A new European coastal flood database for low–medium intensity events

Abstract

Coastal flooding is recognized as one of the most devastating natural disasters, resulting in significant economic losses. Therefore, hazard assessment is crucial to support preparedness and response to such disasters. Toward this, flood map databases and catalogues are essential for the analysis of flood scenarios, and furthermore they can be integrated into disaster risk reduction studies. In this study and in the context of the European Coastal Flood Awareness System (ECFAS) project (GA 101004211), which aimed to propose the European Copernicus Coastal Flood Awareness System, a catalogue of flood maps was produced. The flood maps were generated from flood models developed with LISFLOOD-FP for defined coastal sectors along the entire European coastline. For each coastal sector, 15 synthetic scenarios were defined focusing on high-frequency events specific to the local area. These scenarios were constructed based on three distinct storm durations and five different total-water-level (TWL) peaks incorporating tide, mean sea level, surge and wave setup components. The flood model method was extensively validated against 12 test cases for which observed data were collated using satellite-derived flood maps and in situ flood markers. Half of the test cases represented well the flooding with hit scores higher than 80 %. The synthetic-scenario approach was assessed by comparing flood maps from real events and their closest identified scenarios, producing a good agreement and global skill scores higher than 70 %. Using the catalogue, flood scenarios across Europe were assessed, and the biggest flooding occurred in well-known low-lying areas. In addition, different sensitivities to the increase in the duration and TWL peak were noted. The storm duration impacts a few limited flood-prone areas such as the Dutch coast, for which the flooded area increases more than twice between 12 and 36 h storm scenarios. The influence of the TWL peak is more global, especially along the Mediterranean coast, for which the relative difference between a 2- and 20-year return period storm is around 80 %. Finally, at a European scale, the expansion of flood areas in relation to increases in TWL peaks demonstrated both positive and negative correlations with the presence of urban and wetland areas, respectively. This observation supports the concept of storm flood mitigation by wetlands.