Next forecasting challenge: predicting hurricane's wallop

Despite big improvements in predicting a storm's track, researchers struggle to peg its intensity.

As the remnant of hurricane Dean spends itself over central Mexico, the 2007 Atlantic season's first major storm has earned a place in the record books. It's the third-strongest Atlantic hurricane ever recorded and one of a handful to strike land as a Category 5 storm – the most powerful rating for tropical cyclones.

It's also the kind of storm that can keep even day-shift forecasters at the National Hurricane Center awake at night. That's because, despite strong gains in predicting the track of these storms, forecasters still struggle to figure out how powerful they will become.

"When you look at the statistics for the last 25 years, you see very little improvement" in intensity forecasts compared with track forecasts, says Chris Davis, an atmospheric scientist who focuses on storm-forecast research at the National Center for Atmospheric Research (NCAR) in Boulder, Colo. "It is a very difficult thing to do." Researchers and forecasters are making progress, if glacially. This year, for example, the National Hurricane Center began using new hurricane-forecasting software that takes advantage of a broad range of information from satellites, hurricane-hunter aircraft, and other sensor platforms.

The new model "is designed to assimilate much more data," says Isaac Ginis, a hurricane researcher at the University of Rhode Island's Graduate School of Oceanography in Narragansett, R.I. The results are encouraging enough so that the new model will replace the old model soon, he adds.

Indeed, forecasters were fairly accurate in predicting hurricane Dean's ferocity. By the evening of Aug. 17, the National Hurricane Center said the storm would at least be on the verge of becoming a Category 5 within the next 72 hours. Dean crossed that threshold on Aug. 20.

But compared with track forecasts, which generally rely on large-scale wind patterns in the atmosphere, intensity forecasts rely on data from the ocean as well as the atmosphere. And the processes they must account for take place on time and spatial scales that challenge the current generation of computers and forecast software.

For example, researchers are keenly interested in what happens to sea spray and sea foam under and near a storm's eye walls, where tropical cyclones pack their fastest winds. These processes at the interface of ocean and atmosphere are the toughest to study because they occur where most instruments don't survive for long.

Hurricanes can kick spray 100 feet or more into the air, where the droplets start to evaporate. The question is: What happens to the exchange of heat between the ocean and the atmosphere. The droplets cool as they evaporate and fall back and cool the ocean surface. A cooler surface would tend to sap a storm's strength. But a portion of the airborne droplet evaporates, releasing water vapor, which carries heat to the rest of the storm. In principle, this heat should help sustain a storm's strength. Researchers are divided over whether, on balance, the heating or cooling effect is the most dominant.

Even if researchers zero in on the right answer, they still have to cope with the fact that the processes take place to differing degrees in different parts of the storm and often in locales so small that models can't represent them.

For its part, foam whipped up by high winds may help reduce the friction between the swiftly moving air and the rough ocean surface. That would help keep hurricanes strong.

But scientists are trying to sort out whether the friction rises, then reaches a plateau as winds exceed a certain speed, or whether friction actually decreases as the speed increases. Results from buoys in the path of hurricane Ivan in 2004 suggest that friction does decrease, allowing the storm to spin up and draw more moisture into itself. Here too, the effect can vary widely across a storm's expanse.

As if scientist didn't have enough to think about, Alexander Khain is suggesting another wrinkle to the intensity problem – the presence of tiny particles called aerosols from natural and human causes on land. At a presentation earlier this year at a hurricane and climate conference in Crete, Dr. Khain, a researcher at The Hebrew University in Jerusalem, offered data suggesting that just before hurricane Katrina struck the Louisiana coast in 2005, aerosols got sucked into the outer reaches of the storm, weakening it just before landfall. In principle, as a storm inhales aerosols, this would enhance storm activity at the hurricane's outermost boundary but weaken its center.

The idea remains controversial, notes NCAR's Dr. Davis. Other factors, such as how dry the air might have been, could play a bigger role.

As scientists continue to tackle the intensity-forecast problem, "we have a much better handle on what matters" than 25 years ago, he says. Some aspects may inherently be unpredictable, he says, adding, "There's not an infinite reservoir of improvement out there."

But new models are beginning to reproduce realistic storm structures, although the intensity forecasts they yield are still off. As models start to simulate more processes with increasing levels of accuracy, "you're bound to move closer to your potential for intensity prediction." At the least, he says, "we'll be able to better quantify what the limit" of predictability is.