4D.3 Medium-Term Hurricane Catastrophe Models: Validation and 5 Year Forecast

Monday, 31 March 2014: 4:30 PM
Garden Ballroom (Town and Country Resort )
Alison L. Dobbin, RMS, London, United Kingdom; and P. S. Wilson

The ability to estimate hurricane landfall rates is of great value to insurers and reinsurers for capital, premiums and claims planning. Climate variability is a major source of uncertainty for the industry. Insurers are often interested in quantifying hurricane risk on short to medium term timescales and in recent years, catastrophe models have started to explicitly account for climate variability through 'medium term catastrophe models' that provide a view of hurricane risk on a 1 to 5 year horizon. This is in contrast to 'long term' models that integrate across the climate variability of the longest available observational time series.

In this work we describe a medium term North Atlantic hurricane model that has been used to provide a probabilistic forecast of hurricane landfall activity on a 5 year horizon. By comparing the performance of this ‘medium term' model against a ‘long term' model, we assess the value of the medium term catastrophe model from an insurance industry perspective. Finally, the forecast U.S. landfall rates for the period 2014-2018 will be presented.

- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner