Optimal Decisions That Reduce Flood Damage along the Meuse: an Uncertainty Analysis

In December 1993, the river Meuse ooded and caused a damage of about 250 million Dutch guilders. This prompted the Dutch government to initiate a project to investigate and compare strategies that can reduce future losses due to ooding on the basis of several criteria, including some concerning uncertainties. To obtain decisions that reduce ood damage, one should account for the following uncertainties: the river discharge, the ood damage given the discharge, the downstream water level given the discharge, the costs and the yields of extracting sand and gravel, and the costs of constructing embankments. These uncertainties can best be represented by probability distributions, where Bayes' theorem can be used to update subjective prior information with observations. When the loss is deened as the net present discounted value of the costs of decisions minus the yields of decisions plus the remaining mean ood damage over an unbounded time-horizon, decision theory can be used to obtain optimal decisions with respect to the following three decision criteria: the criterion of minimal expected loss, the criterion of minimal uncertainty in the loss, and the criterion of maximal safety. By using simulation, the present situation and ve strategies have been analysed. The strategy with minimal expected loss and maximal safety is based on widening the summer bed in the south of Limburg, lowering the summer bed in the middle and north of Limburg, and constructing 62 km of embankments around the remaining bottlenecks along the Meuse.