Extreme value analysis of the typhoon-induced surges on the coastal seas of South Korea

Abstract

This research mainly focuses on providing new occurrence probabilities of local abnormal sea level rise (SLR) events, i.e., typhoon-induced surge heights, in any given year for the formulation of coastal management and climate change response policies. The 50- and 100-year return period levels of the typhoon-induced surge height on the seas covering the Korean Peninsula are obtained by adopting extreme value analysis with the two most widely applied probability distributions: the generalized extreme value (GEV) and Weibull distributions. The extreme values used in the above extreme value analysis are obtained from the deterministic Sea, Lake, and Overland Surges from Hurricanes model after validation. The statistical estimation is validated by satisfying the hypothesis testing procedure with respect to the form of a probability distribution using chi-squared (Chi-S) and Kolmogorov–Smirnov (K–S) goodness-of-fit tests. The optimal curves consisting of a bird’s-eye view of the return period levels of the typhoon-induced surge heights are selected by evaluating the statistical performance indicators of the goodness-of-fit tests, namely the weighted sum χ2 and supremum Dn of the Chi-S and K–S goodness-of-fit tests, respectively. In this research, the GEV distribution-based fitting curves are selected as the best-fit curves. The increasing pattern of its inverse cumulative distribution function tends to capture the extreme values of the typhoon-induced surge height. Since the numerically obtained typhoon-induced surge heights were employed to visualize the return period levels of the typhoon-induced surges on the seas of Korea, this approach provides more detailed information for the management of SLR-related natural hazards to coastal populations.