Flood management through mitigation structures distributed over the catchment area: food for thought on design and probabilistic assessment

Abstract

In order to design or diagnose the operation of a set of flood prevention structures, whether large or small, it is necessary to study their operation on the scale of the watershed, and therefore to define scenarios of input (rainfall, flow) on this scale. The approach, which consists in independently defining "single frequency" design floods for each section of the hydrographic network studied, each representative of a flood with a given return period, is deficient when structures are introduced. Indeed, it is difficult to pass on the effect of a structure on the flow quantiles of the downstream sections. However, these probabilistic maps are necessary in order to calculate probabilistic damages, then the expected annual damages (EAD). They must also be estimated in the presence of projected structures in order to evaluate the so-called residual EAD. The difference between the reference and residual EADs represents the risk reduction to be included in a cost-benefit analysis to evaluate the project. We will describe methods to obtain these probabilistic maps in the presence of structure(s), and will present current developments. Obviously, damage reduction is only one indicator among others: protection of people, impacts on geomorphology and ecosystems to name a few. The definition of integrated strategies, multi-objective and therefore multi-criteria, therefore poses new challenges that we will only be able to touch on.