Hurricane School

If the public learned anything from the 2004 hurricane season, a top National Weather Service official says, it should have been this: Don't focus on the skinny black line.

The weather service will use a black line to project the paths of hurricanes again this year, in spite of complaints that it causes needless panic and provides inadequate landfall warnings.

That line got a two-thirds vote of confidence from emergency managers and the news media in a recent survey. The problem lies in how it's interpreted, according to the director of the National Centers for Environmental Prediction. "We're finding that our customers put too much focus on the track, rather than accounting for the range of forecasts," says Louis Uccellini, whose NWS division oversees the National Hurricane Center in Miami. "We have a communications issue."

It's one of several lessons the weather service and its parent agency, the National Oceanic and Atmospheric Administration, learned from Charley, Frances, Ivan and Jeanne, the four hurricanes that swarmed over Florida last August and September.

Here are some others:

• Comparing multiple computer models leads to superior forecast guidance.
• Even as forecasts improve, the public's expectation for accuracy increases.
• Potential three-day and five-day tracks should be explained more clearly.
• The public should be reminded of all hurricane risks-wind, storm surges and flooding.
• "Off season" is a good time to maintain emergency response training and sustain public awareness.
• Successful hurricane preparedness and response requires strong partnerships among federal, state and local agencies.

As the 2005 hurricane season opens, federal forecasters are doing their best to remind people that storms don't always follow predictions and usually cut a path of destruction much wider than a line.

A prime example was Hurricane Charley, the forecasters' nightmare. In mid-August, hurricane warnings were up along almost all of Florida's west coast. The black line went from the southern Gulf of Mexico through Tampa Bay, but Charley went ashore 100 miles south at Charlotte Harbor. The storm took a sudden turn to the right less than 12 hours before landfall. It exploded with destructive power as it approached the coast, going from a Category 2 to a Category 4 in less than five hours. There wasn't enough time for many to get out of harm's way. "You really have to pay attention to that cone of uncertainty," says Uccellini, "and not just the line."

It may not have seemed so last year, but the government's hurricane-tracking record is improving. Yearly averaged forecast errors are 40 percent to 60 percent smaller today than they were in 1970, thanks to major upgrades in meteorological modeling and computer technology. Data from hurricane hunting aircraft also had a positive impact. The hurricane center charted record-low forecast errors in 12-hour to three-day forecasts in 2004. To further refine track area estimates this year, the hurricane center will add wind-speed probabilities to its forecasts.

The same progress hasn't been made in predicting hurricane intensity. The trend during the past 14 years essentially is flat, with a margin of error between 10 knots and 25 knots. Uccellini says there are "some solid science issues" to resolve and more stringent observation requirements. Guessing a storm's force more accurately remains a top government priority. Within the next three years, NOAA plans to outfit an airplane with Doppler radar to take the necessary wind measurements.