<p>Treatment wetlands play an important role in reducing nitrogen content in wastewater and agricultural run-off water. The main removal process is denitrification and the removal efficiency depends on the hydrological and microbiological features of the wetland, especially in terms of water residence times and denitrification rates. The aim of this thesis was to increase the understanding of the coupled hydrological and microbiological processes regulating the denitrification capacity. This was done by applying a broad spectrum of analyses methods, including tracer experiment, water flow modeling, denitrification rate measurements, and analyses of the microbial community structures. The tracer experiment and flow modeling revealed that the wetland design, especially the vegetation, largely can affect the water residence time distributions in wetlands. In the investigated wetland, vegetation dominated the water flow, explaining 60-80% of the variance in water residence times, whereas basin shape only explained about 10% of the variance, but also mixing phenomena significantly affected the residence times and could considerably delay solutes. Measured potential denitrification rates in the wetland exhibited significant spatial variations, and the variations were best described by concentration of nitrogen in sediments and water residence time. Analyses of the denitrifying bacteria populations indicated that a few key populations dominated and that the community diversity increased with decreasing nutrient levels and increasing water residence times. Moreover, it was found that denitrification rates in terms of Menten and first order kinetics can be evaluated by fitting a mathematical expression, considering denitrification and other nitrogen transforming processes to measured product formation in nitrate limited experiments.</p>

O óxido nitroso (N2O) é um importante gás de efeito estufa (GEE) e, nos ecossistemas terrestres, é produzido principalmente pelo processo de desnitrificação. Esse ocorre em condições anaeróbias e, portanto, é fortemente estimulado pelo aumento do teor de umidade do solo. Entretanto, solos sob diferentes usos podem exibir taxas de emissão de N2O distintas, mesmo quando apresentam teores de umidade equivalentes. Ainda não está claro se isso se deve somente ao fato de os mesmos diferirem quanto a atributos físicos e químicos capazes de afetar a atividade dos organismos desnitrificantes ou se também se deve à diferenças com relação ao tamanho de suas populações. O presente trabalho foi desenvolvido com o objetivo de compreender as relações entre os fluxos de N2O, a umidade e a abundância de genes bacterianos envolvidos no processo de desnitrificação (nirK, norB e nosZ) em solos de floresta, pastagem e cultivo de cana-de-açúcar, utilizando um experimento de microcosmos. Amostras de solo foram coletadas na fazenda Capuava, situada no município de Piracicaba, SP. Os microcosmos estabelecidos a partir das mesmas foram mantidos com diferentes teores de umidade (original e ajustados para atingir 60% e 90% da capacidade de campo) e incubados a 30 °C por 30 dias. Ao longo do período de incubação, os fluxos de N2O a partir desses solos foram analisados por cromatografia gasosa. Amostras coletadas do interior dos microcosmos, antes e depois da aplicação dos tratamentos, foram comparadas quanto à estrutura de suas comunidades bacterianas, utilizando a técnica de T-RFLP, e quanto à abundância dos genes 16S rRNA, nirK, norB e nosZ, através da técnica de qPCR. Somente os solos que tiveram sua umidade ajustada para 90% da capacidade de campo exibiram incrementos significativos na produção de N2O. Em tais amostras, também foi verificada a alteração da estrutura das comunidades bacterianas e do número de cópias dos genes norB e nosZ. Apenas este último, no entanto, apresentou uma correlação positiva com a umidade do solo. A abundância dos genes avaliados não apresentou correlações significativas com as taxas de emissão do GEE. Por outro lado, as emissões cumulativas de N2O se correlacionaram positivamente com as quantidades de genes desnitrificantes presentes inicialmente nas amostras de solo. Estes genes se mostraram mais abundantes nas amostras de pastagem e floresta, as quais apresentavam maiores teores de matéria orgânica, carbono, nitrogênio, nitrato e argila do que aquelas provenientes da área cultivada com cana-de-açúcar. Tais resultados demonstram que o conteúdo de água do solo afeta a taxa de emissão de N2O, mas que isso não se deve a alterações na abundância das bactérias envolvidas no processo, como as que carregam os genes nirK, norB e nosZ. Aparentemente, no entanto, quantidade de GEE que o solo é capaz de produzir está relacionada ao tamanho das populações desses organismos desnitrificantes. / Nitrous oxide (N2O) is an important greenhouse gas (GHG) and, in terrestrial ecosystems, it is mainly produced by denitrification. This process occurs under anaerobic conditions and, therefore, is strongly stimulated by the increase of the soil moisture content. However, soils under different uses may exhibit distinct N2O emission rates, even when they have the same moisture content. It is still not clear whether this is due solely to the fact that they differ in relation to physical and chemical properties that affect the activity of denitrifying organisms or whether this is also due to differences in the size of their populations. The aim of this work was to evaluate the relations between N2O fluxes, moisture and abundance of bacterial genes involved in denitrification process (nirK, norB e nosZ) in soil samples from forest, pasture and sugarcane field, through a microcosm experiment. These samples were collected at Fazenda Capuava, located in Piracicaba, SP. Microcosms established from them were maintained with different moisture contents (original and adjusted to achieve 60% and 90% of field capacity) and incubated at 30 °C for 30 days. During the incubation period, the N2O fluxes from soils were analyzed by gas chromatography. Soil samples from microcosms, collected before and after application of the treatments, were compared regarding the structure of their bacterial communities, by using T-RFLP technique, and the abundance of 16S rRNA, nirK, norB and nosZ genes, through qPCR technique. Only samples that had their moisture content adjusted to 90% of field capacity exhibited significant increases in N2O production. In these samples, changes in the structure of bacterial communities and in the copy numbers of norB and nosZ genes were also detected. Only the latter gene, however, showed a positive correlation with soil moisture. The abundance of the quantified genes showed no significant correlations with the gas emission rates. On the other hand, the cumulative N2O emissions were positively correlated with the amounts of denitrifying genes initially present in the samples. These genes were more abundant in pasture and forest soils, which had higher levels of organic matter, carbon, nitrogen, nitrate and clay than those from sugarcane cropping area. These results indicate that soil water content affects the N2O emission rates. However it is not due to changes in the abundance of bacteria involved in the process, such as those that bear the nirK, norB and nosZ genes. Apparently, it is the size of these organisms\' populations that determines the amount of GHG that the soil is able to produce.

Treatment wetlands play an important role in reducing nitrogen content in wastewater and agricultural run-off water. The main removal process is denitrification and the removal efficiency depends on the hydrological and microbiological features of the wetland, especially in terms of water residence times and denitrification rates. The aim of this thesis was to increase the understanding of the coupled hydrological and microbiological processes regulating the denitrification capacity. This was done by applying a broad spectrum of analyses methods, including tracer experiment, water flow modeling, denitrification rate measurements, and analyses of the microbial community structures. The tracer experiment and flow modeling revealed that the wetland design, especially the vegetation, largely can affect the water residence time distributions in wetlands. In the investigated wetland, vegetation dominated the water flow, explaining 60-80% of the variance in water residence times, whereas basin shape only explained about 10% of the variance, but also mixing phenomena significantly affected the residence times and could considerably delay solutes. Measured potential denitrification rates in the wetland exhibited significant spatial variations, and the variations were best described by concentration of nitrogen in sediments and water residence time. Analyses of the denitrifying bacteria populations indicated that a few key populations dominated and that the community diversity increased with decreasing nutrient levels and increasing water residence times. Moreover, it was found that denitrification rates in terms of Menten and first order kinetics can be evaluated by fitting a mathematical expression, considering denitrification and other nitrogen transforming processes to measured product formation in nitrate limited experiments. / QC 20101110

Dans un contexte de destruction des zones humides à l’échelle mondiale, conséquence des activités d’origine anthropique, la restauration écologique de ces milieux et de leurs fonctions est devenue un enjeu écologique et sociétal. L’objectif de ce projet est de recréer des prairies humides à vocation agricole aux caractéristiques pédologiques et floristiques aussi proches que possible de celles des prairies totalement détruites par l’exploitation de matériaux alluvionnaires. L’étude présentée ici se focalise sur l’incidence des matériaux pédologiques utilisés pour la recréation de quatre sols sur les fonctions du sol et les cortèges floristiques associés. L’hypothèse principale est que la recréation d’un sol morphologiquement proche de celui détruit devrait permettre d’orienter la restauration écologique.L’hypothèse sous-jacente est qu’en utilisant différents matériaux pédologiques locaux, on hérite de leurs caractéristiques physico-chimiques et biologiques ce qui permettait de conserver les fonctions pédologiques qui leur sont associées et favorisait le retour d’un cortège floristique compatible avec un usage agricole. La première partie est consacrée à l’étude de deux fonctions remplies par les sols de zones humides à savoir le stockage du carbone organique et la dénitrification. Deux années et demi après la fin des travaux de comblement de ballastières, ces deux fonctions sont conservées au sein des quatre types de sols recréés. Les principaux résultats montrent un niveau d’efficience des matériaux pédologiques testés, fonction de leur sol origine et de leurs caractéristiques physico-chimiques. La deuxième partie est consacrée à l’étude des mécanismes de structuration des communautés végétales. Le suivi de la colonisation spontanée de la végétation a permis d’appréhender la forte contribution de la banque de graines issue des matériaux pédologiques locaux. Malgré la mise en évidence d’un début de trajectoires dynamiques au sein des quatre sols recréés, la similarité entre les communautés obtenues et les communautés cibles des prairies de référence n’excède pas 50 %. Les productions de biomasses aériennes associées à ces communautés sont comparables en quantité à celles des prairies de référence mais pas en qualité. La mise en place d’une gestion par semis associée à une fauche montre dès la première année une production de biomasses de qualité se rapprochant de celles des prairies locales.La dernière partie de ce manuscrit est consacrée à l’effet de trois niveaux d’engorgement des sols sur le processus de dénitrification et sur les traits de réponse d’une espèce prairiale, Holcus lanatus. Placer les quatre matériaux pédologiques dans des conditions identiques d’engorgement permet de souligner l’importance de l’héritage des communautés bactériennes dénitrifiantes sur le processus de dénitrification. Parallèlement, ces conditions expérimentales permettent de mettre en évidence les traits de réponses morphologiques et fonctionnels de l’espèce considérée. À l’issue de ces suivis, le meilleur compromis de restauration alliant sol, végétation et coûts économiques doit prendre en considération l’origine et l’histoire (i.e. gestion) des matériaux utilisés lors de la recréation écologique. / In a worldwide context of wetland destruction, a consequence of anthropic activities, ecological restoration of such habitats and their functions has become a societal and ecological issue. The objective of this project is to recreate agriculture-oriented wet grasslands with pedological and floristical properties as similar as possible to typical grasslands destroyed by alluvial materials extraction. The study presented here focuses on the impact of pedological materials, used in the re-creation of four soils, on soil functions and associated floristic processions. The main hypothesis is that re-creation of a soil morphologically similar to the previously destroyed one should drive ecological restoration. The underlying hypothesis is that different local pedological materials inherit their previous physicochemical and biological properties. This should conserve associated pedological functions and favor the return of a floristic procession compatible with agricultural exploitation. The first part is dedicated to the study of carbon storage and denitrification, two wetlands soils functions. These two functions are retained within the four re-created soils two and a half year after gravel-pit filling. Main results highlight functional efficiency levels of tested pedological material inherited from their respective initial topsoil physico-chemical properties. The second part is devoted to the study of mechanisms structuring plant communities. The high contribution of local pedological materials seed bank during the colonization process and its impact on aforementioned mechanisms was highlighted from our monitoring. Despite demonstration of the start of a dynamic trajectory in the four created soils similarity between obtained and target communities never exceeds 50%. Aerial biomass production associated to these communities is comparable to the production in reference wet grasslands in terms of quantity, but not quality. Implementation of management (sowing and mowing) shows biomass production of comparable quality to reference grassland from the first year onwards. The last part focuses on the effect of three soil waterlogging levels on the denitrification process and the response traits of Holcus lanatus, a meadow species. Pedological materials placement in identical waterlogging conditions highlights the importance of denitrifying bacteria communities inheritance on the denitrification process. These experimental conditions also enabled us to highlight the considered species morphological and functional response traits. To conclude and following our monitoring the best compromise for concurrent restoration of soil and vegetation while considering cost-effectiveness needs to account for topsoils origin and history (i.e. management).

Diversity of the soil microbial community and its functional aspects were investigated in man-influenced environments, such as colliery spoil heaps in post mining sites and upland pasture used for outdoor cattle husbandry. The study was based on the cultivation of bacteria and streptomycetes as well as culture-independent approaches. Cultivated bacteria and streptomycetes were characterized by phenotypic and genotypic means. The culture-independent approaches were based on an analysis of environmental DNA in terms of both qualitative and quantitative parameters.

Denitrification is a microbially-catalyzed reaction which reduces nitrate to N2 through a series of intermediate nitrogen compounds. Nitrate is a nutrient and its release into the environment may lead to eutrophication, depending on the amount that is released and the state of the recipient. The release of nitrate from the mining industry in Kiruna (Sweden) has been identified as an eutrophication risk, and a denitrifying bioreactor is to be constructed at the site to reduce the nitrate release.Since the denitrification rate decreases with temperature and the temperature in Kiruna during large parts of the year drops below 0˚C, the denitrifying bioreactor therefore has to be designed for the site-specific environment in terms of flow rate and hydraulic residence time. Laboratory column studies are used to study and determine the nitrate removal rate in a low temperature environment (5˚C) with pine wood chips as reactive matrix/ electron donor; the input solution had an average concentration of 35 mg NO3-N/L and a high sulfate concentration. Nitrate removal was studied as a function of hydraulic residence time and temperature. Parameters that were monitored include pH, alkalinity and concentrations of ammonium, nitrite and sulfate in the effluent from the columns. On three occasions, samples were gathered along the flow path in the columns (concentration profiles) such that changes in nitrate, nitrite, and occasionally ammonium concentration could be studied in relation to each other. The study concluded that a denitrifying bioreactor utilizing pine wood chips as the reactive matrix is a suitable option for nitrate treatment in a low temperature (5˚C) environment. Under the conditions of the study, effluent nitrate, nitrite, and ammonium concentrations are below limits established in legislation. Nitrate removal rates are given for zero-order nitrate reduction and overall first-order nitrate reduction, as the concentration profiles revealed a decrease in nitrate removal rate as nitrate concentration dropped below 3 mg NO3-N/L. / Nitrat är ett näringsämne som kan orsaka övergödning vid utsläpp, beroende på halterna och recipienten. Växterna som tar upp kväve kommer så småningom att dö och sjunka mot botten där de förmultnar. Förmultningen kräver syre, och vid ökad växtlighet så ökar även konsumtionen av syre då det finns mer organiskt material att bryta ned. Detta leder i slutändan till syrefria områden, där djurliv och växtlighet är mer begränsade. Nitratutsläpp från gruvindustrin i Kiruna har blivit identifierad som en potentiell övergödningsrisk och en denitrifierande bioreaktor ska därmed installeras för att minska utsläppen. Denitrifikation är en mikrobiell reaktion som reducerar nitrat till kvävgas genom en serie av intermediära kväveföreningar. En denitrifierande bioreaktor använder sig utav denitrifikation för att minska nitratkoncentrationer i vatten som passerar genom bioreaktorn som består av huvudsakligen; (1) bakterierna som sköter denitrifikationen, och (2) en kolkälla som fungerar som ”mat” till de denitrifierande bakterierna, Hastigheten varvid nitrat omvandlas till kvävgas genom denitrifikation, minskar med temperatur och den denitrifierande bioreaktorn måste därmed anpassas till omgivningen där den ska placeras med avseende på uppehållstid i reaktorn. Uppehållstiden måste vara tillräcklig för att minska nitratkoncentrationen till önskad nivå, men samtidigt så får uppehållstiden inte vara för lång då andra ämnen kan reagera och bilda ofördelaktiga produkter vid låga nitratkoncentrationer. Kolonnstudier i en låg-tempererad miljö (5˚C) är ett första steg för att studera hastigheten av nitratförbrukning i en sådan omgivning, och används i detta arbete med träflis av tall som kolkälla. Parametrar som påverkar, och varierar som ett resultat av, denitrifikation (exempelvis pH och sekundära föroreningar) övervakas. Hastigheten av nitratförbrukning som fås från kolonnstudierna kan sedan används som riktlinjer för konstruktionen av en denitrifierande bioreaktor i fältskala i Kiruna. Studiens slutsats är att en denitrifierande bioreaktor med träflis av tall som reaktivt medium är ett fungerande alternativ för nitrat reducering i en lågtempererad miljö (5˚C) då nitrat effektivt reduceras till under gränsvärden fastslagna i lag. Även andra potentiella biprodukter (exempelvis nitrit och ammonium) som kan resultera från den miljö som den denitrifierande bioreaktorn ger upphov till är under de gränsvärden som finns fastslagna i lag.