Hurricane forecasting: some progress, but still a gap

Dennis, the off-again, on-again tropical storm, illustrates the major concern of hurricane experts -- rapidly growing vulnerability of many US East Coast and Gulf Coast areas to hurricanes.

The strength and, at times, the course of Dennis have proved difficult to predict accurately. Had it gained hurricane intensity, forecasters say it would have been impossible to have given timely warning to some of the coastal areas at risk. Many of these areas are hard to evacuate. Getting people out often takes 18 to 24 hours. But forecasters can usually offer only 8 to 12 hours' warning.

Yet these hard-to-evacuate areas, such as barrier islands, are being residentially developed at a rapid rate.

At this writing, Dennis was near the southwestern tip of Florida with substantially less than hurricane strength. Whether it would intensify and menace other coastal areas if it moves out over the ocean was unknown.

Uncertain of Dennis's development, forecasters at the National Hurricane Center (NHC) in Coral Gables, Fla., had issued a hurricane "watch" for a broad area. (A "watch" is an advisory that a hurricane may sweep over an area within 24 to 36 hour.) It is less precise than a "warning," which predicts arrival of a hurricane within 8 to 12 hours. Nevertheless, even a "watch" advisory can lead to costly individual and community preparations that may prove unnecessary and that risk creating a "cry wolf" attitude toward hurricane warnings.

Aware of this risk and alarmed by the continuing development of vulnerable areas, hurricane experts in recent years have stressed both their limitations as forecasters and the need for advanced community preparations. These preparations include having detailed evacuation plans where needed and curbing development of areas where evacuation would be difficult.

This has led James P. Walsh, deputy administrator of the US National Oceanic and Atmospheric Administration (NOAA) to warn: "The nation is in the most vulnerable position in history should a major hurrican strike this year. Mushrooming coastal populations, public inexperience and apathy, and limited evacuation routes combine to create catastrophe."

From the mid-1950s to about 1970, hurricane forecasting improved substantially as hunter aircraft and weather satellites gave early warning and meteorologists' knowledge of the storms grew. But, as Charles J. Neumann and Joseph M. Pelissier of NHC document in the June issue of the Monthly Weather Review, there has been virtually no change in forecasting skill for a decade.

There are some signs of limited progress. Intense monitoring of storms that come ashore combined with computer models of specific locations are giving better forecasts flooding.

Also, forecasters now have a way to anticipate the rainfall potential of a storm. Generally, the deeper the convective clouds in a storm, the more rain they produce. Also, the deeper the cloud, the higher and colder its top will be. William Woodley at the University of Wisconsin and Cecilia Griffin of the NOAA have developed a technique for analyzing satellite data to asses clouds for rainfall. As refined and computerized by other meteorologists, this technique now gives forecasters quick, reliable estimates of a hurricane's rainfall potential.

Yet another aid is a new "3-D" technique for using photos from two satellites as stereo pairs.

Developments such as these suggest that forecasting skill may begin to move off its plateau in this decade. However, that cannot be guaranteed. Thus, NOAA says: "Americans in hurricane-susceptible areas must plan for their personal safety. They can't wait for the satellite photos. They have to prepare immediately -- before the need becomes desperate."