Multiscale Optimization for the Management of Runoff Risks in Agricultural Watersheds

This paper presents a novel combined optimization approach to the problem of runoff risk management in agricultural watersheds. Starting with important rains, the runoff risk increases according to the agricultural activities practiced by farmers on their parcels of land. To minimize the risk, we propose to optimize the distribution of crops on the parcels. The redistribution must take into account a number of agronomic, economic, and hydrologic constraints. Two criteria allow the estimation of the runoff risks. On the scale of the digital terrain model, a cellular automaton evaluates a global criterion at the watershed outlet. At the scale of the parcels, a local criterion evaluates the modification of risks generated by a variation of crops on two parcels. We compare the results of these two criteria on the watershed of Villers-Ecalles, Haute Normandie, France.