TRACKING HURRICANES

October 2003
		In recent years, scientists have made leaps and bounds in improving our ability to predict and track hurricanes, mitigating their destructive power. Two scientists, Dr. Naomi Surgi of the National Hurricane Center and Dr. William Gray of Colorado State University's Tropical Meteorology Project, answer your questions about the latest developments in predicting and tracking the devastating storms.

Online NewsHour Special Report: Tracking Hurricanes Forum Introduction How does your field approach view the work being done in chaos theory? Is lack of temperature soundings in the sea the main problem in track prediction? What are the flights through the hurricane like for the observers on board? Are you concerned that such warming might cause not continued warming but a total irreversible change in realities? I was wondering what role Linux and computer clustering have in NOAA's hurricane modeling efforts? Are there any hurricane models that treat a storm as an electrical model where temperature is represented as resistance?	Nick Ketter of Atlanta, Ga., asks: I've been learning about chaos theory lately and the consequences of the theory seem to point to the impossibility of predicting such things as weather in the long term. How does your field approach view the work being done in chaos theory? Dr. Naomi Surgi responds: Chaos theory as you know it applies to long-range computer model forecasts, which are not yet useful for long-range hurricane predictions. Yet, very good seasonal hurricane forecasts are still being made. Here is why. Seasonal hurricane forecasts mainly require an accurate prediction of the prevailing atmospheric conditions during the peak August to October months of the season. It turns out that two dominant climate factors tend to control the seasonal mean circulation during this period: El Niņo/ Southern Oscillation (ENSO) and the tropical multi-decadal signal. Both of these climate factors are linked to slowly varying fluctuations in tropical convection. The seasonal hurricane forecast, therefore, relies heavily on predicting these two climate phenomena. To make the ENSO predictions we use a series of sea-surface temperature forecast models that capture the slowly evolving conditions associated with El Niņo and La Niņa (or ENSO). The prediction of the multi-decadal signal depends mostly on correctly predicting its phase. We know its phase from monitoring its associated atmospheric circulation and oceanic features. Then, because its phase evolves on decadal time scales, forward extrapolation of three to nine months is easily done. Armed with this information the essence of the seasonal outlooks then boils down to predicting the likely atmospheric conditions and activity associated with the combination of ENSO and the decadal signal, which we can ascertain from a variety of statistical analyses based on historical data. Dr. William Gray responds: The atmospheric-ocean is only partly chaotic. There are long-term memory signals to the system that can be used. Chaos theory has little meaning to forecasting real work with weather and climate changes.