Stochastic modeling for flood emergency management in urban areas

The management of emergency during urban flooding requires different numerical models and risk analysis methodologies, compared to those used for non urban areas. From a hydraulic point of view, the flow propagation in a street network involves uncommon aspects in modeling a two-dimensional phenomenon. A complete risk analysis can involve many buildings of remarkable social and cultural interest, (e.g., hospitals, monuments, schools, etc.), and also facilities and transportation networks reliability analysis. A statistical analysis concerning casualties occurred in flood events during the last century in Italy (Guzzetti et al., 1996) has shown that most casualties occurred in the streets, crossing bridges or driving cars. So, knowledge of the flow characteristics in the street network seems to be necessary for a reliable flood risk analysis in urban areas. In this paper hydrological and hydraulic models are used to simulate the propagation of flood waves along the Tiber river valley, including tributaries, and to simulate the inundation of the urban area of Rome. As a result, the flow characteristics in the street network are evaluated. Because of the randomness of rainfall, of the uncertainties in the flood hydrograph evaluation and propagation, and hence in the risk analysis, the described process required a stochastic approach.