Estimating irrigation demand for water management on a regional scale: II. Validation of ADEAUMIS

Current procedures of water management often use simple models and parameterisation to predict demand that may fail in years with extreme climatic and hydrological conditions. The simulation platform ADEAUMIS, presented in a companion paper (), was developed to improve decision-making in strategic water management. In order to estimate irrigation demand over large areas, it includes a spatial data base and a bio-decisional model. The main variables of the data base are irrigated area, weather data and agricultural practices (crops, sowing dates, varieties and irrigation decision rules). ADEAUMIS was applied in the Baı̈se sector, in south-western France, in 1998 and 2000. This paper presents an evaluation of ADEAUMIS. Firstly, the operational validation was conducted. Estimated and true input variables and predicted and observed output variables of the bio-decisional model of ADEAUMIS were compared. Secondly, its credibility was evaluated by the water manager by comparing ADEAUMIS with the current decision support system (DSS) (called RIO), in order to determine if the new platform allows better decisions. The operational validation raises the problem of the quality of validation data. For the 2 years of simulation, data on water consumption and electricity use (for pumping) were used to construct the reference curve of water withdrawal during the irrigation season. This reference curve was compared to the estimations of ADEAUMIS. There was good agreement for irrigation demand over the whole region for one of the three methods used to estimate irrigated area. The quality and spatial coverage of the water withdrawal reference curve and of independent estimations of irrigated area, used as validation data, are discussed. It seems that a better fit to the reference curve could be obtained by improving the set of decision rules used to reproduce the average irrigation practice of the region, in particular the rules for starting irrigation and the rule for delaying irrigation following rainfall.