WEATHER RISK ASSESSMENT AT THE DAWN OF ENSEMBLE NUMERICAL WEATHER PREDICTION

摘要

In the changing world of insurance risk management, catastrophe (CAT) models are moving to the center of quantitative weather risk assessment. These models offer significant benefits for the companies that use them through automation, and the ability to manage risks and set rates proactively. Because these models help insurance companies to better understand the complexities of their risks, they can allocate capital more efficiently than before and remain more competitive. Ensemble NWP modeling provides the means to quantify forecast uncertainty. Using the output from the ensemble members, a probability density function (PDF) of potential losses can be created. Further, once the performance and reliability of an NWP ensemble model forecast system is characterized, its economic value as a function of cost-loss ratios (C/L_p) can be determined. Two classes of ensemble NWP model applications have been defined for CAT loss estimation. "Class 1" applications can provide probabilistic loss forecasts for real-time extreme events. "Class 2" applications involve the characterization of the losses that would be expected within a region's natural climate variability (often referred to as "regional loss profiles"). An example for each was discussed. Other Class 1 applications include medium range (three to seven day) predictions of individual events. Class 2 applications would include alternative regional loss profiles created from seasonal climate predictions. CAT models are continuing to become more sophisticated, and some have begun to incorporate NWP modeling technology. As the cost of computers and networks decreases, the insurance industry should find that the benefits of ensemble NWP modeling technology more cost effective.