Hydrological models, robustness, climate change, SWAT, GR4J


Many studies about climate change impacts assessment are published every year. These studies commonly use a hydroclimatic modelling chain, whose principle is to feed impact models with climate models outputs. An important step in this process is to test the validity of impact models in a climate change context. However, this step is not frequently applied. The aim of this study is to test the robustness of two hydrological models with distinct conceptualizations: a global and empirical model (GR4J) and a semi-distributed and physically-based model (SWAT). They both have been calibrated and validated over climate contrasted periods. Despite a higher decrease of performance between calibration and validation for the GR4J model, both of them show relative robustness. Moreover, the stability of parameters between the two calibration periods shows that their value are not much influenced by the climate of the calibration period, and consequently remains valid during the entire projection period.