Because hurricane dynamics are completely different than those of the typical storm, forecasting their next moves is quite an art. Frontal cyclones are embedded in a stream of air that determines their path and influences their intensity. That stream is absent in the tropics. Also, because tropical systems form in areas of weak temperature contrast, the standard forecast procedures go out the window. Additionally, heat as an element is poorly understood in the scheme of weather equations. Because hurricanes depend so heavily on that heat release, their behavior is often mysterious.
So we may have a storm that has no apparent push and often spins like a top with no particular clear-cut path—an enormous heat engine that's just not well understood. As hurricanes move into northern latitudes they become more predictable, but that tremendous heat release persists and is always ready to throw us some curve balls.
One technique of forecasting simply involves persistence. Whatever path and trend the storm may be on is simply projected to continue for the next 12 to 24 hours. You've probably heard hurricane forecasts with the words, "Little change in path or intensity is expected during the next 24 hours."
Sometimes forecasters rely on climatology. What is the normal path that these storms take? Both persistence and climatology work well when the storm is in an area of a weak jet stream—like the tropics. But as the storm heads northward, the path can change erratically and unpredictably. Numerous computer-forecast methods have been developed using both dynamics and statistics. New computer models deliver a number of solutions or ensembles. The consensus or average solution often delivers the most accurate prediction.
Still one of the most reliable instruments for hurricane tracking is a simple barometer. A hurricane's path is nicely outlined by the pressure drop that occurs ahead of the center. These pressure drops can be mapped out. Where the falls are greatest is where the hurricane is heading. Unfortunately this technique gives only a short-term warning—maybe six hours.
In a hurricane, the strongest winds are found within 50 miles of the center—that's where the most damage is likely to occur. Our ability to predict the position of a hurricane within 50 miles in a 24-hour period is not very great. There is less than a 50-50 chance that such a forecast will be accurate. So, because of the need for allowing enough time to prepare and perform evacuations, hurricane conditions are predicted far more often than they materialize.
Old weather watchers use a one-two-three rule to estimate the potential error in forecasting the position of a hurricane eye. A 24-hour prediction has an average error of about 100 miles. A 48-hour prediction is off by an average of 200 miles. And a 72-hour forecast is wide of the mark by about 300 miles.