Evaluations of extreme wave heights around Hainan Island and their uncertainty induced by decadal variations of input variables

Evaluation of extreme significant wave height (SWH) is essential for marine economy and engineering. However, insufficient sampling timespans may bring considerable uncertainty in such evaluations since SWHs can exhibit significant interdecadal variations under global climate change scenarios. Especially in areas like the South China Sea (SCS), where strong waves are closely related to typhoon events, the frequency of typhoons changed decadally. This study takes the area around Hainan Island in the northwestern SCS as an example, builds a fineresolution numerical model, and applies multiple (Gumbel, Weibull, and Generalized Pareto) distribution models to systematically analyze the uncertainty in the extreme SWH evaluation caused by the decadal variations of input variables. The results show that inputting 1980-2000 and 2001-2021 data subsets can cause considerable differences in the calculated extreme SWHs. It indicates that either timespan subset is sufficient to determine a credible result. If the interdecadal variation of typhoon frequency is included by the Poisson distribution or the long-term trend of SWHs is extrapolated by the non-stationary parameters, the differences will become even more remarkable. This study emphasizes the importance of a sufficient sampling timespan and suitable models to avoid uncertainty in evaluating the extreme SWHs.