Volatility in Tropical Cyclone Losses

Quantification of risk must deal not only with long time-averages but with temporal volatility and recognition of any underlying temporal trends, both of which are often dominated by rare but exceptionally destructive events.

This study will present the results of multiple 100-year simulations of synthetic Atlantic tropical cyclones, forced using output from a global climate model. The generated stochastic tropical cyclone tracks have been converted into insurance losses using a hurricane windfield model and a realistic exposure dataset that returns a reasonable average annual loss for Atlantic hurricane risk.

The work presented will address two topics: firstly, the volatility of results between the 100-year simulations and, secondly, any implication of temporal trends from the same datasets. Both topics will consider the volatility between simulations through the lens of the lifecycle of tropical cyclones in each season: from basin and landfalling storm frequency through to the aggregated seasonal insurance losses to identify the points along the lifecycle of storms where most volatility arises.