1. Cataclysmic cyclones 2. The nature of a storm 3. Tracking the storms 4. Dealing with data Sept. 13, 2004: Ivan is a huge, deadly storm, as it ducks between Cuba and the Yucatan Peninsula. Photo: NOAA Which data mata? If you like data, you have to love meteorology. Satellites download deluges of data, but is more data better? Usually, but quantity can raise problems, says Tim Olander of the University of Wisconsin-Madison. With each technological advance in meteorology, new sources of data have threatened to inundate forecast models. The problem arose with the first weather balloons, the first weather satellites, and looms again with the growing number of instruments perched on modern satellites. The GOES-8 satellite, for example, reads visible light and 18 other wavelengths. But a new series of weather satellites, Olander says, will send "hyperspectral" data on thousands of wavelengths. And that will raise thousands of questions. Before any wavelength is fed into a computer model, should it be averaged, deleted, combined or subtracted, or used as is? Those questions will be tackled with an ancient tactic, he says. "We'll do what we've already done. We'll use trial and error to find out what works best. Using all that data is extremely difficult, it's a very important problem to solve." SLOSH is a computer for estimating storm surge heights and winds. Here's a simulation in the New Orleans area. Movie: NOAA Most improvements will be incremental rather than revolutionary, however. No way does Olander expect to ever predict that a particular tropical depression near West Africa will some day hit Miami Beach rather than Palm beach. Weather is way too chaotic and complicated for that. Yet steady improvements in accuracy have lead the National Hurricane Center to begin issuing five-day forecasts, and those, notes Velden, accurately predicted landfall in Florida for both Charley and Frances. "This is remarkable since just two years ago, the NHC did not even predict out five days due to the uncertainty! Can we make a precise forecast out four or five days? No. But the ability to put out a credible alert to folks and industry that much in advance is a testament to recent gains in research and operational hurricane prediction." Fort Pierce, FL, September 9, 2004 -- Hurricane Frances toppled this gas-station canopy. Photo: Mark Wolfe FEMA Still, the 2004 hurricane season also provides some humbling examples, especially with a slow-moving storm like Ivan. "Look at Jamaica," says Olander. "Ivan was six hours away from the eye going over. Why it veered away we will never know. Everybody was prepared for the worst, and they were very fortunate. But the next storm may come along and hit Jamaica -- hurricanes are very unpredictable." Raise a storm in our hurricane bibliography. There are 1 2 3 4 pages in this feature. Bibliography | Credits | Feedback | Search ©2004, University of Wisconsin, Board of Regents.