Valeur de l'information des réseaux de surveillance de la qualité de l'eau. Impact des paramètres dans un modèle bayésien

The Water Framework Directive imposes increasing requirements on the acquisition of information about water quality. Because of the cost of this information, the rationalization of the water quality monitoring is crucial, what requires a best knowledge of the value of information generated by the water quality monitoring network. This article aims at taking forward this thought. We used the Bayesian frame to estimate the value of additional information according to the parameters: prior probability on the states of the nature, costs resulting from a bad decision and accuracy of the additional information. Then we analyzed the impact of these parameters on this value, particularly the role of the combination of prior probability and costs from bad decision which can increase or decrease the value of information according to the initial level of indecision. The results were illustrated by an empirical study: a river in the ?Bas-Rhin? in France.