Thesis (doctoral)--Swedish University of Agricultural Sciences, 2006. / Thesis documentation sheet inserted. Appendix reprints three papers and manuscripts co-authored with others. Includes bibliographical references. Also partially issued electronically via World Wide Web in PDF format; online version lacks appendix.

Thesis (Ph. D.)--Ohio State University, 2007. / Title from first page of PDF file. Includes bibliographical references (p. 115-119).

Humans have dramatically increased the amount of reactive nitrogen cycling through the biosphere. In coastal systems, excess nitrogen can lead to negative impacts. Thus, it is crucial to understand how nitrogen is cycled within, and eventually removed from, marine systems and the variables that regulate these processes. Sediment denitrification (the microbial conversion of nitrate (NO3^-) to dinitrogen (N2) gas) and water column nitrification (the two step oxidation of ammonium (NH4^+) to nitrite (NO2^-) and then nitrate (NO3^-)) rates were quantified along an in situ gradient of environmental conditions from an estuary to the continental shelf off Rhode Island, USA. Sediment net denitrification rates were directly measured over multiple seasonal cycles using the N2/Ar technique. Denitrification rates ranged from 20-75 μmol m^-2 hr^-1 (mean 44±4), indicating that this process removes ~5% of total reactive nitrogen entering the North Atlantic shelf region per year. Based on model results, these rates also represented a three-fold decrease in sediment nitrogen removal in New England continental shelf sediments over the past century. A literature review of marine water column nitrification observations were compiled to evaluate how ammonium, nitrite, and total oxidation rates vary worldwide. Rates of ammonium, nitrite, and total oxidation differed among estuary, continental shelf, and open ocean environments (p<0.05). This review highlights that as we continue to study marine "nitrification," it is necessary to consider both individual oxidation processes and environment type. Water column ammonium and nitrite oxidation rates were measured using stable isotope tracers off Rhode Island. At all study sites, nitrite oxidation rates (0-99 nM d^-1) outpaced ammonium oxidation rates (0-20 nM d^-1). These oxidation processes responded in dissimilar ways to in situ water column conditions (depth, salinity, dissolved oxygen, and pH), and these relationships varied with location. Nitrous oxide (N2O) production rates up to 10 times higher than ammonium oxidation indicated that ammonium oxidation may be underestimated if this byproduct is not measured. For the first time, the link between sediment metabolism and water column nitrification was also examined, and the results highlight the importance of benthic-pelagic coupling as controlling factor of water column ammonium and nitrite oxidation. / 2019-04-30T00:00:00Z

O principal objetivo desse trabalho foi desenvolver uma nova configuração de biofiltro aerado submerso utilizado no pós-tratamento do efluente de reator UASB, capaz de realizar nitrificação e desnitrificação em um único sistema. Em busca de bases operacionais dessa nova configuração, a pesquisa foi inicialmente conduzida com três reatores seqüenciais. Esses reatores foram dispostos nas seqüências I e II, respectivamente. Os resultados obtidos com esses sistemas seqüenciais levaram a concepção do projeto da nova configuração do biofiltro proposto (BF definitivo). Os resultados da seqüência I mostraram alto potencial para conversão de nitrogênio amoniacal e total, e também alta eficiência na remoção da matéria orgânica carbonácea. Esse sistema, porém, não apresentou potencial para desnitrificação. Na seqüência II foi realizada recirculação entre dois reatores do sistema (anóxico e aeróbio) para aumento de tal potencial. O maior potencial de desnitrificação, nessa seqüência, foi observado para razão de recirculação, Rc, igual a 2,65 e com o uso de 40%, em volume, de esgoto sanitário bruto como fonte de carbono. Sob essas condições operacionais, para concentração média afluente ao sistema igual a 33,74 mg de \'N-\'NH IND.3\'/L as concentrações efluentes médias de \'N-NH IND.3\', \'N-NO IND.2\'POT.-\' e \'N-NO IND.3\'POT.-\' foram, respectivamente, iguais a 0,16, 0,0026 e 9,72 mg/L. Os resultados do BF definitivo mostraram que a nova configuração proposta é viável como unidade de pós-tratamento de efluente de reator UASB, promovendo nitrificação e desnitrificação em um único sistema, além de alta eficiência de remoção da matéria orgânica. Em todas as fases dessa pesquisa, praticamente, todos os resultados obtidos atenderam ao padrão de lançamento de nitrogênio amoniacal estabelecido pela legislação ambiental. / The main objective of this work was developing a new configuration of submerged aerated biofilter used in UASB reactor post-treatment, with nitrification and denitrification in a single system. Searching operational bases of this new configuration, this research was initially driven by three sequential reactors. These reactors were arranged in the sequences I and II, respectively. The results obtained with these sequential systems took the conception of the project of the new configuration of the proposed biofilter (definitive BF). The results of the sequence I showed high ammoniacal and total nitrogen conversion potential, and also high carbonaceous organic matter removal efficiency. This system, however, not presented potential for denitrification. In the sequence II recirculation between two reactors (anoxic and aerated) of the system was made for such potential increasing. The higher denitrification potential, in this sequence, was observed with recirculation reason, Rc, equal to 2,65 and using 40%, in volume, of raw sanitary sewage as carbon source. Under these operational conditions, for average affluent concentration of 33,74 mg of \'N-NH IND.3\'/L the average effluent concentrations of \'N-NH IND.3\', \'N-NO IND.2\'POT.-\' e \'N-NO IND.3\'POT.-\' were, respectively, equal to 0,16, 0,0026 and 9,72 mg/L. The results of the definitive BF showed that the new proposed configuration is feasible as effluent UASB post-treatment unity, with nitrification and denitrification in a single system, besides high organic matter removal efficiency. Practically in all the phases of this research, all the results obtained attended to the launch standard of ammoniacal nitrogen established by the environmental legislation.

A presente dissertação apresenta e discute os resultados do trabalho experimental cujo objetivo foi avaliar a remoção de nitrogênio na forma de mitrato (N-NO3-) pelo processo de desnitrificação biológica em reator anóxico horizontal de leito fixo (RAHLF) contendo matrizes de espuma de poliuretano, em escala de laboratório, utilizando gás metano como fonte de carbono e único doador de elétrons adicionado ao sistema. Para concentrações iniciais de N-NO3- de 20 mg/L e 40 mg/L no substrato sintético, foi possível obter diminuição das concentrações iniciais em 85% e 50%. No entanto, os altos níveis de redução de N-NO3-, obtidos sob condições limitantes de metano, deram suporte à hipótese de que parte da remoção do N-NO3- foi realizada mediante a utilização de compostos reduzidos de enxofre ou nitrogênio, tais como: S0, HS- ou NH4+, provavelmente formados sob condições anóxicas, simultaneamente com o processo de desnitrificação. Foi possível verificar, também, a influência da relação carbono (mg/L CH4 / nitrogênio (mg/L N-NO3-) no estabelecimento das rotas metabólicas de desnitrificação predominantes, quais sejam, a redução dissimilativa do nitrogênio à amônia (RDNA) ou a desnitrificação. Adicionalmente, foram realizados ensaios em reatores tipo batelada, com o objetivo de se medir o consumo de metano. Porém, os resultados não foram satisfatórios, provavelmente em razão da diversidade microbiana presente no inóculo. Foram realizadas análises de microscopia óptica e de fluorescência, assim como de DGGE, para avaliar a diversidade e as alterações nas populações microbianas ao longo do RAHLF e do tempo de experimento. Os diferentes sistemas utilizados apresentaram limitações relacionadas à baixa solubilidade do gás metano no meio líquido, à resistência à transferência de massa da fase gasosa para a líquida e desta última para a biomassa aderida à espuma. / This study presents and discusses experimental work results conducted with the purpose of evaluating nitrate - nitrogen (N-NO3-) removal by biological denitrification process in a lab scale horizontal flow fixed bed anoxic reactor (RAHLF), using methane gas as sole carbon source and electron donor. Support media for microorganisms were polyurethane foam matrixes. For initial N-NO3- concentrations of 20 mg/L and 40 mg/L present in synthetic substrate, it was possible to obtain 85% and 50% removal respectively. These high reduction rates, obtained under limiting conditions of methane, sustained the idea of part of the N-NO3- removal being accomplished by reductive sulfur or nitrogen species utilization, such as: S0, HS- or NH4+, probably formed under anoxic conditions simultaneously to denitrification process. It was possible to verify also carbon (mg/L CH4) / nitrogen (mg/L N-NO3-) ratio effect in denitrification metabolic paths establishment, i.e. dissimilative reduction of nitrogen to ammonia or denitrification itself. In addition, batch tests where conducted with methane consumption measuring purpose. Yet, results where not satisfactory probably due to great microbial diversity present in inoculum. Optical microscopy and fluorescence exams where developed, as well as, DGGE, in order to evaluate diversity and alterations in bacterial populations as a function of reactor\'s length and time. Different systems used in experimental work presented limitations due to low methane gas solubility in bulk liquid and mass transfer resistance from gas to liquid phase and from this to fixed biomass.

La dénitrification autotrophe à l’aide de composés réduit de soufre est une approche intéressante pour le traitement biologique des contaminations azotées et des effluents pauvre en matière organique. La dénitrification autotrophe utilise des composés inorganiques comme sources d'énergie et de carbone. L'absence de matière organique élimine le besoin de post-traitements pour éliminer l'excès de carbone organique et limite la formation sous-produits d’oxydation dans le cadre de la production d’eau potable. Les eaux usées provenant des industries métallurgiques et minières ont généralement un faible pH, des températures basses et des concentrations élevées en métaux lourds. L'élimination biologique de l'azote est un défi parce que les bactéries dénitrifiantes prospèrent habituellement à pH neutre et à températures ambiantes (20-30 °C).Le but de cette thèse était de développer un procédé robuste de dénitrification à base de soufre dans des bioréacteurs à pH acide, températures psychrophiles (< 20 °C) et concentrations élevées en nickel. Le procédé a été optimisé au préalable avec des essais biologiques étudiant l'influence de la source de soufre (S2O32-, S0 biogénique et le synthétisé chimiquement), de la taille des particules de S0 (poudre et lentilles), de la culture dénitrifiante (cultures pures et mixtes de Thiobacillus) et de la température (6-30 °C) sur la cinétique de la dénitrification. L'utilisation de S2O32- et d’une culture pure de T. denitrificans ont permis d’atteindre des rendements de dénitrification les plus élevés. Le soufre élémentaire biogénique a été testé pour la première fois comme donneur d'électrons pour la dénitrification, montrant des taux de dénitrification 1.7 fois plus élevés que ceux obtenue avec de la poudre de S0 synthétisé chimiquement. Les taux de la dénitrification avec le S2O32- augmentent exponentiellement avec la température et les calculs avec l'équation d'Arrhenius donnent une énergie d'activation apparente Ea de 76.6 kJ/mol.Deux réacteurs à lit fluidisé (FBR) ont été utilisés pour étudier la dénitrification avec S2O32- à différents pH (5.25-7.00) et températures décroissantes (3-20 °C). Des rendements de dénitrification > 99% ont été observés pour eaux usés présentant des pH compris entre 5.75 et 5.30. L'addition d'une unité de carbonatation fournissant au biofilm du CO2 comme source de carbone supplémentaire, permettant une dénitrification complète à un pH de 4.75. Dans le même FBR, des taux de charge d'azote élevés (jusqu'à 3,3 kg N-NO3-/m3 d) avec le thiosulfate ont été maintenu à des températures aussi basses que 3 °C. L'impact de deux composés du Nickel (NiEDTA2- et NiCl2) sur la dénitrification à base de soufre a été étudiée dans deux FBR en parallèle à 20 (± 2) °C et des concentrations de nickel variant dans la gamme de 5-200 mg Ni/L. Dans des bioessais discontinues, 25-100 mg Ni/L de NiCl2 ont inhibée l'élimination de NO3- de 7-16%, alors qu'aucune inhibition n'a été observée avec NiEDTA2-. L'EDTA non complexée a inhibée la dénitrification à des concentrations supérieures à 100 mg/L. Les deux composés de Ni ont montré aucun effet négatif sur la dénitrification en FBR aux concentrations testées. Le bilan massique du nickel, la caractérisation de la phase solide et la modélisation thermodynamique ont révélé que des précipités de nickel ont été principalement éliminés avec l'effluent. Les phosphates, sulfures et oxydes de nickel ont été déterminés comme les principaux précipités de nickel et étaient principalement amorphe.Les FBRs se sont révélés être bioprocédés robustes pour l'élimination de l'azote à pH acide, pour des températures psychrophiles et des concentrations élevées de nickel. Les résultats de cette étude sont d'un grand intérêt pour le traitement des eaux souterraines et minières contaminés par les nitrates dans les régions froides du monde et également pour les eaux usées industrielles acides et chargées en métaux lourds / Autotrophic denitrification driven by reduced sulfur compounds is a promising and cost-effective biological nitrogen removal process, recommended for the treatment of organic-deficient waters, e.g. groundwater and several industrial wastewaters. Autotrophic denitrifiers utilize inorganic compounds as sources of energy and carbon. The lack of organics eliminates the need of post-treatments to remove excess organic carbon and limits the formation of harmful organic byproducts (e.g. trihalomethanes, THM), resulting in a clean and safe treatment also for drinking water. Wastewaters from mining and metal-finishing industry commonly feature low pH and temperatures as well as high heavy metal concentrations. Nitrogen removal from these waters is a technical challenge, since denitrifying bacteria usually thrives at circumneutral pH and ambient temperatures (20-30°C).The aim of this study was to develop a robust and efficient sulfur-based denitrification bioreactor process able to tolerate acidic pH, psychrophilic temperatures (< 20°C) and high nickel concentrations. The process was preliminary optimized in batch bioassays investigating the influence of sulfur source, i.e. thiosulfate (S2O32-) and biogenic and chemically synthesized elemental sulfur (S0), S0 particle size (powder and lentils), denitrifying culture (pure and mixed cultures of Thiobacillus) and temperature (6-30°C) on denitrification kinetics. The use of S2O32- and a pure culture of Thiobacillus denitrificans resulted in the highest denitrification rates. Biogenic S0 was tested for the first time as electron donor for autotrophic denitrification, showing 1.7-fold faster NO3- removal than that achieved with chemically synthesized S0 powder. The rates of thiosulfate-driven denitrification exponentially increased with temperature, being modeled according to the Arrhenius equation with an apparent activation energy Ea of 76.6 kJ/mol and a temperature coefficient Q10 of 3.0.Fluidized-bed reactors (FBRs) were used to investigate continuous thiosulfate-driven denitrification under decreasing feed pH (5.25-7.00) and temperatures (3-20°C). Denitrification efficiencies > 99% were observed at feed and effluent pH as low as 5.75 and 5.30, respectively. At lower feed pH values, the denitrification activity rapidly decreased due to an inorganic carbon deficiency. The addition of a carbonation unit providing CO2 as supplemental carbon source to the FBR biofilm allowed complete denitrification even at a pH of 4.75. In the same FBR, high-rate (up to 3.3 kg N-NO3-/m3 d) thiosulfate-driven denitrification was maintained at temperatures as low as 3°C. The impact of two Ni compounds, i.e. NiEDTA2- and NiCl2, on sulfur-based denitrification was investigated in a parallel FBR at 20 (±2)°C and feed Ni concentrations in the range of 5-200 mg Ni/L. Preliminary batch bioassays were carried out to assess Ni and free EDTA toxicity on sulfur-based denitrification. In batch bioassays, 25-100 mg Ni/L of NiCl2 inhibited NO3- removal by 7-16%, whereas no inhibition was observed with NiEDTA2-. Free EDTA inhibited sulfur-based denitrification at concentrations exceeding 100 mg/L. Both Ni compounds showed no detrimental effects on sulfur-based denitrification in FBR at the tested concentrations. Nickel mass balance, solid-phase characterization and thermodynamic modeling revealed that nickel precipitates were mostly washed out with the effluent, due to the slow Ni precipitation kinetics and high upflow velocities in the FBR. Nickel phosphate, sulfide and oxide were indicated as the main nickel precipitates and were mostly amorphous.FBRs were shown to be powerful and robust biofilm systems for nitrogen removal under acidic pH, psychrophilic temperatures and high nickel concentrations. The results of this study are of great interest for the treatment of NO3- contaminated ground and mining waters in cold regions (e.g. Canadian and Scandinavian regions) as well as acidic and heavy-metal-laden wastewaters

Eutrophication is defined as the ‘over enrichment’ of a water body from nutrients, resulting in uncontrolled growth of primary producers, leading to periods of oxygen depletion from decomposition of the algal organic matter. According to the 2010 Water Infrastructure Needs and Investment (a U.S. Congressional Report), 40% of U.S. water bodies are contaminated with pollutants, including nutrients. Non-point sources of nutrient pollution are a major cause of this reduction in water quality. One way to decrease eutrophication is to manage nutrients found in stormwater runoff, before they reach a receiving water body. Bioretention cells containing an internal water storage zone (IWSZ) have been shown to remove higher amounts of nitrogen than conventional cells (without an IWSZ). The IWSZ contains an organic carbon substrate, usually derived from wood chips submerged in water, which supports the biochemical process of denitrification. Characteristics of wood chips that affect nitrogen removal include carbon content (%), leaching of dissolved organic carbon (DOC), and wood chip size and type. However, there is limited information on how the intermittent hydraulic loading that is associated with these field systems impacts their performance. Accordingly, the overall goal of this research is to improve understanding of the effect that the antecedent dry conditions (ADC) have on the performance of a field scale bioretention cell modified to contain an IWSZ. The nine different types of wood chips used in laboratory and field studies identified in the literature were categorized as hardwood and softwood. Literature showed that total organic carbon (TOC) leached from softwood chips is almost double the TOC measured from the hardwood chips, 138.3 and 70.3 mg/L, respectively. The average observed nitrogen removal for softwood chips was found to be greater than the removal for the average of the hardwood chips (75.2% and 63.0%, respectively). Literature also suggests that larger wood chip size may limit the availability of the carbon for the denitrifying organisms and provides less surface area for the biofilm growth. A field study conducted for this research compared the performance of a modified bioretention system designed to enhance denitrification, addition of an IWSZ, with a conventional system that does not contain an IWSZ. Fourteen storm events were completed from January 2016 to July 2016 by replicating storm events previously completed in the laboratory using hydraulic loading rates (HLR) of 6.9 cm/h, 13.9 cm/h, and 4.1 cm/h. The goal was to have results from storm events with ADCs of two, four, and eight days, with the varying durations of hydraulic loading of two, four, and six hours. Synthetic stormwater, simulating nitrogen levels common in urban runoff, was used as the system’s influent to assist in running a controlled experiment. The resultant ADCs ranged from 0 to 33 days, with the average ADC being 9 days. The fourteen sets of influent samples were averaged to obtain mean influent concentrations for the synthetic stormwater. These values were used when calculating the percent nitrogen removal for the four measured nitrogen species (NOx – N, NH4+– N, organic N, and TN). The field storm events were separated into three groups based on HLR and duration to eliminate the affects of both variables on nitrogen removal for these results, since the focus is the ADC. For the low HLR (4.1 cm/hr), there were four storm events (ADCs of 4 to 33 days), as the ADC increased, greater percentages of ammonium – nitrogen, organic nitrogen, and total nitrogen were removed. For nitrate/nitrite – nitrogen, the percent removal was rather consistent for all four storm events, not significantly increasing or decreasing with changes in the ADC. There were five storm events (ADCs of 0 to 28 days) tested with the median HLR (6.9 cm/hr), nitrogen removal for all four species increased as the ADC increased. The increase was significant (p0.05) for nitrate/nitrite – nitrogen. The third group also contained five storm events (ADCs from 0 to 11 days) that were tested with the highest HLR (13.9 cm/hr). Ammonium – nitrogen, nitrate/nitrite – nitrogen, and total nitrogen all increased with the ADC, and organic nitrogen removal decreased with the increasing ADC. As a result, this research concluded that the difference in HLR affects the nitrogen removal efficiency, but overall increasing the ADC increased nitrogen removal for NOx – N, NH4+ - N, organic N, and TN.

This study investigated the effect of excess carbon loading in the anoxic reactor on the nitrogen removal capacity of a biological pre-denitrification system for the treatment of a high ammonia leachate. The influent leachate was low in degradable organic carbon, thus an external carbon source was needed for denitrification requirements. Four different carbon sources were studied: methanol, glucose, acetate, and a waste brewer's yeast. The carbon loading was increased over the duration of the experimental period. The COD:NOx added to the anoxic reactor reached more than three times the carbon loading required to just achieve complete denitrification. All four carbon sources were found to support denitrification, but the glucose system showed erratic behaviour and ultimately failed after reaching a CODrNOx loading of about 23:1. The system using acetate appeared to require the least amount of COD:NOx (5.9:1) for complete denitrification, followed closely by methanol (6.2:1), then the yeast waste (8.5:1), and finally by glucose (9:1). Carbon breakthrough, the bleeding of carbon from the anoxic reactor into the aerobic reactor, was observed to occur just after complete denitrification was reached. The excess carbon did not appear to have any effect on denitrification, except in the case of the glucose system. The unit nitrification was found to decrease as the CODrNOx was increased, even though the ammonia removal remained at 100%. The decrease in nitrification, with respect to the COD:NOx, was most pronounced in the system that used methanol, and about equal in the other three systems. The cause of the decrease in nitrification is suspected to be due to increased ammonia assimilation by the heterotrophs rather than an inhibition of the nitrifiers. Nitrification ceased in the glucose system, but was restored within 12 days after the glucose addition was halted. The cause of the failure of the nitrogen removal process in the glucose system was not determined. Nitrite accumulation was observed in all the systems except the methanol system. The yeast waste system had nitrite accumulation in the aerobic reactor at C0D:N0x loadings over 25:1. Free ammonia inhibition of Nitrobacter is suspected to be the cause of aerobic nitrite buildup. The glucose and acetate systems had nitrite buildup in the anoxic reactor until complete denitrification was achieved. Facultative anaerobic bacteria are suspected of causing this nitrite accumulation. This theory was supported by observations in the glucose system, such as low anoxic pH; this may have been due to volatile fatty acids produced from fermentation. / Applied Science, Faculty of / Civil Engineering, Department of / Graduate

Les eaux souterraines des plaines alluviales agricoles sont particulièrement vulnérables à la contamination en nitrates en raison d’une fertilisation importante, et de la faible profondeur des aquifères. Au sein de ces plaines alluviales, les zones ripariennes, caractérisées par des échanges importants entre les eaux de surface et les eaux souterraines, constituent des environnements propices à l’atténuation des nitrates via le processus de dénitrification. Ce processus naturel dépend de nombreux facteurs et est caractérisé par une forte variabilité spatio-temporelle. La modélisation est une approche intéressante pour étudier la dénitrification au sein des zones ripariennes car elle permet d’intégrer les différents facteurs contrôlant le processus de manière dynamique. L’objectif de cette thèse est d’améliorer la caractérisation et la quantification du processus de dénitrification et d’identifier ses facteurs de contrôle dans les plaines alluviales à travers une approche de modélisation à l’échelle du tronçon de plaine alluviale. L’analyse d’un jeu de données collectées dans un réseau de 25 piézomètres installés dans un méandre de la Garonne (Monbéqui, 6,6 km²) lors de 12 campagnes mensuelles a permis, dans un premier temps, d’identifier les facteurs contrôlant la dynamique des nitrates et le processus de dénitrification. Les données ont ensuite servi à la mise en place et à la validation de l’application du modèle distribué MOHID Land pour simuler l’hydrologie de la zone d’étude. Puis, un module permettant de simuler la dénitrification et prenant en compte les facteurs identifiés au préalable, a été introduit dans le modèle. L’analyse des données récoltées sur le terrain a montré que la dénitrification dans la zone d’étude est contrôlée par la géomorphologie, l’hydrologie et la présence de carbone organique. L’application du modèle hydrologique distribué MOHID Land a permis de simuler correctement l’hydrologie du site étudié en prenant en compte la géomorphologie de la plaine alluviale, les échanges eaux de surface - eaux souterraines et les épisodes de crues. Un module de dénitrification incluant à la fois le carbone organique dissous apporté par la rivière et le carbone organique particulaire présent dans l’horizon supérieur des sols a été implémenté dans le modèle hydrologique. Les résultats indiquent une dénitrification moyenne de 28 kg-N.ha-1.an-1 sur la période simulée dans la zone riparienne de Monbéqui. La dénitrification est plus importante dans les zones ripariennes de basses altitudes et plus globalement dans la partie aval du méandre, caractérisée par de fortes concentrations en nitrates. Les facteurs contrôlant la dénitrification au sein des plaines alluviales ont été identifiés et sont i) la géomorphologie qui détermine la saturation des horizons supérieurs des sols et la disponibilité du carbone organique ; ii) les écoulements souterrains qui contrôlent la répartition spatiale des nitrates, et iii) l’intensité et la fréquence des épisodes de crues qui entrainent des périodes favorables à la dénitrification. Finalement, les résultats obtenus dans la zone d’étude ont été comparés avec des applications du modèle dans des plaines alluviales possédant des caractéristiques contrastées, validant ainsi l’utilisation du modèle dans les environnements de plaine alluviale variés. Les résultats ont mis en avant l’importance de la connectivité hydrologique entre la rivière et la nappe alluviale ainsi que la distribution spatiale des sources de nitrates et de carbone organique comme facteurs expliquant les différences de capacité de dénitrification entre les sites. Ces travaux ouvrent des perspectives pour évaluer l’impact des modifications induites par le changement global sur le processus de dénitrification ainsi que sur leur modélisation à plus large échelle. / Groundwater systems in cultivated floodplains are vulnerable to nitrate contamination due to extensive fertilisationand the shallow depth of the groundwater. Within the floodplain environment, riparian areas, characterised by significantexchanges between surface water and groundwater, are a favourable site for nitrate removal through denitrification.Denitrification is a natural process influenced by numerous factors and characterised by a high spatio-temporal variability.Modelling is an interesting approach for the study of denitrification within riparian areas because it means that floodplainspatial heterogeneity can be taken into account and many of its control factors integrated.The main objective of this thesis was to enhance knowledge of the denitrification process in floodplains and toidentify its control factors using a modelling approach at the scale of the floodplain reach. First, a dataset collected at amonthly time step over the course of a year within a 25-piezometer network of a meander area (Monbéqui, 6.6 km²) wasanalysed in order to identify factors controlling nitrate and denitrification patterns. The data were then used to set up andvalidate the MOHID Land model to simulate the hydrology in the study area. A module designed to simulate thedenitrification process, taking the main denitrification control factors into account, was then integrated into the MOHID Landmodel.The analysis of the dataset collected from the study area showed that denitrification was controlled bygeomorphology, hydrology and organic carbon presence. The application of the MOHID Land model allowed the hydrologyof the study area to be reproduced correctly, particularly the surface water – groundwater exchanges and the influence offlood events. A denitrification module integrating dissolved organic carbon borne by the river and particulate organic carbonfrom soil was added to the model. The results indicated that denitrification in the modelled riparian area was an average ofaround 28 kg-N.ha-1.yr-1 over the simulated period. High denitrification areas were located within the low elevation riparianarea and, more generally, in the downstream area of the meander where the nitrate concentrations were highest. The factors controlling denitrification that were identified in the study area were i) geomorphology, which controlled topsoil layersaturation and organic carbon availability; ii) groundwater flowpaths, which controlled the variability in nitrateconcentrations over the study area; and iii) flood event frequency and intensity, leading to high denitrification periods.Finally, the results were compared with other applications of the model in contrasting floodplains in order to assess themodel’s performance in various floodplain environment. The results indicated that hydrologic connectivity between the riverand groundwater and the spatial location of sources of nitrates and organic carbon were the major factors behind differencesin the denitrification process between floodplains. This study provides insight for an assessment of the impact ofmodifications that could be brought about by global changes to the denitrification process and for the modelling ofdenitrification at larger scales.

In Sweden and other Nordic countries, Free Water Surface (FWS) constructed wetlands are widely used for advanced sewage treatment. This study was performed in an experimental wetlands system in order to research the seasonal dynamics and interferences in N removal under cold climate condition. According to the result of study, as expected, N removal in FWS wetlands is strong related to temperature, the removal rates are higher in late summer and autumn than in spring and early summer. Removal at similar temperatures are quite different when they are in different months. For instance, in the study, the average N removal in June is significant different from N removal in September (P＜0.001), indicating that there are other factors affecting the N removal, different treatment wetlands have different situations. N removal in two different periods (March to July &amp; August to December) were extracted for covariance analysis, it indicated that they are significant different. Besides that, the r2-value of correlation test showed that total N removal rate in FWS constructed wetlands is higher in relation to temperature in autumn and winter (r2-value is 0.4449) than in spring and summer (r2-value is 0.3857). Generally speaking, the dynamics of N removal in FWS wetlands is not steady and variable even at similar temperatures. Finally, I find the temperature of excess heating from district heating which has been used by the residents’ house in Sweden is high enough to heat up the temperature in FWS wetlands. That is a valuable improvement to be put forward.