Analysis of the Future Change in Frequency of Tropical Cyclone-Related Impacts Due to Compound Extreme Events

Abstract

AbstractTropical cyclone-related hazards are often comprised of compound, connected events that individually amplify the total impacts. Often, hazard risk assessments focus on one factor rather than the compound nature of multiple forcing mechanisms. It is possible that extreme event analysis in a univariate context may underestimate the probabilities and impacts of extreme events. In this study, a framework addresses multivariate analysis of risk due to compound hazards related to tropical cyclone characteristics. Combinations of observations and simulations are used to identify possible frequencies of annual chance extreme events forced via connected individual events. The framework emphasizes the statistical dependence of multiple physical variables that contribute to extreme compound events when individual events are not extreme.To make the analysis clear, specific locations are analyzed using both univariate and a joint analysis. The joint analysis is conducted using copula functions to remove the restriction that marginal distributions need come from the same family of probability functions. The primary results suggest that univariate and joint return periods for key tropical cyclone-related hazards could shorten in the future and univariate frequency analysis may underestimate the magnitude of the extreme events because the univariate analyses do not account for the dependence structure between the paired environmental factors.