Constraining Extreme Storm Surges Along the European Coasts From a Large Ensemble of Climate Models

Abstract The storm surge contribution to extreme sea levels along the European coastlines is investigated using hydrodynamic numerical simulations forced by atmospheric pressure and surface winds from a large ensemble of initialized climate models from the Decadal Climate Prediction Project experiment. The outputs, representative of the climate since 1960, amount for a total of 8,000 years of data, thus increasing significantly the sampling size of extreme simulated events compared to typical decadal‐long hydrodynamic hindcasts. The extended DCPP‐forced storm surge data set, once bias‐corrected, provides information on the probability of storm surges that are plausible in Europe in the current climate but for which there is no observational evidence. Our results show that these unprecedented extreme events are on average 20% larger than the observed maxima, with values reaching up to 1 m. The new data set also enables the uncertainties in the probabilities to be constrained significantly (e.g., up to two orders of magnitude for 500‐year return periods). This permits a more robust quantification of coastal hazards and risks, particularly for the most extreme events with return periods that are substantially longer that the observational records.