Page 7: Statistics - Predicting Hurricane Destruction
Two scientists put the hurricane damage models to the test and find they come up short
Florida experienced a record number of hurricanes last year. Was anyone prepared for these monster cyclones? Certainly not the Florida insurance companies.
According to UCF Professor of Statistics Mark Johnson and Chuck Watson, founder of Kinetic Analysis Corp. of Savannah, Georgia, the hurricane damage models used by insurance agencies include too much uncertainty about the risk the insurance companies face each year and how much they should charge policyholders.
Johnson and Watson studied 324 public and private models used by insurance companies to predict wind speeds, damage, and estimated losses. They found vast differences among the models and that most models failed to account for key variables, like soil moisture and leftover debris from prior hurricanes. Johnson and Watson published their results and suggestions for improvement this fall in a Bulletin of the American Meteorological Society article, which received considerable attention.
Johnson and Watson anticipate that better damage models will improve the insurance industry for both the companies and the policyholders. Instead of the current system of rate filing, which uses stacks of vague and contradictory damage estimates to determine rates, soon proposed rates will be calculated through detailed algorithms. “Those tests are essential when you’re dealing with insurance,” Watson explains, “because you’re experimenting with people’s lives and the economic livelihood of insurance companies.”
In addition to the economic benefits, the improved damage models can offer residents new insight into how to best protect their property from harm. Already, as part of their work for the Florida Department of Community Affairs, Johnson and Watson showed that required, state-wide shutters are not the best way to mitigate Florida’s hurricane damage. Even though shutters are valuable in the southern and in coastal areas of Florida, Johnson and Watson found that spending resources on tree maintenance is more likely to reduce damage in the northern part of the state.
“Those tests are essential when you’re dealing with insurance—because you’re experimenting with people’s lives and the economic livelihood of insurance companies.”
Thanks in part to Johnson’s and Watson’s work, Florida legislators are now considering changes to some of the state’s policies regarding hurricane insurance. Johnson and Watson have also analyzed hurricane damage models for other states, including Georgia and North Carolina, and plan to work with the North Caroline Department of Insurance during the next insurance rate filing period.
Mark Johnson and Chuck Watson operate a website [hurricane.methaz.org] that displays real-time damage prediction and tropical cyclone tracking systems for hurricanes and typhoons around the world. The site features interactive online mapping of over 30 hurricane track models, wind, wave, and storm surge forecasts, structure damage predictions, potential for power outages, and a unique “perceived effects” map that indicates what conditions may be like during the worst of the storm using easy-to-understand terms such as “branches breaking,” “moderate structure damage,” or “scattered power outages.” The damage estimates included on the site are based on a more comprehensive database of properties than what insurance company models typically use.
Want to know more?
Hurricane website: http://hurricane.methaz.org
Mark Johnson, mejohnso@mail.ucf.edu
Chuck Watson, ccwatson@mail.methaz.net
| 
Printer Friendly Version of this Article