Research on Threshold Selection Method in Wave Extreme Value Analysis

Abstract

Climate change poses higher requirements on ocean engineering design, and reasonable estimation of design wave heights plays a crucial role in coastal protection and offshore engineering. Extreme value analysis is widely used in frequency calculations of wave parameters, among which the peak over threshold method based on the generalized Pareto distribution is proven to be an effective method, and the different selection of extreme wave samples in this method has a great influence on the calculation results. In this study, long-term significant wave height series were utilized to investigate the long-range correlation of significant wave heights, and thresholds were determined based on the changes of long-range correlations. This approach assumes that extreme events and non-extreme events are generally caused by different physical processes, where extreme events result from massive disturbances leading to abnormal states, and long-range correlations are not affected or minimally affected by extreme events. Thus, thresholds can be determined based on changes of long-range correlations by removing extreme events. Comparing this method to graphical diagnostic techniques, we demonstrated its rationality in determining extreme wave height thresholds. Moreover, the automatic threshold selection offered by this method helps to mitigate errors associated with subjective judgments in traditional approaches.