Is local flood hazard assessment in urban areas significantly influenced by the physical complexity of the hydrodynamic inundation model?

• Published in: Journal of hydrology (Amsterdam) 2020-01, Vol.580, p.124231, Article 124231
• Main Authors: Costabile, Pierfranco, Costanzo, Carmelina, De Lorenzo, Gianluca, Macchione, Francesco
• Format: Article
• Language: English
• Subjects: 2-D inundation models; Flood hazard; Model benchmarking; Urban areas; Vulnerability assessment
• ISSN: 0022-1694; 1879-2707
• DOI: 10.1016/j.jhydrol.2019.124231

•Influence of physical complexity of the 2-D models on flood hazard assessments.•Models benchmarking based on the product between flood velocity V and water depths h.•Similar inundation extent but different prediction of the product hV.•Potential effects of the simulated inundation on buildings, vehicles and people.•Significant variations of the vulnerability class when using simplified models. Flood hazard in urban areas is usually assessed by the estimations of parameters like flood extent, water depths, flow velocities and other related quantities. These hydrodynamic variables can be computed using flood inundation numerical models, the complexity of which has been analyzed in several papers in the literature. However, the question of how detailed a model should be in order to give reasonable results is still an open issue and no shared and general conclusions have been reached so far. One of the main reasons for this is that model benchmarking is very often carried out with reference to the flooded areas extent only, sometimes analyzing also the simulated water depths values and, quite rarely, evaluating the computed velocities. This is in contrast with the main purpose of the simulation models which are, from a practical point of view, advanced tools for flood hazard and vulnerability assessments. Finally, the lack of sufficient data on inundation extent and depths for the calibration and validation of the numerical models has been recognized, especially in complex situations that often characterize the flood evolution in a given area due to bridge obstructions, interaction with buildings and so on. In this paper, the performances of three different approaches to two-dimensional flood modeling (fully dynamic, diffusive and porosity approaches) have been evaluated with reference not only to the hydrodynamic variables, water depths and velocities, but also focusing the attention on their product and, thus, to the consequences associated to their different estimations on the vulnerability assessment. This analysis has been focused first on a flash flood flow experiment in a simplified urban district and then to the event which occurred in the city of Crotone (Calabria, Italy) in 1996. The results highlight that fully-dynamic modeling should be the unavoidable reference tool when the goal of the urban flood mapping activity is not limited to the evaluation of the flood-prone areas extent but involves also the local estimation of flood hazard/vulnerability