In agriculture-dominated areas, water pollution resulting from nutrients migrating from farms to water bodies is a major concern. The migration is further exacerbated by traditional tile drain known as Uncontrolled Tile Drainage (UCTD), which removes excess water from areas to keep the water table low enough for crops to grow. UCTD, commonly used in Ontario, Canada, is believed to contribute to water quality issues, whereas Controlled Tile Drainage (CTD) is an alternative technique in which a structure controls the outlets of the drains so that water only leaves a field when the water table level exceeds a desired threshold. Considered as a Best Management Practice (BMP), CTD has been documented as an efficient practice preventing nutrients from migrating out of agricultural fields. This thesis aims to improve our understanding of the environmental benefits of replacing UCTD with CTD. Three significant contributions were achieved. The first contribution of the thesis is the improvements of the algorithm for calculating nitrates in tile flows in the Soil and Water Assessment Tool (SWAT) model. Researchers have simulated CTD by dynamically changing tile depth to mimic the operation of outlet structure gates, but it has been demonstrated that doing so results in inaccuracies, and so the algorithm in the model has been improved subsequently. The current author proposed and tested a new algorithm for calculating nitrates in tile flows that better represents the dynamics of water and nutrients in soil layers for the SWAT model. A model for the South Nation watershed, located in Ontario, Canada, was then developed and successfully calibrated using the improved SWAT model. The second contribution was the extension of the SWAT model to simulate riverine hydraulic and water quality processes by coupling it with the QUAL2Kw model. In this thesis, a procedure is developed to couple the SWAT model and the QUAL2Kw model to enable continuous simulations of 13 water quality parameters in the South Nation River system. The coupled model was calibrated and verified at various observed locations along the river during the five seasons of growth from 2006 to 2010. The simulation results suggested that CTD also improved the water quality of the river by lowering biologically available N levels of NO2--N, and NO3--N, thereby impeding phytoplankton growth in the river. The third contribution is the verification of the benefits of replacing UCTD with CTD in the future climates. The confirmation was done using the SWAT model alone, and then the coupled SWAT/QUAL2K models, using a matrix of climate change experiments performed with several Global Climate Models and Regional Climate Models. The results suggest that nutrient loading from the watershed will increase in the 2021–2050 period compared to the 1985–2006 period. Thus, pollution from agricultural fields will worsen with the current UCTD approach, while the results also show that CTD would perform effectively and stably in future climate scenarios and could counterbalance the effects of climate change on water quality. To the author’s knowledge, this study is the first attempt to date to assess the environmental effects of CTD on a watershed and river by coupling the SWAT and QUAL2Kw models. The findings expand the current scope of knowledge on the environmental effects of CTD on watersheds and rivers under current and future climate change regimes. Long periods of continuous simulation and a matrix of climate change scenarios also make this study stand out from other studies. It laid a foundation for future investigations.

Agricultural management is essential for achieving optimum crop production and maintaining soil quality. Soil microorganisms are responsible for nutrient cycling and are an important consideration for effective soil management. The overall goal of the present research was to better understand microbial communities in agricultural soils as they relate to soil management practices. For this, we evaluated the differential impact of two contrasting drainage practices on microbial community composition and characterized active denitrifiers from selected agricultural sites. Field drainage is important for crop growth in arable soils. Controlled and uncontrolled tile drainage practices maintain water in the field or fully drain it, respectively. Because soil water content influences nutrient concentration, moisture, and oxygen availability, the effects of these two disparate practices on microbial community composition was compared in paired fields that had diverse land management histories. Libraries of the 16S rRNA gene were generated from DNA from 168 soil samples collected from eight fields during the 2012 growing season. Paired-end sequencing using next-generation sequencing was followed by read assembly and multivariate statistical analyses. Results showed that drainage practice exerted no measureable effect on the bacterial communities. However, bacterial communities were impacted by plant cultivar and applied fertilizer, in addition to sampled soil depth. Indicator species were only recovered for depth; plant cultivar or applied fertilizer type had no strong and specific indicator species. Among indicator species for soil depth (30-90 cm) were Chloroflexi (Anaerolineae), Betaproteobacteria (Janthinobacterium, Herminiimonas, Rhodoferax, Polaromonas), Deltaproteobacteria (Anaeromyxobacter, Geobacter), Alphaproteobacteria (Novosphingobium, Rhodobacter), and Actinobacteria (Promicromonospora). Denitrification in agricultural fields transforms nitrogen applied as fertilizer, reduces crop production, and emits N2O, which is a potent greenhouse gas. Agriculture is the highest anthropogenic source of N2O, which underlines the importance of understanding the microbiology of denitrification for reducing greenhouse gas emissions by altered management practices. Existing denitrifier probes and primers are biased due to their development based mostly on sequence information from cultured denitrifiers. To circumvent this limitation, this study investigated active and uncultivated denitrifiers from two agricultural sites in Ottawa, Ontario. Using DNA stable-isotope probing, we enriched nucleic acids from active soil denitrifiers by exposing intact replicate soil cores to NO3- and 13C6-glucose under anoxic conditions using flow-through reactors, with parallel native substrate controls. Spectrophotometric chemistry assays and gas chromatography confirmed active NO3- depletion and N2O production, respectively. Duplicate flow-through reactors were sacrificed after one and four week incubation periods to assess temporal changes due to food web dynamics. Soil DNA was extracted and processed by density gradient ultracentrifugation, followed by fractionation to separate DNA contributed by active denitrifiers (i.e., “heavy” DNA) from that of the background community (i.e., “light” DNA). Light and heavy DNA samples were analyzed by paired-end sequencing of 16S rRNA genes using next-generation sequencing. Multivariate statistics of assembled 16S rRNA genes confirmed unique taxonomic representation in heavy fractions from flow-through reactors fed 13C6-glucose, which exceeded any site-specific or temporal shifts in putative denitrifiers. Based on high relative abundance in heavy DNA, labelled taxa affiliated with the Betaproteobacteria (71%; Janthinobacterium, Acidovorax, Azoarcus, Dechloromonas), Alphaproteobacteria (8%; Rhizobium), Gammaproteobacteria (4%; Pseudomonas), and Actinobacteria (4%; Streptomycetaceae). Metagenomic DNA from the original soil and recovered heavy fractions were subjected to next-generation sequencing and the results demonstrated enrichment of denitrification genes with taxonomic affiliations to Brucella, Ralstonia, and Chromobacterium in heavy fractions of flow-through reactors fed 13C6-glucose. The vast majority of heavy-DNA-associated nitrite-reductase reads annotated to the copper-containing form (nirK), rather than the heme-containing enzyme (nirS). Analysis of recovered nirK genes demonstrated low sequence identity across common primer-binding sites used for the detection and quantification of soil denitrifiers, indicating that these active denitrifiers would not have been detected in molecular surveys of these same soils.