How good are symmetric triangular synthetic storms to represent real events for coastal hazard modelling

Abstract Coastal risk assessments rely on proper quantification of storm-induced erosion and flooding, and often involve calculations via numerical models. When the real time-series data of a storm are not available as forcing conditions and only bulk information is accessible, synthetic simplified time-evolutions are assumed. The most common approach in coastal studies uses a symmetric triangular storm shape, characterised by the assumptions that the peak of the waves occurs in the middle of the storm, and that the forcing varies linearly. This study aims to investigate this additional source of uncertainty in hazard estimation, using the XBeach-1D model, to assess the differences in simulated erosion and flooding associated with real and synthetic storm definitions. Analysis is performed for real conditions ranging from moderate to extreme at the Northern Adriatic and North-Western Mediterranean coasts, using beach profiles ranging from dissipative to reflective. The storm definitions generate considerable differences in terms of wave power and timing at the peak of the storm. When synthetic storms were applied, coastal hazards were not adequately reproduced in most of the simulated cases. The energy of the storms, profile characteristics, local storm climates, and water levels did not consistently influence the differences between the synthetic- and reality-based outputs.