Scientists Think Like a Hurricane to Beat the Next Katrina more similar news »
When Hurricane Katrina struck New Orleans in late August 2005 and the levees around the city broke, flooding the city and killing hundreds, Ed Link was as surprised as everyone else.
He shouldn't have been. As one of the nation's foremost hurricane experts, Link, a professor at the University of Maryland, had access to the government's most sophisticated mathematical models for predicting damage from big Gulf Coast storms. But those models weren't accurate because the data they were based on were incomplete, out of date or just plain wrong.
As the floodwaters receded and the Army Corps of Engineers rushed to repair the levees, the government asked Link to lead a team of engineers and scientists from the government and private sector -- 300 in all -- to recode those old models. The goal of the vaguely named Interagency Performance Evaluation Task Force was twofold, Link told Wired.com: first, "to get that knowledge built back into the levee repairs so the same vulnerability wasn't built into the system again. The second was to come up with a 'risk assessment' looking forward."
In other words, to have a much better idea, grounded in solid science, of who might be killed or have their property destroyed in future Gulf Coast hurricanes.
The levees have long since been fixed and upgraded, but the risk assessment -- based on a mind-boggling 2 million equations -- is just now nearing completion. As the math came together beginning in 2007, the task force began publishing color-coded, interactive maps in an effort to show Gulf Coast residents what kind of danger they likely faced from hurricanes. The Google Earth-based maps can be found on the Army Corps website.
The ultimate "risk" map, the culmination of the task force's work representing tens of thousands of square miles from Florida to Texas, is slated for release this week.
Gathering the data for the levee upgrade and the risk maps took three years of back-breaking, mind-numbing effort by hundreds of team members using a surprising mix of high technology, old-fashioned detective work, trick psychology and, when all else failed, intuition. The results have revolutionized authorities' understanding of Gulf Coast hurricanes.
But whether the public will pay heed is another matter.
Katrina dissipated on August 30, 2005. In early September, rescuers had just begun going house to house in New Orleans looking for the living and the dead. But Link's team was already on the ground collecting what he called "perishable" data, such as the depths and locations of floodwaters.
For many team members, data collection was dirty, dangerous, thankless work -- and it meant short-shifting their day jobs. "A lot of people just quit what they were doing and basically worked full-time" on the new storm model, Link told Wired.com.
But for one key team member, it wasn't just about sloshing through flooded streets. Don Resio, a scientist working for the Army Corps of Engineers, went hunting for old data sets from decades-old storms, in hopes that historic hurricanes might whisper hints about future ones.
Resio told Wired.com that his hunt mostly involved polite requests to cooperative government agencies like the National Weather Service. But other, equally vital reams of data were locked in the safes of the Gulf Coast oil companies, who, with billions of dollars invested in offshore drilling platforms, were especially concerned with the high winds that come with big storms.
Resio needed that data, but it wasn't his to demand. His solution? "I made 'em feel guilty," he recalled with a laugh.
Slowly, the data came together, culled from more than 150 storms dating back a hundred years. Key figures came from new, high-tech microwave sensors installed aboard "hurricane-hunting" C-130 and P-3 airplanes operated by the Air Force and the National Oceanic and Atmospheric Administration.
A scale-model levee was stress-tested in the world's most powerful centrifuge.
Courtesy Army Corps of Engineers
To create entirely new data from scratch, the task force built a detailed model of New Orleans and flooded it, essentially recreating Katrina on a nonlethal scale. And to zero in on the levees, the team built a miniature earthen levee inside the world's most powerful centrifuge. They added water and spun the centrifuge at speeds duplicating hurricane-force wind and waves, looking for when, where and how the levee would fail.
There were some surprising revelations in the course of the task force's investigation … some of which helped explain why Katrina had taken so many people by surprise. For one, Link's team found that the existing elevation maps of New Orleans were way off and would have to be totally redrawn. "We found things two feet below where people though they were," Link said. Obviously that made the city more vulnerable to flooding.
Also, in tightening up and rewriting the old mathematical models, the task force gained a clearer understanding of the limitations of modern science. "There's a lot we just don't know," Resio told Wired.com. But, as former Secretary of Defense Donald Rumsfeld once said, there are "unknown unknowns," which are bad, and there are "known unknowns," which are somewhat better. Finally the hurricane task force knew the basic outline of what it didn't know.
But when it comes to math, even known unknowns can be tricky. Resio said that for some equations, he and the other researchers needed figures, any figures. So they had to guess. That meant thinking like a hurricane, trying to intuit how wind and water might behave under certain conditions.
Necessary educated guesses aside, Resio told Wired.com that uncertainty is a key parameter of the new storm models -- especially as global warming whips the planet's fundamental weather patterns in unpredictable ways. The team knew they had to capture this unpredictability mathematically and build it into the models.
Spinning at speeds duplicating a hurricane, the scale earthen levee turns to liquid and disintegrates.
Video courtesy Army Corps of Engineers
"We did a bunch of numerical tests to determine variability," Resio said. In other words, they looked at the surprising behaviors of past storms. Were winds unusually fast? Or was the ratio between the size of the storm and wind speed different than the norm? "We added that variability back into the model as a random function," Resio said, so that when officials use the new models to predict hurricane damage, they get a range of predictions. It's one of the new models' greatest strengths, Resio said.
After three years of labor by hundreds of engineers and scientists, emergency managers now have a much better understanding of what kind of damage a major storm might cause. But that doesn't mean that the people most at risk -- Gulf Coast residents -- take these predictions seriously.
Sometimes all the mathematical models and colorful maps in the world won't change a person's mind, which is why many New Orleans residents have rebuilt destroyed homes in exactly the same place, and to the same construction standard, as before Katrina.
To combat public ignorance and complacency, Resio's team includes "risk communicators" -- basically, PR reps for hurricanes. Ironically, the high-tech storm models and sophisticated maps that the risk communicators rely on might actually undermine their work, according to one academic who has studied storms.
"The technologically enhanced discourse of prediction conveys the sense that weather media viewers can be prepared," Marita Sturken, from New York University, wrote in 2006. She called this a technological "selling of preparedness."
Resio is aware of the challenge in making potential hurricane victims believe that they're at risk, even when the world's most sophisticated storm models insist they are. "How do you convince people they need to be concerned?" he said, sighing. "The risk communicators have their hands full."
Mon Jun 09, 2008 more from this source»»
| 
