Made available in DSpace on 2015-03-26T12:31:11Z (GMT). No. of bitstreams: 1 texto completo.pdf: 1048332 bytes, checksum: d2cc8bcdce04c4c44c28d4f3c1b64a84 (MD5) Previous issue date: 2009-09-22 / Conselho Nacional de Desenvolvimento Científico e Tecnológico / The main form of agricultural land degradation is the soil water erosion which is caused by the action of rain and runoff, being affected by a large number of natural and antropic agents. Moreover, much of the water that flows over the soil surface represents a reduction in the volume of water that reaches the water table. The models are important tools for analysis, since they allow a better understanding of hydrological processes and the assesment of management practices and land use. The aim of this work was to develop and calibrate a hydrologic model to support the impact assessment of the use and management of soil on water availability and soil losses. In 2001, Pruski et al. proposed a hydrologic model to estimate the components of water balance in agricultural areas. The intensity of precipitation is obtained by a function derived from the i-d-f equation, resulting in a profile where the peak rainfall intensity occur in the beginning of precipitation and decreases until the end. The infiltration rate is based on the Green-Ampt-Mein-Larson (GAML) equation. A major limitation of the method is related to the precipitation intensity profile, because it is always associated with a recurrence period and does not properly represent the majority of natural rainfall. With regard to infiltration, the GAML equation is applied without adjustment of the input parameters, which can lead to errors in the infiltration estimate. To make easier the application of the hydrological model was developed a software (Simulação Hidrológica-SH2), whose main limitation is the difficulty of making continuous simulation. In this study we implemented some adjustments to the hydrological model: a) replacement of i-d-f equation by a double exponential equation to represent the precipitation profile; b) the parametrization of the GAML equation to estimate the infiltration rate; c) inclusion of the continuity equation for the simulation of the soil water erosion. The software was integrated with a weather generator (ClimaBR) to allow for daily simulation of hydrological processes, and a module for calculating the soil loss was added. For the evaluation of the model were used monthly and annual rainfall, runoff and soil loss data, obtained from four experimental plots located on the campus of the University of Santa Catarina State - UDESC, in the period from 2003 to 2008. The treatments were the soil without vegetation (SC), conventional tillage (PC), minimum tillage (CM) and no tillage (SD). The simulation was performed with two precipitation synthetic series, one set on a monthly basis (SABM) and another set on a daily basis (SABD). The combination of parameters which resulted in the best fit of the GAML equation was: the replacement of hydraulic conductivity (K0) by the constant infiltration rate (Tie); and the estimated matric potential at the wetting front as a function of hydraulic conductivity in the saturated zone ( Ψf(Kw)). The regression coefficients between the total monthly runoff observed and estimated showed the model has a tendency to underestimate the runoff during the months when major events occur, and to overestimate during the months when minor events were observed. With SABD the coefficients of regression equations were better than the SABM. The model underestimated the average annual runoff for SC, PC and CM, but overestimated for SD. However, were not found statistically significant differences between the observed average annual runoff and average annual runoff estimated by the model. The values of r2 were low. There was no statistically significant difference between the average soil loss rate observed and estimated by the model for the soil without vegetation (SC), but was significant for the other soil tillage. / A principal forma de degradação das terras agrícolas é a erosão hídrica do solo, que é causada pela ação da chuva e do escoamento superficial, sendo afetada por um grande número de agentes naturais e antrópicos. Além disso, grande parte da água que escoa sobre a superfície do solo representa uma redução do volume de água que atinge o lençol freático. Os modelos são ferramentas de análise importantes, uma vez que permitem uma melhor compreensão dos processos hidrológicos, a análise do desempenho de práticas de manejo e uso do solo. Assim sendo, o objetivo deste trabalho foi desenvolver e calibrar um modelo hidrológico de suporte à avaliação do impacto do uso e manejo do solo na disponibilidade de água e nas perdas de solo. Em 2001, Pruski et al. propuseram um modelo hidrológico que permite estimar os componentes do balanço hídrico em áreas agrícolas. A intensidade de precipitação é obtida por uma função derivada da equação de chuvas intensas, resultando em um perfil de precipitação onde a intensidade máxima ocorre no início da chuva e decresce até o final. A taxa de infiltração se baseia no modelo de Green-Ampt modificado por Mein-Larson (GAML). Uma das principais limitações do método está relacionada ao perfil de precipitação, pois o mesmo está sempre associado a um período de retorno e não representa adequadamente a maioria das chuvas naturais. No que diz respeito à infiltração, a equação GAML é aplicada sem ajuste dos parâmetros de entrada, o que pode levar a erros consideráveis na estimativa da infiltração. Para facilitar a aplicação do modelo hidrológico foi desenvolvido um software (Simulação Hidrológica-SH2), cuja principal limitação é a dificuldade de fazer simulação contínua. Neste trabalho foram implementadas algumas adequações ao modelo hidrológico: a) substituição da equação de chuvas intensas por uma equação dupla exponencial para representar o perfil de precipitação; b) ajuste dos parâmetros da equação GAML para estimar a taxa de infiltração; c) inclusão da equação da continuidade para a simulação da erosão hídrica. No software foi implementada a integração com um gerador de séries climáticas (ClimaBR) para permitir a simulação diária dos processos hidrológicos, e a inclusão de um módulo para o cálculo da perda de solo. Para a avaliação do modelo foram utilizados dados mensais e anuais de precipitação, escoamento superficial e perda de solo, obtidos de quatro parcelas experimentais localizadas no campus da Universidade do Estado de Santa Catarina UDESC, em Lages (SC), no período de 2003 a 2008. Os tratamentos foram o solo sem cobertura (SC), preparo convencional (PC), cultivo mínimo (CM) e semeadura direta (SD). A simulação foi realizada com duas séries sintéticas de precipitação, uma ajustada em base mensal (SABM) e a outra ajustada em base diária (SABD). A melhor combinação de ajuste da equação GAML foi: a substituição da condutividade hidráulica (K0) pela taxa de infiltração estável (Tie), e a estimativa do potencial matricial na frente de umedecimento em função da condutividade hidráulica na zona de saturação ( Ψf(Kw)). Os coeficientes de regressão entre os valores mensais observados e o estimados mostraram que o modelo apresenta a tendência de subestimar o escoamento superficial nos meses em que ocorrem os maiores eventos, e de superestimar nos meses em que ocorrem os menores eventos observados. Com a SABD os coeficientes das equações de regressão foram melhores do que com a SABM. O modelo subestimou o escoamento médio anual para o SC, PC e CM, mas superestimou para a SD. No entanto, não houve diferença estatística significativa entre o escoamento médio anual observado e o escoamento médio anual estimado pelo modelo. Os valores de r2 para a regressão entre a perda de solo mensal observada e as perdas de solo mensais estimadas pelo modelo foram baixos. Não houve diferença estatística significativa entre a perda de solo média anual observada e a estimada pelo modelo para o solo sem cobertura (SC), mas foi significativa para os demais preparos de solo.

O aumento da ocorrência de desastres hidrológicos relacionados a inundações bruscas tem recebido maior atenção dos diversos órgãos em suas diferentes escalas, com o objetivo de reduzir ao máximo suas causas. Justamente por isso, as medidas não estruturais são medidas de extrema importância na prevenção de tais desastres. Uma dessas medidas deve ser o mapeamento de áreas de perigo de inundações. Portanto, o objetivo do presente trabalho foi propor e avaliar o perigo de inundações bruscas por meio de modelagem hidrogeomorfológica na bacia do arroio Forromeco-RS. Para isso, foi utilizado o modelo CAESAR-LISFLOOD, para representar os processos hidrológicos em escala de bacia e canal. Em escala de bacia foram gerados hidrogramas a partir da criação de chuvas de projeto para diferentes tempos de retorno (TR), considerando como base o maior evento registrado nessa bacia associado a um TR de 22 anos. Esses hidrogramas foram utilizados nas simulações em escala de canal para gerar os diferentes mapas de inundação em termos de profundidade e velocidade do fluxo da água. Para analisar as áreas de perigo de inundação, foram determinados os índices de perigo (IP) associados aos diferentes TR, a partir da profundidade e a velocidade d’água. Através das análises do resultado de IP foi criado o mapeamento final de perigo associado a três tempos de retorno (5, 22, e 100 anos). Além disso, estabeleceu-se três zonas para identificar os níveis de perigo, considerando o cenário mais crítico dos três mapas. Os resultados mostraram que a maior área inundada se encontra em alto perigo, ocupando 77% da área total, o que significa que as pessoas que moram nessa região estão em perigo tanto em casa, como fora delas. Ao mesmo tempo as construções estão em alta possiblidade de serem danificadas. / The increase in the occurrence of hydrological disasters related to flash floods has begun to be more important for several organs at different scales in order to reduce their magnitude and frequency as much as possible. Precisely because of it, non-structural measures are extremely important measures for preventing such disasters. One of these important measures might be the mapping of flood hazard areas. Therefore, the objective of the present work was to propose and evaluate the flash flood hazard by means of hydrogeomorphic modeling of the Forromeco river basin, Rio Grande do Sul state. Thus, the CAESAR-LISFLOOD model was used to represent the hydrological processes at basin and channel scale. At basin scale hydrographs were generated from the creation of hyetographs for different return periods (RP), considering the largest event recorded in this basin. These hydrographs were used in the channel scale simulations to generate the different flood maps in terms of depth and velocity of water flow. In order to analyze the flood hazard areas, the hazard indexes (HI) associated with the different RPs were determined from the depth and water velocity. Through the IP analysis, the final hazard mapping associated with three RPs (5, 22, and 100 years) was created. In addition, three zones were established to identify the hazard levels, considering the most critical scenario of the three maps. The results showed that the largest flood area is in high degree hazard, occupying 77% of the total area. It indicates that people are in danger both inside and outside houses. At the same time buildings are in high possibility of being damaged.

Em regiões úmidas a relação entre os mananciais superficiais e subterrâneos se dá principalmente por meio da descarga dos aquíferos para os rios, essa dinâmica é “alimentada” principalmente pela recarga subterrânea. A área de estudo foi a Bacia do Paraná 3, localizada no Sistema Aquífero Serra Geral, no estado do Paraná. O presente trabalho estudou a recarga subterrânea em duas bacias monitoradas por estações fluviométricas a partir de três diferentes métodos, o balanço hídrico, a separação do escoamento de base usando três diferentes formas de obtenção do parâmetro BFImax (tabelado, curva de permanência e filtro inverso), e a modelagem hidrológica a partir da análise dos reservatórios subterrâneo e subsuperficial do modelo MGB-IPH. Para melhor entender as águas subterrâneas, também se utilizou o TWS do GRACE para avaliar as variações do armazenamento de água na área. Os resultados obtidos mostraram que a maioria dos métodos e suas variações convergiram em grandeza (resultados médio de 491,4±31,9mm para a bacia SFV e 270,1±39,4mm para a bacia SFF), com exceção dos resultados da aplicação da separação do escoamento de base usando o BFImax tabelado (212,0mm para a bacia SFV e 160,6mm para a bacia SFF), além disso, verificou-se que o uso do esvaziamento do reservatório subsuperficial e subterrâneo juntos (612,9±18,1mm para a bacia SFV e 372,8±25,2mm para a bacia SFF) apresenta resultados similares ao uso de filtros e ao balanço hídrico, entretanto, quando se utiliza somente o esvaziamento do reservatório subterrâneo (509,1±17,3mm para a bacia SFV e 53,5±9,7mm para a bacia SFF), somente uma estação converge em resultados, a outra apresenta valores bastante inferiores, esses resultados foram explicados pelas diferentes geomorfologias das bacias. Foi verificado que a aplicação dos filtros de separação do escoamento de base em dados simulados no modelo MGB-IPH não apresenta prejuízos em relação a aplicação dos mesmos em dados observados (<4%). A variação mensal da recarga subterrânea apresentou um comportamento concordante ao TWS do GRACE. Com os resultados obtidos um modelo hidrogeológico conceitual da área foi proposto com base nas características físicas (geomorfologia, solos, características hidráulicas e hidrológicas e dados dos poços) e com base nos resultados obtidos pela aplicação dos métodos de recarga subterrânea. O modelo mostrou que a existência de topos alongados e aplainados em regiões de baixa de declividade favorecem a recarga subterrânea frente a regiões sem topos aplainados e de maior declividade, além disso, a profundidade dos solos também explicou os resultados. / In humid regions, the relationship between groundwater sources and the rivers is mainly governed by the discharge of aquifers into rivers, this dynamics is fed mainly by groundwater recharge. The study area was the Paraná Basin 3, located in the Serra Geral Aquifer System, in the state of Paraná. The present work studied the groundwater recharge in two basins monitored by gauge stations using three different methods: the water-budget method, the hydrograph separation method using three different ways of obtaining the BFImax parameter (Eckhardt pre-defined values, Q90/Q50 ratio and backwards filter) and the hydrological modeling from the analysis of the groundwater and subsurface reservoirs of the MGB-IPH model. To better understand groundwater, TWS obtained from GRACE was also used to assess variations in water storage in the area. The results showed that most of the methods and their variations converged in magnitude (average of 491,4±31,9mm for SFV basin and 270,1±39,4mm for SFF basin), except for the results of the hydrograph separation using the Eckhardt pre-defined values of BFImax (212,0mm for SFV basin and 160,6mm for SFF basin), in addition, it was verified that the use of the emptying of the subsurface and groundwater reservoir together (612,9±18,1mm for SFV basin and 372,8±25,2mm for SFF basin) presents results similar to the use of hydrograph separation and to the water-budget method, however, when only the emptying of the groundwater reservoir is used (509,1±17,3mm for SFV basin and 53,5±9,7mm for SFF basin), only one station converged in results, the other presented lower values, these results were explained by the different geomorphologies of the basins. It was verified that the application of the hydrograph separation in simulated data does not present damages in relation to the application of the same in observed data (<4%). The monthly variation of the groundwater recharge presented a concordant behavior to the TWS. With the results obtained a conceptual hydrogeological model of the area was proposed based on the physical characteristics (geomorphology, soils, hydraulic and hydrological characteristics and hydrodynamic characteristics of the wells) and based on the results obtained by the application of groundwater recharge methods. The model showed that the existence of elongated and flattened tops in low declivity regions favors groundwater recharge compared to regions with no flattened tops and with greater declivity, in addition, soil depth also influenced the results.

Jordbruk syftar till användandet av mark för produktion av livsmedel, foder och råvaror. Reducerad jordbearbetning uppstod ursprungligen under 1900-talet för att öka lönsamheten till följd av stigande priser på drivmedel. För att sänka förbrukningen av drivmedel började bönder att bearbeta åkermarken i grundare skikt. Inom det moderna jordbruket finns det en pågående trend där en övergång från den konventionella plöjningen till reducerade bearbetningsmetoder blir allt vanligare. Idag pågår det kontinuerlig forskning som försöker tydliggöra hur olika bearbetningsmetoder påverkar jordbruksmarkens struktur. Det finns även teorier som pekar på att en övergång från konventionell plöjning till reducerad jordbearbetning kan minska mängden avrinning från åkermark, men trots det är området fortfarande relativt outforskat. Skador till följd av avrinning från kraftiga nederbördstillfällen har under modern tid ökat. I framtida prognoser förutspås det en fortsatt ökning och mer omfattande skador. Med anledning av detta kommer en bättre klimatanpassning att krävas för att minska skador som uppstår till följd av avrinning. Ett exempel på den växande problematiken inträffade 2015 i Hallsberg. Den femte september anlände ett lågtryck som förde med sig över 100 mm nederbörd. Till följd av den extrema mängden nederbörd uppstod det omfattande översvämningar i Hallsberg och på närliggande åkermarker. Från händelsen växte frågan hur en övergång från plöjning till reducerat jordbruk påverkar avrinning fram som studien syftar till att undersöka. Strategin som användes för att nå målet med studien bestod dels av en litteraturstudie samt hydrologiska modelleringar för olika jordbearbetnings metoder i programmet HEC-HMS. I litteraturstudien sammanställdes vetenskapliga studier som jämfört den fysikaliska inverkan som jordbruksmetoder har på åkermark. De hydrologiska modelleringarna baserades på information från Lantmäteriet, mätningar från tidigare undersökningar och SCS runoff curve number method (kurvnummer). Resultatet från litteraturstudien och de hydrologiska modelleringarna sammanvägdes i slutet av studien och från resultatet kunde ingen slutsats dras om hur en övergång från konventionell till reducerad jordbearbetning påverkar avrinning. För att kunna dra en slutsats behövde ett bredare vetenskapligt underlag samt en mer utvecklad modell. / Farming aims towards the utilization of soil for production of food, animal fodder and commodities. In today’s agriculture, we begin to see a gradual transition from the conventional plowing methods to reduced tillage methods. The concept of the reduced tillage originally arose to increase the profitability in agriculture as a response to the rising fuel price. To reduce the fuel consumption farmers began to cultivate the arable soil in more shallow layers. Today researchers are also trying to map how different tillage methods are affecting the soil structure in the arable soil. There are also some scientific theories which point out that a transition from conventional plowing to reduced tillage can reduce runoff from arable land, but despite the interest, the area is still relatively unexplored. Runoff due to heavy rainfall has in modern time increased. In future climate forecasts, the rainfall is predicted to increase even more. As a response to the increased rainfall, the amount of runoff will be higher and with that more extensive damages will occur. In response to these developments, the requirements for climate adjustment will increase in order to minimize the potential damage caused by runoff and flooding events. An example of this growing problem occurred in 2015 in Hallsberg. Between 5 and 6 September, a low-pressure weather system brought down about 100 mm of precipitation. Due to the extreme amount of the rainfall, large floods occurred all over Hallsberg as well as in the nearby arable land. From this event, the question of how a transition from conventional plowing methods to reduced tillage affects the runoff from rainfall emerged. The strategy used for the study consisted of a literature review and hydrological modeling for various soil tillage methods in the program HEC-HMS. In the literature review, scientific studies that compared the physical soil impact from different farming method were compiled. The basis for the hydrological model was created on collected information from the Swedish mapping, cadastral and land registration authority, the literature study and SCS runoff curve method. No clear conclusion could be drawn from the study. To be able to draw a conclusion from the compiled literature and the hydrological model a broader basis of scientific literature and a more developed model was needed.

This thesis deals with the event-based rainfall-runoff modelling of the selected river basin. This thesis is based on my previous work "Factors affecting the water discharge from the river basin during rainfall-runoff events". In that work I described factors, which have effect on water runoff from the river basin. This thesis was solved on basin of Jílecký stream. Water runoff from basin is affected with a number of factors, the most significant are slope of the terrain, soil saturation, geological and pedological conditions, vegetation cover and anthropogenic influence. Calculations of direct runoff were realised with method of CN curves, with using BPEJ map, land use map gained from Corine Land Cover 2006 and maximum daily precipitation sums with 2, 10, 20, 50 and 100 years probabilities recurrence obtained from the rainfall station Netřebice. Calculations were performed in a numerical method using vector data and a raster method performed in ArcMap. In the next step I performed simulations of various scenarios of change in river basic characteristics such as soil saturation change, grassing of arable land with a slope greater than 12°, grassing of all arable land and enlargement the built-up area in the river basin.

Submitted by Emanuel Varela Cardoso (emanuel.varela@ufcg.edu.br) on 2018-09-19T19:31:31Z No. of bitstreams: 1 EDIVALDO AFONSO DE OLIVEIRA SERRÃO – DISSERTAÇÃO (PPGMet) 2018.pdf: 5221512 bytes, checksum: 0bb1b4ee7c18b4cae80a1e0ca7a742f9 (MD5) / Made available in DSpace on 2018-09-19T19:31:31Z (GMT). No. of bitstreams: 1 EDIVALDO AFONSO DE OLIVEIRA SERRÃO – DISSERTAÇÃO (PPGMet) 2018.pdf: 5221512 bytes, checksum: 0bb1b4ee7c18b4cae80a1e0ca7a742f9 (MD5) Previous issue date: 2018-02-23 / CNPq / A bacia amazônica representa a maior extensão de florestas tropicais da Terra, exercendo significativa influência no clima local e global devido aos fluxos de energia e água na atmosfera. Exerce um papel importante no sequestro de carbono, bem como na emissão de água para a atmosfera que é fundamental na manutenção da própria floresta. Vários modelos hidrológicos têm sido aplicados para os biomas do Brasil, desde bases conceituais até a inserção dos SIG’s desenvolvendo os modelos distribuídos de base física. Dentre estes, o Soil and Water Assessment Tool – SWAT é um modelo hidrológico para grandes bacias e foi desenvolvido para simular processos que ocorrem em uma bacia hidrográfica. O objetivo desse trabalho foi utilizar o modelo SWAT para simular a vazão fluvial, alguns processos hidrológicos e risco de erosão da bacia hidrográfica do rio Itacaiúnas, sudeste do Estado do Pará, bem como sua calibração e validação do modelo para a bacia de estudo. Para isto foram necessários dados de uso e ocupação do solo, tipos de solo, declividade. Bem como dados climáticos e fluviométricos de precipitação, temperatura, vento, umidade relativa, radiação solar e vazão fluvial. Os resultados demostraram que para a análise do risco de erosão para a bacia hidrográfica do rio Itacaiúnas foi constatado que ás áreas de maior risco para erosão, foram ás regiões de maior altitude com solos de fácil degradação e com uma dinâmica de uso de solo elevada, sendo estas na jusante da bacia e na porção sul do município de Marabá. Na validação da vazão obtida através do modelo SWAT, foi observado que todos os índices estatísticos aplicados mostraram que a vazão simulada tive um bom desempenho, em relação a vazão observada na estação fluviométrica da ANA. Foi validado ainda a evapotranspiração potencial simulada pelo SWAT, a partir de outros dois métodos, Linacre e Turc e observou-se que todos os índices estatísticos aplicados a evapotranspiração simulada com SWAT tive um bom desempenho em relação aos outros dois métodos de estimar evapotranspiração. Já na comparação dos métodos de Linacre teve um menor desempenho, em relação ao método de Turc, haja visto que na estimativa feita por Turc é levado em consideração radiação solar, o que torna o método mais robusto em relação a Linacre. Logo o modelo SWAT teve um bom desempenho em simular a vazão, evapotranspiração e outros processos hidrológicos para a bacia hidrográfica do rio Itacaiúnas e pode ser uma fundamental ferramenta no monitoramento hidrológico dessa bacia, além de auxiliar os tomadores de decisões em uma boa gestão dos seus recursos. / The Amazon basin represents the largest expanse of tropical rainforests on Earth, exerting significant influence on the local and global climate due to the energy and water fluxes in the atmosphere. It plays an important role in carbon sequestration as well as in the emission of water into the atmosphere that is essential in maintaining the forest itself. Several hydrological models have been applied to the biomes of Brazil, from conceptual bases to the insertion of the GIS, developing the distributed models of physical base. Among these, the Soil and Water Assessment Tool - SWAT is a hydrological model for large basins and was developed to simulate processes that occur in a watershed. The objective of this work was to use the SWAT model to simulate river flow, some hydrological processes and risk of erosion of the Itacaiúnas river basin, southeast of the State of Pará, as well as its calibration and validation of the model for the study basin. For this, data on land use and occupation, soil types and slope were required. As well as climatic and fluviometric data of precipitation, temperature, wind, relative humidity, solar radiation and river flow. The results showed that for the analysis of the erosion risk for the Itacaiúnas river basin, it was found that the areas with the highest risk for erosion were the highest altitude areas with easily degraded soils and a high soil use dynamics, being these downstream of the basin and in the southern portion of the municipality of Marabá. In the validation of the flow obtained through the SWAT model, it was observed that all the applied statistical indices showed that the simulated flow had a good performance, in relation to the flow observed in the fluviometric station of the ANA. It was also validated the potential evapotranspiration simulated by SWAT, from two other methods, Linacre and Turc, and it was observed that all the statistical indices applied to simulated evapotranspiration with SWAT performed well in relation to the other two methods of estimating evapotranspiration. In the comparison of the methods of Linacre had a lower performance, in relation to the method of Turc, since in the estimation made by Turc is taken into consideration solar radiation, which makes the method more robust in relation to Linacre. The SWAT model performed well in simulating flow, evapotranspiration and other hydrological processes for the Itacaiúnas river basin and can be a fundamental tool in the hydrological monitoring of this basin, besides helping decision-makers in a good way. management of its resources.

Ce travail consiste à caractériser la variabilité temporelle et spatiale des séries chronologiques de paramètres hydroclimatiques (pluies, débits) au niveau de trois grand bassins au Maroc ; (bassins de la Moulouya, du Sebou et du Tensift) et à chercher les liens entre cette variabilité hydrologique et les fluctuations climatiques matérialisées par différents indices climatiques, NAO, SOI, WMOI. L’approche d’étude est basée le traitement statistique des séries temporelles, liée aux dimensions temps et espace.Les grands bassins versants d'échelle continentale comme le Tensift, le Sebou et la Moulouya en climat méditerranéen sous influence océanique, intègrent sur des grandes surfaces la réponse hydrologique aux changements climatiques et environnementaux (fluctuations du climat, précipitations, débits) à de larges échelles spatiales et temporelles, mais également les modifications du milieu physique d’origine anthropique (changements d’occupation des sols, aménagements…), ce qui rend parfois difficile l’identification des liens entre la variabilité hydrologique et la variabilité climatique. Les principaux objectifs de ce travail sont de déterminer et de quantifier les relations entre la variabilité hydroclimatique et les fluctuations du climat à l’échelle de chaque bassin étudié et de ses principaux sous-bassins, via l'utilisation de méthodes d’analyses spectrales adaptées à l’étude des processus non stationnaires (analyse en ondelettes continues, analyse de la cohérence par ondelettes). Plusieurs modes de variabilités sont identifiés à partir de l’analyse par station (pluies et débits), du cycle annuel au mode 16-22 ans, cette analyse sera complétée par une analyse par maille, dont les données sont issues d’un fichier (SIEREM) couvrant la période 1940-1999, où on identifie des fréquences de 1an au 8-16 ans, distinguées sur des périodes différentes au niveau de chaque bassin, permettant ainsi une décomposition de la variabilité spatiale des signaux mis en évidence. Trois principales discontinuités sont identifiées en 1970, 1980 et 2000. La contribution des indices climatiques est assez importante elle est entre 55% et 80%. / This work is to characterize the temporal and spatial variability of hydroclimatic time series (rainfall, flow) at three large basins in Morocco; (basins of the Sebou and Moulouya Tensift) and look links between the hydrologic variability and climate fluctuation materialized by various climate indices, NAO, SOI, WMOI. The approach to study is based on statistical analysis of time series, related to time and space dimensions.The great watershed of continental scale as Tensift, Sebou and Moulouya in Mediterranean climate under oceanic influence, integrate over large areas the hydrological response to climate and environmental changes (climate fluctuations, precipitation, flows) not only to large spatial and temporal scales, but also to changes in the physical environment anthropogenic (land use changes, developments ...), which sometimes makes difficult to identify the links between hydrological variability and climate variability. The main objective of this work is to determine and quantify the relationships between hydrological variability and climate fluctuations (regionalised precipitation, climate change indexes) across each studied basin and its main sub-basins, via using spectral analysis methods adapted to the study of non-stationary processes (continuous wavelet analysis, coherence analysis wavelet). Many modes of variability are identified from the station analysis (rainfall and flow rates), the annual cycle to 16-22 years, this analysis will be complemented by a grid analysis, the data come from a (SIEREM) file covering the period from 1940 to 1999, which will allow a better understanding of the spatial variability of signals set highlighted. Which is identified frequencies the 1 year 8-16 years, distinguished different time periods at each basin, three main discontinuities identified in 1970, 1980 and 2000. The contribution of climatic indices is important enough it is between 55% and 80%.

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.

Pouvoir évaluer l'impact du changement climatique sur la ressource en eau, et les systèmes de gestion qui lui sont associés, est une préoccupation majeure de nos sociétés. Une telle évaluation nécessite la mise en place d'une chaîne de simulation qui permet, sur la base d'expériences climatiques futures, i) d'estimer à l'échelle régionale l'évolution possible de la ressource et de sa variabilité, ii) de simuler le comportement des systèmes utilisés pour leur gestion pour iii) estimer les éventuelles modifications de performance. Cette thèse vise à tester la possibilité de mettre en place une chaîne de simulation de ce type pour un système de gestion réel et à identifier quelles sont les composantes à considérer dans ce cas. Pour ce faire, nous chercherons en particulier à apporter des éléments de réponse aux questions suivantes: - Quelles représentations peut-on faire d'un système de gestion opérationnel pour une application en climat modifié ? - Quels éléments d'évaluation peuvent permettre d'estimer l'impact du changement climatique sur ce système de gestion ? - Quelles sont les sources d'incertitudes influençant cette évaluation ? Quelles sont les contributions relatives à l'incertitude totale des différentes méthodes et modèles utilisés ? Nous considérerons plus précisément le système de gestion du barrage de Serre-Ponçon, alimenté par le haut bassin versant de la Durance. Ce barrage, géré par EDF, est l'un des plus grands barrages artificiels européens. Il est multi-usages (irrigation, soutien d'étiage, production d'hydroélectricité, tourisme). Dans un premier temps, nous présenterons le contexte du système de gestion actuel. Nous mettrons ensuite en place un modèle de gestion du barrage visant à reproduire – de façon réaliste du point de vue du gestionnaire actuel (EDF), mais simplifiée pour pouvoir être appliqué sous scénarios futurs - la gestion actuelle du barrage. Nous développerons pour cela i) des modèles permettant d'estimer les différentes demandes en eau et ii) un modèle d'optimisation de la gestion sous contraintes. Ce modèle permettra de simuler la gestion du système au pas de temps journalier sur plusieurs décennies du climat récent, ou de climats futurs modifiés. Nous proposerons ensuite un ensemble d'indicateurs qui permettent de fournir une estimation de la performance d'un tel système à partir des sorties du modèle de gestion obtenues par simulation pour différentes périodes de 30 ans. Nous explorerons la façon dont la performance estimée dépend du modèle choisi pour la représentation du système de gestion actuel, et plus précisément de la façon dont la stratégie utilisée pour l'optimisation de la gestion est élaborée. A ce titre, nous proposerons trois modèles de gestion basés sur trois types de stratégies, obtenues pour des degrés différents de prévisibilité des apports et sollicitations futurs à la retenue. Pour ces simulations, les modèles d'impacts nécessitent des scénarios de forçages météorologiques à l'échelle de bassin versant (e.g. modèle hydrologique, modèle d'usages de l'eau, modèle de gestion de la ressource). Ces scénarios peuvent être obtenus par des méthodes de descente d'échelle statistique (MDES), sur la base des simulations grande échelle des modèles climatiques globaux. Enfin, nous évaluerons les incertitudes liées aux deux types de modèles et estimerons leurs contributions relatives à l'incertitude globale. Nous utiliserons pour cela les scénarios issus de différentes chaines de simulation GCM/MDES produits sur la période 1860-2011 dans le cadre du projet RIWER2030. Nous montrerons que ces deux sources d'incertitudes sont du même ordre de grandeur sur l'estimation des modifications de performance. / Assess the impact of climate change on water resources and management systems associated, is a major concern of our society. This requires the establishment of a simulation chain which allows, on the basis of future climate experiments i) to estimate the possible changes in regional resource and its variability, ii) to simulate the behavior of the systems used to manage them in order to iii) estimate the possible changes in performance. This thesis aims to test the feasibility of establishing a chain simulation of such a management system to identify what are the real components to consider in this case. To do this, we have to provide answers to the following questions: - How can we represent an operational management system in a climate change context? - What elements of evaluation can be used to estimate the impact of climate change on the management system? - What are the sources of uncertainty influencing this assessment? What are the relative contributions to the total uncertainty of these different methods and models used? We consider the system of management of the reservoir of Serre-Ponçon, built on the high basin of the Durance. This dam, operated by EDF, is one of the largest artificial dams Europe. It is multi-purpose (irrigation, low-flow support, hydropower, tourism). As a first step, we will present the context of the current management system. Then, we will establish a management model to reproduce - in a realistic way from the point of view of the current manager (EDF), but simplified to be applied in future scenarios - the current management of the Serre-Ponçon reserve. We will develop for this, i) different models to estimate different water demands and ii) an optimization model with constraints management. This model will simulate the management system in daily time step on several decades of recent climate or future climate change. We then propose a set of indicators to provide an estimate of the performance of such a system from the outputs of the management model obtained by simulation for different periods of 30 years. We will explore how the estimated performance depends on the model chosen to represent the current management system, and more specifically how the strategy used to optimize the management is developed. To this end, we will propose three management models based on three types of strategies, obtained for different degrees of predictability of future inflows and constraints. For these simulations, the impact models require meteorological forcing scenarios at watershed scale (eg hydrological model, model of water use model of resource management). These scenarios can be obtained by statistical downscaling methods (SDM), on the basis of large-scale simulations of global climate models. Finally, we will evaluate the uncertainties associated with the two types of models and will estimate their relative contributions to the overall uncertainty. We have used this scenario from different GCM/SDM simulations over the period 1860-2100 obtained within the RIWER2030 project. We show that these two sources of uncertainty are of the same order of magnitude estimate of changes in performance.

Dans un contexte de changement climatique dont les projections régionales sont incertaines, de forte variabilité inter-annuelle du cycle hydrologique, de forte croissance démographique et de changement d'occupation des sols, les questions relatives au cycle hydrologique et à la ressource en eau actuels et à venir sont cruciales. Dans un tel contexte, ce travail de thèse approfondit la connaissance du fonctionnement hydrologique du bassin versant de l'Ouémé supérieur (situé en climat soudanien au Bénin), en considérant l'ensemble des termes et des processus du cycle hydrologique.Dans un premier temps, le fonctionnement hydrologique de l'Ouémé supérieur est analysé à l'échelle d'un versant grâce au modèle Hydrus 2D. Cette analyse de processus s'appuie sur un ensemble complet de mesures (précipitations, évapotranspiration, humidité du sol, niveau piézométrique, débit en rivière) obtenues dans le cadre de l'observatoire hydrométéorologique AMMA-CATCH. Les simulations montrent que la forêt ripisylve vidange la nappe profonde et la déconnecte ainsi du réseau hydrographique. L'apport d'eau de la nappe profonde permet une transpiration de la forêt ripisylve toute l'année, y compris en saison sèche. Les écoulements en rivière ne sont pas produits uniquement par exfiltration d'écoulements latéraux de subsurface non saturés, les bas-fonds jouent probablement un rôle.À méso-échelle, nous cherchons à quantifier l'impact de la variabilité spatiale de la conductivité hydraulique à saturation sur le bilan hydrologique et l'évapotranspiration. Le modèle numérique utilisé est nTopAMMA, formalisme dérivé de TopMODEL. À partir de mesures de terrain, nous avons mis en évidence que la variabilité spatiale de la conductivité hydraulique à saturation est corrélée à l'occupation du sol du bassin d'étude. La prise en compte de cette variabilité spatiale dans le modèle nTopAMMA montre que l'état hydrique et l'évapotranspiration simulés localement par le modèle dépendent aux trois-quarts de la topographie et pour un quart de la conductivité hydraulique à saturation.Adoptant une approche ascendante, nous confrontons la représentation élaborée à l'échelle du versant à la modélisation hydrologique à méso-échelle. Nous analysons les processus et flux verticaux du modèle nTopAMMA. L'amélioration de la modélisation du cycle hydrologique de l'Ouémé supérieur par le modèle nTopAMMA nécessite (i) la prise en compte des hétérogénéités du bassin versant, (ii) la modification du formalisme de l'évapotranspiration, (iii) la diversification des sources de prélèvements évapotranspiratoires et (iv) l'intégration de la nappe d'altérites. / Understanding how the hydrological cycle and water resources availability evolve in the current context of global change (which encompass climate, environmental and population changes) is a critical issue, particularly in West Africa, where at regional scale, strong interannual and seasonal variabilities overlap with highly uncertain climate predictions. Within this framework, this work aims at improving our knowledge of the behavior of the Upper Oueme catchment in Benin (Sudanian climate), with an analysis of all the hydrological processes and terms of the terrestrial hydrological cycle.First, the hillslope scale is considered by using the Hydrus 2D software and field observations from the hydrometeorological observing system AMMA-CATCH: rainfall, actual evapotranspiration, soil moisture, groundwater level and river runoff. The principal result of this analysis is that the riparian forest transpiration depletes the deep groundwater and disconnects it from the river network. Water supply by the deep groundwater enables the riparian forest transpiration all year long and particularly during the dry season. Seepage of unsaturated subsurface lateral flows contributes to river runoff, but the "bas-fonds" seem to be other important contributors.Then, at mesoscale, we quantify the impact of the spatial variability of hydraulic conductivity on the simulated water balance, with focus on the evapotranspiration term. The numerical model nTopAMMA, especially derived from the TopMODEL hydrological model for the Upper Oueme catchment, is used here. Measurements from a specific field mission evidence the correlation between the hydraulic conductivity spatial variability and the land use one. By taking into account this variability, the simulation results show that evapotranspiration and water storage simulated on the Upper Oueme catchment at local (pixel) scale depend essentially on the topography (75 %) and to a lesser extent on the hydraulic conductivity (25 %).Finally, a bottom-up approach is adopted to analyse the hydrological modelling results at mesoscale, taking advantage of the modelling results at the hillslope scale. Vertical processes and fluxes simulated by nTopAMMA are analyzed. It is shown that an improvement of the Upper Oueme hydrological cycle modelling, with nTopAMMA, requires: (i) the consideration of the catchment heterogeneities, (ii) the modification of the evapotranspiration module, (iii) the diversification of the evapotranspiration sources and (iv) the integration of the deep groundwater reservoir.