Constraining extreme sea levels along the European coasts from a large ensemble of climate models

The storm surge contribution to coastal extreme sea levels along the European coastlines has been explored using a set of hydrodynamic numerical simulations. When forced by high-resolution atmospheric fields, simulated storm surge time series display good correspondence with observations. Because of their length, accuracy and consistency, these numerical data have been widely used to characterise coastal extreme sea levels, in terms of their magnitude and probability of occurrence. These outputs are then often used to infer coastal hazards and risks. However, higher risks associated to the most extreme events, represented by return periods substantially longer than the simulated time span, are generally accompanied by large uncertainties, thus limiting the robustness of long-term coastal risks assessments based solely on these otherwise valuable datasets. One way to reduce these uncertainties is increasing their sample size. Here, we do so by running a number of hydrodynamic simulations forced by mean sea level pressure and surface wind fields from a set of initialised climate models from the Decadal Climate Prediction Project (DCPP) over a domain covering the European coasts (excluding the Baltic Sea) and amounting for a total of 9000 years. Hydrodynamic simulations forced with atmospheric pressure and wind fields from these models, once are biased-corrected, result in a much larger dataset of coastal storm surges. Large datasets also provide information on the probability of extreme sea levels that are plausible in the current climate but for which there is no observational evidence.