For the first time, the National Hurricane Center plans to develop six- and seven-day track forecasts for an entire storm season.
It also plans to do a better job keeping them secret, as they leaked out to the public last year, cutting the experiment short.
“We weren’t able to button them up; people were finding them,” said James Franklin, the center’s top hurricane specialist.
The long-range forecasts are one of four in-house experiments the hurricane center plans to conduct in the upcoming season, which officially starts June 1.
The others include extending the tropical weather outlook from two to five days, developing advisories for disturbances before they become tropical systems and issuing watches and warnings if disturbances threaten land.
On another front, the hurricane center also plans to employ more sophisticated models in hope of improving forecasts – notably intensity predictions, an area where forecasters have struggled for decades.
As one of its major goals within the 10 years, the hurricane center hopes to generate six- and seven-day forecasts to give residents, businesses and the military more time to prepare.
The long-range forecasts are to be honed in-house for about two years because for now they are expected to be in error by 400 miles or more. That means a storm predicted to hit Fort Lauderdale could end up striking southern Georgia.
Currently, the center issues track forecasts out to five days, with an average error of about 275 miles.
Franklin said the hurricane center started to develop six- and seven-day projections last year only to find they were surfacing on some websites. The experiment was immediately ceased.
“We feel pretty strongly that we don’t want these experimental forecasts being used before we know how good they are,” he said.
For the past two years, the hurricane center has been experimenting with extending its tropical weather outlook from two to five days – with good results. The outlook, which can be found on the hurricane center’s website, nhc.noaa.gov, displays disturbances that hold potential to develop and provides the odds that they will grow into tropical storms or hurricanes.
“The skill is there,” Franklin said. “We’re trying to decide how to display that information.”
The hurricane center was less successful when, for the first time last year, it attempted to write advisories for disturbances on an experimental basis. Franklin said it was difficult to predict their track and intensity.
“That wasn’t surprising, because those systems aren’t well defined. You don’t have a center, and you don’t always know where it is,” he said, adding that forecasters hope to do better this year.
Another first this year: A computer model will be fed Doppler radar information directly from one of the hurricane hunter aircraft while it’s investigating a storm. That might lead to improved intensity predictions.
“There’s a fair amount of evidence that Doppler data in high resolution models can improve intensity forecasts,” Franklin said. “That’s exciting.”
Overall, he said the idea of the experiments and new models is to broaden the hurricane center's forecasting capabilities and “aggressively attempt to improve our service.”