Monte Carlo‐based flood modelling framework for estimating probability weighted flood risk

Abstract

AbstractThis study presents a new Monte Carlo‐based flood inundation modelling framework for estimating probability weighted flood risk using a computationally efficient graphics processing unit (GPU) two dimensional (2D) hydraulic model. The 2D flood model is programmed in the GPU framework providing a unique ability to run numerous simulations in a short period of time, permitting the integration of 2D hydraulic modelling into Monte Carlo analysis. The framework operates by performing many 2D flood simulations of randomly sampled input parameters to develop a spatially varied flood hazard map. The probabilistic framework is demonstrated using a 1% annual probability flood event and simulating 1000 different flood simulations by randomly selected peak flows of the Swannanoa River in Buncombe County, USA. The results, in general, display benefits of probabilistic flood risk approach compared with a single simulation approach. The latter approach underestimated 28% of flood risk relative to the former. As the number of simulations increased from 1 to 1000, areas identified as low danger and judgment zone increased by 87.4% and 36.8% respectively, whereas the high danger zone increased by 9.3%. In conclusion, the new Monte Carlo flood risk modelling framework has the ability to provide improved accuracy of flood risk and greater insight into the spatial distribution of flood risk.