Digging into Hurricanes


21 SEPTEMBER 1938 - With little warning from forecasters, the Great New England Hurricane raced toward land at 60 miles per hour (the fastest forward speed of any hurricane in regional history), cutting a path of destruction across Long Island and deep into New England. Sustained winds exceeded 121 mph, and a gust of 186 mph was recorded in Milton, Massachusetts, both records for New England. The storm surge reached 20 feet in Narragansett Bay, and wave heights reached 50 feet off Gloucester. By the time it blew away, New England’s last “intense” hurricane—defined as category 3 or higher on the Saffir-Simpson scale—took the lives of more than 700 people and destroyed more than 8,900 homes and buildings. Property damage was estimated at $400 million in Depression-era dollars.

Most hurricane researchers look at the atmosphere to study their subjects. Some look in the history archives. Jeff Donnelly looks in the mud.

In Succotash Marsh, Rhode Island, on Whale Beach in New Jersey, in Long Beach, New York, Donnelly unearths the history of intense hurricanes in the Northeast. By digging up sediments from back bays and coastal marshes, he can reveal the past and perhaps the future of hurricanes. As a graduate student at Yale and Brown in the 1990s, Donnelly took sediment cores from salt marshes, searching for evidence of past sea-level changes. He sank metal tubes into marshes and pulled up twenty-foot cylinders of mud—much like using suction and a well-placed fingertip to keep a milkshake suspended in a straw. Layered within the cores were sediments containing pollen grains, plant fragments, and sand layers that helped spell out what had happened in the ancient coastal environment.

“As a kid, I was always into dirt,” says Donnelly, an Assistant Scientist in WHOI’s Geology and Geophysics Department. “But I didn’t just want to get dirty. I wanted to figure out the riddles that the Earth could provide the answers to. Earth has a wonderful story to tell if you know how to read its pages.”

Donnelly grew up in southeastern Massachusetts with a rock collection and a love for science and history. “When I was trying to decide what kind of scientist I was going to be, I decided that I wanted my work to be useful and relevant to people,” he notes. “In my view, the coastal zone is ground zero for many of the environmental challenges society is facing.”

His interest in dirt and history became an asset when Donnelly’s advisor at Brown, geologist Tom Webb, got involved with a hurricane research project in 1996. Colleagues working on the Risk Prediction Initiative (RPI), based out of the Bermuda Biological Station for Research, asked Webb if he could find evidence of hurricanes in the records of vegetation changes he was collecting from New England lakes. Webb didn’t see such a signal as he identified and counted pollen grains preserved in the mud of the lakes, but he had a graduate student who was willing to look for one near the coast. Donnelly was suddenly a hurricane researcher.

Webb and Donnelly were funded by RPI to help prove that you could find hurricanes in the coastal geologic record. The fieldwork began at Succotash Marsh (near Kingston, RI) because it sits behind a barrier beach that protects it from constant changes along the surf line. Sifting through the sediments, Donnelly found layers of sand mixed between peat. That sand, he surmised, must have been washed into the marsh by great storms. Essentially, a hurricane or Nor'easter would have had to wash sand from the shoreline all the way into the back bay. And it would have to be a pretty potent storm to move sand that far, over dunes and other coastal formations.

Donnelly used his geologic dating tools—pollen grains, industrial metal residues, carbon 14-to estimate when those sandy sediments were laid down. He then compared his records with aerial photos of the “overwash” fans spread by known historical hurricanes. The sand layers were reliable evidence of the biggest storms. Taking samples from several other sites along the New England coast, he found he had a record of major hurricane activity stretching back 600 to 1,000 years, well beyond the meteorological data. Such a long record would allow scientists to estimate the long-term return rates of intense, land-falling hurricanes.

“I was surprised to see that the concept worked,” Donnelly recalls. “We really could constrain the dates of hurricanes pretty well.” The research was modeled after a study by Kam-biu Liu of Louisiana State University, who had found evidence of ancient Gulf Coast hurricanes in coastal lake deposits in Alabama and Florida.

3 SEPTEMBER 1821 - Perhaps the most destructive hurricane in the history of New York City raised a ten-foot surge of water in just one hour onto the Battery in lower Manhattan—the highest surge ever recorded in New York Harbor. The storm had moved from North Carolina to New York in just ten hours, and hurricane winds were experienced as far west as Philadelphia. Though reliable meteorological records are scarce, eyewitness accounts and historic reconstructions suggest the storm made landfall near Jamaica Bay as a category 3 hurricane. It was the only time in recorded history that New York took a direct hit from an intense hurricane.

After Miami and New Orleans, New York City is considered the most troubling target for a major hurricane disaster. Recent research using theoretical models suggests that a modern category 4 hurricane would drown John F. Kennedy International Airport under 20 feet of water and would flood the Holland and Brooklyn-Battery tunnels and the subways.

That is frightening news to the insurance industry, which has liability for more hurricane-endangered property in New York than in any state besides Florida. After getting hammered by Hurricane Andrew in 1992, insurers decided that they needed to know more about the long-term return rates of major hurricanes. The storm database at the National Oceanic and Atmospheric Administration only goes back to the mid 1800s, so no one really knows whether the 1821 and 1938 hurricanes are hundred-year events or thousand-year events at those locations.

“We need to know how frequent these major hurricanes may be for societal reasons,” says Chris Landsea, an atmospheric scientist in NOAA's Hurricane Research Division. “For instance, when you develop building codes, you need to decide whether you have to prepare for 100 mph winds every 100 years or every 10 years.” “But until we get a few big ones, I am not sure the public and policymakers are going to take notice,” Landsea adds. “It is probably going to take a few $10 billion hurricanes to get their attention.”

While the science policymakers may have other research priorities and interests, insurers have a vested economic interest. So for the past decade, Donnelly and other researchers looking at hurricane history have secured funding from insurance underwriters, by way of the Risk Prediction Initiative.

“At an RPI meeting in Bermuda geologists were giving the long-term perspective on hurricanes over thousands of years,” Webb recalls. “Then Jeff got up and gave a picture of the past 700 years. The insurance folks stood up and said ‘That’s what we need to know. That’s much closer to our time frame.’”

25 AUGUST 1635 - Barely a decade after the European settlement of New England, the Great Colonial Hurricane cuts straight across Buzzards Bay and Cape Cod Bay and pounds southeastern Massachusetts with a storm surge over 20 feet. The leaders of the two English colonies—John Winthrop of Massachusetts Bay and William Bradford of Plymouth-made the first historical record of an intense hurricane striking New England, a storm that nearly wiped out early settlements. Given the size of the storm surge and the tales of extensive destruction of forests, modern researchers believe the hurricane was at least as powerful as the 1938 storm.

Though hurricanes in New England don’t have the notoriety that they do in the Florida Keys, the Carolina Coast, or the Gulf States, the region has taken its share of hurricane hits. Intense hurricanes—category 3 or higher—are relatively rare in New England because the cooler sea surface temperatures and prevailing winds weaken the storms or bend them eastward. But when intense hurricanes do strike New England, the wind speed on the eastern side of the storm is often accentuated by the fast forward motion of the whole system, creating a greater storm surge. Because Long Island and New England jut out into the western Atlantic, they are a mark for fast-moving tropical storms tracking north. Donnelly’s research suggests that a fast-moving category 3 hurricane hits about every ninety to one hundred years.

Donnelly would like to know more about this history and whether there are larger patterns at work in the cycle of hurricanes. He is slowly working up and down the East Coast, “trying to get samples from Virginia to Cape Cod in every suitable marsh” and combining his hurricane research with his interest in long-term climate change. One of the scenarios of current global warming projections is that hurricanes might become more frequent and more intense. Donnelly is trying to figure out if that is historically accurate. So far his data suggests little change in hurricane frequency over the past 700 years, a period when the global climate has cooled with a “Little Ice Age” and warmed in the Industrial Age. “But intense hurricanes are so infrequent here that it is hard to get a statistically valid sample size to compare with climate,” he observes.

In the Gulf of Mexico, where Liu has been working in back-barrier freshwater lakes, the geological evidence for hurricanes extends further back into history, though it is somewhat less precise in the storm-by-storm details. “Our records go back about 5,000 years, and we definitely see a long-term cycle,” he notes. From about 5,000 to 3,400 years ago, hurricane patterns in the Gulf Coast were relatively mild. From 3,400 to 1,000 years ago, the region endured a “hyperactive” period. The past 1,000 years have been relatively placid.

“The good news is that we are living in a quiet period,” Liu says. “The bad news is that if we think we have seen too many catastrophic hurricanes coming our way, we haven't seen anything yet.”

This winter, Donnelly will take core samples from the mangrove wetlands and coastal ponds of Puerto Rico “because there should be a much more extensive record there.” Land-falling hurricanes are much more common in the Caribbean, so he hopes to find a clearer signal of how often hurricanes reach their most frightening potential.

Since 1938, development has exploded along the North Atlantic Coast. A Northeast-bound intense hurricane would threaten billions of dollars worth of property, along with one-seventh of the US population. If a modern version of the Great New England Hurricane were to strike the exact same spots as in 1938, it would result in at least $18 billion dollars of property damage.

“Most people have short memories,” says Donnelly. In fact, it is estimated that three-quarters of the population of the northeastern US has never experienced a hurricane. Donnelly’s research provides evidence to be heeded. “The geologic record shows that these great events do occur,” he says. “We need to make people aware that it can happen again. We’ve got to have better evacuation plans and we need to equip people to react to a big storm.”