Forecasters tweak way to predict hurricane path

[HurricaneJoe22]

Forecasters tweak way to predict hurricane path

They hopes landfall accuracy improves

By Ken Kaye
Staff Writer
Posted March 18 2005

Despite highly accurate track forecasts last summer, the National Hurricane Center plans to discourage residents from putting too much faith in that skinny black line by offering a new supplement: a strike probability map.

The color-coded map will be posted on the center's Web site, http://www.nhc.noaa.gov, by the start of the season June 1. It will display the odds that a given area could face hurricane- or tropical-force winds. It also will try to project a storm's size and intensity.

That way, officials hope residents won't be caught off guard as they were with Hurricane Charley, which initially was predicted to hit Tampa but struck about 100 miles south, in Punta Gorda and Port Charlotte on Aug 13.

"If people had been looking at this probability product rather than focusing on the line, nobody would have come to the conclusion they were safer in Port Charlotte than Tampa," hurricane specialist James Franklin said Thursday.

He noted that the probability of Charley's core hitting Tampa or Port Charlotte was the same: 30 percent 24 hours before landfall.

The hurricane center already provides a strike probability chart, showing the chances that a storm will come within 75 miles of larger coastal cities.

But the map, to be used on an experimental basis, would provide the most comprehensive information as storms approach, Franklin said.

"It considers a whole wide range of things that can happen in the forecast," he said.

The hurricane center is considering getting rid of the skinny black line, or projected path, and replacing it with circles or dots to emphasize a much broader cone of uncertainty.

Bill O'Brien, director of Palm Beach County emergency management, said the map should be a good tool to prompt preparations.

"Any legitimate information we can get, the more logic people have to make good decisions," he said.

On the other hand, Tony Carper, director of Broward County emergency management, said he is concerned the map might show low probabilities when storms are four or five days out -- and persuade residents to let down their guard.

"If you're on the fringes, four or five days out, your probability might be less than 10 percent, and you ultimately could be the one to get the landfall," he said. "That's problematic."

Todd LeDuc, assistant fire chief for the Broward Sheriff's Office, said residents would prefer to know as much as possible.

"As a resident, the more information I have and the simpler it is to digest, the more likely I am to take the appropriate actions," he said. "People tend to wait until the last minute to take precautions."

In 2004, the National Hurricane Center in Miami-Dade County achieved record-breaking accuracy in its track forecasts of nine hurricanes, five tropical storms and one subtropical storm.

The center's average track error was 67 miles when the systems were 24 hours away, a marked improvement over the 10-year average error of 90 miles. Those errors are half what they were 15 years ago.

The steady improvement is the result of more sophisticated computer forecast models, as well as more and better information about the atmosphere that surrounds hurricanes, Franklin said.

The best forecast last year: Hurricane Frances was projected to strike 40 miles from its actual landfall point when the storm was 24 hours out in early September. When the system was five days out, the track error was 144 miles, which is less than half the 10-year average error of 367 miles.

"We have to remind people that we're not perfect every time," Franklin said. "That was a big problem with Charley; people assumed the forecast would be perfect."

Hurricane forecasters concede they have struggled to improve intensity forecasts, which are only slightly better now than they were 15 years ago.

Last year, the average intensity error was about 12 mph when a storm was 24 hours out and 16 mph when a storm was 48 hours out.

Franklin said forecasting the path of last year's storms was "relatively easy."

"Anytime you have a lot of storms in the deep tropics that take relatively straight tracks, your errors tend to be lower," he said.

Ken Kaye can be reached at kkaye@sun-sentinel.com or 954-385-7911.

[StormChasr]

Despite highly accurate track forecasts last summer, the National Hurricane Center plans to discourage residents from putting too much faith in that skinny black line by offering a new supplement: a strike probability map.

Highly accurate track forecasts? Is this article serious? Do they not know of the lef and right biases of the models? I reckon that the news media forgets that the forecasts were for Ivan to strike Tampa Bay at first, and for Jeanne to make a "right angle" turn into Daytona Beach.