Towards objective design of dry dams at watershed scale: how to take into account the spatial structure of the rainfall and its variability
Résumé
Structural measures must be designed at the right spatial scale, and in a coordinate manner, lest the works simply transfer the problems from one area to another. Upstream retention by dry dams is therefore interesting ; moreover, they do not hinder the river natural dynamics. The aim of this study is to test and improve a methodology to design dams distributed on the catchment. To take into account the spatial variability of the rainfall, we used a simulator of stochastic distributed rainfall fields. To obtain hydrographs on a 150km² catchments, a very simple rainfall-runoff model was then designed ; the computation units were 63 subcatchments. A dam can be placed at any subcatchement outlet. This models is a simplistic one, but the focus of the study is the influence of the input (rainfall) on the result (best location for 1 or 3 dams). It is possible to define an efficiency indicator and run an optimizing procedure. Rainfall spatial distribution is strongly impacting the result. Our tests confirm that the definition of the indicator and the choice of the rainfall event or set of rainfall events is subjective and leads to different results. Including or not extreme events in the set of input is a key question. This study illustrates that optimizing on hydrological variables is subjective. Taking into account the cost of damages is more relevant ; besides using annualised damages makes it possible to solve partly the problem of the weight of extreme events, because their effect is weighted by a range of frequency. This tool allowed us to design a promising methodology, to be tested on precise models. Moreover, it is a convincing didactic tool to advocate paying more attention to the rainfall used as input and to work at regime scale.
Origine : Fichiers produits par l'(les) auteur(s)