Optimisation « économique » vs « physique » des réseaux de surveillance de la qualité de l’eau

Monitoring networks are playing an important role in acquiring information to guide policies to improve surface water quality. This has led us to reflections regarding the economic value of this information using the Bayesian method. In this study, we seek to determine the optimal location of stations in a monitoring network, along a river, where the objective is to detect potential accidental pollution. With a theoretical model and a numerical simulation, we compare the economic value of information from two different networks: a ?physically? optimized network where the location of stations minimizes the time required to detect accidental pollution, taking into account only hydrological considerations, and an ?economically? optimized network where the location of stations allows the value of information to be maximized. This comparison is made for three vulnerability scenarios, namely environmental damage, generated by pollution: constant, decreasing or increasing along the river. Our results show that the benefit of optimizing the network economically rather than physically depends on the context, namely the number of stations and the vulnerability scenario. Thus, depending on the additional cost of economic optimization, physical optimization could be recommended.