Statistics Based on Atlantic Hurricane Activity Can Overestimate Risk

New research on the link between the formation of hurricanes in the Atlantic basin and U.S. landfall activity suggests that using Atlantic basin hurricane activity as a proxy for landfall activity can lead to erroneous estimates of both landfall risk and potential insured losses, according AIR Worldwide Corp. Researchers found that a higher number of tropical storms in the Atlantic basin does not translate to an equivalent increase in hurricanes or landfalling hurricanes.

AIR researchers found that a storm’s genesis location, or starting point, greatly influences its probability of making landfall along the North American coastline. The pattern of hurricane genesis locations changes from year to year and by comparing the pattern for a particular season, such as that of the 2004-2005 season, to long-term climatological patterns, one can better understand why in some years the proportion of storms making landfall is high, while in other years it is low.

“By only focusing on the 2004 and 2005 seasons, it is easy to forget that every hurricane season is unique and actual landfall activity is a function of complex interactions between a range of environmental factors such as genesis location, sea surface temperatures and the depth of warm ocean waters, wind shear and atmospheric steering,” said Dr. Peter Dailey, director of research in atmospheric science at AIR Worldwide.

AIR’s research can be used to analyze the landfall probabilities of the two strongest storms of the 2007 season — Category 5 hurricanes Dean and Felix — based on their genesis locations. Dean and Felix, which were the only storms this year to achieve greater than Category 1 status, both took southerly tracks across the Caribbean and eventually made landfall along the coasts of Mexico and Central America.

“Contrary to popular belief, the U.S. did not ‘dodge a bullet’ with respect to Hurricanes Dean and Felix,” stated Dr. Dailey. “Based on where these storms formed and how they would track under typical steering conditions, our research shows that Hurricane Dean had a low chance of making landfall as a hurricane and Felix was much more likely to strike the Mexico or Central America coastline than the U.S.”

Sea surface temperatures in the Atlantic basin have been warmer than average every year since 1995. However, the percentage of Atlantic basin storms that make U.S. landfall as hurricanes has been below the long-term average of 14 percent in nine of those 13 seasons. In 2007, only one of 15 named storms made U.S. landfall as a hurricane, or less than 7 percent. More significantly, total wind energy in 2007 was 33 percent below average despite two Category 5 storms.

“The seasonal forecasters correctly projected that a higher-than average number of tropical storms would form in the basin in 2007,” continued Dr. Dailey. “But it’s much more difficult to predict not only how many of these storms will become hurricanes, but more importantly how many will make landfall as hurricanes. Like many past seasons, the 2007 season showed that an elevated number of tropical storms does not always translate to more hurricanes or more landfalling hurricanes. In 2007, sea surface temperatures were not as warm as some scientists expected and significant wind shear suppression by La Niña did not materialize as they had anticipated. Clearly there’s a danger in assuming that one or two single seasons are indicative of a paradigm shift in hurricane risk. While 2004 and 2005 were both very active seasons, they were not good predictors of activity in 2006 and 2007.”

In addition to a standard view of hurricane landfall risk based on over 100 years of historical data and over 20 years of research and development, the AIR U.S. hurricane model includes an alternative view of landfall risk under warm sea-surface temperature conditions. Under warm ocean conditions, AIR estimates U.S. insured losses to be 15 percent higher than the long-term average. In some areas, such as the Northeast, AIR expects little difference from the long-term average. AIR intends to again submit its U.S. hurricane model for certification under the rigorous standards of the Florida Commission on Hurricane Loss Projection Methodology (FCHLPM) in 2008 for the 12th consecutive year.

Source: AIR Worldwide, www.air-worldwide.com