Impact-based flash-flood forecasting system: Sensitivity to high resolution numerical weather prediction systems and soil moisture

Abstract In recent years, continuous improvements have been made in weather forecasting and flood prediction with great benefit from Early Warning Systems (EWSs). Despite the continuous quest for innovation from the scientific and user communities, EWSs remain based mostly on hazard forecast, and the information on possible consequences and potential impacts is generally missing. In this work, a methodology for quantitative real-time impact assessment for flash floods is presented. The methodology uses a multi-model ensemble approach and considers soil moisture uncertainty. Moreover, the flood forecasting chain, which normally provides only the discharge probability of exceeding a given threshold, is extended to include a fully 2D hydraulic model and a damage estimation model to quantitatively assess impacts in terms of economic losses and the people involved. The procedure was tested on recent flood events occurring in Genoa in northwestern Italy. This paper discusses the potential challenges and opportunities offered by this approach in the decision-making workflow in an operational context.