“OH MY God, this is terrifying…I’ve got windows out…windows broken, trees down…the garage door is gone, the fence is down… I was in LA in an earthquake, this is worse…”
It was 6 am on 20 March and Gillian, a resident of Innisfail on the Queensland coast of Australia, was describing to ABC Radio the advance of tropical cyclone Larry. When the eye of the storm hit Innisfail an hour later, Larry was a maximum-intensity hurricane, with gusts of 290 kilometres per hour – only the third category 5 cyclone to hit Queensland since written records began in 1899.
A hundred kilometres north of Innisfail, Jon Nott’s home in Cairns was also battered by the edge of Larry. For Nott, director of the Australasian Palaeohazards Research Unit at James Cook University, the power of the storm came as no surprise. As a palaeotempestologist, he is part of a growing band of scientists worldwide dedicated to studying storms that occurred hundreds or even thousands of years before records began.
Advertisement
This might seem like an obscure academic exercise, but if you want to predict rare events such as cyclone Larry, a century of records is not nearly enough to go on. To spot long-term trends, you need to extend the record of cyclones – or hurricanes or typhoons, as these tropical storms are known in the Atlantic and north-west Pacific – back as far as possible.
Nott’s findings, for instance, suggest Queensland can expect a lot more Larrys. It appears that the state has been enjoying an unusual lull in cyclone activity that is now coming to an end.
This means the rapidly growing population of Queensland’s coast can expect further batterings this century even if the climate stays the same – which it will not. Global warming is giving the work of researchers like Nott a new significance and urgency. “Palaeotempestology is probably the best hope for really starting to understand links between climate and hurricane activity,” says Kerry Emanuel, a leading hurricane researcher at the Massachusetts Institute of Technology. And already, this young field is throwing up a few surprises.
So how do you study ancient cyclones? A passing hurricane may turn tropical forests into coleslaw but the wind won’t leave much of a lasting mark on the landscape. The storm surges that accompany major hurricanes, however, can dramatically alter landscapes, as the world saw with Katrina last year.
If a surge is big enough to flow over coastal sand dunes into a lake or marsh, it deposits a fan of sand on the bottom (see Diagram). When there are layers of organic matter above and below the sand layer, radiocarbon dating of the organic matter can reveal roughly when it was deposited – and so when the storm struck. In the US and Central America, where written hurricane records go back only 150 years, several teams have been using this technique.
Kam-biu Liu’s team at Louisiana State University in Baton Rouge has examined lake cores up to 5000 years old from a series of sites in the US. Western Lake on the Gulf coast of Florida, for instance, has been hit by 12 category 4 or 5 hurricanes in the past 3400 years. Yet almost all the strikes occurred more than 1000 years ago.
Why was hurricane activity so much higher at that time? An annual area of high pressure in the North Atlantic known as the Bermuda High helps determine whether hurricanes in the Caribbean head west into the Gulf or veer north, perhaps missing land altogether. Liu’s idea is that long-term changes in the position and strength of the Bermuda High account for his findings: there were not necessarily more hurricanes 1000 to 3400 years ago, it is just that more of them hit the Gulf coast.
“Our cores taken from Cape Cod and Virginia Beach have produced very promising results to support the hypothesis,” Liu says. Since major hurricanes are so rare, however, the team will need to study many sites to confirm the idea. Even if Liu is right, no one knows how climate change will affect the Bermuda High, and thus whether the Gulf could see a return to greater hurricane activity.
There are also limitations to the sediment method. Nott has looked at what happens to coastal dunes in front of lakes in Western Australia. “Some of the world’s very biggest cyclones have occurred here in the past 15 years,” he says. “They’ve been phenomenally big, much more intense than Katrina. I’ve seen every tree ripped out, every blade of grass ripped out or buried. And we’ve found that when a seawater inundation goes over a series of dunes it can wipe them out completely. It can totally obliterate them.”
“I’ve seen every tree ripped out, every blade of grass buried”
This means that if a second tropical storm strikes before a dune has grown back to its original height, even a very weak surge could dump just as much sand in a lake. To gauge how intense a storm was, researchers would have to establish the height of dunes at the time, which is difficult. Fortunately Australia offers another method. Where coral reefs are found close to shore, storm surges can pick up coral rubble and deposit it in a ridge on the shore. “It’s just the big events that get recorded in this fashion,” says Nott.
Ridges in exposed areas tend to get swept away by later storms, but on the leeward side of islands they can build up one by one. On Curacoa Island in the Great Barrier Reef, for instance, Nott has found a series of 22 ridges, the oldest deposited over 5000 years ago. Based on coral rubble studies, Nott reported in Nature in 2001 that over the past 5000 years, 10 times as many category 4 or 5 “super-cyclones” have smashed into the Great Barrier Reef as anyone believed. He concluded that a category 5 cyclone should be expected to hit any given location of the Queensland coast about every 250 years. “So the frequency of a cat 5 hitting anywhere on the coast is much higher,” Nott says. “And with a major lull for the past 100 years, Larry was overdue.”
Written in rubble
While coral ridges are a more reliable measure of the intensity of storms than lake sediments, both methods lack precision in dating. To really improve tropical cyclone models, researchers need to know the exact year in which each storm struck, not simply the half-century. The answer, surprisingly, lies in trees and stalagmites.
The rain that falls in a cyclone has a distinctive signature. The oxygen atoms in water molecules come in two isotopes: oxygen-16 and the much rarer, heavier oxygen-18. The very strong uplift in tropical storms and hurricanes separates these isotopes, skewing the normal ratio in the resulting rainfall. Initially the rain is richer in the heavier isotope, but by the time the storm sweeps over land it is richer in the lighter one. Because the rainfall during cyclones is so heavy, it is possible to detect the altered ratio of isotopes among the oxygen atoms incorporated into stalagmites or wood during a particular year.
For instance, Claudia Mora of the University of Tennessee in Knoxville has been able to build a hurricane record going back 227 years by studying isotope ratios in the rings of pine trees in the south-eastern US. Where the record overlaps with historical data, the frequencies match exactly.
In Australia, Nott has gone back even further. In a paper submitted to Earth and Planetary Science Letters, he reports a cyclone record based on analysis of stalagmites in the Chillagoe cave system in northern Queensland. Every year the stalagmites acquire two new layers: an ochre-coloured portion in the rainy season, followed by a white layer in the dry season. “We can go back thousands of years. I’ve got one stalagmite that’s probably three-and-a-half thousand years old,” Nott says.
Combining methods gives a much more accurate picture. For instance, coral rubble reveals that the last huge cyclone to rip through the area where Cairns now stands could have struck anytime between 1800 and 1850. Isotope studies of a 777-year-old stalagmite shows there was only one big cyclone between 1750 and 2006, allowing the exact year – 1821 – to be pinpointed.
The isotope analysis has also thrown up worrying evidence of a previously unknown centuries-scale cycle in cyclones in the region. The results suggest there are periods of high activity for a couple of centuries, followed by relative calm. As yet, neither Nott nor Emanuel know what might be driving the cycle. “It’s the first time we’ve been able to see it, because only now do we have long records at high resolution,” says Nott.
The more detailed record confirms Nott’s earlier finding that eastern Australia is now at the end of a period of low activity. “This is crystal ball stuff, but if we use the past patterns as a guide, we could start going into a higher centennial scale phase of activity very soon,” Nott says. “We’re due.”
Nott and Emanuel now hope to start analysing stalagmites in the US. Amy Frappier, a student at the University of New Hampshire in Durham, plans to conduct similar work in Belize. The findings should help clarify the effect of global warming on hurricane frequency and intensity.
According to a series of recent studies, including one by Emanuel’s team, the number of intense hurricanes has already increased, and this increase in intensity is due largely to the rise in sea surface temperatures (żěè¶ĚĘÓƵ, 3 December 2005, p 36). The link remains controversial. For instance William Gray of Colorado State University, a prominent hurricane forecaster, dismisses any link with climate change and instead blames the record number of Atlantic hurricanes in recent years on a cycle called the Atlantic Multi-Decadal Oscillation.
Emanuel thinks the link with the AMO is an illusion, an artefact of limited data and the methods used to analyse the figures. “There’s no evidence for anything cyclic in the record,” he says. The stalagmite studies could settle this crucial dispute once and for all.
The claimed global increase in intense cyclones is also controversial, with critics questioning the reliability of the data. The problem is that techniques for measuring intensity have changed dramatically over the past 30 years, says John McBride of the Bureau of Meteorology Research Centre in Melbourne, Australia. A storm that was identified as a category 2 in 1975 might have been recorded as a category 3 in 2005, for instance.
Hence the interest in palaeotempestology, in which this is not an issue. The ability to study storms that occurred 5000 years ago is another advantage. “Given the shortness of the instrumental record, the palaeo reconstructions are essential,” says Jeff Donnelly of the Woods Hole Oceanographic Institution in Massachusetts.
It is going to take many years of painstaking work around the world to reach firm conclusions about how climate change affects hurricanes. Already though, comparing Nott’s detailed record of cyclones hitting northern Queensland with other teams’ reconstructions of sea temperatures in the region has thrown up a surprise. The number of super-cyclones peaked between 1600 and 1800. Sea surface temperatures were higher than average between 1700 and 1800, but lower from 1600 to 1700. “So it would appear that high tropical cyclone activity was occurring during both higher and lower sea surface temperatures,” Nott says.
Total damage control
At first this might seem to contradict the idea that warmer seas mean more intense cyclones, but Nott points out that his record only includes cyclones that made landfall. Even if fewer intense cyclones formed in the Coral Sea off Queensland during the 1600s, there is evidence that trade winds were stronger at the time, which would have blown more of the cyclones that did form onto the coast – just as Liu thinks the variation in the number of intense hurricanes hitting the Gulf coast may be due to changes in their paths rather than in their frequency. “What it shows is that the steering wind environment may be an equally important factor in terms of the tropical cyclone hazard to coastal communities, which is what we are concerned about,” Nott says.
Put another way, even if global warming increases the overall number of intense hurricanes, some communities might be hit far less often. Others, of course, may have to endure many more super-cyclones than usual, so you would think coastal planners would take a keen interest. Yet while many insurance companies take the research seriously, funding some of Liu’s work for instance, politicians appear less interested. When Nott published his 2001 study showing that category 4 and 5 cyclones are 10 times as common along the north Queensland coast as thought, he expected planners in coastal towns such as Cairns to sit up and take notice.
“But Cairns city council went into total damage control,” Nott claims. “They actually set aside $15,000 in their budget to find someone to discredit my Nature paper.” A council spokeswoman confirmed that it allocated A$15,000 to pay for a “review”, but says the money was never spent, because the Queensland Department of Emergency Services decided not to alter its estimate of the frequencies of large storm surges, despite Nott’s findings.
Nott has also acted as an expert witness in a court case over a building development in an area that was flooded by 7 metres of water in 1918 when a category 4 cyclone hit. The developers claimed it was a once-in-10,000-years event. “They based this calculation on 40 years of historical data, a period in which very few cyclones crossed the Queensland coast,” says Nott. “I demonstrated to the court that these sorts of cyclones in fact strike much more frequently. The judge said I won the scientific debate, but that it was up to the politicians to set standards – but they won’t, because they fear litigation by developers.”
Tallying typhoons
Coral rubble ridges and lake sediments are not the only way to find out about past storms. Historical documents can also reveal a lot. Shipping logs have proved a rich source of information about Atlantic storms, but their use is mostly limited to improving the existing database of hurricanes going back to about 1850 (żěè¶ĚĘÓƵ, 27 September 2003, p 14).
In China, records go back much further. The earliest record of a typhoon, found in 2003, dates from AD 470. In 2001, Kam-biu Liu’s team at Louisiana State University in Baton Rouge published a 1000-year history of typhoon landfalls in Guangdong province, southern China, based on reports in semi-official local gazettes. Liu was surprised to find that periods of high typhoon activity coincided with cold periods – possibly because storm tracks shifted south, with Japan getting fewer strikes and China suffering more.
