ORLANDO – With the last two Atlantic hurricane seasons relatively quiet for the U.S. coast, you might think that weather experts, marina owners and insurance companies would be happy. But they're not.

In fact, the quiet is what worries them. Not because it portends a big one – despite the science behind predicting the number of named storms each year, no one can predict where they'll make landfall – but because the perceived lull may make coastal residents complacent.

After two years without a major hurricane, some may forget the damage wrought by storms like Katrina, Rita and Wilma.

"Our biggest fear is that there will be a warning, you'll evacuate, it will be a pain in the butt, nothing happens," says Steve Letro, the meteorologist in charge of the National Weather Service in Jacksonville, Fla. "Then there'll be a second warning, you'll evacuate, and it will be a pain in the butt but nothing happens. Time three you don't evacuate – and you get hit."

At the two-day BoatU.S. Hurricane Preparation Symposium here, much of the talk centered on the need for boat and marina owners to stay prepared, regardless of the predictions for the pending hurricane season.

Collective memory is short, the experts say, and the pain of the 2004 and 2005 seasons, when Katrina, Rita and other storms punished U.S. cities, is already foggy history for too many boaters. But Katrina alone caused roughly $750 million in damage to recreational boats and destroyed 75 percent of the matinas along the 150 miles of Gulf Coast, according to estimates by BoatUS.

"It really doesn't matter how many storms there are," Letro says. "Last year, we had 15 named storms. It was busy season, including two Category 5 hurricanes. How much did we hear in the U.S? Not much – but don't tell that to the people in Central America."

THE 2008 PREDICTIONS

For what it is worth – and Letro says it's not much – early predictions for the 2008 hurricane season, which officially starts June 1, show 13 named storms and seven hurricanes, including three major ones.

The probability for a landfall on the U.S. coastline is 60 percent, a bit higher than the average of 52 percent for the last 100 years.

STEVE LETROThe "steering currents" that drove Hurricane Wilma in 2005 (above) were well defined, making prediction of where she would strike much easier. For Hurricane Ophelia (below), the currents made successfull prediction far more difficult.

STEVE LETRO

Letro's annual presentation, which focuses on major weather events and their prediction, is always a big hit at the symposium, which ended yesterday. His major point at last year's conference was that we are in a period of unusually active and strong tropical cyclone activity – and such cycles generally last 25 to 40 years.

This year, an equally clear message came through: Hurricane forecasting is limited by inherent error and for those on coasts, the bigger area covered by a storm warning is far more important that what is commonly called smaller "cone of death" that estimates the probability of a strike.

FORECASTING STORMS

A storm's path is driven by many factors, and "steering currents" are among the hardest to pin down. That accounts for much of the forecasting error, Letro says.

Take two real-life examples from 2005: Hurricanes Wilma and Ophelia. Prediction in one case was successful while less so in the other.

The steering currents driving Hurricane Wilma were well defined, and all the track models showed the storm would move toward the Yucatan and then curve back across the southern part of Florida. And that's just what happened.

"If the stream is well-defined and in the middle, the storm will be pretty easy to predict," Letro says. "With Wilma, the timing was off but the forecast itself was very, very good."

Earlier that season, Ophelia was different. The steering currents were not as defined, and the forecast tracks over time resembled what Letro and his colleagues call "the squashed spider syndrome."

"The forecasts were all over the place," he says.

DECREASING ERRORS

Of course, the science of forecasting has improved over time.