Floods are a major natural hazard in the Mediterranean region, causing deaths and extensive damages. Recent studies have shown that intense rainfall events are becoming more extreme in this region but, paradoxically, without leading to an increase in the severity of floods. Consequently, it is important to understand how flood events are changing to explain this absence of trends in flood magnitude despite increased rainfall extremes. A database of 98 stations in southern France with an average record of 50 years of daily river discharge data between 1959 and 2021 was considered, together with a high-resolution reanalysis product providing precipitation and simulated soil moisture and a classification of weather patterns associated with rainfall events over France. Flood events, corresponding to an average occurrence of 1 event per year (5317 events in total), were extracted and classified into excess-rainfall, short-rainfall, and long-rainfall event types. Several flood event characteristics have been also analyzed: flood event durations, base flow contribution to floods, runoff coefficient, total and maximum event rainfall, and antecedent soil moisture. The evolution through time of these flood event characteristics and seasonality was analyzed. Results indicated that, in most basins, floods tend to occur earlier during the year, the mean flood date being, on average, advanced by 1 month between 1959–1990 and 1991–2021. This seasonal shift could be attributed to the increased frequency of southern-circulation weather types during spring and summer. An increase in total and extreme-event precipitation has been observed, associated with a decrease of antecedent soil moisture before rainfall events. 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 antecedent soil moisture and flood event runoff coefficients that is remaining stable over time, with dryer soils producing less runoff and a lower contribution of base flow to floods. In a context of increasing aridity, this relationship is the likely cause of the absence of trends in flood magnitudes observed in this region and the change of event types. These changes in flood characteristics are quite homogeneous over the domain studied, suggesting that they are rather linked to the evolution of the regional climate than to catchment characteristics. Consequently, this study shows that even in the absence of trends, flood properties may change over time, and these changes need to be accounted for when analyzing the long-term evolution of flood hazards.

Neural networks are used to forecast hydrogeological risks, such as droughts and floods. However, uncertainties generated by these models are difficult to assess, possibly leading to a low use of these solutions by water managers. These uncertainties are the result of three sources: input data, model architecture and parameters and their initialization. The aim of the study is, first, to calibrate a model to predict Champagne chalk groundwater level at Vailly (Grand-Est, France), and, second, to estimate related uncertainties, linked both to the spatial distribution of rainfalls and to the parameter initialization. The parameter uncertainties are assessed following a previous methodology, using nine mixed probability density functions (pdf), thus creating models of correctness. Spatial distribution of rainfall uncertainty is generated by swapping three rainfall inputs and then observing dispersion of 27 model outputs. This uncertainty is incorporated into models of correctness. We show that, in this case study, an ensemble model of 40 different initializations is sufficient to estimate parameter uncertainty while preserving quality. Logistic, Gumbel and Raised Cosine laws fit the distribution of increasing and decreasing groundwater levels well, which then allows the establishment of models of correctness. These models of correctness provide a confidence interval associated with the forecasts, with an arbitrary degree of confidence chosen by the user. These methodologies have proved to have significant advantages: the rigorous design of the neural network model has allowed the realisation of models able to generalize outside of the range of the data used for training. Furthermore, it is possible to flexibly choose the confidence index according to the hydrological configuration (e.g., recession or rising water table).

Artificial neural networks (ANNs) adjust to the underlying behavior in the dataset using a training rule or optimizer. The most popular first-and second-order optimizers, Adam (AD) and Levenberg–Marquardt (LM), were compared with the aim of predicting extreme flash floods of a runoff-dominated hydrological system. A fully connected multilayer perceptron with a shallow structure was used to reduce complexity and limit overfitting. The inputs of the ANN were determined by rainfall–water level cross-correlation analysis. For each optimizer, the hyperparameters of the ANN were selected using a grid search and the cross-validation score on a novel criterion (PERS PEAK) mixing the persistency (PERS) and the quality of flood-peak restitution (PEAK). For an extreme and unseen event used as a test set, LM outperformed AD by 25% on all performance criteria. The peak water level of this event, 66% greater than that of the training set, was predicted by 92% after more training iterations were done by the LM optimizer. This shows that the ANN can predict beyond the ranges of the training set, given the right optimizer. Nevertheless, the LM training time was up to five times longer than that of AD during grid search.

The risk of drought impacting the drinking water and agricultural production is worrying in the developed countries, especially in a changing climate context. To manage and prevent this phenomenon, real-time monitoring and predictive systems are emerging as the key solutions. In the field of artificial intelligence, neural networks are one of these predictive systems. This family of parameterized models is a composition of neuronal functions, which apply a non-linear transformation from their inputs to their outputs. These networks are able to learn a hydro(geo)logical system behaviour using a database composed of observed inputs (rainfall, evapotranspiration, etc.) and outputs (groundwater level, discharge, etc.), thanks to an algorithm minimizing a cost function between observed and simulated outputs. However, it remains difficult to assess the uncertainty generated by these models, possibly leading to misinterpretations by the end users. These uncertainties are mainly of three types. The first is related to the input data. Indeed, hydrosystems are surface elements whereas meteorological inputs are punctual elements. The interpolation error can, therefore, be significant because of the lack of knowledge between gauging stations. The second is the neural network model architecture itself. It is possible to deal with this source of uncertainty using regularization methods. Finally, the neural networks are submitted to uncertainties related to parameter initialization, before the training step. The initial parameters may have an important impact on the results. In this paper, we address the prediction of theBlavetgroundwater level (Bretagne, France). In order to assess uncertainties, we will first focus on the parameters initialization of the model. Neuronal models are optimized using cross-validation and early stopping. Then, an ensemble model is realized, in which each member is the result of a unique set of parameters initialization. The purpose of the study is to define how many initializations are necessary to obtain a reasonable confidence interval for forecasts, with the smallest interval and the higher rate of observed points inside this interval. The best model will be determined using cross-validation scores thereby ensuring optimal robustness. We show that, in this case study, an ensemble model of 20 different initializations is sufficient to estimate uncertainty while preserving quality. In the second part, the resulting ensemble model will be used to estimate the global model uncertainty using probability density functions (pdf) applied to the distribution of groundwater level data and cross-validation scores of forecasts. It reveals that the groundwater level predictions are composed of two mixed distributions. Therefore, we will use the expectation-maximization algorithm (EM) to obtain parameters of mixed models. Mixed normal and mixed Gumbel laws, among five mixed distributions assessed, give the best groundwater distribution and are able to generate an abacus drawing uncertainty of model.