Aug. 23, 2010: An image released by NOAA made from the GEOS East satellite shows Hurricane Irene as it passes over Puerto Rico and the Dominican Republic.AP/NOAA
The outlined area in this satellite photo indicate the current position of systems being monitored by the National Hurricane System. Color indicates the probability that the formation will become a tropical cyclone within 48 hours.National Hurricane Center
Here's a forecast you might not expect.
Two Colorado State University climatologists, who have independently been tracking and predicting the severity of hurricanes for nearly 30 years, are abandoning their long-range forecasting efforts.
Simply put, the advance predictions just weren't accurate.
"We have suspended issuing quantitative forecasts at this extended-range lead time, since they have not proved skillful over the last 20 years," Philip J. Klotzbach and William M. Gray wrote in their annual December report intended to predict the severity of the upcoming year's Atlantic hurricane season.
Klotzbach and Gray will continue forecasting -- just not in December, months in advance of the hurricane season.
"From a computation perspective, there's enough chaos in weather that I don't know if you'll ever be able to predict it months in advance," Klotzbach told FoxNews.com.
The National Oceanic and Atmospheric Administration (NOAA), Colorado State and others traditionally issue hurricane predictions in late May just prior to the start of the season, and update them in early August, just prior to the historic peak. Twenty years ago, Klotzbach decided to push the envelope, to see how early he could predict the season.
The results weren't good.
In December 2010, the duo predicted 180 total days of tropical cyclone activity; the season actually saw 137. The numbers for 2009 were off as well: 136 days of cyclone activity were predicted, but only 66 cropped up.
"Beyond five days or so, that's where the crystal ball gets a little fuzzy," Chris Vaccaro, a spokesman for NOAA, told FoxNews.com.
"You can predict seasonal trends fairly well months in advance. But the devil's in the details. Which ones will stay over open water and which will hit land?"
And that type of more immediate forecasting has actually improved dramatically, Vaccaro and Klotzbach both said.
"Hurricane forecasting has come a long way," Vaccaro said. "Once a storm does form, our forecast skills are quite high in being able to predict where that storm will hit."
The challenge lies in understanding the forces at work, he said. The long-term seasonal outlooks are based on large-scale, slow-moving climate factors -- things like ocean temperatures, the presence of a La Nina or El Nino force, weather patterns, and so on. But the birth and eventual path of an individual storm is dependent on short-term weather patterns -- the position of the jet stream or the presence of a cold front.
"Those day-to-day weather patterns are very fluid and have a tremendous impact on the strength and track of a specific storm," Vaccaro told FoxNews.com.
Colorado State's hurricane experts tell a similar story. Despite advances in supercomputers, hurricane-hunter aircraft, satellites, weather buoys and more, long-term prediction simply isn't feasible today. But short-term forecasts get better every year.
"Our April forecasts have shown reasonably good skills and are getting better over the years," Klotzbach said. "When you're forecasting the weather a week from now versus tomorrow, you're going to put a lot more faith in tomorrow's forecast."
"Our knowledge of how storms work still isn't perfect," Klotzbach admitted.
Jeremy A. Kaplan is Science and Technology editor at FoxNews.com, where he heads up coverage of gadgets, the online world, space travel, nature, the environment, and more. Prior to joining Fox, he was executive editor of PC Magazine, co-host of the Fastest Geek competition, and a founding editor of GoodCleanTech.