages Were placed for two weeks in a distilled water (approximately 5 litres) bath that was kept in the cool room at a temperature of 10\u00C2\u00B0C. This colder temperature and two week time period were selected to mimic the insitu groundwater temperature and sampler residence time. The bag enclosed waters were then analyzed for TOC. The results of this testing are presented in Table B4. The obtained results show that the dialysis membrane as prepared for the field sampling contributes some TOC into the enclosed waters. It is not known whether this TOC is glycerol, or some other form that has leached from the membrane. Table B 4 . TOC (mg/l) data from Series 3 Testing Membrane Bag Room Temp. 10\u00C2\u00B0C No. Bath1 Bath2 1 2.6 3.6 2 2.9 3.8 3 2.9 4.4 Note 1: - bath water contained same chemistry as for Series 1 - bags left in bath for one week Note 2: - bath water contained distilled water - bags left in bath for two weeks 277 APPENDIX C MEASUREMENT OF DISSOLVED NITROGEN & OXYGEN BY DIRECT INJECTION USING GAS CHROMATOGRAPHY Under the conditions of the laboratory column testing performed as part of this research, a technique was required by which representative samples of the column effluent (and influent) could be collected and analyzed for dissolved N 2 and O z . The low flow rates adopted (0.5 to 1.25 ml/min) posed a constraint on large volume sample collection, due to the time required for sample discharge and collection, and the associated potential for atmospheric contamination. Smith et a l . , (1991), reported the use of gas chromatography (GC) for quantification of dissolved N 2 as part of their investigations into groundwater denitrification, but did not provide specific details on sample size or injection procedure. The use of gas chromatography for the measurement of dissolved N 2 and 0 2 was demonstrated by Swinnerton et al. (1962a,b), who described a procedure including the stripping of the dissolved gas from solution prior to injection into the GC apparatus. Swinnerton et al. developed a glass injection chamber containing a fritted glass divide, to be used for the stripping process. This technique was reported to be effective for water samples as small as l -2ml. Alternate means of \"stripping\" dissolved gases include diffusion across a gas permeable membrane (Bilstad et a l . , 1978), or sample agitation in a helium headspace, with subsequent analysis of the headspace (Fendinger & Adams, 1986). However, both these techniques also require large water samples. As part of the current research, a glass sampling chamber was manufactured according to the description of Swinnerton et al. (1962a,b) and connected to a Hewlett 278 Packard 5750 Research Chromatograph equipped with a thermal conductivity detector. Molecular Sieve 5A was selected for the GC column packing, as this material was shown by Mindrup (1978) and Supina (1974) to be effective for separating N 2 and 0 2 . However, attempts at using the chamber stripping approach, particularly for small water samples, were not successful in resolving the N 2 - 0 2 content. A poorly manufactured injection septum, as well as an inappropriate frit size, led to this failure. This stripping approach was therefore abandoned. The injection of small water samples directly into the GC column, without any prior stripping, was then investigated on the assumption that gas stripping would occur within the upgradient portion of the column. These investigations showed, surprisingly, that water samples as low as 100/A could be directly injected to provide effective and repeatable N 2 and 0 2 peak separation. The adopted \"favourable\" injection and GC column conditions consisted of: - column diameter & type: 2mm stainless steel - column length: 2m - column temperature: 30\u00C2\u00B0C - column packing: Molecular Sieve 5A - carrier gas: Helium - carrier gas flow rate: 20ml/min - injection port temperature: 120\u00C2\u00B0C - detector type: thermal conductivity (TC) - detector temperature: 200\u00C2\u00B0C Though this approach was effective, it was found that such direct injection could only be performed for a total of ten (10) lOO^tl water samples at any given time before peak separation deteriorated - presumably as a result of water accumulation in the GC column. However, in accordance with Supina (1974), it was found that the column could be re-activated after each set of injections by raising the GC column temperature to 190\u00C2\u00B0C 279 (selected to be no greater than the TC detector temperature) for a period of at least 12hrs, and then lowering to 30\u00C2\u00B0C prior to the next set of injections. The effectiveness of the developed direct injection technique in quantifying the dissolved N 2 and 0 2 was investigated by injecting various small-sized samples of air as a standard set, and comparing the measured N 2 and 0 2 peaks (measure of gas mass) with peaks obtained from injection of water samples. Additionally, as nitrate and alcohol were added to the flow waters in the experiments, the significance of these additives towards N 2 -0 2 separation was investigated by injecting 100/xl samples of water containing various concentrations of each additive. Results of A i r Injections The data obtained from the series of air injections (ambient air from near the GC was sampled for this series) is shown on Table CI , and the averages plotted on Figure CI . The volume size range of the air samples was based on trial & error selection to encompass the range of N 2 -0 2 peak areas that could be expected from injection of lOOjitl samples of effluent water, and which would be free of dissolved 0 2 and possibly saturated with dissolved N 2 . Figure CI shows that the peak separation and subsequent mass detection results in a linear relationship between the injected volume, V, and measured N 2 -0 2 peak area, P. The lines of best fit (visual) plotted on Figure CI have the following relationships: 280 P \u00C2\u00BB 2 = 1200V a i r (CI) Po2 = 336V a i r (C2) By considering that, at STP, 1 mole of ideal gas occupies a volume of 22.421, or 0.02242m3, (Anderson et a l . , 1983), and that l m 3 of air contains 0.78m3 N 2 and 0.21m3 0 2 (Manahan, 1991), it can be computed that the mass densities of N 2 and 0 2 in air are: 9.74 x 1Q-* \"?N? =1 Z u E 15-O z 10-COLUMN 1 EFFLUENT NOx & TOC 80 medium sand TIME (days) 120 NOx TOC 0.8-0.7-0.6-0.5-6 o-4-z <8 t 0.3-s z 0.2-0.1-COLUMN 1 EFFLUENT NO-2, NH-4 & IC medium sand -12 -10 i t %*VaV*^ ! ! I \u00E2\u0080\u00A2 I ^ A A A A > M ' A A * A * * * > * 90 95 100 105 110 115 120 TIME (days) \u00E2\u0080\u00A214 NH-4 \u00E2\u0080\u00A2 NO-2 -*\u00C2\u00AB- IC FIGURE G4 340 COLUMN 1 EFFLUENT NOx & TOC medium sand TIME (days) -m- NOx -o- TOC COLUMN 1 EFFLUENT NO-2, NH-4 & IC medium sand 0.8-T R 1 4 TIME (days) - S - NH-4 - A - NO-2 -*\u00C2\u00AB- IC F I G U R E G 5 341 30-25-20-E 15-COLUMN 2 EFFLUENT NOx & TOC medium sand -90 -80 -70 -60 TIME (days) NOx TOC 0.8-0.7-0.6-0.5-6 0-4-z \u00C2\u00AB3 7 0.3-0.2-0.1 COLUMN 2 EFFLUENT NO-2. NH-4 & IC medium sand 10 20 30 40 TIME (days) -7 -6 a, -4 u -3 -2 NH-4 -jfc- NO-2 - * \u00C2\u00AB - IC F I G U R E G6 342 COLUMN 2 EFFLUENT NOx & TOC medium sand TIME (days) -m- NOx -O- TOC COLUMN 2 EFFLUENT NO-2, NH-4 & IC medium sand 0.8-r- - \u00E2\u0080\u0094 g 80 90 100 110 120 130 1 40 150 160 TIME (days) -SB- NH-4 -Jkr NO-2 . -X- IC F I G U R E G 7 343 25 2 0 -Z E 1 5 -O z COLUMN 3 EFFLUENT NOx & TOC fine gravel 30 40 SO TIME (days) NOx - B - TOC 0 .7 -COLUMN 3 EFFLUENT NO-2, NH-4 & IC fine gravel - 3 5 0.6- - 3 0 0 . 5 -M j=. 0 .4 -r\u00C2\u00AB 6 z \u00C2\u00AB 0 . 3 -- 2 5 a, -15 \u00C2\u00A3 0 .2 -0 . 1 -10 \u00E2\u0080\u0094I\u00E2\u0080\u0094 10 - 1 \u00E2\u0080\u0094 20 \u00E2\u0080\u0094i\u00E2\u0080\u0094 30 40 50 TIME (days) NH-4 -A- NO-2 -a*- IC F I G U R E G8 344 30-E 15 O Z COLUMN 3 EFFLUENT NOx & TOC fine gravel 110 120 130 TIME (days) 140 \u00E2\u0080\u0094i\u00E2\u0080\u0094 150 100 90 -80 70 60 50 \"a s o p 40 E\" -30 -20 10 -0 160 NOx - S - TOC 0.7-COLUMN 3 EFFLUENT NO-2, NH-4 & IC fine gravel 100 110 120 130 TIME (days) 140 \u00E2\u0080\u0094I\u00E2\u0080\u0094 150 \u00E2\u0080\u00A235 -30 -25 \u00E2\u0080\u00A220 ^ a, is a 10 -0 160 S- NH-4 -+r NO-2 IC F I G U R E G 9 345 30-25 -2 0 -E 15-10-COLUMN 4 EFFLUENT NOx & TOC fine gravel 30 40 50 TIME (days) 100 90 I-80 70 NOx - B - TOC 1.2-\u00C2\u00A7 0.8-I 6 0 .6-z <* z \u00C2\u00B0'4~ 0.2-COLUMN 4 EFFLUENT NO-2, NH-4 & IC fine gravel 10 20 30 40 50 TIME (days) 16 \u00E2\u0080\u00A214 \u00E2\u0080\u00A212 \u00E2\u0080\u00A210 \u00E2\u0080\u00A28 H - 6 -4 NH-4 -A- NO-2 IC F I G U R E G 1 0 346 30-25-20-Z E 15-o z 10-80 COLUMN 4 EFFLUENT NOx & TOC fine gravel \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 90 100 110 120 TIME (days) \u00E2\u0080\u0094 i \u00E2\u0080\u0094 130 140 150 100 90 80 -70 \u00E2\u0080\u00A260 ~ \u00E2\u0080\u00A2a \u00E2\u0080\u00A2so & u o \u00E2\u0080\u00A240 H -30 -20 10 -0 160 NOx TOC 1.2-COLUMN 4 EFFLUENT NO-2, NH-4 & IC fine gravel 90 100 110 120 130 TIME (days) 140 150 16 \u00E2\u0080\u00A214 \u00E2\u0080\u00A212 10 \u00E2\u0080\u00A2a he a. o \u00E2\u0080\u00A26 -2 -0 160 NH-4 NO-2 IC FIGURE GJLL 347 20-COLUMN 5 INFLUENT/EFFLUENT NOx medium sand -1.2 18-16-I* 12^ O Z 10-r0.8 -0 .6 B, 6 z h0.4 -0.2 10 15 20 TIME (days) 25 30 35 40 effluent NOx - a i r influent NOx effluent NO-2 35-30-25-=5, \u00C2\u00AB H S, o ID-S ' COLUMN 5 INFLUENT/EFFLUENT TOC & IC medium sand El -35 -30 -25 -20 e -is a -10 El 35 40 TIME (days) effluent TOC -air influent TOC -*\u00C2\u00AB - effluent IC F I G U R E G12 348 COLUMN 6 INFLUENT/EFFLUENT NOx medium sand 2 0 R 3 . 5 TIME (days) - B - effluent NOx TIT influent NOx effluent NO-2 COLUMN 6 INFLUENT/EFFLUENT TOC & IC medium sand TIME (days) -S- effluent TOC T\u00C2\u00ABV influent TOC - * \u00C2\u00AB - effluent IC FIGURE G.13 "@en . "Thesis/Dissertation"@en . "1996-05"@en . "10.14288/1.0050360"@en . "eng"@en . "Civil Engineering"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "Investigations into the distribution of non-point source nitrate in two unconfined aquifers and the role for carbon addition in the control of nitrate concentrations in groundwater"@en . "Text"@en . "http://hdl.handle.net/2429/6301"@en .