Did the 2006 hurricane season come as a surprise? No, say the modeling vendors. By their calculations, what happens during any given season is never unexpected. Why? Because they're in the business of storm modeling, not storm forecasting.
"The storms in 2006 were in the range of what was expected," says Tom Larsen, senior vice president of product management at Eqecat Inc. "2006 was on the low end of the expectations. But then, similarly, 2004 and 2005 hurricane seasons were on the high end."
Larsen said Eqecat would once again next spring, at the beginning of the hurricane season, remind its clients that the Atlantic has entered into "an era of higher-than-average activity." "We are reaffirming that we are in a period of higher activity," he says. "But within that higher activity, there is still a range of outcomes . . . 2006 is not that surprising."
At AIR Worldwide Inc., Eqecat's rival, Peter Dailey, director of AIR's atmospheric sciences, says his company's message to clients has always been the same. While there are some "key drivers" of hurricane activity, he says, there are just too many factors involved in trying to gauge whether a storm will make landfall.
It's hard enough to accurately predict how many hurricanes any given season will spawn. But to predict the number of landfalls brings in a whole other set of variables. "Landfall in itself has its own climate factors that enter into the likelihood," says Dailey.
The difficulty with which to predict the number of hurricanes, of course, never stops the forecasting industry from issuing bulletins every spring on the number of storms expected during the upcoming Atlantic hurricane season. "What happened this season is I don't know of any example of a seasonal forecast that did not predict elevated levels of activity," says Dailey. "They all did."
Those forecasts, left to experts with government weather offices like the National Oceanographic & Atmospheric Administration's National Hurricane Center, and researchers employed by hurricane research departments like the ones located at Colorado State or University College London, predicted an active 2006 season.
The forecasters concluded the season would be active because of the idea that warmer-than-average sea surface temperatures in the Atlantic would persist, says Dailey.
Manuel Lonfat, senior meteorological modeler with Risk Management Solutions Inc., says it's important to remember that forecasters use statistical, not meteorological, methods to predict the Atlantic hurricane season.
Statistical methods make use of historical data to determine relationships between the state of the climate and hurricane activity. Using those relationships, forecasters make predictions about hurricane activity for the upcoming season.
Still, those environmental conditions are difficult to predict, says Lonfat.
But the forecasting world may soon change. Some researchers have started using methods based on models that compute the future state of the climate from the current state by solving the equations of motion, according to Lonfat.
"Although very promising, such methods are still in early stages of development, when it comes to their application to seasonal forecasting of hurricanes," he says.
December 1, 2006
Copyright 2006© LRP Publications