Author(s): Daniel Gonzalez-Marco; Agustin Sanchez-Arcilla; Panagiotis Prinos
Keywords: Extremes; Distributions; Probability; Uncertainty
Abstract: This paper presents a summary of results obtained in Task 2 of the EU (Framework Program 6) FLOODsite project. The paper deals with the estimation of extremes for risk assessment in fluvial, coastal and estuarine environments. The approach, techniques and conclusions are, however, general enough to be of application to other geophysical variables. The paper starts with a review of the data sets considered in the analyses, since the “length” and properties of these data (related to the climatology of the studied environment and the observational gear and policy) will determine many of the conclusions that follow. The selection of analysis technique and the proposed Probability Distribution Function will also depend on the data characteristics which are reviewed in the next section. The paper presents, by way of example, the conventional and Bayesian extreme analyses applied to the longest directional wave data series available in the Spanish Mediterranean coast. The obtained results, corresponding to a buoy off the Ebre Delta coast, show how the introduction of a priori information improves the estimated median values and allows defining the uncertainty intervals associated to the probabilistic distribution assumed. The paper ends with some remarks on present practice and how to improve the extreme estimation for the various sites considered in the EU (Framework Program 6) FLOODsite project.