9
Peter Webster remembers well the 2004 Miami conference on Hurricanes and Tropical Meteorology, sponsored by the American Meteorological Society, where he had a run-in with Bill Gray. Regarded as one of the most talented tropical meteorologists of his generation for his studies of the dynamics and forecasting of monsoons, Webster had just returned from a research trip to Bangladesh, and he was still recovering from the malaria that he had contracted there. The lanky Gray put his arm around the much shorter Australian scientist—Webster resembles a slightly weather-beaten version of Sting—and told him he really ought to get working on hurricanes.
“He said, ‘You know, you can really solve these problems,’” Webster remembers. Gray was referring in particular to the still incompletely resolved question, grappled with in his own work and by Riehl even earlier, of why some disturbances grow into hurricanes while others do not.
Before long, Webster did what Gray suggested. He got into hurricane research. Gray, however, would abhor the result.
Webster had followed the Trenberth-Landsea spat, and his initial instinct had been to side with Landsea. Webster felt especially skeptical about what he viewed as Trenberth’s attempt to “look regionally and infer globally” about trends in tropical cyclones. “Only 11 percent of hurricanes form in the Atlantic Ocean,” Webster explains. And so, suspecting a global overview of hurricanes would show that the 2004 Atlantic anomalies weren’t matched elsewhere, he set out to test what he called the “Trenberth hypothesis.” For help, Webster turned to a student of Gray’s who also happened to be an Australian friend going back decades—Greg Holland.
As Webster related this story to me in his office in Georgia Tech’s airy Ford Environmental Science and Technology Building, it suddenly seemed as though it was happening all over again. Webster’s cell phone went off with an oddly festive ring tone. It was Holland on the line, calling from Boulder, Colorado.
Not only are Webster and Holland golfing buddies who talk on the phone regularly. Before ever setting out to work together on hurricanes, they had been involved in working on the Aerosonde, a small remote-operated aircraft designed to obtain hard-to-get meteorological data—for instance, from the perilous air-sea boundary region of a hurricane, where better observations could help resolve some of the remaining mysteries about hurricane intensification (particularly concerning the role of sea spray). The Aerosonde can also stay in storms far longer than manned aircraft, and thus can take more continuous observations. Along with a colleague, Holland had conceptualized these mini-planes back in 1992 while still in Australia working for that nation’s Bureau of Meteorology Research Centre. Then he started a company to produce them, which he ran for half a decade. (The first Aerosonde flight into a tropical cyclone came in 2005—into Tropical Storm Ophelia, which performed a flirtatious dance with the Carolina coast before dying out at sea.)
In Holland, Webster had brought on board a longtime hurricane expert (particularly on storms affecting Australia) who had studied under Gray but originally trained as a mathematician, and who in 1997 had published a thermodynamic theory of the maximum potential intensity achievable by hurricanes under different environmental conditions that is the leading rival to Emanuel’s account (although the two versions have far more similarities than differences). A white-bearded scientist with bushy eyebrows and a very quotable Australian wit, Holland also happens to be a rare witness to one of the world’s most infamous tropical storms, having lived through Cyclone Tracy as it tore apart the city of Darwin, Australia, early on Christmas Day in 1974. Holland was on duty at the time, working as a forecaster at the ten-story Darwin Tropical Analysis and Regional Forecasting Centre. He was on the eighth floor when Tracy hit. The building “was jump ing around, and I literally mean jumping around,” he remembers. “It was moving around so much—and it wasn’t just swaying, it was actually literally sharp moves—that there were times when it was actually hard to walk. You’d take a step and the building would walk out from underneath you.” Although an exceedingly tiny hurricane of a type sometimes called a “midget storm,” Tracy—whose true intensity remains unknown—uprooted every tree and entirely destroyed most of the buildings in Darwin with its winds. Holland’s building took a “fair hiding” but was one of the few that remained standing.
In 2004, Holland moved to the National Center for Atmospheric Research in Boulder, Colorado, to take over the directorship of the Mesoscale and Microscale Meteorology Division. There, he helped Hai-ru Chang, a senior research scientist at Georgia Tech who works with Webster, to assemble a data set on global hurricane intensities. Later, after the data had been compiled, Webster’s partner Judith Curry, a climate scientist who is also chair of the Georgia Tech School of Earth and Atmospheric Sciences, joined the project. Their cross-disciplinary team thus brought together a hurricane specialist (Holland), a theoretically inclined tropical meteorologist who had previously paid little attention to hurricanes (Webster), and a climate researcher (Curry) whose best-known work to that point involved studies of Arctic climate and weather (in which she had employed Holland’s Aerosondes).
The team had gone into the project expecting to disprove Trenberth. But the data weren’t cooperating. Instead, they increasingly suggested that Trenberth had been on to something. No one on the Webster team had had much public involvement in the politico-scientific quagmire that is the American global warming fight. But they were about to venture into an arena where most researchers never go—the scientific equivalent of appearing on Hardball with Chris Matthews or The O’Reilly Factor.
Later, reflecting on the experience, Holland remarked that he planned on becoming a “bloody hermit on a mountaintop” the next time one of his papers on hurricanes and climate came out.
Although officially submitted on June 22, well before the peak of the hurricane season, the Webster groups study appeared in Science two months after Emanuel’s work came out, only weeks after Katrina, and just days before Hurricane Rita moved into the Gulf of Mexico. The timing, which the scientists could never have planned, had an incalculably massive impact on the amount of attention the results received. As Curry later put it in a blog comment: “When our paper was published right between Katrina and Rita, the #$%^& hit the fan.”
Emanuel’s study had limited itself to two major ocean basins. But the Science study surveyed every regularly active basin using so-called best track records from the world’s hurricane forecasting centers spanning the official satellite era, which began in 1970—a fact that lent at least some consistency to the data. This more conventional methodology had little overlap with Emanuel’s PDI approach. It certainly didn’t require cubing wind-speed measurements taken over the lifetime of each storm. Instead, the satellite-era records could be expressed in the familiar Saffir-Simpson categories of hurricane strength.
In some basins, particularly the Atlantic, the “best track” records reflected measurements from airplane reconnaissance missions, the most reliable means of sampling storm intensities. But that wasn’t possible for every basin. Seeking to cut budgets, the U.S. Department of Defense had terminated military aircraft reconnaissance of Northwest Pacific typhoons in 1987. In other basins there had never been reconnaissance in the first place. Forecasters had instead learned to determine storm strength by examining satellite images and employing a cloud pattern recognition scheme known as the Dvorak technique. As Webster and his colleagues acknowledged, this methodology had been created specifically for an operational setting in the mid-1970s, and had undergone changes over time in its manner of application. In short, the “best track” records were the best available, but only in a far-from-perfect world.
Examining these records from the satellite era, Webster’s team did not find any trend in the total number of tropical cyclones. Instead there had been a large upturn in Atlantic storm frequency since 1995, counterbalanced by a downturn in other basins. When it came to hur ricane intensity, though, a different story emerged. The study detected a “thirty-year trend” toward more frequent and intense hurricanes. Category 4 and 5 storms had apparently almost doubled in number and in proportion to weaker storms—a trend, again, closely following rising sea-surface temperatures. In conclusion, Webster and his co-authors noted that their findings were “not inconsistent with recent climate model simulations that a doubling of CO? may increase the frequency of the most intense cyclones,” citing the work of NOAA’s Thomas Knutson. And that was pretty much all they said about it.
![[Image]](images/StormWorld_169_1.jpg)
Compared to Trenberth or Emanuel, the Webster group hadn’t pointed the finger very directly at global warming. That’s partly because the authors didn’t agree among themselves about how to address the contentious subject. Holland in particular had previously linked his name with those of Landsea and Gray on papers debunking a strong hurricane-climate link. As Curry remembers: “Peter when he was first writing it essentially said, ‘Well, it looks like global warming to me.’ And then Greg Holland in his revision crossed it out and said, ‘Of course this has nothing to do with global warming.’ And my reaction was, ‘Oh my gosh, this is a tar baby.’” So the group wound up with a fairly weak statement, hardly one that directly attributed the changes they had detected to human influences. You had to read between the lines. But that wasn’t hard to do—and Holland, following Emanuel, was even then in the process of becoming the second major climate convert in the longtime hurricane-research community.
As for their take-home message to the public, Webster, Holland, and Curry had hoped to issue a warning for coastal cities while remaining relatively agnostic about the climate issue. Storms like Katrina, they stressed, might not be “once-in-a-lifetime events,” but rather might become more frequent. Knowing how the risk of intense hurricanes making landfall was changing could have enormous implications for how we decide to protect low-lying areas (or whether we decide to abandon them). As their original press release put it, a risk-assessment study should be undertaken for “all coastal cities in the southern and southeastern U.S. . . . the southeastern U.S. needs to begin planning to manage the increased risk of category-5 hurricanes.”
But that message got lost in the subsequent furor. As Webster, Curry, and Holland would later write, “Even senior scientists are ill-prepared for their first major experience with mixing politics, science, and the media.” With the press already in high gear over Katrina, and with Rita soon to make landfall, their paper immediately got sucked into the sharp and ever-intensifying global-warming debate that followed upon those storms’ destruction. That included radio and television showdowns (Curry versus Landsea on The News Hour with Jim Lehrer on PBS; Curry versus Gray on The Diane Rehm Show; Webster versus Patrick Michaels on CNN’s Lou Dobbs Tonight, and so on). It included congressional hearings. And it included environmentalist advertising campaigns that used arguments about hurricanes to raise concern about global warming, coupled with strong backlash from the community of climate “skeptics.”
So instead of getting to talk about coastal policy, “Webster and his coauthors found themselves drawn into the global-warming crossfire and baited into conflicts with colleagues like Gray. As Curry recalls of her experience with television interviews by satellite: “You’re strapped in like you’re in an electric chair with all these wires. You’re staring at this camera. People start asking you things you are totally unprepared for. We didn’t know how to be effective in that environment.” On one occasion she got a press call from Hustler magazine. “The guy was intelligent and asked good questions,” she remembers.
And their immersion in a high-stakes, very public battle over the content of the science was just beginning.
The disturbance that would become Hurricane Rita formed east of the Turks and Caicos Islands on September 17, arising from the interaction of a tropical wave and an old cold front. The thick cloud cluster moved through the Florida straits on September 20, passing just south of Key West as it became a hurricane and then intensified to Category 2 status. The warm waters of the Gulf lay in wait. Just like Katrina, Rita soon hit the infamous Loop Current and rapidly began to intensify. By September 21, the storm had transmogrified into a Category 5 monster, having spun up from a tropical storm in just thirty-six hours—a rapid deepening that the forecasters did not anticipate.
The next day Rita reached peak intensity, with maximum sustained winds approaching 180 miles per hour. Pressure is estimated to have dropped all the way down to 895 millibars (26.42 inches), making Rita the strongest known hurricane in the Gulf and the third-most intense storm ever detected in the Atlantic basin. But soon that rank would drop to fourth.
Rita was also a very large hurricane, occupying much of the breadth of the western Gulf at its height. Luckily, the storm ultimately targeted a relatively less populated area just east of the Louisiana-Texas border for landfall. Like Katrina, Rita also weakened considerably as it approached shore, so that maximum sustained winds at landfall were only about 115 miles per hour, or Category 3 strength. Despite the storms weakened state, however, Rita drove a large surge, possibly as high as fifteen feet in places. The storms final destination fell very close to the site struck in 1957 by Hurricane Audrey, and the surge destroyed entire coastal communities in Louisiana’s Cameron Parish, with the circle of destruction spreading into the city of Lake Charles, with its mix of refineries and casinos, and Vermilion Parish. The ultimate damage total was on the order of $10 billion.
Rita’s impact cannot be judged solely from this tale of destruction, however. Psychologically, a hurricane even stronger than Katrina, appearing only weeks later, shocked Americans. It seemed as though every time they saw a satellite picture of the Gulf on television they found themselves staring into a huge vortex. Furthermore, Rita’s earlier track had pointed straight at Houston, Texas, prompting the largest evacuation in that state’s history. “Unless the storm turns south or north in the next 24 to 48 hours we are set up for a truly horrific event,” Houston Chronicle science reporter and hurricane-beat specialist Eric Berger wrote on his weblog on the evening of September 21. “I am not going to sugar-coat this, my friends.”
Some two million people crowded the interstates as Rita approached, some of them Katrina evacuees who had fled to Houston less than a month earlier. The evacuation was a disaster: People sat in traffic for as long as sixteen or eighteen hours to get to nearby cities like Austin. Cars broke down or ran out of gas, leaving their drivers stranded along roadsides in 100-degree heat. CNN filmed the chaos—including images of standing traffic on one side of highways and no cars at all on the other—and its reporters demanded to know why more lanes weren’t open. Finally, Texas officials implemented a “contra-flow” regime they had never rehearsed. In the end many of the more than one hundred deaths attributed to Rita arose indirectly, in the context of the evacuation. Twenty-four nursing-home evacuees died when their bus burst into flames along the highway outside Dallas.
As this disorganized mobilization suggested, Houston, like so many U.S. coastal cities—including nearby Galveston, situated on a barrier island and protected after the 1900 storm by a fifteen-foot sea wall—faces a catastrophe if directly hit by a powerful hurricane. Located across Galveston Bay from the Gulf, Houston could find itself staring down a giant wall of water if a strong hurricane approaches leading with its right front quadrant. “Within an hour or two, a storm surge, topping out at 20 feet or more, would flood the homes of 600,000 people in Harris County,” wrote Berger in a February 2005 warning article for the Chronicle that, had Rita taken a more southerly course, would have been unendingly praised for its clairvoyance (just as all the articles predicting the destruction of New Orleans later were). The sea wall protecting Galveston Island would valiantly fight off the surge, Berger noted, but fail as water crept behind it and engulfed the island from the bay side. Near the Port of Houston, meanwhile, chemical and water treatment facilities would also find themselves underwater—leaking God knows what. Damages could reach $40 or $50 billion, and hundreds or even thousands of people could perish.
So while Houston may not be quite as vulnerable as New Orleans to a hurricanes devastation—at least it would drain more quickly—it will someday experience a storm that changes its history forever. It’s only a matter of time.
Rita drove the public and media dialogue about global warming and hurricanes—keyed, once again, to land-falling storms—to an even more frantic pitch. Everyone had an opinion on the issue. On the one hand, a who’s who of Democratic leaders, including Al Gore and Jimmy Carter, cited the recent findings by Emanuel and the Webster group to warn that global warming had contributed to the hurricane problem and to argue that action on greenhouse gas emissions must follow promptly. Skeptics responded by continuing their campaign to dispute the scientific evidence and insisting that no serious cuts in emissions were needed. At one point in Congress, New Mexico Republican Pete Domenici called California Democrat Dianne Feinstein’s suggestion of a hurricane-global warming link “nuts.”
Public opinion, too, seemed on the move. In an ABC survey taken just after Rita, 39 percent of respondents said they thought the hurricanes were the result of climate change, a considerable increase since polls taken just after Katrina. Posing the question differently—“Thinking about the increase in the number and strength of hurricanes in recent years, do you think global warming has been a major cause, a minor cause, or not a cause of the increase in hurricanes?”—a USA Today!CNN Gallup poll found that 36 percent of the public viewed global warming as a major cause, 29 percent viewed it as a minor cause, and 30 percent felt it was not a cause. Yet the poll’s question wording conflated the different scientific issues of hurricane frequency and hurricane intensity, and did not specify whether the question referred to the undisputed upturn in Atlantic hurricanes or to a global increase. Assuming the latter, that means the question assumed at face value the validity of the Emanuel and “Webster results. Yet many hurricane specialists, particularly Landsea and Gray, strongly disputed those results. Or as Gray barked at Curry when the two appeared together on The Diane Rehm Show on September 21, 2005, the day of Rita’s rapid intensification in the Gulf: “I question whether the authors are that expert in the analysis of this data that they can make these conclusions.”
Unmistakably, the Emanuel and Webster papers suggested a very different view of the hurricane-climate relationship than had existed before. Previously the consensus had been: Sure, global warming might strengthen hurricanes—but not by that much, and not in a measurable way in the near future. Now, however, a new vision was coming to life, and it was much more disturbing.
One thing hadn’t changed. The relationship between hurricane frequency and global warming remained murky. Some modeling studies generated more storms, and some fewer, for a doubling of CO2 concentrations. And except in the Atlantic, which had seen a dramatic upswing in storm numbers since 1995, there didn’t appear to be an appreciable trend in storm frequency. Neither was there a very reliable theory to explain why there should or shouldn’t be one. Given that large and intense hurricanes do much more to cool down the oceans than weak tropical storms, it’s perfectly conceivable that global warming might lead to stronger storms but fewer of them in total—precisely the result produced in one GCM study using Japan’s Earth Simulator, which ranks among the world’s most powerful supercomputers.
Any individual hurricane, however, can cause harm to life and property in four basic ways. And now it seemed possible that each of these destructive storm characteristics might be worsened in the future—or, perhaps, had already worsened—due to global warming.
The deadliest aspect of a hurricane is its storm surge, which can easily exceed twenty feet in a land-falling Category 5 storm—and that’s not counting the huge waves that ride atop such surges. Hurricanes Camille of 1969 and Katrina of 2005 presented dramatic examples of deadly surges, the worst the U.S. Gulf Coast has seen. But for sheer carnage, neither can begin to rival the 1970 storm that struck present-day Bangladesh, whose surge killed over 300,000 people, even more than the 2004 tsunami.
Considering that winds play the most central role in driving the wall of water that hurricanes propel toward shorelines, this destructive aspect of storms should worsen if the average hurricane grows more intense. Furthermore, one of the most certain outcomes of global warming is a rise in sea level, caused both by the melting of ice and by thermal expansion of seawater. If all hurricanes ride atop higher seas, then barring some dramatic retreat from coastal areas, all hurricanes pose a greater threat to human lives and property in those areas—whether or not the storms have independently intensified.
After storm surges comes the next chief source of hurricane damage: winds, which can be powerful enough to tear down trees and all but the sturdiest of human-built structures. Some famous examples of purely wind-driven hurricane destruction occurred in 1974’s Cyclone Tracy, 1992’s Hurricane Andrew, and 2004’s Charley. The Emanuel and Webster work most obviously suggested that the average hurricane’s maximum sustained wind speed is increasing due to global warming—a particularly troublesome outcome in that wind damage increases steeply as velocity rises.
Third, in addition to storm surges and wind destruction, hurricanes cause heavy rains and, in some cases, catastrophic flooding. The canonical example here is Hurricane Mitch, which came ashore as a very weak storm and yet killed more than 11,000 people through rainfall alone. Another major rainfall killer was Tropical Depression Jeanne of 2004, and there have been many, many others. In a globally warmed world, hurricane precipitation, too, is expected to increase, because the air converging into the storm center should carry more water vapor—thus potentially upping the chance that a given hurricane will create a rainfall-related disaster. Thomas Knutson’s modeling study of hurricane intensification under global warming scenarios predicted an eventual 18 percent increase in rainfall near the storm center for increased concentrations of atmospheric CO2.
Changes to the final major source of hurricane damage—hurricane-spawned tornadoes—are probably the most speculative. Such tornadoes were discussed in early empirical work by Gray, which noted that among many other factors, hurricanes making U.S. landfall from the Gulf of Mexico tend to generate many more tornadoes than those striking from the Atlantic—presumably because hurricanes approaching from the Gulf lead with their powerful right front quadrant.
Most U.S. land-falling hurricanes generate at least some tornadoes, and although these tend to be fairly weak whirlwinds on the Fujita scale (in the F0 or F1 range) and no rival to Great Plains tornadoes, they can sometimes be stronger. Tropical Storm Beryl of 1994 spawned three F3 tornadoes, with winds up to 206 miles per hour. But the fundamental fact is this: Stronger hurricanes tend, on average, to spin off more numerous tornadoes at landfall. As Gray and a co-author put it in 1974: “There appears to be a direct relationship between tropical cyclone intensity and tornado incidents.” If so, it stands to reason that these incidents should increase in number if hurricanes grow more intense (and continue to make landfall at similar frequencies). There’s some evidence that, in conjunction with more active Atlantic hurricane seasons since 1995, the number of hurricane-spawned tornadoes has also gone up, reaching a high number of 302 in 2004 followed by 194 in 2005. So if global warming is contributing to the Atlantic hurricane uptick since 1995, it may also be linked to increased tornadic activity.
Temporarily assuming, for the sake of argument, the validity of the Emanuel and Webster studies, what does all of this add up to? Fifty years from now, even if all four sources of hurricane damage have indeed worsened as suggested above, we still wont be able to point to a particular storm and say, “Aha, global warming!” Rather, we’ll be in the same position we’re in now, forced to argue over statistics to detect trends.
So far, data on damage from U.S. hurricanes that has been “normalized”—i.e., adjusted for changes in population and wealth, as well as for inflation—does not show any trend over time. The 1926 Miami hurricane, if it happened today, would be expected to cause considerably more damage than 2005’s Hurricane Katrina. However, U.S. damage levels may not be a very good indicator of whether hurricanes themselves are changing in response to global warming. Any damage study must inevitably limit itself to the relatively small number of U.S. land-falling storms; moreover, damage levels depend on countless factors besides storm strength itself, ranging from storm track to tide at the time of landfall. Perhaps most important of all is societal change. If more people are living in the paths of hurricanes, and putting valuable homes and possessions along those paths, then storms will inevitably cause much more damage—and that’s precisely what we’ve seen happen in the United States.
It stands to reason, however, that if storms also change in a major way, then barring some other major countervailing change, that will eventually show up in damage statistics. In turn, those analyzing such statistics, including insurance and reinsurance companies, could find themselves shocked by the new world in which they’re trying to write coverage. Whether or not global warming can be blamed, that already seems to be happening. In the wake of the 2005 Atlantic hurricane season, Risk Management Solutions, a “catastrophe modeling” firm that works for insurance companies, revised its estimate of hurricane risks to Florida and the Gulf Coast upward after calling together a panel of scientists including Emanuel and Knutson.
Finally, it’s not enough merely to consider that all four vectors of hurricane destruction could worsen as a result of global warming. We must also weigh still more ambiguous evidence about how the world and its storms may change—evidence about the possible linkage between the occurrence of South Atlantic hurricane-type storms and global warming, for instance. If, as some scientists suspect, the South Atlantic is a “marginal” basin that generally can’t support hurricanes, but with slight climatic chang es could begin to host them more regularly, this could have serious consequences for Brazil and perhaps other nations in the region. Similarly, we must weigh the possibility that by raising sea-surface temperatures, global warming would increase the chance that already-formed hurricanes will be able to strike regions where they have never or only rarely made landfall in the past, such as Southern California.
Such possibilities are hard to express in terms of a statistical likelihood—many have not been studied in detail—but not, for that reason, acceptable to ignore. Now, the Emanuel and Webster studies, combined with hurricanes Katrina and Rita, had brought them all into sharp focus.