Improving Flood Damage Assessment Models in Italy

Abstract:

The use of Stage-Damage Curve (SDC) models is prevalent in ex-ante assessments of flood risk. To assess the potential damage of a flood event, SDCs describe a relation between water depth and the associated potential economic damage over land use. This relation is normally developed and calibrated through site-specific analysis based on ex-post damage observations. In some cases (e.g. Italy) SDCs are transferred from other countries, undermining the accuracy and reliability of simulation results. Against this background, we developed a refined SDC model for Northern Italy, underpinned by damage compensation records from a recent flood event. Our analysis considers both damage to physical assets and production losses from business interruptions. While the first is calculated based on land use information, production losses are measured through the spatial distribution of Gross Value Added (GVA). An additional component of the model assesses crop-specific agricultural losses as a function of flood seasonality.

Our results show an overestimation of asset damage from non-calibrated SDC values up to a factor of 4.5 for tested land use categories. Furthermore, we estimate that production losses amount to around 6 per cent of the annual GVA. Also, maximum yield losses are less than a half of the amount predicted by the standard SDC methods.