Probabilistic flood risk assessment over large geographical regions

We develop a probabilistic model to estimate the rate of flood‐induced losses for a set of properties distributed over a large geographical region (e.g., a portfolio of insured properties within a country). The use of detailed physically based models over large areas becomes difficult due to the vast amount of data needed and the high implementation cost. The proposed model allows one to incorporate results from such detailed models but can also be used in regions that have not been studied in much detail. Minimal required information includes the rate and spatial extent of severe precipitation, the topography and river network from which regions at risk of flooding can be identified, and information on historical floods with an approximate delineation of the flooded area, and associated aggregate losses for at least a few major events. An application to river flood loss from residential buildings in Belgium is presented.