Rainfall Data Influence 1992 Hurricane Forecast
THIS year's Atlantic hurricane season ended Nov. 30, less than two weeks ago. Yet - are you ready for it? - we already have a forecast for the season that starts next June. Hurricane activity in 1992 should be well below the long-term average.In an era when long-range weather forecasts usually look only three months ahead at most, William M. Gray believes he can "make surprisingly skillful forecasts of Atlantic tropical cyclone activity" over half a year in advance. That's a long way for any forecaster to stick his neck out. However, Dr. Gray - professor of atmospheric science at Colorado State University in Fort Collins, Colo., has unusual credibility in this field. His hurricane season forecasts made in late May and updated in late July have been essentially on target in all but one of the years (1989) since he began making them in 1984. This reflects the remarkable fact that certain persistent ocean and atmospheric conditions strongly influence Atlantic tropical storm activity even though they have no comparable influence in the Pacific or Indian Oceans. For example, that activity is depressed in the presence of El Nino - a 12-to-18-month anomalous sea surface warming in the eastern equatorial Pacific. Other factors include departures from average of sea-level air pressure over the Caribbean and Gulf of Mexico region and stratospheric wind patterns. Gray plugs these factors into a forecast equation that predicts the number and nature of Atlantic tropical storms. That equation missed the forecast for 1989, seriously underestimating the intensity of that hurricane season. This alerted Gray and his colleagues to a factor they had overlooked - West African rainfall. When Africa's Western Sahel region broke its long-term drought in 1989, tropical storm activity picked up. Gray's method regained its predictive skill by including West African rainfall. Now ongoing research has convinced Gray and his colleagues that these rainfall data have very long-range predictive power. Gray's scheme also takes account of stratospheric winds that circle the globe above the equator and reverse direction every 12 to 16 months. He says these factors "can be used to forecast between 44 and 51 percent of season-to-season variability ... of Atlantic seasonal tropical cyclone activity as early as late November of the previous year. Gray anticipates a mild 1992 hurricane season. He notes that drought has returned to the Sahel. This, he says, should help "keep the lid on next year intense hurricane activity." He forecasts eight tropical storms strong enough to receive names, four of which should be hurricanes. He expects one hurricane to be "intense," meaning it reaches sustained low-level winds of at least 96 knots (178 kilometers or 111 miles an hour). Like his regular forecasts, this doesn't say where or when such storms may strike. Gray also warns readers that "this seasonal forecast is a statistical one, which may fail in some years." So take it with at least a small grain of salt. We have only to wait a year to see if he's right.