1. Cataclysmic cyclones 2. The nature of a storm 3. Tracking the storms 4. Dealing with data Errors in forecast tracks of Atlantic hurricanes. Straight lines show the trend toward greater accuracy. Graph: Hurricane Research Division Five-day map of Hurricane Ivan's projected path, made Sept. 14, 2004, shows New Orleans within the "cone of uncertainty" along the Gulf Coast. Map: NOAA Cloudy picture is clearing Hurricanes having such destructive habits, it's no surprise that emergency managers want better predictions. Since evacuations are slow, costly and disruptive, it's critical to know where the storm will be, and how strong it will be, a couple of days in the future. Key West, Fla., Aug. 12, 2004: A boarded-up storefront carries a message for Hurricane Charley. AP Photo/J. Pat Carter Predictions must start from data on present conditions, and scientists have some hair-raising technology for gathering data. They fly above storms and parachute miniature weather stations into the maelstrom to get precise data. They peer down from satellites to get the big picture. Then they crank up computers that would humiliate that box on your desk. Satellite data is always valuable over the oceans, where hurricanes form and travel, but where observers are uncommon. Satellite data can take many forms, from simple, visual-light images of the clouds to data from individual wavelengths of light. Data on water vapor are particularly useful, says Christopher Velden, a hurricane researcher at the Space Science and Engineering Center at the University of Wisconsin-Madison. "Only in the water vapor channel can you get information in clear, cloud-free areas." Temperatures at altitude, as seen in different wavelengths (click on image to start animation) Images: CIMSS/AMSU Instruments on the Geostationary Operational Environmental Satellites measure heat radiation released by water vapor in the upper atmosphere. By tracking movement, researchers calculate wind speed and direction. Thus the satellite can see invisible winds in the hurricane's environment that are crucial to steering the storm. A newer source of data, the NOAA-15 and NOAA-16 satellites, gives temperature readings from oxygen molecules in the air. Again, the advantage is more, better data to feed computer models of hurricane movement. "We send it to forecasters, so they visually have information on current behavior," says Velden. "It's very important to know how it is doing now, so that can be fed into prediction models. The better they are initialized, hopefully the better the prediction will be." Where will you be? Over the past decade or so, improved software, computers, satellites and aircraft have meant a steady improvement in track forecasts, but the improvement is incremental, not revolutionary. Today's three-day forecast is as accurate as a two-day forecast during the late 1980s. Average errors in the hurricane tracks are now: 24-hours: 80 miles 48-hours: 110 miles 72 hours: 230 miles Track record? Instead of offering one probable track, modern hurricane predictions are shaped like a funnel. A line down the center indicates the most likely track, but the cone shows the probability of a strike based on similar hurricanes in the past. Anybody within that cone, stresses Timothy Olander, a hurricane researcher at the Space Science and Engineering Center at the University of Wisconsin-Madison, could come face-to-face with nature's worst storm. The improved accuracy comes from better computer models and better knowledge of atmospheric conditions in and near the storm. Once the data are gathered, "We know pretty much what moves a hurricane," said Edward Zipser, a meteorologist now at the University of Utah, and who talked to us in 1998. "If you get accurate data to start, the models are far more capable of forecasting the flow patterns tomorrow or the next day." As Zipser indicated, predicting a hurricane's path is relatively straightforward, although it does involve massaging a gob of data. Generally speaking, tropical cyclones move in response to "steering winds" in their environment. The direction of these winds depends on the high and low pressure zones near the storm. Latitude also plays a role. Between about 10 and 30 degrees north latitude, steady trade winds blow toward the west, pushing Atlantic hurricanes from their origin near Africa toward the Caribbean and United States. North of 30 degrees, prevailing westerlies push the storms back toward the east. "Steering winds" move hurricanes. Trade winds blow hurricanes westward, but as the storms move north, they may get caught by westerly winds and cycle back toward the east. Map: USGS How wild will you be? Even if you can predict where a hurricane will go, how strong will it be when it gets there? Forecasting intensity is more problematic than path prediction, says Velden. The changes in intensity may result from changing ocean conditions, Velden explains. If the storm goes over a deep pool of warm water, it gets a surge of power. Atmospheric conditions such as wind shear (a change in wind speed and/or direction with altitude) also matter. "The atmospheric environment must be just right," says Velden. "A lack of vertical wind shear between the surface and upper-level winds allows it to convect without tilting over, and feed over the energy source." In those conditions, the hurricane's heat engine can be self-sustaining. If wind shear is too great, however, the storm can be blown apart. Wind-speed forecasts show slight improvements in accuracy. Graph: Hurricane Research Division At any rate, a run-of-the-mill hurricane can become a rip-roaring monster with little warning. This year, Velden says, "Charley jumped two categories in 24 hours," an increase in maximum winds of 40 to 60 miles per hour. "It was a contracting eye, much like a twirling skater who brings in the arms; it gets smaller, but the spin gets faster." Damage goes up exponentially as wind speeds increase, making a category 4 storm many times as destructive as a category 2 storm. Building damage, for example, increases from "moderate damage to houses" in category 2 to "extreme structural damage." Will you rain cats and dogs? Another problem is predicting rain intensity. "What's the difference between a storm that drops five inches -- and one that drops 20?" Zipser asked. "It's not related to the rates of wind." In fact, he notes, tropical storm Charlie -- which was too sluggish to qualify as a hurricane -- dropped 15 to 20 inches of rain on Del Rio, Texas in 1997. Better precip predictions call for more understanding of the physics of convective systems in thunderstorms and hurricanes, Zipser noted. Precisely modeling the condensation of water particles into bazillions of rain droplets is not easy, Zipser said. "We have more to learn in terms of fundamental understanding of what controls the intensity." What's new in hurricane data? There are 1 2 3 4 pages in this feature. Bibliography | Credits | Feedback | Search ©2004, University of Wisconsin, Board of Regents.