< Back to Articles

Digging up dirt to study storms of the past

Nature Network Boston

June 13, 2007

Late last month, just before the start of the hurricane season, the National Oceanic and Atmospheric Administration predicted as many as 17 “named” Atlantic storms this year (the average is 11). The news raised a perennial question: is global warming to blame for the growing number of hurricanes hitting the United States?

Many researchers turn to climate models for an answer, but Jeff Donnelly, a geophysicist at Woods Hole Oceanographic Institution on Cape Cod, takes a different approach. He’s a specialist in the new field of paleotempestology, which uses geologic data such as sediment cores to figure out what factors, such as sea surface temperatures, have affected hurricane development over the past 5,000 years.

Scientists have been collecting instrument data on hurricanes only during the past few decades, so geologic data like Donnelly’s that reach further back in time can strengthen or modify theorized relationships between climatic variables and hurricane activity.

Donnelly had a paper published in Nature last month, which suggested that fluctuations in sea surface temperatures aren’t the only—or even the most important—contributor to hurricane activity, at least at Donnelly’s Caribbean research site. He says that other variables, such as El Niño—a periodic oscillation in tropical air currents that alters weather around the world—affect storm development and that scientists need to consider how climate change will affect those factors as well when predicting future hurricane frequency.

“This work is very important,” says James Elsner, a hurricane modeler at Florida State University. “There’s only limited data now, but when there’s enough research to be able to piece together what made certain times more active than others, that will give us a good sense of what might happen in the future.”

Buried treasure
Donnelly’s paper also illustrates his unique approach. He and his team have traveled around the world, from the kettle ponds of Cape Cod to French Polynesia, to look for signs of past hurricanes in the sediment. His technique is based on the fact that, during a hurricane, monster waves bring sand and other debris from beaches into coastal lagoons, marshes, and ponds.

In these areas, Donnelly drives dozens of long, clear, plastic tubes, measuring five centimeters or more in diameter, down several meters into the muck to extract material deposited by past storms. By radiocarbon dating seeds or bits of wood found at different layers in the sediment cores, he can gauge the time and frequency of previous hurricanes. He then compares this record to cores that document other historical variables, such as sea surface temperature, to figure out which factors are most strongly associated with hurricane activity over thousands of years.

Ocean to ocean
Donnelly is currently studying sediment cores from Japan, the Yucatan peninsula in Mexico, and Grenada. “We really want to map out the activity not just across the western North Atlantic, but also in the Pacific and other places,” says Donnelly.

With a recent grant from the U.S. Department of Energy, Donnelly hopes to take the next step in his work: assessing the impact of these historical storms on the ecology of the coastline. He’ll compare the hurricane record from various New England ponds to sediment cores containing pollen and charcoal to see how active hurricane eras changed the vegetative landscape and frequency of fires.

“It’s fine to say, ‘Things are going to change in the future.’ We really have to say, ‘Who cares? What does that mean in terms of ecosystems and land forms where people are building their houses?’” says Donnelly. “That’s the future of paleotempestology.”

< Back to Articles