Mediterranean catchments are particularly prone to short-duration, high-intensity rainfall events that generate flash floods with significant impacts. Analyzing this type of event requires sub-daily hydrometric data in order to adequately capture their dynamics. This study investigates trends in flood characteristics across 38 Mediterranean basins in southern France, with an average size of 200km², using hourly discharge, radar rainfall, and reanalysis-derived soil moisture data over the period 1997–2024. Flood events are identified using a peaks-over-threshold approach and classified according to response time (flash-floods versus slower onset floods) and antecedent soil moisture conditions. Trends in flood peaks and direct runoff volumes are assessed using regional quantile regression after applying a spatiotemporal declustering procedure.Results indicate increasing magnitudes for both flash floods and slow-onset floods under saturated soil conditions. These increases are more pronounced for flood volumes than for peak discharges, with trend magnitudes approximately twice as large. However, these results should be interpreted with caution given the pronounced spatiotemporal variability of flood processes in Mediterranean environments, given that the detected trends are not statistically significant based on a regional bootstrap assessment.Overall, the results suggests that flood volume provides a more sensitive indicator of change than peak discharge in Mediterranean catchments. This underscores the importance of considering hourly data and process-based flood classification to improve the detection of evolving flood hazards in regions impacted by flash floods.

This study introduces a stochastic continuous hourly areal rainfall model designed to simulate areal rainfall time series for hydrological risk management. The rainfall model, named SCHYPRE (Simulation of Continuous HYetographs for Predictive Risk Estimation), extends an established at-site event-based rainfall model to a basin-scale and continuous rainfall model, integrating both extreme event modeling and continuous simulation of seasonal and long-duration rainfall patterns. The rainfall model parameters were calibrated using a 28.5 years dataset of hourly rainfall observations at a 1 km resolution. This dataset enabled the computation of areal rainfall time series across 2108 catchments in France, encompassing a wide range of climatic regimes from continental and Mediterranean to mountainous environments. The evaluation framework demonstrates the rainfall model’s ability to reproduce observed areal rainfall statistics, including mean and extreme values, seasonality, autocorrelation, and intermittency of rainfall. Frequency analysis conducted over durations from one hour to one year shows good agreement between the simulations and the adapted law. An advantage of rainfall modeling is its robustness in estimating extreme return levels. Unlike traditional probabilistic methods, which are more sensitive to sampling variability, the stochastic rainfall model whose parameters are calibrated on large observational datasets of internal variables, ensures a robust estimation of return levels across all return periods, including extremes. Additionally, rainfall modeling inherently avoids quantile-crossing inconsistencies, a common issue in independent duration-based probabilistic modeling.

The Réal Collobrier hydrological observatory, located in south-eastern France and managed by INRAE (formerly Cemagref) since 1966, is a benchmark site for regional hydro-climatology. Created by the Ministry of Agriculture, its initial objective is to improve understanding of hydrological processes in Mediterranean regions underlain by metamorphic soils. The observatory's catchment, situated in the Maures massif near the Mediterranean coast, is densely instrumented. Flow measurements are collected at the outlets of ten small forested nested catchments (ranging from 1.57 to 70 km²), including four headwater streams. A dense network of 15 rain gauges records rainfall data at a fine temporal scale. The dataset also includes climatological data, water temperature and soil data. The vegetation is dominated by forest communities on crystalline substrates (maquis of heath, cork oak, maritime pine, and chestnut). The geological formations are predominantly crystalline, with metamorphism increasing from east to west (from gneiss to schists and phyllites) [1]. Direct human influence has been negligible over the past 60 years, with land use and land cover remaining almost unchanged, except for a wildfire in 1990 that affected one small sub-catchment. All data presented in this article are available in the INRAE hydrological observatory's open database (https://bdoh.inrae.fr/). The article describes the long-term dataset collected at the observatory and the validation procedures. The raw dataset underwent quality control, gap filling, and homogenization procedures to ensure temporal consistency and to improve data reliability. This rigorous quality control process results in a robust dataset used for research purposes. This well-documented hydro-climatic information can significantly advance understanding of hydrological processes [2-6], help validate and evaluate models in Mediterranean environments [7-9]. Given the non-perennial nature of rivers in this area, these data are particularly useful for studying the origin of intermittent flow, as well as the start and end dates of flow period [10]. The hydrological dataset now spans 58 years, offring the opportunity to evaluate long-term hydrometeorological trends [11,12]. Since 2019, observations of soil moisture at several depths have been added, providing valuable information on soil water availability and vegetation dynamics. The Réal Collobrier catchment area is part of the SOERE-RBV (Long-Term Observation and Experimentation System for Environmental Research - Mountain Basin Networks), which belongs to the OZCAR research infrastructure [13], certified by AllEnvi (National Research Alliance for the Environment) (http://www.ozcar-ri.org/real-collobrier/).

Most studies on flood trends rely on daily discharge data at the station scale, limiting their ability to disentangle contrasting trends in different flood-generation mechanisms. This is particularly true for capturing short-duration processes such as flash floods. Here, we propose a process-based regional framework to investigate trends across different flood types using hourly data. We analyze 829 small to medium-sized catchments (<500 km(2)) across France using hourly discharge, radar rainfall, and reanalysis-based soil moisture data. Flood events are extracted using a peaks-over-threshold approach and classified into four categories based on response time (flash vs. slow) and antecedent soil moisture conditions (saturated vs. non-saturated). Catchments are grouped into four homogeneous hydro-climatic regions, and a spatio-temporal declustering procedure is applied to account for spatial and temporal dependence between concurrent events. Trends in flood peaks and direct runoff volumes are then assessed using regional quantile regression. Results reveal strong contrasts between regions and flood types. Only a few trends were detected for flood peaks, whereas flood volumes exhibit more frequent significant increasing trends. The strongest increases are found for flash flood volumes in mountainous and Mediterranean regions. For slow-onset floods, significant positive volume trends are observed in basins located in temperate regions, in line with increases in both rainfall totals and duration. Overall, flood volumes show a systematically higher sensitivity to change than peak flows. These findings demonstrate that flood trend detection critically depends on temporal resolution, flood type, and regional context, and that volume-based metrics provide complementary and often stronger signals than peak-based indicators for assessing evolving flood hazards.

A large transient multi-scenario and multi-model ensemble of future streamflow and groundwater projections in France developed in a national project named Explore2 was recently made available. The main objective of Explore2 is to provide rich and spatially-consistent information for the future evolution of hydrological (surface and groundwater) resources and extremes in France to support adaptation strategies. The Explore2 dataset was obtained using a nested multi-scenario multi-model approach to estimate future uncertainty and to assess local climate at the catchment scale: three greenhouse gas (GHG) emission scenarios, a set of 17 combinations of Global Climate Models and Regional Climate Models (GCM/RCM), and two bias correction methods provide the meteorological forcing for nine surface hydrology models and four groundwater hydrology models (one to simulate groundwater recharge and three to simulate groundwater level). In this paper, we present the methodology underlying the dataset, the evaluation of the hydrological models against daily observations of streamflow and groundwater level, and the key messages on the impact of climate change on both mean river flows and groundwater recharge. This large set of hydrological projections shows a high model agreement on the decrease in seasonal flows in the South of France under the RCP8.5 high-emission scenario, confirming its hotspot status. The surface hydrological models agree on the decrease in summer flows across France under the RCP8.5 scenario, with the exception of northern part France. This area may indeed benefit from more active winter recharge that may counterbalance decrease in summer precipitation and increase in evapotranspiration. In addition to northern France, annual groundwater recharge is projected to increase slightly in the north-east while remaining unchanged elsewhere by the end of the century, according to the RCP8.5 scenario. In the mountainous areas, winter flows will increase as a result of higher air temperature and the high degree of agreement between the models holds regardless of the RCP considered. Unsurprisingly, the higher the GHG emission scenario, the higher the median changes. Most of these changes are organised in France along a north-south gradient, regardless of the RCP considered.

This contribution presents a regionalization approach to estimate spatially distributed hydrologic parameters based on: (i) the SMASH (Spatially distributed Modelling and ASsimilation for Hydrology) hydrological modeling and assimilation platform (Jay-Allemand, 2020; Jay-Allemand et al., 2020) underlying the French national flash flood forecasting system Vigicrues Flash (Javelle et al., 2019); (ii) the variational assimilation algorithm from (Jay-Allemand et al., 2020), adapted to high dimensional inverse problems; (iii) spatial constraints added to the optimization problem, based on masks derived from physiographic maps (e.g., land cover, terrain slope); (iv) multi-site global optimization, which targets multiple independent watersheds. This method gives a regional estimation of the spatially distributed parameters over the whole modeled area. This study uses a distributed rainfall-runoff model with 4 parameters to calibrate, with a spatial resolution of 1×1 km2 and a 15 min time step. Performances of the calibrated hydrological model and the parameters robustness are evaluated on two French study areas with 20 catchments in each, in spatio-temporal extrapolation based on cross-validation experiments over a 12-year period. Several spatial regularization strategies are tested to better constrain the high dimensional optimization problem. The model parameters are calibrated based on the Nash-Sutcliffe Efficiency (NSE) computed for multiple calibration basins in the study area. Results are discussed based on the Nash-Sutcliffe Efficiency and the Kling-Gupta Efficiency criteria obtained on calibration and validation catchments for two subperiods of 6 years. Further work aims to improve the global search of prior parameter sets and to better balance the adjoint sensitivity with respect to the spatial constraints resolution and catchment characteristics. This will ensure a better consistency of simulated fluxes variabilities and enhance the applicability of the regionalization method at higher spatial scales and over larger domains.

Accurate and high-resolution hydrological models are crucially needed, especially for important socioeconomic issues related to floods and droughts, but are faced with data and model uncertainties which can be reduced by maximizing information integration from multisource data. This work focuses on improving the integration of satellite and in situ land surface data into spatially distributed hydrological models. The Hybrid Data Assimilation and Parameter Regionalization (HDA-PR) approach incorporating learnable regionalization mappings, based on neural networks into the differentiable spatially distributed hydrological model SMASH, is modified to account for satellite-based moisture maps in addition to discharge at gauging stations and basin physical descriptors maps. Regional optimizations of a spatially distributed conceptual model are performed on a flash-flood-prone area located in the South of France, and their accuracy and robustness are evaluated in terms of simulated discharge and moisture against observations. In general, the integration of satellite-derived soil moisture data alongside traditional observed streamflow measurements during calibration procedures has demonstrated notable improvements in hydrological performance, both in terms of simulated discharge and moisture. This is achieved thanks to an improved learning of regionalization of model conceptual parameters with HDA-PR integrating satellite-based moisture through the RMSE metric adapted to a spatially distributed model with variational data assimilation. This study provides a solid foundation for advanced data assimilation of multi-source data into learnable spatially distributed differentiable geophysical models.

Study region: This study is carried out for 1929 gauged catchments in France, ranging from 1 to 10,000 km 2 , where quality hydrometric observations are available for flood frequency analysis. Study focus: The regional estimation of hydrological hazards is studied for flood risk management and prevention in hydrology. For gauged catchments, flow quantiles can be estimated from observations using statistical approaches based on suitable probability distributions or simulation approaches based on rainfall-runoff transformation models. For ungauged catchments, the lack of hydrological observations means that we have to extrapolate our knowledge of hazards from gauged catchments to ungauged catchments, using regionalization methods. It is therefore necessary to combine regionalization methods with the implemented hazard estimation approach. In this paper, two popular machine learning methods, Random Forest and Neural Networks, are tested and compared as regionalization methods. A classical regionalization method using multiple linear regression is also applied as a benchmark to evaluate the performance of all configurations. All these regionalization methods are applied to a simulation-based approach (the SHYREG method) and to a statistical-based approach using generalized extreme value distribution (GEV). New hydrological insights:

Floods are a major natural hazard in the Mediterranean region, causing deaths and extensive damages. Recent studies have shown that intense rainfallevents are becoming more extreme in this region but, paradoxically, without leading to an increase in the severity of floods. Consequently, it isimportant to understand how flood events are changing to explain this absence of trends in flood magnitude despite increased rainfall extremes. Adatabase of 98 stations in southern France with an average record of 50 years of daily river discharge data between 1959 and 2021 wasconsidered, together with a high-resolution reanalysis product providing precipitation and simulated soil moisture and a classification of weatherpatterns associated with rainfall events over France. Flood events, corresponding to an average occurrence of 1 event per year (5317 events intotal), were extracted and classified into excess-rainfall, short-rainfall, and long-rainfall event types. Several flood event characteristics havebeen also analyzed: flood event durations, base flow contribution to floods, runoff coefficient, total and maximum event rainfall, and antecedentsoil moisture. The evolution through time of these flood event characteristics and seasonality was analyzed. Results indicated that, in mostbasins, floods tend to occur earlier during the year, the mean flood date being, on average, advanced by 1 month between 1959-1990 and1991-2021. This seasonal shift could be attributed to the increased frequency of southern-circulation weather types during spring and summer. Anincrease in total and extreme-event precipitation has been observed, associated with a decrease of antecedent soil moisture before rainfallevents. The majority of flood events are associated with excess rainfall on saturated soils, but their relative proportion is decreasing over time,notably in spring, with a concurrent increased frequency of short rain floods. For most basins there is a positive correlation between antecedentsoil moisture and flood event runoff coefficients that is remaining stable over time, with dryer soils producing less runoff and a lowercontribution of base flow to floods. In a context of increasing aridity, this relationship is the likely cause of the absence of trends in floodmagnitudes observed in this region and the change of event types. These changes in flood characteristics are quite homogeneous over the domainstudied, suggesting that they are rather linked to the evolution of the regional climate than to catchment characteristics. Consequently, thisstudy shows that even in the absence of trends, flood properties may change over time, and these changes need to be accounted for when analyzing thelong-term evolution of flood hazards.

The estimation of storage and fluxes in surface hydrology is an essential scientific question related to major socio-economic issues, especially when forecasting extreme floods and droughts with the undergoing climate change. Advanced spatially distributed modeling tools are critically needed to perform reliable and skillful local forecasts. Nevertheless, hydrological modeling remains a challenging task because of limited observations of physical processes and modeling uncertainties. In particular, given the spatial sparsity of constraining discharge data, hydrological modeling is faced with the challenge of producing predictions at ungauged locations based on the regionalization of the model parameters. Despite the overparameterization problem in spatially distributed modeling, Jay-Allemand et al. (2020) presented promising results for estimating the spatial variability of the distributed parameters within a catchment using only downstream discharge observations. However, providing better spatial constrains on the estimated parameters patterns, inside or outside calibration catchments in a regionalization perspective, remains a challenge. This contribution presents a regionalization approach based on: (i) the SMASH (Spatially distributed Modelling and ASsimilation for Hydrology) hydrological modeling and assimilation platform (Haruna et al., 2021) underlying the French national flash flood forecasting system Vigicrues Flash (Javelle et al., 2019); (ii) the variational assimilation algorithm from Jay-Allemand et al. (2020), adapted to high dimensional inverse problems; (iii) spatial constraints added to the optimization problem, based on masks derived from physiographic maps (e.g., soil occupation and nature, bedrock type, terrain slope); (iv) multi-objective optimization which targets independent watersheds. This method gives a regional estimation of the distributed parameters over the modeled area. Performances of the model and the parameters robustness are evaluated on a large sample of French catchments and flash floods in spatio-temporal extrapolation based on cross-validation experiments. Effects of the spatial constraints (regularization and multi-objective optimization) are discussed in the light of adjoint sensitivity maps. Further work aims to improve the global search of prior parameter sets and to better balance the adjoint sensitivity with respect to the spatial constraints resolution and catchment characteristics. This will ensure a better consistency of simulated fluxes variabilities and enhance the applicability of the regionalization method at higher spatial scales.

This contribution presents improvements of conceptual models in SMASH (Spatially distributed Modelling and ASsimilation for Hydrology) platform, underlying the French national flash flood forecasting system Vigicrues Flash [1], based on: (i) the 3-parameters model formulation and variational data assimilation algorithm of [2] that showed promising results (i) hypothesis testing on a large sample of catchments and flash floods; (ii) comparison of the SMASH model performances in uniform and distributed calibration to GR models; (iii) a new wrapped Python interface automatically generated by the f90wrap library [3]. Multiple tests have allowed us to converge on two parsimonious distributed model structures that have comparable performances to the GR models in spatially uniform calibration. These two structures, mainly based on GR operators at the pixel scale, differ in the production operator, with the 6-parameters structure being GR production and the 7-parameters structure being VIC production. Furthermore, the use of distributed calibration applied to these formulations via adjoint model resolution shows significantly better calibration performances without being less robust in spatio-temporal validation. Immediate work deals with improving the regional calibration scheme by tayloring the global search of semi-distributed prior parameter sets, with multi-gauge constrains, improving physiographic regularizations in the forward-inverse SMASH assimilation chain, using Python librairies. References [1] P. Javelle, et al. Flash flood warnings: Recent achievements in france with the national vigicrues flash system UNDRR GAR, 2019. [2] M. Jay-Allemand, et al.. On the potential of variational calibration for a fully distributed hydrological model: application on a mediterranean catchment. HESS, 2020, https://doi.org/10.5194/hess-24-5519-2020 [3] J. R. Kermode. f90wrap: an automated tool for constructing deep python interfaces to modern fortran codes. 2020. https://doi.org/10.1088/1361-648X/ab82d2

Abstract. Reducing uncertainty and improving robustness and spatio-temporal extrapolation capabilities remain key challenges in hydrological modeling especially for flood forecasting over large areas. Parsimonious model structures and effective optimization strategies are crucially needed to tackle the difficult issue of distributed hydrological model calibration from sparse integrative discharge data, that is in general high dimensional inverse problems. This contribution presents the first evaluation of Variational Data Assimilation (VDA), very well suited to this context but still rarely employed in hydrology because of high technicality, and successfully applied here to the spatially distributed calibration of a newly taylored grid-based parsimonious model structure and corresponding adjoint, over a large sample. It is based on the Variational Data Assimilation (VDA) framework of SMASH (Spatially distributed Modelling and ASsimilation for Hydrology) platform, underlying the French national flash flood forecasting system Vigicrues Flash. It proposes an upgraded distributed hourly rainfall-runoff model structure employing GR-based operators, including a non-conservative flux, and its adjoint obtained by automatic differentiation for VDA. The performances of the approach are assessed over annual, seasonal and floods timescales via standard performance metrics and in spatio-temporal validation. The gain of using the proposed non-conservative 6-parameters model structure is highlighted in terms of performance and robustness, compared to a simpler 3-parameters structure. Spatially distributed calibrations lead to a significant gain in terms of reaching high performances in calibration and temporal validation on the catchments sample, with median efficiencies respectively of NSE = 0.88 (resp. 0.85) and NSE = 0.8 (resp. 0.79) over the total time window on period p2 (resp. p1). Simulated signatures in temporal validation over 1443 (resp. 1522) flood events on period p2 (resp. p1) are quite good with median flood (NSE; KGE) of (0.63; 0.59) (resp. (0.55; 0.53)). Spatio-temporal validations, i.e. on pseudo ungauged cases, lead to encouraging performances also. Moreover, the influence of certain catchment characteristics on model performance and parametric sensitivity is analyzed. Best performances are obtained for Oceanic and Mediterranean basins whereas it performs less well over Uniform basins with significant influence of multi-frequency hydrogeological processes. Interestingly, regional sensitivity analysis revealed that the non conservative water exchange parameter and the production parameter, impacting the simulated runoff amount, are the most sensitive parameters along with the routing parameter especially for faster responding catchments. This study is a first step in the construction of a flexibe and versatile multi-model and optimization framework with hydbrid methods for regional hydrological modeling with multi-source data assimilation.

Regional flood estimation is an important issue in hydrology to anticipate and reduce the damages caused by extreme rainfall events. Approaches based on event simulation are particularly suitable to address this. As research has demonstrated the seasonality of rainfall characteristics, many flood frequency estimation approaches take into account rainfall seasonality to include seasonal fluctuations. For an event-based approach, since its hydrological model is initialized for each rainfall event, its performance is very sensitive to the initial states of the model. The seasonality of its hydrological model could thus become a decisive factor. Due to the complexity of the regionalization method, very few flood frequency estimation approaches based on event simulation have been regionalized at a large scale and do not consider the seasonality of hydrological parameters. This is the case for the SHYREG method studied in this article. Using data from HYDRO French database and SAFRAN, we discuss several adapted configurations considering the seasonality of both rainfall and hydrological parameters during its calibration and regionalization phase. Tests were carried out on 1929 catchments throughout France. Rather than calibrating a constant annual parameter for the hydrological model, we calibrated “winter” and “summer” parameters based on different observed flow quantiles (“seasonal”, “annual”, or “both”). Criteria on flood quantiles were calculated for different samplings. We also discuss the representativeness of seasonal parameters for the regionalization procedure and hydrological coherence observed from this seasonal parameterization. It seems that calibrating parameters based on seasonal flow quantiles helps reproduce annual quantiles, while the opposite is not possible. Among all the calibration configurations, calibration performed on both seasonal and annual flow quantiles makes the largest improvement compared to the initial annual parameterization method. It can correctly restitute seasonal flood quantiles for both calibration and validation catchments, with an obvious improvement in terms of estimating flood frequency in ungauged sites. It shows that the seasonality of hydrological parameters is worth considering for a regional flood estimation approach.

Low water levels are a seasonal phenomenon, which can be long, short, and more or less intense, affecting entire watercourses. This phenomenon has become a concern for many countries who seek better understanding of the processes that affect it and learn how to optimally manage water resources (pumping, irrigation). Consequently, a lumped rainfall model at daily time step (GR) has been defined, calibrated, and regionalised over French territories. The input data come from SAFRAN, the distributed mesoscale atmospheric analysis system, which provides daily solid and liquid precipitation and temperature data throughout the French territory. This model could be improved, in particular to more accurately simulate the hydrological response of watersheds interacting with groundwater. The idea is to use piezometric data from the ADES bank, available in France, and to use it for the calibration phase of the hydrological model. The analysis was carried out across ten French catchments that are representative of various hydrometeorological behaviours and are located in a diverse hydrogeological context. Each catchment must be represented by a piezometer that closely represents the main aquifer that interacts with the basin. This piezometer is located on part of the watershed that is most covered in terms of its drainage network, and closest to its outlet. Different signal processing methods are used to characterise the relationship between the fluctuation of river flow, piezometric levels and rainfall time series. Potential processing methods will be carried out in the temporal domain. To quantify groundwater table inertia and that of the catchment area, correlograms were calculated from daily chronicles of flows and piezometric levels. A cross-correlatory analysis was set up to see, in more detail, the correlations between the flow rates (especially base flows) and piezometric level time series. This type of analysis makes it possible to study relationships between various observations, and tests were carried out to take this information into account during the phase of the calibration of hydrological model parameters. These different analyses will hopefully help us to use piezometric data to consolidate the quality and robustness of the modelling.

Une approche conceptuelle parcimonieuse a été développée pour la quantification et la gestion de la ressource en eau sur des territoires dépourvus d'information. La base de données Web LoiEau présentée dans cet article, constitue le résultat de l'application de cette méthode à l'exutoire de plus de 130 000 bassins versants non instrumentés sur le territoire national. L'approche régionale exploite une information hydrométéorologique récente et de bonne qualité la plus exhaustive possible (données climatiques de la réanalyse SAFRAN et données hydrologiques de la banque Hydro) et résulte d'avancées méthodologiques réalisées sur la détermination d'une structure bien optimisée d'un modèle hydrologique journalier adapté à une gamme de fonctionnements hydrologiques variés, contraint seulement par deux paramètres pour permettre sa régionalisation et son utilisation sur des bassins non jaugés. La base de données Web LoiEau fournit des chroniques hydrologiques simulées de 1958 à 2018 au pas de temps journalier, à partir desquelles de multiples indicateurs hydrologiques sont extraits, permettant de caractériser la ressource en eau dans son ensemble (étiage, saisonnalité, bilan). Cette base est diffusée pour les services de l'État via une interface Web. L'étude des incertitudes liées à l'échantillonnage des données observées a permis d'établir des intervalles de confiance pour chaque indicateur hydrologique. Un indice de confiance en la méthode est aussi proposé pour chaque bassin afin de qualifier le degré d'applicabilité des résultats dans le cas où les spécificités locales les rendent inappropriés.A parsimonious conceptual approach has been developed for the quantification and management of water resources in territories without information. The LoiEau Web database, presented in this article, is the result of applying this method to the outlet of more than 130 000 ungauged catchments over the French territory. The regional approach exploits recent and good quality hydro-meteorological information that is as exhaustive as possible (the SAFRAN reanalysis and the Hydro database) and is the result of methodological advances made in determining a well-optimized structure of a daily hydrological model adapted to a varied range of hydrological process, constrained by only two parameters to allow its regionalization and use on ungauged basins. The LoiEau Web database provides simulated hydrological data from 1958 to 2018 at daily time steps, from which multiple hydrological indicators are extracted to characterize water resources (low flow, seasonality, mean annual streamflow). This database is distributed via a Web interface. Uncertainties were calculated by resampling of the observed data and allowed to calculate confidence intervals for each hydrological index on each catchment. A confidence index in the method is also proposed for each catchment in order to qualify the degree of applicability of results in the event that local specificities make them inappropriate.Mots clés : modélisation pluie-débitméthode régionaleindice d'étiagebase de données sur l'eauKeywords: regionalized daily rainfall-runoff modellow-flow indiceshydrological database RemerciementsNous remercions l'Agence Française pour la Biodiversité (AFB) pour son soutien financier ainsi que sa chargée de mission Claire Magand pour ses conseils avisés durant le suivi du projet. Nous remercions aussi Météo-France pour la mise à disposition des données SAFRAN pour la mise en œuvre de ce travail de recherche autour de la modélisation hydrologique.

L'observatoire hydrologique du Réal Collobrier situé dans le sud-est de la France à proximité du littoral méditerranéen (massif des Maures) et géré par Irstea depuis 1966, constitue un site de référence en hydro-climatologie régionale. En raison du réseau dense de mesures des pluies et des débits, ce site offre une occasion unique d'évaluer les tendances hydrométéorologiques méditerranéennes à long terme. La végétation est composée de forêts principalement calcifuges sur des sols cristallins (maquis de bruyère, chêne-liège, pin maritime et châtaignier). L'influence humaine directe a été négligeable au cours des 50 dernières années. L'occupation du sol est demeurée presque inchangée, à l'exception notable d'un incendie de forêt en 1990 qui a touché un petit sous-bassin hydrographique. Par conséquent, les changements dans la réponse hydrologique des bassins versants sont causés par des changements dans les conditions climatiques et/ou physiques. Cette étude examine les changements sur la période d'observation de 50 ans, à l'aide de séries de précipitations et de débits. L'analyse utilise plusieurs indices climatiques décrivant des modes de variabilité distincts, à des échelles de temps interannuelles et saisonnières. Des indices hydrologiques décrivant les épisodes de sécheresse, en particulier en termes de durée et de sévérité sont également utilisés. Les tendances sont évaluées à l'aide du test statistique de Mann-Kendall. L'analyse montre qu'il existe une tendance marquée à la diminution des ressources en eau du bassin versant en réponse aux tendances climatiques, avec une augmentation de la sévérité et de la durée de la sécheresse. Mais les changements sont variables d'un sous-bassin à l'autre en fonction de leur propre fonctionnement hydrologique.

Calibration of a conceptual distributed model is challenging due to a number of reasons, which include fundamental (model adequacy and identifiability) and algorithmic (e.g., local search vs. global search) issues. The aim of the presented study is to investigate the potential of the variational approach for calibrating a simple continuous hydrological model (GRD; Génie Rural distributed involved in several flash flood modeling applications. This model is defined on a rectangular 1 km2 resolution grid, with three parameters being associated with each cell. The Gardon d'Anduze watershed (543 km2) is chosen as the study benchmark. For this watershed, the discharge observations at five gauging stations, gridded rainfall and potential-evapotranspiration estimates are continuously available for the 2007–2018 period at an hourly time step. In the variational approach one looks for the optimal solution by minimizing the standard quadratic cost function, which penalizes the misfit between the observed and predicted values, under some additional a priori constraints. The cost function gradient is efficiently computed using the adjoint model. In numerical experiments, the benefits of using the distributed against the uniform calibration are measured in terms of the model predictive performance, in temporal, spatial and spatiotemporal validation, both globally and for particular flood events. Overall, distributed calibration shows encouraging results, providing better model predictions and relevant spatial distribution of some parameters. The numerical stability analysis has been performed to understand the impact of different factors on the calibration quality. This analysis indicates the possible directions for future developments, which may include considering a non-Gaussian likelihood and upgrading the model structure.

A parsimonious conceptual approach has been developed for the quantification and management of water resources in territories without information. The LoiEau Web database, presented in this article, is the result of applying this method to the outlet of more than 130 000 ungauged catchments over the French territory. The regional approach exploits recent and good quality hydro-meteorological information that is as exhaustive as possible (the SAFRAN reanalysis and the Hydro database) and is the result of methodological advances made in determining a well-optimized structure of a daily hydrological model adapted to a varied range of hydrological process, constrained by only two parameters to allow its regionalization and use on ungauged basins. The LoiEau Web database provides simulated hydrological data from 1958 to 2018 at daily time steps, from which multiple hydrological indicators are extracted to characterize water resources (low flow, seasonality, mean annual streamflow). This database is distributedviaa Web interface. Uncertainties were calculated by resampling of the observed data and allowed to calculate confidence intervals for each hydrological index on each catchment. A confidence index in the method is also proposed for each catchment in order to qualify the degree of applicability of results in the event that local specificities make them inappropriate.

The Real Collobrier hydrological observatory in southeastern France close to the Mediterranean coast, managed by Irstea since 1966, constitutes a benchmark site for regional hydro-climatology. Because of the dense network of stream gauges and raingauges available, this site provides a unique opportunity to evaluate long term hydro-meteorological Mediterranean trends. The vegetation is composed of forest mainly calcified on crystalline soils (maquis of heath, cork-oak, maritime pine and chestnut). Direct human influence has been negligible over the past 50 years. The land use/land cover has remained almost unchanged, with the notable exception of a wildfire in 1990 that impacted a small sub-catchment. Therefore changes in the hydrological response of the catchments are caused by changes in climate and/or physical conditions. This study investigates changes in observational data using up to 50-year daily series of precipitation and streamflow. The analysis used several climate indices describing distinct modes of variability, at inter-annual and seasonal time scales. Trends were assessed by the Mann-Kendall test. The analysis also used hydrological indices describing drought events based on daily data for a description of low flows, in particular in terms of timing and severity. The analysis shows that there is a marked tendency towards a decrease in the water resources of the Real Collobrier catchment in response to climate trends, with a increase in drought severity and duration. But the changes are variable among the sub-catchments according to their own hydrological functioning.