Identifying Modelling Issues through the Use of an Open Real-World Flood Dataset

The present work deals with the reconstruction of flood wave that hit Mandra town (Athens, Greece) on 15 November 2017, using framework forensic hydrology. flash event was caused by a huge storm high level spatial and temporal variability, which part Medicane Numa-Zenon. included: (a) post-event collection 44 maximum water depth traces in town; (b) hydrodynamic simulation employing HEC-RAS MIKE FLOOD software. derived open dataset (which also includes additional data required for modeling) is shared community possible use as benchmark case model developers. With regards to modeling issues, we investigate calibration strategies computationally demanding cases, test whether calibrated parameters can be blindly transferred another simulator (informed modeling). Regarding calibration, it seems coupling an initial screening phase simple grid-search algorithm efficient. On other hand, informed concept does not our study area: every numerical has its own dynamics while are grey-box nature. As result, modeler should always skeptical about their global use.