This work introduces diagnostic methods for land surface model (LSM) evaluation that enable developers to identify structural shortcomings in model parameterizations by evaluating model 'signatures' (characteristic temporal and spatial patterns of behavior) in feature, cost-function, and parameter spaces. The ensemble-based methods allow researchers to draw conclusions about hypotheses and model realism that are independent of parameter choice. I compare the performance and physical realism of three versions of Noah LSM (a benchmark standard version [STD], a dynamic-vegetation enhanced version [DV], and a groundwater-enabled one [GW]) in simulating high-frequency near-surface states and land-to-atmosphere fluxes in-situ and over a catchment at high-resolution in the U.S. Southern Great Plains, a transition zone between humid and arid climates. Only at more humid sites do the more conceptually realistic, hydrologically enhanced LSMs (DV and GW) ameliorate biases in the estimation of root-zone moisture change and evaporative fraction. Although the improved simulations support the hypothesis that groundwater and vegetation processes shape fluxes in transition zones, further assessment of the timing and partitioning of the energy and water cycles indicates improvements to the movement of water within the soil column are needed. Distributed STD and GW underestimate the contribution of baseflow and simulate too-flashy streamflow. This work challenges common practices and assumptions in LSM development and offers researchers more stringent model evaluation methods. I show that, because of equifinality, ad-hoc evaluation using single parameter sets provides insufficient information for choosing among competing parameterizations, for addressing hypotheses under uncertainty, or for guiding model development. Posterior distributions of physically meaningful parameters differ between models and sites, and relationships between parameters themselves change. 'Plug and play' of modules and partial calibration likely introduce error and should be re-examined. Even though LSMs are 'physically based,' model parameters are effective and scale-, site- and model-dependent. Parameters are not functions of soil or vegetation type alone: they likely depend in part on climate and cannot be assumed to be transferable between sites with similar physical characteristics. By helping bridge the gap between the model identification and model development, this research contributes to the continued improvement of our understanding and modeling of environmental processes. / text

Les travaux présentés dans cette thèse portent sur la notion de performance environnementale comme cadre analytique pour l’étude des impacts des activités agricoles sur l’environnement. Après avoir mis en évidence un manque de formalisation de cette notion dans la littérature, nous la définissons comme la distance entre l’état environnemental d’un écosystème à un moment donné et un état environnemental à atteindre pour cet écosystème, dit « de haute performance environnementale ». Dans cette étude, l’état environnemental est approché par trois critères liés à la ressource « eau » : les concentrations en nitrates et en matières en suspension, et les débits.Suite à la description de l’état initial, nous proposons une méthode d’optimisation de la performance environnementale des activités agricoles basée sur un couplage entre un modèle agro-hydrologique (SWAT) et un modèle d’optimisation multicritère (Weighted Goal Programming). La démarche d’optimisation considère la réaffectation spatialisée de systèmes de culture sur l’espace considéré comme facteur d’amélioration, les différents critères de l’optimisation étant examinés simultanément.La méthode est implémentée sur le bassin versant de la Boulouze (coteaux de Gascogne). Les résultats montrent une amélioration de la performance environnementale du système étudié. L’analyse de la nouvelle occupation des sols souligne qu’au-delà des évolutions quantitatives des surfaces allouées aux systèmes de culture, la spatialisation des changements d’un système vers un autre influe également sur la performance environnementale à l’échelle du bassin versant. Enfin, nous interrogeons l’applicabilité de cette méthode et mettons en évidence l’intérêt qu’elle présente comme outil à la fois de modélisation de scénarios et de communication. / We introduce in this thesis the notion of environmental performance as an analytical framework aimed at studying the impacts of agricultural activities on the environment. We outline first a need to formalize this notion and then define it as the distance between the “environmental” state of a system at a specific time and a state of “high environmental performance” to achieve. Focus is on water management at the watershed level with three criteria: nitrogen, sediments (water quality) and water flows (water quantity).Our aim is to optimize the environmental performance of agricultural activities at the watershed scale. First the initial status of the watershed is described, then a method is proposed. This method is based on a coupling between an agro-hydrological model, the Soil and Water Assessment Tool (SWAT), and the Weighted Goal Programming optimization method. Our purpose is the reallocation of farming systems within the watershed when considering the optimization criteria simultaneously.We implement this method on the Boulouze watershed (coteaux de Gascogne, in the southwestern part of France). Results outline an improvement of the environmental performance. The analysis of the new land-use plan emphasizes the fact that the changes of farming systems are not only due to evolutions of the surfaces where they are implemented, but that the environmental performance at watershed scale is also affected by their location.Finally we explore the applicability of the method used highlighting its attributes as a relevant tool for modeling scenarios and for communicating.

This diploma thesis deals with morphometric analysis and hydrological regime of Skalnaté pleso lake and Hincové Oká lakes, especially the lake level fluctuation, during the period 2011 - 2014. One of the tasks was the bathymetric mapping of observed locations. For Hincové Oká lakes was it the first mapping of bathymetry at all. The hydrological regime and lake level fluctuation regime of observed locations shows interesting differences from the other Tatra lakes whose hydrological regime is controlled by precipitation amounts. The annual culmination of Hincové Oká lakes (the largest lake) is shifted into the spring caused by melting snowpack. When there is no rainfall for a couple of days Skalnaté pleso lake often dries up, even during summer months, when levels of most Tatra lakes reach their maximum values. However summer drying is shortterm. Lake level fluctuation regime is very dynamic. After a strong rainfall event its level rises very steeply, up to 0,14 m per hour. Powered by TCPDF (www.tcpdf.org)

The object of this work is an analysis of the influence of extreme hydrological events (floods and droughts) in terms of interaction between groundwater and surface water in the river floodplain. The aim of this thesis is also the evaluation of the hydrological regime of the upper Lužnice basin and evidence of retention potential preserved natural floodplain. The selected objectives of the work were resolved on the basis of the values of groundwater levels and surface water in the area of interest. Based on data from underground wells and a digital terrain model, it was possible to create an image of the distribution of groundwater levels in river floodplains during selected hydrological situations. Data from surface flow was obtained from several habitats in the flow direction Lužnice. This has enabled graphically illustrate individual flood waves and characterize their size, speed and transformation after the passage of a preserved natural floodplain of Lužnice river. Carried out research has been confirmed retention capacity of the floodplain, which is most significant at flow rates slightly in excess of the flow capacity of the river bed. However, extreme floods or in periods with many culminations in a relatively short time leads to reduces water retention capability of the floodplain. In...

La gestion des eaux pluviales urbaines connait une évolution majeure depuis les années 1960 : d'une stratégie basée sur l'évacuation rapide et totale des eaux pluviales de la ville, on passe progressivement à une stratégie de contrôle à la source (CS). Ces 10 dernières années ont vu une généralisation du CS, ainsi qu'en témoigne l'augmentation du nombre de réglementations de rejet. Ces dernières imposent, dans les nouvelles parcelles urbanisées, la réalisation d'ouvrages ayant pour fonction de réguler et/ou de retenir les eaux pluviales : les techniques alternatives (TA). Une systématisation de ces réglementations ne va pas manquer de générer dans l'avenir de nombreux bassins versants urbains fortement équipés en TA, et dont le comportement hydrologique sera déterminé par les réglementations appliquées. Il apparaît que le choix d'une réglementation par une collectivité est un choix complexe et, d'une collectivité à l'autre, des logiques différentes sont appliquées, dont la cohérence globale peut être discutée. Cette thèse questionne et analyse les politiques de CS, et en particulier les réglementations de rejet, dans une perspective de compréhension de leurs effets hydrologiques dans le long terme. Elle comprend deux analyses complémentaires : nous étudions d'abord l'évolution des politiques de CS dans six collectivités en France, afin de déterminer les logiques appliquées dans le choix des réglementations. Nous présentons ensuite une analyse hydrologique des conséquences de ces réglementations sur le comportement de bassins versant urbains "modèles". Cette analyse, centrée sur la modélisation hydrologique des bassins, permet de questionner les politiques de CS actuelles, mais aussi de discuter des outils que l'hydrologie urbaine peut fournir pour permettre la mise en place de politiques de CS cohérentes, en phase avec le fonctionnement hydrologique des bassins versants / Urban stormwater management is facing a major evolution since 1960: from a strategy based on the complete and fast conveyance of stormwater outside of the urban area, to a strategy based on source control (SC). Recently, SC is becoming a common practice, in connection with the increasing diffusion of flow regulations. These regulations prescribe, for all new urban development, to build facilities (Best Management Practices, BMP) to reduce the flow-rate or the volume of stormwater entering the sewer system. A wide application of flow regulations will determine the future hydrological behaviour of many urban catchments. The choice of a regulation is a complex task for local authorities, involving several logics, whose global coherence can be discussed. This thesis analyses SC policies and regulations and discusses them in terms of their long-term hydrological effects. Two complementary approaches are developed. The first is based on the analysis of the evolution of SC policies in six French authorities, and it aims to define and discuss the logics applied. The second consists in the hydrological modelling of the consequences of SC regulations on the behaviour of two experimental catchments. This hydrological analysis allows (i) to evaluate the pertinence of actual SC policies and (ii) to discuss which tools urban hydrology can provide to support the development of coherent SC policies, according to the hydrological functioning of urban catchments

Ph.D. (Geography) / South Africa is characterised by limited natural water resources and is water-stressed. The Upper Vaal Water Management Area (WMA) is part of the Vaal River system, which is deemed to be the ‘workhorse’ of South Africa as it is located in the economic heart of the country. The WMA supplies water to all of the major economic entities in the economic hub of the country and is fully exploited in terms of the water that it makes available. Owing to its prime importance in terms of the economic development of the country, the quality of the water in the Upper Vaal WMA is an important factor which needs to be investigated and monitored intensively. The degradation in the quality of the water in this region will in its turn reduce the availability of water in the Upper Vaal WMA and have widespread environmental, social, as well as economic consequences and impacts. The aim of this study is thus to quantify and predict future hydrological responses, in terms of water quality, towards land cover changes in the case of the Upper Vaal WMA for the period, July 2000 to June 2012. In so doing, model equations can be formulated and predictions can be made for future hydrological changes in terms of future concentrations of the selected water quality parameters linked to specific land cover change scenarios. The quality of the water in the Upper Vaal WMA varies significantly across all of the secondary and tertiary catchments and could, therefore, be described as being of a heterogeneous nature. Its quality as a whole is below average and in some cases, poor. The quality of the water in the secondary catchment of the Wilge River, on the other hand, is for the most part of an ideal to acceptable standard. However it has been degraded to a certain extent as it displays tolerable concentrations of phosphate, COD and Chlorophyll a and unacceptably high concentrations of nitrate. The Vaal River, Vaal River Barrage and the Mooi River catchments are significantly degraded owing to the mainly tolerable to unacceptably high concentrations of most of the selected water quality parameters that they contain.....

Sub-Saharan Africa is today facing a big challenge regarding food deficiency and water scarcity due to climate change. One of these countries is Rwanda, a small landlocked country in the middle of Africa. Rwanda strongly depend on agriculture, both in the aspect of reducing poverty and hunger but also because their economy security depend on it. Because of increasingly fluctuating rainfalls their agriculture becomes more dependent on irrigation and the availability to water resources. To investigate how the climate change will affect the amount of water resources in the coming decades, this study is focusing on the watershed and marshland of Muvumba P8 in Nyagatare, Rwanda. A hydrological model was created, in a software called Soil and Water Assessment Tool (SWAT), with soil, land use and slope maps for the watershed. Calibrating the model was done with help of Climate Forecast System Reanalysis (CFSR) data and run for nine different climate model datasets. An uncertainty had to be taken into account regarding both the measured local data and the downloaded data. To be able to compare the amount of water resources and the irrigation requirements for the rice crop the farmers were growing on the marshland, the crop water requirements for rice was estimated with FAO’s program called CROPWAT. The irrigation system on the marshland allows a double cropping of rice every year and consist of a system depending on elevation differences to create natural fall. There was three reservoirs along the marshland but to limit the project, only the first reservoir was taken into account. This was complemented with existing data and field survey. Six out of nine climate models showed a decrease in median discharge over the coming 30 years compared to the CFSR historical median discharge. This means that less water in general will reach the outlet of the watershed in the years to come. At the same time all climate models indicate an increase in irrigation requirements for the rice crops. The seasons are probably going to change, a longer and drier season between June and August and a rainier season between September and November are projected.

Muitos rios brasileiros ainda não são monitorados. A carência de informações relevantes costuma gerar diversos problemas para a administração pública. Por vezes, essa falta de dados pode ser compensada por meio de modelagem hidrológica, que requer um conhecimento prévio sobre a bacia a ser estudada. Atualmente, a mudança no uso e cobertura do solo em bacias hidrográficas é um dos temas mais abordados nas pesquisas hidrológicas, podendo a avaliação dos impactos causados por essa ser feita através de modelagem hidrológica. O presente trabalho utilizou o framework JAMS (Jena Adaptable Modelling System) na elaboração de um modelo hidrológico para a bacia do Ribeirão da Onça, a qual está localizada em zona de afloramento do Sistema Aquífero Guarani, um dos mais importantes aquíferos transfronteiriços do mundo. A bacia é predominantemente rural, logo, seu uso e ocupação do solo variam ao longo do tempo. Sendo assim, objetivou-se avaliar os efeitos das alterações de uso do solo no regime hidrológico da bacia por meio da elaboração de múltiplos modelos hidrológicos no JAMS/J2K, os quais foram calibrados independentemente (de 2008 a 2015) e validados por um modelo atual (2017), a fim de se gerar uma parametrização capaz de predizer os efeitos hidrológicos decorrentes de alterações no uso e ocupação do solo da bacia. Previamente à elaboração dos modelos, procedeu-se à validação dos dados hidrológicos e climatológicos coletados e disponíveis da bacia, por meio da análise das correlações entre as medições dos equipamentos, feita por programa elaborado em Python. Em seguida, elaboraram-se mapas de uso do solo para os anos de 2008, 2011, 2013, 2015 e 2017. Então, os modelos com seus conjuntos de URHs (Unidades de Resposta Hidrológica) específicos foram calibrados, e os parâmetros ótimos encontrados para cada modelo foram aplicados nos demais para identificação de quais poderiam ser os mais representativos. Durante esta etapa, verificou-se que o modelo de 2013 comportava-se anomalamente, devido à estiagem ocorrida em seu período de calibração, o que justificou sua retirada das etapas seguintes. Em seguida, um programa criado em Python permitiu que os arquivos .dat gerados durante a calibração fossem analisados visualmente, por meio de gráficos de dispersão e plots violinos, embasando a tomada de decisão relativa à redução da amplitude de variação dos parâmetros para uma posterior busca manual por parâmetros representativos temporalmente. Os parâmetros encontrados produziram modelos com NSE variando de 0,25 a 0,39; PBIAS de -20,98 a -0,74 e R2 de 0,45 a 0,65, sendo que o modelo de validação (2017) obteve NSE de 0,26; PBIAS de -1,78 e R2 de 0,45; indicando que a parametrização encontrada, apesar de não conseguir produzir modelos que eficientemente reproduzam os picos no escoamento, é capaz de estimar satisfatoriamente o escoamento de base, mesmo em condições de mudanças de uso e ocupação do solo. Com isso, realizou-se a aplicação do conjunto de parâmetros em cenários hipotéticos de uso e ocupação do solo, que produziu resultados coerentes com o que se espera. Além disso, também se realizou a setorização do modelo, utilizando os dados de vazão de duas calhas Parshall instaladas na bacia. No entanto, não houve êxito na setorização, evidenciado pelos valores de NSE (-35,41 e 88,08), PBIAS (118,63 e 118,81) e R2 (0,15 e 0,12). O fraco desempenho na setorização pode ser corrigido calibrando-se o modelo espacialmente, em adição à calibração temporal, ou seja, com múltiplos dados de entrada de vazão, localizados em pontos distintos na bacia. / Many Brazilian rivers are not monitored. The lack of relevant information may generate multiple problems for public administration. Sometimes this data may be supplied through hydrological modelling, which requires a previous knowledge regarding the watershed that will be studied. Currently, the land use and cover change in watersheds is one of the most researched themes in the hydrological scientific community, with the impacts caused by it being able t obe assessed by hydrological modelling. In that sense, this work has used the framework JAMS (Jena Adaptable Modelling System) in the elaboration of a hydrological model for the Ribeirão da Onça catchment, which is located in an outcrop zone of the Guarani Aquifer System, one of the largest transboundary aquifers in the world. The catchment is mainly rural; thus, its land use and occupation vary heavily over time. In this context, the intent of this work was to evaluate the effects of the land use/cover changes in the hydrological processes through the elaboration of multiple J2K/JAMS hydrological models, which were calibrated individually (2008 through 2015) and validated with a recent model (2017), with the goal of creating a model that is able to predict long term hydrological effects of land use/cover changes in the catchment. Before creating the models, a validation of the hydrological and climatic data available in the catchment was done, through the analysis of the correlations between the measures, achieved with the use of a Python software. After that, land use maps were elaborated for the years 2008, 2011, 2013, 2015 and 2017. Then, the models with their HRU (Hydrological Response Units) specific sets were calibrated, then the optimal parameters for each model were applied on the other ones, to assess which were the most representative. During this step, the 2013 model was deemed anomalous, due to the drought that occurred during its calibration period, which justified its removal from the following stages. After that, an application made in Python allowed the .dat files generated during calibration to be visually analysed, through scatter and violin plots, aiding the decisions regarding the new maximum and minimum values for the parameters, preceding a manual search for the globally representative parameters set. The set that was found produced models with NSE varying from 0.25 to 0.39, PBIAS from -20,98 to -0,74 and R2 from 0,45 to 0,65, being that the validation model (2017) achieved an NSE of 0.26; PBIAS of -1.78 and R2 of 0.45, meaning that the parameters found, despite not being able to effectively reproduce the runnof peak values, is able to satisfactorily estimate the base runoff, even in land use and cover change conditions. With that, the parameters set found was applied in hypothetical scenarios of land use and cover, which produced coherent results with what was expected. A model sectorization was also attempted, using discharge data from two Parshall flumes installed in the catchment. However, there was no success in the attempt, evidenced by the NSE (-35.41 and 88.08), PBIAS (118.63 and 118.81) and R2 (0.15 and 0.12) values. The weak performance of the sectorization may be fixed by calibrating the model spatially, in addition to the calibration in regards to time. In other words, with multiple input discharge data, located in distinct points inside the catchment.

L’objectif de cette thèse est de mettre en évidence la relation climat/bilan hydrologique. Pour une meilleure analyse de la relation pluie/débit, une spatialisation pluviométrique est faite à partir de l’interpolation, en tenant compte de l’altitude et des coordonnées géographiques des stations.L’impact de la variabilité pluvio-hydrologique et de la dynamique des états de surface sur la sédimentation est aussi examiné.La variabilité des écoulements et la dynamique sédimentaire dans le bassin-versant Mono-Ahémé-Couffo pendant la période 1961-2000 dans un contexte de gestion intégrée des ressources en eau (GIRE) ont été analysées, en fonction des lames d’eau précipitées, des effets de lâchers d’eau du barrage de Nangbéto mis en service depuis septembre 1987, de la dynamique des états de surface et du substratum géologique (socle et grès) en place. L’approche combinée de l’utilisation des données météorologiques, hydrologiques, d’occupation du sol, sédimentologiques, associées aux études antérieures et aux observations de terrain, ont permis de développer des méthodes statistiques efficaces pour atteindre les objectifs fixés pour cette étude.Les résultats montrent que les déficits pluviométriques des décennies 1970 et 1980 ont multiplié par 4les déficits de l’écoulement, avec pour conséquence un déficit de la recharge. Mais, avec la légère reprise pluviométrique (3 %) de 1988 à 2000 par rapport à la sous-période 1965-1987, on assiste à un excédent d’écoulement de 0,1 % sur le Couffo et 37 % sur le Mono. Les effets du barrage de Nangbéto se traduisent par une augmentation de 97 % des débits d’étiage alors que les débits de crue ont connu une baisse de 3 %. Ceci montre le rôle joué par le barrage de Nangbéto sur le débit d’écoulement et l’atténuation des inondations et par conséquent une pérennisation de l’écoulement en aval du bassin.L’utilisation du modèle GR2M se révèle assez efficace à simuler les débits moyens malgré l’influence des lâchers d’eau du barrage de Nangbéto.En revanche les faibles coefficients d’écoulement et la faible corrélation entre écoulement et recharge,montrent que la précipitation ne constitue pas le seul paramètre expliquant le fonctionnement hydrologique du bassin, mais qu’interviennent aussi le substratum géologique et les états de surface.Ces derniers déterminent l’aptitude des sols à l’infiltration, au ruissellement et à l’érosion. De 1956 à2000, on note une évolution régressive des formations végétales naturelles au profit des formations anthropiques, du fait des pratiques agricoles. Cette pression humaine a entraîné une diminution de plus de 60 % des formations de forêts et de savanes arborées, qui jouent un rôle modérateur sur l’écoulement, facilitent l’infiltration et ralentissent les apports solides vers les lits des cours d’eau.Leur dégradation entraînerait donc une augmentation de l’écoulement et de l’évaporation au détriment de l’infiltration, accélérant le déficit des réservoirs souterrains ainsi que l’érosion et donc la sédimentation.À cet effet, la variation pluvio-hydrologique, associée à la dégradation actuelle des unités paysagiques, a contribué à l’augmentation des charges solides dans le bassin, dont les plus importantes parviennent au lit en période des hautes eaux, et occupent seulement 8 % de ce temps.Ces apports sont à la fois d’origine continentale et marine, dont les plus importants sont ceux drainés par les cours d’eau. Ils ont contribué à la réduction des superficies des plans d’eau de 4 % dans la basse vallée. Avec les apports du Mono avant la construction du barrage de Nangbéto, le système lagunaire se comble de 1,6 mm.an-1 et pourrait engendrer, la destruction de la biocénose et donc la disparition des ressources halieutiques. / This thesis aims to highlight the relationship between climate and water balance process on the basin of the complex river-lagoon Mono-Aheme-Couffo. A better analysis rain/flow is made using rainfall spatialization from interpolation, taking into account altitude and geographical coordinates of hydrometeorological stations. Also, pluvio-hydrological variability and surface states dynamics impacts on sedimentation are examined. Runoff variability and sediment dynamics in the watershed Mono-Aheme-Couffo over the period 1961-2000 in a context of integrated water resources management (IWRM) were analyzed according to the rainfall depths, effects of water releases from the Nangbeto dam operational since september 1987,the dynamic of surface state and geological bedrock (crystalline basement and sandstone). Thisapproach combined the use of meteorological, hydrological, land use land cover, sedimentological data, associated with a previous studies and field observations data have enabled us to develop efficient statistical methods for achieving the objectives of this study.The results show that rainfall deficits of the 1970s and 1980s have multiplied by 4 deficit of flow, due to recharge lack. But, slight high rainfall (3%) of 1988-2000 period compared to the sub-period 1965-1987, caused an excess flow of 0.1 % on Couffo basin and 37 % on Mono basin. Effects of the dam Nangbeto result in an increase of 97 % of low flows whereas rising flows fall of 3 %. This shows the role of the dam Nangbeto on stream flow and flood reduction and consequently perpetuation of the flow on the downstream of the basin. The GR2M model proves effective enough to simulate the average flows despite the influence of water releases from the dam Nangbeto.In contrast, the low runoff coefficients and the low correlation between runoff and recharge, show that precipitation is not the only factor explaining the hydrological basin functioning, but also geological bedrock and surface conditions play also key role. These determine the suitability of soils for infiltration, runoff and erosion. From 1956 to 2000, there was a regressive evolution of natural forestland and increase of agriculture land. Human pressure on the catchments induce decrease of over60 % of the forest and savannah land, which play a moderating role on the flow, make easier infiltration and reduce sediment transport towards the rivers beds. Natural vegetation degradation could only lead to runoff and evaporation increase, decrease of infiltration and groundwater shortage and so accelerated erosion and sedimentation process.To this effect, pluvio-hydrological change linked to land use land cover change has contributed to increase of sediment transport which largest part reach the river bed during the period of high waterhold only 8 % of this time. These solid charge contributions are both marine and continental origins;the most important are drained by rivers. They have contributed to the water bodies area reduction of 4 % in the lower valley. Mono sediment inputs before the dam Nangbeto drive the lagoon system closing of 1.6 mm/year and could cause the biocenose destruction and therefore disappearance of fishstocks.

Les problèmes environnementaux liés à l’excès d’azote d’origine agricole restent une préoccupation majeure en France malgré une réglementation contraignante. Pour concilier maintien de la production agricole et limitation de cet excès, des stratégies novatrices d’atténuation des flux d’azote ont été identifiées, puis leurs effets ont été simulés à l’échelle de paysages agricoles à l’aide de modèles agro-hydrologiques spatialisés. Pour cela, la méthodologie mise en oeuvre a consisté, à partir d’enquêtes de fermes dans deux contextes contrastés (Bretagne et Gascogne),à simuler des scénarios : i) d’optimisation des pratiques agricoles en accord avec le 5ème programme d’actions de la directive Nitrate et ii) d’aménagements paysagers du territoire et en particulier des zones environnementales (prairies fauchées non-fertilisées), en variant l’emprise et la localisation. Les résultats montrent l’intérêt de placer ces zones en position d’interception (i.e. en zone riparienne humide), plutôt qu’en tête de thalweg, notamment sur le site breton où circulations d’eau sub-superficielles dominent. Cette solution limiterait les pertes nitriques sans augmenter les autres émissions d’azote. / Environmental issues related to the excess of reactive nitrogen of agricultural origin activity, are still a major concern in France despite restrictive regulations. To reconcile sustained agricultural production and reduced nitrogen emissions, we identified innovative mitigation strategies and simulated their effects were simulated using distributed agro-hydrological models. The approach consisted in farm surveys in two contrasted situations (NW and SW of France) a basis to simulate the following scenarios :i) optimisation of agricultural practices according to the 5th action programme of the Nitrate Directive ii) conversion of agricultural land into environmental zones (unmanaged grasslands), with varying extension and in different landscape positons. Results show the interest of locating them in an interception position (i.e. in humid zones around the stream network) rather than in headwater position, especially in the NW catchment were subsurface flow is the dominant pathway. This solution would reduce nitrate losses without increasing other nitrogen emissions.