Hurricane Ike approaching the Texas coast, Sept. 12, 2008. (Credit: NOAA Environmental Visualization Laboratory)
Hurricane Ike approaching the Texas coast, Sept. 12, 2008. (Credit: NOAA Environmental Visualization Laboratory)

Future of Hurricane Forecasting Looking Up

UA researchers, like Tom Galarneau and Xubin Zeng, are helping improve the ability to predict hurricanes.
July 5, 2016

Like so many hurricanes before her, Gloria was conceived as a tropical disturbance off the west coast of Africa. She spun westward toward the Bahamas, all the while feeding off the Atlantic Ocean's warm surface temperatures, then arced northwest. Days later, on Sept. 27, 1985, she slammed into North Carolina's Outer Banks. Ten hours later she slammed into western Long Island, NY.

Before she spun out to sea, Gloria triggered floods, eroded beaches, destroyed homes, felled trees, killed more than a dozen and knocked out power to millions.

The 1985 Atlantic hurricane season gave rise to 11 tropical cyclones, seven of which became hurricanes – large swirling storms that generate winds of 74 miles per hour or more and form over warm ocean surface waters – according to the National Hurricane Center's 1985 annual summary. The summary noted that six hurricanes and two topical storms made landfall that year. Every U.S. coastal state was placed under a hurricane warning at one time or another that season.

During that hurricane season, Tom Galarneau, was living with his parents in Albany, NY, when Gloria made her appearance there. Galarneau was six years old.

"When I was a little kid, Hurricane Gloria came up the coast in 1985," recalls Galarneau, now an assistant professor in the University of Arizona's Department of Hydrology and Atmospheric Sciences. "We didn't get the brunt of the hurricane, but we got the wind and the rain, and it was dark, and my teacher was repeatedly yelling at me for not paying attention. That was my first memory of being really interested in the weather."

Now, thanks to scientific and technological improvements, like better computational models and new and improved satellites, frequency and characteristics of hurricanes are becoming more predictable both seasonally and in the short term.

At the UA, Galarneau studies large-scale weather patterns that favor hurricane formation within a one- to two-week period. In contrast, his colleague Xubin Zeng, the Agnese N. Haury Chair in Environment and a UA professor of atmospheric sciences, studies seasonal hurricane forecasting; that is, predicting the number of hurricanes that are expected to form during the Atlantic hurricane season, which runs from June 1 to Nov. 30. 

Last year, Zeng and Kyle Davis, then a master's student in the UA Department of Atmospheric Sciences, and Elizabeth A. Ritchie, a UA atmospheric sciences professor, came up with a better model for predicting the number of hurricanes expected in an upcoming season.

Using data from the 1950 to 2013 hurricane seasons, the researchers tested their new model by hindcasting the number of hurricanes that occurred each season from 1900 to 1949. The model improves the accuracy of seasonal forecasting for the North Atlantic and Gulf of Mexico by 23 percent.

"You need to look at the historical data to assess today's risk," says Zeng.

For the 2016 season, the researchers' model predicts eight hurricanes in the North Atlantic, including the Gulf of Mexico and Caribbean Sea, with a range of six to 11. Six is the historical average for Atlantic hurricanes since 1950.

Unlike last year, El Niño played no role in this year's hurricane prediction, but ocean surface temperatures, as always, did. Surface temperatures over the tropical Atlantic are running warmer than average this year, says Zeng, and that favors hurricane activity.

So does weaker than average wind speed over the Atlantic. Weaker winds limit the mixing of the cold, deep waters with the relatively warm, shallower ocean waters. This in turn leads to warmer surface temperatures boosting the likelihood of hurricane activity.

In fact, the latest data point to a more active hurricane season than last.

As for Galarneau, that's something that concerns him when it comes to planning and public safety, but doesn't displease him when it comes to his research, medium-range hurricane forecasting.

Like seasonal hurricane forecasting, Galarneau's medium-range forecasting relies heavily on computational models. And like Zeng, Galarneau is on the lookout for warm surface ocean temperatures, about 80 degrees Fahrenheit, very high relative humidity, and very low wind shear over the Atlantic Ocean.

"In the Atlantic, weak cyclonic circulation systems move out from Africa into the Eastern part of Atlantic," says Galarneau. "When these systems move over really high sea surface temperatures, thunderstorms begin to form within the area of circulation, and then the thunderstorms themselves can spin up the circulation to form a hurricane."

One of the challenges of forecasting hurricanes in the short-term is getting a good read on the moisture levels in and around those thunderstorms, says Galarneau. Galarneau has a website that assesses the accuracy of the United States global computer model's forecast.

"The site looks at the forecasts for the past 60 days and assesses how well the model is doing, and if there are problems, where are the problems are coming from," says Galarneau.

And if there are problems, Galarneau says it's usually related to how well the model is measuring atmospheric moisture.

"Moisture is the fuel for thunderstorms," says Galarneau. "And the problem with moisture in the tropics is that it's hard to measure. A lot of times in the tropics, you get these blobs of moisture that are formed over very warm oceans, particularly in the western pacific. They call it the western pacific warm pool. Thunderstorms feed off this moisture, but one of the issues with our numerical model is that it takes all this moisture and quickly rains it out in the first day of its forecast."

In other words, the model sees the atmosphere as being too dry, when, in fact, it's full of moisture.

But Galarneau says new weather satellites will soon come online, one before the end of this year, and will provide new data, including more moisture observations, which will help forecasters get a better and faster idea of what the atmosphere looks like.

With that development in mind, Galarneau is beginning to focus on the development of hurricanes in the southwest Caribbean Sea, a haven for high humidity.

"I'm looking at that area because it's relatively understudied," says Galarneau. "And 90 percent of storms in that area make landfall. But it takes only one storm to make landfall to cause catastrophic damage."