Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Water table height and nitrate leaching in undisturbed soil columns Elder, Linda A. 1988

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1988_A6_7 E42.pdf [ 8.12MB ]
Metadata
JSON: 831-1.0058894.json
JSON-LD: 831-1.0058894-ld.json
RDF/XML (Pretty): 831-1.0058894-rdf.xml
RDF/JSON: 831-1.0058894-rdf.json
Turtle: 831-1.0058894-turtle.txt
N-Triples: 831-1.0058894-rdf-ntriples.txt
Original Record: 831-1.0058894-source.json
Full Text
831-1.0058894-fulltext.txt
Citation
831-1.0058894.ris

Full Text

W A T E R T A B L E H E I G H T A N D N I T R A T E L E A C H I N G IN U N D I S T U R B E D S O I L C O L U M N S by LINDA A. ELDER B.Sc.(Agr.), The University of Cuelph, 1982 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF BIO-RESOURCE ENGINEERING We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 1988 © LINDA A. ELDER, 1988 ln 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. 1 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 BIO-RESOURCE ENGINEERING The University of British Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date: April 1988 ii A B S T R A C T Water table control by subsurface drainage has been shown to affect leaching losses of nitrate-nitrogen: a concern both for economic use of fertilizer, and for maintenance of water quality. The effect of water table height on leaching of N O j ' - N was investigated in this study in nineteen 15cm x 100cm undisturbed cores of silty clay loam. The experiment simulated fertilization fol lowed by rainfall, then rapid water table rise and fall, under conditions similiar to those experienced in the early spring in the Lower Fraser Valley. In the first part of the experiment, a concentrated solution of KNO3 and K G (equivalent to 35 kg/ha of N and 22 kg/ha of Cl) was applied to the columns, fol lowed by intermittent leaching with distilled water. Leachate from two depths in each column was collected before and after a period of static water table, and analyzed for NO3", N0 2 ", N H 4 + , and C l " . This procedure was repeated without nutrient addition in the second part of the experiment. Chloride was used an an inert tracer to follow anion movement and retention within the columns. There was no significant difference in the leachate NO3" concentration or leachate N/CI ratio from any of the four water table heights tested (15, 35, 55, and 75 cm above drain depth). The NO3" concentrations and N/CI ratios decreased with depth in the soil columns, indicating removal of N from the percolating soil solution, either by denitrification or immobilization. The variability in leachate concentrations among all columns was very high (eg. for a typical sample time, N 0 3 " - N ranged from 0.01 to 15.72 mg/L, and C l " ranged from 4.8 to 14.5 mg/L), as was the variability in constant head satiated hydraulic conductivities (range: 1 to 1468 cm/day; CV = 181%), and drainable porosity (range: 2.7 to 10.4%; CV = 39%). Cross sections of columns leached with 1% methylene blue solution did not reveal differences in patterns of water transmission between low and high conductivity columns. Indications were that penetration of dye was greater in columns with higher conductivities, and that preferential f low occurred in all columns examined. Leachate concentrations and N/CI ratios correlated significantly with iii hydraulic conductivity: Spearman's correlation coefficients were always > 0.8 for samples obtained from the bottom of the columns. However, even when the conductivity was included as a covariate in an analysis of covariance, there was no significant effect of water table height on nitrate leaching. iv T A B L E O F C O N T E N T S ABSTRACT ii TABLE OF CONTENTS...., iv LIST OF TABLES vi LIST OF FIGURES vii ACKNOWLEDGEMENT viii I. INTRODUCTION 1 II. LITERATURE REVIEW 3 A. Soil Nitrogen Cycling : 3 1. Mineralization, Immobilization & Nitrification 3 2. Denitrification 5 B. Water and Solute Movement 10 1. Displacement 10 2. Preferential Flow 12 C. Seasonal Leaching Patterns 17 D. Fertilization 19 1. Quantity 19 2. Timing 22 E. Drainage 24 F. Study Methods 31 1. Size and Nature of Experimental Unit 31 2. Conservative Tracers 35 G. Summary 35 III. MATERIALS AND METHODS 37 A. Overview of the Experiment 37 B. Site and Soil Description 37 C. Acquisition and Preparation of the Column 40 D. Field Samples 41 E. Storage Conditions During the Experiment 41 F. Column Set-up 42 G. Physical Measurements 47 1. Edge Flow 47 2. Hydraulic Conductivity 49 3. Dye Tests 49 V H. Run Procedures 50 1. Run 0 50 2. Run 1 51 3. Pre-Flush 52 4. Run 2 52 5. Runs 3 and 4 53 I. Nutrient Application Rates 54 J. Chemical Analyses 55 K. Experimental Design and Statistical Analyses 58 IV. RESULTS AND DISCUSSION 60 A. Field Samples 60 B. Physical Measurements 60 1. Drainable Porosity 60 2. Hydraulic Conductivity 62 3. Dye Tests 67 C. Runs 0 to 2 Leachate Concentrations 74 1. Ortho-Phosphate 74 2. Chlorides 77 3. Nitrate and Nitrite Nitrogen 79 4. Ammonium Nitrogen 82 D. Runs 3 and 4 Leachate Concentrations 82 1. Distribution of the Data 82 a. Chloride 82 b. Nitrogen 89 c. N/CI Ratios 91 2. Recovery of Solute 91 3. Trends with Time 93 4. Effect of Hydraulic Conductivity 101 a. Chloride 101 b. Nitrogen 107 5. Effect of Water Table 108 V. SUMMARY AND CONCLUS ION 113 VI. LITERATURE CITED 117 VII. APPENDIX A. Summary of ANOVA results 126 VIII. APPENDIX B. Summary of ANCOVA results 136 vi LIST O F T A B L E S Table I. Soil physical and chemical characteristics 39 Table II. Nutrient application rates 56 Table III. KCI-extractable nitrate and ammonium in field samples 61 Table IV. Physical characteristics of soil columns 63 Table V. Runs 0 - 2 ortho-phosphate concentrations 75 Table VI. Runs 0 - 2 chloride concentrations 78 Table VII. Runs 0 - 2 nitrate + nitrite concentration 80 Table VIII. Runs 0 - 4 leachate ammonia concentration 83 Table IX. Runs 3 and 4 chloride concentration 84 Table X. Runs 3 and 4 nitrate plus nitrite concentration 85 Table XI. Runs 3 and 4 N/CI ratios 90 Table XII. Solute recoveries 92 Table XIII. Wilcoxon matched-pairs test for trends with time 96 Table XIV. Runs 3 and 4 overflow chloride concentrations 97 Table XV. Runs 3 and 4 overflow nitrate + nitrite concentrations 98 Table XVI. Runs 3 and 4 overflow N/CI ratios 99 Table XVII. Spearman correlation coefficients for concentration with hydraulic conductivity 102 Table XVlll. Summary of ANOVA and A N C O V A results 111 vii LIST O F F I G U R E S Figure 1. Column configuration 43 Figure 2. Halfway sampler design 45 Figure 3. Flow-separating bottom 48 Figure 4. Relationship between drainable porosity and hydraulic conductivity 66 Figure 5. Dye patterns in column 1 68 Figure 6. Dye patterns in column 6 69 Figure 7. Dye patterns in column 7 70 Figure 8. Relationship between Run 4 Final Drainage CI- concentration and water table height 86 Figure 9. Relationship between Run 4 Final Drainage N 0 3 - concentration and water table height 87 Figure 10. Relationship between Run 4 Final Drainage N/CI ratios and water table height 88 Figure 11. Mass balance for nitrate for Runs 3 & 4 Grand Total 94 Figure 12. Relationship between Run 4 Final Drainage CI- concentrations and hydraulic conductivity 103 Figure 13. Relationship between Run 4 Final Drainage N 0 3 -concentrations and hydraulic conductivity 104 Figure 14. Relationship between Run 4 Final Drainage N/CI ratios and hydraulic conductivity 105 viii A C K N O W L E D G E M E N T S I wish to extend my sincere appreciation to Dr. S.-T. Chieng for continued support and advice, and to Dr. A.A. Bomke for consultation and encouragement. I would also like to thank Dr. P. Liao, Mr. N. Jackson and Mr. J. Pelke for technical assistance, and Dr. K.V. Lo and Professor L.M. Staley for reviewing this work. In addition, my thanks are extended to fellow students who helped with the dirty, heavy work throughout the research - Chris Lague and Buff Mitchell, and especially to Paul Richard, who provided much encouragement and stimulating discussion. Of course, my deep gratitude goes to my family, for unquestioning encouragement throughout this work. Finally, very special thanks are extended to Richard Senger, for unending assistance at all stages of this work, and continuous encouragement and support. This study was financially supported, in part, by a postgraduate scholarship from the Natural Sciences and Engineering Research Counci l of Canada and by a University of British Columbia Graduate Fellowship, awarded to the author and gratefully acknowledged. 1 I. INTRODUCTION Nitrogen in agriculture has a long history of study: reviews on N in agricultural soils commonly cite work dating centuries into the past (Tisdale et al., 1985; Russell, 1973). In spite of the great efforts expended to understand the fate of this essential nutrient, our understanding of the N cycle is still imperfect in many areas, and inadequate enough in some to present problems for efficient management. In current times agriculture is faced with two great challenges: to produce high yields in a cost-effective manner and, increasingly, to minimize the negative environmental impact of agricultural activities. Nitrogen management plays a prominent role in both of these concerns. As an essential and often limiting nutrient, N must be managed for greatest availability at the time it is needed most by a growing crop, in order to ensure maximum return on the fertilizer investment. Nitrogen is present in the environment in many different compounds, several of which are very mobile and may be " lost " from the soil-crop system (eg. nitrates, dinitrogen, nitrogen oxides and ammonia gas). When removed from the agricultural system, some of these N compounds may present environmental concerns. Nitrogen oxides produced by denitrification have recently become a concern as potential contributors to acidic precipitation. Nitrates are potential pollutants of both ground water and surface waters, presenting potential health threats to humans and livestock and being implicated in cultural eutrophication of certain surface waters. The 10 mg NO3-N/L recommended maximum concentration of nitrate in drinking and surface waters is often exceeded in ground water in agricultural areas and in subsurface drainage discharge (Rosswail & Paustian, 1984; Zwemnan et al., 1972; Lucas & Reeves, 1980). Nitrate has been called "the major potential soluble water pollutant in humid regions" (Walter et al., 1979). Nitrate leaching is dependent on soil moisture content, ln humid, north temperate areas such as the lower coastal mainland of British Columbia, regulation of soil moisture by subsurface drainage is an essential component of successful agricultural operations. While 2 many studies have examined the occurrence of N loss related to the presence of drainage systems, relatively few have investigated the intensity of drainage as a management option for N control. Water table depth, a design parameter in drainage planning, is relatively easily manipulated and measured, yet it is a main factor in very few experiments. The few studies which have manipulated water table height have shown the potential for control of NO3" leaching by drainage management. Existing work also indicates that there is the potential for nitrate leaching from the crop root zone in B.C.'s lower mainland (Kowalenko, 1987; Richard & Chieng, 1985). Water table depth was therefore chosen as the principle variable for this study on the effect of drainage intensity on nitrate leaching. Objectives The specific objectives of this study were: 1. To investigate the effect of water table height in undisturbed soil cores on the leaching of applied nitrate-nitrogen; and 2. To investigate the behavior of N 0 3 _ - N in the saturated zone below the water table by comparing nitrate concentrations at two depths in the profile. 3 II. LITERATURE REVIEW A. Soil Nitrogen Cycling The soil N cycle has been extensively researched and literature on the subject is overwhelming in volume. This discussion of N cycling will be confined to the principle factors affecting leaching losses of nitrate in cold, wet, heavy soils, and is not an attempt to provide an exhaustive review of the literature. 1. Mineralization, Immobilization & Nitrification Mineralization is the conversion of organic N to inorganic N of the NH3, NH^"*" and NO3' forms. Aminization followed by ammonification produces ammonia (NH3) from organic compounds such as proteins. Ammonia is soluble and exists in pH-dependent equilibrium with the ammonium ion (NH^" 1") in aqueous systems. Nitrification is the oxidation of ammonium to nitrate (NO3") via nitrite (NO2"): 2 N H 4 + + 3 0 2 - - > 2 N 0 2 " + 2H 2 0 + 4 H + (1) 2 N 0 2 " + 0 2 --> 2 N 0 3 " (2) The conversion is accomplished in two distinct steps by separate groups of microorganisms. Immobilization is the conversion of the inorganic N compounds discussed above into organic forms of N via uptake by soil microorganisms (plant assimilation of N and N 2 fixation are usually not considered to be immobilization; Jansson & Persson, 1982). Aminization, ammonification and immobilization are activities of heterotrophic soil microorganisms. Nitrification is most significantly accomplished by autotrophic bacteria, but various heterotrophic microorganisms are capable of nitrifying activity. The balance between the opposing activities of inorganic N production and uptake is often termed "net mineralization" or the mineralization-immobilization turnover (MIT) (Jansson & Persson, 1982). If net mineralization is very low there will be, obviously, little NO3" available for leaching. Immobilization will be significant under conditions favouring rapid growth of the general soil microbial biomass. However, immobilization of N requires a carbon source. NH^"*" is usually 4 available for nitrification because the microbial growth constituting immobilization is often limited by the availability of C and an energy source (Schmidt, 1982). The balance between the 3 forms of inorganic N resulting from ammonification and nitrification is usually toward accumulation of the nitrate ion, as the rate of conversion of NO2' to NO3" is faster than either N H 4 + or N0 2 " formation (for well-drained neutral to slightly acid soils; Tisdale et al., 1985). Given the presence of adequate populations of appropriate microorganisms, the factors affecting mineralization of N include pH, temperature, aeration and moisture content. Each factor will be discussed briefly below. Generally, conditions favorable for nitrification also favor ammonification (Tisdale et al., 1985), but nitrification often exhibits narrower requirements than ammonification. pH. According to Belser (1982), nitrification is "much more" sensitive to pH than ammonification. There appears to be quite a range of reported values for the extreme limits and even optimum activity of nitrifiers. The best summary is perhaps by Schmidt (1982), who said that "Most observations indicate an arbitrary lower limit for nitrification of pH 4.0, obvious nitrification in the pH 4-6 range, and pH independent nitrification in the range 6 to 8." Temperature. The upper temperature limit for nitrification appears to be lower than the optimum temperature for ammonification (Belser, 1982). The absolute responses to temperature seem to vary regionally, indicating an adaptation of local nitrifying populations to regional climate (Belser, 1982; Schmidt, 1982; Russell, 1973). At the lower temperature extreme, Tisdale et al. (1985) report "significant" nitrate formation in Iowa soils between 0 - 2 C, and Malhi and Nyborg (1986) found increases of mineral N of 0.35 to 0.70 kg N/ha/day in continuously frozen Black Chernozems, in a field study in north-central Alberta. At the upper temperature extreme, Schmidt (1982) reports an Australian study finding nitrification up to 60 C. Optimum ranges vary also. The often-quoted study of Mahendrappa et al. (1966) indicates optima varying from 20 - 25 C for soils in the northwest U.S. to 30 - 35 C for soils in the southwest U.S. 5 The temperature coefficient (Q-JQ) f ° r mineralization is 2 for 5 - 35 C (Tisdale ef al., 1985). However, it has been necessary to use a Q 1 f J of 3 to accurately model a Swedish clay loam overlaying fine sand in a cold-temperate and semi-humid climate (Johnsson et a/., 1987). Aeration. The reactions of nitrification are oxidations and the most important nitrobacteria are obligate aerobes (Tisdale et a/., 1985). Thus, the aeration status of a soil is very important to the occurrence of nitrification. While most reviews emphasize the need for well-aerated soil, Russell (1973) cites studies indicating nitrification is possible even at very low C>2 concentrations (0.3 or 0.4% C^). Aeration is reduced usually in response to increasing moisture levels or intense microbial activity. In contrast to nitrification, ammonification is not inhibited by anaerobic conditions (Belser, 1982). Moisture Content. Again, soil nitrifying response to moisture content varies with soil type (Russell, 1973). At high moisture contents the effect on nitrification is primarily through reduced aeration as C>2 diffuses much more slowly through water than through air. In some cases, as moisture content increases, nutrients may solubilize and become mobile, stimulating respiratory consumption of 0 2 (Tiedje ef al., 1984). Nitrification appears to proceed most rapidly at moisture tensions between 1/3 and 1 bar; above 15 bars the nitrification rate declines (Russell, 1973; Tisdale ef al., 1985). Nitrification is also more sensitive to moisture stress than is ammonification, at either moisture excess or extreme deficit (Russell, 1973; Belser, 1982). 2. Oenitrification Denitrification is the microbially-mediated reduction of nitrates and nitrites to the gaseous forms dinitrogen (N 2 ) and nitrous oxide ( N 2 0 ) : 2 H N O s + 4 H + --> 2 H N 0 2 + 2 H 2 0 2 H N 0 2 + 2 H + --> 2NO + 2 H 2 0 2NO + 2 H + - > N 2 0 + H 2 0 N 2 0 + 2 H + --> N 2 + H 2 0 6 While gaseous products can result at various stages of the N cycle, denitrification in the strictest definition is the use of N oxides as terminal electron acceptors in respiration (Firestone, 1982; Colburn & Dowdell, 1984). The microorganisms capable of this activity are primarily facultative anaerobes, most being chemoheterotrophs, some being lithotrophs (using reduced S or H 2 as electron donors) and a few being capable of autotrophic activity (Firestone, 1982; Tisdale et al., 1985). Denitrification has been recognized as a component of the N cycle for a very long time. However, the conditions under which it proceeds and the nature of the reaction products have made intensive study of denitrification difficult until quite recently. For example, it was not until the mid-1970's that the acetylene reduction technique for inhibiting conversion of to N 2 allowed accurate quantification of denitrification (Yoshinari & Knowles, 1976; Balderston et al., 1976). Denitrification in soils is strongly affected by many edaphic and climatic factors, which will be discussed briefly below. The information on factors affecting denitrification has been generated largely with small, disturbed soil samples or even pure cultures in laboratory studies. While the trends in response are likely transferable to field situations, the laboratory studies are oversimplifications of very complex systems, and therefore laboratory results are not necessarily directly transferable to the field. For example, commonly air-dried and ground soils are used to study denitrification potential. However, it has been shown that denitrification capacity is significantly different in fresh and air-dried soils and that the response changes with the duration of storage in the air-dried state (Patten ef al., 1980). Carbon Supply. Both laboratory and field studies indicate the importance of carbon availability to denitrification. There is general agreement in the literature that the form of C is important - the C must be readily "available" to denitrifying organisms - but there is not a general consensus on which form of C most accurately predicts denitrification response in all soils. Water-soluble organic carbon and mineralizable carbon were found by Burford and Bremner (1975) to have highly significant relationships with denitrification capacity, while 7 Stanford et al. (1975) found glucose-C correlated better than total organic C. More recently, Beauchamp et al. (1980) measured denitrification rates on the A, B, and C horizons of 9 soil series representing 3 natural drainage classes and 7 textural groups. In their study, in contrast to Burford and Bremner's or Stanford's results, denitrification rate correlated more highly with total organic C than with water-extractable, mineralizable, or 0.1 N Ba(OH)2-extractable C. No estimate of C correlated well for soils low in available C, and, in poorly-drained soils the regression of mineralizable or water-soluble C on denitrification was non-significant. Thus until more information is available, it may be necessary to determine the most appropriate analysis of C for denitrifying activity in each different soil type. The importance of available C to denitrification is illustrated by the study of Gilliam and Gambrell (1978). When the content of available C was low enough to be limiting (0.1% C, analytical method not specified) the denitrification rate was almost the same at 5, 15 and 25 C. The response to temperature increased considerably when the C content was higher (0.9% C). Aeration. N oxides are used as electron acceptors by denitrifying bacteria only when the energetically more efficient electron acceptor, 0 2 , is not available in sufficient quantity. Therefore denitrification requires low or nonexistent 0 2 levels. A variety of studies have shown that denitrification can occur at soil atmospheric 0 2 concentrations of 4 to 17% (Firestone, 1982). A number of parameters can be measured to estimate soil oxygen status, including 0 2 level in the soil atmosphere, various measures of soil water content (eg. % air-filled porosity, water potential), and redox potential. Redox potentials between 300 and 650 mV appear to support "significant" denitrification in some soils (Firestone, 1982). Not only is the 0 2 content of the soil atmosphere important, but oxygen may also be carried deep in the profile as dissolved 0 2 in percolating water, especially in the colder temperatures of winter (Gilliam & Gambrell, 1978). Moisture Content. Increasing the moisture content of a soil may result in dramatic increases in denitrification. However, the effect of high moisture content is primarily to reduce oxygen availability. 8 pH. While denitrification is relatively insensitive to pH in the range pH 6 - 8, it does appear to decrease under acid conditions (Firestone, 1982), although the extent of inhibition is unclear. Several studies have measured significant rates of denitrification at a pH lower than 5 (Gilliam & Gambrell, 1978; Firestone, 1982). For example, Gilliam & Gambrell (1978) measured the effect of pH and temperature on denitrification in the topsoil and subsoil of two acidic North Carolina coastal soils, with and without addition of a C energy source. Although denitrification was considerably slower without carbon amendment and in the subsoils compared to the topsoils, denitrification proceeded readily. The authors concluded that pH < 5 would not be a "serious limiting factor in NO3" reduction" in their soils. Temperature. As with nitrifiers, temperature tolerance of denitrifying organisms appears to vary with climatic region (Firestone, 1982). Firestone's review (1982) reports minimum temperatures for denitrification of 2.7 to 10 C, and maximums of about 75 C. Nitrate Concentration. Obviously, nitrite or nitrate must be present before denitrification can occur. The dependence of denitrification rate on nitrate concentration has been found to be first order at lower N concentrations (<40 mg/L) and zero order at higher concentrations (Firestone, 1982; Kanwar ef al., 1982). First order and zero order kinetics described the disappearance of nitrate equally well in a study by Gilliam & Gambrell (1978), although first order kinetics were better when denitrification was rapid. Soil Texture. The influence of soil texture on denitrification results from the environment created by the textural type. Some studies have found evidence for increased denitrification losses with increasing clay content of the soil (eg. Avnimelech & Raveh, 1976; Devitt et al., 1976; Dowdeil et al., 1980), because the increased water holding capacity and poorer natural drainage of clay result in greater nutrient retention and lower oxygen status. Values of denitrification coefficients used in modeling N transformations are typically higher for heavier soil fractions (eg. Rijtema, 1980). Devitt ef al. (1976) found only a low potential for denitrification in coarse-textured profiles, and that changes in the CI/NO3" ratio and the redox potential were related to the position of the clay layer. The correlation between denitrification 9 and increased clay content is not unconditional, however. Cast et al. (1974) studied 2 soils, finding no evidence of denitrification in the subsoil with the higher clay content and possibly lower aeration. They speculated that the root distribution and exudation may have played a role in the difference in denitrification between the soils (presumably implying a C-limited situation in the soil with the higher clay content). Interaction of factors. The complex and dynamic nature of soil make it very difficult to predict a quantitative effect of any single factor on denitrification, or many other nutrient cycling reactions. The magnitude of influence that a particular factor has on denitrification will often depend on interaction with a number of other factors. For example, temperature affects C» 2 solubility, diffusion and consumption, possibly resulting in complex interactions. Craswell (1978) found that there was an interaction between temperature and water content in denitrification: as the temperature of incubation decreased, the soil water content required for denitrification to occur increased. Also, Bailey and Beauchamp (1973) found that, at a particular water content (1/2 field capacity or field capacity), denitrification occurred at a high nitrate concentration but not at a low nitrate concentration. These two studies may be interpreted to mean that the inhibitory effect of 0 2 on denitrification may be less at higher temperatures and higher nitrate concentration. In the field the relationship between parameters is not always clear. In an intensive study of a small area of cultivated silt loam, Burton & Beauchamp (1985) found the denitrification rate to be highly variable, and that there were no consistently significant statistical relationships between N 2 0 production and 12 measured soil physical and chemical parameters: air-filled porosity related negatively to denitrification rate; weak positive relationships existed for water-soluble and total organic C with denitrification. Examples of strong unexpected correlations also occurred, such as to extractable K + . As with many soil chemical and microbiological phenomena, denitrification reactions occur at a micro rather than a macro scale. It is conditions at microsites, rather than average conditions within the bulk of the soil, that will determine whether a specific reaction will 1 0 proceed. Microsites may be saturated soil aggregates of sufficient diameter to have anaerobic interiors (Smith, 1980) or they may be sites of very intense microbial activity, where 0 2 consumption exceeds supply. Local accumulations of readily-available C may promote such intense microbial activity and become "hot-spots" for denitrification (Parkin, 1987). B. Water and Solute Movement As nitrate is a negative ion for which most soils have little or no adsorption capacity, nitrate moves readily with the soil water (Bohn et a/., 1979). Thus, understanding the movement of water in soil is essential to understanding the leaching of nitrate from soils. 1. Displacement In the simplest view of soil water movement, water added to the top of a mass of soil infiltrates the soil pores, displacing existing soil water ahead of it. Water is held approximately until field capacity is reached, then the "wetting front" moves deeper into the soil under the influence of gravity. This model of water movement has been called miscible or uniform displacement, plug flow, or Darcian flow, and has been popular for a considerable period of time (Thomas & Phillips, 1979). To understand solute and water movement more completely, consider the fate of a particular " p lug " of water or solute, already present in a saturated soil matrix before the initiation of leaching. In true plug flow, the particular plug will arrive at the bottom of the soil column after the passage of a volume of water equivalent to the pore volume of the soil between the plug's original position and the collecting point. At the collecting point, the concentration of solute in the plug (C) would be unchanged with respect to its initial concentration (Co): i.e. C/Co = 1 (Greenwood & Burns, 1979). In reality, the plug concentration is subject to attenuation by several processes, including molecular diffusion and hydrodynamic dispersion. Diffusion is relatively more important at low flow rates, such as when hydraulic conductivity is low or precipitation or irrigation rates are low. Dispersion is a mixing resulting from flow velocity variations due to the non-uniform nature of soil pores, and also from density differences between resident and 11 invading solutions. Dispersion is affected by pore geometry (diameter, tortuosity and the presence of dead-end pores) and total pore length, as well as by electrostatic phenomena such as anion exclusion (Greenwood & Bums, 1979; Starr & Parlange, 1976). Due to dispersion and diffusion, real breakthrough curves do not look like step-change functions, but usually are sigmoidal with a gradual approach to a maximum C/Co (of less than 1) at approximately 1 pore volume. In the simplest cases with considerable uniformity of the soil pores, the breakthrough curve will be symmetric about the midpoint (Starr & Parlange, 1976). Soil columns prepared in the laboratory by drying, grinding, sieving and packing soil commonly exhibit plug flow of mobile solutes. The results of Ritchie et al. (1972) are typical of many studies in the literature. In their study, disturbed columns were prepared by packing 17 cm diameter columns with Houston Black clay that had been air-dried, passed through a 1 mm sieve and mixed. Fluorescein-tagged water was applied to the core until fluorescein appeared in the effluent. At that point, the distribution of fluorescein throughout the column was very uniform, except near the bottom. (Fluorescein adsorbs to soil particles, marking the transmission route of percolating water). The uniform distribution of fluorescein indicated that the fluorescein moved as a plug and that the entire soil volume participated in water movement. Uniform displacement is not confined to laboratory columns of disturbed soil, but also exists in some undisturbed soils which have little or no structure. For example, Tyler and Thomas (1981) applied an extremely concentrated slug of chloride to the surface of almost-saturated, undisturbed cores of Bruno sandy loam, then leached the cores with deionized water. The breakthrough curves for this "virtually structureless" soil showed a slow approach to maximum C/Co at about 1 pore volume, indicating that almost complete displacement of soil water had occurred. In a field situation where displacement is a significant mechanism of soil water movement, nitrate (and other mobile, non-adsorbing solutes) will be lost by leaching whenever internal drainage occurs in the soil and nitrogen is present as nitrate. 12 2. Preferential Flow Many soils in their natural state are not structureless, but consist of aggregates of various sizes and shapes, penetrated and divided by roots, and old root channels, the tunnels of macroinvertebrates (such as earthworms) and perhaps also by the constantly-changing cracks of shrink-swell clays undergoing changes in moisture content. In soil with moderately-well to well developed structure, infiltrating water is often observed to penetrate these cracks or channels first or fastest, rather than to uniformly displace water already held in the matrix. Such a flow pattern has been called channel flow, crack flow, macropore flow, preferential flow and bypassing flow. It is evidenced by peak leachate concentrations in pulse-input experiments appearing well before 1 pore volume of effluent, and by breakthrough curves that start earlier and are more diffuse than the ideal steep sigmoidal curve for plug flow (Omoti & Wild, 1979; White, 1985a; Starr & Parlange, 1976). The concept of bypassing flow is not new, and is usually attributed to the work of Lawes and others with the Rothamsted drain gauges in 1882 (discussed by, the example, Thomas and Phillips, 1979). According to Thomas and Phillips, Lawes and colleagues recognized that the drainage from a structured soil consisted first of relatively unaltered infiltrating water and later of water more like that held in the soil matrix. In spite of this early recognition of bypassing flow, the concept was largely ignored, especially in modeling solute movement in soils, until the early to mid-1970's. From that time the literature is filled with examples of how miscible displacement failed to predict water and solute movement through soils with macropores. Typically, miscible displacement theory could predict solute movement in disturbed laboratory columns, but consistently underestimated the depth of solute penetration in structured soils (Shuford etal., 1977; Coles & Trudgill, 1985). According to theory, there are two structural domains within the soil: the intraped soil matrix and the interped cracks or channels. The intraped volume (i.e. the interior of soil peds or aggregates) consists of networks of very small pores relative to the pore sizes comprising 13 the interped space. Macropores within peds are sometimes distinguished separately from interped cracks. When water addition to the soil surface exceeds the infiltration rate of the peds, or when peds become saturated, the macropores or interped spaces begin to conduct water. This occurs even if the water potential is too low for the spaces to fill with water completely (White, 1985a). Bouma ef al. (1981) have noted that there are many observations in the literature of macropore flow occurring even in unsaturated clay soils. Within the finer pores of the peds, water movement is relatively slow and solutes are subject to the same dispersive and diffusive movement discussed briefly in the previous section. Water in the macropores is subject to dispersion and convection, but may flow at considerably higher rates than within the peds. The water held within the soil peds is often referred to as immobile or stagnant water because of its relatively slow rate of movement down the soil profile; the water in the macropores is called mobile water (Lawes ef al., 1982, in Addiscott ef a/., 1978). Work with dye tracers illustrates and justifies the mobile/immobile designations. For example, Ritchie ef al. (1972) showed that under steady-state flow conditions, large portions of the total porosity of large (21 and 73 cm diameter) undisturbed cores of a swelling clay did not transmit the water and solute that was added to the tops of the cores. At depths of 20, 35 and 50 cm, the fluorescein-tagged area was only 60%, 10% and 2%, respectively, of the total cross-sectional area of the cores. These results are especially significant in view of the dye patterns obtained with disturbed cores of the same soil, as discussed in the previous section. Diffusive exchange between intraped water and interped water is possible but there is evidence to show that under a wide range of conditions, it does not occur to a considerable extent over short time intervals. Thus the soil matrix concentration is relatively unchanged by macropore flow (Thomas & Phillips, 1979). For example, Shuford ef al. (1977) report that 27 cm of rain from a hurricane infiltrated through a silt loam soil without causing a significant change in soil solution nitrate 14 concentration at 30, 60 or 120 cm depth. Thus there was not a significant amount of diffusion out of the soil matrix into the macropores. Also, considering the large volume of water, if displacement had been the primary method of water movement, the soil solution concentration should have decreased markedly. Even more convincing evidence for the lack of interaction between interped water and matrix water (and for the existence of bypassing flow) is provided by the presence in soil water discharge of chemicals normally quickly adsorbed to soil (White, 1985a). For example, following storm events Bottcher et al. (1981) found 2 soil-adsorbing pesticides and phosphorus in tile drainage from a silty clay soil, and Shaffer et al. (1979) found that surface-applied Cd quickly appeared at 112 cm depth following irrigation of a structured silty clay soil. In fact, even when there are high lateral hydraulic gradients between flowing macropores and dry soil peds, there appears to be little movement of water into the peds from the channels (Bouma ef al., 1981). These findings are somewhat contradicted by more recent work by Cermann ef al. (1984). In a column study of undisturbed silt loam soils they found lateral penetration of bromide into the soil matrix surrounding macropores not only deep in the profile (where there may have been positive pressure from ponded macropores) but also in the surface soil layers. The bromide profiles were measured 5 days after sprinkling, a much longer time interval than the 300 minutes employed by Bouma ef al. (1981). Bouma and Dekker (1977) distinguish between the rapid adsorption of water from macropores by dry peds and subsequent slow penetration of that water within the peds. Additionally, they found more lateral penetration into peds at the surface of the soil than deeper. Although the net diffusive exchange may be small, solute movement across the macropore/ped interface is continuous, in a direction determined by the diffusion gradient. Where the leaching solution is concentrated, solute will move into the peds as water flows down the pores, but as the solute concentration in the pore decreases as the pulse passes, 15 the net solute movement will be back into the pores. This solute will then be available for leaching in the next macropore flow event (Addiscott et al., 1978; Thomas & Phillips, 1979). Thus the concentration of leachate moving through a soil, or discharging from a column or into a drainage system depends on the location of solute with respect to the peds and channels or cracks, at the commencement of leaching. Where a dilute solution leaches a soil (i.e. the soil solution is more concentrated than the leachant), it is typical to see a large volume of dilute leachate from the macropore flow, which usually ends shortly after the irrigation or rainfall ceases, sometimes followed by leachate of increasing concentration (depending on the pedal soil moisture status). This effect is illustrated clearly by data from Coles and Trudgill (1985), showing sharp drops in the nitrate concentration of soil water discharged from plots of weakly-structured silt and clay loam soils. The decreases in nitrate concentration correspond to equally sharp peaks in soil water discharge due to intense rainfall. When the solution applied to the soil is concentrated itself, the leachate concentration pattern changes. Such conditions arise in the laboratory when leaching is done with a pulse of concentrated solution followed by a solution more dilute in the ionic species under consideration. In the field, analogous situations include fertilization followed immediately by rain or irrigation, or perhaps chemigation. In these cases, the concentrated pulse flows down the macropores, with only slight attenuation by diffusion into the ped matrix. Water from the finer pores arrives later, and may be of lower concentration than the early pulse, depending on the ion concentration in the soil water relative to the added solution. It is this type of leaching pattern that was said earlier to be indicative of bypassing flow. Examples of pulse-input experiments with results of the type just discussed include work by Addiscott et al. (1978) and Tyler and Thomas (1981). In the experiment of Tyler and Thomas (1981), already briefly described, a slug of chloride was leached through undisturbed cores of silty clay loam and silt loam, both of which had "distinct structure". In these soils the CI" breakthrough curves peaked well before only 16 0.2 pore volumes. The peaks rose very steeply but dropped off more gradually. Fluorescein patterns in these soils indicated that the dye and water flowed through distinct channels below approximately 15 cm, but above that - in the tilled layer - the dye was more dispersed. Leaching losses of NO3" may be increased by macropores in yet other ways. Kissel ef al. (1974) speculate that nitrification may be greater on the surfaces of macropores than within the rest of the soil mass; this N would be readily available to leaching by rain or irrigation. On a field scale, bypassing flow may cause rapid leaching loss of at least a portion of applied nitrate if application is followed by rainfall or irrigation. Losses are potentially great on strongly structured soils such as cracking clays, but have also been measured on weakly structured soils (Coles & Trudgill, 1985). Nitrate in the plow layer is also vulnerable to leaching with rainfall if the plow layer overlies a structured profile (Kissel ef al., 1974). The amount of macropore flow relative to displacement through the soil matrix depends primarily on the soil structure, including pore size and geometry (see especially Bouma & Dekker, 1977; Bouma & Wosten, 1979), soil water content, tillage (as it affects macroporosity and hydraulic conductivity) and, very importantly, the rate of water addition (Thomas & Phillips, 1979). In reviewing the literature, Thomas and Phillips (1979) found that ratio of macropore flow to displacement ranged from almost 100% to less than 50%. Antecedent moisture content may have a profound effect on the proportion of bypassing flow following a precipitation event. White (1985a) cites studies indicating that the proportion of bypassing flow increases the drier the soil is initially. The proportion of applied water and solute which is lost as bypassing flow increases with increasing rain or irrigation intensity (Bouma ef al., 1981; Cermann ef al., 1984; Kneale & White, 1984), and with increasing volume applied at any particular intensity (Bouma & Dekker, 1978). lf the water carries solute, such as recently applied nitrate, this implies that more solute will be lost as precipitation or irrigation intensity increases. In field studies, Coles and Trudgill (1985) have found that surface-applied nitrate and chloride did penetrate the soil profile more deeply with increasing rainfall intensity. 17 It should be evident at this point that prediction-of solute leaching in structured soils is not a simple matter, but depends on a number of factors. White (1985a) very adequately summarizes the situation: Whether macropore flow that bypasses a fraction of the resident water within soil peds increases or decreases the amount of solute leached depends on: 1. the location of the solute (whether recently applied fertilizer or a soil-generated solute held on the exterior or within peds); 2. the ratio of macropore to matrix flow; 3. the saturated hydraulic conductivity of the matrix; 4. the antecedent water content of the soil; 5. the contact area between the bypass flow and the relatively static water of the matrix; and 6. the rate of solute diffusion between mobile and immobile water volume. It is the interaction of all of these factors that produces the variable responses found in the literature for nitrate leaching in structured soils. C. Seasonal Leaching Patterns Disregarding fertilization practices for the moment, there may naturally be distinct seasonal patterns to nitrate leaching. The patterns are simply due to the responses of the microorganisms and the plant communities involved in the N cycle to changing weather (temperature and precipitation). When the information presented in the previous two sections is integrated with climatic variation, the seasonal pattern of nitrate leaching can be summarized as follows. In the spring there is typically a " f lush" of nitrate production and a subsequent increase in leaching if moisture conditions are favorable. The flush is due to increased mineralization in warm spring weather (Kissel ef al., 1976; Cooke & Williams, 1970) and the leaching may occur with spring rains (Kissel ef al., 1976) or as a result of infiltrating snowmelt. In addition, under spring conditions plant uptake of NO3" is usually slow, and transpiration is low so that there is little to prevent leaching (Kolenbrander, 1969). Spring leaching losses will be greater for uncropped land than land under cover. Therefore leaching may be somewhat less under winter sown crops. There is considerable evidence in the literature that leaching losses of 18 nitrate are much less under perennial crops than under annual crops (Bergstrom, 1987; Bolton etal., 1970; Kolenbrander, 1981). The amount of NO^" leached will be reduced if denitrification is significant. Denitrification may also be favored by the warm moist conditions of spring, if a suitable C source and microbial population are available. If microbial activity is high, reducing conditions may develop very quickly following soil saturation from heavy spring rains (Verdegem ef al., 1981). Rapid peaks in the groundwater table, sometimes even rising near the soil surface, may be seen in response to spring rainfall in humid regions. Over the growing season NO3" leaching is usually minimal (Khanif ef al., 1984), because of NO3" uptake by plants, active microbial growth, and, with increased evapotranspiration, less percolating water. This generalization may be misleading in some situations. For example, Cameron ef al. (1978) found that 21 - 44% of spring-applied CI" was lost by summer leaching on a fallow, structured clay loam near Ottawa, Ontario. Nitrate was also lost by both leaching and denitrification over the summer, but losses were masked by mineralization and/or nitrification. (Significant denitrification during late spring and summer was suggested by the results of an N study conducted simultaneously on the same study site; Kowalenko, 1978.) In the fall after harvest there may be another, small, flush of mineralization (Cooke & Williams, 1970), depending on soil temperatures and the organic matter available for decomposition. Nitrate, either residual from the summer or newly produced, may be leached if sufficient rain is received before the soil freezes. If there is no freeze-up, leaching may continue throughout the winter. Organic matter breakdown and mineralization may also occur throughout the winter in cold, unfrozen, soil (Benoit, 1973). Of course, denitrification may also occur in cold, wet soils, although leaching may predominate as a loss mechanism if there is sufficient percolating water and the temperatures are low enough to restrict or inhibit denitrification. Major leaching losses usually occur between late fall and early spring (Cameron ef al., 1978; Kowalenko, 1987). In humid climates 19 leaching and/or denitrification of NO3" from the root zone may be so severe over winter that little or no NO3" may remain in the soil by spring (eg. Kowalenko, 1987). D. Fertilization The two major issues concerning fertilization and nitrate loss involve the rate and timing of N application. 1. Quantity A great number of studies have found that N leaching from the crop root zone increases as application rates exceed recommended rates. Only a few of the studies available in the literature will be discussed. In a review of primarily European studies, Kolenbrander (1981) presents evidence of increasing nitrate losses in tile drainage with increasing fertilization rate over a wide range of conditions, and emphasizes that leaching losses are usually greater for arable land than grassland, except at very high fertilization rates where the two are comparable. In a 5 year study on moderately to moderately-well drained soil in Iowa, Burwell et al. (1976) found that, for similarly managed, contoured watersheds in continuous corn, fertilization at 448 kg N/ha annually increased N discharge 3 times over that for N fertilization at the recommended rate of 168 kg/ha annually. Average annual losses were 20.7 kg N/ha for the high rate fertilization compared with 6.8 kg/ha for the normal rate (average annual water-weighted N concentrations in drainage were 21.0 and 5.8 ppm, respectively). Miller (1979) reported that the concentration and annual losses of NO-^'-N through tile drainage on clay and sand soils in Ontario increased considerably when the recommended rate of application was exceeded. Drainage concentrations of 10 mg/L were not exceeded as long as fertilization was at or below the recommended rates. O n organic soils Miller reported very high annual drainage losses of nitrate (15 - 43 ppm; 37 - 245 kg/ha), even though application rates (about 80 kg N/ha per year) did not greatly exceed recommended rates (from 40 - 70 kg/ha per year). 20 In a two-part study, Baker et al. (1975) first established two plots with very similar nitrate leaching characteristics. Then, in the second phase of the study (Baker & Johnson, 1981), the plots received different amounts of N fertilizer, being fertilized 3 times at 2 year intervals (one at 100, 90, and 90 kg/ha and the other at 250, 240, and 90 kg/ha). The plot receiving less N lost an average of 27 kg/ha, while the other plot lost an average of 48 kg/ha. After the first fertilization, drainage concentrations of NO^ ' -N were about 2 times higher in the plot receiving more N; after the second fertilization they were about 4 times higher. On a Brookston clay in Ontario, Bolton et al. (1970) found that fertilization at 129 kg N/ha increased N loss through tile drains for continuous corn or a corn/oats/alfalfa rotation, compared with plots receiving no fertilizer. However, the six year average losses had a maximum of only 15 kg/ha/year even though fertilization exceeded recommended rates. Johnston et al. (1965) studied four tile-drained, irrigated fields under different cropping systems on silty clays in California. They found that, for each system, the quantity of N lost in drainage correlated with the amount of N applied, and that large percentages of applied N were sometimes lost (ranging from 1 - 70%). Cast et al. (1978) found that for continuous corn on a Webster clay loam, nitrate in tile effluent was not very different for applications of 20 or 112 kg/ha/year (the latter is the recommended rate), or for applications of 224 or 448 kg/ha in the first year of the study. However, the two higher fertilizer rates had increasing losses in the second and third years, to 59 and 120 kg/ha/year in year 3 for the 224 and 448 kg/ha rates, respectively. It must be emphasized that many factors interact to produce nitrate leaching and so leaching (or loss including denitrification) is not a simple function of fertilization rate. The work of Hanway and Laflen (1974), Baker et al. (1975) and Sheppard and Bates (1986) in North America and of Jaakola (1984) in Finland illustrate this point. Hanway and Laflen (1974) concluded that there was no relationship between losses or concentrations of a number of nutrients (including inorganic N) in surface runoff or tile 21 drainage and the quantity of fertilizer applied. They also found large variability between the 4 terraced sites they studied for 3 years. Sheppard and Bates (1986) conducted a 3 year study on 3 soils (a sandy loam, silt loam, and clay loam) in the semi-humid conditions of southern Ontario. They found that for six rates of N application ( N H 4 N 0 3 at 0, 56, 112, 168, 224 and 336 kg N/ha) to corn, soil NO3"-N contents in early fall were always significantly dependent on the rate of fertilization in the spring, being higher at fertilization rates both above and below those producing optimum yield. By spring sampling (late March to mid-April), there was generally little difference in the NO3" content of plots for any rate of fertilization for the sandy loam and silt loam soils. Increasing NO3" concentrations with depth were interpreted to mean that leaching was the predominant loss mechanism for N 0 3 ' - N . In contrast to the other two soil types, the clay loam soil had relatively constant N 0 3 " -N contents at fall and spring sampling, and in two of the three years, the spring soil N0 3 " content was dependent on the fertilization rate. N0 3 " retention in the clay loam was attributed to the relatively impermeable nature of the soil. This study illustrates the importance of soil texture and structure as interacting factors with fertilization rate. Baker et al. (1975) found the concentration and annual losses of nitrate from 4 adjacent tile drains to be very variable - annual losses over 4 years ranged from 0 to 93 kg N/ha. Additionally, N 0 3 " -N concentrations were unexpectedly high (almost always > 10 mg/L) in spite of low levels of applied N (112 kg N/ha applied in 2 of 5 years). The authors speculated that spatially-variable hydraulic conductivity was important in influencing nitrate leaching at this site. They concluded that with such a large amount of variability, nitrate loss cannot be attributed solely to the use of fertilizers, but that all factors including soil moisture, quantity and distribution of organic matter, temperature, tillage, and quantity and distribution of precipitation are also important determinants of leaching. Many of the studies of the effect of fertilization rate on leaching are actually quite difficult to interpret conclusively, because they have not made attempts to control or measure denitrification. 22 2. Timing Many studies have shown that the timing of fertilizer application is critical to controlling leaching losses of nitrate. In many fertilizers N is not present as nitrate and therefore nitrification must occur before nitrate can leach. Leaching of fertilizer nitrate is greatest when application is timed so that N is in the nitrate form either prior to precipitation or irrigation, or when the soil profile is saturated or nearly so, i.e. at the times of year when NO^" leaching would be greatest anyway. Again, the number of studies dealing with seasonal timing of fertilization and N loss is very great; relatively fewer actually measure NO3" leaching on a short-interval time basis. First, the generalizations. Tyler and Thomas (1981), working with near-saturated, strongly-structured soil columns, concluded that if rain or irrigation should occur shortly after fertilization of nearly-saturated soils, there will be considerable fertilizer loss, at any time of year. According to Kolenbrander (1981), in the Netherlands, N applied in fall or winter is more liable to leach than when applied in spring or summer. Cooke and Williams (1970) reviewed work in Great Britain on N losses dating back to the late 1800 s. They concluded that less N is lost from grassland when fertilization is periodic over the growing season, than when N is applied as a single application in autumn, winter, or sometimes early spring. As already discussed, these are normally the seasons with the highest leaching. Burns and Greenwood (1982), on the basis of reviews of field work and predictions from their model for N loss throughout England and Wales, suggest that nationwide there may be a 4 4 % loss of N from autumn residues from the top meter of all arable soils, but that losses of N applied after April 1st are likely to be insignificant. However, they warn that more nitrate will be leached in the years that rain exceeds evapotranspiration later into the spring. More specifically, some field studies find leaching greatest for fall N applications. For example, in Finland, jaakola (1984) found that application of N in autumn (September or 23 November) increased leaching of N through drains more than May applications of N did for a winter wheat crop on clay soil. In a lysimeter study with grass on sandy soil in Holland, with application in August through March, leaching losses peaked at 35 - 4 0 % of applied N for November and December applications, reducing to only about 1 0 % loss for August and February applications (Kolenbrander, 1981). Kissel et al. (1974) also found losses of NO3" and Cl" from the first rainfall following December applications to a structured, swelling clay soil in Texas, although the losses did not peak until late April or early May, at which time accumulated drainage was 60 mm. Spring fertilization can also result in significant leaching of NO3", as illustrated by the following studies. A lysimeter study conducted by Carstang (1980) in the U.K. showed that, of 8 weekly applications of fertilizer throughout March and April, the greatest losses of N under grass occurred from the early March applications. This study is somewhat unique because it addresses the effect of timing within a season, rather than just comparing applications between seasons. This is an important consideration, as the next two studies also show. Bottcher et al. (1981) attributed high NO3" losses in one year of their study to the occurrence of heavy rain shortly after spring fertilization. Where urea was applied, rains occurring about 2 weeks after application caused a nitrate pulse in drainwater: the two weeks of warm weather had allowed mineralization of the urea to occur. In a Kentucky study data from pan lysimeters installed under no till and conventional till corn indicated that large amounts of surface applied NO3" and Cl " could be lost from the structured silt loams, even in May and June, following precipitation. Tyler and Thomas (1977) therefore recommended delaying the application of N for 4 - 6 weeks after corn planting in May. Finally, in apparent contrast to these other studies, is work by Phillips et al. (1981). In a 6 year study of the effects of rates and timing of liquid manure application on corn yield, 24 manure was applied at rates of approximately 224, 560, and 897 kg N/ha/year in the spring (early May), fall (late September), winter (early December), or a spring-fall split application, to a sandy clay loam soil near Ottawa, Ontario. While the NO^ ' -N content of tile drain water increased with increasing N application rate, there appeared to be little effect (especially at the two higher application rates) of time of application on spring or fall drain effluent concentration. Part of the explanation for a lack of effect of timing may be in the nature of manure itself. Drain effluent from the 897 kg N/ha manure treatment was similar in concentration to that from a 134 kg N/ha chemical fertilizer (unspecified composition) treatment, which is not a particularly high application rate. Also, only drain concentrations were compared in the study, total losses were not specified. In south coastal British Columbia there have been few studies of nitrate leaching (Kowalenko, 1987). On a moderately-well to well-drained Brunisol in the Lower Fraser Valley of British Columbia, Kowalenko (1987) found that leaching of NH4NO3 applied in late May was quite limited over the spring and summer. Leaching became significant in October, so that November-applied fertilizer and residual NO3" from the spring application were leached extensively over the winter. Fairly large amounts of leaching appeared to continue into March or April in some years. E. Drainage In humid areas artificial drainage is often an essential component of a successful agricultural water management system. Subsurface drainage allows manipulation of soil moisture to provide improved growing conditions for crops and an extended working season. Profound alterations in the edaphic environment result from installation of artificial, subsurface drainage. Physical changes include lowered or controlled water table, improved aeration, improved structure, increased percolation (Lowrance et al., 1984) and increased hydraulic conductivity and soil moisture storage capacity (Walter et al., 1979; Hundal et al., 1976). The improved aeration may be responsible for improved permeability through promotion of 25 aggregation, crack formation and proliferation of worms which create burrows (Leyton & Yadev, 1960). While it is accepted that drainage can profoundly affect nutrient movement from agricultural land, it is.not always possible to predict the exact nature of the effect (Lowrance et al., 1984). Usually with drainage there is a decrease in surface runoff and an increase in subsurface flow of water, resulting in a decreased loss of sediment-associated, surface-transported nutrients and increased leaching of soluble nutrients such as nitrate (Walter ef al., 1979; Bottcher ef al., 1981). Typically, the predominant form of N in drain effluent is the NO3" ion, the percent composition of NO3" in drain total N for some typical studies being: 95% (Lowrance ef al., 1984; Burwell ef al., 1976); 7 5 % (Bottcher ef al., 1981); and 60 - 84% (Sharpley & Syers, 1981). There are a number of proposed mechanisms for increased nitrate leaching under subsurface drainage, most of which are interrelated to some degree and are difficult to examine in isolation. Of prime importance is the way in which water movement is affected by drainage. First, as already mentioned, drainage seems to increase both the rate and quantity of water flow through soil. The greater the volume of water moving through the root zone, the larger the potential loss of nitrate. In some agricultural watersheds, tile flow is a significant source of nitrate in streamflow (Lowrance et al., 1984). For example, in 4 agricultural watersheds in Iowa, tile discharge accounted for 62 - 88% of the average annual streamflow and 84 - 95% of the average annual soluble N (ie. NO3") stream load (Burwell ef al., 1976). Tile flow is usually discharged directly to drainage ditches or streams, rather than filtering through a mass of soil. Thus subsurface drainage effectively "short-circuits" the normally tortuous path of nitrate from the soil in a field to a stream. The more direct delivery may decrease the chance of nitrate uptake, immobilization or other reactions that would normally decrease the amount of nitrate reaching streams. This is clearly illustrated by the work of Thomas and Barfield (1974), who found that tile drainage entering a stream had a 26 much higher NO3" concentration than the subsurface seepage entering the same stream from below the tile outlets (an average > 15 ppm in tile versus slightly > 3 ppm in seepage). The same study of Thomas and Barfield (1974) also points out the difficult nature of prediction of the impact of drainage on nutrient flux. In spite of the high drain NO3" concentration, tile drainage contributed a relatively small proportion of the total NO3" load to the stream, because the volume of tile flow was small relative to the volume of deeper seepage entering the stream. This study also brings out another point - that subsurface drainage does not intercept all of the water percolating through the soil. Therefore the impact of tile drainage on groundwater or surface water quality depends on the concentration of nitrate in the tile effluent relative to that in seepage water and on the proportion of percolate that is intercepted by drains. It has already been mentioned that tile drainage often has a higher nitrate concentration than soil water at comparable depths in more poorly drained areas. The difference in drain water and soil water NO3" concentrations is usually attributed to an altered balance between nitrification and denitrification. As the intensity of drainage increases and aeration improves, the amount of denitrification should decrease. For example, Colburn and Dowdell (1984) cite unpublished data which indicates that for an arable clay soil with denitrification actually measured, mole-drained plots lost an average of 16 - 50 kg N/ha from denitrification compared to 24 - 110 kg N/ha lost from the undrained plots. Thus, soil farther from drain lines, or undrained sections of a field, should have a greater potential for denitrification than the well-drained soil near a drain line. Similarly, denitrification should be higher in the saturated soil beneath a drain line, resulting in lower NO3" concentrations there (Gilliam et al., 1974; Gambrell ef al., 1975; Baker ef al., 1975). The study of Gambrell ef al. (1975) in the North Carolina Coastal Plain illustrates many aspects of the nitrification/denitrification balance. Comparing a tiled, moderately well-drained soil with a poorly drained soil, they found that the drained soil to 3 m depth had well-oxidized 27 conditions (Eh = 500 - 700 mV), over 200 kg N0 3"-N/ha, and only 2 - 5 mg/L water-soluble C in the soil water. In contrast, the poorly-drained soil below 1 m was saturated, had reducing conditions, very low N0 3 ", and 1 0 - 1 5 mg/L water-soluble C in the soil water. That denitrification was occurring in the poorly-drained soil was supported by decreasing NO^-N/Cl ratios with depth. It cannot, however, be universally assumed that saturation implies denitrification and reduced leaching. Verdegem and Baert (1984) found that there was no significant change in Cl/NO^-N ratios below the permanently saturated zone in a sandy soil, explained, perhaps, by very low levels of C (<1%) and redox measurements indicative of oxidized conditions. Improved aeration also promotes general soil microbial activity, including mineralization of organic C and N (van Burg ef al., 1980). This is nicely illustrated by the study of Benoit (1973). In that study, 6 years after artificial drainage was installed in a silt loam soil in Vermont, there was a highly significant decrease in total soil N in the drained plots compared with the control plots, and the decrease was greater with more intense drainage. In spite of similar management practices before and after drainage, average annual losses over 6 years were 1000 kg N/ha. The losses were attributed to increased rates of organic matter breakdown from improved soil aeration on the naturally poorly to somewhat poorly drained soils. The increased availability of C and NC>3" may actually promote denitrification (and thus perhaps decrease leaching) at the wet, anaerobic sites which are likely to remain in a drained but inherently wet soil (Colburn & Dowdell, 1984). Additionally, the increased mineralization combined with increased percolate volume in a drained soil may result in increased leaching of C, NO3" and microorganisms to the saturated zone near or below the drains. There, it is more likely that anaerobic conditions will be encountered, and denitrification may be enhanced (Verdegem & Baert, 1984). The uniqueness of each field situation makes generalizations and predictions of nitrate status very difficult. Therefore, it is not surprising to find at least one study where drain NO3" 28 concentrations were measured to be lower than NO3" concentrations in shallow wells located between the drain lines (Richard & Chieng, 1985). The authors of that study speculated that the reason for the unusual relative concentrations was the enhancement of denitrification near the drains by possible increased C availability from the envelope of wood shavings around the drain lines. The influence of moisture content (through aeration) on the nitrification/denitrification balance, and the relative ease with which soil moisture may be manipulated by subsurface drainage, has led some researchers to speculate that NO3" leaching to groundwater or loss through drain lines to surface waters could be controlled by drainage design. The intensity of drainage (depth and spacing of drain lines) in a particular soil determines the water table height and therefore the location of the anaerobic zone. Water table height may also be manipulated by providing for submergence of drain lines through outlet water level control. There are very few studies of nitrate leaching where water table height or drainage design are actually the variables under investigation. This is unfortunate, given the necessity of drainage in humid areas, the coincidence of concern for NO3" pollution in many of the same areas, and the fact that there are often alternative drainage designs that will meet the same objectives (Skaggs, 1978). A few of the studies with water table as a primary variable will be discussed below. In a study by Meek et al. (1970), 6 large columns (38 cm diameter, 305 cm length) were placed in an underground field laboratory. The soil in the columns was repacked by horizon, with the top 60 cm clay, followed by 30 cm of silty clay and then silt loam. Simulated tile lines were installed at 180, 240, and 300 cm depths, with the water table controlled to 175 cm, creating submerged zones of 5, 65, and 125 cm. Solution samples from various depths in the columns indicated that a diminishing NO3" peak moved downward in the soil with each irrigation to approximately 160 cm. Below the water table there was a rapid decrease in NO3" -N concentration: water that had passed through the 65 and 125 cm submerged zones 29 contained a maximum of 1.1 ppm NO^ ' -N, compared to peaks of up to 40 ppm higher in the profile. In an 8 year field lysimeter study with grassland soil, van Dijk (1980) showed clearly that, at the same level of fertilization, leaching losses of N increased the lower the water table was maintained. At 242 kg N/ha and 300 kg CI7ha per year, N leaching losses were 5, 7, 22, and 30 kg/ha per year for water tables 85, 60, 35, and 10 cm above drain height, respectively, with corresponding N/CI ratios in the drainage water of 0.024, 0.031, 0.095, and 0.150 (initial N/Cl = 0.81). Controlled drainage may be useful in reducing NO^" delivery to surface waters in yet another way. In a naturally poorly-drained soil, Gilliam et al. (1979) measured a 50% decrease in NO^" to surface waters when less intensive drainage forced more water to seep through deeper horizons where denitrification occurred. Drainage control was achieved by water control structures in the outlet ditches. The decreased NO3" delivery to surface waters was directly due to decreased tile flow, not to a decreased NO3" concentration in tile drainage water. The water that was forced to leave the fields by deep seepage rather than through the tile lines was forced to pass through a deep, highly reduced zone where denitrification was believed to occur (no NO3" was measured in this deep soil zone). In the same study the moderately well-drained sites showed similar reduction in NO3" loss through tile lines by reduction in tile flow volume. However, the authors concluded that overall reduction of NO3" delivery to surface waters could not be certain, because the large volume of NC^ ' -bearing water leaving the field by deep seepage passed through oxidized soil layers in which there was no evidence of denitrification. In spite of the paucity of field or even laboratory studies relating N leaching to water table height and drainage design, some of the N management computer models allow for simulation of the effects of various drainage designs. Of course, if model predictions are to be applicable to a particular soil, the assumptions underlying the model must also apply to the particular situation. For example, in Ritjema's (1980), model of N emission from grassland 30 farms, developed in the Netherlands, for any particular texture, for any particular precipitation surplus, N loads to surface waters are highest for poor drainage and lowest for good drainage. The author's justification for this is as follows: in poorly drained soils, nitrate and precipitation will be transported by shallow interflow to surface waters, with a very short residence time in the soil. Better drainage status will result in deeper percolation, longer residence time, and increased opportunity for denitrification. However, the study of van Dijk (1980) was conducted under conditions similar to those simulated by Ritjema's model and clearly contradicts the model predictions. Skaggs and Gilliam (1981) presented an extension of DRAINMOD which simulated NO3" transport in a poorly drained soil in North Carolina. DRAINMOD is a verified water management model designed for soils with shallow water tables. Using constant values for N03 _ -N concentration in drainage water, surface runoff and seepage water of 11, 1, and 0 mg/L, respectively, they modeled various combinations of surface and subsurface drainage, including the use of controlled drainage (raised outlet water levels to raise the soil water table during non-growing seasons). They determined that drainage objectives for trafficability and crop protection could be met by drainage designs which produced considerably different NO3" outflows. Surface drainage water characteristically carries lower concentrations of nitrate than subsurface drainage water. Skaggs and Gilliam found that good surface drainage in combination with subsurface drainage produced about 50% less nitrate outflow than when surface drainage was poor. Controlled drainage with raised water tables also reduced nitrate losses. Simulations with this model were extended to six soil series by Deal et al. (1986). Based on further field studies, the estimates of NO3" content of the various flow components were refined for each soil series for each combination of surface and controlled or uncontrolled subsurface drainage. Once again, results indicated that total NO3" loss can be reduced by using controlled water tables with drainage (and thereby forcing increased deep seepage through reduced soil layers). 31 While results of these simulations are interesting, the greatest weakness of the model is that it is basically just proportioning flow through different paths - it does not attempt to simulate N transformations. The results also obviously depend on denitrification actually occurring in deep soil layers. As has already been shown, this can not be a universal assumption. F. Study Methods 1. Size and Nature of Experimental Unit Soil nutrient studies may be undertaken on many size scales, ranging from small pots of soil to field-size plots. Each study technique has attendant advantages and disadvantages, and limitations which must be recognized if experimental results are to be correctly interpreted. One of the major considerations in choosing the size and type of study system is the variability that can be expected among experimental units. Variability in nutrient content of soils is well-documented. Lund (1982) found significant spatial variability in saturation extracts of N 0 3 " - N and C l " below 4 agricultural fields in California, whether management had been constant for a considerable time or had been recently changed. Solution NO^ ' -N concentrations ranged from means of 40 - 136 mg/l, and C l " ranged from 0.16 - 0.25 mg/l (standard deviations were not reported, but appeared to be quite high in some cases). The author could not relate the variations to any soil or field characteristic. White et al. (1987) found substantial variability in soil NO3" and N H 4 + concentrations under grazed and ungrazed grassland: C V s for soil NO3" ranged from 50 - 200%. They attributed the high variability to physically small samples, and found that most of the variance occurred within relatively short distances. Other authors finding considerable variability in N 0 3 " -N content of soils include Omoti and Wild (1979), White (1985b), Kissel et al. (1974), and Dowdell and Webster (1980). All authors recommended using large volumes of undisturbed soil to study leaching. 32 However, a larger size of experimental unit does not necessarily reduce variability in leaching. Dowdell and Webster (1980) used 45 cm diameter undisturbed-core lysimeters and found CV's of 50 to 105% between lysimeters within a treatment (although results were complicated by slightly different N uptake by the grass in each lysimeter). Two important studies on variability in leaching involved tile-drained fields where each drain line was monitored separately. Guitjens et al. (1984) and Baker et al. (1975) both found substantial variability from drain to drain in both discharge and drain water composition. The study by Baker et al. (1975) was conducted on a silt loam soil in Iowa. Only 4 drain lines were studied, over a period of 5 years. Guitjens et al. (1984) studied 15 drain lines over 3 years, monitoring 18 water quality parameters. Unfortunately they report results only on an annual basis, but there is considerable variability among drain lines for many parameters. Water management varied among some of the lines in some of the years, but variability in NO3" losses did not appear to be due to water management. The pattern of nitrate losses among lines was similar for the first 2 years, but changed in the last year of the study. Smaller sized study units, such as lysimeters or laboratory cores, are often used in leaching studies, perhaps because of their lower relative cost, especially if replication is desired, and the greater ease with which external variables may be controlled. Laboratory cores or lysimeters may be made of either disturbed or undisturbed soil. Disturbed cores made from homogenous mixtures of soil can offer the advantage of uniformity in chemical and physical characteristics in all replicates, so that it may be easier to determine differential effects of imposed treatments. However, the natural soil structure is completely destroyed in disturbed cores. With recent recognition of the importance of physical structure to leaching processes in soils, disturbed cores have serious disadvantages if experimental results are to be interpreted in terms of processes actually occurring in the field. Therefore the objectives of a particular investigation must determine whether disturbed or undisturbed cores are appropriate. 33 Laboratory-scale undisturbed cores may be obtained, either by directly pushing or rotating some kind of sleeve over a column of soil in a drilling-type procedure, or by carving out a soil monolith and then encasing it in some kind of material. The encased-monolith method (see, for example, Murphy et al., 1981; Buchter et al., 1984) has many advantages, including suitability to a wide range of soil types and virtually no restrictions on size of sample. Proponents of this method apparently feel that boundary flow is reduced in encased monoliths compared with cores in rigid pipes, because the container wall will conform to the irregular shape of the soil surface. Obtaining undisturbed cores by drilling or pressing tubing into the ground has several potentially serious drawbacks. Compression or compaction of the core has been reported by many authors, sometimes to a considerable degree. For example, Turner (1974) reported "typical" compression of "less than" 10%. When the core casing is pounded into the ground, compaction of the surface layers may result in decreased infiltration (Tuttle et al., 1984). Compaction may be minimized throughout the length of the core by the use of a double sleeve or hollow tube system with rotary capability (Bausch et al., 1977; Runge, 1965). In the double sleeve system an outer sleeve (effectively the drill stem) is rotated into the ground. The bit attached to the outer sleeve has a hollow center and cuts a soil core of much smaller diameter than the outer sleeve, and slightly smaller diameter than the inner sleeve. As the outer sleeve rotates into the ground, the core is pushed into the inner sleeve, which does not rotate. The inner sleeve may be either a rigid plastic or steel tube or heat-shrinkable tubing. One of the major problems with undisturbed cores taken in rigid tubes is that the core does not necessarily fit perfectly in the tube, commonly resulting in "gaps" and channels at the soil-tube boundary which may seriously affect hydraulic conductivity measurements and nutrient movement studies (McNeal & Reeve, 1964; Weber & Peeper, 1977; Mielke, 1973). Using heat-shrink tubing to enclose an undisturbed core may prevent or at least minimize boundary flow problems (Bondurant ef al., 1969; Mielke, 1973), because the shrinking tubing 34 will conform to the irregular shape of the soil core as it cools. However, heat-shrink tubing of large diameter is very expensive. For cores that are pushed directly into rigid containers, Weber and Peeper (1977) recommend using cores of at least 23 cm diameter for herbicide leaching studies to reduce the impact of boundary-flow on experimental results. They further recommend using the divided-bottom technique of McNeal and Reeve (1964) for cores less than 23 cm diameter. In the divided bottom technique a smaller diameter ring, sealed to a bottom and drain tube, is pushed into the center of the bottom surface of the soil core. Only leachate collected from the central region is used in experimental analysis; leachate draining down the tube sides is excluded. It is difficult to compare leaching studies conducted on different types of experimental units. Only rarely does a single study attempt to address the nature of differences in leaching responses from different study units: a recent study by Bergstrom (1987) is quite a thorough attempt to make such a comparison. In this Swedish study 4 management alternatives (barley, with and without fertilization, unfertilized lucerne ley and fertilized grass ley) were compared on 4 different types of study units: tile drained plots (0.36 ha each); large lysimeters (27 m 2 ) composed of soil reconstituted layer by layer; medium lysimeters (1.2 m diameter) composed of disturbed soil, and small undisturbed core lysimeters (0.295 m internal diameter). Bergstrom found that crop yield varied up to 50% in the different units, which may have complicated leaching results. Drainage volumes varied among units, as expected (tile drains do not intercept all of the water flowing through the profile while a lysimeter does). Drainage volumes were also different between disturbed and undisturbed soil. There were variations in yearly fluxes of nitrate leached, which were related to the varying drainage volumes. Thus, there were differences in leaching loss of nitrate between lysimeters and tile drained plots, and also between disturbed and undisturbed soils. Bergstrom's conclusion was that although there were differences among study methods, each method itself produced consistent results. 35 2. Conservative Tracers Because N 0 3 " is subject to uptake, immobilization and denitrification in soil, as well as leaching, the amount of N removed from the soil solution by such processes must be known before the effect of any treatment on leaching can be evaluated. A complete mass balance would provide the necessary information, but is not always possible. For example, measurement of gaseous N losses may complicate an experimental set-up. In some circumstances, soil sampling is undesirable, such as in a leaching column study where soil sampling would effectively destroy the column. To circumvent these two problems, a conservative tracer of NC>3" is often employed. Ions like C l " and Br" move in a very similar manner to the N 0 3 " ion, but are considered to be chemically and microbiologically inert with respect to the soil (Bowman, 1984; Verdegem et al., 1981). C l " is more widely employed than Br". The recovery of C l " in the leachate from a column or field plot will indicate what proportion of applied solute remains in the soil mass. If no NO-j" was immobilized or retained on passage through a mass of soil, the proportion of N 0 3 " - N to C l " in the leachate (N/CI) would not change from the proportion applied (N Q/CI 0), and (N/CI) = (N 0 /CI 0 ). If (N/CI) < (N 0/CI Q), N 0 3 " - N is being removed from the solution as it passes through the soil, either by immobilization, denitrification or plant uptake. G. Summary To summarize very briefly, from the literature it is known that substantial quantities of NC» 3"-N may be lost by leaching from certain soils, especially in humid areas during the rainy season. Losses are affected by the form of N fertilizer applied and the timing of fertilizer application with respect to the occurrence of precipitation or irrigation. Leaching losses of solute are also strongly affected by soil structure: leaching patterns may be different when preferential flow occurs than when displacement is the dominant mechanism of water movement. Thus, study techniques which preserve the natural soil structure are more likely to produce results indicative of what is actually occurring in the field. Subsurface drainage affects 36 the water table height and soil moisture status, and is often associated with increased leaching losses of nitrate, due to reduced denitrification or increased mineralization and nitrification. While there has been considerable work on leaching losses over the winter (from fall to spring planting) there has been much less investigation into leaching of nitrate during the spring itself. In the spring, after nitrate starts to appear in the soil due to increased mineralization and nitrification, and perhaps fertilizer addition, precipitation may cause significant leaching of NO^ ' -N . Where artificial drainage is installed, water table fluctuations may occur in response to spring rainfall. The fate of nitrate under these conditions is less well understood, and is the subject of this research. 37 111. M A T E R I A L S A N D M E T H O D S A. Overview of the Experiment Twenty four 1 m deep by 15 cm diameter undisturbed cores of soil were isolated within PVC pipe and removed from an uncultivated grassland site. Soil samples were collected at the time of core recovery to determine soil nitrogen status. The cores were fertilized with a concentrated nutrient solution (KNO3, KCI, I ^HPO^, then leached with successive additions of water to simulate fertilization followed by precipitation. The cores were held at approximately 8 C for the duration of the experiment, to simulate early spring conditions. Leachate was collected from 2 depths (45 cm and the core bottom) and analyzed for N 0 3 " -N, CP, N H 4 + -N, and ortho P0 4-P. A preliminary saturation and drainage of the columns, (subsequently called "Run 0"), was conducted to determine soil N and Cl" status and approximate drainable porosities. Four leaching trials were then run, each lasting approximately 1 0 - 1 4 days. Considerable difficulty was experienced in Runs 1 and 2 with the column design and experimental procedure. These runs are discussed only briefly in the following sections. There was a delay of approximately 1 month between Runs 2 and 3 due to unavailability of a cold room. A successful column design was employed for Runs 3 and 4, which form the foundation of this thesis. B. Site and Soil Description The columns were taken from the Boundary Bay Water Control Project site in Delta, B.C. The columns were removed from an area which had not been cultivated for several decades. The area is poorly drained and overgrown with natural grasses and some wildflowers. The soil at the Boundary Bay site is classified as a Humic Luvic Cleysol (Luttmerding, 1981). It has developed on moderately fine to fine textured deltaic deposits of both marine and freshwater origin, overlaying sandy deposits at depths of over 100 cm. Surface and subsurface textures vary from silty clay loam to silt loam (Luttmerding, 1981). The drainage at 38 the Boundary Bay site is moderately poor to poor. Where there is no sub-surface drainage, water tables typically are at or near the surface for much of the winter. The A horizon in the areas sampled generally appeared very dark in colour and rich in organic matter; in some places however, the A horizon was less organic and appeared to be a firm, very dark grey. The A-B horizon boundary generally occurred at approximately 30 cm, but was irregular, ranging from 20 to approximately 40 cm. The top of the B horizon was firm and appeared partially leached. Reddish-brown mottles were common throughout the profile, as were concretions, primarily around old root channels. The concretions were yellow to reddish-brown in colour and an average of several millimeters in diameter. Fine roots were also abundant. Roots and concretions were observed to the maximum depth excavated, approximately 1.2 m. Occasional peat-like deposits of variable size occurred throughout the B horizon. The B horizon of the site excavated for this experiment was weakly prismatic, grading to a more massive structure deeper in the profile. In some places the prismatic structure was stronger. The soil is acid (Driehuyzen, 1983; Luttmerding, 1981) and, in the samples excavated, had a pronounced odour of sulphur below about 70 cm. Table I lists some of the physical and chemical characteristics of soils sampled at the Boundary Bay Water Control project site (Driehuyzen, 1983). The 10-year mean March and April air temperatures at the nearby climate station Delta Ladner South are 6.0 C and 8.8 C, respectively (Envt. Canada, 1982). Soil temperatures at the Boundary Bay site are only available for April to October. For April, the mean temperatures at 5, 10, and 20 cm depth were 11, 10 and 9 C, respectively, for an undrained site, and 9.8, 10, and 9.9 C, respectively, for a drained site (BCMAF, unpublished). March soil temperatures may be roughly estimated by examining data from other climate stations. The 30-year mean soil temperatures at the UBC climate station for 5, 10 and 20 cm depth are 7.5, 6. and 6.9 C, respectively, for March, and 11.4, 10.6 and 10.1 C, respectively, for April (Envt. Canada, 1982). Table I. Soil physical and chemical characteristics a. Physical properties Depth Particle Size (%) Texture Bulk Density (cm) Sand Silt Clay CSSC* (g/cm3) o - 30 3.5 73.3 23.2 s i l — 30 - 50 11.2 65.4 23.4 s i l 1.40 50 - 70 21.7 58.8 19.5 s i l 1.40 90 - 110 33.1 50.0 16.9 1 1.37 * Canadian System of Soil Classification b. Chemical properties Depth PH PH C Total C/N N (cm) (H20) (CaCl2) (%) (%) 0 - 3 0 4.4 4.3 5.2 0.421 12.5 30 - 50 5.6 5.2 1.8 0.135 13.6 50 - 70 5.8 5.2 0.4 0.035 11.4 90 - 110 4.5 3.9 0.4 0.033 12.2 Depth Exchangeable Cations (meq/lOOg) CEC Base (meq/ (cm) Ca Mg K Na lOOg) (%) 0 - 30 7.9 6.1 0.5 0.2 28.9 50.5 30 - 50 6.6 10.7 0.3 0.4 22.7 50.5 50 - 70 5.1 9.1 0.3 0.4 17.1 87 .1 90 - 110 2.7 3.8 0.2 0.4 14.8 48.6 (Data from Driehuyzen, 1983) 40 C. Acquisition and Preparation of the Columns A truck-mounted Ciddings Model CRSP hydraulic soil probe was used to obtain the columns. As the standard tooling supplied by Ciddings was not adequate, new tooling was designed. A press plate of 1.25 cm (1/2") steel plate was made which fitted to the drill head providing a pressing surface perpendicular to the drill mast. Thus the drill was used as a hydraulic press to push sections of PVC pipe into the ground. One meter sections of 15 cm (6") internal diameter Class 160 (0.64 cm wall) PVC pipe were used. The soil cutting edge was sharpened to an approximately 20 degree angle to provide good soil penetration without weakening the cutting edge itself. Initial trials in July, 1985 indicated that it was very difficult or impossible to obtain columns in moderately dry soil. Therefore columns were not placed until the soil moisture increased; all columns were placed within a 1 week period in October, 1985. Care was taken during placement to avoid vertical compression of the soil core. With high soil moisture, pushing an unvented pipe into the ground developed considerable pressure in the free space above the soil. If the pressure was not relieved, several vertical centimeters of core compression could occur. To avoid compression, the drill was "backed off" and the pipe vented every few centimeters of depth, resulting in a vertical compression of approximately 1 cm (2 cm maximum) per column (i.e. less than 2% of the column length). Untrafficable conditions required that the columns remained in situ after placement until February, 1986. Over winter the columns were covered with dead grass to avoid possible deleterious effects of raindrop impact. A backhoe was used to remove the columns from the ground. The columns were tipped slightly into a trench dug by the backhoe, then lifted by chain to transportation racks, thus remaining vertical during the entire process. 41 O. Field Samples After the columns were removed soil samples were taken from the trench wall at 3 depths: the A horizon (0 - 30 cm), the top of the B horizon (30 - 50 cm), and the bottom of the B horizon (50 - 100 cm). Where possible, samples were taken adjacent to each column; a few centimeters of soil wall was removed prior to sampling to avoid sampling soil which may have been subject to surface water draining down the outside of the pipe walls. Samples were placed in plastic bags and stored immediately in ice chests. E. Storage Conditions During the Experiment During the experiment, the columns were kept in a temperature-controlled room at approximately 7.8 C. Air temperatures were obtained as hourly readings from copper-constantan thermocouples at various locations in the room. Soil temperatures were recorded hourly by thermocouples at 3 locations within one column: 5 cm depth from the soil surface, 7.6 cm in from the column wall; 30 cm deep, 2.5 cm in; and 30 cm deep, 7.6 cm in. Air temperatures were consistent throughout the room (maximum difference between locations = 1.3 C) and air and soil temperatures matched closely after an initial 1 week equilibration (maximum difference = 4.5 C, but usually less than 1 C difference). During the 2 weeks of hourly recording, a Hg thermometer in a water bath provided a reasonable estimate of soil core temperature. Therefore, for the duration of the experiment a Hg thermometer in water was used as the temperature reference, and thermocouple readings were taken only infrequently. The average soil temperature was approximately 7.8 C, with temperatures ranging between 6.4 and 8.9 C approximately 86% of the time. Soil temperatures were also consistent between locations within the column, with less than 1 C as the maximum difference at any particular time. The temperature was chosen to simulate temperatures in the field for the period late March to mid-April, based on the climate data presented earlier. Throughout the experiment, the columns were kept covered with black polyethylene to reduce evaporative loss of water, and the cold room was kept dark to prevent growth of 42 vegetation. Vegetation was removed from the surface of the columns prior to starting the experiment. F. Column Set-up This section will describe the physical set-up of individual columns, as used in Runs 3 and 4, and as illustrated in Figure 1. The set-up used in earlier runs, where different from this arrangement, will be noted in the description of the particular run. Using a portable steel collar and lever arm, each column was seated into a Schedule 40 slip-fit PVC cap. The lever arm helped control the heavy columns (approximately 50 kg each), to avoid excessive pounding while seating the caps. "Plas-tyton" (an inert wax-based product for potable water pipes) was used as a lubricant and sealant inside the cap. A sand filter of HCI-washed and distilled water-rinsed sand filled the space in the cap below the column bottom. Nylon "no-see-um" netting screens prevented loss of sand through the drain outlets. The sand was used to support the bottom of the soil column to prevent its collapse, and to reduce the volume of water which would be held in the cap while the experiments were running. Many of the columns had deformed from a circular shape during placement in the ground, and so could not be seated completely within the caps unless they were pounded. Pounding the columns was avoided because it may have altered the natural structure of the columns. Extra sand was used to fill in the space when a column could not be completely seated. This resulted in considerably different volumes of water being held in the sand filters in caps of different columns. The clear PVC drain tubes included short splices of soft latex tubing which could be clamped shut with pinch clamps. Water table control was achieved in a manner similar to that used by Meek ef al. (1970), by fixing the water table control tube into an "overflow bottle" at the pre-selected water table height. Thus when the water table within the soil core rose above the level of the opening of the water table control tube, leachate flowed into the collection bottle. The "drain tube" was fixed vertically along the column and left open to the air, to monitor the height of the water table within the column. In this way, the height of the 43 S a m p l e c o l l e c t i o n b o t t l e W a t e r l e v e l c o n t r o l t u b e ( 1 . 2 5 cm OD PVC t u b i n g ) 15 cm ID s c h e d u l e 40 PVC p i p e H a l f - w a y s a m p l e r W a t e r t a b l e h e i g h t S c h e d u l e 40 PVC s l i p - f i t c a p S a n d f i l t e r P o l y e s t e r s c r e e n 1.25 cm m a l e t h r e a d e d t o 1.25 cm i n s e r t a d a p t e r ( 2 ) D r a i n t u b e ( 1 . 2 5 cm OD PVC t u b i n g ) F i g u r e 1. Column c o n f i g u r a t i o n 44 water table control tube and overflow bottles could be adjusted as necessary to achieve the desired water table within the column. Throughout the experiment, there was generally very good agreement between water levels in the 2 tubes: a discrepancy of more than 1 or 2 cm usually indicated a bubble in the lines or air in the sand filter. All fittings and tubing used in the experiment were PVC, all were acid-washed and distilled water-rinsed prior to use. Reddish yellow deposits and slime built up in the tubing of some columns during the experiment; when this occurred, tubing and fittings were removed and cleaned. Soil water samplers were installed at approximately the half-way point in the columns: 45 cm from the column bottom, at a slight upward angle. A simple gravity sampler, shown in Figure 2, was designed and fabricated for this experiment. The choice of soil water sampling system may be critical to the quality of results in a study of soil water, and therefore merits some discussion here. There are two main types of soil water sampler: gravity and suction. Gravity (zero tension) samplers are usually some kind of impermeable tray inserted into a horizontal tunnel to collect soil water draining through the roof of the tunnel. Suction (tension) samplers involve some kind of porous cup or plate in intimate contact with the soil, to which varying levels of vacuum or suction are applied. Water held by the soil matrix at tensions less than the applied vacuum is drawn into the sampler. The porous samplers most readily available and most commonly used are made of ceramic (heat-fused clay). In spite of their wide use, suction samplers and particularly ceramic samplers, were deemed to be unsuitable for use in this study. The purpose of the half-way samplers in this study was to provide access to water leaching through the profile at a point above the drain level so that chemical changes, if any, could be observed in the leachate as it moved down the column. Ceramic suction samplers have failed to accurately sample several chemical constituents in solution around them. Only chemical constituents of relevance to this study will be mentioned here. An extensive study by Zimmermann et al. (1978) showed that ceramic P o l y e s t e r s c r e e n f i 1 t e r £ 0.1 cm p e r f o r a t i o n s (4) 2„2 cm OD PVC p i p e (0.1 cm wal 1) if Column i n t e r i o r ( s o i l ) Column w a l l S o l v e n t w e l d s e a l • S c h e d u l e 40 PVC a d a p t e r 2.2 cm f e m a l e s l i p - f i t t o 1.25 cm f e m a l e t h r e a d e d PVC 1.25 cm m a l e t h r e a d e d t o 1.25 cm •v>x i n s e r t a d a p t e r 1.25 cm ID l a t e x t u b i n g F i g u r e 2. H a l f - w a y s a m p l e r d e s i g n 46 samplers exhibited poor and variable recovery of ammonium, phosphate, nitrite and nitrate. In their study they tested samplers in pure solutions; recoveries ranged from 11.1% of a standard solution (for ammonium) to 96.5% of a standard solution (for nitrate). Comparing the composition of samples from gravity and ceramic suction samplers in soil under a deciduous forest, Haines et al. (1982) found significant differences (at the 0.05 level) for NO3" , NH^" 1 " and Cl". At 30 cm depth, gravity samplers yielded 3.4 times the NO3", 5.1 times the NH^"*", and 1.6 times the Cl* concentration of suction samplers. The concentration of phosphate in the filtrate from a ceramic sampler is affected by leaching from, sorption to, and screening by the sampler and by diffusion rate through the sampler wall (Hansen & Harris, 1975). Ceramic samplers are subject to plugging after prolonged use. Plugging may exacerbate the difference in composition of filtrate and soil solution, and is a factor in variable nitrate recovery (Hansen & Harris, 1975). There are also purely physical factors which may contribute to sampling error. Application of a vacuum to soil alters the nature of water flow patterns around the sample point (Cochran et al., 1970; Tadros & McCarity, 1976). In experimental units as small as those used in this study, the flow alteration caused by a suction sampler may affect a significant part of the soil volume. The nature of the vacuum (strength and duration of application) appear to alter the chemistry of the filtrate. This is likely due to the different pore sizes sampled at different levels of vacuum: the chemical composition of water held in fine pores may be different from that held in large pores and cracks. Several studies have established that suction samplers fail to sample the rapidly draining water in large pores, which seems to bypass the sampler. Instead, water in fine pores is sampled (Shaffer et al., 1979; Quin & Forsythe, 1976; Severson & Crigal, 1976). For example, Shuford et al. (1977), working with a silt loam soil, found that nitrate peaks of several hundred ppm that were intercepted by gravity samplers were "missed" by suction samplers. If the objective of the study is to sample rapidly leaching water which may be lost from the profile, then suction samplers are inappropriate. 47 In summary, ceramic suction samplers were not chosen because of their reported poor and highly variable recovery of chemical constituents of interest to this study, and because results from ceramic samplers may not be directly comparable to gravity sampling, which is the method of sampling at the bottom of the columns. G. Physical Measurements 1. Edge Flow To address the concern that large amounts of water may be leaching down the inside of the PVC pipe and failing to move through the soil mass, a device was designed and built to partition flow from the column bottom. The cross section of the column bottom can be divided into two equal areas: an inner circle of 10.78 cm diameter, and an outer ring of 2.23 cm thickness. The flow-separating device illustrated in Figure 3 consisted of a galvanized ring of 10.8 cm diameter sealed to a perforated PVC plate. The galvanized ring was inserted into the column bottom, so that the soil mass was separated into two equal areas. The number and size of perforations in both zones of the plate were equal. Cemented to the side of the plate opposite the galvanized ring was a small, 10.2 cm I.D. PVC pipe. The pipe was sealed and had a sloping bottom with a drain hole located in the lowest part of the bottom. A PVC fitting attached a drain tube to the drain hole, thus making a completely sealed chamber to collect leachate draining from the inner 50% area of the soil bottom. An outer chamber made of larger pipe collected leachate from the outer 50% of the soil area. To test for equality of flow through the two zones, a column was saturated by siphoning up slowly through the bottom over 4 days time, then allowed to drain. Leachate for volume comparisons was collected from each chamber over a period of 12 hours and measured with a 250 mL graduated cylinder. During the first hour of drainage the flow rate from the inner zone was about twice as great as from the outer zone. For the period of time from 1 to 12 hours after the start of drainage, the flow rates from the two zones were almost identical. Since the nutrient leaching experiments in this study were all of several days duration, the flow separating bottom was deemed unnecessary and was not used. 48 11 &//// iW4 ISl -15 cm ID schedule 40 PVC pipe -soil -11 cm ID galvanized tube ,-0.3 cm PVC plate with 0.16 cm perforations ± -11cm ID schedule 40 PVC pipe 15 cm ID PVC pipe 1.25 cm ID PVC tubing 1.25 cm threaded. 1.25 cm insert fitting (2 ) F i g u r e 3 . F l o w - s e p a r a t i n g bo t tom 49 2. Hydraulic Conductivity After Run 4 was completed, hydraulic conductivity was determined for all columns. A constant head procedure was used to measure satiated hydraulic conductivity (K s a t ) . However, the large size of the columns presented considerable problems for conductivity measurements by standard methods (Klute, 1965). Columns were "wet up" by siphoning through the column bottom, very slowly over a period of several days. It was hoped that a slow rise in water table from the bottom of the column would minimize the inevitable air entrapment within the soil. Water tables were allowed to rise 5 cm above the soil surface of each column. The test consisted of opening the drainage tube of a column and collecting leachate for specific time intervals while maintaining a 5 cm head. Initially after opening the drainage tubes, the conductivity was quite high, but dropped as leaching continued; measurements used in the calculation of K$a^. were recorded after the outflow rates had reached relative equilibrium. 3. Dye Tests After the hydraulic conductivity tests were conducted a series of dye tests were made to qualitatively evaluate leaching patterns as they related to soil pore distributions. There are a number of different dye compounds that have been used with greater or lesser success to trace the movement of water in a soil matrix. For this experiment, a dye was required that was acid-stable (the soil pH was quite low), anionic (so that it would act similarly to the nitrate ion), inexpensive (the large size of the columns could require large quantities of dye) and easy to detect. Methylene blue (MeBl) has been used to mark water transmission routes in soils (White, 1985c; Bouma & Dekker, 1978). As it was readily available, relatively cheap, and easily detected under visible light, it was chosen for this experiment. Preliminary investigations showed that MeBl stains showed clearly on the B horizon of Ladner soils, but not against the darker A horizon. Tests on isolated blocks of soil of maximum 20 cm height indicated that 0.1% MeBl solution applied to the top of the block stained pores throughout the full height of 50 the block. Therefore, following the hydraulic conductivity determinations, dye tests were undertaken on the soil columns, using 0.1% MeBl. Dye tests were run on 4 columns initially, (# 6, 7, 15, and 17), one from each water table treatment level. The procedure followed was as described for Run 3, including a preflush, except that no half-way samples were collected, and 0.1% MeBl in distilled water was used instead of nutrient stock and distilled water. In total, 850 mL dye solution was added to each column over 54.5 hours (The quantity of water added and duration of precipitation are discussed in the next section on Run Procedures). After incubation and drainage, the columns were cut into cross-sections. Sections were cut every 5 cm (the smallest practical height) from the top of the soil until no dye was found, then every 20 cm to allow observation of the soil core. Only 2 columns (#15 and #17) were sectioned because dye penetration was extremely poor. Dye was observed at a maximum depth of only a few centimeters below the surface. The dye tests were therefore repeated, using columns 6 and 7 again, and column 1. Dye concentration was increased to 1% MeBl. Columns 6 and 7 received 800 mL dye solution in total over.70 hours, while column 1 received only 590 mL total due to extremely poor infiltration. After incubation and drainage, the columns were sectioned in 5 cm slices from the soil surface until the maximum depth of dye penetration was reached. H. Run Procedures I. Run 0 Prior to addition of any nutrients, the soil columns were leached with distilled water to check for differences in background concentrations of the constituents of interest. Water was added to the columns "from the bottom up" by siphoning through the drain tube, from buckets placed roughly level with the soil surfaces, creating an initial head of about 1 m. The time for water levels to reach the soil surfaces was extremely variable, ranging from 15 minutes to approximately 4 days. After allowing the water tables to remain at the surface for 2 days, the drain tubes were opened and leachate collected. NO^ + N 0 2 , NH^" 1 ", CI", ortho-51 P 0 4 and pH were measured on all samples. Methods of chemical analysis are described in a later section. 2. Run 1 At the start of Run 1 the water table indicator tubes were open and functional, the drainage tubes were closed and the water tables were at <0 cm. 25 mL concentrated nutrient solution was added followed by two additions of 65 mL distilled water and 9 additions of 70 mL distilled water at roughly 3 hour intervals (interrupted by two overnight breaks). Total precipitation was 785 mL over a period of 54 hours. Some of the columns with very slow infiltration rates received only 715 mL due to ponding of water on the soil surface. Water table control in Run 1 was attempted by draining leachate from individual columns as the water table in each column approached the desired treatment level. This was an extremely inefficient and inaccurate method of water table control: almost constant attention was required to prevent the water tables from greatly exceeding the desired levels. Unfortunately, 17 of the 21 columns in Run 1 were leaking quite steadily by the end of the run. Although accurate volumes of leachate could not therefore be obtained, the experiment was continued to get some idea of leachate nutrient concentrations. Columns were drained 4 days after the last precipitation event. For each column, all leachate was combined and analyzed for pH, NO3 + N 0 2 , N H 4 + , and ortho -P0 4 . At this point it is appropriate to discuss the quantity and duration of precipitation in the experiment. The amount of water added in each run was determined primarily by the water holding capacity of the soil: the amount of water added had to be sufficient to raise the water table to the maximum desired height. Assuming 6% drainable porosity, approximately 820 mL of water would be held under a 75 cm water table in columns of the diameter used in these experiments. The actual amount of water added varied from the calculated amount because some of the columns required more water to obtain 75 cm water tables. As the amount of nitrate leached is dependent on the total amount of water passing through a volume of soil (Kolenbrander, 1981), the amount of water added was kept the same for all 52 columns, except in the few cases where infiltration was so poor that water addition was stopped for specific columns before the others. The length of time over which water was added varied from the planned two day period because of slow infiltration rates. 3. Pre-Flush Between each of Runs 1, 2, and 3 the water tables in all columns were raised to the soil surface by siphoning distilled water through the drainage tube, as described under Run 1. Experimentation with two preliminary columns showed that repeated leachings reduced infiltration rates into the soil. It was hoped that reversing the direction of water flow between runs would prevent any reduction in infiltration from becoming severe, whether it was due to slaking of clay particles blocking pores in the soil, or due to air entrapment from repeated addition of water to the relatively small enclosed volumes of soil. Usually between 1/2 hour to 4 days was required to raise the water tables by siphoning, the time being characteristic of each column. After 2 days incubation with surface water tables, the drainage tubes were opened and leachate collected for analysis. Columns were always allowed to drain for 48 hours prior to starting another run. It was assumed that after a 48 hour drainage, the moisture content in the columns was close to field capacity. Preflush leachate was collected, measured in volume, and analyzed for pH, NO^ + NO2, N H 4 " 1 " , Cl " , and ortho-PC^. 4. Run 2 In Run 2 the water table was pre-set by siphoning through the bottom of the columns over a 2 day period. After an additional 2 days of adjusting the water tables, the desired treatment levels had still not been achieved, but the water table in each column was reasonably close to the desired treatment level. Then, 25 mL concentrated nutrient solution and 25 mL distilled water was added to each column, followed 3 hours later by 80 mL distilled H2O. Examination of the water tables after an overnight 12 hour equilibration showed that the water table rise had been so variable that the columns were breaking out of their treatment groupings, and so no further "precipitation" was added. 53 The columns were incubated for 2 days, then drained in two batches. Leachate draining in the first 20 minutes was collected separately from that draining over the next 24 hours. The two batches, henceforth referred to as "20 minute" and "24 hour" leachate, were analyzed separately for pH, NO3 + N 0 2 , N H 4 + , or tho -P0 4 , and CI". Unfortunately, leaking remained a problem throughout the experiment. 5. Runs 3 and 4 There was a considerable delay between Run 2 and Run 3 during which the columns spent approximately 2 weeks at room temperature while a new cold room was built and while the half-way samplers were modified. The set-up in Runs 3 and 4 was markedly different from the previous runs, and finally achieved good water table control. Except with variations in timing, the same procedure was followed in Run 4 as in Run 3. Run 4 is, in effect, a continuation of Run 3, with no flushing between the runs. The runs commenced with <0 cm water tables and columns at field capacity. The drainage tube for each column opened into an overflow bottle at the desired water table level, as described in the section on column configuration. Run 3 was started with addition of 25 mL concentrated nutrient solution and 25 mL distilled water. Precipitation events of 80 mL distilled water followed at, on average, every 3 hours, with the exception of 2 overnight breaks. Total water addition was 850 mL over approximately 49 hours. Samples were taken from the half-way samplers 1 hour after the last precipitation event. Columns were incubated for 2 days, then drained for 48 hours. The overflow bottles in the lowest 2 treatments started filling after only a few precipitation events, while for some columns with high water tables there was no overflow volume at all. Overflow bottles were changed either as they filled or once per day. Each overflow bottle for each column was analyzed separately to check for trends in leachate quality over time. Samples from the overflow bottles are henceforth referred to collectively as 54 " O F " and as OF1, OF2, OF3 in reference to columns where multiple bottles were obtained. Samples were analyzed for pH, NO3 + N 0 2 , N H 4 + and Cl " . In Run 4 no nutrient was added. Precipitation events consisted of 80 mL additions of distilled water, very roughly every 5.5 hours, interrupted by 3 overnight breaks. The timing of the water additions was dictated by avoidance of ponding due to decreasing infiltration rates. Total addition of water was 880 mL (except for column 1 which received only 560 mL due to ponding) over 93 hours. Samples from the overflow bottles were handled as in Run 3. Samples were obtained from the half-way samplers approximately 4 hours after the last precipitation event and approximately 1 day after that. These samples are referred to as " H W 1 " and "HW2" , respectively. Due to the very long time period over which the columns received water, it was difficult to determine the true "incubation period": the water table in many columns had reached the desired level long before the end of precipitation. Therefore, the columns were drained 36 hours after the last precipitation event, creating periods of incubation ranging from about 110 hours to 28 hours. In both Runs 3 and 4, this drainage was collected in one batch and is referred to as "Final Drainage" or "FD". The total amount of solute leached in all sample events for each run was calculated and divided by the total volume of leachate for the run to obtain an overall concentration for each run; these values are referred to as "Run 3 Total " (R3TOT) and Run 4 Total (R4TOT). "Grand Total" (GRTOT) refers to the overall total for Runs 3 and 4 combined. All samples were analyzed for pH, NO3', NO2", and Cl " . I. Nutrient Application Rates Nutrient application rates were set to be equal to rates used at the Boundary Bay Water Control site, which were based on recommended fertilization rates. The fertilization rate used in the first spring fertilization at Boundary Bay was 35 kg/ha of N for forage on March 29 (Richard & Chieng, 1985). 55 Nutrient application rates for all runs are shown in Table 11. Nitrogen was applied as KNO3, chloride as K G , and phosphate as K2HPO4. All chemicals were analytical grade. All solutions were made with single distilled water. J. Chemical Analyses Soil samples obtained in the field were stored at 2 C until analyzed. 2 M KCI extractions were performed according to Keeney and Nelson (1982). In brief, approximately 20 g field-wet soil was weighed into a polypropylene bottle. After addition of 100 mL 2 M (analytical grade) KCI the bottle was closed and shaken on a shaking table for 1 hour. Samples were allowed to settle, then poured through Whatman #42 (no ash, quantitative) filter paper. The high clay content of many of the B horizon samples caused extremely slow settling following shaking. Such samples were centrifuged for 5 minutes, then filtered. Filtrate was stored at approximately 2 C until analysis for NO3 + N 0 2 and NH^"*". All samples were extracted within 48 hours of removal from the field. Due to the large number of samples (52), duplicate extractions were performed on only 8 of the samples. Gravimetric moisture determinations were made simultaneously with the KCI extractions. Samples were weighed on an analytical balance prior to and after drying at 105 C for 24 hours. Moisture determinations were made in duplicate. All leachate samples were stored between 1 - 4 C in polyethylene bottles until analyzed. Analyses were usually performed within 48 hours for nitrate and nitrite and within a week for all other constituents. Total volume and pH were usually measured immediately. Samples were shaken thoroughly before analysis, and filtered through Whatman #42 filter paper if the sample was cloudy or contained sediment that did not settle immediately. For samples that appeared to be clear, preliminary tests indicated that for NO3 + N 0 2 , N0 2 ", NO3" and C l " analyses, filtration did not affect the analytical result at the level of precision in this experiment. For both soil extracts and column leachate, samples were usually analyzed as duplicates, with several samples run in triplicate (or greater) for each analytical batch, to check on Table II. Nutrient application rates Mass per Column (mg)* Field Rate (kg/ha) ** Run N CI P N CI P 0 0.0 0.0 0.0 0 0 0 1 64.0 40.0 29.2 35 22 16 2 128.0 46.4 58.4 70 26 32 3 63.5 39.5 0.0 35 22 0 4 0.0 0.0 0.0 0 0 0 Notes: * Mass per column, as measured in the nutrient solution applied to the soil columns. ** Mass applied as equivalent field rate. 57 precision. The method of standard addition was also used to check accuracy. For each analytical batch, standards were run at least twice (at the beginning and end of a run), but usually once per approximately 40 samples. Standard absorbances were used to generate regression lines of absorbance versus concentration; the regression curves were used to calculate sample concentrations. Measurement of pH was performed at room temperature with a Good 201 ATC digital pH meter. All chemical analyses were made on a Technicon Autoanalyzer II, using the following methods: 1) ( N 0 3 + N 0 2 ) -N , and N 0 2 " -N only - Technicon Industrial Method #33-69W (Technicon Corp., 1969) and Technicon Industrial Method #161-71W (Tech. Indust. Syst., 1973b), respectively; hydrazine sulphate reduction of NO3" to N0 2 "; diazotization of N 0 2 " with sulfanilimide followed by coupling with N-1-NEDD. 2) N H 4 + -N - Technicon Industrial Method #90-70W for ammonia in the range of 0 - 10 mg/l (ppm), and Method #154-71W for ammonia in the range of 0 -140 ug-N/l (ppb) (Technicon Indust. Syst., 1973a); both utilizing the Berthelot Reaction. 3) CI" - Technicon Industrial Method #99-70W (Technicon Indust. Syst.,1971); formation of ferric thiocyanate for colorimetric determination following liberation of thiocyanate from mercuric thiocyanate by formation of mercuric chloride. 4) Orthophosphate-P - Technicon Industrial Method #155-71W (Tech. Indust. Syst., 1973c); formation of phosphomolybdenum blue on reaction with acidified ammonium molybdate. As recommended by the U.S. Environmental Protection Agency (EPA, 1972), analytical precision will be reported as the range of standard deviations obtained for replicate measurements of actual samples spanning the range of concentrations encountered. Three to 6 replicates were analyzed at each concentration. Analytical precision was as follows: 1) (NO3 + N 0 2 ) -N : for 0.46 to 2.15 mg N0 3 ' -N/L, the standard deviation was +/-0.006 to +/-0.05. 2) NH 3 -N : fo r 17 to 125 ug NH3-N/L, the standard deviation was +/-2.0 to +/-10.2. 3) CI": for 1.1 to 9.8 mg Cl'/L, the standard deviation was +/-0.04 to +/-0.49. Precision was improved in Runs 3 and 4, with standard deviations ranging from + /-0.08 to +/-0.18 for 3.2 to 5.9 mg CI7L. 4) Orthophosphate-P: for 0.89 to 2.64 mg P/L, the standard deviation was +/-0.03 to +/-0.19. 58 K. Experimental Design and Statistical Analyses Columns were assigned treatments by random draw. The main experiment was a single factor completely random design. The treatment variable - water table height - had four levels: 15 cm, 35 cm, 55 cm and 75 cm above the column bottom (equivalent to height above drain depth). Each treatment was replicated 5 times (ie. 5 columns per treatment); however, during Run 1 a column in the 15 cm treatment stopped infiltrating water and was discarded, reducing the number of replications to 4 for the lowest water table. Statistical tests were performed on data from Runs 3 and 4, for the following chemical constituents: (NO^ + N 0 2 ) -N and C l " concentration, and the N/CI ratio. The N/CI ratio is simply the ratio of the concentration of (NO3 + N 0 2 ) -N to the concentration of CI" recovered in a particular sample. All statistical tests and data transformations, except as noted, were run on SPSSx, Version 2.1 on the U.B.C. mainframe computer. The data for leaching loss of each chemical constituent for each sampling event for each run was analyzed separately. Results of successive runs or even sample event events could not be analyzed together or the assumption of independence of observations would be violated. However, an overall "Run Total" concentration was calculated and statistically analyzed for each chemical constituent for each run. Application of most statistical techniques requires that the data being tested meet certain assumptions. Analysis of variance and covariance require: (i) error effects to be normally distributed (with a mean of zero) and independent; (ii) homogeneity of variance for all treatments; (iii) independence of treatment means and variances; and (iv) additivity of the main effects (Little & Hills, 1978; Kirk, 1982). Moon and Schreier (1985), among others, have pointed out that many soil properties are not normally distributed and that parametric techniques may often be inappropriate for soil data. Testing for normality is perhaps one of the most difficult tests to make. The 59 Kolmogorov-Smirnov type tests were used because they are probably the more efficient goodness of fit test when compared with the chi-squared test, and are appropriate for small sample sizes (Conover, 1972; Pfaffenberger & Patterson, 1977). SPSSx 'NPAR TESTS K-S (normal)' was used to generate the greatest absolute difference between a sample cumulative distribution and a normal distribution with sample mean and variance. This value was then tested for significance using Lilliefors modification for unknown population mean and variance (Lilliefors, 1967; Conover, 1972). Homogeneity of error variance was checked with the Bartlett-Box F test. Kozak (1986, personal communication) recommended performing the test on the observations after subtraction of the treatment means. SPSSx 'ONEWAY' was then used to generate the Bartlett-Box F tests. Hicks (1982) recommends accepting homogeneity of variance if the calculated probability is "well above" 0.05. If means and variances are related, the population may not be normally distributed, and a transformation of the data may be indicated (Little & Hills, 1978). Independence of means and variance is discussed further in a later section. Correlation analysis was run on SPSSx 'NPAR CORR' for Spearman's rho (r s). Spearman's rho is a nonparametric rank correlation coefficient which measures the strength of a relationship, not necessarily linear, between 2 variables (Siegel, 1956). An r $ of 0 indicates no correlation at all between variables, while r g of +/- 1 indicates perfect correlation. Trends in leachate concentrations were tested with the Friedman 2-way analysis of variance for related samples (Siegel, 1956) and the Wilcoxon, matched-pairs signed-rank test, the non-parametric equivalent of the paired t-test. The Wilcoxon tests were run on SPSSx "NPAR TESTS WILCOXON ' . 60 IV. R E S U L T S A N D D I S C U S S I O N A. Field Samples The objectives for analyzing soil samples taken from near the columns were to obtain an estimate of the levels of easily-extractable (potentially mobile) N compounds in the soil, to obtain an indication of the variability of N content that may be expected within the columns, and to identify columns which may have very different N contents from the other columns. The results from 2 M KCI extracts of the soil samples are presented in Table III, where the samples are identified by the ID numbers of the nearest leaching columns. The extractable N H 4 + -N content of all B horizon samples was very similar, both from above and below 50 cm. The extractable N H 4 + -N content of the A horizon, however, was considerably more variable (the standard deviation was much higher for the A horizon samples than for either set of B horizon samples). As expected, the NO^ ' -N content of the field samples was very low. from all depths only 2 of 49 samples analyzed had a content greater than 1 mg NO-j'-N/kg dry soil. Most samples had N contents at the limit of detection of the analytical procedure used (0.04 mg NO^'-N/L of extractant). Thus there was no reason to discard any columns from the experiment on the basis of initial N content. Based on a bulk density of 1.39 g/cm 3 (Driehuyzen, 1983) the 1 m profile at the site from which the columns were taken contained an average of 74 kg extractable N H 4 + -N/ha (range 57 - 98 kg N/ha) and an average of 6 kg NC>3"-N/ha (range 4 - 22 kg N/ha). B. Physical Measurements 1. Drainable Porosity The drainable pore volume (or drainable porosity) is defined as "..the volume of water that can be drained from a unit volume of soil when the soil-moisture pressure is decreased from atmospheric pressure to some specific negative pressure", such as that created by a falling water table (Luthin, 1978). The volume drained is usually not a linear response to T a b l e I I I . KCI - e x t r a c t a b l e n i t r a t e and ammonium i n f i e l d sample C l o s e s t N i t r a t e - N Ammonium - N Column A H o r i z B H o r i z B H o r i z A H o r i z B H o r i z B H o r i z ID # Top Bottom Top Bottom 4,6 0.65 0.31 *0.27 7.14 3.75 3.46 18 0.45 *0.25 *0.17 7 .16 3.46 2.47 12 0.43 *0.24 *0.25 11.28 3.47 3.83 2 0.34 *0.26 0.35 10.95 3.32 3.45 14 0.60 *0.25 0.58 11.22 3.62 3.39 8 0.59 *0.24 0.35 14.07 3.78 3.25 16 0.33 *0.28 *0.24 7.67 3.51 3.19 13 *0.30 *0.25 0.26 8.37 3.59 3.23 3 0.49 *0.24 0.30 12.97 3.88 3.52 9 *0.29 0.28 *0.25 8.01 3.81 3.47 5 0.35 *0.24 *0.25 7.02 3.56 3.62 19 *0.31 *0.25 ns 6.30 3.62 ns 21 *0.28 0.39 2.89 7.48 3.49 4.53 10,11 *0.28 0.35 ns 8.78 3.58 ns 7,15 *0.30 0.35 *0.25 15.71 3.47 3.33 1,20 1.84 *0.24 *0.25 9.13 3.78 3.45 17 0.62 0.26 0.25 10.87 3.82 3.25 Mean 0.50 0.28 0.46 9.65 3.62 3.43 s .d. 0.37 0.05 0.68 2.74 0.16 0.42 Not e s : V a l u e s a r e mg N/kg d r y s o i l A H o r i z o n = 0 - 30 cm B H o r i z o n Top = 30 - 50 cm B H o r i z o n Bottom = 50 - 100 cm ns = no sample * = sample a t d e t e c t i o n l i m i t ; a c t u a l v a l u e i s l e s s than or e q u a l t o t h e v a l u e shown. Means and s t a n d a r d d e v i a t i o n s ( s . d . ) were c a l c u l a t e d u s i n g the t a b l e d v a l u e s f o r samples a t t h e d e t e c t i o n l i m i t . 62 changing suction, and so to compare drainable porosities between soil samples the same suction should be applied to each sample. For this experiment, an estimate of the drainable porosity of each column was obtained by allowing free drainage from a saturated column (ie. water tables at the soil surface). As the length of all soil columns was quite similar (0.88 - 0.99 m) the difference in suction created by the falling water tables should have been minimal between columns. The volume of water held in the soil was calculated by averaging the volumes drained in Run 0, and Runs 2 and 3 preflushes, then subtracting the volume of water held in the sand filter and tubing of each column. Volumes calculated in this way are only estimates of the true water holding capacity, because air entrapment within the columns is practically unavoidable. Average drained volumes for each column were then divided by the volume of soil in each column. From the data presented in Table IV it appears that there is quite a range in drainable porosities among columns. Expected drainable porosities for this soil for 100 cm suction are around 6% (Chao, 1987); almost half of the columns have drainable porosities greater than 1% different from this. The variation in drainable porosity likely is due to texture differences between the soil columns, and to differences in bulk density induced by compaction during placement. 2. Hydraulic Conductivity During the measurement of hydraulic conductivity, the columns required 4 - 5 hours of leaching before the volume of outflow stabilized. Hydraulic conductivity was calculated for the average volume using Darcy's law for 1 dimensional flow (Klute, 1965): T a b l e IV . P h y s i c a l c h a r a c t e r i s t i c s o f s o i l co lumns Column D r a i n a b l e K s a t * L o g ( K s a t ) P o r o s i t y # (%) (cm/day) 7 ns 35 1.54 9 5.9 214 2 .33 14 5.6 148 2.17 16 6.6 199 2 .30 M e a n ( s . d . ) 6.0 ( 0 . 5 ) 142 2 .09 5 7.4 744 2.87 8 2.8 1 0 10 7.1 1 0 13 7.3 159 2 .20 20 2.7 8 0 .90 M e a n ( s . d . ) 5.5 ( 2 . 5 ) 109 1.19 1 3.9 3 0 .48 6 5.6 166 2 .22 11 4.0 27 1.43 17 3.7 35 1.54 19 10.4 1468 3.17 M e a n ( s . d . ) 5.5 ( 2 . 8 ) 185 1.77 2 7.1 77 1.89 4 6.2 84 1.92 15 ns 49 1.69 18 9.9 268 2 .43 21 3.2 12 1.08 M e a n ( s . d . ) 6.6 ( 2 . 8 ) 83 1.80 O v e r a l l M e a n ( s . d . ) 5.8 ( 2 . 3 ) 121 1.69 ( 0 . 8 8 ) N o t e s : s . d . - S t a n d a r d d e v i a t i o n * - Means f o r K s a t a r e c a l c u l a t e d f r o m l o g v a l u e s 64 K = (Q/At) ( L / 4H) where A = cross sectional area (cm^) K = hydraulic conductivity (cm/min) Q = volume of water (cmr) collected t = time (min) to collect water L = length of the soil core (cm) AH = the hydraulic head difference (cm) between the soil column bottom surface and the top of the constant depth of water above the soil column. From the K$a^ values reported in Table IV, it is obvious that the columns vary significantly in hydraulic characteristics, but large variation in saturated hydraulic conductivity measurements is to be expected (Warrick & Nielsen, 1980; Klute, 1965). In a comparison of different studies, Warrick & Nielsen (1980) found that the coefficient of variation (CV) for saturated hydraulic conductivity values ranged from 86% to 190%. The CV for the K s a t measurements in this experiment is 181%, indicating that the data are not more variable than other studies. The magnitude of measured K $ a j . is very high. The hydraulic conductivity of a clayey soil is characteristically of the order of magnitude of 1 0 - 4 - 1 0 - ^ cm/sec (Hillel, 1980). Eleven of 19 columns have conductivities that fall within this range; 7 are in the range of 1 0 - ^ cm/sec and 1 is at 10 cm/sec. According to Hillel, sandy soils have conductivities in the range 10 __) - 1 0 cm/sec, and therefore the latter higher experimental conductivities may be artificially high. Driehuyzen (1983) reports average auger hole method conductivities at the Boundary Bay site of 28 - 98 cm/day, with a range of individual measurements from 25 to 124 cm/day, for the 0 - 1 m profile. Approximately 1/3 of the soil columns used here have conductivities that are comparable to field data, but several of the columns have extremely high conductivities, or very low conductivities. Both inherent soil structural variability and the design and nature of the laboratory columns could have contributed to the variability and magnitude of measured conductivities. With soil columns as large as in this experiment, determination of hydraulic conductivity can 65 be quite difficult. The limitations of the simple method of hydraulic conductivity determination used here include: 1) Air entrapment within the soil, resulting from wetting the samples up without using a vacuum (Klute, 1965). However, air entrapment should have been minimized by slow wetting from the bottom of the soil core. The measurements thus estimate "satiated" rather than "saturated" hydraulic conductivity . 2) Possible edge flow within the cores (Klute, 1965). 3) Variation in the head. Although the head above the soil surface was kept constant, the relatively large surface area (182 cm^) often varied in micro-relief, so that the actual head experienced may not have been exactly 5 cm. When the head was set, an attempt was made to compensate for varying surface levels. All of the factors above could contribute to variations in measured K s a t . Also, there is a fairly strong positive correlation between drainable porosity and K s a t (Spearman's r $ = 0.711, significant at <=< =0.01), indicating that variations in texture and structure may have contributed to the variations in hydraulic conductivity. The relationship between K s a t and drainable porosity is illustrated in Figure 4. Columns with coarser textures would be expected to have higher saturated conductivities than finer textured soils (Hillel, 1980), and also to have lower moisture retention at a particular level of suction, and therefore a higher drainable porosity. Additionally, the presence of large cracks or channels within the soil core can influence hydraulic conductivity (Klute, 1965). However, the purpose of the K g a t determinations was not to determine accurate field conductivities, but rather to measure conductivities actually occurring in the experimental columns, including effects resulting from column design, and to study the influence of the actual conductivities on the results of the leaching experiments. The Kolmogorov-Smirnov goodness-of-fit test with Lilliefors modification was used to test the distribution of hydraulic conductivity data. For the raw K s a t data, the greatest deviation of the sample cumulative frequency distribution (CDF) from a normal CDF with 66 12-10-in O CL 6-_© D 4 -2-0-0 1 2 3 4 Log Satiated Hydraulic Conductivity (cm/day) F i g u r e 4 . R e l a t i o n s h i p between d r a i n a b l e p o r o s i t y and h y d r a u l i c c o n d u c t i v i t y 67 sample mean and standard deviation is 0.32, which is significant at cx <0.01 (Lilliefors, 1967) and leads to strong rejection of the null hypothesis of normality of distribution. Square root and log transformations of the data both yielded maximum deviations of 0.13 and acceptance of the assumption of normality of distribution of the transformed data (not significant even at <*= 0.20). According to Warrick & Nielsen (1980), normal and log-normal distributions are commonly assumed for soil physical properties; therefore, the log transformation was chosen over the square root transformation for the conductivity data. Spearman's rank correlation coefficient was used to test whether the random assignments of columns to treatment groups had resulted in a random distribution of hydraulic conductivities across treatment groups as well. R$, with correction for tied ranks and observations, was calculated as -0.054, which was not significant as a 2-tailed test at even the 0.20 level of significance. Thus the null hypothesis of no correlation between variables can not be rejected: the calculated r g indicates that there was no tendency for columns with higher conductivities to be assigned to the lower or higher water table heights. As K g a t was measured after the leaching tests were conducted, this was a fortunate but not predictable result of a random assignment procedure. 3. Dye Tests A blue methylene blue stain in the soil indicates that infiltrating water passed through that particular pore space. MeBl adsorbs readily to soil; water in dye solution flowing down channels may penetrate into soil peds, but the MeBl will remain on the ped surface as the uncolored water penetrates further into the ped (Bouma & Dekker, 1978). MeBl therefore indicates the initial paths of solution penetration, but may not indicate subsequent redistribution of water and solutes such as NO-j". Figures 5 - 7 show the patterns of MeBl observed at various depths in columns 1, 6 and 7. The maximum depth of penetration was 76 cm for column 1, 80 cm for column 6, and 60 68 F i g u r e 6 . Dye p a t t e r n s i n co lumn 6 70 71 F i g u r e 7 ( c o n t , ) . Dye p a t t e r n s i n c o l u m n 7 \ 74 cm V 70 cm <$ - h o l e , c o m p l e t e l y s t a i n e d 65 cm ( d y e q u i t e p a l e a t t h i s d e p t h ) J 72 cm for column 7. As mentioned previously, dye penetration in columns 15 and 17 was very poor, so no cross-sections are shown. The dye patterns illustrate two points quite clearly. First, boundary flow occurred to some extent in all of the columns leached with dye, however, in no case were the edges the only dyed areas nor were they the major dyed areas, in the lower part of the dyed depth of column 7, dye stains along the soil/pipe boundary were often discontinuous, appearing for up to several centimeters and then disappearing again. The large dye stain at the edge of the 88 cm section of column 6 was continuous to the surface, but disappeared a few millimeters below the exposed section. Thus it appears that most of the water added to the tops of the columns does enter the soil matrix, and, at least in some cases, water and dye travelling down the pipe surface may re-enter the soil. The second point to note is that the dye stains in the soil indicate that water penetrating the columns flows preferentially down larger channels or cracks, rather than by piston-type displacement. Although the tops of all three columns were uniformly stained (except for a few small areas of high microrelief), the only cross-section showing any evidence of plug or piston flow is column 7, 88 cm. In column 7, at 88 cm (5 cm below the surface), a very large area of soil was stained and therefore still involved in water transport. The soil at this depth was crumbly, consisting of loosely packed aggregates of up to a few millimeters in diameter, and with very few distinct channels. By only 82 cm depth, the flow of water and dye had concentrated within a much smaller area, and, in this section and those below it, the dyed areas were often obviously pores or root channels. The dye patterns in column 7 may be comparable to the results of Tyler and Thomas (1981) who found displacement-type flow in the unstructured plow layer, followed by preferential flow in the lower, more structured, profile. There was rapid attenuation of the soil area involved in water transport in both columns 1 and 6. For column 1, by 90 cm (6 cm from the top), the stains are quite concentrated and 73 by 85 cm the stains are distributed through a large number of distinct channels of small area, indicating that only a small area of the soil is involved in water transport. Dye penetration in column 6 was quite poor and very concentrated. Most of the dye and water appeared to channel around the large stone shown in the 88 cm section (4 cm below the surface). By 85 cm, there was only one weakly-colored stain on the boundary, and 2 strongly dyed roots. Dyed roots were a common observation in all three columns. The patterns of dye penetration result from differences in structure and perhaps texture between the columns. Column 6 appeared to be clayey with a massive structure throughout the column, except for irregular patches of fine crumbly organic material. This is in contrast to columns 1 and 7, which had more organic, crumbly tops, grading into silty and clayey bottoms with large distinct channels throughout and cracks in the upper regions. Both columns 6 and 1 exhibited poor infiltration throughout the experiment, but 1 was much worse than 6 (column 1 received less dye solution than 6 or 7 because of ponding). The deeper dye penetration in column 7 may therefore result from good infiltration (good surface structure) and an abundance of conducting channels. The slow infiltration in columns 6 and 1 may have allowed more adsorption of dye near the surface, resulting in less dye delivery below the surface. Bouma and Dekker (1978) also note that dye adsorption to soil increases as pore water velocity decreases. Columns 15 and 17 also had poor infiltration throughout the experiment and this may have been responsible for the very poor dye penetration in those columns. The poor dye penetration in column 6 was somewhat surprising because of the relatively high measured K s a t for this column. However, K s a t was measured under saturated conditions and constant head, while the dye experiments were conducted under mixed saturated and unsaturated conditions with intermittent water application. The flow patterns of water in the soil columns under the two different sets of experimental conditions are likely to be quite different. Also, the dye tests were conducted after K t was measured, and the 74 extended saturation and ponding in the conductivity determinations may have damaged surface structure and reduced the infiltration rate in some columns (such as 6). C. Runs 0 to 2 Leachate Concentrations The mean pH of the leachate increased throughout the experiment, from 4.41 (+/-0.70) in Run 0 to 5.66 (+/-0.42) in Run 4 Final Drainage. The low pH was expected in view of the low soil pH's in Table I. 1. Ortho-Phosphate Table V lists the concentrations of orthophosphate-P measured in the leachate of Runs 0 to 2, including the preflush. Mean orthophosphate-P concentrations in leachate were generally low throughout Runs 0 to 2, with only a few columns showing P concentration greater than 1 mg/L. In Run 0 (the leaching of untreated columns) P levels were more variable than in subsequent runs - there were higher standard deviations of the mean for each treatment group. This reflects variations in initial P status of the columns, with respect to easily leached P. Nonetheless, all the groups exhibited similar mean orthophosphate-P levels, except for the 35 cm water table treatment, which included column 10 with much higher than average orthophosphate-P levels. Phosphate concentrations in leachate from Runs 1 and 2 remained low, with phosphate levels dropping markedly in Run 1 compared to Run 0. A few columns (5, 13, 19 and 2) showed increased P leaching in the Run 2 preflush compared with Run 1, and many showed slight increases again in Run 2. The objective of the phosphate leaching tests was to check for obvious boundary or crack flow. Large amounts of boundary flow would indicate columns that were unsuited for leaching studies. The phosphate ion is strongly retained in many soils, especially if the content of clay (particularly 1:1 clays) and Al and Fe hydroxides is high, and the soil is acidic (Tisdale et at, 1985). The Ladner soil would be expected to retain P. The presence of the phosphate ion in high concentrations in leachate following nutrient addition would likely Table V. Runs 0-2 ortho-phosphate concentration Water Column Run 0 Run 1 Pre- R u n 2 7 ns 0.20 0.14 0.22 0.21 15 9 0.86 0.42 0.24 0.35 0.29 cm 14 1.32 0.63 0.39 0.32 0.23 16 0.89 0.55 0.34 0.45 0.36 mean 1.02 0.45 0.28 0.34 0.27 s.d. 0.26 0.19 0.11 0.09 0.07 5 0.54 0.15 0.39 0.49 0.43 35 8 1.56 0.22 0.14 ns ns cm 10 7.63 2.64 0.45 0.22 0.32 13 1.75 0.41 0.59 1.42 0.81 20 1.57 0.42 0.25 0.13 0.20 mean 2.61 0.77 0.36 0.57 0.44 s.d. 2.85 1.05 0.17 0.60 0.26 1 0.91 0.30 0.18 0.21 0.18 55 6 2.14 1.00 0.19 0.31 0.29 cm 11 0.12 0.11 0.12 0.20 0.16 17 0.96 ns 0.22 0.22 0.20 19 0.69 0.43 0.53 0.48 0.48 mean 0.96 0.46 0.25 0.28 0.26 s.d. 0.74 0.38 0.16 0.12 0.13 2 0.78 0.27 0.68 ns ns 75 4 1.00 1.64 1.04 0.39 0.25 cm 15 0.40 0.38 0.22 ns ns 18 0.88 0.35 0.33 0.55 0.43 21 1.37 0.82 0.25 ns ns mean 0.89 0.69 0.50 0.47 0.34 s.d. 0.35 0.57 0.35 Notes: ns - no sample s.d. - standard deviation Concentrations are mg-P/L 76 indicate (low bypassing the bulk of the soil matrix, by, for example, flowing down large cracks or channels (Thomas and Phillips, 1979; Bottcher ef al, 1981), or the soil-pipe interface or boundary. To evaluate whether "high concentrations" of phosphate ion are present in the leachate, the initial concentration of phosphate applied and the potential diluted concentration of phosphate in the leachate must both be estimated and then compared to measured leachate concentrations. It is difficult to express an "initial concentration" for phosphate in these experiments, because nutrient was applied in a concentrated slug (of 1167 mg P/L for Run 1) and then diluted. However, if the total mass added was diluted by the total volume of water added, then the PO4-P concentration would be approximately 37 mg/L for Run 1. For Run 2, the initial concentrated application was at 2334 mg-P/L, with a potential dilution to approximately 450 mg-P/L due to precipitation added. However, the calculation of initial concentrations, is even more complicated in Run 2 because pre-set water tables were present with variable quantities of water in each column. The concept of "potential dilution" is based on the plug flow theory of water movement. In the hypothetical case of a column of an inert matrix, an inert solute, and complete mixing of the solute within the total volume of solvent, the leachate concentration would be equal to the mass of applied solute divided by the total volume of solvent. Similarly, if the solute was applied in an initial "batch" (as in these experiments), and the solvent applied in aliquots over an extended time period (also as in these experiments), then, the leachate first reaching the column bottom would be expected to have the highest concentration, followed by successively more dilute leachate. If the column were perfectly drained (retaining no leachate in dead-end pore spaces or narrow capillary pores), then if all of the leachate was combined, the final concentration of this mixed leachate would be expected to have a concentration equal to the added mass of solute divided by the total volume of solvent added. 77 The concentrations referred to above as "potential dilutions" are calculated in this way. Leachate concentrations below this potential dilution result from a number of possible causes, including non-plug flow, imperfectly draining columns, and solutes interacting chemically or physically with the soil matrix. The leachate concentrations of phosphate are well below both the application concentrations and potential dilutions for all columns, indicating that there is likely no major problem with boundary (edge) flow or large artificial channels. The slight rise in phosphate concentrations in Run 2 leachate is likely the result of phosphate applied in Run 1 finally reaching the column bottoms through larger natural channels. In view of the low levels of leachate phosphorus, and because considerable difficulty was experienced with the phosphate analyses, phosphate addition and analysis was discontinued after Run 2. 2. Chlor ides Chloride concentrations for Runs 0 - 2 are presented in Table VI. The leaching of Run 0 was intended to provide a comparison of initial chloride levels in the columns. The chloride concentrations in Run 0 were quite high for all columns. Dilute HCI was used to wash the caps, sand and tubing prior to installation. It is possible that the leachate CI" concentrations were artificially high due to insufficient rinsing of the tubing and sand prior to installation. This idea is supported by a significant, negative correlation between the pH and chloride concentration in Run 0 (Spearman's r s = -0.569, significant at oc = 0.02), and the lack of a significant correlation between pH and CI" in later Runs (eg. for Run 2, 20 minute r g = 0.364, not significant at <*= 0.10). In the Run 2 preflush, the chloride concentrations dropped considerably for all columns. The concentrations were still quite variable between columns, as evidenced by the high standard deviations, but there appears to be little difference in CI" concentration between the treatment groups. 78 Table VI Runs 0-2 chloride concentrations Water iColumn Run 0 Pre- R u n 2 Table # flush 20 Min 24 Hour Total 7 ns 5.8 12.1 4.2 9.1 15 9 24.7 6.4 9.6 6.6 9.0 cm 14 16.8 9.6 9.9 5.5 9.5 16 33.4 8.9 9.8 5.2 9.3 Mean 25.0 7.7 10.4 5.4 9.2 s.d. 8.3 1.9 0.2 0.7 0.2 5 24.1 11.5 10.6 11.2 11.2 35 8 35.1 6.3 ns ns ns cm 10 113.0 6.2 9.2 5.2 8.6 13 24.4 11.5 13.7 8.7 10.9 20 47.7 8.6 9.1 7.3 9.1 Mean 48.8 8.8 10.6 8.1 10.0 s.d. 37.1 2.6 2.2 2.5 1.3 1 28.1 9.1 9.4 7.8 9.0 55 6 48.6 11.5 13.0 4.0 10.4 cm 11 13.4 6.0 8.2 7.0 7.4 17 24.5 8.9 9.5 4.8 8.7 19 31.6 8.9 9.3 5.9 9.2 Mean 29.2 8.9 9.8 5.9 8.9 s.d. 12.8 1.9 1.8 1.6 1.1 2 30.4 13.5 ns ns ns 75 4 35.6 10.0 10.6 10.5 10.8 cm 15 19.3 9.1 ns ns ns 18 31.4 8.8 9.5 5.2 9.2 21 15.9 5.7 ns ns ns Mean 26.5 9.4 10.0 7.9 10.0 s.d. 8.5 2.8 Notes: s.d. - standard deviation ns - no sample Concentration in mg/L 79 In Run 2, chloride concentrations generally increased in the 20 minute leachate compared with the preflush; 20 minute concentrations were also slightly higher than the 24 hour leachate concentrations. The 20 minute leachate is expected to be composed of water held under low tension - drainage from the sand filter and the largest pores. A higher concentration in this leachate compared to the 24 hour leachate indicates the occurrence of preferential flow. There is no evidence of an effect of water table height on CI" leaching in any sample time of Run 2. Nutrient solution applied in Run 2 had a mean concentration of 1857 mg/L CI", with a minimum potential dilution in the applied precipitation of 357 mg/L CI". When the water held under the pre-set water table is accounted for (by dividing the mass applied by the total volume drained), the potential dilutions range from 231 - 33 mg Cl"/L, with an average potential dilution of 91 mg CI7L. The leachate concentrations are at least 4 times below the potential dilutions indicating that most of the applied chloride was retained within the columns. 3. Nitrate and Nitrite Nitrogen As Table VII illustrates, the NO3 + N 0 2 concentration in the leachate from Run 0 (no nutrient addition) was very low, with most (12 of 18) columns having concentrations of NO3 + NC>2 less than 0.19 ppm, the limit of detection. From the data, it appears that, except for columns 18 and possibly 13, the columns are all very similar with respect to initial status of nitrate. In Run 1, considerable difficulty was experienced in maintaining water tables close to desired levels because of poor experimental design and because almost all of the columns were leaking by the end of the run. Leaking occurred from the bottoms and through the half-way samplers, the latter potentially allowing nutrient to leak out before it had passed through all of the soil mass. Thus, the results of analyses on the leachate are not useful for determining a treatment effect, but do seem to indicate that, because of the low N concentrations, applied N is not simply leaching through the columns unaltered. 80 Table VII. Runs 0-2 nitrate + nitrite concentration Water Column Run 0 Run 1 Pre- R u n 2 Table # flush 20 Min 24 Hour Total 7 ns 0.18 0.91 *0.14 0.14 *0.10 15 9 *0.19 0.18 2.34 0.69 0.94 0.78 cm 14 *0.19 0.18 0.31 0.14 *0.14 *0.10 16 *0.19 0.17 0.71 0.16 *0.14 *0.09 Mean 0.19 0.18 1.07 0.28 0.34 0.27 s .d. — 0.01 0.88 0.27 0.40 0.34 5 0.34 0.18 2.12 0.37 0.42 0.39 35 8 *0.19 0.18 *0.14 ns ns ns cm 10 *0.19 0.18 *0.14 *0.14 *0.14 *0.14 13 0.76 0.24 2.06 0.71 0.69 0.70 20 *0.19 0.17 *0.14 *0.14 *0.14 *0.14 Mean 0.33 0.19 0.92 0.34 0.35 0.34 s .d. 0.25 0.03 1.07 0.27 0.26 0.27 1 *0.19 *0.16 0.14 *0.14 *0.14 *0.14 55 6 0.19 0.18 1.45 0.31 0.31 0.31 cm 11 *0.19 *0.16 *0.14 *0.14 *0.14 *0.14 17 *0.19 *0.16 1.71 0.14 0.31 0.24 19 0.28 0.18 2.97 2.71 2.15 2.54 Mean 0.21 0.17 1.28 0.69 0.61 0.67 s .d; 0.04 0.01 1.19 1.13 0.87 1.05 2 *0.19 *0.16 *0.14 ns ns ns 75 4 *0.19 0.18 0.36 *0.14 0.31 *0.16 cm 15 *0.19 0.18 0.14 ns ns ns 18 2.68 0.58 9.82 5.95 8.25 6.57 21 0.22 *0.17 0.14 ns ns ns Mean 0.69 0.25 2.12 3.05 4.28 3.37 s .d. 1.11 0.18 4.31 — — — Notes: Concentrations in mg/L s.d. - standard deviation ns - no sample * - at least one analytical replicate was at the limit of detection; actual value is less than or equal to the value shown. Minimum detectable concentrations were used to calculate means and standard deviations. 81 Reduction in N concentration can be considered in the same manner that phosphate concentrations were considered. For Run 1, each column received 25 mL nutrient stock at a concentration of 2560 mg/L, followed by a total of 785 mL precipitation, for a "potential di lution" to 81.5 mg/L. Measured leachate concentrations are 2 orders of magnitude below this potential dilution, and so indicate considerable physical retention or chemical or microbial conversion of NO^ + N 0 2 in the columns. The N concentrations in the preflush to Run 2 were much more variable than in the initial leachings, reflecting the retention of added N. The preflush procedure should have "picked up" some of the nitrate held in the upper parts of the column, and nitrate held in dead end pores and within the soil peds as the entire column became saturated, and should have promoted mixing of this " immobi le" water with water leaching out through larger pores, thus removing some of the solute held within the matrix. There are quite distinct differences in the leachate concentrations of the columns in the preflush, with columns 9, 5, 13, 6, 17, 19, and 18 having much higher concentrations than the others. With only one exception (column 17), these are the columns with the highest hydraulic conductivities. If the higher hydraulic conductivities result from larger drainage channels or greater continuity of soil pores, then solution in the upper regions of the columns (ie. where the nitrate is likely located) had a better opportunity to drain directly out of those columns with higher hydraulic conductivities. Where lower conductivities exist, the draining solution is subject to more dispersion. This point is elaborated further in the discussion on the effect of hydraulic conductivity on leachate quality in a subsequent section. In Run 2 it appears that the columns with higher water tables also had higher leachate N concentrations in both the 20 minute and 24 hour leachate samples. However, the standard deviations are large, indicating quite variable responses within treatments. As already mentioned, the poor design and the significant amount of leaking make any conclusions uncertain. Spearman's correlation coefficients for the 20 minute, 24 hour and run total concentrations against the measured initial water tables were 0.14, 0.34, and 0.25, 82 respectively, and against the measured final water tables were -0.26, +0.03, and -0.16, respectively. All coefficients were non-significant. Thus, there was no effect of water table on leachate N concentration. It appears that the trends exhibited by the treatment means are misleading, and result from a few columns in the two highest water table treatments having very high leachate concentrations. 4. Ammonium Nitrogen The Nh^" 1 " concentration in the leachate was usually less than 1 mg-N/L for all Runs in the experiment. Because the concentrations were so low, and were variable without apparent pattern, N H 4 + is not discussed further; data is presented in Table VIII. D. Runs 3 and 4 Leachate Concentrations 1. Distribution of the Data The concentration of C l " and NO3 + NO2 and the N/CI ratios for all sample times for Runs 3 and 4 are presented in Tables IX - XI. Scatterplots of chloride and nitrate concentration and the N/Cl ratios for Run 4 Final drainage are shown in Figures 8 - 1 0 . Only one sample event is shown as all plots are very similiar. The most obvious feature of the tables and scatter plots is that for each sample time, for each constituent of interest, there is almost no apparent difference in means (for each sample time the range within each treatment group is wide, and the 95% confidence intervals for the means are generally overlapping), a. Chloride For the chloride concentrations, normality of distribution was accepted for all sample times and homogeneity of variance was accepted for all sample times except R3TOT and GT. Independence of means and variances was accepted as there was no obvious relation between the two. Therefore the raw chloride concentrations were used for all statistical analyses. T a b l e V I I I . R u n s 0 - 4 L e a c h a t e A m m o n i a C o n c e n t r a t i o n s C o l R u n 0 R u n 1 P r e - R u n 2 P r e R u n 3 R u n 4 W a t e r T a b l e i f l u s h 2 0 m i n 2 4 h o u r f l u s h O F H W F D O F H W 1 H W 2 F D 7 2 5 5 2 4 8 3 0 9 3 4 9 1 3 8 1 7 6 3 1 4 5 1 5 9 4 0 3 3 6 3 2 9 5 3 4 9 3 2 9 2 9 1 8 1 0 4 3 9 3 0 3 2 c m 1 4 4 9 1 2 7 5 3 0 9 2 7 2 2 5 5 2 3 1 8 4 1 3 2 3 1 1 6 4 8 4 3 3 6 3 6 9 5 1 1 3 8 3 7 3 1 7 3 1 4 0 3 5 M e a n 3 4 4 3 0 7 3 0 5 3 6 0 3 2 9 3 4 1 5 - - 7 2 3 3 - - - - 3 6 5 1 0 3 0 2 4 8 5 5 8 7 8 1 4 0 3 1 8 5 1 1 3 7 5 9 2 2 6 3 5 8 3 6 3 9 8 3 2 2 1 3 8 3 2 5 5 3 3 3 5 1 0 9 1 8 3 4 5 0 8 6 , 3 0 c m 1 0 2 0 7 2 5 5 1 8 7 2 1 1 2 3 4 2 3 5 2 8 3 0 3 3 1 3 1 2 7 3 6 2 6 7 6 7 1 4 0 5 8 5 5 4 5 2 3 3 3 4 2 3 0 2 0 2 6 1 2 2 8 2 0 7 2 4 1 2 0 7 7 1 4 3 5 1 2 6 M e a n 6 2 7 4 6 8 3 8 8 6 0 4 3 9 1 5 9 2 4 - 2 2 4 0 - - - - 8 9 1 3 2 2 2 5 5 2 5 5 2 9 5 3 1 2 4 3 3 6 5 1 2 1 3 5 3 0 3 2 5 5 6 3 0 2 2 5 5 2 6 1 3 7 6 2 9 2 1 9 1 1 8 6 2 2 3 4 1 2 0 3 6 3 0 c m 1 1 2 8 8 2 4 8 2 4 1 2 8 2 2 5 5 7 1 1 3 0 3 5 3 7 1 4 4 3 0 1 7 3 2 9 2 4 1 2 4 1 2 2 8 4 6 1 2 5 1 1 7 8 5 1 6 8 2 7 7 2 1 1 3 1 0 8 1 9 7 6 0 3 9 6 4 4 3 5 6 5 5 1 8 8 3 9 2 3 9 1 0 0 1 1 8 6 5 M e a n 4 0 0 2 3 1 2 8 8 3 5 2 3 2 1 3 9 3 7 7 5 4 0 5 5 6 3 3 5 3 5 3 2 3 3 6 3 1 2 2 0 7 4 5 0 4 2 7 3 5 6 5 1 9 2 2 3 1 0 2 8 2 1 7 1 7 5 4 2 7 5 2 8 8 3 2 2 3 6 3 2 8 8 1 1 1 9 1 4 6 1.1 3 0 1 1 8 4 5 3 0 c m 1 5 3 5 6 2 6 1 2 2 8 5 1 1 3 0 2 0 7 6 6 5 . 4 0 6 8 5 8 3 5 1 8 1 9 8 4 3 6 3 4 4 3 6 4 6 4 0 6 2 1 3 2 2 1 2 3 7 9 5 5 7 1 2 1 3 6 3 3 7 6 1 7 4 7 2 9 3 8 3 0 1 0 1 8 5 3 0 M e a n 6 6 3 3 2 0 2 7 5 3 9 4 2 2 4 5 9 2 1 1 8 6 2 6 2 0 9 3 6 5 6 7 0 0 . OJ N o t e s : B l a n k s i n d i c a t e n o s a m p l e C o n c e n t r a t i o n s a r e u g / L N H 3 - N 84 T a b l e I X . R u n s 3 a n d 4 c h l o r i d e c o n c e n t r a t i o n W a t e r C o l u m n P r e - R u n 3 R u n 4 G r a n d T a b l e # f l u s h O F H W F D T O T O F H W 1 H W 2 F D T O T T o t a l 7 7 . 1 9 . 1 n s 13 .6 9 . 7 13. 6 n s n s 17 .0 14 .0 11 . 7 15 9 6. 2 7 .3 25.2 8 . 1 7 .5 9. 1 n s n s ' 1 1 .5 9. .5 8. .5 c m 14 7 . ,5 7 .7 n s 9 .4 7 .9 9. 9 n s n s 11 .5 10. .2 9. . 1 16 6 .5 6 . 3 n s 7 .8 6. .5 8. 2 n s n s 9 9 8. 5 7 5 M e a n 6. 8 7 6 9 . 7 7 9 10. 2 12 . 5 10. 6 9. 2 LCL 5 9 5. . 7 5. .4 5. 8 6. 5 7 .5 6. 7 6. 3 UCL 7 . a 9 4 14. .0 10. 0 13. 9 17 . 4 14. .4 12. 1 5 7 . 5 9 . . 0 n s 1 0 . 2 9 5 1 0 1 n s n s 11 . . 1 1 0 . . 5 9 . 9 3 5 8 4 . 0 6 . 2 1 5 . 2 8 . 7 8 . . 3 1 2 . 3 2 1 . 3 2 3 . 1 1 4 . 3 1 3 . . 7 1 1 . 0 c m 1 0 4 . 3 6 . 1 n s 8 . . 1 8 . , 0 1 0 . 8 n s n s 1 2 . 8 1 1 . 9 9 . 9 1 3 8 . 6 9 . 3 n s 9 . 9 9 . . 4 1 0 . 1 n s n s 11 . . 3 1 0 . . 4 9 . 9 2 0 5 . 5 6 . 9 n s 9 . 5 8 . . 6 11 2 n s n s 1 4 . . 0 1 2 . . 1 1 0 . 3 M e a n 6.0 7.5 - - 9.3 8.7 10.9 - - - - 12.7 11.7 • 10.2 L C L 3.5 5.5 - - 8.2 7.9 9.7 - - - - 10.9 10.0 9.6 U C L 8.5 9.4 - - 10.4 9.6 12.1 - - - - 14.5 13.4 10.8 1 3 .2 4.8 30 .2 6 .0 6 . 3 6.8 28 .9 19. 1 7 8 9 2 7 . 5 55 6 7 , 7 8.4 23 .4 11 . 7 10. . 1 12.7 32 .7 32.0 15. .7 14. 8 12. 4 c m 11 4. . 7 5.6 18 .8 7 .8 7 . ,0 9.0 17 . 3 n s 11 . 3 9 8 8 .4 17 7 4 7 . 3 18 .6 9 .0 8 6 9.2 31 .0 31.0 11 8 11 . 2 9 . 9 19 9 ,0 n s 11 .0 9 .2 9. .2 n s 13 .4 13.4 10 8 10. .9 10. 1 M e a n 6.4 6.5 20.4 8.7 8.2 9.4 24.7 23.9 11.5 11.2 9.7 L C L 3.5 3.9 11.6 6.2 6.3 5.5 13.9 9.3 8.0 8.5 7.3 U C L 9.4 9.2 29.2 ' 11.3 10.2 13.3 35.5 38.4 15.0 13.9 12.0 2 9 . 2 8.5 15 .5 11 .6 1 1 . 1 n s 19 .4 26.8 12 . 3 12 8 12 .0 75 4 7 .9 7.8 25 . 3 10 . 5 10 . 2 11.2 27 8 22.2 14. .8 14 . 4 12 . 4 c m 15 8 . 2 10. 7 14 . 4 14 . 5 13. .0 13.8 22 6 n s 16 . 1 15. 2 14 . 1 18 6 .6 n s 14 2 a .2 8 . 4 n s 13. .0 13.9 9 .2 9 4 8 .9 21 1 .6 2.6 14. .6 4 .8 4 . 5 5.8 18. . 7 24.2 7 . 8 ' 8 .2 6. . 3 M e a n 6.7 7.4 16.8 9.9 9.4 10.3 20.3 21.8 12.0 12.0 10.7 L C L 3.0 2.0 10.9 5.4 5.4 0.1 13.6 12.9 7.6 8.2 6.9 U C L 10.4 12.8 22.8 14.4 13.4 20.4 27.0 30.6 16.4 15.9 14.6 N o t e s : C o n c e n t r a t i o n s a r e m g - C l / L n s - n o s a m p l e L C L . U C L - l o w e r a n d u p p e r 9 5 % c o n f i d e n c e i n t e r v a l f o r t h e m e a n . T a b l e X . R u n s 3 a n d 4 n i t r a t e p l u s n i t r i t e c o n c e n t r a t i o n W a t e r C o l u m n P r e - R U N 3 R U N 4 G r a n d T a b l e » f l u s h O F HW F D T O T O F H W 1 H W 2 F D T O T T o t a l 7 1 . 0 9 0 8 8 n s 2 . 8 1 1 . . 1 2 2 . 7 4 n s n s 4 . 9 6 3 . 0 3 2 . 0 2 1 5 9 1 1 . 5 1 1 3 . 9 1 0 . 1 8 1 4 . 2 5 1 3 9 2 1 4 4 1 n s n s 1 6 . 0 7 1 4 . 6 8 1 4 3 1 c m 1 4 1 . 3 9 0 . . 1 1 n s 0 . 1 3 0 . . 1 2 1 .  1 1 n s n s 0 . 3 6 0 . 9 7 0 . 5 5 1 6 1 . 7 4 2 . . 2 8 n s 2 . 1 5 2 . . 1 6 3 . 8 7 n s n s 4 . . 6 6 4 . 0 0 3 . 0 6 i l e a n 3 . 9 3 4 . . 3 0 4 . 8 4 4 . . 3 3 5 . . 5 3 ' 6 . . 5 1 5 . . 6 7 4 . 9 9 L C L - 4 . 1 1 - 6 . . 0 0 - 5 . 3 1 - 5 . . 9 3 - 4 . 0 5 - 4 1 6 - 4 . 1 0 - 5 . 0 4 U C L 1 1 . 9 8 1 4 . . 5 9 1 4 . 9 8 1 4 . . 5 9 1 5 . . 1 2 1 7 . 1 9 1 5 . 4 3 1 5 . 0 1 5 9 . 0 8 9 . 9 5 n s 1 1 . . 6 4 1 0 . 5 5 11 . . 5 7 n s n s 1 2 . 2 4 1 1 . 8 3 11 . . 1 5 3 5 8 0 0 6 0 . 0 5 0 . 0 4 0 . 0 3 0 . 0 3 0 . 0 4 0 . 0 6 0 . 0 6 0 . 0 3 0 . 0 4 0 . 0 4 c m 1 0 0 . 0 2 0 . 0 1 1 . 9 5 0 , 0 1 0 . 0 5 0 0 3 n s n s 0 . 0 3 0 . 0 3 0 . 0 4 1 3 1 3 . . 2 9 1 9 . 8 2 n s 1 5 . . 2 3 1 8 . 8 4 1 5 . 3 4 n s n s 1 5 . 5 4 1 5 . 4 0 1 7 . 1 2 2 0 0 . 0 1 0 . 0 1 n s 0 . 6 2 0 . 4 0 0 . 3 4 8 . 1 7 n s 2 . 7 0 1 . 2 5 0 . 8 1 M e a n 4 . 4 9 5 . 9 7 0 . 9 9 5 . 5 0 5 . 9 7 5 . 4 6 4 . 1 1 - - 6 . 1 1 5 . 7 1 5 . 8 3 L C L - 3 . 3 2 - 5 . 0 2 - - - 3 . 6 2 - 4 . 5 6 - 3 . 7 4 - - - - - 2 . 9 3 - 3 . 4 1 - 3 . 9 5 U C L 1 2 . 3 0 1 6 . 9 6 - - 1 4 . 6 3 1 6 . 5 1 1 4 . 6 7 - - 1 5 . 1 5 1 4 . 8 3 1 5 . 6 1 1 0 . 0 2 0 . . 0 4 0 . 0 4 0 . . 0 4 0 . 0 4 0 . 0 3 2 3 . 5 9 n s 0 . 0 4 1 . 7 4 0 . 7 2 5 5 6 4 . 9 1 4 . . 7 9 n s 7 . 7 7 5 . 8 6 5 9 6 4 3 . 4 2 4 3 . 4 2 11 . 2 8 9 8 9 7 . 8 8 c m 11 0 . . 3 5 0 . . 2 8 1 3 . 0 3 1 6 7 1 . 1 5 1 . 0 2 0 0 6 2 4 . 11 2 . 7 4 1 . . 7 4 1 . 4 5 1 7 2 8 4 1 . . 6 5 n s 2 . 7 2 2 . 2 1 2 . 4 1 3 9 . 1 3 0 . 0 4 4 . 8 2 4 . 6 2 3 . 4 6 1 9 1 1 . . 0 0 n s 2 2 3 9 1 0 . 0 2 1 0 . 3 6 n s 2 3 2 0 2 2 . 7 2 1 1 . 2 8 1 1 . 9 2 11 . 1 4 M e a n 3 . 8 2 1 . 6 9 1 1 . 8 2 4 . 4 4 3 . 9 2 2 . 3 5 2 5 . 8 8 2 2 . 5 7 6 . 0 3 5 . 9 8 4 . 9 3 L C L - 1 . 7 3 - 1 . 7 9 - 1 6 . 1 6 - 0 . 8 3 - 1 . 3 0 - 1 . 7 7 4 . 7 1 - 5 . 6 6 - 0 . 2 8 0 . 1 4 - 0 . 6 0 U C L 9 . 3 8 5 . 1 6 3 9 . 7 0 9 . 7 2 9 . 1 5 6 . 4 8 4 7 . 0 4 5 0 . 8 1 1 2 . 3 4 1 1 . 8 2 1 0 . 4 6 2 2 . 8 9 1 . 3 9 1 4 2 1 6 . 2 0 5 . 4 2 n s 1 9 . 6 6 1 9 . 1 3 4 . 9 9 5 . 7 8 5 . 6 1 7 5 4 4 . 9 3 3 . 0 2 2 4 . . 3 1 6 . 3 7 5 . 9 3 2 . 6 8 3 7 . 5 6 3 7 . 6 1 1 0 . 2 8 9 . 8 2 7 . 9 0 c m 1 5 0 . 7 0 0 . 4 5 3 . 2 3 2 . 8 9 1 9 7 1 . 3 2 4 . 9 2 6 . 4 2 3 . . 2 2 2 . 9 4 2 . . 4 4 1 8 1 2 . 6 8 n s 3 3 . 1 3 1 5 . 7 2 1 6 . . 2 6 n s 3 3 . 9 7 3 6 . 0 8 1 0 . . 5 6 1 1 . 9 0 1 3 9 2 2 1 0 . 0 1 0 . 0 1 7 . 7 8 0 . 0 3 0 . . 2 2 0 . 0 3 6 . 5 0 1 7 . 1 2 0 . . 0 4 0 6 9 0 . . 4 5 M e a n 4 . 2 4 1 . 2 2 1 6 . 5 3 6 . 2 4 5 . 9 6 1 . 3 4 2 0 . 5 2 2 3 . 2 7 5 . 8 2 6 . 2 3 6 . 0 6 L C L - 2 . 0 9 - 0 . 9 0 1 . 3 9 - 1 . 0 9 - 1 . 7 8 - 1 . 9 5 1 . 7 7 6 . 7 5 0 . 1 6 0 . 4 4 - 0 . 4 5 U C L 1 0 . 5 7 3 . 3 3 3 1 . 6 7 1 3 . 5 8 1 3 . 7 0 4 . 6 3 3 9 . 2 7 3 9 . 8 0 1 1 . 4 8 1 2 . 0 1 1 ? 5 7 Notes: C o n c e n t r a t i o n s a r e m g ( N 0 3 + N 0 2 ) - N / L n s - n o s a m p l e L C L . U C L - L o w e r a n d u p p e r 9 5 % c o n f i d e n c e i n t e r v a l s f o r t h e m e a n . 86 20 -i CD J , C o c < D O c o o <D 15-10-15 I 35 55 l 75 Water Table Height (cm) F igure 8. R e l a t i o n s h i p between Run 4 F i n a l Drainage C I " concen t ra t i on and water t a b l e h e i g h t u means, +/- 1 standard d e v i a t i o n 87 CO E, c o c O c o o <D 1 5-i 10 5 -0 -- 5 ' 15 3 5 I 5 5 7 5 Water Table Height (cm) Figure 9. Relationship between Run 4 Final Drainage NO3" concentration and water table height. means, +/- 1 standard deviation 88 1.5-1 1-0.5-0--0.5- 15 35 55 75 Water Table Height (cm) F i g u r e 10. R e l a t i o n s h i p b e t w e e n Run 4 F i n a l D r a i n a g e N/Cl r a t i o s and w a t e r t a b l e h e i g h t o O means, +/- 1 s t a n d a r d d e v i a t i o n 89 b. Nitrogen For NO3 + N 0 2 concentrations the decisions were more complicated. Normality was rejected at the 0.05 level for R3PRE and R3FD. Error effects did seem to be distributed around zero for all sample times. Homogeneity of variance was rejected for R30F. From examination of means and variances it appears that for most sample times variance increases with increasing mean, indicating that a transformation of the data may be needed. Unfortunately, there is no definitive method to determine which transformation is appropriate, although many suggestions exist in the literature (Little & Hills, 1978; Kirk, 1982; Hicks, 1982; LeClerg et al, 1962; Jeffers, 1959). From comparisons of mean to variance, mean to standard deviation and mean-squared to standard deviation, it was unclear which, if any, transformation was most appropriate. Kirk (1982) suggests computing the transformations and then choosing the transformation which produces the smallest ratio of maximum treatment range to minimum treatment range. The range ratios for log, reciprocal, and square root transformations and for the untransformed data were compared, and no transformation appeared to be consistently appropriate. Similarly, there was no consistent and marked improvement in tests for normality or homogeneity of variance for the transformed data. Therefore the NO3 + NO2 concentration data were used untransformed, accepting normality and homogeneity of variance, but recognizing that treatment means and variances may not, in all cases, be independent, and therefore may limit the validity of significance testing in parametric ANOVAs. Nitrite was measured as a separate component of the leachate in Run 4. O n average, NO2" comprised about 5% of total NO3 + N02- The low NO2" concentration was expected because NO2" is not usually present in high concentrations in soil (Schmidt, 1982), and it is not acid-stable (Nelson, 1982; van Cleemput & Baert, 1984). Nitrite alone is not discussed further. 90 T a b l e X I . R u n s 3 & 4 N / C l r a t i o s W a t e r C o l u m n R u n 3 R u n 4 G R A N D T a b l e # O F H W F D T O T A L O F H W 1 H W 2 F D T O T A L T O T A L 7 0 . 1 0 n s 0 . , 2 1 0 . 1 1 0 . 2 0 n s n s 0 . 2 9 0 . 2 2 0 . , 1 7 1 5 9 1 . 9 0 0 . 0 1 1 . . 7 6 1 . 8 7 1 . . 5 8 n s n s 1 . . 4 0 1 . 5 4 1 , 6 8 cm 1 4 0 . 0 1 n s 0 , . 0 1 0 . 0 1 0 . 1 1 n s n s 0 . 0 3 0 . 1 0 0 . 0 6 1 6 0 . 3 6 n s 0 . . 2 8 0 . 3 3 0 . 4 7 n s n s 0 . 4 7 0 . 4 7 0 . 4 1 i l e a n 0 . 5 9 0 . . 5 7 0 . 5 8 0 . 5 9 _ _ 0 . 5 5 0 . 5 8 0 . 5 8 L C L - 0 . 8 1 - 0 . 7 2 - 0 . 8 0 - 0 . 4 8 - 0 . 4 0 - 0 . 4 7 - 0 , . 6 1 U C L 2 . 0 0 1. 8 5 1 9 6 1. . 6 7 1, . 3 5 1. . 6 3 1, . 7 7 5 1 . 1 0 n s 1 . 1 4 1 . 1 1 1 . 1 5 n s n s 1 . , 1 0 1 . 1 3 1 . 1 2 3 5 8 0 . 0 1 0 . 0 0 0 . 0 0 0 . 0 0 0 . 0 0 0 . 0 0 0 . 0 0 0 , . 0 0 0 . 0 0 0 . 0 0 c m 1 0 0 . 0 0 n s 0 . 0 0 0 . . 0 1 0 . 0 0 n s n s 0 . 0 0 0 . 0 0 0 . 0 0 1 3 2 . 1 4 n s 1 . . 5 4 2 . 0 0 1 . 5 2 n s n s 1 , . 3 8 1. . 4 8 1 . 7 3 2 0 0 . 0 0 n s 0 . . 0 7 0 . 0 5 0 . 0 3 n s n s 0 . 1 9 0 . . 1 0 0 . 0 8 M e a n 0 . 6 5 - - 0 . 5 5 0 . 6 3 0 . 5 4 - - - - 0 . 5 3 0 . 5 4 0 . 5 9 L C L - 0 . 5 4 - - - 0 . 3 6 - 0 . 4 8 - 0 . 3 7 - - - - - 0 . 2 8 - 0 . 3 3 - 0 . 4 0 U C L 1 . 8 4 - - 1 . 4 6 1 . 7 5 1 . 4 6 - - - - 1 . 3 5 1 . 4 2 1 . 5 8 1 0 . 0 1 0 . 0 0 0 . 0 1 0 . 0 1 0 . 0 0 0 . 8 1 n s 0 . 0 0 0 . 1 9 0 . 1 0 5 5 6 0 . 5 7 n s 0 . 6 7 0 . 5 7 0 . 4 7 1 . 3 3 1 . 3 6 0 7 2 0 . 6 7 0 . . 6 3 cm 1 1 0 . 0 5 0 . 6 9 0 . 2 2 0 . 1 6 0 . 1 1 0 . 0 0 n s 0 . 2 4 0 . 1 8 0 . 1 7 1 7 0 . 2 3 n s 0 . 3 0 0 , . 2 5 0 . 2 6 1 . 2 6 0 . 0 0 0 . 4 1 0 . 4 1 0 . 3 4 1 9 n s 2 . 0 4 1 . 0 9 1 , . 1 3 n s 1 . 7 3 1 . 7 0 1. . 0 4 1 . . 0 9 1 . 1 1 M e a n 0 . 2 2 0 . 9 1 0 . 4 6 0 . 4 2 0 . 2 1 1 . 0 3 1 . 0 2 0 . 4 8 0 . 5 1 0 . 4 7 L C L - 0 . 1 9 - 1 . 6 6 - 0 . 0 8 - 0 . 1 3 - 0 . 1 1 0 . 2 1 - 1 . 2 1 - 0 . 0 2 0 . 0 3 - 0 . 0 4 U C L 0 . 6 2 3 . 4 8 0 . 9 9 0 . 9 7 0 . 5 3 1 . 8 4 3 . 2 5 0 . 9 9 0 . 9 8 0 . 9 8 2 0 . 1 6 0 . 9 2 0 . 5 3 0 4 9 n s 1 . 0 1 0 . 7 1 0 . 4 1 0 . . 4 5 0 . 4 7 7 5 - 4 0 . 3 9 0 . 9 6 0 . 6 1 0 . 5 8 0 . 2 4 1 . 3 5 1 . 7 0 0 . 6 9 0 . 6 8 0 , 6 4 cm 1 5 0 . 0 4 0 . 2 2 0 . 2 0 0 . 1 5 0 . 1 0 0 . . 2 2 n s 0 . 2 0 0 , . 1 9 0 , 1 7 1 8 n s 2 . 3 3 1 . 9 3 1 . 9 5 n s 2 . 6 1 2 . 5 9 1 . 1 5 . 1 . 2 7 1 . . 5 6 2 1 0 . 0 0 0 . 5 3 0 , . 0 1 0 . 0 5 0 . 0 1 0 . 3 5 0 . 7 1 0 . 0 1 0 . 0 8 0 . . 0 7 M e a n 0 . 1 5 0 . 9 9 0 . 6 6 0 . 6 4 0 . 1 2 1 . 1 1 1 . 4 3 0 . 4 9 0 . 5 3 0 . 5 8 L C L 0 1 2 - 0 . 0 1 - 0 . 2 8 - 0 . 3 0 - 0 . 1 8 - 0 . 0 9 - 0 . 0 1 - 0 . 0 7 - 0 . 0 5 - 0 . 1 5 U C L 0 4 2 2 . 0 0 1 . 5 9 1 . 5 9 0 . 4 1 2 . 3 0 2 . 8 7 1 . 0 5 1 . 1 2 1 . 3 2 N o t e s : R a t i o s a r e ( N 0 3 + N 0 2 1 - N / C l L C L , U C L - L o w e r a n d u p p e r 9 5 % c o n f i d e n c e i n t e r v a l s f o r t h e m e a n , n s - N o s a m p l e 91 c. N/CI Ratios The N/CI ratios for the two runs are shown in Table XI. The distribution of this data was also difficult to determine. Normality was rejected ate* = 0.05 for R3FD, R3TOT, R4TOT, and GT, and homogeneity of variance was rejected for only R40F. As there was no clear relationship between means and variance, and because deviations from normality are not considered to seriously impair the F test (Hicks, 1982; Kirk, 1982), the N/CI ratios were used untransformed. 2. Recovery of Solute The mass of CI" leached generally increased from Run 3 to Run 4 (as did the concentration of CI" in the leachate). It is somewhat difficult to calculate a percent mass recovery for Runs 3 and 4 because mass recoveries are low overall and therefore not complete from previous runs. The recoveries in Table XII were calculated on the basis of mass applied at the start of Run 3. The total recovery of CI* was less than 60% in all columns, averaging about 40%. The actual recovery of the CI" applied in Run 3 is likely less than this amount, because of incomplete recovery from applications in earlier runs and because of the unmeasured and likely variable leaching of native chloride. As the chloride recovery is not 100%, the nitrate recoveries must be interpreted with caution, and must be corrected for solute retained within the columns. The percent recoveries of (NO3 + N C ^ - N in Table XII indicate the approximate proportion of applied nitrate recovered in the leachate. The N recoveries are much lower than the CI" recoveries for each column (thus the N/Cl ratios are also less than the initial ratio), the difference being due to denitrification, N uptake, and variability in native content of NO3" and CI". If the percent of CI" leached is considered as the maximum potential percent of NO3" leached, a very rough mass balance for NO3" can be calculated as: i) % retained within the column: equivalent to % CI" not leached; ii) % leached: as calculated from N recovery in leachate; iii) % "transformed" (denitrification and uptake): 100% - % leached - % retained. 92 Table XII. S o l u t e r e c o v e r i e s Run 3 Run 4 Grand Run 3 Run 4 Grand Run 3 Run 4 Grand 10 To ta l To ta l To ta l To ta l To ta l T o t a l T o t a l T o t a l T o t a l 7 21.2 26.7 47.9 1.48 3.58 5.06 0.07 0.13 0.11 9 15.3 20.6 35.9 17.71 19.78 37.50 .1.16 0.96 1.04 14 15.1 20.1 35.2 0.14 1.19 1.32 0.01 0.06 0.04 16 14.1 17.7 31.8 2.89 5.18 8.08 0.21 0.29 0.25 Mean 16.4 21.3 37.7 5.56 7.43 12.99 0.36 0.36 0.36 5 20.8 20.4 41.1 14.39 14.29 28.68 0.69 0.70 0.70 8 16.6 27.5 44.0 0.04 0.05 0.09 0.00 0.00 0.00 10 14.6 21.5 36.1 0.06 0.04 0.10 0.00 0.00 0.00 13 19.8 21.9 41.7 24.71 20.12 44.83 1.25 0.92 1.07 20 19.1 25.2 44.2 0.54 1.61 2.16 0.03 0.06 0.05 Mean 18.2 23.3 41.4 7.95 7.22 15.17 0.39 0.34 0.36 1 7.5 7.5 15.0 0.03 0.87 0.90 0.00 0.12 0.06 6 21.0 30.0 51.0 7.37 12.44 19.81 0.35 0.41 0.39 11 14.1 20.1 34.2 1.44 2.23 3.67 0.10 0.11 0.11 17 17.6 23.8 41.3 2.71 6.11 8.82 0.15 0.26 0.21 19 18.8 22.2 41.0 13.13 15.02 28.15 0.70 0.68 0.69 Mean 15.8 20.7 36.5 4.93 7.33 12.27 0.26 0.32 0.29 2 22.0 27.9 49.9 6.69 7.78 14.47 0.30 0.28 0.29 4 19.9 28.8 48.7 7.18 12.19 19.37 0.36 0.42 0.40 15 27.2 30.2 57.3 2.55 3.62 6.17 0.09 0.12 0.11 18 15.6 20.4 36.0 18.82 16.06 34.88 1.21 0.79 0.97 21 10.2 17.0 27.2 0.31 0.90 1.20 0.03 0.05 0.04 Mean 19.0 24.9 43.8 7.11 8.11 15.22 0.40 0.33 0.36 Notes: C h l o r i d e and n i t r a t e r e c o v e r i e s are percent of mass a p p l i e d in Run 3. Ad jus ted N/Cl r a t i o i s (N/Cl ) / (No /C lo ) . 93 These rough mass balances for total quantities leached in both Runs 3 and 4 are shown in Figure 11, which illustrates that in most cases leaching was not the major N loss mechanism. This conclusion is also clearly supported by the low N/Cl ratios at all stages of sampling. Considering that the physical characteristics of the columns largely controlled the retention of CI", it is useful to look at the leached nitrate as a fraction of that which could potentially leach. The adjusted recoveries, calculated by dividing the leachate N/CI ratio by N 0 / C l 0 (where N 0 /C I Q = 1.61), are also shown in Table XII. The adjusted ratios do not contain additional information over the unadjusted ratios, but are intuitively easier to understand as they are based on a scale of 1 rather than 1.6. From the adjusted recoveries it is very clear that in most columns only a fraction as much nitrate leached as chloride. The columns for which the adjusted N/Cl's are the highest, have generally shown higher nitrate concentrations in the leachate in previous sample times (eg. #18, 13, 9, 5), emphasizing again the influence of natural variability between the columns. 3. Trends with Time This experiment was not designed to generate breakthrough curves, however, it is still informative to examine the patterns in leachate concentration with time. In the following analyses, all of the columns are considered together, regardless of assigned water table height. For most columns, both CI" and NO3" concentrations increased progressively from the start of Run 3 through to the end of Run 4, as can be seen from the data in Tables IX - XI. The trends in leachate concentration between R30F, R3FD, R40F, and R4FD were tested with the Friedman 2-way analysis of variance for related samples (Siegel, 1956). For chloride, the test statistic (X r 2 ) was 45.08, and for nitrate X r 2 was 19.88; both are significant at o<= 0.001, indicating that at least one of the 4 sample times had a significantly different mean concentration than the other sample times for both nitrate and chloride. For the N/Cl ratios X r z was 5.23, which was not significant at<*= 0.10, but was significant at tx = 0.20. In other words, while the nitrate and chloride concentrations alone changed significantly with sample time, the relative concentrations of the two did not change. 7 9 14 16 5 8 10 13 20 1 6 11 17 19 2 4 15 18 21 Column ID # Legend EZ3 % Converted • • % Leached E323 % Retained Water Table Height (cm) F i g u r e 1 1 . Mass b a l a n c e f o r n i t r a t e f o r Runs 3 & 4 Grand T o t a l 95 To test which of the sample times had concentrations that were significantly different from the others, Wilcoxon matched-pairs tests were performed (Siegel, 1956). Results of the Wilcoxon tests are summarized in Table XIII. From the first part of Table XIII it can be seen that the CI" concentrations increased significantly at each sample time (R30F < R3FD < R40F < R4FD). For nitrate, the concentrations increased significantly within a Run, but not between Runs (R30F < R3FD and R40F < R4FD, but R3FD <fc R40F). R30F and R40F are actually each composed of separate samples, which are presented in Tables XIV to XVI. The trends in leachate composition were investigated in more detail by analyzing the trends between the separate OF events. For CI", concentrations increased significantly within a run with successive OF samples and also from the last OF to FD events. These successively increasing concentrations are consistant with a displacement-type movement of water in the columns. The nitrate leaching patterns are not as clear as the CI" patterns. The OF concentrations increased progressively and significantly within Run 3 and also Run 4. Not only did the concentration of NO3 + N 0 2 increase as the OF progressed, but it increased more than the chloride concentration increased, as shown by the significant increase in N/Cl ratios from R30F1 to R30F2 and from R40F1 to R40F2. There was no consistent trend between the nitrate concentration in the FD and the various OF events. Although the first OF samples in Run 3 had significantly lower NO3 + N 0 2 concentrations than R3FD, the difference between R30F2 and R3FD was not significant. Similarly, R40F1 was significantly less than R4FD, but R40F2 and R40F3 were not. Also, the N/Cl for OF1 in both runs was significantly less than the FD N/Cl in each run, but the later OF events had N/Cl ratios that were not significantly different from the FD's. The N/Cl ratio also decreased significantly between R3FD and R40F1, as did the NO3 + N 0 2 concentration. The N/Cl ratios are key to understanding the trends in concentration. Except for columns 9 and 13, all N/Cl ratios are very much lower than the N/Cl of the nutrient solution T a b l e X I I I . W i l c o x o n m a t c h e d - p a i r s t e s t f o r t r e n d s w i t h time Sample Events # o f W i l c o x o n T P a i r s CI N0 3+N0 2 N/Cl R30F w i t h R3FD 17 0** 27** R3FD R40F 16 0** 49 — R40F R4FD 16 0** 10** — R30F1 w i t h R30F2 7 0** 0** 0* < R40F1 R40F2 9 0** ]_** o** < R40F2 R40F3 5 0* 0* 1.5 R30F1 wi t h R3FD 17 0** 18** 30* < R30F2 R3FD 7 0** 13 6 R3FD w i t h R40F1 16 37 +8** g** > R40F1 w i t h R4FD 16 0** 3** 10** < R40F2 R4FD 9 0** 10 22 R40F3 R4FD 5 0* 4 1 Notes: W i l c o x o n T v a l u e , c a l c u l a t e d a c c o r d i n g t o S i e g e l , 1956. For CI and NO3+NO2 a o n e - t a i l t e s t was used ( t e s t i n g the h y p o t h e s i s t h a t t h e l a t e r sample event has t h e h i g h e r c o n c e n t r a t i o n ) . For N/Cl r a t i o s a t w o - t a i l e d t e s t was used (no d i r e c t i o n of d i f f e r e n c e was p r e - s u p p o s e d ) . S i g n i f i c a n t l y d i f f e r e n t N/Cl r a t i o s a r e f o l l o w e d by t h e d i r e c t i o n o f d i f f e r e n c e f o r t h e p a i r s as the y appear i n the f i r s t column. *, ** - s i g n i f i c a n t a t 0.05 and 0.01, r e s p e c t i v e l y . + - r e s u l t i s s i g n i f i c a n t but i n the o p p o s i t e d i r e c t i o n t o t h a t s p e c i f i e d . Table XIV. Runs 3 Se 4 overflow chloride concentrations Water Column R u n 3 R u n 4 Table # 0F1 OF2 0F1 OF2 OF 3 OF 4 15 cm 7 9 14 16 8.6 6.7 7.3 5.9 9.3 7.8 7.9 6.5 12.0 8.3 8.8 7.8 13.1 9.4 10.1 8.3 14.3 9.7 10.3 8.7 — Mean 7.1 7.9 9.3 10.2 10.8 — 35 cm 5 8 10 13 20 9.0 6.2 6.1 9.0 6.9 9.1 9.4 10.0 11.8 10.8 9.3 11.2 10.2 12.8 10.0 10.4 10.3 Mean 7.4 9.3 10.6 11.0 — — 55 cm 1 6 11 17 19 4.8 7.7 5.6 7.3 9.2 6.8 11.9 8.7 9.2 13.3 9.2 — — Mean 6.4 — 9.1 11.3 — — 75 cm 2 4 15 18 21 8.5 7.8 10.7 — 11.2 13.8 — — — 2.6 — 5.8 — — — Mean 7.4 — 10.3 • — — — Notes: Concentrations are in mg-Cl/L Blanks indicate no sample Table XV. Runs 3 & 4 overflow nitrate plus nitrite concentration Water Column R u n 3 R u n 4 Table # 0F1 0F2 0F1 OF2 0F3 OF 4 7 0.65 0.97 0.95 1.99 3.75 __ 15 9 11.95 15.41 7.88 16.64 17.36 — cm 14 0.00 0.17 0.10 1.26 1.50 — 16 1.58 2.71 1.20 4.42 4.53 — Mean 3.54 4.81 2.53 6.08 6.78 — 5 8.64 11.32 11.47 12.19 __ 35 8 0.05 — 0.04 0.04 — — cm 10 0.01 — 0.03 — — — 13 18.27 20.71 6.06 14.30 19.44 18.80 20 0.01 0.34 Mean 5.40 16.01 3.59 8.84 1 0.04 0.03 55 6 3.92 5.71 5.32 6.45 — — cm 11 0.28 — 0.78 1.19 — — 17 1.65 — 2.41 — — — 19 Mean 1.47 — 2.14 3.82 2 1.39 75 4 3.02 — 2.91 cm 15 0.45 — 1.32 18 21 0.01 — 0.03 Mean 1.22 — 1.42 Notes: Concentrations are mg NO3-N/L Blanks indicate no sample Table XVI. Run 3 & 4 overflow N/Cl ratios Water Column R u n 3 R u n 4 Table # 0F1 0F2 OFl 0F2 OF 3 OF 4 15 cm 7 9 14 16 0.08 1.79 0.00 0.27 0.10 1.98 0.02 0.42 0.08 0.95 0.01 0.15 0.15 1.77 0.13 0.53 0.26 1.80 0.14 0.52 — Mean 0.54 0.63 0.30 0.65 0.68 — 35 cm 5 8 10 13 20 0.96 0.01 0.00 2.03 0.00 1.24 2.20 1.14 0.00 0.00 0.66 0.03 1.19 0.00 1.43 1.86 1.82 Mean 0.60 1.72 0.37 0.87 — — 55 cm 1 6 11 17 19 0.01 0.51 0.05 0.23 0.62 0.00 0.45 0.09 0.26 0.48 0.13 — — Mean 0.20 — 0.20 0.31 — 75 cm 2 4 15 18 21 0.16 0.39 0.04 — 0.26 0.10 — — — 0.00 — 0.01 — — — Mean 0.15 — 0.12 — — — Notes: Blanks indicate no sample 100 used (N Q /CI 0 = 1.61), indicating a loss of NC>3"-N as water leaches down the column. The increased N/CI ratio from OF1 to OF2, and OF1 to FD in both runs, indicates a decreased loss of NO3", rather than an input of NO3" from, for example, mineralization. If displacement was a significant mechanism of water movement, then the "decreased loss" of NO3" could result from the differing residence times of solute in the soil column. The early leachate samples in each run would be composed of solution that had been resident for the longest time, and therefore had the most opportunity to experience denitrification or immobilization. The early samples may also have been composed partially of dilute water from bypassing flow. If the small volume of nutrient solution was largely absorbed into soil peds at the surface, subsequent precipitation that became bypassing flow may not carry much of the added nutrient with it down into the column. If denitrification were occurring it would be expected to occur during the incubation prior to the FD. This may explain the (nonsignificant) trend to decreasing NO3 + N 0 2 concentrations (R40F3 > R4FD) and N/CI ratios (R30F2 > R3FD, R40F2 > R4FD, R40F3 > R4FD) between the later OF events and the FD events. The NO3 + N 0 2 concentration and N/CI ratio for R3FD were significantly higher than for R40F1, even though Cl " concentrations increased (but not significantly) between these two sample times. After the FD in Run 3, internal drainage and water redistribution should have continued, likely creating a saturated zone near the column bottom. Denitrification may have continued in this zone and in anaerobic microsites higher in the column. This lower NO3 + N 0 2 water would have been the first displaced after leaching started in Run 4. Also, there is the possibility of dilution from bypassing flow at the start of leaching (as already mentioned). The occurrence of bypassing flow in some columns is supported by the larger size of the Wilcoxon T for the R3FD with R40F1 C l " comparison than for any other C l " comparison. Although the C l " concentration increased in most columns from R3FD to R40F1, in almost half of the columns (7 of 16), the concentration actually decreased, and so the magnitude of the T value increased. A decrease in the concentration of an inert solute under 101 these conditions is not compatible with displacement as the sole mechanism of water movement, but is better explained by dilution from bypassing flow. Finally, the two sets of HW samples in Run 4 allow another comparison of changes in concentration with time (refer to Tables IX to XI). Unfortunately, only 11 columns produced a sufficient quantity of sample for analysis and in all but one of those the samplers were under the water table. The concentration changes between HW1 and HW2 were extremely variable: for both CI" and NO3 + N 0 2 the concentration increased in about as many columns as it decreased. In half of the columns for which an N/Cl ratio could be computed, the ratio did not change from HW1 to HW2, indicating an absence of denitrification or immobilization at that depth within those columns, over the duration of the incubation period. 4. Effect of Hydraulic Conductivity Because of the wide range in hydraulic conductivities of the columns, correlation analysis was performed on hydraulic conductivity and leachate concentration paired by soil column, regardless of treatment group. Spearman correlation coefficients were calculated for most sample times of Runs 3 and 4 for NO3 + N0 2, C F , and N/Cl (see Table XVII). From scatter plots of leachate concentration versus K s a t it appeared that, for all chemical constituents except CI", leachate concentration was related to K $ a t , and that the relationship was non-linear to poorly linear. Plotting concentration against log K s a t improved linearity (see Figures 12 - 14). As all scatterplots for each of nitrate, chloride, and N/Cl are very similar, only one typical sample event is shown, a. Chloride There is generally no significant correlation between log K s a t and chloride concentration in the leachate. The preflush chloride concentrations and Run 3 OF samples were the only exceptions, both showing significant positive correlations. The results of the chloride correlations are somewhat surprising, especially in view of the correlation results of other chemical constituents. T a b l e XVII Spearman c o r r e l a t i o n c o e f f i c i e n t s f o r c o n c e n t r a t i o n w i t h h y d r a u l i c c o n d u c t i v i t y PRE R30F R3HW R3FD R3T0T R40F R4HW1 R4HW2 R4FD R4T0T GT Cl 0.G17 ** 0.563 * -0.252 0.224 0.27 -0.035 -0.236 -0.283 -0.267 -0.153 0.004 N03+N02 0.865 *** 0.859 *** 0.702 ** 0.837 *** 0.838 *** 0.882 *** 0.525 0.636 * 0.844 *** 0.840 *** 0.834 *** N/Cl -- 0.861 »** 0.758 ** 0.861 *** 0.849 *** 0.907 *** 0.843 *** 0.881 ** 0.874 *** 0.852 *** 0.828 *** Notes: *_ **_ »** _ s i g n i f i c a n t at 0.05, 0.01, and 0.001, r e s p e c t i v e l y ; s i g n i f i c a n c e l e v e l s a r e f o r t w o - t a i l e d t e s t s 103 20-18-J , 16-ion 14-o 12-Concenl 10-8-loride 6-_c CJ 4-2-A A A 0 1 2 3 4 Log Satiated Hydraulic Conductivity (cm/day) Figure 12. Relationship between Run 4 Final Drainage CI" concentrations and hydraulic conductivity. 104 18-16-12 "p 10-"c fl> R-I 6 H "5 4H 2-0 A A A A A A A A A A A A A A A 0 1 2 3 4 Log Satiated Hydraulic Conductivity (cm/day) F i g u r e 13. R e l a t i o n s h i p b e t w e e n Run 4 F i n a l D r a i n a g e N 0 3 _ c o n c e n t r a t i o n s and h y d r a u l i c c o n d u c t i v i t y . 105 o o 2-1.5-1 -0.5-0-A A A A A A A A A A 0 1 2 3 4 Log Satiated Hydraulic Conductivity (cm/day) F i g u r e 14. R e l a t i o n s h i p b e t w e e n Run 4 F i n a l D r a i n a g e N/C l r a t i o s and h y d r a u l i c c o n d u c t i v i t y . , 106 The significant positive correlation between high K s a t and leachate Cl " in the overflow from Run 3 is logical. C l " was applied at the start of Run 3, just prior to the commencement of "precipitation". The overflow samples are composed of the first leachate reaching the column bottoms - likely a combination of (i) soil water from the lower parts of the columns displaced by mass flow, (ii) soil water from nearer the column surface draining through large, continuous cracks or channels, and perhaps (iii) a component of the precipitation water running directly down the column walls or through large cracks. The second and especially third sources of soil water could be expected to have higher chloride concentrations than the first source. In the case of (iii) above, the precipitation water may be expected to pick up recently-applied chloride which has not had time to diffuse into smaller soil pores. If the high hydraulic conductivities result from a significant contribution of flow from large cracks or, especially, edge flow, then the appearance of high chloride in the overflow samples is logical. Continuing this reasoning, in columns with low conductivities (and likely relatively less bypassing flow), the applied chloride would not be picked up by water rapidly draining down the column, but would be carried deeper into the soil mass, presumably with more opportunity for diffusion and dispersion of the chloride within the pore network. The leachate in these columns should be composed of soil water from low in the column displaced by piston-type flow from above, and perhaps water from larger cracks which are not continuous with the surface or upper regions. Soil water from lower in the profile in these columns would be expected to have lower chloride concentrations, as the slowly penetrating, recently-applied chloride would not likely have reached the column bottoms by the overflow events in Run 3. Run 4 OF did not correlate significantly with K s a t , likely because the Cl" applied at the start of Run 3 had penetrated the soil peds at the surface by the time that Run 4 started. Any bypassing flow in R40F1 would not have "picked up " much of the Cl", therefore, as it may have done at the start of Run 3. The significant correlation between chloride and K s a j . in the preflush may result from similar effects as described above: from the low percent mass recoveries of chloride it 107 appears that much of the chloride, even chloride applied in earlier runs, is retained in the upper regions of the soil columns. When the columns were held saturated for a few days in the preflush, chloride-rich soil water had time to mix with the preflush water. When the columns were drained, those with high K s a t had higher chloride concentrations in the leachate, as the chloride within the upper regions of the column could drain directly down larger cracks. In low K s a t columns more of the water and chloride had to move through the soil mass, and may have been more subject to diffusive exchange with relatively immobile water, thus lowering the chloride concentration in leachate reaching the column bottom. The lack of a significant correlation for other sample times might be explained, at least partially, by natural variability in the native soil chloride content. Native chloride leaches out of the soil matrix and if present in sufficient quantity, could obscure the leaching patterns of applied chloride. The small size of all correlation coefficients may support this argument: perhaps higher chloride application rates would have resulted in more distinct leaching patterns. b. Nitrogen For all sample times there is a significant and positive correlation between leachate NO3 + NC>2 concentration and hydraulic conductivity. Except for the HW samples, Spearman's r s is always over 0.8, indicating quite a strong correlation. All Spearman correlation coefficients for the N/Cl ratios with log K s a t are positive and significant at equal to or greater than the 0.01 level. The explanation for the positive correlation between leachate chloride concentration and log Ksa^. also applies to leaching of NO3 + NO2, with some additional considerations. As has already been discussed, the nitrogen anions are usually free to move with soil water. However, they are also subject to chemical and/or microbial reaction. The much stronger correlation between nitrate and K g a t than between CI" and K s a j . was likely due, at least in part, to the non-conservative nature of the nitrogen anions. 108 The conductivity affects the residence time of solutes in the saturated soil zone. In columns with high conductivities it is plausible that some of the applied NO3" leached through large channels too quickly to interact with the soil. Where the K g a t was lower, more of the applied NO3" entered the soil matrix, where it may have been subject to transformations including denitrification or immobilization. Thus, in the lower K $ a t columns, the proportion of NO3" reaching the column bottoms would have been relatively less than the chloride proportion for the same columns, as evidenced by the strong, positive correlations between K j a^ and N/CI ratios. Even if, for example, the denitrification potential is high, if the residence time is short, very little NO3" will be denitrified. The interactions leading to retention of N within the columns or denitrification therefore appear to be additive to the purely physical effects of K s a t on leaching, resulting in stronger correlations between K $ a t and NO3 + NC>2 concentration than between K $ a t and C l " leachate concentration. A second factor which may contribute to the differences in correlation of NO3 + N 0 2 and Cl " with K s a t may be the rate of application relative to the native content of Cl" or NO3 + N0 2. The NO3 + N 0 2 content of the soil was very low at the start of the experiment, so that the levels added were relatively high. However, chloride was likely added at a much lower rate relative to its initial concentration, which, as discussed, may have obscured leaching patterns resulting from the effect of hydraulic conductivity. Significant correlation coefficients of the magnitude obtained here, however, argue in favour of including K s a (. as a covariate in subsequent analyses of variance. 5. Effect of Water Table There were two ways in this experiment in which the effect of water table on leachate concentration could be evaluated. In the first way, the half-way samples allow a comparison of soil water composition at different depths with respect to the water table. Unfortunately, most of the H W samples were obtained from below the water table so that concentrations in unsaturated and saturated zones could not be compared. 109 There did not appear to be any difference in HW nitrate concentration or N/Cl ratio between the 75 and 55 cm water tables. This was a little unexpected, because some denitrification was expected in the larger saturated zone in the 75 cm treatment. From the data in Tables IX to XI, it is obvious that, for the columns where HW samples were obtained, the concentrations of CI" and N O j + N 0 2 , and usually the N/Cl ratios, were much higher in the HW samples than either the OF or FD. The lower nitrate content and N/Cl ratio of samples obtained from the column bottom (OF and FD) could indicate denitrification occurring between the HW sample point (45 - 50 cm) and the column bottom. The change in CI" concentration between the same two sample points, however, means that N " loss" can not be attributed solely to denitrification: attenuation of the CI" concentration indicates that there is considerable dispersion of solute as it moves below the HW point, accompanied by solute retention within the soil matrix. Denitrification below the water table in this soil has also been suggested by the data of Richard and Chieng (1985), who found that, at the Boundary Bay site, both well water and drain water NO3" concentrations in tile drained field plots were significantly greater than well water nitrate concentrations from an undrained plot, especially in the spring. In the discussion on trends in leachate concentration with time it was noted that in 4 of the 8 columns where HW N/Cl ratios could be calculated, the N/Cl ratio did not change between HW1 and HW2 in Run 4, indicating no loss of N at the HW point during incubation of the column. Three of the four columns with constant N/Cl ratios had 35 or 55 cm water tables, meaning that the samplers were above or just below the water table. Considering that the water used was not deoxygenated, there jnay have been too much oxygen in solution near the top of the water table for substantial denitrification to have occurred. Gilliam and Gambrell (1978) suggested that cold, percolating rainwater may carry oxygen deep into the soil profile in the field. The second way in which the effect of water table was evaluated (and the original main objective of the experiment), was to compare leaching from columns with different water 110 tables. It is obvious from the very similar means and very wide and overlapping confidence intervals in Tables IX to XI that water table height had very little effect on leachate concentration in this experiment. The results of the analyses of variance for each sample time are presented in Appendix A; the calculated F probabilities are summarized in Table XVIII. For the CI" concentrations, the NO3 + N 0 2 concentrations, and the N/Cl ratios there is no significant difference in leachate composition for any of the 4 water tables tested. In all cases the error mean squares were very large relative to the treatment mean squares, indicating that there was a very large amount of variability in response within a treatment group (water table level). This variability may be due partially to inherent differences between experimental units (soil columns), and perhaps also to the effects of another independent variable (such as hydraulic conductivity). Covariance analysis can be used to remove the effect of an independent variable on the dependent variable so that the effect of a treatment factor on the dependent variable can be assessed. In this experiment the strong correlation between K s a (. and leachate composition, and the reasonable assumption that K g a^ affects solute time in residence in the soil and hence opportunity for microbial transformation, indicates that covariance analysis may be appropriate. In addition to the basic assumptions of analysis of variance, covariance analysis also assumes that (Hicks, 1982): the independent variable is not affected by the treatment factor (reasonable for the short term in this case); the regression is linear and of non-zero slope; and the regression coefficients are homogenous. Linearity of the regression will be assumed. Homogeneity of the regression coefficents can be tested indirectly by treating the covariate as another independent variable and testing the significance of the interaction between the two independent variables (Norusis, 1985): if the interaction is not significant, homogenous slopes can be assumed (ie. the effect of hydraulic conductivity on leaching can be assumed to be the same for all water tables). Tests for interaction of water table and K s a(. were nonsignificant for all sample times, therefore homogeneity of slopes can be assumed and a common slope for the covariate (K t) can be used. T a b l e X V I I I . Summary o f ANOVA and ANCOVA r e s u l t s Sample F P r o b a b i l i t y E v e n t ANOVA ANCOVA CHLORIDE Run 3 OF 0.886 FD 0.891 — TOT 0.667 — Run 4 OF 0.846 — FD 0.913 — TOT 0.805 — Grand T o t a l 0.707 — NITRATE Run 3 OF 0.607 0 .151 FD 0.968 0 .268 TOT 0 .940 0 .208 Run 4 OF 0.633 0 .177 FD 0.999 0 .392 TOT 0.999 0 .349 G rand T o t a l 0 .988 0 .249 N/CI RATIO Run 3 OF 0.635 0 .248 FD 0.975 0 .504 TOT 0.963 0 .404 Run 4 OF 0 .579 0 .212 FD 0.997 0 .274 TOT 0.998 0 .424 G rand T o t a l 0 .989 0 .404 112 From the analysis of covariance (ANCOVA) tables in Appendix B, it can be seen that the regression is significant for all sample times for N O j + .N02 concentration and N/CI ratios, as expected from the previous correlation analysis with hydraulic conductivity. The calculated significance levels for the F tests for the effect of water table are summarized in Table W i l l for comparison to the A N O V A results. Although the F probabilities for the significance of the treatment effect decrease markedly when was included as a covariate, there is still no statistically significant effect of water table on leachate concentration. From the results, though, it seems reasonable to speculate that if the hydraulic conductivities had been constant, there may have been an effect of water table on nitrate leaching. From the treatment means adjusted for K $ a t in Appendix B, the 35 cm water table has the highest NO3 + N 0 2 concentration, and the highest N/CI ratio, in every sample time. The slightly greater (but not significantly greater) leaching losses from the 35 cm water table may perhaps be explained by the balance between denitrification and leaching. Denitrification may have been greater in the 55 and 75 cm water table columns due to the larger saturated zone, which also extended higher into the soil profile where the carbon content was likely higher also. This would agree with the results of van Dijk's (1980) lysimeter study, where the greatest leaching losses of nitrate corresponded to the lowest water table levels, and with the findings of Richard and Chieng (1985) discussed previously. 113 V. SUMMARY AND CONCLUSION These experiments were designed to study the effect of water table height on nitrate leaching losses over the relatively short term of approximately one week, under spring conditions. It is clear from the results that there was no significant effect of water table height on nitrate leaching. Chloride concentrations in the leachate increased with time during the experiment, and reached a maximum of about 10 - 15 mg CI7L in the Run 4 drainage after incubation. Nitrate concentrations in the leachate also increased with time within a run, but decreased between runs, presumably due to denitrification and dilution from the initial precipitation applied. Nitrate concentrations reached maximums of up to 16 mg N/L. A few columns consistently exceeded the recommended safe drinking water level of 10 mg NO-^'-N/L, even at the very modest fertilization rate of 35 kg N/ha. It is difficult to estimate whether the high nitrate or low nitrate columns are representative of what would actually occur in this soil in the field. The leachate concentrations are equivalent to losses of up to about 15 kg N/ha, and average about 5 kg N/ha, during Runs 3 and 4. However, the amount of water that passed through the columns was much greater than would occur normally in the field in that time period, and part of the N lost was likely from N added in earlier Runs (in other words, because of retained fertilizer, the actual fertilization rate was greater than 35 kg/ha of N). Comparing the solution concentrations of nitrate and chloride and the N/Cl ratios at 45 - 50 cm depth with leachate collected from the column bottoms indicated that there was considerable dilution or dispersion occurring between the two depths. There was also some loss of NO^", probably by denitrification, as evidenced by a lower N/Cl ratio in leachate obtained from the column bottoms than from the half way sample point. Variability is a common characteristic of natural systems and was the predominant characteristic of the leaching columns used in this study. Nutrient content and leaching are very variable characteristics in soils. The field soil samples showed that A horizon N H 4 + contents were quite variable. The leaching experiment was conducted at cool temperatures 114 not only to simulate early spring conditions, but also to reduce microbial activity (mineralization and immobilization should have been slow under the saturated conditions and below 10 C over the short duration of the experiment), in an effort to reduce a potential source of variability between columns. When these soil columns were first removed from the field and drained, and especially later between Runs 2 and 3 when the columns warmed up to room temperature for about one month, it is likely that mineralization and nitrification occurred within the columns, creating additional and variable sources of NO-j ' among the columns. Native nitrate concentrations in the field soil samples were very low, with only a few exceptions, and thus should not have been a major source of variability in this experiment. However, because most of the field soil samples and the Run 0 leachate samples were at or near the detection limit of the analytical method used, it was not possible to calculate a true initial N/Cl ratio by which later leachate ratios could be corrected. If this correction were made, it might reduce some of the variability in the experiment (expressed as high experimental error in the ANOVAs). The small size of the cores used in this study may have contributed to variability in leachate composition in two ways. First, by emphasizing inherent soil structural or chemical variability. For example, a column with a few large channels in which bypassing flow occurs, may perform quite differently than one without any large channels, in which only displacement operates. In those columns that consistently had higher leachate nitrate concentrations than the others, bypassing flow may have been operational. Dye tests with methylene blue indicated that preferential flow occurred, at least in the upper regions of some columns. Also, the soil was observed to be very heterogenous both in the field and in the cores that were sectioned: some cores contained large organic deposits, some appeared to have larger reduced zones than others, etc. This heterogeneity likely contributed to variation in leaching of native C l " and NO^". 115 Secondly, there may have been effects on leachate composition resulting from edge effects (boundary flow), as discussed in the literature review. While this cannot be discounted, the dye stains, leachate ortho-phosphate concentrations and the observation that the sectioned cores appeared to fit tightly against the column walls, were indications that boundary flow was not significant. The hydraulic characteristics of the columns were also quite variable: infltration rates were observed to be quite variable; drainable porosity ranged from approximately 3% to 10.5%, with a CV of 39%; and the satiated hydraulic conductivity ranged from extremes of approximately 1 to 1400 cm/day, with a CV of 181%. Hydraulic conductivity correlated significantly and positively with leachate nitrate concentrations and the N/CI ratios, indicating the importance of hydraulic characteristics to solute movement. In columns with low K $ a t the solute residence time increased and NO^'was removed from the soil solution, by denitrification or immobilization, as evidenced by decreasing N/CI ratios. When an analysis of covariance was applied with K s a t as a covariate, the F ratios for the effect of water table on nitrate leaching improved considerably, but were still not significant. The indication is clear, though, that there may have been an effect of water table height on leachate quality that was obscured by variation in physical characteristics between columns. Recommendations for Further Study Based on work in this study, the following recommendations are made. 1) If the Giddings Probe is used to take undisturbed cores, the following suggestions should improve the quality of the soil columns that are isolated. a) If the pipe is pushed into the ground, the press plate used should be perforated to prevent vertical compression (as described in the Materials and Methods). b) It is possible that the variability in physical properties, such as drainable porosity and K s a t , experienced in this experiment was at least partially due to unequal compaction of the soil columns during isolation. If so, column placement by a rotary-action, double-sleeve 116 technique (as described in the Literature Review) should reduce compaction of the cores generally, and should therefore reduce the effects of unequal compaction. 2) If further studies are undertaken on the effect of water table on nitrate leaching, it is recommended that: a) Undisturbed soil be used. Physical properties correlated strongly with leaching, indicating the importance of natural soil structure to solute movement. Undisturbed soil is therefore necessary if the objective is to study solute movement as it occurs in the field. b) Larger experimental units be used. Larger units -- either large field lysimeters or tile drained plots -- are recommended for two reasons. First, to reduce the relative importance of both the edge effect (boundary flow), which may occur in any lysimeter, and compacted zones, which may arise in lysimeters that are pushed or rotated into place. Secondly, in small cores the structural features may differ drastically among experimental units; in larger units it is more likely that each unit will contain at least some of the influential structural features. 117 VI. LITERATURE CITED Addiscott, T.M.; D.A. Rose; J. Bolton. 1978. Chloride leaching in the Rothamsted drain gauges: influence of rainfall pattern and soil structure. Jour. Soil Sci. 29: 305 - 314. Avnimelech, Y. & J. Raveh. 1976. Nitrate leakage from soils differing in texture and nitrogen load. J. Env. Qual. 5:79 - 82. Bailey, L.D. & E.G. Beauchamp. 1973. Effects of moisture, added NO3 and macerated roots on NO3 transformation and redox potential in surface and subsurface soils. Can J. Soil Sci. 53:219 - 230. Baker, J.L.; K.L. Campbell; H.P. Johnson; J.J. Hanway. 1975. Nitrate, phosphorus and sulphate in subsurface drainage water. J. Env. Qual. 4:406 - 412. Baker, J.L. & H.P. Johnson. 1981. Nitrate-nitrogen in tile drainage as affected by fertilization. J. Env. Qual. 10:519 - 522. Balderston W.L.; B. Sherr; W.J. Payne. 1976. Blockage by acetylene of nitrous oxide reduction in Pseudomonas perfectomarinus. Appl. Environ. Microbiol. 31:504 - 508. Bausch, W.; A.B. Onken; C W . Wendt; O.C. Wilke. 1977. A self-propelled high-clearance soil coring machine. Agron. J. 69:122 - 124. Beauchamp, E.G.; C. Gale; J.C. Yeomans. 1980. Organic matter availability for denitrification in soils of different textures and drainage classes. Commun. Soil Sci. Plant Anal. 11:1221 -1233. Belser, L.W. 1982. Inhibition of nitrification in N.S. Subba Rao (Ed). Advances in agricultural microbiology. Butterworth Scientific, London, England, pp.267 - 293. Benoit, G.R. 1973. Effect of agricultural management of wet sloping soil on nitrate and phosphorus in subsurface water. Water Resour. Res. 9:1296 - 1303. Bergstrom, L. 1987. Nitrate leaching and drainage from annual and perennial crops in tile-drained plots and lysimeters. J. Env. Qual. 16:11 - 18. Bohn, H.L.; B.L. McNeal; G.A. O 'Connor. 1979. Soil Chemistry. Wiley-lnterscience, N.Y., N.Y. Bolton, E.F.; J.W. Aylesworth; F.R. Hore. 1970. Nutrient losses through tile lines under three cropping systems and two fertility levels on a Brookston clay soil. Can. J. Soil Sci. 50:275 -279. Bondurant, J.A.; R.V. Worstell; C.E. Brockway. 1969. Plastic casings for soil cores. Soil Sci. 107:70 - 71. Bottcher, A.B.; E.J. Monke; L.F. Huggins. 1981. Nutrient and sediment loadings from a subsurface drainage system. Trans. ASAE. 24:1221 - 1226. Bouma, J. & L.W. Dekker. 1978. A case study on infiltration into dry clay soil, I. Morphological observations. Geoderma 20:27 - 40. 118 Bouma, J.; L.W. Dekker; C.J. Muilwijk. 1981. A field method for measuring short-circuiting in clay soils. J. Hydrol. (Neth.) 52:347 - 354. Bouma, ). & J.M.M. Wosten. 1979. Flow patterns during extended saturated flow in two undisturbed swelling clay soils with different macrostructures. Soil Sci. Soc. Amer. Jour. 43:16 - 22. Bowman, R.S. 1984. Evaluation of some new tracers for soil water studies. Soil Sci. Soc. Amer. Jour. 48:987 - 993. Buchter, B.; J. Leuenberger; P.J. Wierenga; F. Richard. 1984. Preparation of large core samples from stony soils. Soil Sci. Soc. Amer. Jour. 48:1460 - 1462. Burford, J.R. & J.M. Bremner. 1975. Relationships between the denitrification capacities of soils and total, water-soluble and readily decomposable soil organic matter. Soil Biol. Biochem. 7:389 - 394. van Burg, P.F.J.; M.L. 't Hart; H. Thomas. 1980. Nitrogen and grassland - past and present situation in the Netherlands, in Proc. Int. Symp. Eur. Grassland Fed. on The role of nitrogen in intensive grassland production. Pudoc, Wageningen, Neth. pp. 1 5 - 3 3 . Burns, I.G. & D.J. Greenwood. 1982. Estimation of the year-to-year variations in nitrate leaching in different soils and regions of England and Wales. Agric. & Environ. 7:35 - 45. Burton, D.L. & E.G. Beauchamp. 1985. Denitrification rate relationships with soil parameters in the field. Commun. Soil Sci. Plant Anal. 16:539 - 549. Burwell, R.E.; G.E. Schuman; K.E. Saxton; H.G. Heinemann. 1976. Nitrogen in subsurface discharge from agricultural watersheds. J. Env. Qual. 5:325 - 329. Cameron, D.R.; C.G. Kowalenko; K.C. Ivarson. 1978. Nitrogen and chloride distribution and balance in a clay loam soil. Can J. Soil Sci. 58:77 - 88. Chao, E.C.Y. 1987. Water table depth simulation for flat agricultural land under subsurface drainage and subirrigation practices. M.A.Sc. Thesis, Department of Bio-Resource Eng., University of British Columbia. van Cleemput, O. & L. Baert. 1984. Nitrite: a key compound in nitrogen loss processes under acid conditions? Plant & Soil 76:233 - 241. Cochran, P.H.; G.M. Marion; A.L. Leaf. 1970. Variations in tension lysimeter leachate volumes. Soil Sci. Soc. Amer. Proc. 34:309 - 312. Colburn P. & R.J. Dowdell. 1984. Denitrification in field soils. Plant & Soil 76:213 - 226. Coles, N. & S. Trudgill. 1985. The movement of nitrate fertilizer from the soil surface to drainage waters by preferential flow in weakly structured soils, Slapton, S. Devon. Agric. Ecosyst. & Environ. 13:241 - 259. Conover, W.J. 1971. Practical nonparametric statistics. John Wiley, N.Y., N.Y. Cooke, G.W. & R.J.B. Williams. 1970. Losses of nitrogen and phosphorus from agricultural land. Water Treat. & Examin. 19:253 - 276. 119 Craswell, E.T. 1978. Some factors influencing denitrification and nitrogen immmobilization in a clay soil. Soil Biol. Biochem. 10:241 - 245. Deal, S.C.; J.W. Gilliam; R.W. Skaggs; K.D. Konyha. 1986. Prediction of nitrogen and phosphorus losses as related to agricultural drainage system design. Agric. Ecosyst. & Environ. 18:37 - 51. Devitt, D.; J. Letey; L.J. Lund; J.W. Blair. 1976. Nitrate-nitrogen movement through soil as affected by soil profile characteristics. J. Env. Qual. 5:283 - 288. van Dijk, T.A. 1980. Nitrogen losses from soils under grass as affected by residence time in the groundwater - lysimeter results, in Proc. Int. Symp. Eur. Grassland Fed. on The role of nitrogen in intensive grassland production. Pudoc, Wageningen, Neth. p. 170. Dowdell, R.J.; J Morrison; A.E.M. Hood. 1980. The fate of fertilizer nitrogen applied to grassland: uptake by plants, immobilization into soil organic matter and losses by leaching and denitrification, in Proc. Int. Symp. Eur. Grassland Fed. on The role of nitrogen in intensive grassland production. Pudoc, Wageningen, Neth. pp. 129 - 136. Dowdell, R.J. & C P . Webster. 1980. A lysimeter study using nitrogen-15 on the uptake of fertilizer nitrogen by perennial ryegrass swards and losses by leaching. Jour. Soil Sci. 31:65 - 75. Driehuyzen, M.G. 1983. Boundary Bay Water Control Project; summary of results. B.C. Ministry of Agriculture and Food, Unpublished report. Envt. Canada. 1982. Canadian climate normals, 1951 - 1980, Temperature and precipitation, British Columbia. Environment Canada. E.P.A. 1972. Handbook for analytical quality control in water and wastewater laboratories. U.S. E.P.A., Cincinnati, Ohio. Firestone, M.K. 1982. Biological denitrification, in F.j. Stevenson (Ed). Nitrogen in agricultural soils, Agronomy #22. ASA-CSSA-SSSA, Madison, Wisconsin, pp. 289 - 326. Gambrell, R.P.; J.W. Gilliam; S.B. Weed. 1975. Denitrification in subsoils of the North Carolina coastal plain as affected by soil drainage. J. Env. Qual. 4:311 - 316. Garstang, J.R. 1980. The date of nitrogen application to upland pasture in the spring in a high-rainfall maritime climate, in Proc. Int. Symp. Eur. Grassland Fed. on The role of nitrogen in intensive grassland production. Pudoc, Wageningen, Neth. p. 159. Cast, R.G.; W.W. Nelson; J.M. MacGregor. 1974. Nitrate and chloride accumulation and distribution in fertilized, tile-drained soils. J. Env. Qual. 3:209 - 213. Gast, R.G.; W.W. Nelson; G.W. Randall. 1978. Nitrate accumulation in soils and loss in tile drainage following nitrogen applications to continuous corn. J. Env. Qual. 7:258 - 261. Germann, P.F.; W.M. Edwards; L.B. Owens. 1984. Profiles of bromide and increased soil moisture after infiltration into soils with macropores. Soil Sci. Soc. Amer. Jour. 48:237 -244. Gilliam, J.W.; R.B. Daniels; J.F. Lutz. 1974. Nitrogen content of shallow groundwater in the North Carolina coastal plain. J. Env. Qual. 3:147 - 151. 120 Gilliam, J.W. & R.P. Gambrell. 1978. Temperature and pH as limiting factors in loss of nitrate from saturated Atlantic Coastal Plain soils. ). Env. Qual. 7:526 - 532. Gilliam, J.W.; R.W. Skaggs; S.B. Weed. 1979. Drainage control to diminish nitrate loss from agricultural fields. J. Env. Qual. 8:137 - 142. Greenwood, D.J. & I.G. Burns. 1979. Nitrate leaching: principles and estimation. ADAS Quarterly Review 35:246 - 259. Guitjens, J.C; P.S. Tsui; D.F. Thran. 1984. Quantity and quality variations in subsurface drainage. Trans. ASAE. 27:425 - 428. Haines, B.L.; J.B. Waide; R.L. Todd. 1982. Soil solution nutrient concentrations sampled with tension and zero-tension lysimeters: report of discrepancies. Soil Sci. Soc. Amer. Jour. 46:658 - 661. Hansen, E.A. & A.R. Harris. 1975. Validity of soil-water samples collected with porous ceramic cups. Soil Sci. Soc. Amer. Proc. 39:528 - 536. Hanway, J.J. & J.M. Laflen. 1974. plant nutrient losses from tile-outlet terraces. J. Env. Qual. 3:351 - 356. Hicks, CR. 1982. Fundamental concepts in the design of experiments, 3rd edition. Holt, Rinehart & Winston, N.Y., N.Y. Hillel, D. 1980. Fundamentals of soil physics. Academic Press, N.Y., N.Y. Hundal, S.S.; C O . Schwab; C S . Taylor. 1976. Drainage system effects on physical properties of a Lakebed clay soil. Soil Sci. Soc. Amer. Jour. 40:300 - 305. Jaakola, A. 1984. Leaching losses of nitrogen from a clay soil under grass and cereal crops in Finland. Plant & Soil 76:59 - 66. Jansson, S.L. & J. Persson. 1982. Mineralization and immobilization of soil nitrogen, in F.J. Stevenson (Ed). Nitrogen in agricultural soils, Agronomy #22. ASA-CSSA-SSSA, Madison, Wisconsin, pp. 229 - 252. Jeffers, J.N.R. 1959. Experimental design and analysis in forest research. Almqvist & Wiksell, Stockholm, Sweden. Johnson, W.J.; F. Ittihadieh; R.M. Daum; A.F. Pillsbury. 1965. Nitrogen and phosphorus in tile drainage effluent. Soil Sci. Soc. Amer. Proc. 129:287 - 289. Johnsson, H.; L. Bergstrom; P.E. Jansson. 1987. Simulated nitrogen dynamics and losses in a layered agricultural soil. Agric. Ecosyst. & Environ. 18:333 - 356. Kanwar, R.S.; H.P. Johnson; D. Kirkham. 1982. Transport of nitrate and gaseous denitrification in soil columns during leaching. J. Hydrol. 55:171 - 184. Keeney, D.R. & D.W. Nelson. 1982. Nitrogen - Inorganic forms, in A.L. Page (Ed). Methods of soil analysis, Part 2. Chemical and microbiological properties. Agronomy #9 (2nd Ed.). ASA - SSA, Madison, Wisconsin. 121 Khanif, Y.M.; O. van Cleemput; L. Baert. 1984. Seasonal fluctuation of mineral nitrogen in the rootzone of sandy soils. Pedologie 34:23 - 33. Kirk, R.E. 1982. Experimental design: procedures for the behavioral sciences, 2nd Ed. Brooks/Cole Pub., Monterey, California. Kissel, D.E.; J.T. Ritchie; E. Burnett. 1974. Nitrate and chloride leaching in a swelling clay soil. J. Env. Qual. 3:401 - 404. Kissel, D.E.; S.J. Smith; D.W. Dillow. 1976. Disposition of fertilizer nitrate applied to a swelling clay soil in the field. J. Env. Qual. 5:66 - 71. Klute, A. 1965. Laboratory measurement of hydraulic conductivity of saturated soil, in C A . Black (Ed). Methods of soil analysis, Part 1. Physical and mineralogical properties. ASA, Madison, Wise. pp. 210 - 221. Kneale, W.R. & R.E. White. 1984. The movement of water through cores of a dry (cracked) clay-loam grassland topsoil. J. Hydrol. (Neth.) 67:361 - 365. Kolenbrander, C.J. 1969. Nitrate content and nitrogen loss in drainwater. Netherlands J. of Agric. Sci. 17:246 - 255. Kolenbrander, C.J. 1981. Leaching of nitrogen in agriculture, in J.C. Brogan (Ed). Nitrogen losses and surface run-off from land-spreading of manures: proceedings. ECSC, EEC, EAEC. NijhoffA/V. Junk Pubs., The Hague. Kowalenko, C C . 1978. Nitrogen transformations and transport over 17 months in field fallow microplots using ^ N . Can J. Soil Sci. 58:69 - 76. Kowalenko, C C . 1987. The dynamics of inorganic nitrogen in a Fraser Valley soil with and without spring or fall ammonium nitrate applications. Can. J. Soil Sci. 67:367 - 382. LeCIerg, E.L.; W.H. Leonard; A.C. Clark. 1962. Field plot technique. Burgess Publishing, Minneapolis, Minnesota. Leyton, L. & J.S.P. Yadav. 1960. Effect of drainage on certain physical properties of a heavy clay soil. J. Soil Sci. 11:305 - 313. Lilliefors, H.W. 1967. On the Kolmogorov-Smirnov test for normality with mean and variance unknown. Am. Stat. Assoc. Jour. (1967):399 - 402. Little, T.M. & F.J. Hills. 1978. Agricultural experimentation, design and analysis. Wiley, N.Y., N.Y. Lowrance, R.R.; R.L. Todd, L.E. Asmussen. 1984. Nutrient cycling in an agricultural watershed: II. Streamflow and artificial drainage. J. Env. Qual. 13:27 - 32. Lucas, J.L. & C M . Reeves. 1980. Nitrate in groundwater and land irrigation of sewage. Prog. Wat. Tech. 13:81 - 88. Lund, L.J. 1982. Variations in nitrate and chloride concentrations below selected agricultural fields. Soil Sci. Soc. Amer. Jour. 46:1062 - 1066. Luthin, J.N. 1978. Drainage engineering. Krieger Publishing, Huntington, N.Y. 122 Luttmerding, H.A. 1981. Soils of the Langley - Vancouver map area. Report #15, Vol. 6. B.C. Soil Survey. Mahendrappa, J.K.; R.L. Smith; A.T. Christianson. 1966. Nitrifying organisms affected by climatic region in western United States. Soil Sci. Soc. Amer. Proc. 30:60 - 62. Mahli, S.S. & M. Nyborg. 1986. Increase in mineral N in soils during winter and loss of mineral N during early spring in north-central Alberta. Can J. Soil Sci. 66:397 - 409. McNeal, L. & R.C. Reeve. 1964. Elimination of boundary-flow errors in laboratory hydraulic conductivity measurements. Soil Sci. Soc. Amer. Proc. 28:713 - 714. Meek, B.D.; L.B. Grass; L.S. Willardson; A.J. Mackenzie. 1970. Nitrate transformations in a column with a controlled water table. Soil Sci. Soc. Amer. Proc. 34:235 - 239. Mielke, L.N. 1973. Encasing undisturbed soil cores in plastic. Soil Sci. Soc. Amer. Proc. 37:325 - 326. Miller, M.H. 1979. Contribution of nitrogen and phosphorus to subsurface drainage water from intensively cropped mineral and organic soils in Ontario. J. Env. Qual. 8:42 - 48. Moon, D.E. & H. Schreier. 1985. Soil variability: does it invalidate our interpretations? in The role of soil analysis in resource management, Proc. 9th B.C. Soil Sci. Workshop, B.C. Min. Environ. Tech. Rep. #16. pp. 99 - 109. Murphy, J.B.; E.H. Crissinger; W.C. Little. 1981. Fiberglass encasement of large, undisturbed, weakly cohesive soil samples. Soil Sci. 131:130 - 134. Nelson, D.W. 1982. Caseous losses of nitrogen other than through denitrification, in F.J. Stevenson (Ed). Nitrogen in agricultural soils, Agronomy #22. ASA-CSSA-SSSA, Madison, Wisconsin, pp. 327 - 363. Norusis, M.J. 1985. SPSSx advanced statistics guide. McGraw-Hill Books, N.Y., N.Y. Omoti, U. & A. Wild. 1979. Use of fluorescent dyes to mark the pathways of solute movement through soils under leaching conditions: 2. Field experiments. Soil Sci. 128:98 - 104. Parkin, T.B. 1987. Soil microsites as a source of denitrification variability. Soil Sci. Soc. Amer. Jour. 51:1194 - 1199. Patten, D.K.; J.M. Bremner; A.M. Blackmer. 1980. Effects of drying and air-dry storage of soils on their capacity for denitrification of nitrate. Soil Sci. Soc. Amer. Jour. 44:67 - 70. Pfaffenberger, R.C. & J.H. Patterson. 1977. Statistical methods for business and economics. Irwin-Dorsey, Ltd., Georgetown, Ont. Phillips, P.A,; J.L.B. Culley; F.R. Hore; N.K. Patni. 1981. Pollution potential and corn yields from selected rates and timing of liquid manure applications. Trans ASAE. 24:139 - 144. Quin, B.F. & L.J. Forsythe. 1976. All-plastic suction lysimeters for the rapid sampling of percolating soil water. N.Z. Jour, of Sci. 19:145 - 148. Richard, P. & S. Chieng. 1985. Nitrogen and phosphorus leaching under drained and undrained conditions in agricultural lands. ASAE. Paper #PNR-85-210. 123 Ritchie, J.T.; D.E. Kissel; E. Burnett. 1972. Water movement in undisturbed swelling clay soil. Soil Sci. Soc. Amer. Proc. 36:874 - 879. Ritjema, P.E. 1980. Nitrogen emission form grassland farms - a model approach, in Proc. Int. Symp. Eur. Grassland Fed. on The role of nitrogen in intensive grassland production. Pudoc, Wageningen, Neth. pp. 137 - 147. Rosswall, T. & K. Paustian. 1984. Cycling of nitrogen in modern agricultural systems. Plant & Soil 76:3 - 21. Runge, E.C.A. 1965. The hydraulic soil-coring machine - a new method of sampling Illinois soils. Illinois Res. Spring:10 - 11. Russell, E.W. 1973. Soil conditions and plant growth. Longman, N.Y., N.Y. Schmidt, E.L. 1982. Nitrification in soil, in F.J. Stevenson (Ed). Nitrogen in agricultural soils, Agronomy #22. ASA-CSSA-SSSA, Madison, Wisconsin, pp.253 - 288. Severson, R.C. & D.F. Grigal. 1976. Soil solution concentrations: effect of extraction time using porous ceramic cups under constant tension. Water Resour. Bull. 12:1161 - 1170. Shaffer, K.A.; D.D. Fritton; D.E. Baker. 1979. Drainage water sampling in a wet, dual-pore system. J. Env. Qual. 8:241 - 246. Sharpley, A.N. & J.K. Syers. 1981. Amounts and relative significance of runoff types in the transport of nitrogen into a stream draining an agricultural watershed. Water, Air & Soil Poll. 15:299 - 308. Sheppard, S.C. & T.E. Bates. 1986. Changes in nitrate concentration overwinter in three southern Ontario soil profiles. Can J. Soil Sci. 66:537 - 541. Shuford, J.W.; D.D. Fritton; D.E. Baker. 1977. Nitrate-nitrogen and chloride movement through undisturbed forest soil. J. Env. Qual. 6:255 - 259. Siegel, S. 1956. Nonparametric statistics. McCraw Hill, N.Y., N.Y. Skaggs, R.W. 1978. Analysis of combination surface-subsurface drainage systems for humid regions. ASAE Paper #78-2541. Skaggs, R.W. & J.W. Gilliam. 1981. Effect of drainage system design and operation on nitrate transport. Trans ASAE. 24:929 - 934, 940. Smith, K.A. 1980. A model of the extent of anaerobic zones in aggregated soils, and its potential application to estimates of denitrification. Jour. Soil Sci. 31:263 - 277. Stanford, G.; R.A. VanderPol; S. Dzienia. 1975. Denitrification rates in relation to total and extractable soil carbon. Soil Sci. Soc. Amer. Proc. 39:284 - 289. Starr, J.L. & J.Y. Parlange. 1976. Solute dispersion in saturated soil columns. Soil Sci. 121:364 -372. Tadros, V.T. & J.W. McGarity. 1976. A method for collecting soil percolate and soil solution in the field. Plant & Soil 44:655 - 667. 124 Technicon Corporation. 1969. Nitrate and nitrite in water, Indust. method #33-69W. Technicon Corp., Tarrytown, N.Y. Technicon Industrial Systems. 1971. Chloride in water and waste water, Indust. method #99-70W. Technicon Industrial Systems, Tarrytown, N.Y. Technicon Industrial Systems. 1973a. Ammonia in water and seawater, Indust. method #154-71W. Technicon Industrial Systems, Tarrytown, N.Y. Technicon Indust. Systems. 1973b. Nitrite in water and seawater, Indust. method #161-71W. Technicon Industrial Systems, Tarrytown, N.Y. Technicon Indust. Systems. 1973c. Orthophosphate in water and seawater, Indust. meth. #155-71W. Technicon Industrial Systems, Tarrytown, N.Y. Thomas, C W . & B.J. Barfield. 1974. The unreliability of tile effluent for monitoring subsurface nitrate-nitrogen losses from soils. J. Env. Qual. 3:183 - 185. Thomas, C W . & R.E. Phillips. 1979. Consequences of water movement in macropores. J. Env. Qual. 8:149 - 152. Tiedje, J.M.; A.J. Sexstone; T.B. Parkin; N.P. Revsbech; D.R. Shelton. 1984. Anaerobic processes in soil. Plant & Soil 76:197- 212. Tisdale, S.L.; W.L. Nelson; J.D. Beaton. 1985. Soil fertility and fertilizers. Macmillan Co., N.Y., N.Y. Turner, C O . 1974. A device for obtaining uncontaminated and undisturbed soil samples. Down to Earth 30:18 - 21. Tuttle, CL. ; M.S. Golden; D.L. Sirois. 1984. A portable tool for obtaining soil cores in clayey or rocky soils. Soil Sci. Soc. Amer. Jour. 48:1453 - 1455. Tyler, D.D. & C W . Thomas. 1977. Lysimeter measurements of nitrate and chloride losses from soil under conventional and no-tillage corn. J. Env. Qual. 6:63 - 67. Tyler, D.D. & C W . Thomas. 1981. Chloride movement in undisturbed soil columns. Soil Sci. Soc. Amer. Jour. 45:459 - 461. Verdegem, L & L. Baert. 1984. Losses of nitrate nitrogen in sandy and clayey soils, 1. Leaching of nitrate-nitrogen to the subsoil; hints to chemical nitrate-nitrogen reduction; nitrate-nitrogen losses by artificial drainage. Pedologie 34:235 - 255. Verdegem, L.; O. van Cleemput; J. Vanderdeelen. 1981. Some factors inducing the loss of nutrients out of the soil profile. Pedologie 31:309 - 327. Walter, M.F.; T.S. Steenhuis; D.A. Haith. 1979. Nonpoint source pollution control by soil and water conservation practices. Trans ASAE. 22:834 - 840. Warrick, A.W. & D.R. Nielsen. 1980. Spatial variability of soil physical properties in the field, in D. Hillel (Ed). Applications of soil physics. Academic Press, N.Y., N.Y. pp. 319 - 344. Weber, J.B. & T.F. Peeper. 1977. Herbicide mobility in soils, in B. Truelove (Ed). Research methods in weed science, 2nd Ed. Southern Weed Sci. Society, pp. 73 - 78. 125 White, R.E. 1985a. The influence of macropores on the transport of dissolved and suspended matter through soil, in B.A. Stewart (Ed). Advances in soil science, vol. 3. Springer-Verlag, N.Y., N.Y. pp. 95 - 120. White, R.E. 1985b. A model for nitrate leaching in undisturbed structured clay soil during unsteady flow. Jour, of Hydrology 79:37 - 51. White, R.E.; R.A. Haigh; J. H. Macduff. 1987. Frequency distributions and spatially dependent variability of ammonium and nitrate concentrations in soil under grazed and ungrazed grassland. Fertilizer Research 11:193 - 208. Yoshinari, T. & R. Knowles. 1976. Acetylene inhibition of nitrous oxide reduction by denitrifying bacteria. Biochem. Biophys. Res. Commun. 69:705 - 710. Zimmermann, C.F.; M.T. Price; J.R. Montgomery. 1978. A comparison of ceramic and teflon in-situ samplers for nutrient pore water determinations. Estuar. & Coastal Mar. Sci. 7:93 - 97. Zwerman, P.J.; T. Greweling; S.D. Klausner; D.J. Lathwell. 1972. Nitrogen and phosphorus content of water from tile drains at two levels of management and fertilization. Soil Sci. Soc. Amer. Proc. 36:134- 137. APPENDIX A. Summary of ANOVA Results CHLORIDE CONCENTRATION - RUNS 3 & 4 ANALYSIS OF VARIANCE 127 Variable PRE By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES 1.9293 75.7556 77.6849 MEAN SQUARES .6431 5.0504 F F RATIO PROB. .1273 .9424 Variable R30F By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS • TOTAL D.F. 3 13 16 SUM OF SQUARES 2.7847 56.6721 59.4568 MEAN SQUARES .9282 4.3594 F F RATIO PROB. .2129 .8857 Variable R3FD By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES 3.9332 95.4186 99.3518 MEAN SQUARES 1.3111 6.3612 F F RATIO PROB. .2061 .8906 CHLORIDE CONCENTRATION - RUNS 3 & 4 ANALYSIS OF VARIANCE 128 Variable R3TOT By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES 6.1937 58.1785 64.3722 MEAN SQUARES 2.0646 3.8786 F F RATIO PROB. .5323 .6671 Variable R40F By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 12 15 SUM OF SQUARES 4.7942 71.1674 75.9617 MEAN SQUARES 1.5981 5.9306 F F RATIO PROB. .2695 .8462 Variable R4FD By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES 4.1303 119.3699 123.5002 MEAN SQUARES 1.3768 7.9580 F F RATIO PROB. .1730 .9130 CHLORIDE CONCENTRATION - RUNS 3 & 4 ANALYSIS OF VARIANCE 129 Variable R4T0T By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES 5.4145 82.3882 87.8028 MEAN SQUARES 1.8048 5.4925 F F RATIO PROB. .3286 .8047 Variable GRTOT By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES 5.9739 63.3699 69.3438 MEAN SQUARES 1.9913 4.2247 F F RATIO PROB. .4714 .7068 130 NITRATE PLUS NITRITE CONCENTRATION - RUNS 3 & 4 ANALYSIS OF VARIANCE Variable PRE By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES 1.3355 418.7393 420.0748 MEAN SQUARES .4452 27.9160 F F RATIO PROB. .0159 .9971 Variable R30F By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 13 16 SUM OF SQUARES 66.9070 458.8065 525.7135 MEAN SQUARES 22.3023 35.2928 F F RATIO PROB. .6319 .6074 Variable R3FD By Variable WT SOURCE BETWEEN GROUPS D.F. 3 SUM OF SQUARES 9.1841 MEAN SQUARES 3.0614 F F RATIO PROB. .0835 .9680 WITHIN GROUPS 15 549.8317 36.6554 TOTAL 18 559.0158 131 NITRATE PLUS NITRITE CONCENTRATION - RUNS 3 & 4 ANALYSIS OF VARIANCE Variable R3TOT By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES 16.8036 638.9011 655.7047 MEAN SQUARES 5.6012 42.5934 F F RATIO PROB. .1315 .9398 Variable R40F By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 12 15 SUM OF SQUARES 52.0533 352.5867 404.6400 MEAN SQUARES 17.3511 29.3822 F F RATIO PROB. .5905 .6329 Variable R4FD By Variable WT SUM OF MEAN F F SOURCE D.F. SQUARES SQUARES RATIO PROB. BETWEEN GROUPS 3 1.0965 .3655 .0103 .9985 WITHIN GROUPS 15 533.5767 35.5718 TOTAL 18 534.6731 132 NITRATE PLUS NITRITE CONCENTRATION - RUNS 3 S. 4 ANALYSIS OF VARIANCE Variable R4T0T By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES .9637 503.8728 504.8365 MEAN SQUARES .3212 33.5915 F F RATIO PROB. .0096 .9987 Variable GRTOT By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES 4.8023 556.3460 561.1482 MEAN SQUARES 1.6008 37.0897 F F RATIO PROB. .0432 .9876 133 N/Cl RATIOS - RUNS 3 & 4 ANALYSIS OF VARIANCE Variable R30F By Variable WT SUM OF MEAN F F SOURCE D.F. SQUARES SQUARES RATIO PROB. BETWEEN GROUPS 3 .8526 .2842 .5864 .6346 WITHIN GROUPS 13 6.3004 .4846 TOTAL 16 7.1530 Variable R3FD By Variable WT SUM OF MEAN F F SOURCE D.F. SQUARES SQUARES RATIO PROB. BETWEEN GROUPS 3 .0988 .0329 .0695 .9753 WITHIN GROUPS 15 7.1060 .4737 TOTAL 18 7.2048 Variable R3TOT By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES .1585 8.6072 8.7658 MEAN SQUARES .0528 .5738 F F RATIO PROB. .0921 .9633 N/Cl RATIOS - RUNS 3 & 4 ANALYSIS OF VARIANCE 134 Variable R40F By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 12 15 SUM OF SQUARES .6314 3.6939 4.3253 MEAN SQUARES .2105 .3078 F F RATIO PROB. .6837 .5789 Variable R4FD By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES .0143 4.2516 4.2660 MEAN SQUARES .0048 .2834 F F RATIO PROB. .0168 .9969 Variable R4T0T By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES .0124 4.7776 4.7900 MEAN SQUARES .0041 .3185 F F RATIO PROB. .0130 .9979 135 N/Cl RATIOS - RUNS 3 S. 4 ANALYSIS OF VARIANCE Variable GRTOT By Variable WT SOURCE BETWEEN GROUPS WITHIN GROUPS TOTAL D.F. 3 15 18 SUM OF SQUARES .0488 6.2858 6.3346 MEAN SQUARES .0163 .4191 F F RATIO PROB. .0388 .9894 APPENDIX B. Summary of ANCOVA Results CHLORIDE CONCENTRATIONS - RUNS 3 & 4 * * * * * * * ' A N A L Y S I S O F C 0 V A R I A N C E . . . T e s t s o f S i g n i f i c a n c e f o r R30F u s i n g UNIQUE Sums o f Squares S o u r c e o f V a r i a t i o n Sum of Squares DF Mean S q u a r e F S i g . o f F WITHIN CELLS 37.76820 12 3. 14735 Regres s ion 18.90392 1 18. 90392 6 .00630 .031 CONSTANT 74 . 20408 1 74. 20408 23 .57669 .000 WT 4.22417 3 1 . 40806 .44738 . 724 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R30F RUN 3 OVERFLOW C o v a r i a t e B Be ta S t d . E r r . T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL LGKS 1.4362462128 .5775522070 .58604 2.45078 .031 .15938 2.71311 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R30F RUN 3 OVERFLOW F a c t o r Code Obs. Mean A d j . Mean E s t . Mean Raw Res i d . S t d . Res i d . WT 15 CM WT 7. 58177 6 .83113 7 .58177 .00000 .00000 WT 35 CM WT 7. 48086 8 .00990 7 .48086 .00000 .00000 WT 55 CM WT 6. 54492 6 .75297 6 .54492 .00000 .00000 WT 75 CM WT 7 . 40312 7 .28442 7 .40312 .00000 .00000 CHLORIDE CONCENTRATIONS - RUNS 3 & 4 • ' " • • • " A N A L Y S I S O F C 0 V A R I A N C E ' * ' * * * * * * * * * * * * * T e s t s o f S i g n i f i c a n c e f o r R3FD u s i n g UNIQUE Sums of Squares S o u r c e o f V a r i a t i o n Sum of Squares DF Mean S q u a r e F S i g . o f F WITHIN CELLS 88.33633 14 6 .30974 Regres s ion 7.08224 1 7 .08224 1 . 12243 . 307 CONSTANT 221 .99518 1 221.99518 35. 18295 .000 WT 3.79528 3 1 .26509 . 20050 . 894 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R3FD RUN 3 FINAL DRAINAGE C o v a r i a t e B Be ta S t d . E r r . T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL LGKS .7646071294 .2724387165 .72170 1.05945 .307 - .78329 2.31251 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R3FD RUN 3 FINAL DRAINAGE F a c t o r WT WT WT WT Code 15 35 CM WT CM WT 55 CM WT 75 CM WT Obs. Mean 9.71140 9.27724 8.73004 9.89574 A d j . Mean 9.41139 9.65850 8.67241 9.81212 E s t . Mean 9 .71140 9 . 27724 8. 73004 9 . 89574 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .-00000 .00000 .00000 CHLORIDE CONCENTRATIONS - RUNS 3 & 4 • • ' • A N A L Y S I S O F C O V A R I A N C E ' ' * * * * * * * * * * * * * ' ' T e s t s of S i g n i f i c a n c e f o r R3TOT u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 50.52907 14 3.60922 Regress ion 7.64944 1 7.64944 2. . 11942 . 167 CONSTANT 188.20025 1 188.20025 52. .14431 .000 WT 8.77387 3 2.92462 .81032 . 509 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R3T0T RUN 3 TOTAL C o v a r i a t e B Beta S t d . E r r . LGKS .7756414806 .3626046783 .53279 T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL 1.45582 .167 - .36707 1.91835 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R3T0T RUN 3 TOTAL F a c t o r Code Obs. Mean WT WT WT WT 15 CM WT 35 CM WT 55 CM WT 75 CM WT 7.89925 8.74140 8.24320 9.43640 A d j . Mean 7.58430 9.15788 8. 17413 9.34096 E s t . Mean 7.89925 8.74140 8.24320 9.43640 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 CHLORIDE CONCENTRATIONS - RUNS 3 & 4 * * A N U V S I S O F C 0 V A R I A N C E * * * * * * * * * * * * * * * * T e s t s of S i g n i f i c a n c e f o r R40F u s i n g UNIQUE Sums o f Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 70.88172 11 6.44379 Regress ion .28572 1 .28572 .04434 .837 CONSTANT 294.03653 1 294.03653 45.63097 .000 WT 5.01405 3 1.67135 . 25937 . 853 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e .. R40F RUN 4 OVERFLOW C o v a r i a t e B Beta S t d . E r r . T - V a l u e S i g . o f T Lower 9 5% CL Upper 95% CL LGKS .1773484158 .0633620165 .84223 .21057 .837 1 .67638 2.03107 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R40F RUN 4 OVERFLOW F a c t o r Code Obs. Mean Adj Mean Est . Mean Raw Res i d . S td . Res i d. WT 15 CM WT 10. 21265 10 .11633 10.21265 .00000 .00000 WT 35 CM WT 10. 88120 10 . 94289 10.88120 .00000 .00000 WT 55 CM WT 9. 41285 9 .43491 9"41285 .00000 .00000 WT 75 CM WT 10. 25440 10 . 25059 10.25440 .00000 .00000 CHLORIDE CONCENTRATIONS - RUNS 3 & 4 * * * * * * * A N A L Y S I S O F C 0 V A R I A N C E * * * * * * * * * * * * * * * * T e s t s of S i g n i f i c a n c e f o r R4FD u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum o f Squares DF Mean Square F S i g . o f F WITHIN CELLS 118.66541 14 8.47610 R e g r e s s i o n .70422 1 .70422 .08308 . 777 CONSTANT 535.12845 1 535.12845 63.13380 .000 WT 3.65163 3 1.21721 . 14360 .932 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term -Dependent v a r i a b l e . . R4FD RUN 4 FINAL DRAINAGE C o v a r i a t e B Beta S t d . E r r . T - V a l u e S i g . o f T Lower 95% CL Upper 9 5% CL L'GKS -.241 1058673 -.0768082332 .83647 - .28824 . 777 2.03516 1 . 55295 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R4FD RUN 4 FINAL DRAINAGE F a c t o r Code Obs. Mean Adj . Mean Es t . Mean Raw R e s i d . S t d . Res i d . WT 15 CM WT 12.45532 12 . 54992 12 .45532 .00000 .00000 WT 35 CM WT 12.67460 12 . 55438 12 .67460 .00000 .00000 WT 55 CM WT ' 1 1 . 4 7 2 8 0 11 .49097 1 1 .47280 .00000 .00000 WT 75 CM WT 12.01518 12 .04155 12 .01518 .00000 .00000 CHLORIDE CONCENTRATIONS - RUNS 3 & 4 A N A L Y S I S 0 F C O V A R I A N C E * * * * * * T e s t s of S i g n i f i c a n c e f o r R4T0T u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 81 .43401 14 5.81672 Regres s ion .95423 1 .95423 . 16405 .692 CONSTANT 499 .69361 1 499.69361 85.90650 .000 WT 4.41537 3 1 .47179 .25303 .858 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R4T0T RUN 4 TOTAL Covar i a t e B Beta S t d . E r r . T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL LGKS . 2739515593 1076204502 .67637 .40503 .692 1 . 72462 1 . 17672 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R4T0T RUN 4 TOTAL F a c t o r Code WT WT WT WT 15 CM WT 35 CM WT 55 CM WT 75 CM WT Obs. Mean 10.55925 1 1.71080 11.18740 12.01320 A d j . Mean 10.67049 11.56370 11.21180 12.04691 E s t . Mean 10.55925 11.71080 11.18740 12.01320 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 CHLORIDE CONCENTRATIONS - RUNS 3 & 4 * * * * * * * * * * * * * * * " ' " A N A L Y S I S O F C 0 V A R I A N C E * * * * * * * * * * * * * * * * T e s t s of S i g n i f i c a n c e f o r GRTOT u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 62.32541 14 4 .45182 R e g r e s s i o n 1.04447 1 1 .04447 . 23462 .636 CONSTANT 319.49472 1 319 .49472 7 1 .76729 .000 WT 6.72694 3 2 .24231 . 50369 .686 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . GRTOT GRAND TOTAL C o v a r i a t e B Beta S t d . E r r . T - V a l u e S i g . o f T Lower 95% CL Upper 9 5% CL LGKS .2866118030 .1283826334 .59172 .48437 .636 .98250 1 .55572 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . GRTOT GRAND TOTAL F a c t o r Code WT WT WT WT 15 CM WT 35 CM WT 55 CM WT 75 CM WT Obs. Mean 9.20075 10.20720 9.66200 10.72600 A d j . Mean 9.08437 10.36110 9.63648 10.69073 E s t . Mean 9 . 20075 10.20720 9.66200 10.72600 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 NITRATE PLUS NITRITE CONCENTRATIONS - RUNS 3 & 4 * * * * * * * * * * * * * * * ' - ' A N A L Y S I S O F C O V A R I A N C E T e s t s of S i g n i f i c a n c e f o r R30F u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 236.33330 12 19 .69444 Regress ion 222 .47302 1 222 .47302 11.29623 .006 CONSTANT 61.18280 1 61 .18280 3.10660 . 103 WT 125.33894 3 41 .77965 2 . 121 3.9 .151 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R30F RUN 3 OVERFLOW Covar i a t e B Beta S t d . E r r . T - V a l u e S i g . o f T Lower 95% CL Upper 9 5% CL LGKS 4.9271023360 .6963442279 1 .46597 3.36099 .006 1 .73303 8.12117 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R30F RUN 3 OVERFLOW F a c t o r Code WT WT WT WT 15 35 55 75 CM WT CM WT CM WT CM WT Obs. Mean 4.29712 5.96844 1.68855 1.21880 A d j . Mean 1.72199 7.78335 2.40226 .81159 E s t . Mean 4.29712 5.96844 1 .68855 1 .21880 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 NITRATE PLUS NITRITE CONCENTRATIONS - RUNS 3 & 4 . . . . . . . . . . . . . . . . . ' A N A L Y S I S O F C 0 V A R I A N C E * * * * * * * * * T e s t s of S i g n i f i c a n c e f o r R3FD u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 210.03360 14 15 .00240 R e g r e s s i o n 339.79810 1 339 .79810 22.64958 .000 CONSTANT 49.24364 1 49 .24364 3.28238 .092 WT 65.71704 3 21 .90568 1.46015 . 268 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R3FD RUN 3 FINAL DRAINAGE C o v a r i a t e B Beta S t d . E r r . T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL LGKS 5.2961898711 .7861321984 1.11284 4 .75916 .000 2.90938 7.68300 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R3FD RUN 3 FINAL DRAINAGE F a c t o r Code WT WT WT WT 15 CM WT 35 CM WT 55 CM WT 75 CM WT Obs. Mean 4.83567 5.50468 4.44180 6.24220 A d j . Mean 2.75762 8.14553 4.04263 5.66296 E s t . Mean 4.83567 5.50468 4.44180 6 . 24220 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 NITRATE PLUS NITRITE CONCENTRATIONS - RUNS 3 & 4 • * * * * * ' ' * * * * * ' A N A L Y S I S 0 F C O V A R I A N C E T e s t s o f S i g n i f i c a n c e f o r R3T0T u s ing UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 269.05848 14 19 .21846 R e g r e s s i o n 369.84235 1 369 .84235 19.24412 .001 CONSTANT 66 .00612 1 66 .006 12 3.43452 .085 WT 99.22403 3 33 .07468 1 . 72098 . 208 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R3T0T RUN 3 TOTAL C o v a r i a t e B Beta S t d . E r r . T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL LGKS 5 . 5253703638 . 7608369477 1 .25954 4.38681 .001 2 . 82392 8.22682 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R3T0T RUN 3 TOTAL F a c t o r Code WT WT WT WT 15 CM WT 35 CM WT 55 CM WT 75 CM WT Obs. Mean 4.32907 5.97430 3.92462 5.96036 A d j . Mean 2. 16109 8.72942 3.50818 5.35606 E s t . Mean 4.32907 5.97430 3.92462 5 . 96036 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 NITRATE PLUS NITRITE CONCENTRATIONS - RUNS 3 & 4 * * * * * * * * * * A N A L Y S I S O F C 0 V A R I A N C E * * * * » * a * • * * * * * * * » T e s t s of S i g n i f i c a n c e f o r R40F u s i n g UNIQUE Sums of Squares Source o f V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 141.66063 11 12 . 87824 R e g r e s s i o n 210.92603 1 210 .92603 16 . 37848 .002 CONSTANT 44.75581 1 44 .75581 3.47531 .089 WT 76.08511 3 25 .36170 1 .96935 . 177 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R40F RUN 4 OVERFLOW C o v a r i a t e B Beta S t d . E r r . T -Va lue S i g . o f T Lower 95% CL Upper 95% CL LGKS 4.8186206203 .7734498309 1.19065 4 .04703 .002 2. 19801 7.43923 A d j u s t e d and E s t i m a t e d Means Var i a b l e . . R40F F a c t o r WT WT WT WT Code 15 CM WT 35 CM WT 55 CM WT 75 CM WT RUN 4 OVERFLOW Obs. Mean 5.53220 5.46470 2.35425 1.34307 Ad j . Mean 2.91509 7. 14098 2.95357 1.23967 E s t . Mean 5.53220 5.46470 2.35425 1.34307 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 NITRATE PLUS NITRITE CONCENTRATIONS - RUNS 3 & 4 A N A L Y S I S 0 F C O V A R I A N C E T e s t s of S i g n i f i c a n c e f o r R4FD u s i n g UNIQUE Sums of Squares Source o f V a r i a t i o n WITHIN CELLS Regress ion CONSTANT WT Sum of Squares DF Mean Square F S i g . o f F 192.43485 14 13.74535 341 . 14151 1 341 .14151 24. .81869 .000 29.83986 1 29 .83986 2. .17091 . 163 44.28009 3 14 . 76003 1 .07382 . 392 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R4FD RUN 4 FINAL DRAINAGE C o v a r i a t e B Beta S t d . E r r . T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL LGKS 5.3066488933 .7995930180 1.06520 4.98184 .000 3.02202 7.59127 A d j u s t e d and E s t i m a t e d Means Var i a b l e . . R4FD F a c t o r WT WT WT WT Code 15 CM WT 35 CM WT 55 CM WT 75 CM WT RUN 4 FINAL DRAINAGE Obs. Mean 6.51252 6.11058 6.03150 5.82006 A d j . Mean 4.43036 8.75664 5.63155 5.23968 E s t . Mean .51252 . 11058 .03150 . 82006 Raw R e s i d . .00000 .00000 .00000 : o o o o o S t d . R e s i d . .00000 .00000 .00000 .00000 NITRATE PLUS NITRITE CONCENTRATIONS - RUNS 3 & 4 . . . . . . . . . . . . . . . . . . A N A L Y S I S O F C O V A R I A N C E * * T e s t s o f S i g n i f i c a n c e f o r R4TOT u s i n g UNIQUE Sums of Squares * * * * Source o f V a r i a t i o n Sum of Squares DF Mean Square F WITHIN CELLS 172.22747 14 12 . 30196 Regress ion 331.64531 1 331 .64531 26 . 95873 CONSTANT 31 . 75148 1 31 . 75148 2. .58101 WT 43.98663 3 14.66221 1 . 19186 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . . R4T0T RUN 4 TOTAL S i g . o f F .000 . 130 . 349 Covar i a t e LGKS 5.2322682861 Beta S t d . E r r . T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL .8112906701 1.00772 5.19218 .000 3.07092 7.39361 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R4T0T RUN 4 TOTAL F a c t o r Code Obs. Mean WT WT WT WT 15 CM WT 35 CM WT 55 CM WT 75 CM WT 5.66910 5.71038 5.98352 6.22790 A d j . Mean 3.61612 8.31935 5. 58917 5.65566 E s t . Mean 5.66910 5.71038 5.98352 6.22790 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 NITRATE PLUS NITRITE CONCENTRATIONS - RUNS 3 & 4 * * * * * * * * * * * ' • • ' " ' • A N A L Y S I S O F C O V A R I A N C E ' * * * * * * * * * * * * * * * T e s t s of S i g n i f i c a n c e f o r GRTOT u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 205.06203 14 14.64729 Regress ion 351 .28364 1 351.28364 23.98284 .000 CONSTANT 48.08890 1 48.08890 3.28313 .091 WT 67 . 57376 3 22.52459 1.53780 . 249 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . GRTOT GRAND TOTAL C o v a r i a t e B Beta S t d . E r r . LGKS 5.3849544476 .7946147168 1.09959 T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL 4.89723 .000 3.02656 7.74335 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . GRTOT GRAND TOTAL F a c t o r Code Obs. Mean WT WT WT WT CM WT CM WT 15 35 55 CM WT 75 CM WT 98680 83224 93032 06280 Adj 2. 8 . 4 . 5. Mean 8739 1 51734 52446 47386 E s t . Mean 4.98680 5.83224 4 .93032 6.06280 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 N/Cl RATIOS - RUNS 3 & 4 * * * * * * * * * * * * * A N A L Y S I S O F C O V A R I A N C E * * * * * * * * * * * * * * * * T e s t s o f S i g n i f i c a n c e f o r R30F u s i n g UNIQUE Sums of Squares Source o f V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 3.46462 12 . 28872 Regress ion 2.83580 1 2.83580 9 . 82203 .009 CONSTANT .69105 1 .69105 2. 39352 . 148 WT 1 . 36078 3 . 45359 1 . 57106 . 248 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R30F RUN 3 OVERFLOW C o v a r i a t e B Beta S t d . E r r . LGKS .5562768334 .6708927209 .17750 T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL 3.13401 .009 .16954 .94301 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R30F RUN 3 OVERFLOW F a c t o r Code WT WT WT WT 15 CM WT 35 CM WT 55 CM WT 75 CM WT Obs. Mean . 59500 . 65020 .21225 . 14950 A d j . Mean . 30426 . 855 1 1 . 29283 . 10353 E s t . Mean . 59500 '.65020 .21225 .14950 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 N/Cl RATIOS - RUNS 3 & 4 . . . . . . . . . . . . . . . . . . A N A L Y s i S O F C 0 V A R I A N C E * * * * * * * * * * * * * * * * * T e s t s o f S i g n i f i c a n c e f o r R3FD u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 3.29556 14 . 23540 R e g r e s s i o n 3.81047 1 3.81047 16.18741 .001 CONSTANT .55272 1 . 55272 2.34804 . 148 WT .57939 3 .19313 . 82044 . 504 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R3FD RUN 3 FINAL DRAINAGE C o v a r i a t e B Beta S t d . E r r . T -Va lue S i g . o f T Lower 95% CL Upper 95% CL LGKS .5608446731 .7322775902 .13940 4 .02336 .001 .26187 .85982 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R3FD F a c t o r WT WT WT WT Code 15 CM WT 35 CM WT 55 CM WT 75 CM WT RUN 3 FINAL DRAINAGE Obs. Mean . 56525 .54900 .45660 .65480 A d j . Mean .34519 .82865 .41433 .59346 E s t . Mean .56525 .54900 .45660 .65480 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 N/Cl RATIOS - RUNS 3 & 4 * * * * * * * * * * * * * * * . * * * A N A L Y S I S O F C O V A R I A N C E * * * * * * * * * * * * * * * * T e s t s of S i g n i f i c a n c e f o r R3T0T u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 4.18404 14 .29886 Regress ion 4.42320 1 4.42320 14.80025 .002 CONSTANT .73065 1 . 73065 2.44481 . 140 WT .9361 1 3 .31204 1.04409 .404 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R3T0T RUN 3 TOTAL C o v a r i a t e B Beta S t d . E r r . T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL LGKS .6042562894 .7168633365 .15707 3.84711 .002 .26738 .94113 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R3T0T RUN 3 TOTAL F a c t o r Code Obs. Mean WT WT WT WT 15 CM WT 35 CM WT 55 CM WT 75 CM WT .58125 .63500 .42160 .64320 A d j . Mean .34416 .93630 .37606 .5771 1 E s t . Mean .58125 .63500 .42160 .64320 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 N/Cl RATIOS - RUNS 3 & 4 • A N A L Y S I S 0 F C O V A R I A N C E T e s t s of S i g n i f i c a n c e f o r R40F u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 1 .62965 1 1 .14815 Regress i on 2.06423 1 2.06423 1 3 .93339 .003 CONSTANT .43514 1 .43514 2 .93715 .115 WT .78355 3 . 261 18 1 . 76295 .212 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R40F RUN 4 OVERFLOW C o v a r i a t e B Beta S t d . E r r . LGKS .4766909204 .7475456206 .12771 T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL 3.73275 .003 .19561 .75777 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R40F RUN 4 OVERFLOW F a c t o r Code WT WT WT WT 15 35 55 75 CM WT CM WT CM WT CM WT Obs. Mean .59050 .54200 .21250 .11367 A d j . Mean .33160 .70783 .27179 . 10344 E s t . Mean .59050 . 54200 .21250 . 1 1367 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 N/Cl RATIOS - RUNS 3 & 4 * * * * * * * * * * * * * * * * * " A N A L Y S I S O F C O V A R I A N C E * * ' . . . . . . . * * * * * * T e s t s of S i g n i f i c a n c e f o r R4FD u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 1.39250 14 .09946 Regress ion 2.85914 1 2.859 14 28.74534 .000 CONSTANT .33943 1 .33943 3.41259 .086 WT .42863 3 .14288 1 .43646 .274 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R4FD RUN 4 FINAL DRAINAGE C o v a r i a t e B Beta S t d . E r r . T - V a l u e LGKS .4858146367 .8200481115 .09061 5.36147 S i g . o f T Lower 95% CL Upper 95% CL .000 .29147 .68016 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R4FD RUN 4 FINAL DRAINAGE F a c t o r Code WT WT WT WT 15 CM WT 35 CM WT 55 CM WT 75 CM WT Obs. Mean . 54950 .53580 .48420 .49180 A d j . Mean .35888 .77804 .44758 .43867 E s t . Mean .54950 .53580 .48420 .49180 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 N/Cl RATIOS - RUNS 3 & 4 A N A L Y S I S O F C O V A R I A N C E T e s t s of S i g n i f i c a n c e f o r R4T0T u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 1 .82243 14 .13017 Regress ion 2.95517 1 2.95517 22. 70178 .000 CONSTANT .31290 1 .31290 2.40373 . 143 WT .38856 3 .12952 .99498 .424 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . R4T0T RUN 4 TOTAL C o v a r i a t e B Be ta S t d . E r r . T - V a l u e S i g . of T Lower 95% CL Upper 95% CL LGKS 4939058517 .7864776377 .10366 4.76464 .000 .27158 .71624 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . R4T0T RUN 4 TOTAL WT WT WT WT F a c t o r 15 35 55 75 Code CM WT CM WT CM WT CM WT Obs. Mean .58100 . 54420 .50700 . 53500 A d j . Mean . 38721 .79048 . 46977 .48098 E s t . Mean .58100 .54420 . 50700 . 53500 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 cn N/Cl RATIOS - RUNS 3 & 4 * * * * * * * " * * * * * * * * * * A N A L Y S I S O F C O V A R I A N C E * * * * * * * * * * ' * * * * * * T e s t s o f S i g n i f i c a n c e f o r GRTOT u s i n g UNIQUE Sums of Squares Source of V a r i a t i o n Sum of Squares DF Mean Square F S i g . o f F WITHIN CELLS 2.68966 14 .19212 Regres s ion 3.59614 1 3.59614 18 .71828 .001 CONSTANT .48488 1 .48488 2 . 52385 . 134 WT .60174 3 . 20058 1 . .04404 . 404 R e g r e s s i o n a n a l y s i s f o r WITHIN CELLS e r r o r term Dependent v a r i a b l e . . GRTOT GRAND TOTAL C o v a r i a t e B Beta S t d . E r r . LGKS .5448426295 .7563760293 .12593 T - V a l u e S i g . o f T Lower 95% CL Upper 95% CL 4.32646 .001 .27474 .81494 A d j u s t e d and E s t i m a t e d Means V a r i a b l e . . GRTOT GRAND TOTAL WT WT WT WT F a c t o r Code 15 CM WT 35 CM WT CM WT CM WT 55 75 Obs. Mean .58025 .58740 .46860 .58220 A d j . Mean .36647 .85908 .42754 .52261 E s t . Mean . 58025 .58740 .46860 .58220 Raw R e s i d . .00000 .00000 .00000 .00000 S t d . R e s i d . .00000 .00000 .00000 .00000 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058894/manifest

Comment

Related Items