Résumé
Distributed hydrological models allow us to represent several processes at various scales on the watersheds. Thus, verification of such models should be conducted on each of the represented processes, at various scales and not only on the discharge at the outlet. This kind of multicriteria verification is of importance in order to evaluate the global behaviour of the models. Nevertheless, the aim is not to represent all the plot scale processes but to evaluate which degree of complexity should be introduced in spatialised catchment scale models in order to well represent the main processes of the watershed. The REW concept (Reggiani et al. 1998, 1999) is interesting in this kind of evaluation procedure as it allows the description of the main local processes at the sub-watershed scale. A validation strategy of the REW-v4.0 model is undertaken on the Donga catchment, in Benin. The goal of this study is to evaluate the model ability to reproduce the discharge at the outlet, at intermediary stations, and also to reproduce the watertable dynamics and the vadose zone dynamics at the sub-catchment scale. For this work of watershed modelling, a strategy of observation has been undertaken since 1998. A network of 20 gauges allows a quite fine description of the spatial and temporal variability of rainfall. An automatic meteorological station measures climatic variables needed to compute a potential evapotranspiration rate. Soils were characterised with the help of Beerkan infiltration measures and a pedology map. We follow 6 discharge stations, 18 wells and several soil moisture stations on typical vegetation stands; all these data allow the verification of the model behaviour.