CORKY PERRET lost everything when Hurricane Katrina hit. His house on the beachfront out on Highway 90 between Gulfport and Biloxi, Mississippi, was reduced to firewood by the winds and washed away by a 10-metre storm surge. “Nothing is left, it was totally destroyed,” he says.
Some of his neighbours are going for good. Perret, a fisheries official, would like to stay and rebuild. But he is still unsure what to do. His house was built during a period when there were few hurricanes. For the next few years at least, that era is over. And it could be that the quiet days will never return. Whatever happens, the effects of future hurricanes will probably be worse, he says, because the offshore islands that provided some protection have been partly washed away.
Perret is one of the many people around the Gulf of Mexico trying to cope with an uncertain future after an extraordinary year in the Atlantic. There were so many tropical storms that for the first time ever forecasters ran out of names. Wilma became the most powerful Atlantic hurricane ever recorded. Katrina brought New Orleans to its knees, while consumers across the world reeled as oil prices shot up.
Advertisement
So what’s going on? Are hurricanes becoming more destructive as global warming kicks in? Not long ago, the world’s leading hurricane experts all agreed on the answer: the upsurge in Atlantic hurricanes is just part of a normal long-term cycle. There is no sign that global warming is making hurricanes anywhere in the world stronger or more frequent. As for the future, we haven’t got a clue, because our models are not good enough to predict how climate change will alter tropical storms.
That was the official consensus in 1996. Two years later the top researchers, including William Gray of Colorado State University and Kerry Emanuel of the Massachusetts Institute of Technology, felt confident enough to go a little further. In a paper in the Bulletin of the American Meteorological Society in 1998, they concluded that a doubling of carbon dioxide levels in the atmosphere (CO2 levels are now up about a third) would not affect the frequency of tropical cyclones or alter the areas they affect, and would increase their intensity only slightly, by around 10 per cent or so.
Now this consensus has been shattered. Alarm has grown with a flurry of papers claiming that, contrary to expectations, the surge in temperatures over the past 30 years has already made hurricanes more severe. Not more frequent, but more intense, with stronger winds, longer durations, more rainfall and even less predictable paths.
And if the odds on powerful hurricanes are already shortening, there might be worse to come. “Future warming may lead to an upward trend in tropical cyclone destructive potential,” Emanuel wrote in one paper, “and – taking into account an increasing coastal population – a substantial increase in hurricane-related losses in the 21st century.”
Such claims have created a schism among the high priests of hurricane forecasting. Many, like Gray, see no upward trend and no human fingerprint. They accuse the authors of the latest papers of bias. So who is right?
Worldwide there are about 85 tropical cyclones each year, of which about two-thirds reach hurricane force. The number does not vary much year by year, but their distribution does. Global conditions that encourage hurricanes in the Atlantic tend to discourage them in the Pacific, and vice versa. This year the Atlantic has been battered, but the Pacific has been relatively peaceful (see Map). But the world’s hurricane cockpit, with more than a third of the world total, is the western North Pacific. Last year Japan alone was hit by 10 typhoons (the local name for hurricanes), three times the usual number.
Tropical cyclones start off as clusters of thunderstorms, which often form as warm, humid air rises from the surface of a tropical ocean. As the air rises, the water vapour condenses, releasing latent energy that heats the air and sends it even higher. This latent energy is what drives hurricane formation.
If enough storm clouds form in close proximity, they can create what Emanuel calls a “pillar” of humid air, extending from the ocean surface right up to the boundary with the stratosphere. The low pressure at the base of the pillar sucks in more air, which picks up energy in the form of water vapour as it flows over the sea surface and releases it as it rises, lowering the pressure still further. If this happens more than four degrees north or south of the equator, the Coriolis forces acting on the inward-flowing air start the cyclone spinning.
When conditions are favourable, a tropical cyclone can rapidly gain strength, turning from a depression into a storm and finally a hurricane as wind speeds pick up (see Diagram). Its power is staggering: Chris Landsea of the US National Oceanic and Atmospheric Administration (NOAA) in Miami has calculated that a typical hurricane can release over 10 million trillion joules a day – the equivalent of about a million Hiroshima bombs. Luckily for us, most of this energy goes into driving air upwards rather than sideways.
On the face of it, global warming can only make things worse. The initial pillar of humid air generally forms only when the temperature of the sea surface exceeds 26 °C. As the oceans warm, larger areas will exceed the threshold. And every degree above the threshold seems to encourage stronger hurricanes. When Katrina hit category 5 back in August, the surface of the Gulf of Mexico was around 30 °C. This has encouraged the view that a warmer world will have more hurricanes, stronger hurricanes and hurricanes in places till now outside their range.
How to kill a hurricane
But the world is not that simple. As Gray, one of the founders of hurricane science, pointed out decades ago, a hurricane can form only when a whole set of atmospheric conditions are just right. One of these is the difference between the sea temperature and the air high above it. This is what drives the convection currents that are necessary for the initial storm clouds to form. If, as climate models suggest, global warming raises average temperatures high in the atmosphere as well as at the ocean surface, then the sea surface may have to get even warmer before it triggers hurricanes – and the hurricane-generating potential of the tropics would remain largely unchanged.
Nor are hurricane formation zones certain to expand. Hurricanes usually develop in or near the band of rising air called the intertropical convergence zone, which moves across the equator with the seasons. And for every spot where warm air rises, other air has to fall. Beyond 20 degrees north or south, descending air generally blocks storm formation. It is possible, though, that warmer seas will allow hurricanes to travel further out of the tropics after they have formed.
There are other complications, too. Even when conditions are just right, no hurricane will form unless an atmospheric disturbance provides a trigger. And even then, most tropical storms fizzle out instead of growing into hurricanes. So, while it is important to know if global warming will more frequently create the conditions needed for hurricanes to form, it is equally important to know whether it will be more likely to trigger them and, once formed, to kill them.
There are two ways to kill a hurricane. The first is to shut off its fuel – the supply of warm, humid air that comes from warm ocean waters. This happens most obviously when a hurricane passes over land. But it can also happen at sea. As the storm grows, its waves stir up the ocean, mixing the warm surface water with the generally cooler water beneath. The surface cools and, particularly for slow-moving hurricanes, that can be the end.
This means that a hurricane can grow very intense only when the warmth extends for tens of metres below the surface. Katrina strengthened as it moved towards New Orleans because the water was warm to a depth of more than 100 metres. Leaving Katrina aside, here, for once, the effect of global warming looks clear-cut: warmth is penetrating ever deeper into the world’s oceans and setting up ideal conditions for hurricane growth.
But the second way by which hurricanes are defused could have the opposite effect. Winds can disrupt the pillar of humid air. Even quite small invasions of wind can mess up a potential hurricane by introducing dry air, preventing thunderclouds forming or lopping off their tops. Climate models are not good at replicating small-scale features like hurricanes, but most suggest that global warming will increase wind speeds in the upper atmosphere to levels that would disrupt hurricanes.
With all these uncertainties and contradictions, it is not surprising that different computer models have predicted everything from fewer hurricanes to more hurricanes as global warming kicks in. It doesn’t help that even the best models are not that good at predicting hurricanes in today’s world. “If we can understand why the world sees about 85 named storms a year and not, for example, 200 or 25, then we might be able to say more about what we’d expect in a global warming scenario,” says Peter Webster of the Georgia Institute of Technology in Atlanta. “Without this understanding, forecasts are merely statistical extrapolation.”
“If we can understand why the world sees 85 storms a year, and not 25 or 200, then we could say what global warming will do”
It gets worse. Another complication is whether climate change will encourage or discourage the natural climatic cycles that are known to influence the virulence of hurricane seasons. El Niño, the periodic reversal of wind and ocean currents in the equatorial Pacific, is a key player here. During an El Niño event, the warm waters of the tropical western Pacific spread east and the zone where thunderstorms can form spreads with them. Result: more hurricanes for the year or so that the event lasts. El Niños have the opposite effect in the tropical Atlantic, where they disrupt hurricane formation by increasing westerly winds high in the atmosphere.
But Atlantic hurricanes also have longer-term cycles. The last upsurge ran from the 1940s to the 1960s. The current one has been running for a decade and could last another 20 years. The driver for this cycle appears to be changes in the Atlantic circulation system known as the ocean conveyor, which powers the Gulf Stream.
There are other factors at work, too. Atlantic storms mostly form in the waters off west Africa and are more likely when this region is wet. This may be because during dry years in west Africa, dry, dusty winds blow over the Atlantic, disrupting hurricane growth. But none of this helps much with predicting hurricanes in a warmer world. No one is clear what global warming will do to El Niño. Climate change may or may not shut down the ocean conveyor. The Sahara could get wetter, but then again it might get drier.
Faced with such overwhelming uncertainty, until recently researchers all agreed that predicting exactly how global warming will change hurricane patterns in specific regions is impossible. As for the global picture, theory and the latest models suggested only a small increase in intensity in the future.
Kevin Trenberth of the National Center for Atmospheric Research in Boulder, Colorado, was the first to break ranks. In October 2004, after a record Atlantic hurricane season, he sparked controversy when he claimed that global warming was already making a difference: “High sea surface temperatures…make for more intense storms and this is consistent with the evidence that we’re seeing.”
Trenberth’s claim infuriated Landsea, who was then working with Trenberth on a chapter for the next Intergovernmental Panel on Climate Change report on, among other things, hurricanes. Landsea promptly resigned from the IPCC: “I cannot…continue to contribute to a process that I view as both being motivated by preconceived agendas and being scientifically unsound,” he wrote. Ouch.
But this August, Emanuel published a study suggesting that not only is Trenberth right, but the trend is already far greater than expected. After trawling records of hurricane intensity worldwide for the past half-century, Emanuel concluded that on average storms are lasting 60 per cent longer, with wind speeds up to 15 per cent higher (Nature, vol 436, p 686). This may sound small, but the damage done by a hurricane is proportional not to the wind speed, but to the wind speed cubed. The results suggest the destructive power of a typical hurricane has increased by 70 per cent. “This work implies that global tropical cyclone activity is responding in a rather large way to global warming,” Emanuel says. “I was one of the sceptics myself a year ago,” he adds.
A month later, Webster and Greg Holland of NOAA also published evidence supporting Trenberth (Science, vol 308, p 1853). “We went about this business because we did not believe his allegation that what has happened with Atlantic hurricanes is due to global warming,” says Webster. “We finished up thinking his conclusions were partly right.”
While there has been no overall increase in the number of hurricanes, Webster and Holland conclude that the frequency of the strongest storms has almost doubled since the early 1970s. The trend, they say, is global and clearly connected to the worldwide rise in sea surface temperatures. That makes it extremely unlikely to be caused by natural cycles, which are relatively short-term and confined to single ocean basins. “We can say with confidence that the trends in sea surface temperatures and hurricane intensity are connected to climate change,” Curry declared.
“We went about this because we did not believe what has happened in the Atlantic is due to global warming. We finished up thinking it is in part”
The controversy has grown bitter. In a detailed response posted on his website in mid-October (), Gray claims that Emanuel’s calculations “are not realistic”, while Webster’s findings are “not physically plausible”.
Webster points out that Gray has contributed to only one paper on global warming and hurricanes. “We are responding to a person who has not done the research.”
So what is the controversy about? Put simply, Gray believes that both papers are based on bad data. In particular, he says, measurements from flights through typhoons between 1973 and 1986 in the western Pacific, the largest area of hurricane activity, cannot be relied upon. During this period, a new method for converting raw measurements into estimates of wind speeds was adopted that, all researchers now accept, produced systematic underestimates. This polluted Emanuel’s calculations, Gray says, and provided most of the baseline for Webster’s claims of an increase in category 4 and 5 hurricanes.
“If we disregard this anomalous period…we see little difference,” Gray says. Exclude the bad data, he concludes, and there are no trends to speak of outside the North Atlantic – and there the recent upsurge is due to a well-established natural cycle.
Neither Emanuel nor Webster deny that there are problems with data consistency, but they both say they have made appropriate corrections. And Emanuel did tweak his corrections to Atlantic data in September after discussions with Landsea (). “Gray has not brought to my attention any difficulties with the data of which I was not already aware and had corrected for,” Emanuel says. The more reliable records of air pressure in the eye of hurricanes provide a means of checking the validity of the wind speed measurements, he points out.
Webster says his figures all come from the period after 1970, when satellite measurements can corroborate those from aircraft. He accuses Gray of “grasping at thin air”.
Super Typhoon Tip
Part of the reason for the dispute may be the different perspectives of the protagonists. Gray’s forecasts are largely based on what happened when conditions were similar in the past. The business of climate researchers is to disentangle long-term trends – to look for differences, not similarities. Likewise, Gray’s main interest is the 10 per cent of hurricanes that occur in the Atlantic, where everyone acknowledges that, whatever part climate change is playing, the natural cycle is still the dominant factor driving storm frequency and intensity. Webster and Emanuel are equally concerned with the other 90 per cent of hurricanes around the world.
So where does that leave us? There is as yet nothing extraordinary about recent hurricanes. The storm that hit Galveston, Texas, in 1900 killed more people than Katrina in 2005. Both pale compared with the East Pakistan hurricane of 1970, which probably killed half a million people in what is now Bangladesh. And the largest and most powerful hurricane ever recorded occurred in 1979. Typhoon Tip had a central pressure of just 870 millibars, wind speeds of 300 kilometres per hour and a diameter of 2174 kilometres. Luckily, it had weakened greatly before it swept over Japan.
It is also pointless to argue about whether any recent individual hurricanes have been made worse by global warming – it is simply impossible to know. In fact, Emanuel estimates that it will take 50 years to detect a clear trend in the intensity of hurricanes hitting the US, because the numbers are so low. But in the North Atlantic as a whole, the large upswing in hurricanes in the last decade is “unprecedented, and probably reflects the effect of global warming”, he declared in his Nature paper.
Evidence, then, is starting to emerge of a human fingerprint in hurricane trends. It is not yet proof, but neither can it be ignored. And perhaps the most striking finding of both Emanuel and Webster is the strong link between sea surface temperature and storm strength (see Graphic). At least for now, the simplistic view that warmer seas mean stronger hurricanes is holding true.
What worries a growing number of researchers is the possibility that this trend will continue, that hurricanes will respond to the increasing amounts of energy swilling around in a warmer world by becoming substantially stronger. We might be, to put it simply, throwing more fuel on the fire.
Even small increases in wind speed, rainfall or the level of storm surges can have huge implications for the damage a hurricane causes, and for the likely death toll. On the other hand, the impact could be greatly reduced if we prepare for the worst: stopping development in the most vulnerable areas and making buildings and infrastructure more wind and flood-proof. Perhaps the question planning authorities should be asking is not, “Are Emanuel and Webster right?”, but “Can we afford to bet on them being wrong?”. It’s a question occupying Corky Perret, too.
