Les émissions élevées de CO2 de la filière industrielle de fabrication des tuyaux en fonte résultent pour l'essentiel de l'emploi massif de carbone fossile, charbon et coke, comme combustible et agent réducteur dans les procédés. Substituer du carbone issu de biomasse au carbone fossile en vue de réduire ces émissions de CO2, avec application au cas d'une usine lorraine, a été l'idée de départ de ce travail. Différents types de biomasse ont été envisagés. Le bois et la paille seraient localement disponibles en quantité suffisante pour autoriser une substitution partielle de 20 % du coke. Cette biomasse doit cependant être traitée thermiquement (séchage et pyrolyse) avant de satisfaire les spécifications techniques imposées par les procédés métallurgiques (agglomération, cokerie, haut fourneau). Six scénarios d'usage de la biomasse au haut fourneau ou à l'agglomération ont été sélectionnés. Pour les évaluer sur la base de leur mérite environnemental, nous avons réalisé une analyse de cycle de vie comparative en suivant une approche du type berceau à la porte. Point original de notre étude, l'inventaire de cycle de vie est issu de modèles systémiques des principaux procédés, spécifiquement développés et intégrés à un logiciel de diagrammes de flux. Grâce à ces modèles et notamment celui du haut fourneau, on a pu déterminer les taux de remplacement du coke par trois types de biomasse (charbon de bois en blocs, charbon de bois pulvérisé, bois torréfié pulvérisé) et prédire les modifications de fonctionnement qu'entrainerait l'injection de biomasse. Les résultats de l'analyse de cycle de vie montrent que l'injection de bois, carbonisé ou torréfié, sous forme pulvérisée aux tuyères du haut fourneau serait une solution intéressante d'un point de vue environnemental : baisse des émissions de gaz à effet de serre (-15 %) et des impacts sur la santé humaine et sur les écosystèmes. A l'heure actuelle cependant, le coût économique d'une telle solution demeure excessif / High CO2 emissions from the industrial production of pig iron pipes are mainly due to the massive use of fossil carbon, coal and coke, as a fuel and/or a reducing agent in the processes. The use of biogenic carbon as a substitute for fossil carbon in order to reduce the CO2 emissions, in the case of an existing plant in Lorraine, was the starting point of our work. Different types of biomass were investigated. Local resources for wood and straw would be sufficient for a partial substitution for 20% of the coke. However, this kind of biomass should first undergo a thermal pretreatment step (drying and pyrolysis) to fit the technical requirements for the metallurgical processes (sinter plant, coke oven and blast furnace). Four scenarios of using biomass in the blast furnace or at the sinter plant were selected. To evaluate their environmental performances, we produced a comparative life cycle assessment, using a cradle-to-gate approach. An original feature of our study lies in the use of systems modeling for the determination of the life cycle inventories of the main processes, which were modeled using a flowsheeting software. From these models, particularly the model of the blast furnace, we could evaluate the replacement ratio of three biomass types (charcoal lumps, charcoal fines and torrefied wood fines) for the coke and predict the influence of biomass injection in the blast furnace on its operation. The results of the life cycle assessment show that the injection of pulverized, torrefied or carbonized, wood through the tuyeres of the blast furnace would be an attracting solution from an environmental point of view. It would reduce greenhouse gas emissions by 15%, as well as mitigate the impacts on human health and ecosystems. Yet, the costs associated to such a solution are currently too high

Neste trabalho foi desenvolvido um método de análise de erva mate mediante as técnicas de espectroscopia no infravermelho próximo (NIR), espectrometria de emissão óptica com plasma indutivamente acoplado (ICP OES) e espectrometria de massa com plasma indutivamente acoplado (ICP-MS), com o intuito de classificar a erva mate por país de origem da América do Sul. Para isto, foram utilizados métodos de análise multivariada supervisionada e não supervisionada, como a análise hierárquica por agrupamentos (HCA), a análise por componentes principais (PCA), o método do k-ésimo vizinho mais próximo (kNN), soft independent modeling class analogy (SIMCA), análise discriminante por mínimos quadrados parciais (PLS-DA) e análise discriminante por máquinas de vetor de suporte (SVM-DA). Foram analisadas 54 amostras de erva mate provenientes da América do Sul (Argentina, Brasil, Paraguai e Uruguai). As amostras foram cominuídas em moinho criogênico e diretamente analisadas para a obtenção dos espectros de reflexão no NIR. Para a determinação dos elementos Ca, K, Mg, Al, Fe, Mn, Ba, Cu, P, Sr e Zn por ICP OES e Li, Be, Ti, V, Cr, Ni, Co, As, Se, Rb, Mo, Ag, Cd, Sb, La, Ce, Pb, Bi e U por ICP-MS, foi realizada a decomposição das amostras por via úmida em bloco digestor em sistema fechado. A classificação da erva mate por país de origem foi possível, sendo que a melhor classificação para todos os métodos de análise multivariada foi obtida por NIR e para as concentrações dos elementos avaliados, a SVM-DA foi o melhor. Conclui-se que a erva mate pode ser avaliada por país de origem mediante a técnicas de ICP OES, ICP-MS e NIR. / A method of yerba mate analyzis using near infrared spectroscopy (NIR), inductively coupled plasma optical emission spectrometry ( ICP OES ) and inductively coupled plasma mass spectrometry (ICP-MS) was developed in order to classify yerba mate by country of origin in South America. To this end, principal components analysis (PCA), k-nearest neighbor (KNN), soft independent modeling class analogy (SIMCA), partial least square discriminant analysis (PLSDA), discriminant analysis by vector machine (SVM-DA) were used for multivariate analysis of the NIR spectra and element concentrations found in 54 samples of yerba mate from South America (Argentina, Brazil, Paraguay and Uruguay ). The samples were ground in cryogenic mill and directly analyzed in order to obtain the reflectance spectra in the NIR region. For elements determination (Ca, K, Mg, Al, Fe, Mn, Ba, Cu, P, Sr and Zn by ICP OES and Li, Be, Ti, V, Cr, Ni, Co, As, Se, Rb, Mo, Ag, Cd, Sb, La, Ce, Pb, Bi and U by ICP-MS), the samples were acid digested in closed vessel in digestion block. The classification of the yerba mate by country of origin was possible, whereas the best classification using all multivariate analysis methods was obtained for NIR. For the element concentrations, the best classification was obtained by SVM-DA. It was concluded that yerba mate classification by country of origin can be assessed through ICP OES, ICP-MS and NIR.

Orientador: Varlei Rodrigues / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Física Gleb Wataghin / Made available in DSpace on 2018-08-28T03:02:40Z (GMT). No. of bitstreams: 1 Rodrigues_KevinLiu_M.pdf: 15717954 bytes, checksum: 0231043634d153e2ddc45e00a8630aaa (MD5) Previous issue date: 2015 / Resumo: O estudo de nanopartículas é desafiador devido à complexidade destes sistemas, que estão na interface entre o pequeno e o grande. A física dos sólidos deixa de valer em muitas das partículas menores devido ao alto grau de confinamento, e a fração de átomos de superfície para átomos no núcleo. Devido à natureza pequena dos \textit{clusters}, as medidas realizadas nestes sistemas são medidas de alto grau de precisão. Com o objetivo de caracterizar feixes de fontes de agregados de maneira rápida e acessível, um circuito capacímetro é utilizado para medir a influência da deposição de nanopartículas na região dielétrica de um capacitor, uma proposta em aberto na área de nanopartículas. Como essa medida captura a contribuição de inúmeros agregados diferentes, todos depositados na região dielétrica do capacitor, um filtro seletor de massas pode ser utilizado para estreitar a largura da distribuição de tamanhos e obter a influência causada por esse tamanho específico apenas. Neste trabalho será apresentado o projeto para a implementação de um filtro com esse propósito, original de Issendorf e Palmer, cujo princípio de operação é o tempo de voo das partículas, e tem resolução esperada de uma parte em quarenta. Medidas de capacitância também foram realizadas na tentativa de caracterizar a variação da capacitância em função da quantidade de material depositado, com resultados promissores, mas que requerem um maior refinamento das medidas. Os nanoagregados foram produzidos depositados e analisados no Grupo de Física de Nanossistemas e Materiais Nanoestruturados (GFNMN) da física da Unicamp / Abstract: The study of nanoparticles is a challenging one because of the complexity of these systems, which lie on the interface between the small and the big. By that it is meant that many crystals and solids approximations cease to work, and confinement effects begin to appear, as well as surface phenomena which gain strength as surface to volume atoms increase. Due to the small particle sizes, measurements become increasingly difficult, requiring higher and higher precision techniques. With the objective of quickly characterizing beams produced by nanoaggregates sources, a capacitance measuring circuit was created with the intent of measuring the influence of the aggregates on the dielectric constant of the medium between the electrodes, a technique which has not been applied to cluster beams yet. In combination with the capacitance measurements, a mass selecting device was built in order to make the capacitance measurements reflect only a narrow mass distribution, allowing the study of properties of a single cluster size provided they are small enough. In this study will be discussed the project for such a mass filter, original by Issendorf and Palmer, and its implementation on a cluster source. Its principle of operation is the time of flight mass spectrometry, and its projected resolution of about one part in forty. Capacitance measurements were also performed to try and characterize the capacitance change as a function of the density of deposited material. The results were promissing but require further refinement for a strong correlation. All the particle beams were produced at the cluster source operating at the GFNMN group at the department of applied physics at IFGW Unicamp / Mestrado / Física / Mestre em Física / 2011/15018-7 / FAPESP

Como resultado da crescente impermeabilização das bacias, as redes de drenagem pública são submetidas a vazões de pico cada vez maiores. Na tentativa de contornar ou amortizar este problema, pesquisadores têm estudado soluções que compensem os efeitos da urbanização. Estas soluções são denominadas técnicas compensatórias, ou BMP\'s (Best management Practices), cujo foco é manter o ciclo hidrológico mais próximo do natural da bacia, pelo menos no que se refere ao escoamento superficial. A aposta em medidas de controle na fonte de geração do escoamento, isto é, no lote, cresceu, sendo estabelecido seu uso por meio de leis e decretos. Assim, com o propósito de redução e detenção do escoamento pluvial, antes do mesmo ingressar no sistema de microdrenagem, os microrreservatórios na saída dos lotes são cogitados como estruturas compensatórias. O presente trabalho busca dimensionar e avaliar o impacto desses microrreservatórios a princípio em escalas de lote e loteamento e, em seguida, em escala de macrodrenagem. O estudo foi aplicado a uma bacia da cidade de São Carlos, estado de São Paulo, em processo de urbanização, que vem sofrendo grande pressão imobiliária para a construção de novos empreendimentos. As legislações locais sancionadas, nos últimos anos, estabelecem a implantação de poço de infiltração e reservatório de captação e reaproveitamento de águas pluviais, sem maiores detalhes sobre métodos construtivos ou estudos preliminares de viabilidade. Os resultados aqui descritos foram produzidos considerando um loteamento existente na bacia do córrego do Mineirinho, São Carlos, SP. Foram produzidos volumes para o dimensionamento dos reservatórios para as diferentes áreas de lote, assumindo os percentuais de impermeabilização de 50% e 70%, segundo os diferentes métodos de estimativa constantes da literatura. As vazões máximas e os volumes resultantes apresentaram significativa variação, quando calculados utilizando os métodos de estimativa constantes da literatura. Através da modelagem hidráulico-hidrológica, concluiu-se também que o reservatório de detenção na fonte é eficiente na redução da vazão de pico gerada no lote para vazão de pico em nível de pré-urbanização. Além disso, a definição do diâmetro do orifício de descarga mostrou-se determinante ao dimensionamento do reservatório, de modo a garantir sua eficiência. Os microrreservatórios dimensionados para duração da chuva crítica do lote (90 minutos), apesar de maiores, exibiram maior capacidade de detenção, sem extravasar quando submetidos a diferentes durações de chuvas. No loteamento, o uso desses dispositivos possibilitou a atenuação da vazão de pico de descarga em, pelo menos, 44%. Em escala de bacia, diferentes posicionamentos do loteamento, com e sem microrreservatórios de lote, foram instrumentais em demonstrar a capacidade de redução das vazões de pico locais. Este estudo comprovou que os microrreservatórios de lote são uma boa opção na redução das vazões escoadas pelo canal de macrodrenagem, quando da ocorrência de eventos de precipitação na bacia do córrego do Mineirinho. / As a result of the growing waterproofing of basins, the public drainage systems are under increasing flow rate peaks. In an attempt to mitigate these issues, researchers have studied solutions to address the effects of the urban development. These solutions are well-known as BMP\'s (Best Management Practices) and aim to preserve the hydrological cycle near to the natural basin, at least for the superficial run-off. The increasing search for solutions that controls the source of the run-off was responsible for the establishment of new laws and rules for on-site run-off control. Thus, for reduction and detention of run-off before it goes into the micro-drainage system, micro-reservoirs are considered as BMP\'s solutions. The present study aims to sizing and assess the impact of micro-reservoirs, initially on site and on residential land scale, and subsequently, on macro-drainage scale. The study was applied to Mineirinho Stream Watershed, in São Carlos city, São Paulo state. These basins are in urbanization process, and suffering from market pressure of constructions of real estate ventures. Recent local legislations have imposed the implementation of infiltration pits and rain water tanks, but without more details about construction methods or even preliminary feasibility studies. The results described in this work, regarding to sizing of micro-reservoirs to different allotment size ranges, used 50% and 70% of waterproofing percentages, according estimation methods from the literature. Peak flows and sizing showed a significant variation when calculated by these methods. Hydraulic-hydrologic SWMM software modelling showed that micro-reservoir is efficient in reducing the hydrograph peak flow from urbanization to preurbanization level. Furthermore, the definitions of the orifice diameter have demonstrated to be decisive to dimensioning of the reservoir, in order to ensure its efficiency. The microreservoirs sized for a 90 minutes critical rain, showed more detention capacity, without overflow when under different rainfall durations. The use of micro-reservoirs in the allotments of a residential division, was able to reduce the hydrograph outside peak in 44%, at least. In a basin-scale, moving the same residential division, by using or not micro-reservoirs in the allotments, were conclusive in demonstrate that this BMP is a good option to reducing the hydrograph river flow, being able to reduce the peak flow in the discharge point. The study comproved that the use of micro-reservoirs in both basin positions can be a feasible solution to recurring flooding problems in the Mineirinho Stream.

La diffusion des techniques et du savoir-faire est un thème fort en matière d’hydraulique somptuaire. Les recherches menées sur la communauté des fontainiers français au service du roi d’Espagne Philippe V permettent d’approfondir cette problématique. Installés sur le domaine royal de La Granja de San Ildefonso dans la Sierra de Guadarrama près de Ségovie, les fontainiers mettent en œuvre les innovations techniques déjà appliquées sur le chantier de Versailles. Cette étude de cas ouvre la réflexion, d'une part, sur les transferts techniques et la circulation des savoirs entre France et Espagne à l'époque moderne, à une période où des liens étroits se sont tissés entre la cour madrilène et la cour versaillaise, et d'autre part, sur la place de l'Espagne comme centre d'innovation technologique.Les jardins et leurs fontaines deviennent objet de démonstration du pouvoir, expression de la domination politique et de la puissance royale. La symbolique de faste associée aux effets d'eau des fontaines de La Granja de San Ildefonso nécessite l'installation d'un nouveau réseau de techniciens spécialisés. Le roi d'Espagne fait alors appel à des fontainiers issus de Sceaux, Versailles, Trianon, Marly et Chantilly. Ces experts hydrauliciens participent pleinement au succès de l'opération, aux côtés des jardiniers, sculpteurs et architectes, prenant part à la conception des systèmes d’alimentation en eau des jardins et à l'entretien de ces infrastructures au quotidien.Tout au long du XVIIIe siècle, les innovations techniques mises en œuvre sur les jardins royaux français, transmises par l'intermédiaire de cette communauté de fontainiers expatriés, vont s'agréger à la tradition locale hydraulique espagnole, déjà le fruit de plusieurs transferts technologiques. Les jardins de San Ildefonso doivent être replacés dans une perspective à dimension européenne de diffusion des savoir-faire du personnel technique royal français. / The study about the community of French hydraulic engineers in Spain in the service of King Philip V is to question the history of technology focusing on royal palaces hydraulics and the circulation of knowledge. The arrival of the French hydraulic engineers in La Granja de San Ildefonso changed the hydraulic technologies in Spain. From this perspective, hydraulics is an expression of political domination. This case study allows to investigate technological innovations and their transfers between France and Spain during modern times, a period when the links between both the Madrid and the Versailles courts were very close.A new network of specialized technicians was necessary to give full dimension to the political domination which was perfectly conveyed by the fountains of La Granja de San Ildefonso with their highly symbolic and breath-taking water effects. That is why the King of Spain hired hydraulic engineers from Sceaux, Versailles, Trianon, Marly and Chantilly. It was thanks to these experts who designed water supply systems and daily routines to take care of the installations, by the side of gardeners, sculptors and architects, that the project ended up being a great success.During the whole 18th century, technical innovations were implemented in the French royal gardens, taught through this community of expatriated hydraulic engineers and enriched the local hydraulic Spanish tradition which was already the result of various technological transfers. The San Ildefonso gardens must be placed on a European perspective regarding the expertise of the French royal technical workforce.

[pt] Os codificadores universais via recorrência de padrões surgiram nos anos 70, quando foram propostos os codificadores lz77 e lz78. Devido a baixa complexidade computacional e ao bom desempenho, quando aplicados na compressão de arquivos de dados, estes codificadores se tornaram extremamente populares. Embora estes codificadores sejam universais, i.e., possuam taxas de compressão que convergem para entropia, recentemente foi mostrado que as taxas de compressão não convergem da forma mais rápida possível, nem mesmo para a classe de fontes sem memória. A redundância de um codificador universal C, dada por R, mede a rapidez com que a taxa de compressão converge para a entropia. Para fontes com número de estados finito, enquanto os melhores resultados de codificadores universais via recorrência de padrões apresentam uma redundância da ordem de 1/{log n}, existem codificadores que atingem uma redundancia de {log n} /n. Portanto, os codificadores via recorrência de padrões não são ótimos. Embora seja conhecido codificadores ótimos segundo o critério da redundância, tais codificadores possuem uma alta complexidade computacional e são pouco úteis na prática. Dentre os codificadores via recorrência de padrões, o codificador lz78 possui uma das mais baixas redundâncias para a classe de fontes com número de estados finito. De fato, não existe outro codificador desta classe, tal que a redundância seja melhor que a do lz78. Tomando como base o lz78, este trabalho propõe novas técnicas para acelerar a convergência da taxa de compressão (diminuir a redundância) dos codificadores universais via recorrência de padrões, para fontes com número de estados finito. Estas técnicas dão origem a novas versões do lz78. As redundâncias das novas versões são estabelecidas, considerando a classe de fontes com número de estados finito. Estas versões são aplicadas na compressão de arquivos de dados, e os resultados obtidos são comparados com os resultados de versões anteriores. / [en] The string matching encoders were proposed about 20 years ago, when the lz77 and lz78 were introduced. They became extremely popular because of the relationship between their low complexity and their good performance. Although these encoders are universal, that is, their compression rates converge to the source entropy, it was shown, recently, that those rates do not converge as fast as possible, even for the class of memoryless source. The redundancy of an unversal encoder C, denoted by Rc, measures how fast the compression rate converge to the entropy. In the class of Finite State Machine (FSM) source, while the best result of the string matching encoders is O 1/ {log n}, there are encoders which achieve a redundancy of O {log n}/n. Therefore, the string matching encoders are not optimal. Even though optimal encoders are known, in general, those encoders have a high complexity and are not useful in practice. Considering the class of string matching encoders, the lz78 has one of the best results over the class of FSM source. In fact, there is no encoder based on the string matching, which archives a better redundancy. This work makes use of the lz78 encoder to propose new techniques to improve the performance of sting matching encoders over the class of FSM. These techniques bring up new versions of the lz78. For the class of FSM, the redundancies of this versions are stablished. These versions are used to compress data files, and their perfoermances are compared to the performances of older versions.

No description available.

Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-05-05T18:23:07Z No. of bitstreams: 1 yuricalillouresdeoliveira.pdf: 2613361 bytes, checksum: b14d216c9e91ffe085e35724d58bc704 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2016-06-07T15:42:11Z (GMT) No. of bitstreams: 1 yuricalillouresdeoliveira.pdf: 2613361 bytes, checksum: b14d216c9e91ffe085e35724d58bc704 (MD5) / Made available in DSpace on 2016-06-07T15:42:11Z (GMT). No. of bitstreams: 1 yuricalillouresdeoliveira.pdf: 2613361 bytes, checksum: b14d216c9e91ffe085e35724d58bc704 (MD5) Previous issue date: 2016-01-12 / CNPq - Conselho Nacional de Desenvolvimento Científico e Tecnológico / Esta dissertação apresenta um estudo de estratégias de alimentação de cargas hospitalares críticas utilizando conversores estáticos de energia. As cargas hospitalares escolhidas foram os equipamentos de raio x, tomografia computadorizada e ressonância magnética, sendo comparadas as seguintes formas de integração à rede: (i) estrutura com um retificador trifásico conectado à rede e três inversores monofásicos conectados às cargas ou (ii) estrutura com três retificadores monofásicos conectados à rede e três inversores monofásicos conectados às cargas. Por questões de patente, houve dificuldade em se encontrar os circuitos de front-end das fontes de alimentação das cargas hospitalares. Logo, as mesmas foram construídas dentro do ambiente de simulação do software PSIM utilizando dados de medição obtidos na literatura. O controle dos retificadores trifásico e monofásico foi realizados utilizando duas malhas de controle em cascata, com a malha interna controlando a corrente consumida da rede e a malha externa regulando a tensão no barramento CC. Para os três inversores monofásicos de saída foram implementados os controles PI RES e PI SRF visando uma comparação quanto ao desempenho. Como forma de validar o modelo matemático do sistema e o algoritmo de controle, foram feitas simulações utilizando o software PSIM. A análise feita comprovou que os resultados de simulação se mostraram coerentes com a modelagem matemática do sistema. / This master thesis presents a study about feeding strategies of critical hospital loads through power static converters. The chosen hospital loads were the x ray equipment, the computed tomography and the magnetic resonance. The following forms of grid integration were compared: (i) structure with a three-phase rectifier connected to the network and three single-phase inverters connected to the loads or (ii) structure with three single-phase rectifiers connected to the network and three single-phase inverters connected to the loads. Due to patent issues, it was difficult to find the front-end circuits for power supplies of hospital loads. Therefore, the loads were implemented in PSIM software environment using measurement data obtained from literature. The control of three-phase and single-phase rectifiers was implemented using two cascaded control loops, with the inner loop controlling the current drawn from the network and the outer loop regulating the DC bus voltage. For the single-phase inverter, PI RES and PI SRF controls were implemented, targeting a comparison in terms of performance. In order to validate the mathematical model of the system and the control algorithm, simulations were performed using the PSIM software. The analysis showed that the simulation results were consistent with the mathematical modeling of the system.

La fonte et l’effondrement du pergélisol riche en glace dans la région subarctique du Québec ont donné lieu à la formation de petits lacs (mares de thermokarst) qui émettent des gaz à effet de serre dans l’atmosphère tels que du dioxyde de carbone et du méthane. Pourtant, la composition de la communauté microbienne qui est à la base des processus biogéochimiques dans les mares de fonte a été très peu étudiée, particulièrement en ce qui concerne la diversité et l’activité des micro-organismes impliqués dans le cycle du méthane. L’objectif de cette thèse est donc d’étudier la diversité phylogénétique et fonctionnelle des micro-organismes dans les mares de fonte subarctiques en lien avec les caractéristiques de l’environnement et les émissions de méthane. Pour ce faire, une dizaine de mares ont été échantillonnées dans quatre vallées situées à travers un gradient de fonte du pergélisol, et disposant de différentes propriétés physico-chimiques. Selon les vallées, les mares peuvent être issues de la fonte de palses (buttes de tourbe, à dominance organique) ou de lithalses (buttes de sol à dominance minérale) ce qui influence la nature du carbone organique disponible pour la reminéralisation microbienne. Durant l’été, les mares étaient fortement stratifiées; il y avait un fort gradient physico-chimique au sein de la colonne d’eau, avec une couche d’eau supérieure oxique et une couche d’eau profonde pauvre en oxygène ou anoxique. Pour identifier les facteurs qui influencent les communautés microbiennes, des techniques de séquençage à haut débit ont été utilisées ciblant les transcrits des gènes de l’ARNr 16S et des gènes impliqués dans le cycle du méthane : mcrA pour la méthanogenèse et pmoA pour la méthanotrophie. Pour évaluer l’activité des micro-organismes, la concentration des transcrits des gènes fonctionnels a aussi été mesurée avec des PCR quantitatives (qPCR). Les résultats montrent une forte dominance de micro-organismes impliqués dans le cycle du méthane, c’est-à-dire des archées méthanogènes et des bactéries méthanotrophes. L’analyse du gène pmoA indique que les bactéries méthanotrophes n’étaient pas seulement actives à la surface, mais aussi dans le fond de la mare où les concentrations en oxygène étaient minimales; ce qui est inattendu compte tenu de leur besoin en oxygène pour consommer le méthane. En général, la composition des communautés microbiennes était principalement influencée par l’origine de la mare (palse ou lithalse), et moins par le gradient de dégradation du pergélisol. Des variables environnementales clefs comme le pH, le phosphore et le carbone organique dissous, contribuent à la distinction des communautés microbiennes entre les mares issues de palses ou de lithalses. Avec l’intensification des effets du réchauffement climatique, ces communautés microbiennes vont faire face à des changements de conditions qui risquent de modifier leur composition taxonomique, et leurs réponses aux changements seront probablement différentes selon le type de mares. De plus, dans le futur les conditions d’oxygénation au sein des mares seront soumises à des modifications majeures associées avec un changement dans la durée des périodes de fonte de glace et de stratification. Ce type de changement aura un impact sur l’équilibre entre la méthanogenèse et la méthanotrophie, et affectera ainsi les taux d’émissions de méthane. Cependant, les résultats obtenus dans cette thèse indiquent que les archées méthanogènes et les bactéries méthanotrophes peuvent développer des stratégies pour survivre et rester actives au-delà des limites de leurs conditions d’oxygène habituelles. / The thawing and collapse of ice-rich permafrost in the subarctic region of Quebec has given rise to thaw ponds (thermokarst ponds) that emit the greenhouse gases carbon dioxide and methane to the atmosphere. However, the microbial community composition that underlies biogeochemical processes in thaw ponds has been little investigated, particularly concerning the diversity and activity of micro-organisms involved in the methane cycle. The objective of this thesis study was to determine the phylogenetic and functional diversity of micro-organisms in subarctic thaw ponds, and the relationships with environmental properties and methane emission. To that aim, we sampled ten thaw ponds in four different valleys located across a permafrost degradation gradient with distinct physico-chemical properties. Depending on valley, the ponds were derived either from the thawing of a palsa (peat-mound) or lithalsa (mineral-mound), which influenced the nature of organic carbon available for microbial remineralization. During summer, the ponds were observed to be well-stratified; there were with strong physico-chemical gradients down the water column, with an upper oxic layer and a bottom low oxygen or anoxic layer. To identify the factors influencing microbial community composition, we used high throughput sequencing techniques targeting transcripts of 16S rRNA gene, and additionally targeted genes involved in the methane cycle: mcrA for methanogenesis and pmoA for methanotrophy. As a proxy of microbial activity, we also measured the concentration of functional gene transcripts using with quantitative PCR (qPCR). The results showed a striking dominance of micro-organisms involved in the methane cycle, namely methanogenic Archaea and methanotrophic Bacteria. The pmoA analyses implied that methanotrophic Bacteria were not only active in the surface, but also in the bottom waters where oxygen concentrations were minimal; this was unexpected given their need for oxygen in methane consumption. In general, the microbial community properties were largely determined by the origin of the ponds (palsa versus lithalsa), and much less so by the extent of permafrost degradation. The key environmental variables pH, phosphorus and dissolved organic carbon likely contributed to the differentiation of microbial community between the palsa and lithalsa valleys. With intensification of climate warming, these microbial communities will face changing conditions that are likely to modify their taxonomic composition, and these responses are likely to differ between ponds in the two landscape types. Oxygenation of the ponds will likely be subject to major shifts in the future associated with changes in the duration of the ice-free season and the extent of stratification. Such changes will impact the balance between methanogenesis and methanotrophy, and thereby affect the net rates of methane emission. However, the results obtained here indicate that methanogenic Archaea and methanotrophic Bacteria have strategies to survive and remain active beyond the limit of their usual oxygen preferences.

MODIS est une méthode fiable et précise utilisée couramment pour suivre l'évolution du couvert nival au-dessus de bassins versants alpins. Toutefois, cette méthode de télédétection possède quelques limitations importantes, tel que l'inhabilité à distinguer la neige humide de la neige sèche, qui pourrait être mieux prise en compte par l'utilisation d'une méthode de télédétection complémentaire telle que l'imagerie par radar à synthèse d'ouverture (RSO). Le site d'étude utilisé pour le projet est le bassin versant de la rivière Nechako, situé dans la chaîne Côtière de la Colombie-Britannique, qui est caractérisé par un manteau neigeux pouvant atteindre plusieurs mètres d’épaisseur en montagne. Quinze images RADARSAT-2 en mode ScanSAR Wide ont été obtenues en polarisation VV et VH entre les mois de mars et juillet 2012. Elles ont été traitées à l'aide d'un algorithme basé sur la méthode de Nagler et Rott pour distinguer la neige humide de la neige sèche, mais qui utilise un seuil graduel plutôt que le seuil de -3 dB fréquemment utilisé. Les cartes de neige humide qui découlent de cette technique correspondent mieux aux incertitudes retrouvées sur le bassin en raison de la présence importante de forêts de conifères et de régions montagneuses. Les cartes ont été combinées au produit de neige de MODIS, afin d'utiliser son habileté à détecter le couvert nival avec précision pour corriger les zones de bruit des images RSO, causées entre autres par des sols gorgés en eau. Afin d'aider l'analyse des images RSO, une modélisation fine du manteau neigeux a été effectuée avec le logiciel Crocus afin de procéder à une analyse détaillée de l’évolution des caractéristiques du manteau neigeux, notamment du contenu en eau liquide de la neige, tout au long de l’hiver. La modélisation a été effectuée à l'emplacement de trois coussins à neige sur le bassin versant et est réalisée grâce à l'utilisation de données du North American Regional Reanalysis (NARR). À partir des résultats du modèle Crocus et de l'équivalent en eau observé aux coussins à neige, une relation a été établie entre la détection de neige humide en montagne par RADARSAT-2 et le ruissellement reçu au réservoir de la rivière Nechako. Avec le jeu de données actuel, le ruissellement maximal reçu au réservoir a été prévu avec une précision de 10 jours. Il est prévu que davantage d'années d’images radar pourraient permettre de confirmer et de réduire cet intervalle.