Estimating irrigation demand for water management on a regional scale: I. ADEAUMIS, a simulation platform based on bio-decisional modelling and spatial information

Abstract In most countries, planning and management of water resources has become a very important issue. Part of this problem, accurate estimation of water demand by agriculture is a key need for water management. In south-western France, the main water manager (CACG) uses a decision support system (DSS) called RIO, based on a simple representation of the agricultural and physical system. This DSS is satisfactory for average years but fails in years characterised by extreme weather conditions. To improve RIO, the ADEAUMIS simulation platform has been developed. ADEAUMIS is a bio-decisional model linked to a spatialised input data base. The bio-decisional model works on a daily time step and includes a model of plant development coupled with a model of action. The database includes information relative to irrigated area, weather and agricultural practices. Soil information has been shown to be unnecessary. The methods used to collect these data had to respect the operational constraints of the water manager, namely early delivery and cheap acquisition. At the present stage of development, one method per type of input data is used except for the estimation of irrigated area where one can choose between three different methods. Irrigation demand is calculated for individual simulation units and then aggregated over the study area. Applied to the Baise sector of the Neste system for 2 years, 1998 and 2000, ADEAUMIS showed its ability to simulate a realistic dynamics of water withdrawal over the region, but also its great sensitivity to estimation of irrigated area. This platform, developed for strategic water management, could be easily used for real-time water management or planning. Extending the objective from quantitative to qualitative water management would certainly require more complex modelling.