Use of dynamical weather and climate models for tropical cyclone risk assessment

The insurance and reinsurance industry is heavily reliant on quantitative data supplied by catastrophe (CAT) models for tropical cyclone risk assessment. Traditional CAT models use synthetic hazard data generated by sampling from a probability distribution fitted to historical data. The limitations of this approach have led to increasing interest in moving away from a purely statistical approach towards combined statistical and dynamical methods that provide an independent assessment of risk.

This powerful new approach to CAT modeling is in its infancy and two examples of how dynamical models can be used to improve risk assessment will be discussed here in detail. The first is the generation of new and independent event sets by extracting tropical cyclone information from global or regional climate model simulations. The second is the use of very high resolution dynamical simulations of landfalling hurricanes to generate a detailed inland wind swath. These wind swaths can then be used to inform the development of simple wind field models that can be run many times to generate a probabilistic wind swath.