Team Trying to Predict Detailed Hurricane Locations and Losses

June 6, 2007

With hurricane season once again upon us, you may find it beneficial to check out an interesting new alternative information source.

Mark Johnson, a Univ. of Central Florida professor, and Chuck Watson, a Georgia researcher, have teamed to form Kinetic Analysis Corp.Their emphasis, in work sponsored by the Florida State Board of Administration, the Florida Department of Community Affairs, the World Bank, and other public and private groups, is on improving predictions of where a hurricane will hit each year, the expected dollar losses, and impacts on specific sectors, such as the offshore oil and gas industry, electricity providers, and Toxics Release Inventory (TRI) facilities. They are attempting to provide publicly available information that previously has been pursued primarily by private groups, such as the insurance industry, and kept private. Organizations funding KAC hope to use the information to improve emergency preparedness and response, building standards, land use plans, and similar efforts.

Johnson and Watson have been working on their prediction model for about a decade, but they keep adjusting it, so there is no way to accurately determine a long-term track record. And their predictions for 2006 raise questions about the model, since it was close on some counts, but way off on others. They say the 2006 busts were caused by the timing and strength of El Nino and La Nina in the central Pacific, and that they have changed their model to help remedy this. As some evidence of the usefulness of their model in other years, they say their predictions of dollar losses from Katrina, made before landfall, were pretty close, though moderately understated, despite strong protestations before landfall from many people that their dollar predictions were far too high.

With the latest model, they have tracked multiple factors around the world for 2007 from January to early May, and added that to data on tracks that hurricanes have taken since 1851. Their county-specific predictions for 2007, extending from south Texas to Cape Cod, MA, are that areas with much greater likelihood than normal of being hit include New Orleans and much of Louisiana, the Brownsville, TX, area, much of Florida, and most of the South and North Carolina coast. See UCF's May 23, 2007, release and KAC's 2007 forecast; Mark Johnson, 954-401-6249; and Chuck Watson, 912-398-9753. They also are predicting much higher odds of serious disruption of Gulf of Mexico oil and gas operations and facilities.

As noted earlier, there isn't a substantial track record for the accuracy of these predictions, so you may want to simply acquire the prediction for your audience area, then see how it pans out this year, before possibly using the information next year. Or you may want to fold this information in to a story in the days before a hurricane approaches, especially if the odds had been high that a hurricane would hit.

In addition, you may want to compare KAC's active tracking of each storm to that of NOAA's. In the past, the predicted tracks have been substantially different at times, KAC says. KAC updates its site frequently throughout the day as each storm approaches, as does NOAA.

You may also want to keep tabs on KAC's predictions for specific damage amounts and sectors affected.

Like KAC, NOAA is predicting an active hurricane season. However, NOAA doesn't attempt to predict, until shortly before landfall, where those storms will hit.

For many more hurricane resources, see SEJ's main hurricane page, and search the SEJ site for additional resources that haven't been linked to this page.


SEJ Publication Types: 
Topics on the Beat: