Is climate change causing an upsurge in us tornadoes? 30 July 2008 by Chris Mooney

• 
  30 July 2008 by Chris Mooney
  
• 
  Magazine issue 2667.
  

FROM the start of 2008, something seemed amiss with the weather. During an unseasonably warm January, two tornadoes struck Kenosha County, Wisconsin, damaging 105 homes. It was only the second known January tornado strike in Wisconsin. Just a few days later, a tornado struck Vancouver in Washington - only the third January tornado in that state since 1950.

Then on 5 and 6 February, a storm system spawned 84 tornadoes across several southern states as their primary elections took place. Five of the tornadoes were powerful EF4s on the so-called Enhanced Fujita scale. The Super Tuesday outbreak was the worst for 23 years, and left 57 dead. Once again, it came unusually early in the year. Unusually warm weather seemed partly to blame.

By the end of that month, 2008 already seemed to be one for the record books. Altogether there were 148 tornadoes across the US in February, more than double the previous record for this month, set in 1971.

March seemed more normal, apart from an EF2 tornado hitting downtown Atlanta for the first time. Its 10-kilometre rampage racked up $200 million in damage and struck a CNN building, ensuring massive media coverage.

In May the drama continued. It is usually a very active month, but for this May the monthly count was a staggering 447 tornadoes. With half the year gone, tornadoes have already killed 119 people. That is nearly as many as in the whole of 1998, when 129 deaths made it the deadliest year since the 1970s.

Inevitably, in the face of such extreme weather, many are wondering what part global warming might have to play. Interviewed after the deadly Super Tuesday tornado outbreak, Senator John Kerry cautiously drew a connection (and was promptly criticised by the conservative right for doing so). Some climate bloggers have been far less hesitant in pointing the finger.

After all, it's not the first time global warming has been linked to tornadoes. The record number of US tornadoes in 2004 - 1819 - was mentioned by Al Gore in An Inconvenient Truth. So was he right to hint at a connection? Was Kerry?

There is no doubt that any increase in the strength or frequency of tornadoes could have dramatic implications. In the US, tornadoes kill 60 people per year and cause $400 million in damage, on average. It has been estimated that the US can expect a catastrophic $1 billion tornado once per decade.

One recent study concluded that if one of the most powerful tornadoes, rated EF4 or EF5, were to rip through Chicago, it could kill 13,000 to 45,000 people and rack up over $40 billion in damage. Other exposed major cities besides Atlanta include Oklahoma City, St Louis and Houston. We definitely don't want to raise the odds on these worst-case scenarios.

If a strong tornado ripped through Chicago, it could kill up to 45,000 people

Of course, the US is certainly not the only country at risk. Of the 10 deadliest tornadoes on record, just two occurred there. Six of them occurred in Bangladesh, including one in 1989, reported to have killed as many as 1300 people, which would make it the deadliest on record.

Determining the impact of ongoing climate change on weather events remains one of the knottiest aspects of global warming, and within this murky arena, the tornado issue is among the murkiest. It is entirely possible, for instance, that as the planet warms we could see an increase in severe thunderstorms but not in the kind that produce tornadoes. As politicians and bloggers stampede to link the 2008 tornado season to global warming, it remains questionable whether science can supply much justification.

The first question is whether these storms are changing in number or intensity. It is not an easy one to answer. While their size varies dramatically, in the US the average tornado is just 150 metres across and rakes along the ground for 8 kilometres. They can easily go unnoticed, especially in remote regions.

Even the severe thunderstorms that spawn tornadoes only measure tens of kilometres across and last for just a few hours. When it comes to such small-scale and relatively rare meteorological phenomena, historical records cannot be taken at face value. "Tornado reports, even more so than hurricane reports, are target-of-opportunity observations," says Harold Brooks of the National Severe Storms Laboratory in Norman, Oklahoma.

Over the past half-century the number of tornado reports in the US has grown rapidly, by an average of 14 per year, meaning that the total number of reports has doubled. No tornado expert thinks this huge rise reflects a real increase, though. Instead, it is due to more people living in tornado-prone regions, and to improved observation and reporting systems. These have led to the more frequent detection of the commonest and weakest tornadoes, rated EF0 and EF1 (see "More tornadoes?").

If you look only at EF4 and EF5 tornadoes, the number reported has, if anything, fallen slightly. But these numbers are also questionable. The original Fujita scale was only adopted in the 1970s, and retrospectively applied to earlier tornadoes based on written reports. It was based on damage to structures - a tornado that missed buildings would be classed as an F0 no matter what its wind speeds - yet did not fully take into account differences in the strength of buildings. In an attempt to solve such issues, the Enhanced Fujita scale was introduced last year.

The bottom line is that historical records of tornadoes do not currently support the idea that global warming is leading to more tornadoes, or more intense ones. This does not mean that global warming is not affecting tornadoes, just that we cannot say either way.

Harbinger of change

"Even if this year of tornadoes does set a record, one would not want to conclude it is a harbinger of climate change," says Anthony Del Genio of NASA's Goddard Institute for Space Studies in New York. "It's sort of like judging a baseball player who comes up from the minors and hits a home run at his first at bat. Is that a fluke, or is this guy going to be the next great baseball player?" The same applies to 2004: maybe there would not have been so many tornadoes without global warming, but there's no way to prove this.

It will be many years yet before any unambiguous trend can be discerned in the US tornado record. In the meantime, the only way researchers can begin to answer the question of whether climate change will affect the frequency or strength of tornadoes is to turn to theory and computer models. The trouble is that even the best global climate models divide the atmosphere up into cubes many tens of kilometres across. This means they cannot simulate thunderstorms, let alone smaller phenomena like tornadoes.

"There is insufficient evidence to determine whether trends exist in... small scale phenomena such as tornadoes, hail, lightning and dust storms," concluded the Nobel prizewinning Intergovernmental Panel on Climate Change (IPCC) in its most recent report, released early last year.

Since that report the science has inched forward just enough to supply talking points for eager activists, if not to provide definitive understanding. The latest research is based on a seminal 2003 study by Brooks and his colleagues, who sought to analyse the large-scale meteorological features that create suitable environments for severe thunderstorms not just in the US, but around the globe. Different criteria exist for defining such storms, but Brooks settled on thunderstorms that produce either large hail (greater than 5 centimetres in diameter), wind gusts over 120 kilometres per hour, or tornadoes of EF2 strength or greater.

Severe storms

Such storms depend on two factors in particular, Brooks's team concludes: the "convective available potential energy" and vertical wind shear. CAPE reflects the instability of the atmosphere, and is at its highest when there's very warm and humid air near the surface and very cool air aloft. Wind shear, which is what makes storms rotate, is the difference in wind speed or direction between the ground and 6 kilometres up. When CAPE and wind shear are both high, you get the sort of atmospheric conditions that can lead to tornadoes.

On this basis, Brooks identified environments across the globe that favour severe thunderstorms, including parts of Europe, equatorial Africa, Brazil, Argentina and areas south of the Himalayas. But far and away the most conducive areas were central and eastern parts of the US, particularly the Great Plains. Here the Rocky mountains create a very unstable atmosphere and warm, moist air sweeps in from the Gulf of Mexico.

Even though high levels of CAPE and wind shear are clearly conducive to tornadoes, these two factors alone do not guarantee that severe thunderstorms will form or that, once formed, they will spawn tornadoes. Only a third of severe thunderstorms produce tornadoes.

In fact, Brooks found, several other factors also determine whether tornadoes form. Not all of those are understood, but important ones include the altitude of the cloud base and the magnitude of the wind shear in the very lowest layer of the atmosphere, between the ground and 1 kilometre up.

Brooks's work left many unanswered questions, but as CAPE and wind shear are large enough features to be simulated by general climate models, it did suggest a way to use models to predict how global warming will affect small-scale events. Two studies published in 2007, too recent to be included in the latest instalment of the IPCC report, looked at how these basic factors might change as a result of global warming.

The first, by Del Genio and his colleagues, suggested that global warming would create stronger convective updraughts - related to CAPE - in severe thunderstorms, due to more low-level water vapour and warmer surface temperatures. This would also mean more lightning; the study predicts a 6 per cent rise in dangerous strikes over the western US.

In the central and eastern parts of the country, the team predicts that severe thunderstorms will not occur more frequently, and that any increase in updraft speed is likely to be offset by a decline in wind shear. However, when strong storms do occur, they could be even more intense.

What about tornadoes? Del Genio says it is "plausible" that their number or other attributes could change as well, but his study couldn't directly predict this. "Telling the difference between severe storms in general, and the particular ones that produce tornadoes, that's the most difficult thing to do," he adds.

Similar conclusions emerged from the second study, by Jeff Trapp of Purdue University, Indiana, in collaboration with Brooks and others. They found that global warming should increase CAPE but decrease overall wind shear. Higher CAPE more than makes up for lower wind shear, however: they predict that by the end of the century there will be an increase in the number of days in the US when conditions favour severe thunderstorms (see above).

For Trapp, as for Del Genio, there is still not enough information to make a firm prediction. "What we did is sort of the low hanging fruit," Trapp says. "Trying to separate out this frequency of severe thunderstorms from the frequency of tornadoes explicitly is much higher in the tree."

The weather, however, doesn't seem willing to wait. Del Genio's paper got little media attention when it was published. "It was kind of ignored last year, and I was a little bit surprised," he says. "And now this year, of course, it's going crazy."

With the media spotlight on them, Brooks worries that he and his fellow tornado experts might become as divided over global warming as hurricane specialists have become. It got ugly after Hurricane Katrina. Politicians and the media promptly pounced on studies suggesting global warming was making hurricanes worse, even though the science was still uncertain. Researchers themselves became polarised over the issue.

The hurricane connection

As if things were not complicated enough, the two issues are linked. Hurricanes can spawn tornadoes as they move over land, and more than 300 of the record 1819 tornadoes in 2004 were caused by hurricanes, with Frances and Ivan each spawning over 100. So more or stronger hurricanes worldwide will mean more tornadoes, unless something happens to reduce the number making landfall. However, the number of hurricanes making it ashore in the US is so low it will take another 50 years to detect any long-term change due to global warming.

Given such complexity, it's not surprising that tornado researchers are reluctant to jump to any conclusions on the global warming issue. But if we see another dramatic tornado strike on a major city, will the media be content with cautious answers? Brooks is not convinced. "I'm hoping all who are involved in the research side can at least talk to each other enough that, if we get misquoted in the political argument, we're aware of what's going on and can say, 'Wait, that's not what I said.'"

That may all depend on the weather.

The distinguishing feature of tornadoes is that they extend from the base of a cloud down to the ground, unlike minor whirlwinds such as dust devils. The strongest tornadoes are spawned by supercells - violent thunderstorms with a strong, rotating updraught (visible in the picture opposite).

Tornadoes vary in size from a metre or so across to several kilometres, and last for anything from seconds to hours. The "Tri-State Tornado" of 1925 is the longest-lasting tornado on record at over 3 hours, during which it travelled 350 kilometres across Missouri, Illinois and Indiana. It killed 695 people.

Tornadoes occur all over the world but conditions in the US's "Tornado Alley" favour supercell formation. As a result, the US has 1270 tornadoes each year, more than any other country; Canada is a distant second with around 100. By contrast, there are only 300 tornadoes per year in the whole of Europe. However, by land area the Netherlands has more tornadoes than any other country, followed by the UK.

• 
  From issue 2667 of New Scientist magazine, page 38-41.