Made available in DSpace on 2015-05-14T12:09:36Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 5605900 bytes, checksum: 3d25379d8d6eb2b9a5d187b806036974 (MD5) Previous issue date: 2014-05-27 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / This study analyzed the influence of the soil s usage and occupation on the surface temperature variance and on radiation balance in Joao Pessoa city. For that, it was used remote sensing techniques, Geographic Information Systems and SEBAL algorithm. For the soil s usage and occupation classification, and the determination of emissivity, albedo, NDVI, IAF and SAVI were used, and for surface temperature and radiation balance, Landsat 5-TM satellite s images was used, referring to november 5, 1991; August 26, 2006; and October 8, 2010.The usage and occupation of João Pessoa s soil analyses showed that during the years of 1991, 2006 and 2010, the city had a considerable increase of the urban mesh above herbaceous vegetation and bare soil. The surface temperature has shown greater values above urban area and bare soil on the last three years, above all on 1991 which obtained maximum temperature values, with average of 31.4°C above urban area and 31.1°C above bare soil. The radiation balance images and statistic values has shown that the highest values of Rn was founded above arboreal/shrubby vegetation and water for all the years, above all on 2010 when was obtained the highest values of Rn for all soil usages, which the averages were 717.5W/m² above water, and 703.2W/m² above arboreal/shrubby vegetation. The relationship between the albedo of soil usages (urban and herbaceous vegetation) with the temperature and radiation balance, has shown that it was satisfactory, indicating the method efficiency for estimate the quoted parameters with the targets analyzed for João Pessoa city. The temperature and radiation balance results found on this study, obtained through SEBAL and satellite images, allows to conclude that the values has shown match variances with those showed in the literature for similar soil usages / Este trabalho analisou a influência do uso e ocupação do solo na variabilidade da temperatura da superfície e no saldo de radiação na cidade de João Pessoa. Para tanto, foram utilizadas técnicas de sensoriamento remoto, Sistemas de Informações Geográficas e o algoritmo SEBAL. Para a classificação do uso e ocupação do solo e a determinação dos valores de emissividade, albedo, NDVI, IAF, SAVI, temperatura de superfície e saldo de radiação à superfície, foram utilizadas imagens do satélite Landsat 5-TM referente aos dias 5 de novembro de 1991, 26 de agosto de 2006 e 8 de outubro de 2010.A análise do uso e ocupação do solo da cidade de João Pessoa apontou que no decorrer dos anos de 1991, 2006 e 2010 a cidade teve um considerável crescimento da malha urbana sobre áreas de vegetação herbácea e solo exposto. A temperatura da superfície apresentou maiores valores sobre área urbana e solo exposto nos três anos, sobretudo em 1991 que obteve valores máximos de temperatura, com média de 31,4ºC sobre área urbana e de 31,1ºC sobre solo exposto. As imagens do saldo de radiação e os valores estatísticos mostraram que os maiores valores de Rn foram encontrados sobre vegetação arbórea/arbustiva e água para todos os anos, sobretudo em 2010 onde foram obtidos os maiores valores de Rn para todos os usos do solo, cujas médias foram de 717,5W/m² sobre água e 703,2W/m² sobre vegetação arbórea/arbustiva. A relação entre o albedo dos usos do solo (urbano e vegetação herbácea) com a temperatura e o saldo de radiação, mostrou-se satisfatórios, indicando a eficiência do método na estimativa dos parâmetros citados com os alvos analisados para a cidade de João Pessoa. Os resultados de temperatura e saldo de radiação encontrados nesta pesquisa, obtidos através do SEBAL e das imagens de satélite, permitem concluir que os valores apontaram variações condizentes com os apresentados na literatura para usos do solo similares

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / This research was conducted with the objective of studying the dynamics of soil temperature and water content as a function of the different amounts of black oat residue (Avena sativa Schreb.) in the soil cover, through existing relationships with meteorological variables, as well as the simulation of these dynamics using Hydrus-1D. The experiment was carried out in an Argissolo Vermelho Distrófico arênico (Rhodic Paleudalf) in Santa Maria - RS. The treatments consisted of the use of 3, 6 and 9 Mg ha-1 of oat straw in soil cover and an uncovered soil treatment (0 Mg ha-1). Soil temperature and water content at different depths, in addition to meteorological conditions, were measured during the period from December 1, 2014 to December 6, 2015. The effect of straw on soil temperature was analyzed at different depths, correlating covered soil temperature with uncovered soil temperature. The equations used for estimating soil surface temperature were adjusted by the least square method, where the surface temperature is estimated as a function of air temperature and incident solar radiation, and its performance compared to the sinusoidal equation. The hydraulic and thermal parameters required for the simulation of soil temperature and water content were optimized by the reverse solution and validated on uncovered soils. The optimized parameters were used to simulate the temperature and water content in soils with different amounts of straw with two contour conditions for the soil surface temperature, one measure and the other estimated. The straw affected the soil temperature down to 50 cm deep. The uncovered soil temperature at which inversion of the straw effect occurred was 19.6 ° C at the surface and 15.8 ° C at 50 cm depth. The presence of straw keeps the soil warmer than the soil without straw at the beginning of the day, but prevents it from warming similar to the ground without straw for the rest of the day, causing in turn that, bare soil gets warmer during the day and cools more at night compared to the covered soil. This temperature was similar in all amounts of straw, but decreased in depth. The temperature of the soil surface can be estimated as a function of air temperature and solar radiation. In uncovered soil the proposed equation presents better performance compared to the sinusoidal equation. In covered soils the sinusoidal equation had better performance. The optimal hydraulic and thermal parameters of the soil improved the simulations of temperature and soil water content. The soil temperature with different amounts of oat straw can be simulated using Hydrus-1D, and the use of the contour condition with the measured surface temperature performed better than the estimated surface temperature. / Esta pesquisa foi conduzida com o objetivo de estudar a dinâmica da temperatura e do conteúdo de água do solo em função de diferentes quantidades de resíduos de aveia preta disposta em cobertura por meio das relações existente com variáveis meteorológicas, bem como a representação destes comportamentos por meio de simulação utilizando o Hydrus-1D. O experimento foi conduzido em um Argissolo Vermelho Distrófico arênico em Santa Maria - RS. Os tratamentos consistiram da utilização de três diferentes níveis de cobertura do solo (3, 6 e 9 Mg ha-1) com palha de aveia (Avena sativa) e um tratamento com solo descoberto (0 Mg ha-1). A temperatura do solo e o conteúdo volumétrico de água, em diferentes profundidades, além das condições meteorológicas, foram mensurados durante o período de 01/12/2014 a 06/12/2015. O efeito da cobertura do solo com resíduos vegetais na temperatura do solo foi analisado nas diferentes profundidades correlacionando a temperatura do solo com palha e a temperatura do solo descoberto. O método dos mínimos quadrados foi usado para gerar equações de estimativa da temperatura da superfície do solo em função da temperatura do ar e da radiação solar e o seu desempenho foi comparado com a equação senoidal. Os parâmetros hidráulicos e térmicos, necessários para a simulação da temperatura e do conteúdo de água do solo, com o Hydrus-1D, foram otimizados por meio da solução inversa e validados em solos descobertos. Os parâmetros otimizados foram usados para a simulação da temperatura e do conteúdo de água em solos com diferentes quantidades de palha com duas condições de contorno para a temperatura da superfície do solo, uma medida e outra estimada. A palha exerce influência sobre a temperatura do solo até 50 cm de profundidade. A temperatura do solo descoberto em que ocorre a inversão do efeito da palha foi de 19,6 °C na superfície e 15,8 °C a 50 cm de profundidade. A presença de cobertura de palha mantém o solo mais aquecido que o solo sem palha no início do dia, mas impede que ele se aqueça igual ao solo sem palha no restante do dia. Assim resultando em uma percepção de que a palha aquece e resfria o solo em relação à temperatura do solo desnudo. Essa temperatura foi semelhante em todas as quantidades de palha, porém decresceu em profundidade. A estimativa da temperatura da superfície do solo pode ser estimada em função da temperatura do ar e da radiação solar. Em solo descoberto as equações propostas tem melhor desempenho que a equação senoidal. Em solos cobertos a equação senoidal teve melhor desempenho. Os parâmetros hidráulicos e térmicos do solo otimizados melhoraram as simulações da temperatura e conteúdo de água do solo. A simulação da temperatura do solo com diferentes quantidades de cobertura vegetal pode ser simulada usando o Hydrus-1D, sendo que o uso da condição de contorno com a temperatura da superfície medido teve melhor desempenho que o utilizando a temperatura da superfície estimada.

A gestão de recursos hídricos tornou-se cada vez mais complexa devido ao rápido crescimento sócio-econômico e as mudanças ambientais nas bacias hidrográficas nas últimas décadas. Modelos computacionais são importantes ferramentas de suporte na gestão de recursos hídricos e tomada de decisões devido a sua funcionalidade, provendo informações importantes sobre os principais processos físicos, químicos e biológicos, e permitindo melhorar o entendimento desses processos, os quais ocorrem em diferentes escalas espaciais e temporais. Na presente tese, o objetivo foi compreender o funcionamento hidrológico do sistema integrado bacia hidrográfica - lagoa, e os efeitos na hidrodinâmica do lago, utilizando como suporte o acoplamento da modelagem hidrológica - hidrodinâmica, e o uso de técnicas de sensoriamento remoto para o monitoramento de parâmetros de qualidade da água (e.g., clorofilaa, temperatura da superfície d’água e níveis da água). A área de estudo é a bacia hidrográfica da Lagoa Mirim, localizada no sul do Brasil, possuindo uma área total de 58.000 km2 (56% no Uruguai e o restante no Brasil). Foram propostos e testados modelos empíricos para estimativa de clorofila-a emumlago raso subtropical, baseados em imagens do sensor MODIS e técnicas estatísticas. Além disso, foi desenvolvido e avaliado o acoplamento da modelagem hidrológica-hidrodinâmica de grande escala e o sensoriamento remoto. O modelo hidrológico distribuído de grande escala MGBIPH acoplado com o modelo hidrodinâmico IPH-ECO foi utilizado para simular a bacia hidrográfica e os principais componentes hidrodinâmicos da Lagoa Mirim. O modelo mostrou bom desempenho quando comparado com observações de vazões, além de dados provenientes de sensoriamento remoto, através de altimetria espacial. As simulações mostraram importantes aspectos sobre a estrutura de fluxo, campos de velocidade e níveis d’água na lagoa, assim como a influência de grandes rios, forçantes externas como o vento (intensidade e direção) e o impacto do estressor antrópico (retiradas para irrigação) no sistema. As simulações permitiram avaliar aspectos relacionados com as variações espaciais e temporais (diurna, mensal, sazonal e inter-anual) da temperatura da superfície da água, a dinâmica dos fluxos de calor (sensível e latente) e os efeitos de eventos meteorológicos de pequena escala como frentes frias, os quais têm um impacto significativo sobre a temperatura superficial da água e os fluxos de calor na lagoa. Quanto aos modelos empíricos para estimativa de clorofila-a a partir do MODIS, os resultados mostram que um simples e eficiente modelo desenvolvido a partir de análise de regressão múltipla, apresentou ligeiras vantagens sobre os modelos de redes neurais artificiais, modelos multiplicativos não paramétricos e modelos empíricos (e.g., Appel, Kahru, FAI e O14a) usualmente utilizados na estimativa de Chl-a em ambientes aquáticos. Resultados também indicam que é inapropriado generalizar um único modelo desenvolvido a partir do conjunto total de dados, para estimar concentrações de Chl-a na lagoa, o que corrobora a heterogeneidade espacial na distribuição de Chl-a e as diferenças entre regiões (litoral e pelágica). A modelagem hidrológica-hidrodinâmica de grande escala apoiada por informação de sensoriamento remoto, mostrou ser uma abordagem promissora para o entendimento da estrutura e funcionamento de lagoas rasas de grande porte e longo prazo, úteis para a gestão integrada dos recursos hídricos. / The last decade, the water resource management is being complex due to the rapid socioeconomic development and environmental changes in river basins. Computations models are important support tools in water resource management and make decision providing important information and allowing a better comprehension of the physical, chemical and biologic processes, which occur in di erent temporal/spatial scales. In this thesis, the objective was to understand the hydrological functioning of the integrated basin- lake system and its e ects on hydrodynamics, using hydrodynamic - hydrodynamic modeling and water quality monitoring (e.g., chlorophyll-a, water surface temperature and water levels) from remote sensing techniques. The study area is the Lake Mirim basin, located between Brazil and Uruguay (basin total area 58.000 km2). Empirical models were proposed and tested to chlorophyll-a estimation in a shallow subtropical lake, based on MODIS imagery and statistics techniques. In addition, we developed and assessed the coupling of large scale hydrological/hydrodynamic modeling and remote sensing techniques. The large-scale distributed hydrological model MGB-IPH coupled with the hydrodynamic model IPHECO were used to simulate the river basin and the hydrodynamic components of the Lake Mirim. The coupled model showed good performance when compared to in-situ measurements and satellite altimetry data. The simulations showed important aspects relate to flow structure, velocity fields and lake water levels, as well as the influence of large rivers, external forcing as such the wind (intensity and direction), and the impact of anthropogenic stressors (irrigation withdrawals) in the system. The simulations allowed assessing the spatial and temporal variations (diurnal, monthly, seasonal and inter-annual) in the water surface temperature, heat fluxes dynamics (sensible and latent) and the e ects of short time-scale events, as such cold fronts passages over the lake, which cause strong impacts on the water surface temperature and heat fluxes in the lake. Regarding the empirical models developed to chlorophyll-a estimation from MODIS imagery, the results showed that a simple and e cient model developed from multiple regression analysis, performed best in comparison with artificial neural network models, non-parametric multiplicative models, and empirical models (e.g., Appel, Kahru, FAI and O14a) common used in the Chl-a estimation in aquatics environments. Results also indicated that is inappropriate to generalize a single model developed from the total datasets to estimates Chl-a in the lake, which corroborates the spatial heterogeneity (Chl-a distribution) and the di erences among regions (littoral and pelagic). The synergy between large-scale hydrological-hydrodynamic modeling, in situ measurements and remote sensing techniques provided a promising approach to improve the comprehension of the structure and ecosystem functioning of large shallow lakes in long-term time scale, useful to water resources management.

Interannual variation of Indian summer (June-September: JJAS) monsoon rainfall (ISMR) depends on its relative intensity during early (June-July: JJ; contribution 52%) and late (August-September: AS; contribution 49%) phases. Apart from variations in sea surface temperature (SST), the primary reasons behind the variability during JJ and AS can be very different due to change in climatic conditions on account of post-onset processes. Here, using a high resolution general circulation model with seasonally varying climatological SST, mechanisms those govern the intensity of rainfall during JJ and AS are investigated. There is no significant relation-ship between intensity of precipitation over Indian region in JJ and AS. Moreover, the factors determining early monsoon (JJ) precipitation are different than that for late monsoon (AS). In absence of interannual SST variation, pre-monsoon soil moisture do not play a significant role for the interannual variation of monsoon precipitation over India. A large scale oscillation of the ITCZ is noticed on interannual time scale spanning from around 60◦E to 150◦E that brings spatially coherent flood and drought over this region. Early monsoon precipitation has a larger dependency on spring snow depth over Eurasia and phase of the upper tropospheric Rossby wave in May. However, late monsoon precipitation over India is mainly governed by the intensity and time scale of the intraseasonally varying convective cloud bands. This study suggests that early monsoon (JJ) precipitation over Indian region is more correlated with pre-monsoon signatures of land-atmosphere parameters. However, in later parts after the onset (AS), the monsoon intensity is primarily driven by its internal dynamics and characteristics of intraseasonal oscillation.

Made available in DSpace on 2014-08-20T14:25:45Z (GMT). No. of bitstreams: 1 dissertacao_catia_cristina_braga_rodrigues.pdf: 1202246 bytes, checksum: 0067166e90de2215d4ec858dba3e1ba5 (MD5) Previous issue date: 2012-04-18 / The periods of large variations in rainfall in the Mato Grosso do Sul State (centralwest region, Brazil), in particular the great floods or droughts, generate large economic losses to the state. The summer drought may further compromise the winter period, which often short rains occur. The main objective of this research is to verify bimonthly rainfall variation in the state of Mato Grosso do Sul during the summer (maximum variation) and the winter (minor variations) and their relationships with the bimonthly SST anomalies combined of the South Pacific and South Atlantic oceans. The combined anomalies were represented by Principal Oscillation Patterns (POPs), which represent the predominant patterns of ocean circulation. Monthly precipitation data from January to December (from 1981 to 2009) and monthly data of Sea Surface Temperature (SST) were used. From the rainfall climatology of the marking periods were selected two periods: December/January (rainier) and July/August (dry season). Homogeneous regions were generated by the K-means Method, three for December/January period and two for the July/August period. The regional accumulated rainfall presented no significant trend at 5% probability. The regional accumulated precipitation anomalies of two-month periods showed no significant correlations with SST anomalies in the Central Pacific (Niño 3.4) along the studied period. During the two months December/January, the largest excess precipitation in the region occurred with a predominance of strong El Nino events and droughts in major events La Niña and neutral. In regions 2 and 3 there is not a predominant sign of some events. During the July/ August period, the greatest excesses of precipitation occurred in the region with a predominance of La Niña events, with less relation respect region 2, both regions in a coherent signal (Neutral and La Niña). In cases of severe dry period, both regions have low relation with the events signals. The settings of multiple regressions of bi-monthly precipitation accumulated in homogeneous regions with the scores of the first six patterns of SST showed significant increase in both studied marking periods. The correlation coefficient between the bimonthly observed accumulated precipitation and regional estimates were significant at 1% probability in all regions. This occur especially in the region that represents the northwest quarter of the state in December/January and in the region that represents the major portion of Mato Grosso do Sul State in the July/August period. / Os períodos de grandes variações da precipitação no Mato Grosso do Sul, em especial as grandes cheias ou grandes estiagens, trazem grandes prejuízos financeiros ao Estado. As estiagens de verão podem comprometer ainda mais o período de inverno, no qual já é normal ocorrer poucas chuvas. Esta pesquisa tem por proposta principal verificar as variações bimensais da precipitação no estado de Mato Grosso do Sul nos períodos de verão (variações máximas) e inverno (variações mínimas) e suas relações com as anomalias bimensais de TSM combinadas dos oceanos Pacífico Sul e Atlântico Sul. As anomalias combinadas foram representadas pelos Padrões de Oscilação Principal (POP), os quais representam os padrões predominantes de circulação oceânica. Foram usados dados de precipitação mensal de janeiro a dezembro, período de 1981 a 2009 e dados médios mensais de Temperatura na Superfície do Mar (TSM). A partir da climatologia da precipitação foi selecionado os bimestres dezembro/janeiro (mais chuvoso) e julho/agosto (menos chuvoso) e geradas pelo método K-means regiões homogêneas, sendo três para o bimestre dezembro/janeiro e duas para o bimestre julho/agosto. As precipitações acumuladas regionais não apresentaram tendência temporal significativas a 5% de probabilidade. As anomalias de precipitação acumulada regionais dos bimestres não apresentaram correlações significativas com as anomalias de TSM na região do Pacífico Central (Niño 3.4) ao longo do período estudado, mas durante o bimestre dezembro/janeiro, os maiores excessos de precipitação da região 1 ocorreram com forte predomínio de eventos El Niño fortes e a maiores estiagens em Eventos La Niña e Neutros. Nas regiões 2 e 3 não existem um sinal predominante de alguns dos eventos. Durante o bimestre julho/agosto, os maiores excessos de precipitação ocorreram na região 1 com predomínio de eventos La Niña, com menor relação na região 2, mas ambas regiões com coerência do sinal (Neutro e La Niña). Nos casos de fortes estiagens, ambas as regiões apresentam baixa relação com os sinais dos eventos. Os ajustes das regressões múltiplas das precipitações acumuladas bimestrais nas regiões homogêneas com os escores dos seis primeiros padrões de TSM apresentaram aumento de significância em ambos bimestres estudados. O coeficiente de correlação entre as precipitações acumuladas regionais bimestrais observadas e estimadas foram significativas a 1% de probabilidade em todas as regiões, especialmente na região que representa o noroeste do Estado no bimestre dezembro/janeiro e a região que representa a maior parte do estado do Mato Grosso do Sul no bimestre julho/agosto.

La caractérisation des dynamiques de sous-mésoéchelle (<10km) à la surface de l'océan et leurs impacts sur les processus océaniques globaux sont des enjeux scientifiques majeurs. L'imagerie satellitaire est un outil essentiel dans ce contexte, qui présente toutefois des limitations liées aux instruments de télédétection. Dans le cas des images de température de surface des océans (SST), les mesures satellitaires des structures océaniques sont limitées par la résolution grossière des capteurs micro-ondes (~50km) d'une part, et par la sensibilité aux conditions climatiques (e.g., couverture nuageuse) des instruments de mesure infrarouge haute-résolution. Dans cette thèse, nous nous intéressons à l'analyse, la modélisation et la reconstruction des structures turbulentes haute-résolution capturées par imagerie satellitaire de SST, et proposons quatre contributions principales. Dans un premier temps, nous développons une méthode de filtrage conjointe Fourier-ondelettes pour le prétraitement d'artefacts géométriques dans les observations satellitaires infrarouges. Dans un deuxième temps, nous nous focalisons sur la caractérisation de la variabilité géométrique de champs de température de surface (SST) en utilisant des modèles de marches aléatoires appliqués aux lignes de niveaux. En particulier, nous considérons des processus aléatoires de type schramm Loewner (SLE). Nous nous intéressons ensuite à la modélisation stochastique des variabilités inter-échelles de champs de SST. Des modèles stochastiques de textures multivariées sont introduits. Ces modèles permettent de reproduire des propriétés statistiques et spectrales similaires à celles des données ayant servi à les calibrer. Nous développons ensuite des méthodes de super-résolution de champs de SST conditionnellement à une observation basse-résolution. Nous utilisons des modèles multivariés de textures formulés dans le domaine des ondelettes, en exploitant l'apprentissage d'à priori statistiques (i.e., covariances et covariances croisées) des différentes sous-bandes à partir d'images haute-résolution. Des contraintes supplémentaires imposées sur la phase de Fourier des différentes sous-bandes simulées permettent la reconstruction de structures géométriques marquées tels que les fronts. Nous démontrons la pertinence de la méthode proposée sur des images satellitaires de SST obtenues à partir du capteur Modis/Aqua. / The characterization of sub-mesoscale dynamics (<10 km) in the ocean surface and their impact on global ocean processes are major scientific issues. Satellite imagery is an essential tool within this framework. However, the use of remote sensing techniques still raise challenging. For instance, regarding Sea Surface Temperature (SST) images, satellite measurements of oceanic structures are limited by the coarse resolution of microwave sensors (~50km) on one hand, and by sensitivity to climatic conditions (eg., Cloud cover) of high-resolution infrared instruments on the other hand. In this thesis, we are interested in analysis, modeling and reconstruction of high-resolution turbulent structures captured by satellite SST imagery. In this context, we propose four main contributions. First, we develop a joint Fourier-Wavelet filtering method for the pre-processing of geometrical noises in satellite-based infrared observations, namely the striping noises. Secondly, we focus on the characterization of the geometric variability of sea surface temperature (SST) fields using random walk models applied to SST isolines. In particular, we consider the class of Schramm Loewner evolution curves (SLE). We then focus on the stochastic modeling of the cross-scale variabilities of SST fields. Stochastic multivariate texture-based models are introduced. These models are designed to reproduce several statistics and spectral properties that are observed on the data that are used to calibrate the model. We then develop our framework for stochastic super-resolution of SST fields conditionally to low-resolution observations. We use multivariate texture-based models formulated in the wavelet domain. These models exploit the formulation of statistical and spectral priors (i.e., covariances and cross-covariances) on wavelet subbands. These priors are directly learned from exemplar high-resolution images. Additional constraints imposed on the Fourier-phase of the different simulated subbands allow the reconstruction of coherent geometric structures such as the edge information. Our method is tested and validated using infrared high-resolution satellite SST images provided by Aqua Modis sensor.

Reconstruire des champs géophysiques à partir d'observations bruitées et partielles est un problème classique bien étudié dans la littérature. L'assimilation de données est une méthode populaire pour aborder ce problème, et se fait par l'utilisation de techniques classiques, comme le filtrage de Kalman d’ensemble ou des filtres particulaires qui procèdent à une évaluation online du modèle physique afin de fournir une prévision de l'état. La performance de l'assimilation de données dépend alors fortement de du modèle physique. En revanche, la quantité de données d'observation et de simulation a augmenté rapidement au cours des dernières années. Cette thèse traite l'assimilation de données d'une manière data-driven et ce, sans avoir accès aux équations explicites du modèle. Nous avons développé et évalué l'assimilation des données par analogues (AnDA), qui combine la méthode des analogues et des méthodes de filtrage stochastiques (filtres Kalman, filtres à particules, chaînes de Markov cachées). Des applications aux modèles chaotiques simplifiés et à des études de cas de télédétection réelle (température de surface de lamer, anomalies du niveau de la mer), nous démontrons la pertinence d'AnDA pour l'interpolation de données manquantes des systèmes dynamiques non linéaires et à haute dimension à partir d'observations irrégulières et bruyantes.Motivé par l'essor du machine learning récemment, la dernière partie de cette thèse est consacrée à l'élaboration de modèles deep learning pour la détection et de tourbillons océaniques à partir de données de sources multiples et/ou multi temporelles (ex: SST-SSH), l'objectif général étant de surpasser les approches dites expertes. / Reconstructing geophysical fields from noisy and partial remote sensing observations is a classical problem well studied in the literature. Data assimilation is one class of popular methods to address this issue, and is done through the use of classical stochastic filtering techniques, such as ensemble Kalman or particle filters and smoothers. They proceed by an online evaluation of the physical modelin order to provide a forecast for the state. Therefore, the performanceof data assimilation heavily relies on the definition of the physical model. In contrast, the amount of observation and simulation data has grown very quickly in the last decades. This thesis focuses on performing data assimilation in a data-driven way and this without having access to explicit model equations. The main contribution of this thesis lies in developing and evaluating the Analog Data Assimilation(AnDA), which combines analog methods (nearest neighbors search) and stochastic filtering methods (Kalman filters, particle filters, Hidden Markov Models). Through applications to both simplified chaotic models and real ocean remote sensing case-studies (sea surface temperature, along-track sea level anomalies), we demonstrate the relevance of AnDA for missing data interpolation of nonlinear and high dimensional dynamical systems from irregularly-sampled and noisy observations. Driven by the rise of machine learning in the recent years, the last part of this thesis is dedicated to the development of deep learning models for the detection and tracking of ocean eddies from multi-source and/or multi-temporal data (e.g., SST-SSH), the general objective being to outperform expert-based approaches.

Porites lobata is an important reef building coral in the tropical eastern Pacific and the dominant Porites species in the Galápagos archipelago. Following the 1982-83 El Niño-Southern Oscillation the Galápagos Islands experienced 97-99% coral mortality, leaving many areas throughout the archipelago denuded of corals. Because very few long term assessments have been conducted on the growth and resilience of P. lobata to natural disturbances in the Galápagos Islands (Glynn et al., 2001; Glynn et al., 2009), benthic surveys were performed on a uniquely dense aggregation of P. lobata colonies at Devil’s Crown, Floreana Island between 1993 and 2011. Annual changes in live tissue area were calculated for the majority of the population (n=17) using Coral Point Count with Excel extensions (CPCe 3.6) software to determine growth and recovery trends for this aggregation. Total live tissue area (n=10) increased from 1993 to 2011, however due to high interannual variability this increase was not significant. Within this overall pattern, a general trend of decline was observed in live tissue cover from 1993 to 2000, with increases in tissue area observed from 2000 to 2011. Severe bleaching (85-100%) was observed during the 1998 survey, followed by 42% tissue loss (n=10), coinciding with sea water warming associated with the very strong 1997-1998 El Niño-Southern Oscillation event. Subsequent regrowth of coral tissue was observed during the 2001 survey with continued recovery through 2009. Multiple comparison testing revealed a significant difference between the impacted state (1999) and the recovered state (2009), (p = 0.002, Dunn’s method, n=17), suggesting this aggregation required a period of ten years to recover from this disturbance. During this recovery period the moderately strong 2007-2008 La Niña, with accompanying stressful low temperatures, occurred but did not interrupt tissue regrowth. Warmer than average sea surface temperatures occurred during the warm months from 2008 to 2011, during which time a cool period occurred from 2010 to 2011. While the magnitude and duration of temperature anomalies during warming were not as great as those observed during the 1997-98 ENSO, low temperatures observed during the cool period were similar to those experienced throughout the 2007-08 La Niña. During this time total live tissue cover was reduced by 19% (n=10); however it is unknown whether this was due to warming or the following cool period. Based on results from the 1997-98 El Niño and 2007-08 La Niña, this reduction in live tissue was most likely caused by elevated sea surface temperatures. Data on the growth and resilience of P. lobata populations at Devil’s Crown will be used for conservation and management of this important resource.

Reproductive success in loggerhead (Caretta caretta) sea turtles is strongly dependent on the effective placement and internal conditions of their nests. Embryos rely on optimal incubation conditions for proper development and growth, which determines how many hatchlings will emerge from the nest. The internal microclimate of each nest is delicately balanced and can be easily influenced by external environmental conditions. This study was designed to examine several environmental variables and determine their effects on sea turtle nesting numbers, hatching success, and incubation conditions in Broward County Florida. Over a span of 25 years (1991-2015), the Broward County Sea Turtle Conservation Program has collected data on each sea turtle nest laid in Broward County. This data was analyzed and plotted to visualize nesting and hatching trends, and regressions were fitted to make comparisons to historic air temperature, sea surface temperature, precipitation, and lunar illumination data. These regressions were tested for significance, and each environmental variable was found to have varying levels of impact on sea turtle nesting and hatching behavior. Of the environmental variables considered in this study, analyses suggest that sea turtles are most responsive to temperature, with sea surface temperature serving as the best proxy for predicting nesting behaviors. Air temperature over the incubation period was found to be the best indicator for hatch success percentage. Air temperature, sea surface temperature, and precipitation averages all significantly affected the length of the incubation period. The regression models created in this study could be used to examine the interactions between climatic variables, and to indicate what impacts can be expected by these various environmental factors. This information could be used to estimate the future effects of climate change on sea turtle reproduction, and to predict general reproductive success and future population trends.

No description available.