Tropical Cyclone Damage Assessment for Predicting Effects of Climate Change on Caribbean Storms

Caribbean island nations are particularly susceptible to tropical cyclone (TC) impacts due to their tropical locations and proximity to favorable cyclone-forming conditions. Accurate TC risk models are essential to promoting resilience in vulnerable island nations and informing policy decisions for emergency procedures, allocation of aid, and economic relief such as insurance and reinsurance. The overarching goal of this study is to construct an open-source TC risk model that will accurately predict the effect of climate change on asset losses caused by Caribbean storm impacts. This project focuses on wind speed as the main variable used in hazard and vulnerability calculations. The hazard component of the model was derived here from IBTrACS data for historical TC tracks and applying an empirically-derived formula to calculate wind swaths. For the exposure values, the LitPop data, based on night light from space and the World Bank’s produced capital stock, were utilized to represent exposed assets. Vulnerability was modeled using Emanuel’s formula for wind-related vulnerability. Combining the hazard, exposure and vulnerability components allowed us to calculate asset losses for all storms that made landfall in the Caribbean. The simulated asset losses were compared with actual losses estimated by the EM-DAT data set. The results show good agreement between simulated and actual losses for some storms, but severe underestimates for others. We interpret this as a consequence of modeling hazard based on wind only, while EM-DAT represents total losses, including those from rainfall, storm surge, and flooding. We conclude that the model is potentially useful for assessment losses due to wind in Caribbean TCs.