Statistical Assessment of Storm Surge Scenarios Within Integrated Risk Analyses

This paper summarizes the results from calculating exceedance probabilities of different storm surge scenarios developed within the XtremRisK project, which were then used as boundary conditions for integrated risk analyses for the city of Hamburg in the Elbe Estuary and Sylt Island off the coastline of Schleswig-Holstein in northern Germany. A stochastic storm surge model is developed to simulate a large number of synthetic and physically consistent storm surge scenarios, before the resulting samples are used to calculate bivariate joint exceedance probabilities of the storm surge heights (total water levels with tides included) and intensities. The Copula theory is exploited and functions from the Archimedean family are used to build the statistical models. The latter are extended to the three-dimensional case to also take into account wave conditions where appropriate. The uncertainties associated with the results from the multivariate extreme value analyses are briefly discussed and (where possible) quantified and recommendations of how to exploit the presented methodologies in future applications are given.