The Inside Track on Hurricane Season
By Michelle Nery

Having survived two highly active hurricane seasons, Hampton Roads residents – indeed, everyone who lives along the Atlantic Coast – are keeping a wary eye on the 2005 season, which started June 1 and continues through Nov. 30. Robert Tuleya, an adjunct professor in the Center for Coastal Physical Oceanography at Old Dominion University, is doing more than that, however. He’s using lessons learned from previous blustery seasons and a new hurricane prediction model to improve the accuracy of future prognostications.

Tuleya joined the university more than two years ago after retiring from the National Oceanic and Atmospheric Administration, where he worked for 31 years. At NOAA, he was based in the Geophysical Fluid Dynamics Laboratory (GFDL) creating one of the world’s most trusted hurricane prediction models. He still works with NOAA and the National Weather Service on projects to improve modeling and forecasting, including the creation of a new hurricane forecasting system to be implemented in 2007.

From his office in Old Dominion’s Crittenton Hall, Tuleya studied GFDL models on his computer screen of the 15 tropical storms that formed in the Atlantic Ocean including the nine hurricanes – among them Charley, Frances, Ivan and Jeanne – which ravaged the East Coast in 2004.

Predicting with Accuracy
For nearly two decades, Tuleya has worked with his colleagues to develop computer models to better understand the basic mechanisms that produce hurricanes and to predict the movement and intensity of individual storms. “This is not an exact science to say the least,” he admits. “But there have been incredible improvements since I started working in the 1970s.”

In 1995, while working for NOAA, Tuleya, along with GFDL colleagues Yoshio Kurihara and Morris Bender, developed the GFDL hurricane model that continued to be one of the best prediction systems for the National Hurricane Center in both the Atlantic and Eastern Pacific Basins for the 2004 season.

Based on its impressive performance in tracking Hurricane Emily in 1993, the system was adopted as the operational hurricane prediction model by the National Meteorological Center. While conventional models predicted that Emily would make landfall in the vicinity of Georgia, the GFDL forecast system correctly forewarned that the hurricane would strike North Carolina’s Outer Banks before veering back out to sea.

More than 300 forecasts were made for the Atlantic during the 2004 season, and the GFDL model proved to be either the best or next best in providing track guidance for forecasts from one to five days. In addition, more than 175 forecasts were made for the 2004 season in the Eastern Pacific, and the GFDL model proved to be the best track forecast model for periods from one to four days.

“As far as forecast intensity, there were signs of improvement using the GFDL model during the 2004 season,” Tuleya said. “For most forecast periods, the GFDL model was competitive with statistical forecast methods. Over the years, statistical methods have been superior to dynamic models such as the GFDL model in forecasting hurricane intensity.”

Tuleya expects that a new high-resolution version of the GFDL model will be completed soon. The model will be able to simulate storm scale more realistically. “The resolution near the storm within the multiple grid system will improve to nine kilometers from the current 18 kilometers horizontal resolution.” Test experiments conducted last fall indicated “improved track and intensity forecasts for the 2005 season.”

Combining Research and Forecasting
He is currently developing the Hurricane Weather Research and Forecasting (HWRF) system with the National Weather Service, the National Center for Atmospheric Research, and several NOAA laboratories and universities, including Old Dominion. “The whole idea is to combine research and forecasting together,” he says. “It is popular now to transition research from universities and research institutes to operational centers to make it more fruitful.”

Tuleya ran the prototype HWRF model during the 2004 season and completed more than 120 forecasts. He has been evaluating the results and plans to use them to improve the high-resolution model as well as to develop a movable “nested” grid system to zero in on the internal storm structure this year. The current plan is to have the HWRF operational by the 2007 season.

In addition to using models to predict present-day storms, Tuleya is involved in determining the impact of global warming on future hurricane activity. He has worked with GFDL climatologist Thomas Knutson for several years. “We merged two models, a climate model with an operational hurricane model, to simulate hurricanes at very high resolution,” Tuleya says. Most of his current research in this area has focused on the increasing intensity of storms. “There has been a lot of evidence from modelers like us and theorists that there will be higher-intensity storms. Our models show a 5 percent to 10 percent increase,” he explains. “It can make a difference in the destructive nature of hurricanes. With the number of people living on coastlines, it is an important issue.”

Tuleya’s research indicates that most intense storms are moving half a category higher, or approximately 8 mph. “The rainfall intensity associated with hurricanes is also increasing,” he says. “There is a bit more wind intensity, which is showing a 5 percent to 10 percent increase, while rain is showing a 20 percent increase.” Tuleya and Knutson published a paper in the September 2004 Journal of Climate on the impact of global warming on hurricane intensity and precipitation, which was quoted widely in the media, from The New York Times to the International Herald Tribune. On the other hand, this study is not without controversy. Given the increased hurricane activity this past season, it is natural to ask whether the occurrence of strong storms could be due to global warming. There is disagreement among experts.

“From our standpoint, the small .9 degree Fahrenheit warming observed in the Atlantic since 1900 implies only a 2-3 miles per hour intensity increase to date. Such a small increase is hard to detect. It is difficult to attribute the upswing in strong hurricane activity this past season to global warming. Season-to-season variability is very large.”

Interestingly, Tuleya himself once learned the consequences of not paying attention to the results of his computer modeling. He and his wife were in their hometown of Princeton, N.J., as Isabel approached Hampton Roads in 2003. His wife warned him that they should head south to prepare their home in Portsmouth, but Tuleya told her not to worry about the storm. “I’m kind of embarrassed because an 80-foot tree fell on my house,” he admits.