Risk: Windstorm modelling - A hazy outlook?

(From Reinsurance)

Byline: Dr Jayanta Guin, vice-president of research and modelling at AIR.

As the 2003-04 European winter storm season gets underway it is an excellent time to re-examine this complex and potentially costly peril.

The insurance industry will undoubtedly remember 1999, the year Anatol, Lothar, and Martin paralysed much of the continent. Insured losses from Lothar and Martin cost roughly EUR8bn in France alone.

Using a methodology developed by Denmark's Riso Labs for estimating return period winds, it is estimated that the wind speeds associated with Lothar and Martin have a return period of around 100 years. Clearly, this is not the worst case scenario and reinsurers must be prepared to handle larger losses.

Decisions on the management of (re)insurance portfolios worth hundreds of millions of dollars are made daily based on data derived from catastrophe models. Most models consist of three primary components: hazard, building vulnerability, and loss estimation. Modellers must incorporate a detailed understanding of both the science (meteorology, in the case of extratropical cyclones) and the impact of the danger to structures. Once total damages are estimated, insured losses are determined by applying policy conditions.

Physical modelling