Ten years ago, Hurricane Katrina roared ashore in southeastern Louisiana, and the resulting storm surge floods—smashing height records at 10 meters—produced the costliest and one of the deadliest natural disasters in U.S. history. But because the Gulf Coast is a historical hotspot for hurricanes, the region’s risks were well-studied. Now, a new method of risk assessment finds even areas that haven’t been historical hurricane hotspots may have cause for concern. The technique—a model that takes into account the physics of a storm as well as the shape of a region’s sea floor and coastline—offers city managers a better way to anticipate, and prepare for, disaster, the authors say.
The world’s well-known hurricane hotspots—whether New Orleans, Japan, or the Philippines—tend to command public attention when it comes to forecasting, because they experience strong storms relatively frequently. Furthermore, the memory of a powerful storm’s impact can motivate both residents and city planners to better prepare for their region’s vulnerability. But historical records of hurricanes (or tropical cyclones, as they are generally called around the world) “are incomplete and don’t go back very far,” says Kerry Emanuel, an atmospheric scientist at the Massachusetts Institute of Technology in Cambridge. “We have in the Atlantic maybe 50 good years, and another 50 iffy years, of hurricane record. And in the rest of the world that historical record is even worse.”
In fact, there are a number of places around the world that may be highly vulnerable to the strong winds, rainfall, and storm surges of tropical cyclones, even though such storms haven’t struck them in recorded history. Risk managers use the term “black swan” to refer to a truly unpredictable, unavoidable event with powerful consequences. But Emanuel and his colleague Ning Lin, a civil engineer at Princeton University, suggest that for many such regions, the risk from a tropical cyclone may belong to a slightly less silent category of risk: the “grey swan.” Such events, they say, may not be predicted from history alone, but are somewhat foreseeable based on other available data, particularly storm physics and the geophysical setting.
By coupling hurricane models with hydrodynamic models—which simulate how water moves in, for example, a coastal region—Lin and Emanuel assessed the threat of storm surge from tropical cyclones for three low-lying and vulnerable coastal regions: Tampa, Florida, where the most recent strong hurricane was a 1921 storm that flooded the city; Cairns, Australia, which has been affected by several recent cyclones; and the Persian Gulf, which has not been hit by a tropical cyclone in recorded history. The most destructive aspect of a tropical cyclone tends to be its storm surge, which is complicated by factors including the storm’s intensity, size, and the shape of the coastal sea floor. This combination of factors makes storm surge a good metric for identifying a grey swan storm, the scientists say. To predict the likelihood of such storms, they used a two-part model. The first part, a deterministic model, simulates storms in many different environments over tens of thousands of years. The second part involved a statistical analysis of this large sample.
For all three regions, they saw the potential for severe “grey swan” storms, Lin and Emanuel report today in Nature Climate Change. In Tampa, for example, a one-in-10,000-year-strong tropical cyclone could generate a storm surge of 6 meters, whereas a storm surge of about 3.5 meters—similar to the 1921 storm—could happen every 100 to 300 years. Similarly, a 10,000-year storm striking Cairns could produce a storm surge of about 5.7 meters, and one bearing down on Dubai, United Arab Emirates, could produce a storm surge of 4 meters.
These aren’t disasters that will strike every year, or even every decade, Lin says. “We emphasize that this is about extreme storms.” But she notes that these statistical frequencies are also based on the climate of 1980 to 2005. A 10,000-year event in the late 20th century could become a 700-to-2500-year event by the late 21st century because of climate change, Lin says. Other factors not included in the model, such as rising sea levels, may further exacerbate this threat. The goal, she says, is to give policymakers more tools to effectively predict risk and to plan accordingly. To that end, the researchers provide risks for the impact of storms that might recur every 10 to 10,000 years. “We want to remind people that the risk is there—and may increase in future due to climate change and sea level rise.”
Putting numbers on such “grey swans” can be crucial to getting decision-makers to pay closer attention, says Erwann Michel-Kerjan, executive director of the Wharton Risk Center at the University of Pennsylvania. “People are really bad at probability; they don’t know what to do with it,” he says. But, he adds, Hurricane Katrina and 2012 Superstorm Sandy, have dramatically changed the landscape of storm risk assessment. For one thing, scientists are now focusing on the impact of storm surge on coastal communities, rather than just wind damage. And, as no one wants another Katrina on their watch, managers are also more likely to take the probabilities seriously. “You cannot just look at the past. You have to look at a set of possibilities in the future,” Michel-Kerjan says. “Just because it hasn’t happened yet doesn’t mean that it will not happen.”