The future of modeling hurricane catastrophes

AIR Worldwide developed the industry's first "CAT Model" (catastrophe model) over 20 years ago. Today, they are widely used to assess risk from natural events such as hurricanes, floods, and winter storms. The first such model, the AIR U.S. Hurricane Model, estimates the financial impact of hurricane landfalls along the U.S. coastline by simulating damage to individual properties from wind and storm surge across a wide range of plausible storms. Advancements in this and other models evolve in parallel with improvements in the data, the science, and through innovative application of the models.

Data serves as the building blocks for each of four CAT model components: event generation, local hazard calculation, damage calculation, and financial impact to exposure. Future CAT models will advance as HURDAT, wind speed observations, and land use data improve in coverage and accuracy. Exposure data, in particular, has seen dramatic improvements over the last decade, and the sensitivity of damage estimates to high-quality exposure information has become increasingly apparent. As more is learned about the dynamics of hurricane wind fields, the latest science and engineering can be incorporated into CAT models. The models are also becoming more versatile. By coupling a CAT model with projections of future climate, for example, the financial impact of climate change can be quantified probabilistically, making use of ensembling techniques used operationally.

As model development moves forward, the future of catastrophe modeling will be guided by original research made available to all stakeholders and peer-reviewed by outside experts. AIR has embarked on several such projects applying catastrophe modeling techniques to pressing real-world problems such as the relationship between SSTs and hurricane landfalls and the influence of climate change on flood risk. As this published work is vetted across industry and the scientific community, not only will CAT models evolve, but cross-discipline collaborations will become more common further reinforcing their credibility.