Insurance

In this article we describe the difficulties and opportunities related to the use of climate modeling results in the insurance industry. Specific problems like reliable estimates of climate sensitivities, the probabilistic nature of climate predictions, the necessity of spatial downscaling, and problems with long causal chains are discussed.

In the second part of the paper we propose solutions to the outlined problems. Those include a so called two-tier approach and probabilistic forecast systems to deal with necessarily imperfect prediction schemes.

In the third part of the paper we describe for different time scales the type of usable forecasts and list some potential benefits for insurance companies. The paper finishes with the outline of an integrated forecast system and conclusions.