Predicting hurricanes and tropical storms isn't an exact science, as illustrated by the 2006 storm season. Or more precisely, predicting hurricane landfall isn't something done with certainty, and is a bit like gazing into a hazy crystal ball with bifocals (there were five hurricanes in 2006, but all stayed at sea). But one thing forecasting fortune-tellers have agreed on is that slow hurricane seasons will be the exception, not the rule, for the next several years.

Those predictions took a step towards fact last month, when Tropical Storm Andrea formed a full three weeks before the official start of the 2007 storm season. The event was a rare one for the month of May, according to the Atlantic Oceanographic and Meteorological Laboratory, because 97 percent of tropical activity takes place during the traditional tropical storm season, which starts every June 1st and continues until the end of November. Since 1851, only 18 tropical storms and four hurricanes have occurred in May, and none of them made landfall. Hurricane Alma was the earliest hurricane hit U.S. shores when it made landfall as a Category 1 hurricane on June 9, 1966.

Andrea's early start supports predictions made by Colorado State University hurricane experts Philip J. Klotzbach and William M. Gray. In April, they issued a revised prediction from their initial Dec. 2006 report, and increased the number of anticipated named storms from 14 to 17 named storms. And instead of seven hurricane-strength storms, they are predicting nine, with five storms expected to reach Category 3 strength or higher. The predictions are based on statistical data from the two months prior to the forecast date in order to give current conditions more weight over data compiled over the course of years and decades, a new technique adopted for this year's reports.