Probabilistic flood risk associated with dam-break

Assessment of flood risk associated with dam-break is crucial in dam emergency action planning to minimize loss of life. It is typically performed deterministically based on a single precautionary scenario. For example, the hypothetical collapse of concrete dams is assumed total and instantaneous with the reservoir level at the spillway crest level. However, a probabilistic approach could account for the uncertainties affecting flood hazard and risk assessments, both in dam-break scenario parameters and in the damage model. In this paper, a probabilistic method is proposed based on a set of dam-break scenarios characterized by different breach widths and reservoir levels. A probability, conditional to the dam-break event, is associated with each scenario. A hydrodynamic model based on the 2D shallow water equations is used to produce flood hazard maps for each scenario. Only the direct risk of loss of life is considered; therefore, the exposure is expressed through dasymetric maps of the population at risk for two social time scenarios (daily weekday and night/weekend). Vulnerability is assessed using a hazard-damage function available in the literature (with an associated confidence interval). Accordingly, the spatial distribution of the fatality rate of the population at risk in the study area is calculated as a function of flood hazard, flood severity understanding, and warning time. The following probabilistic maps are produced: a) the probabilistic inundation map, which identifies potentially floodable areas in the event of a dam-break with the corresponding probabilities; b) a probabilistic flood hazard map, which displays the spatial distribution of the probability-weighted averages of the flood hazard indexes calculated for each dam-break scenario, coupled with a map of the associated uncertainty expressed through a suitable coefficient of variation; c) a probabilistic flood risk map (in terms of loss of life), which provides the spatial distribution of a loss of life risk index with the associated uncertainty expressed in terms of confidence interval. The probabilistic maps offer more detailed and comprehensive information than the deterministic ones. The probabilistic method is applied to the case study of the hypothetical collapse of the Mignano concrete gravity dam (Arda River, northern Italy).