La désagrégation de l'information de pluie

Rainfall disaggregation

Cemagref Lyon, France

Abstract

Rabifall data is an essential input to hydrological models used to estimate runoff in natural or urban watersheds.Severe storms are often associated with floods and its material damages and losses of human lives. Generally, rainfall analysis in hydrology has been developed from the analysis of observed data, in contrast to the analysis of rainfall generation mechanisms observed in meteorology. The punctual observation provided by rain gauges network and the rainfall fields images obtained by radar and satellite highlighted the strong space-time rainfall variability and the uncertainties of its modeling. Despite the complexity of the process and the seeming chaos, rainfall measurements allowed a better understanding of rainfall systems organization and to progress in the space-time representation of rainfall. Several rainfall models were thus developed. They try to reproduce rainfall at different scales, by respecting its statistical and geometric properties as observed in data. Among the available models, we can distinguish the models based on the disaggregation of rainfall into space and time scales smaller than those of data acquisition. These models try to meet the needs of knowledge of rainfall at small scales required by hydrological models, which is a function of its application. In this context, some approaches and principles of rainfall disaggregation models are herein pointed out.