As catastrophe models incorporate more physical science and reduce their dependence on statistical projections of past events, they are better able to represent evidence of climate change in weather events, according to Dr Claire Souch, director international model management, for Risk Management Solutions (RMS)
This development is important, Souch said, because the world we face is going to be different than it was on an average of the last 100 years. If catastrophe models can incorporate changing weather patterns, they will help the insurance industry plan for its long term survival.
“At first the models dealt with events outside the normal loss experience across historical claims. Now the two are merging and we are moving from a catastrophe view to incorporate some of the background loss,” she explained.
Both Souch and Tom Larsen, senior vice president, EQECAT, said that catastrophe models have improved continuously. Larsen said that data quality, too, is getting better many places. “These improving models mean that insurers are better able to map their risks and understand them on a finer basis,” he said.
Correlations in weather patterns are one aspect that catastrophe models may in future be able to represent as they draw more on physical science. David Stephenson, reader in statistical climatology, University of Reading, argued that weather hazards are not stationary and that events may be inter-linked “The assumption in the models is that events are independent, but a lot of meteorological events are not independent.” Catastrophe and climate models should be brought together to get a more dynamic view of weather events, he urged.
The two modellers also emphasised the importance of not looking for single number answers from the models, but said users should take the “good, solid clues” offered by their probability curves to develop strategy. “Catastrophe models are one tool we can use to help us adapt to a changing climate,” said Souch.
Lee Coppack is editor of Catastrophe Risk Management.