De-clustering of S H -Time Series for Applying the Peaks-Over-Threshold Method

In this work, a new de-clustering method is presented and applied for the estimation of design values of S H by means of the “Peaks-OverThreshold” (POT) method. An essential requirement of the POT method refers to the independency among exceedances of a certain threshold level u . Usually, the independency is approached by applying to the exceedances of u , a subjectively selected time separation interval. The proposed De-Clustering Algorithm (DeCA) is based on a predefined percentage of wave energy reductions for discriminating independent sea-state clusters. In this way the subjective factors involved in the time separation interval selection are avoided and the physics of the phenomenon is properly taken into account.