Influence of rainfall spatial variability on hydrological modelling: a simulation approach

The link between rainfall space-time variability and hydrological modelling at a catchment outlet remains an open research issue in hydrology and it puts the following question : is it relevant to consider the spatial distribution of rainfall in hydrological modelling ? Radar images provide a detailed knowledge on rainfall spatial variability which was not available with rain gauge networks. Thus, different case studies, linked to different approaches, have been carried out in order to determine the interest of radar images for flood runoff modelling. However, the results of such analyses are sometimes contradictory and highlight the difficulty to evaluate the contribution of radar data to hydrological modelling since rainfall measurement errors and modelling errors can not be distinguished from the influence of rainfall variability. In order to overcome errors on rainfall data, to control the rainfall variability and the hydrological behaviour of catchments, we suggest proceeding by simulation. In this way, we have developed a simulation chain composed of a rainfall simulator (based on the turning-band method), a catchment simulator (using the Diffusion-Limited Aggregation method) and a distributed hydrological model (composed of a production function and a transfer function which is based on the Hayami model).