Floods are considered the most frequent natural disaster world-wide and may have serious socio economic impactsin a community. In order to accomplish flood risk mitigation, flood risk analysis and assessment are requiredto provide information on current or future flood hazard and risks. Hazard and risk maps involve different data,expertise and effort, depending also on the end-users. More or less advanced deterministic approaches can be used,but intuitively probabilistic approaches seem to be more correct and suited for modelling flood inundation giventypical uncertainties. Two very important matters remain open for research: the calibration of hydraulic models(oriented towards the estimation of effective roughness parameters) and the uncertainties (e.g. related to data,model structure and parameterisation) affecting flood hazard mapping results. Both matters are strictly connectedand the performance measures represent the “metric” of this connection.Here, we test the ability of different performance measures based on binary and distributed information to calibrateand evaluate model predictions in a credible and consistent way and to reduce the uncertainty in probabilisticflood inundation maps for two hydraulic models: a two-dimensional inertial finite element model and a recentlydeveloped version of the LISFLOOD-FP model which solves a reduced form of the full shallow water equations ina highly efficient manner. These models are applied to the Imera river basin in Sicily probabilistic flood inundationmaps constructed for each performance measure calibration. Through a comparison of the resulting hazard maps,the influence these measure data on calibration and derivation of probabilistic flood mapping will be shown.
|Number of pages||1|
|Publication status||Published - 2013|