Neste estudo, trata-se de avaliar fontes de precipitação baseadas em estimativas por satélite e técnicas de assimilação de dados para previsão de vazões por meio do modelo hidrológico distribuído MGB-IPH. A insuficiente representatividade espacial dos pluviômetros torna difícil a correta representação dos campos de precipitações. Por outro lado, as estimativas de satélite, embora forneçam uma descrição espacial mais consistente, são potencialmente menos acuradas. Sendo assim, procura-se utilizar métodos que combinem os dados de ambas as fontes para gerar um campo de precipitação mais consistente. Neste trabalho, implementaramse dois modelos de combinação pluviômetro-satélite, CHUVSAT e MERGEHQ, através de uma metodologia de interpolação. Por outro lado, as técnicas de assimilação de dados acoplados aos modelos de previsão hidrológica são também de interesse neste estudo, pois minimizam as incertezas associadas ao processo de calibração de parâmetros, às variáveis de estado e dados de entrada do modelo hidrológico. Para esse propósito, escolheu-se a bacia do rio Tocantins e implementou-se particularmente a técnica de assimilação de dados de tipo sequencial chamado na literatura de filtro de partículas, conjuntamente com o método de filtro Kalman por conjunto e o método de assimilação AsMGB atualmente acoplado ao modelo MGB-IPH. O estudo mostra que a precipitação combinada utilizada como dado de entrada na simulação hidrológica permitiu reproduzir adequadamente os hidrogramas observados para o período de calibração e validação. Já para o caso das vazões resultantes, durante a etapa de previsão, a precipitação combinada mostrou-se com melhor desempenho em termos estatísticos que os métodos sem combinar, sobretudo após 24 horas de antecedência. Finalmente, a técnica de assimilação de dados por filtro de partículas conseguiu absorver os erros da simulação melhorando as medidas de desempenho na etapa de previsão sendo superior ao modelo de previsão sem considerar assimilação. / The objective of this study is to evaluate precipitation sources based on satellite estimates and data assimilation techniques for prediction of flows by means of the distributed hydrological model MGB-IPH. The insufficient spatial availability of rain gauges makes difficult to represent precipitation fields appropriately. In contrast, satellite estimates, although providing a more consistent spatial description, are potentially less accurate. Thus, raingauge satellite merging methods that combine data from both sources to generate a more consistent precipitation field are used herein. For this purpose, two models namely CHUVSAT and MERGEHQ were implemented using an interpolation technique. On the other hand, data assimilation techniques coupled with hydrological forecasting models are also assessed in this study. The assimilation process minimizes the uncertainties associated with the parameter calibration procedure, variable state and hydrological input data. In this manner, the sequential data assimilation technique namely particle filter in conjunction with the Kalman filter method and the assimilation method AsMGB, which is currently coupled to the MGBIPH model, were implemented and applied to the Tocantis basin. The obtained results showed that the combined precipitation used as input data in the hydrological simulation allowed reproducing adequately the observed hydrograms for the periods of calibration and validation. In the case of the resulting flows during the forecast stage, the merging precipitation was shown to perform better in statistical terms than the uncombined methods, especially after 24 hours in advance. Finally, the data assimilation technique by particle filter was able to absorb all simulation errors, improving the performance measures in the forecasting stage, thus being superior to the forecasting model without considering assimilation.

Neste trabalho foi proposta uma estratégia de modelagem hidrológica para a transformação chuva vazão da Bacia do Ribeirão da Onça (B.R.O) utilizando-se técnicas de auto calibração com análise de incertezas e de ponderação de modelos. Foi utilizado o modelo hidrológico Soil and Water Assessment Tool (SWAT), por ser um modelo que possui uma descrição física e de maneira distribuída dos processos hidrológicos da bacia. Foram propostas cinco diferentes séries de precipitação e esquemas de interpolação espacial a serem utilizados como dados de entrada para o modelo SWAT. Em seguida, utilizou-se o método semiautomático Sequential Uncertainty Fitting ver.-2 (SUFI-2) para a auto calibração e análise de incertezas dos parâmetros do modelo e produção de respostas com intervalos de incerteza para cada uma das séries de precipitação utilizadas. Por fim, foi utilizado o método de ponderação bayesiana de modelos (BMA) para o pós-processamento estocástico das respostas. Os resultados da análise de incerteza dos parâmetros do modelo SWAT indicam uma não adequação do método Soil Conservation Service (SCS) para simulação da geração do escoamento superficial, juntamente com uma necessidade de maior investigação das propriedades físicas do solo da bacia. A análise da precisão e acurácia dos resultados das séries de precipitação em comparação com a resposta combinada pelo método BMA sugerem a última como a mais adequada para a simulação chuva-vazão na B.R.O. / This study proposed an approach to the hydrological modeling of the Ribeirão da Onças Basin (B.R.O) based on automatic calibration and uncertainty analysis methods, together with model averaging. The Soil and Water Assessment Tool (SWAT) was used due to its distributed nature and physical description of hydrologic processes. An ensemble, composed by five different precipitation schemes, based on different sources and spatial interpolation methods was used. The Sequential Uncertainty Fitting ver-2 (SUFI-2) procedure was used for automatic calibration and uncertainty analysis of the SWAT model parameters, together with generation of streamflow simulations with uncertainty intervals. Following, the Bayesian Model Averaging (BMA) was used to merge the different responses into a single probabilistic forecast. The results of the uncertainty analysis for the SWAT parameters show that the Soil Conservation Service (SCS) model for surface runoff prediction may not be suitable for the B.R.O, and that more investigations about the soil physical properties at the Basin are recommended. An analysis of the accuracy and precision of the simulations produced by the precipitation ensemble members against the BMA simulation supports the use of the latter as a suitable framework for streamflow simulations at the B.R.O.

The spatial patterns of land surface and subsurface characteristics determine the spatial heterogeneity of hydrological processes. Soil depth is one of these characteristics and an important input parameter required by distributed hydrological models that explicitly represent spatial heterogeneity. Soil is related to topography and land cover due to the role played by topography and vegetation in affecting soil-forming processes. The research described in this dissertation addressed the development of statistical models that predict the soil depth pattern over the landscape; derivation of new topographic variables evaluated using both serial and parallel algorithms; and evaluation of the impacts of detailed soil depth representation on simulations of stream flow and soil moisture. The dissertation is comprised of three papers. In paper 1, statistical models were developed to predict soil depth pattern over the watershed based on topographic and land cover variables. Soil depth was surveyed at locations selected to represent the topographic and land cover variation at the Dry Creek Experimental Watershed, near Boise, Idaho. Explanatory variables were derived from a digital elevation model and remote sensing imagery for regression to the field data. Generalized Additive and Random Forests models were developed to predict soil depth over the watershed. The models were able to explain about 50% of the soil depth spatial variation, which is an important improvement over the soil depth extracted from the SSURGO national soil database. In paper 2, definitions of the new topographic variables derived in the effort to model soil depth, and serial and Message Passing Interface parallel implementations of the algorithms for their evaluation are presented. The parallel algorithms enhanced the processing speed of large digital elevation models as compared to the serial recursive algorithms initially developed. In paper 3, the impact of spatially explicit soil depth information on simulations of stream flow and soil moisture as compared to soil depth derived from the SSURGO soil database has been evaluated. The Distributed Hydrology Vegetation Soil Model was applied using automated parameter optimization technique with all input parameters the same except soil depth. Stream flow was less impacted by the detailed soil depth information, while simulation of soil moisture was slightly improved due to the detailed representation of soil depth.

Submitted by CAIO VINICIUS FERREIRA MARMONTEL (caioomarmontel@gmail.com) on 2018-05-30T19:16:14Z No. of bitstreams: 1 tese_caiovfmarmontel.pdf: 3393744 bytes, checksum: c80daca7ae4e8871496c00cb8b548ba6 (MD5) / Approved for entry into archive by Maria Lucia Martins Frederico null (mlucia@fca.unesp.br) on 2018-06-04T13:57:05Z (GMT) No. of bitstreams: 1 marmontel_cvf_dr_botfca.pdf: 3241299 bytes, checksum: 194185ca33e9c13c6ba04b1dabe49aae (MD5) / Made available in DSpace on 2018-06-04T13:57:05Z (GMT). No. of bitstreams: 1 marmontel_cvf_dr_botfca.pdf: 3241299 bytes, checksum: 194185ca33e9c13c6ba04b1dabe49aae (MD5) Previous issue date: 2018-05-03 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / A água disponível para o consumo humano vem se tornando cada vez mais escassa, um exemplo, foi a crise hídrica iniciada em 2014 na Região Metropolitana de São Paulo. Diante desse problema, uma possibilidade de melhorar a compreensão do comportamento hidrológico, é a utilização de tecnologias, ou seja, as bases computacionais, em especial a integração do SIG com a modelagem hidrológica, por possuir um grande potencial de estimar quantitativamente a vazão. O presente estudo teve como objetivo testar a aplicabilidade do modelo hidrológico SWAT para estimação de vazões em um trecho da bacia hidrográfica do rio Paraibuna - Vale do Paraíba, região da Mata Atlântica, no estado de São Paulo. O rio Paraibuna utiliza suas águas para gerar energia elétrica e é um dos formadores do rio Paraíba do Sul, importante corpo d’água da região sudeste do Brasil. Os dados climáticos e fluviométricos foram coletados no LHFWE/PESM-NC com uma série histórica de 22 anos. O trecho da bacia hidrográfica estudada tem elevação de 974 a 1662 metros, cinco tipos de solos e três tipos de usos da terra. As avaliações dos dados observados e simulados na calibração e validação foram feitas por meio da avaliação dos gráficos-visuais de dispersão, hidrogramas e índices estatísticos pelo SWAT-CUP. Os dados gerados do balanço hídrico pelo SWAT, após a calibração foram valores próximos com a realidade. Na calibração e validação foi verificado boa aderência entre os dados simulados e observados, ou seja, os valores simulados reconheceram os picos e recessões dos valores observados. Os índices estatísticos (NS, PBIAS, RSR e R²) calculados para a calibração e validação do SWAT foram qualificados como “muito bom” para a estimação das vazões. Os resultados confirmaram a aplicabilidade do modelo, dessa forma, pode servir como ferramenta para planejamento e gestão de políticas públicas dos recursos hídricos em bacias hidrográficas, na região da Mata Atlântica. O modelo hidrológico SWAT mostrou-se muito bom e apto para estimação de vazões e do balanço hídrico no trecho da bacia hidrográfica do rio Paraibuna. / The water available for human consumption is becoming increasingly scarce, an example was the water crisis started in 2014 in the Metropolitan Region of São Paulo. Faced with this problem, one possibility to improve the understanding of hydrological behavior, is the use of technologies, that is, the computational bases, especially the integration of GIS with hydrological modeling, because it has a great potential of quantitatively estimate the flow rate. The present study had as objective to test the applicability of the SWAT hydrological model for flow estimation in the stretch of the Paraibuna river basin - Paraiba Valley, in the Atlantic Forest region, in the state of São Paulo. The Paraibuna River uses its waters to generate electricity and is one of the creators of the river South Paraiba, an important body of water in the southeastern region of Brazil. Climatic and fluviometric data were collected at the LHFWE/PESM-NC with a historical series of 22 years. The stretch of the studied basin has elevation of 974 to 1662 meters, five types of soils and three types of land use. The evaluations of the data observed and simulated in the calibration and validation were done by means of the evaluation of the visual graphics of dispersion, hydrograms and statistical indexes by SWAT-CUP. The data generated from the SWAT water balance after calibration, were close to reality. In the calibration and validation, in good adherence was observed between the simulated and observed data, that is, the simulated values recognized the peaks and recessions of the observed values. The statistical indices (NS, PBIAS, RSR and R²) calculated for SWAT calibration and validation were qualified as "very good" for the estimation of flows. The results confirmed the applicability of the model, so it can serve as a tool for planning and management of water resources public policies in watersheds, in the Atlantic Forest region. The SWAT hydrological model was very good and suitable for estimation of flow and water balance in the stretch of the Paraibuna river basin. / 1513206

Neste estudo, trata-se de avaliar fontes de precipitação baseadas em estimativas por satélite e técnicas de assimilação de dados para previsão de vazões por meio do modelo hidrológico distribuído MGB-IPH. A insuficiente representatividade espacial dos pluviômetros torna difícil a correta representação dos campos de precipitações. Por outro lado, as estimativas de satélite, embora forneçam uma descrição espacial mais consistente, são potencialmente menos acuradas. Sendo assim, procura-se utilizar métodos que combinem os dados de ambas as fontes para gerar um campo de precipitação mais consistente. Neste trabalho, implementaramse dois modelos de combinação pluviômetro-satélite, CHUVSAT e MERGEHQ, através de uma metodologia de interpolação. Por outro lado, as técnicas de assimilação de dados acoplados aos modelos de previsão hidrológica são também de interesse neste estudo, pois minimizam as incertezas associadas ao processo de calibração de parâmetros, às variáveis de estado e dados de entrada do modelo hidrológico. Para esse propósito, escolheu-se a bacia do rio Tocantins e implementou-se particularmente a técnica de assimilação de dados de tipo sequencial chamado na literatura de filtro de partículas, conjuntamente com o método de filtro Kalman por conjunto e o método de assimilação AsMGB atualmente acoplado ao modelo MGB-IPH. O estudo mostra que a precipitação combinada utilizada como dado de entrada na simulação hidrológica permitiu reproduzir adequadamente os hidrogramas observados para o período de calibração e validação. Já para o caso das vazões resultantes, durante a etapa de previsão, a precipitação combinada mostrou-se com melhor desempenho em termos estatísticos que os métodos sem combinar, sobretudo após 24 horas de antecedência. Finalmente, a técnica de assimilação de dados por filtro de partículas conseguiu absorver os erros da simulação melhorando as medidas de desempenho na etapa de previsão sendo superior ao modelo de previsão sem considerar assimilação. / The objective of this study is to evaluate precipitation sources based on satellite estimates and data assimilation techniques for prediction of flows by means of the distributed hydrological model MGB-IPH. The insufficient spatial availability of rain gauges makes difficult to represent precipitation fields appropriately. In contrast, satellite estimates, although providing a more consistent spatial description, are potentially less accurate. Thus, raingauge satellite merging methods that combine data from both sources to generate a more consistent precipitation field are used herein. For this purpose, two models namely CHUVSAT and MERGEHQ were implemented using an interpolation technique. On the other hand, data assimilation techniques coupled with hydrological forecasting models are also assessed in this study. The assimilation process minimizes the uncertainties associated with the parameter calibration procedure, variable state and hydrological input data. In this manner, the sequential data assimilation technique namely particle filter in conjunction with the Kalman filter method and the assimilation method AsMGB, which is currently coupled to the MGBIPH model, were implemented and applied to the Tocantis basin. The obtained results showed that the combined precipitation used as input data in the hydrological simulation allowed reproducing adequately the observed hydrograms for the periods of calibration and validation. In the case of the resulting flows during the forecast stage, the merging precipitation was shown to perform better in statistical terms than the uncombined methods, especially after 24 hours in advance. Finally, the data assimilation technique by particle filter was able to absorb all simulation errors, improving the performance measures in the forecasting stage, thus being superior to the forecasting model without considering assimilation.

Les étiages extrêmes entraînent souvent des conséquences importantes sur de multiples secteurs socio-économiques. Les récentes études liées au changement climatique semblent indiquer que ces événements risquent de devenir plus sévères et plus fréquents au cours des prochaines décennies. Malheureusement, le peu de données hydrométéorologiques disponibles avant les années 1970 ne permet pas de remettre dans un contexte historique les derniers événements observés ni ceux projetés par les études d'impact. Ces travaux de thèse s'attachent ainsi à améliorer l'état des connaissances sur les étiages extrêmes historiques ayant touché le territoire français depuis la fin du XIXe siècle grâce à une reconstruction hydrométéorologique. Ils proposent aussi un nouveau cadre méthodologique pour l'étude de ces extrêmes.Des informations sur la situation synoptique atmosphérique depuis la fin du XIXe siècle ont récemment été mises à disposition de la communauté via des réanalyses globales étendues comme la Twentieth Century Reanalysis (20CR). Ces travaux introduisent la méthode SCOPE qui permet de reconstruire la météorologie locale sur l'ensemble de la France à partir de la réanalyse 20CR. Elle produit un ensemble de 25 chroniques météorologiques spatialement homogènes de précipitations, température et évapotranspiration sur la période 1871-2012 et sur une grille de 64 km² recouvrant la France. Ces séries constituent le jeu de données SCOPE Climate, qui est ensuite utilisé comme forçage d'un modèle hydrologique sur un large échantillon de plus de 600 bassins versants français faiblement anthropisés. Un ensemble de 25 reconstructions de débits journaliers, appelé SCOPE Hydro, est ainsi produit entre 1871 et 2012. SCOPE Climate et SCOPE Hydro ont tous deux montré de bonnes performances en comparaison de données indépendantes de leur construction.Des événements spatio-temporels d'étiage extrême sont finalement identifiés et caractérisés à partir des chroniques de débit de SCOPE Hydro. Ces événements sont tout d'abord définis localement comme une période où le débit se trouve en-dessous d'un seuil mixte, combinaison d'un seuil fixe et d'un seuil variable avec la saison. Un regroupement spatial des événements locaux au sein du même événement spatio-temporel d'étiage extrême est effectué à l'échelle de la France, suivant une méthode spécialement mise au point dans le cadre de ces travaux. A l'issue de ces étapes, un événement peut être étudié localement ou à l'échelle nationale, grâce à des caractéristiques d'étendue, de durée ou de sévérité. Cette méthodologie permet d'identifier des événements exceptionnels d'étiage extrême anciens et peu connus (1878, 1893, 1942-1949), ou relativement récents mais peu documentés (1972, 1978, 1985), en plus de ceux connus (1921, 1976, 1989-1990, 2011). L'étude de l'évolution de ces événements sur 140 ans montre une plus grande proportion du territoire français touchée par des étiages extrêmes après les années 1940. Une comparaison des événements reconstruits à des sources documentaires faisant mention des sécheresses passées montre une bonne cohérence entre reconstructions et réalité.Ce travail de thèse contribue aux questions scientifiques d'actualité, notamment dans le cadre de la variabilité climatique et du changement climatique d'origine anthropique et de leurs conséquences hydrologiques. Ces travaux proposent deux méthodes innovantes sur la descente d'échelle statistique multivariée et l'identification spatio-temporelle des événements d'étiage extrême. Ils ont par ailleurs produit deux jeux de données hydroclimatiques ensemblistes de référence utilisables pour tout type d'étude climatique et hydrologique requérant une profondeur historique importante. / Extreme low-flow events have significant consequences on numerous socio-economic fields. Hydroclimate projections for the 21st century suggest an increase in low-flow severity and frequency. Nevertheless, projected events as well as recent observed events can hardly be put into a sufficiently long historical perspective due to the lack of hydrometeorological data before the 1970s. This work proposes to improve the knowledge on past extreme low-flow events having affected France since the end of the 19th century thanks to a hydrometeorological reconstruction. It also provides a new methodological framework to study these extreme events.Information on the atmospheric synoptic situation since the end of the 19th century have recently been released to the scientific community through extended global reanalyses like the Twentieth Century Reanalysis (20CR). This work introduces the SCOPE method that provides local meteorological reconstructions on the entire France derived from the 20CR reanalysis. SCOPE produces a 25-member ensemble of spatially coherent meteorological series of daily precipitation, temperature and evapotranspiration over the 1871-2012 period and on a 64 km² grid covering France. This dataset, called SCOPE Climate, is then used as forcings to run continuous hydrological simulations over more than 600 near-natural French catchments leading to a 25-member ensemble of daily streamflow time series -- called SCOPE Hydro -- between 1871 and 2012. SCOPE Climate and SCOPE Hydro show a relatively high skill during validation experiments against independent data.Spatio-temporal extreme low-flow events are finally identified and characterised from SCOPE Hydro reconstructed series. The events are first locally identified based on deficit characteristics under a novel combination of a fixed threshold and a daily variable threshold. A spatial matching procedure at the scale of France is developed in order to spatially assemble local extreme events into the same spatio-temporal event. After these steps, an event can either be studied at the local or national scale through its spatial extent, duration or severity characteristics. This work identified past and little known exceptional extreme events (1878, 1893, 1942-1949) or recent but poorly documented events (1972, 1978, 1985) besides well-known events (1921, 1976, 1989-1990, 2011). The evolution of these events since 1871 shows that a greater proportion of the French territory is affected by extreme low-flow events since the 1940s. A good coherence is found between reconstructed events and documentary sources on historical droughts.This work contributes to timely scientific issues, especially within the context of climate change and its hydrological impacts. This work proposes two innovative methods on multivariate statistical downscaling and spatio-temporal identification of extreme low-flow events. It also produced two reference hydroclimatic datasets that may be used in any study requiring long hydrometeorological series.

I detta examensarbete har det undersökts om digitala höjdmodeller, framställda från LIDAR-data som insamlats av Lantmäteriet, kan användas för komplettera och lokalisera nya bäckar och diken som inte finns med i Lantmäteriets databaser idag. I Lantmäteriets verksamhet karteras bäckar och diken först och främst idag med hjälp flygfotografier i stereo. Då metoden fungerar bra på öppen mark kan det vara svårt att urskilja bäckar under tät vegetation som till exempel barrskog. Metoden som använts för att lokalisera bäckar och diken går ut på att beräkna flödesriktningar och flödesackumuleringar i höjdmodellerna. I projektet testas hur metoden påverkas av höjdmodeller av olika upplösningar och ursprung. Förutom Lantmäteriets egna höjdmodeller testas en modell som tagit fram genom en så kallad B-spline interpolation av laserdata. Detta för att undersöka om Lantmäteriets modeller duger för att göra de hydrologiska beräkningar som krävs för att identifiera bäckar och diken. För att evaluera noggrannheten på bäckarna mäts ett antal bäckar och diken in under olika terräng- och vegetationstyper mot vilka en areaavvikelse per meter referensbäck och andelen som ligger under godkänt noggrannhetskrav i plan beräknas. Studien har visat att Lantmäteriets bäcklinjer under tät vegetation, främst barrskog, har en lägre noggrannhet än på öppen mark. Lantmäteriets egna höjdmodeller med 1 meters upplösning har visat sig prestera bäst eller likvärdigt med övriga modeller under dessa förhållanden och genererar bäckar med större noggrannhet än reda karterade bäcklinjer. Dock är metoden beroende av kompletterande information om bäckarnas utformning från till exempel flygfotografier eller fältstudier, eftersom genererade bäckar måste väljas ut manuellt. Detta gör att inga nya bäckar med absolut säkerhet kan lokaliseras med denna metod. Det rekommenderas därför att metoden endast används för att komplettera redan karterade bäckar och diken. / This Master thesis investigated whether digital elevation models, derived from LiDAR data collected by the Swedish mapping and surveying agency, Lantmäteriet, which is a public authority that manages geographic information in Sweden, can be used to map streams and dikes not included in their databases today. Today Lantmäteriet identifies streams and dikes mainly with the photogrammetric methods. The methods works well on open land, but it can be difficult to distinguish smaller streams and ditches covered by dense vegetation. The method used to locate streams and dikes is to calculate flow directions and flow accumulations in the height models. The project tests how the methods are influenced by elevation models with different resolutions and origins. In addition to the national elevation model, a new model interpolated from LiDAR-data was tested, using a so-called B-spline method. This was to find out if the model produced by Lantmäteriet is useful to make the hydrological calculations required to identify streams and dikes. In order to evaluate the accuracy of the streams, a number of reference stretches of streams and dikes were surveyed in areas with different terrain and vegetation types. The area deviation per meter reference stretch and the proportion below approved accuracy was calculated. The results shows that Lantmäteriet’s mapped watercourses under dense vegetation, mainly coniferous forest, have a lower accuracy than on open fields. Lantmäteriet’s height models with 1 meter resolution have been shown to perform best or equivalent to other models under these conditions, and generate broads with greater accuracy than finding marked pelvic lines. However, the method is dependent on supplementary information about the streams distribution from, for example, aerial photographs or field studies, as generated streams must be manually selected. This means that watercourses cannot be mapped accurately with this method. It is therefore recommended that the method should only be used for supplement already mapped streams and dikes.

Dans un contexte où des millions de personnes dépendent de la ressource en eau exposée au caprice de la mousson en Afrique de l'Ouest, ce travail de thèse vise à mieux appréhender les processus hydrologiques en zone de socle, et notamment à évaluer le rôle de la redistribution latérale de l'eau dans le sol par les interactions entre les réservoirs souterrains, la végétation et l'atmosphère, par la caractérisation et la modélisation à l'échelle de la parcelle et du versant. Ce travail s'appuie sur le dispositif expérimental déployé dans le petit bassin versant de l'Ara dans le cadre de l'observatoire AMMA-CATCH. La mise en oeuvre du modèle ParFlow-CLM permet de simuler les transferts dans la zone saturée et la zone non saturée par la résolution de l'équation de Richards en 3D, en étant conditionné par un forçage atmosphérique en surface. Après avoir identifié les paramètres influents pour les transferts verticaux, une caractérisation spatiale de ces paramètres a été menée. La configuration du modèle a ensuite été évaluée en 1D. Il est montré que le modèle reproduit de manière pertinente les séries temporelles du bilan d'énergie et la distribution de l'eau dans le sol. L'effet de la variabilité spatiale des paramètres hydrodynamiques est ensuite étudié à l'échelle de la parcelle. Enfin, en incluant les géométries de socle connues et une distribution de végétation, les transferts horizontaux souterrains sont mis en évidence avec la formation de zones sèches ou humides en relation avec des distributions spatiales d'évapotranspiration. / In West Africa, millions of people rely on water resources exposed to the monsoon variability. In this context, the aim of this thesis is to better understand hydrological processes in bedrock areas, and more particularly to estimate the role of lateral redistribution of soil water by the interactions between underground reservoirs, vegetation and atmosphere, using hydrogeological prospection and modeling at field and catena scale. This work is supported by the experimental device implemented in the small Ara catchment in the framework of the AMMA-CATCH observatory. The use of the ParFlow-CLM model allows the simulation of transfers in the saturated and the vadose zone by solving the Richards equation in 3D. The model was forced using observed atmospheric forcing at the surface. We first identify influential parameters for vertical water transfers. Then a spatial characterization of these parameters is carried out. The 1D version of the Parflow-CLM model is assessed using observed data. We show that the model provides relevant times series of the surface energy balance and of soil water distribution as compared to the observations. The impact of the spatial variability of the hydraulic parameters at the field scale (<1 ha) is studied. Finally, the bedrock geometry and the spatial distribution of vegetation are taken into account in the modelling. This allows the identification of horizontal subsurface lateral fluxes, which generate wet and dry patterns, which are related to the spatial distribution of evapotranspiration.

Neste estudo, trata-se de avaliar fontes de precipitação baseadas em estimativas por satélite e técnicas de assimilação de dados para previsão de vazões por meio do modelo hidrológico distribuído MGB-IPH. A insuficiente representatividade espacial dos pluviômetros torna difícil a correta representação dos campos de precipitações. Por outro lado, as estimativas de satélite, embora forneçam uma descrição espacial mais consistente, são potencialmente menos acuradas. Sendo assim, procura-se utilizar métodos que combinem os dados de ambas as fontes para gerar um campo de precipitação mais consistente. Neste trabalho, implementaramse dois modelos de combinação pluviômetro-satélite, CHUVSAT e MERGEHQ, através de uma metodologia de interpolação. Por outro lado, as técnicas de assimilação de dados acoplados aos modelos de previsão hidrológica são também de interesse neste estudo, pois minimizam as incertezas associadas ao processo de calibração de parâmetros, às variáveis de estado e dados de entrada do modelo hidrológico. Para esse propósito, escolheu-se a bacia do rio Tocantins e implementou-se particularmente a técnica de assimilação de dados de tipo sequencial chamado na literatura de filtro de partículas, conjuntamente com o método de filtro Kalman por conjunto e o método de assimilação AsMGB atualmente acoplado ao modelo MGB-IPH. O estudo mostra que a precipitação combinada utilizada como dado de entrada na simulação hidrológica permitiu reproduzir adequadamente os hidrogramas observados para o período de calibração e validação. Já para o caso das vazões resultantes, durante a etapa de previsão, a precipitação combinada mostrou-se com melhor desempenho em termos estatísticos que os métodos sem combinar, sobretudo após 24 horas de antecedência. Finalmente, a técnica de assimilação de dados por filtro de partículas conseguiu absorver os erros da simulação melhorando as medidas de desempenho na etapa de previsão sendo superior ao modelo de previsão sem considerar assimilação. / The objective of this study is to evaluate precipitation sources based on satellite estimates and data assimilation techniques for prediction of flows by means of the distributed hydrological model MGB-IPH. The insufficient spatial availability of rain gauges makes difficult to represent precipitation fields appropriately. In contrast, satellite estimates, although providing a more consistent spatial description, are potentially less accurate. Thus, raingauge satellite merging methods that combine data from both sources to generate a more consistent precipitation field are used herein. For this purpose, two models namely CHUVSAT and MERGEHQ were implemented using an interpolation technique. On the other hand, data assimilation techniques coupled with hydrological forecasting models are also assessed in this study. The assimilation process minimizes the uncertainties associated with the parameter calibration procedure, variable state and hydrological input data. In this manner, the sequential data assimilation technique namely particle filter in conjunction with the Kalman filter method and the assimilation method AsMGB, which is currently coupled to the MGBIPH model, were implemented and applied to the Tocantis basin. The obtained results showed that the combined precipitation used as input data in the hydrological simulation allowed reproducing adequately the observed hydrograms for the periods of calibration and validation. In the case of the resulting flows during the forecast stage, the merging precipitation was shown to perform better in statistical terms than the uncombined methods, especially after 24 hours in advance. Finally, the data assimilation technique by particle filter was able to absorb all simulation errors, improving the performance measures in the forecasting stage, thus being superior to the forecasting model without considering assimilation.

Les régions méditerranéennes sont particulièrement soumises au risque de crues rapides. Ces crues sont provoquées par des pluies intenses pouvant être stationnaires et qui affectent des bassins versants à la topographie marquée et aux temps de réponses courts. Ces crues peuvent être très violentes, dévastatrices voire meurtrières. Il est donc important de disposer de systèmes de prévision adaptés pour l'anticipation des pluies et des débits des cours d'eau méditerranéens. Le système couplé ISBA-TOP a été particulièrement conçu pour simuler la réponse hydrologique de bassins versants à dynamique rapide. Il s'agit d'un couplage entre le modèle de surface ISBA et une version du modèle hydrologique TOPMODEL adaptée aux crues rapides. Une première étape du travail de thèse a été d'évaluer l'apport d'une meilleure représentation des processus hydrologiques dans le sol pour la simulation des débits avec ISBA-TOP. Une version d'ISBA-TOP basée sur la version diffusive d'ISBA (ISBA-DF qui découpe verticalement les colonnes de sol en de multiples couches) a donc été comparée à la version originale d'ISBA-TOP qui utilisait l'approche ISBA-3L où 3 couches de sol sont représentées. Sur un échantillon d'une vingtaine de cas d'étude, on montre une légère amélioration des simulations de débits avec la nouvelle version d'ISBA-TOP. Cette nouvelle version permet de s'affranchir du calage de deux paramètres gérant la répartition verticale de l'eau dans les colonnes de sol qu'il était nécessaire de réaliser pour chacun des bassins étudiés. La simulation de petits bassins versants non jaugés, pour lesquels on ne dispose pas d'observations pour réaliser le calage des paramètres, peut donc être envisagée. Il n'en reste pas moins que la modélisation hydrologique avec ISBA-TOP (comme avec tout autre modèle) reste entachée d'incertitude. La seconde partie de cette thèse vise à identifier ces incertitudes et à les quantifier. Dans le cas des crues rapides, les données de précipitation utilisées en entrée du modèle hydrologique constituent la source majeure d'incertitude mais elle n'est pas la seule. La connaissance des conditions initiales d'humidité des sols est également incertaine et le modèle hydrologique, lui-même, est entaché d'incertitude de par la `` paramétrisation '' des processus physiques et de par sa structure. Afin d'explorer ces deux sources d'incertitude et de les hiérarchiser, une étude de sensibilité du modèle ISBA-TOP à ses paramètres et à ses conditions initiales d'humidité des sols a été menée. Cette étude de sensibilité a d'abord été conduite pour un cas idéalisé, d'interprétation plus facile, avant de confirmer et d'approfondir les résultats sur des cas réels. Les simulations de débit avec ISBA-TOP se sont avérées particulièrement sensibles à trois des paramètres hydrodynamiques : le contenu en eau à saturation, la conductivité hydraulique à saturation et le taux de décroissance de la transmissivité avec la profondeur. Cette sensibilité, toutefois, est très dépendante des conditions initiales d'humidité des sols. Sur la base des résultats de l'étude de sensibilité, une méthode de perturbation des paramètres du modèle hydrologique a été conçue afin de tenir compte des erreurs de modélisation. Elle est complétée par une méthode de perturbations des conditions initiales en humidité des sols afin de constituer un système de simulation hydrologique d'ensemble. Evalué sur six cas réels, il a été montré que cet ensemble a un apport pour la simulation des débits par rapport à la version déterministe du modèle. / Intense precipitation events in the Mediterranean often lead to devastating flash floods. The affected watersheds are characterized by steep slopes and a short response time. Flash floods can be violent and destructive. Dedicated meteorological and hydrological systems are thus necessary to anticipate those flash floods. The ISBA -TOP coupled system was developed to simulate the hydrological reaction of fast responding rivers. It is a coupling between the ISBA surface scheme and a version of the TOPMODEL hydrological model dedicated to mediterranean flash floods simulation. A first part of the thesis aims at assessing the benefit of a better representation of hydrological processes within the soil for discharge simulation with ISBA-TOP. Moreover, it would allow the use of ISBA-TOP for any watershed even ungauged. A version of ISBA-TOP based on ISBA-DF (the diffusive version of ISBA that discretizes soil columns in multiple layers) has been compared to the original ISBA-TOP that relies on ISBA-3L (with only 3 soil layers). On 18 study cases, better discharge simulations are obtained basically using the new ISBA-TOP version. So this improved representation of hydrological processes in the soil allows to improve discharge simulations and to be confident into the results obtained for small ungauged catchments. The second part of this PhD work concerns the uncertainty that affects ISBA-TOP simulations. For flash floods, rainfall data used to drive an hydrological model are the major source of uncertainty. But initial soil moisture knowledge is also uncertain and the hydrological model himself is affected by uncertainty. The sensitivity of ISBA-TOP model to its parameters and initial soil moisture is investigated to document those two sources of uncertainty. First an academic case is used to verify some preliminary hypotheses and then real cases are studied to properly consider the different data heterogeneities. Discharge simulations with ISBA-TOP are sensitive to three hydro dynamical parameters : the saturated soil water content, the saturated hydraulic conductivity and the rate of decrease of the transmissivity with depth. This sensitivity is found very dependant on initial soil moisture conditions. Perturbation methods varying the 3 parameters that have the highest impact on discharge simulations as well as initial soil moisture allow to design an ensemble prediction system. This ensemble has been assessed for 6 real cases. Using this ensemble-based approach for discharge simulation lead to better results than using the deterministic version. The skill of the ensemble with 30 to 50 members is close to the skill of this ensemble with 100 members whereas it outerperfoms an ensemble with 10 members. The last part of the thesis is the conception of a complete hydrometeorological ensemble prediction system (HEPS). The hydrological ensemble defined in the second part of the thesis is used to take the uncertainty that affects the hydrological modelling and initials soil moisture into account. The uncertainty that affects precipitation fields is sampled using the AROME ensemble prediction system (AROME-EPS). The skill of this complete chain is better than an HEPS based on AROME-EPS and the deterministic version of ISBA-TOP. But both HEPS exhibit a low bias for discharge simulation on the study cases sample. This bias can come from a low bias that is found for the AROME-EPS rainfall forecasts. A simple bias correction applied on rainfall forecasts improves the bias on discharge forecasts.