Design of dry dams at watershed scale: lessons learnt from sensitivity analyses using a simple but consistent rainfall-runoff model
Résumé
We investigate the assessment of the overall efficiency of a set of dry dams on flood mitigation. To ensure the spatial consistence of our flood scenarios, we used a stochastic rainfall simulator to generate rainfall fields representative of the climatology of a 150 km² catchment near Lyon. The influence of the variability and spatial structure of the rainfall on the choice of the best locations for a set of dams was previously studied using a simplistic rainfall-runoff model. The estimated optimum is highly dependent to the choice of the sub-sets of events, confirming that assessments based on a limited number of scenarios are heavily flawed. A large set of events, representative of the regime, is indispensable. For further investigations of this approach, a cascade-of-reservoirs conceptual model now computes the rainfall-runoff transformation and can optionally be chained to a 1D-hydraulic models. This increases both results consistence and computation-time, however sensitivity analyses remain accessible. We studied optimisation of outlet dimensions in addition to dam location. The effect of a given solution at a given point is represented over a wide range of flood probability, versus peak flood return period, but also versus peak volume return period, which strongly influences the mitigation potential.
Origine : Fichiers produits par l'(les) auteur(s)
Loading...