Читать книгу Six Degrees: Our Future on a Hotter Planet - Mark Lynas - Страница 8

America's slumbering desert

It would be easy to walk right past them. Not many hikers pass this way, and those that do are unlikely to give a second thought to a few old stumps rooted in the river bed. In any case, this lonely spot, where the West Walker River canyon is at its narrowest as it plunges down the eastern flanks of California's Sierra Nevada, is not a place to linger-the area is notorious for sudden downpours and flash floods. The river runs almost the width of the entire gorge, and there's no place to climb to safety if the heavens open.

But these stumps have a story to tell. Dead trees can talk, in a way. An astute hiker or an observant angler would be puzzled: what are they doing in a river bed, a place now treeless because of the constant flowing water? Investigated by scientists in the early 1990s, the tree stumps were found to be Jeffrey pines-a common enough species for the area, but one that certainly doesn't normally root in rivers. What's more, these trees were old. Very old. Tissue samples revealed that the stumps dated from medieval times, and grew during two specific periods, centred on AD 1112 and 1350.

The mystery deepened when similar old stumps were revealed in Mono Lake, a large saltwater body a hundred kilometres to the south of Walker River, near the state border with Nevada. It's a spectacular location, famous for broad skies and sunsets, with little to interrupt the gently rolling arid landscape other than a few extinct volcanoes. The Mono Lake tree stumps belonged not just to pines, but also to other native species like cottonwoods and sagebrush, all rooted far below current-day natural lake levels and only revealed thanks to water diversion projects that supply far-away Los Angeles. Again, carbon dating revealed the same two time intervals as for the Walker River trees. Clearly, something significant had happened back in medieval times.

More evidence came from deeper in the mountains, hidden in two locations today famous for their giant sequoia groves-Yosemite and Giant Sequoia National Parks. These enormous trees, which in terms of total wood volume stand as the largest living organisms on Earth, are also among the oldest. Some living trees are up to 3,000 years old. And because each annual growth cycle leaves a clear ring, these monumental plants are also an excellent record of past climate. Over a decade ago, scientists sampling wood from dead giant sequoias noticed old fire scars on the edges of some of their rings. These scars were especially frequent during this same medieval period-between AD 1000 and 1300-as the old trees in West Walker River and Mono Lake were growing. Wildfires had raged in both national parks twice as frequently as before, and there can only be one plausible explanation-the woods were tinder-dry.

Raging wildfires, dry rivers and lakes-the pieces of the jigsaw were beginning to make sense. The area we now call California had in medieval times been hit by a mega-drought, lasting at different periods for several decades, and altering both landscape and ecosystems on a scale that dwarfs today's drought episodes. But just how geographically widespread was this event? Evidence from another lake, far away on the Great Plains of North Dakota, provides a partial answer. Moon Lake, like Mono Lake in California, is a closed basin, making it saline. Salinity fluctuates with the climate: in sequences of wet years, more fresh water ends up in the lake and salt levels go down. The converse is also true: in dry years, more water evaporates, leaving a more concentrated salty brine behind. Canadian scientists have now reconstructed long-term records of Moon Lake's saltiness by sampling the remains of tiny algae called diatoms-whose type and number fluctuate with salinity levels-from old lake sediments. Lo and behold, back before AD 1200, a series of epic droughts had swept the Great Plains, the return of which-the scientists agreed-‘would be devastating’.

An insight into the devastating nature of such a drought was gained by a team of biologists working in northern Yellowstone National Park, a good 1,500 kilometres to the south-west of Moon Lake, in Wyoming. They drilled into sediments spilled out by rivers, only to discover a peak in muddy debris flows-the product of flash floods-about 750 years ago. These flash floods had poured off mountainsides denuded of forest cover by frequent fires: so rather oddly, these flood debris flows are actually a classic sign of drought. It appeared that the whole of the western United States had been struck at the same time.

The effect on Native American populations in this pre-Columbian era was indeed devastating. Whole civilisations collapsed, beginning in the Chaco Canyon area of modern-day New Mexico. One of the most advanced societies on the continent at their peak, the Pueblo Indian inhabitants of Chaco Canyon erected the largest stone building on the North American continent before the European invasion, a ‘great house’ four storeys high, with over 600 individual rooms-much of it still standing today. Yet when the big drought came in AD 1130, they were vulnerable-population growth had already diminished the society's ecological base through the overuse of forests and agricultural land. Most people died, whilst the survivors went on to eke out a living in easily defended sites on the tops of steep cliffs. Several locations show evidence of violent conflict-including skulls with cut marks from scalping, skeletons with arrowheads inside the body cavity, and teeth marks from cannibalism.

Indeed, the whole world saw a changing climate in medieval times. The era is commonly termed the ‘Medieval Warm Period’, a time when-so the oft-told story goes-the Vikings colonised Greenland and vineyards flourished in the north of England. Temperatures in the North American interior may have been 1 to 2°C warmer than today, but the idea of a significantly warmer world in the Middle Ages is actually false. Recent research piecing together ‘proxy data’ evidence from corals, ice cores and tree rings across the northern hemisphere demonstrates a much more complicated picture, with the tropics even slightly cooler than now, and different regions warming and then cooling at different times. However small the global shift, the evidence is now overwhelming that what the western US suffered during this period was not a short-term rainfall deficit, but a full-scale mega-drought lasting many decades at least. As recently as 2007 US scientists reported tree-ring studies reconstructing medieval flows in the Colorado River at Lees Ferry, Arizona, showing that the river lost 15 per cent of its water during a major drought in the mid-1100s. For sixty years at a time, the river saw nothing but low flows-none of the floods that normally course down the Colorado arrived to break the dry spell. Indeed, the remarkable coincidence of these dates with evidence from New Mexico suggests that this was the very same drought that finished off the Chaco Canyon Indians.

To see the worst that even such a small change in climate can do, consider that most undramatic of places-Nebraska. This isn't a state that is high up on most tourists' ‘to do’ lists. ‘Hell, I thought I was dead too. Turns out I was just in Nebraska,’ says Gene Hackman in the film Unforgiven. A dreary expanse of impossibly flat plains, Nebraska's main claim to fame is that it is the only American state to have a unicameral legislature. Nebraska is also apparently where the old West begins-local legend in the state capital Lincoln insists that the West begins precisely at the intersection of 13th and O Streets, a spot marked by a red brick star.

But perhaps the most important Nebraska fact is that it sits in the middle of one of the most productive agricultural systems on Earth. Beef and corn dominate the economy, and the Sand Hills region in central Nebraska sports some of the most successful cattle ranching areas in the entire United States.

To the casual visitor, the Sand Hills look green and grassy, and in pre-European times they supported tremendous herds of bison-hence their high productivity for modern-day beef. But, as their name suggests, scratch down a few centimetres and the shallow soil quickly gives way to something rather more ominous: sand. These innocuous-looking hills were once a desert, part of an immense system of sand dunes that spread across thousands of kilometres of the Great Plains, from Texas and Oklahoma in the south, right through Kansas, Colorado, Wyoming, North and South Dakota, to as far north as the Canadian prairie states of Saskatchewan and Manitoba. These sand dune systems are currently ‘stabilised’: covered by a protective layer of vegetation, so not even the strongest winds can shift them. But during the Medieval Warm Period, when temperatures in the Great Plains region may only have been slightly warmer than now, these deserts came alive-and began to march across a fertile landscape which today is a crucial food basket for humanity. This historical evidence indeed suggests that even tiny changes in temperature could tip this whole region back into a hyper-arid state.

People who remember the 1930s Dust Bowl might think they have seen the worst drought nature can offer. In the toughest Dust Bowl years, between 1934 and 1940, millions of acres of Great Plains topsoil blew away in colossal dust storms. One, in May 1934, reached all the way to Chicago, dumping red snow on New England. Hundreds of thousands of people, including 85 per cent of Oklahoma's entire population, left the land and trekked west. All this took only an average 25 per cent reduction in rainfall-enough for ploughed farmland to blow away, but the giant dunes stayed put. What awoke the dunes from their long slumber nearly a thousand years ago was drought on an altogether different scale-with dramatically less rainfall, sustained over decades rather than just years.

In a world which is less than a degree warmer overall, the western United States could once again be plagued by perennial droughts-devastating agriculture and driving out human inhabitants on a scale far larger than the 1930s calamity. Although heavier irrigation might stave off the worst for a while, many of the largest aquifers of fossil water are already overexploited by industrialised agriculture and will not survive for long. As powerful dust and sandstorms turn day into night across thousands of miles of former prairie, farmsteads, roads and even entire towns will find themselves engulfed by blowing sand. New dunes will rise up in places where cattle once grazed and fields of corn once grew. For farmers, there may be little choice other than to abandon agriculture completely over millions of square kilometres of what was once highly productive agricultural land. Food prices internationally would rise, particularly if serious droughts hit other areas simultaneously. And although more southerly parts of the United States are expected to get wetter as the North American monsoon intensifies, residents here may not welcome an influx of several million new people.

Further east, however, agriculture may actually benefit from warmer temperatures and higher rainfall. Rather as California offered sanctuary of a sort to displaced Okies' during the Dust Bowl, the Midwest and Great Lakes areas will need to provide jobs and sustenance to those who can no longer scratch a living from the sandy soils far out west, once the rains stop falling and the desert winds begin to blow.

Already the day after tomorrow?

Just as farmers on the High Plains of North America are watching their fields and grasslands blowing away in the relentless heat, their kinfolk across the Atlantic may be grappling with another problem: extreme cold. One of the most counter-intuitive projected impacts of global warming is the possible plunging of temperatures throughout north-west Europe as the warm Atlantic current popularly known as the Gulf Stream stutters and slows down. This is the scenario fictionalised in an exaggerated form by the Hollywood disaster epic The Day After Tomorrow, where a collapse in the Atlantic current triggers a new ice age, flash-freezing New York and London (although the good guy still gets the girl). Real-world scientists were quick to lambast the film for flouting the laws of thermodynamics, but they also acknowledged that the reality of a slowdown in the North Atlantic Ocean current may still be pretty scary, especially for those who live in a part of the world which is used to a mild maritime climate far out of keeping with its high northern latitude.

A short technical aside is required here. Only a small part of the great current that delivers warm water into the North Atlantic is actually the real Gulf Stream: it, as its name suggests, is a stream of warm subtropical water heading north-east out of the Gulf of Mexico, which eventually becomes part of the much larger system of currents known to scientists as the Atlantic Meridional Overturning Circulation. The MOC is partly driven by the cooling and sinking of water at high latitudes off the coast of Greenland and Norway, where freezing Arctic air lowers its temperature and squeezes fresh water out as sea ice, leaving behind a heavy, salty brine which quickly sinks to the bottom of the ocean. From there it begins a return journey south-eventually surfacing (1,200 years later) in the Pacific. Scientists have long feared that a freshening and warming of the Norwegian and Greenland seas-due to higher rainfall, run-off from melting land glaciers and the disappearance of sea ice-could stop this water sinking, and shut down the great ocean conveyor. Hence the famous ‘Shutdown of the Gulf Stream’ scenarios familiar from newspaper headlines and the Hollywood movie.

Far-fetched it may seem, but Atlantic circulation shutdown has always been more than just a theory. It has happened before. At the end of the last ice age, 12,000 years ago, just as the world was warming up, temperatures suddenly plunged for over a thousand years. Glaciers expanded again, and newly established forests gave way once more to chilly tundra. The period is named the ‘Younger Dryas’, after an arctic-alpine flowering herb, Dryas octopetala, whose pollen is ubiquitous in peaty sediment layers dating from the time. In Norway temperatures were 7-9°C lower than today, and even southern Europe suffered a reversal to near-glacial conditions. On the other side of the Atlantic, cooling also occurred, and there is evidence of rapid climate change from as far afield as South America and New Zealand.

The culprit seems to be the sudden shutting-off of the Atlantic circulation due to the bursting of a natural dam holding back Lake Agassiz, a gigantic meltwater lake which had pooled up behind the retreating North American ice sheets. When the dam broke, an enormous surge of water (the lake's volume was equivalent to seven times today's Great Lakes) is thought to have poured through Hudson Bay and out into the Atlantic. This freshwater surge diluted the North Atlantic seas and stopped them being salty enough to sink, interrupting the deep ocean current and triggering climatic destabilisation across the world.

Obviously today there are no gigantic ice lakes waiting to flood into the North Atlantic, but global warming could still interrupt the formation of deep water by melting sea ice and causing greater freshwater run-off from Siberian rivers. Despite the rapidly melting ice cap, however, for many years there was no evidence that changes in the Atlantic MOC were actually happening, and many oceanographers had begun to pooh-pooh the theory. That was until the RSS Discovery, a scientific research vessel owned by the British government, began a routine cruise across the Atlantic in 2004. The ship's scientific team set themselves the task of sampling seawater at various depths on a line drawn between the Canary Islands in the east and Florida in the west, aiming to repeat similar measurements taken in 1957, 1981, 1992 and 1998. They had not expected to discover anything terribly exciting; in fact the team leader Professor Harry Bryden confided to one journalist: ‘In 1998 we saw only very small changes. I was about to give up on the problem.’

But 2004 was different. Bryden and his colleagues found that less warm water was flowing north at the surface and less cold water was flowing south at depth. Overall, the Atlantic circulation had dropped by 30 per cent, equivalent to the loss of 6 million tonnes of water flow per second. No wonder Professor Bryden admitted that he was ‘surprised’. Suddenly the slowing-down of the great Atlantic current system was no longer just a hypothesis postulated for the distant future. It was already happening.

The media reaction was instantaneous. ‘Current that warms Europe weakening’, warned CNN. NPR's All Things Considered show led with Atlantic Ocean's heat engine chills down'. In Europe, the response was one of understandable concern. Alarm over dramatic weakening of Gulf Stream', reported the UK's Guardian newspaper on 1 December 2005. ‘Global warming will bring cooler climate for the UK’ was the Telegraph's take on the same story. A couple of paragraphs down, the paper reported one expert as confirming that ‘an average temperature drop of a degree or two within decades would herald more extreme winters’.

Older readers would have shuddered at the thought of a return to winters as bitter as that of 1962-3, when the UK was blanketed in snow for more than three months, and temperatures hit a low of–16°C in southern England. In places the sea froze, and ice floes appeared in the river Thames at London's Tower Bridge. That season was about 2.7°C colder than average-almost exactly the temperature drop predicted for London in one modelling study investigating the possible result of a 50 per cent drop in the warm Atlantic current. Was Europe's new ice age just around the corner?

Apparently not. Almost exactly a year later, and with much less fanfare, Science magazine reported that ‘a closer look at the Atlantic Ocean's currents has confirmed what many oceanographers suspected all along: there's no sign that the ocean's heat-laden “conveyor” is slowing’. Instead of just the snapshot data generated by a few irregular ship cruises, nineteen permanent instrument-laden sensors had now been stretched across the Atlantic between West Africa and the Bahamas-and they were able to deliver a much more consistent picture. A year of continuous monitoring, Harry Bryden now reported to a conference in Birmingham, showed that his original 30 per cent decline was just a part of random natural variability after all, the sort of thing that happens constantly from one year to the next.

This result was a triumph for the modellers, most of whom had for years been pouring cold water on the European ice age theory. They agreed that huge volumes of freshwater would need to surge into the North Atlantic in order to shut off the Gulf Stream-far more than currently being generated by melt from Greenland or higher precipitation in Siberia. Rather than plunging overnight, the ocean circulation might decline by a stately 25 to 30 per cent or so, but only after at least 100 years of sustained greenhouse gas emissions. Even then, it wouldn't cool Europe-it would simply moderate the otherwise rapid rise in temperatures.

As the IPCC concluded in 2007: ‘it is … very unlikely that the MOC [Atlantic Meridional Overturning Circulation] will undergo a large abrupt transition during the course of the 21st century’. Although all of them showed some weakening by 2100, none of the models assessed by the IPCC supported the collapse scenario. And even with this MOC slowdown, the IPCC reported that ‘there is still warming of surface temperatures around the North Atlantic Ocean and Europe due to the much larger effects of the increase in greenhouse gases’. The IPCC's judgement was final: there would be no new ice age for Europe.

Africa's shining mountain

The amateur adventurer Dr Vince Keipper had waited years for this day. Nearing the summit of Kilimanjaro, the highest point on the African continent, Keipper and his group were looking forward to panoramic views of the surrounding Kenyan and Tanzanian plains. They had climbed through the steep and treacherous Western Breach and past the towering ice cliffs of the Furtwängler Glacier. The weather was perfect, with only a few clouds far beneath. Then, not far from the top of the 5,895-metre peak, a loud rumbling sound from behind them brought the group to a sudden halt. ‘We turned around to see the ice mass collapse with a roar,’ remembered Keipper. ‘A section of the glacier crumbled in the middle, and chunks of ice as big as rooms spilled out on the crater floor.’ Keipper and his group knew they had had a lucky escape: they might have been buried had the collapse happened only a few hours earlier. They also knew that the event they had just witnessed had a powerful symbolic resonance: right in front of their eyes, the highest peak in Africa was melting.

Kilimanjaro has become something of a poster child for the international climate change campaign. The Swahili words kilima and njaro translate as ‘shining mountain’, testament to the power of this massive volcano to inspire awe in onlookers through the ages. A recent aerial photo of the crater, with little more than a few ice fragments encrusting its dark sides, was the centrepiece for a touring global warming photography exhibition sponsored by the British Council in 2005. During the 2001 UN climate change conference in Marrakech, Morocco, Greenpeace sent a team to Kilimanjaro to hold a press conference by video link from beside one of the mountain's disappearing glaciers. Kilimanjaro's international celebrity status has also attracted the attention of climate change deniers, who suggest that deforestation on the mountain's lower slopes is more to blame for glacial retreat than global warming.

None of the contrarian rhetoric cuts any ice, so to speak, with Lonnie Thompson, a glaciologist at Ohio State University and a man who is deservedly one of America's most celebrated natural scientists. Thompson pioneered the drilling of ice cores in inaccessible mountain regions, bringing back ice tens of thousands of years old from glaciated peaks as remote and far apart as Peru's Nevado Huascarán and Tibet's Dasuopu, often pushing himself to the edge of human endurance in the process. In 1993 Thompson and his drilling team camped for 53 days at 6,000 metres between the two peaks of Huascarán, perhaps setting a world record for high-altitude living. (I stayed there for one night in 2002-one of the most freezing, wind-blasted and wretched nights of my life.) At one point a gale blew Thompson's tent, with him inside, towards a precipice-until he jammed his ice axe through the floor. ‘I don't understand,’ he once remarked, ‘why anyone would want to climb a mountain for fun.’

As Thompson was one of the first to recognise, this mountain ice contains a unique record of climate variations down the ages-preserved in layers of dust, isotopes of oxygen and tiny bubbles of gas trapped within the frozen layers of water. Once carried down in freezer boxes and analysed in the laboratory, these icy signatures trace everything from droughts to volcanic eruptions from decades and centuries past. They also tell a story about past temperature changes: the two isotopes of oxygen, ¹⁶O and ¹⁸O (which have different atomic weights due to the presence of two more neutrons in the latter's nucleus), vary in abundance with water temperature, so their proportions in ice cores are a good ‘proxy’ record of ancient climates.

Thompson and his team also drilled on three of Kilimanjaro's remaining glaciated areas, and in October 2002 concluded that 80 per cent of the mountain's ice had already melted during the past century. The news made international headlines, along with Thompson's prediction that the rest of the ice would be gone by between 2015 and 2020. As he readily admitted, this prediction was not based on complex computer modelling or any other advanced techniques. ‘In 1912 there were 12.1 square kilometres of ice on the mountain,’ he told journalists from CNN. ‘When we photographed the mountain in February of 2000, we were down to 2.2 square kilometres. If you look at the area of decrease, it's linear. And you just project that into the future, sometime around 2015 the ice will disappear off Kilimanjaro.’

If there was an urgency in Thompson's voice, this was because he knew that recent melting had already begun to destroy the unique record of past climate preserved in Kilimanjaro's glaciers.

In their analysis of dust layers in the ice, the scientific team found evidence of a marked 300-year drought four thousand years ago; a drying so severe that it has been linked to the collapse of several Old World civilisations across North Africa and the Middle East. The ice also indicated much wetter conditions even longer ago, when huge lakes washed over what is now Africa's dry Sahel. Close to the surface Thompson's team discovered ice containing a layer of the radionuclide chlorine-36, fallout from the American ‘Ivy’ thermonuclear bomb test on Eniwetok Atoll in 1952. With this precise time control, the scientists could tell that ice which would have preserved a record of climate fluctuations since the 1960s had already melted away.

Moreover, the oldest ice at the base of the cores proved to be over 11,000 years old, showing that at no time since the last glacial epoch has the peak of Kilimanjaro been free of ice. This discovery made Thompson's ice cores even more valuable, for the simple reason that within as little as ten years the sawn-up circular cores in Ohio State University's walk-in freezer will be the only Kilimanjaro ice left anywhere in the world. With this in mind, Thompson and his team have already decided that some of the ice will be kept intact for future generations of scientists to dissect with new technologies, possibly unlocking climatic secrets still undreamt of today.

The efforts of climate change deniers to suggest that there is something special about the disappearance of Kilimanjaro's glaciers are undermined by similar changes taking place in mountain ranges right across the world, not least in the Rwenzori Mountains of Uganda, nearly a thousand kilometres to the north-west. In this remote region, where Uganda borders the Democratic Republic of the Congo, the fabled ‘Mountains of the Moon’ generate such heavy rainfall (about 5 metres per year) that the cloud-shrouded peaks are only visible on a few days out of every year, and form the main headwaters of the river Nile. At the top of the highest peak, the 5,109-metre Mount Stanley (named after the explorer, who passed by in 1887), ice and snow deny the summit to all but the most determined mountaineers. Yet as at Kilimanjaro, glacial retreat in the Rwenzoris has been profound: the three highest peaks have lost half their glacial area since 1987, and all the glaciers are expected to be gone within the next two decades.

Elsewhere in the world, disappearing mountain glaciers pose a major threat to downstream water supplies. But Kilimanjaro's ice cap is so small that its final disappearance will make little difference to the two major rivers-the Pangani and the Galana-which rise on its flanks. Instead, the crucial water link for Kilimanjaro is not the glaciers, but the forests. The montane forest belt at between 1,600 and 3,100 metres provides 96 per cent of the water coming from the mountain-this lush tangle of trees, ferns and shrubs not only captures Kilimanjaro's torrential rainfall like a giant sponge, but also traps moisture from the clouds which drape themselves almost permanently around the mountain's middle slopes. Much of this water drains underground through porous volcanic ash and lavas, and emerges in waterholes-vital for local people as well as for wild animals-far away on the savannah plains.

So is Kilimanjaro's water-generating capacity safe from global warming? Not quite: rising temperatures and diminishing rainfall increase the risk of fires, which have already begun to consume the upper reaches of montane forest. By the time the glaciers have disappeared, so will the higher forests, depriving downstream rivers of 15 million cubic metres of run-off every year, according to one estimate. In contrast, the loss of glacial water input will likely add up to less than 1 million cubic metres annually: significant, but not catastrophic. The diminishing water supply will affect everything from fish stocks to hydroelectric production downriver in poverty-stricken Tanzania. Much of the mountain's world-famous biodiversity (Kilimanjaro hosts twenty-four different species of antelope alone) will also be threatened by the weather changes.

As the snows disappear, so will much of the wildlife and the verdant forests that tourists currently trek through on their arduous journey to the roof of the African continent.

Ghost rivers of the Sahara

Far to the north of Kilimanjaro, in the Sahel, another drought-stricken area could by this time be experiencing some blessed relief. The Sahelian region of North Africa has long been synonymous with climatic disaster: during the 1970s and 80s famines struck the area with such severity that they sparked massive humanitarian relief efforts like Band Aid and Live Aid. Reporting from Ethiopia's refugee camps in 1984, the BBC's Michael Buerk spoke of a ‘biblical famine’ as the camera swept slowly over the dead and dying. Over 300,000 people perished during earlier famines in the 1970s.

The Sahel is an immense area, stretching in a wide belt east to west across northern Africa from Senegal on the Atlantic coast to Somalia on the Indian Ocean. For the most part savannah and thorn scrub, it is a climatic transition zone between the hyper-arid Sahara to the north and the lush tropical forests which grow nearer to the equator in the south. Intermittent rains mean that nomadic cattle herding has long been a dominant way of life, with people wandering far and wide through the seasons in search of grazing for their livestock. It is often assumed that global warming will further desiccate the Sahel, allowing the Saharan dunes to march south into Nigeria and Ghana, and displacing millions in the process. Although the forecasts are tentative and uncertain, both palaeoclimatic studies and computer models suggest that the reverse might be true. As other parts of Africa shrivel in the heat, could the Sahel end up as a refuge?

For clues to how the area's climate might alter, we need to venture north into the great Sahara. Here, the world's largest desert has also seen the highest temperature ever recorded on Earth: a truly blistering 58°C. The Sahara covers an area so huge that the entire contiguous United States would comfortably fit inside. This desert doesn't just have sand dunes, it has sand mountains, some reaching to nearly 400 metres in height. It is so completely uninhabitable that only a sprinkling of people get by in a few dwindling oases and at the desert's edge.

But scattered over this enormous area are clear signs that a very different Sahara existed many thousands of years ago. Neolithic paintings and rock carvings have been discovered in places where settled human existence is utterly impossible today. This ancient art depicts elephants, rhinoceroses, giraffes, gazelles and even buffalo-all animals which currently roam only hundreds of kilometres to the south. In Egypt's hyper-arid Western Desert, where less than 5 mm of rain falls on average each year, arrowheads and flint knives for hunting and butchering big game have been unearthed by archaeologists. At one site in south-western Libya, archaeologists even discovered tiny flint fish-hooks-again in an area where no trace of surface water persists now.

Other indications of a wetter past have also been discovered. Although anyone crossing Egypt's dry Safsaf Oasis by camel would today see little more than rock and dunes, radar pictures taken from the space shuttle Endeavour in 1994 clearly show whole river valleys buried beneath the sands. These ghostly watercourses even include major tributaries to the Nile flowing out through modern-day Sudan, all long-dry and forgotten beneath the dust. In southern Algeria, huge shallow lakes once gathered, supporting plentiful populations of fish, birds and even Nile crocodiles. The carbon dating of freshwater snails and desiccated vegetation preserved in these old lake beds shows that between five and ten thousand years ago the desert edge retreated 500 kilometres further north, and at different times almost disappeared altogether.

On the borders of what is today Chad, Nigeria and Cameroon, an immense lake, over 350,000 square kilometres in area, extended across the southern Sahara. Nicknamed Lake Mega-Chad, after its modern-day remnant Lake Chad, this gigantic inland sea was the largest freshwater body that Africa had seen for the last two and a half million years. It would have been only slightly smaller than today's largest lake, the Caspian Sea. Strange ridges of sand, which today lie marooned far away in the desert, reveal the shores of the old lake, as do the shells of long-dead molluscs which once thrived in its warm, shallow waters. The flat landscape between the marching dunes testifies to the erosive power of its long-vanished waves.

Common sense suggests that a major lake in such an arid area could only have been maintained by much higher rainfall, and longer-term records do indeed show that the Saharan region has experienced repeated wet and dry episodes over cycles of many thousands of years. The coldest periods of the ice ages tended to be the driest in the Sahara, whilst warm interglacials brought rain-allowing life to emerge once again. During the early Holocene epoch, 9,000 to 6,000 years ago, the northern hemisphere summer sun was slightly stronger than today, thanks to a small cyclical shift in the Earth's orbit around the sun. The increased heating warmed up the giant North African landmass to such an extent that it powered a monsoon-just like the one that brings annual summer rains to the Indian subcontinent today.

Monsoons are based on the simple principle that land surfaces heat up quicker in the summer than the surrounding oceans. This creates an area of low pressure as the hot air in the continental interior rises, sucking in cooler, moister air from the neighbouring seas. These rain-bearing winds bring torrential summer downpours to monsoonal climates such as India's, where agricultural life revolves with this annual cycle. The African summer monsoon is weaker and less generally recognised, but is still the only source of reliable rainfall for the Sahel. Climate models project that land surfaces will warm much faster than the oceans during the twenty-first century, potentially adding a boost to summer monsoons. So with one degree of global warming, this monsoon could begin to gain power and penetrate once again far into the African continent, greening the Sahara.

But will it actually happen? Before anyone makes plans to move large-scale food production to the central Sahara, a note of caution needs to be sounded. During the early Holocene, an additional monsoon driver was the difference in the distribution of solar heat between the two hemispheres. This time the whole globe is heating up, so the past is not a perfect analogue for the future. Moreover, it would be wrong to get the impression that the more humid Sahara was some kind of verdant wonderland-rainfall totals mostly only reached 100 mm or so, enough to support only the barest savannah-type vegetation, and wetter phases would also have been interspersed with long droughts. However, computer models can help negotiate a way through the conflicting possibilities-and the answer they provide holds profound implications for all the inhabitants of North Africa.

The preliminary stage is set by Martin Hoerling and two other climate scientists based in Boulder, Colorado, who used sixty different model runs to confirm that whilst southern Africa dries out with global warming, northern Africa does indeed begin to get wetter. Indeed, the long-term drying trend which caused such misery and devastation during the second half of the twentieth century goes into full reversal after about 2020 (with one-degree global warming or less), when the Sahel sees a long-term recovery in its rainfall. By 2050 the recovery is in full swing, with 10 per cent more rainfall right across the sub-Saharan zone.

This conclusion is supported by a second study, which projects heavier rains on both the West African coast and into the Sahel as a warmer tropical Atlantic Ocean supplies huge amounts of water vapour to form rain-bearing clouds. With more plentiful rains, crop production can potentially increase, offsetting declines elsewhere-assuming, that is, that temperatures are not so high that people who once died from famine now die from heatstroke.

However, computer modellers based in Princeton, New Jersey, have come up with a rather different long-range forecast. Their model accurately simulates the terrible 1970s and 1980s drought-but after a short interlude of higher rainfall, it projects even fiercer drought conditions for the Sahel region in the second half of the twenty-first century.

So why the divergence between the different models produced by the Princeton and the Boulder teams? The Princeton researchers admit that they are stumped. ‘Until we better understand which aspects of the models account for the different responses in this region,’ they caution, ‘we advise against basing assessments of future climate change in the Sahel on the results of any single model.’ Nevertheless, they insist, ‘a dramatic 21st century drying trend should be considered seriously as a possible future scenario’.

This latter finding also chimes with global studies, which suggest stronger droughts affecting ever-larger areas as the world warms up. One of the most wide-ranging analyses was undertaken by Eleanor Burke and colleagues from the Hadley Centre at Britain's Meteorological Office, who used a measure known as the ‘Palmer Drought Severity Index’ to forecast the likely incidence of drought over the century to come. The results were deeply troubling. The incidence of moderate drought doubled by 2100-but worst of all, the figure for extreme drought (currently 3 per cent of the planet's land surface) rose to 30 per cent. In essence, a third of the land surface of the globe would be largely devoid of fresh water and therefore no longer habitable to humans.

Although these figures are based on global warming rates of higher than one degree by 2100, they do indicate the likely direction of change. As the land surface heats up, it dries out because of faster evaporation. Vegetation shrivels, and when heavy rainfall does arrive, it simply washes away what remains of the topsoil. It may seem strange that floods and droughts can be forecast to affect the same areas, but with a higher proportion of rainfall coming in heavier bursts, longer dry spells will affect the land in between. This, then, is the most likely forecast for the Sahel: whilst rainfall totals overall may indeed rise, these increases will come in damaging flash-flood rainfall, interspersed with periods of intensely hot drought conditions.

According to some historians, the greener Sahara of 6,000 years ago was the geographical basis for the mythical Garden of Eden, its original inhabitants expelled not by God for bad behaviour, but by a devastating drying of the climate. Whilst scientists continue to argue over the specifics of the likely climatic future of the Sahara and Sahel, one thing seems clear: humanity will not be returning to Eden any time soon.

The Arctic meltdown begins

Over recent years a new phrase has entered the scientific lexicon: ‘the tipping point’. Originally popularised by Malcolm Gladwell's bestselling book of the same name, the understanding that social or natural systems can be non-linear is a crucially important one. An oft-used analogy is of a canoe on a lake: wobble it a little, and stability can return with the boat still upright. Cross the point of no return, the ‘tipping point’, and the boat will capsize and find a new stability-this time upside down, with the ill-advised canoeist floundering underneath.

Scientists have increasingly realised that the Earth's climate is a good example of a non-linear system: over the ages it has been stable in many different states, some much hotter or colder than today. During the ice ages, for example, global temperatures averaged five degrees cooler than now for tens of millennia. Moreover, the system can ‘tip’ from one state into another with surprising rapidity. Episodic sudden warmings embedded within the last ice age saw temperatures in Greenland rise by as much as 16°C within just a few decades. The reasons why the climate flipped so rapidly are still not completely understood, but it is clear that even tiny changes in ‘forcings’-from greenhouse gases or the Sun's heat-have in the past led to dramatic responses in the climate system. In contrast, our relatively stable climate is highly unusual-the Holocene period, during which all of human civilisation has come about, has seen very little change in global temperatures. Until now.

Scientists have established beyond reasonable doubt that the current episode of global warming, of about 0.7°C in the last century, has pushed Earth temperatures up to levels unprecedented in recent history. The IPCC's 2007 report confirmed that no ‘proxy records’ of temperature-whether from tree rings, ice cores, coral bands or other sources-show any time in the last 1,300 years that was as warm as now. Indeed, records from the deep sea suggest that temperatures are now within a degree of their highest levels for no less than a million years.

The part of the globe most vulnerable to this sudden onset of warming, and the part which will likely see the first important ‘tipping point’ crossed, is the Arctic. Here, temperatures are currently rising at twice the global rate. Alaska and Siberia are heating up particularly rapidly; in these regions the mercury has already risen by 2-3°C within the last fifty years.

The impacts of this change are already profound. In Barrow, Alaska, snowmelt now occurs ten days earlier on average than in the 1950s, and shrubs have begun to sprout on the barren, mossy tundra. Scientists based in Fairbanks, Alaska, have documented a sudden thawing of underground ice wedges on the state's normally cold North Slope, with new meltwater ponds dotting the landscape. These ice masses had previously remained frozen for at least the past three thousand years, indicating how far outside previous historical variability current warming is moving.

In other parts of the state entire lakes are draining away into cracks in the ground as the impermeable permafrost layer thaws underneath them. More than 10,000 lakes have shrunk or disappeared altogether in the last half-century, highlighting an alarming drop in the state's water table. In 2007, Canadian researchers reported that in Ellesmere Island, Nunavut, ponds which existed for millennia have now become ephemeral as their water evaporates away in the summer heat. Water-dependent species from insect larvae and freshwater shrimps to nesting birds are being wiped out as a result. Vegetation that once grew on these thin, waterlogged soils is now so desiccated that it easily catches fire.

Arctic mountain glaciers are also responding. On the Seward Peninsula of Alaska, the Grand Union Glacier is retreating so quickly that it is projected to disappear entirely by the year 2035. Other, much larger glaciers elsewhere in Alaska are also thinning rapidly. In the decade up to 2001 alone, the biggest Alaskan glaciers are estimated to have lost 96 cubic kilometres of ice, raising global sea levels by nearly 3 mm. Across the entire Arctic, glaciers and ice caps have lost 400 cubic kilometres of volume over the past forty years.

Perhaps the clearest bell-wether of change is found out at sea. The Arctic ice cap has been in constant retreat since about 1980, with each successive summer seeing more and more of its once-permanent ice disappearing. Each year on average 100,000 square kilometres of new open ocean is revealed as the ice which once overlay it melts away. In September 2005 alone, an area of Arctic sea ice the size of Alaska vanished without trace. Even in the pitch blackness of the winter months, the sea ice cover has been ebbing-both 2005 and 2006 saw the ice extent fall far below average.

Here is where the tipping point comes in. Whilst bright white, snow-covered ice reflects more than 80 per cent of the Sun's heat that falls on it, the darker open ocean can absorb up to 95 per cent of incoming solar radiation. Once sea ice begins to melt, in other words, the process quickly becomes self-reinforcing: more ocean surface is revealed, absorbing solar heat, raising temperatures and making it more difficult for the ice to re-form during the next winter. Climate models differ about exactly where the Arctic sea ice tipping point may lie, but virtually all of them agree that once we are past a certain threshold of warming the disappearance of the entire northern polar ice cap is pretty much unavoidable.

These models suggest that we have not yet reached this critical tipping point-but it may not lie very far away. One model run projects a sudden collapse in sea ice cover after 2024, with four million square kilometres of ice melting away in the following ten years. In this simulation, reported by a US-based team led by Marika Holland of the National Center for Atmospheric Research in Boulder, Colorado, the whole ocean becomes virtually ice-free in summertime by 2040. Whilst other model runs examined by the same team don't cross the tipping point until 2030 or 2040, one simulates a collapse in sea ice production beginning as early as 2012.

Even so, Holland's team emphasises that ‘reductions in future greenhouse gas emissions reduce the likelihood and severity of such events’-in other words, all is not yet lost. Another team, led by NASA's Jim Hansen, reaches a similar conclusion. Despite major changes already in the system, Hansen and co-authors write, ‘it may still be possible to save the Arctic from complete loss of ice’-but only if other atmospheric pollutants (such as soot, which darkens the ice surface and speeds melting) are reduced as well as carbon dioxide. Implement a dramatic programme of emissions reductions, and we ‘may just have a chance of avoiding disastrous climate change’, the team concludes. We may not have much time left, however: at the time of writing, 10 August 2007, a new historic sea ice minimum has just been reported for the Arctic. With a whole month of summer melting still left to go, the expectation is that the previous record low, recorded in 2005, will be ‘annihilated’. Particularly worrying is that dramatic ice extent reductions are being recorded for every sector of the Arctic basin, whereas in previous years only certain areas were affected. Perhaps this is what a tipping point looks like.

But why is Arctic sea ice so important? As the following chapter will show, without it emblematic Arctic species like polar bears and seals are doomed to extinction. But the impacts will also hit closer to home, far away from the once-frozen north. As Ted Scambos, lead scientist at the US National Snow and Ice Data Center in Colorado, explains: ‘Without the ice cover over the Arctic Ocean we have to expect big changes in the Earth's weather.’

These big changes are inevitable because of how the world's climate works. Most mid-latitudinal weather is generated by the contrast between polar cold and equatorial heat: the reason the UK gets year-round rainfall is because of its location on this unstable boundary between these competing air masses-the so-called ‘polar front’. The nor'easter storms which barrel up the eastern US coastline in winter are also generated by this temperature contrast. But with the Arctic warming up, this contrast will lessen and the zone where it takes place will migrate north as rising temperatures contract the world's weather belts towards the poles. In the UK places like Cornwall and Wales which are accustomed to bearing the brunt of stormy winter weather may find themselves in the doldrums for weeks and months at a time, with a much drier overall climate. Only Scotland is likely to hang on to the wetter weather indefinitely. And as chapter 3 will show, the result in the western US is also likely to be drought-but on a scale never before experienced in human history.

Nor are these predicted changes just conjecture: they are already under way. Satellite measurements over the past 30 years have shown a marked 1° latitudinal contraction of the jet streams towards the poles in both hemispheres. Given that these high-altitude wind belts-narrow corridors of rapidly moving air at the top of the troposphere-mark the boundaries between the different air masses, their gradual movement shows that the location of the world's typical climate zones is already starting to shift in response to rising global temperatures.

What we have so far witnessed is still only the beginning. As one group of scientists warned recently: ‘The Arctic system is moving toward a new state that falls outside the envelope of recent Earth history.’ As future chapters show, this new ice-free Arctic will see extreme levels of warmth unlike anything experienced by the northern polar regions for millions of years.

Danger in the Alps

When the Englishmen Craig Higgins and Victor Saunders left the Hornli hut at 4 a.m. on 15 July 2003, they had no idea that they would end the day being part of the biggest-ever rescue on Switzerland's iconic Matterhorn. The ascent began straightforwardly, with the two climbers scaling three rock towers, after which steep slabs led up to a small bivouac hut midway up the Hornli ridge. Higgins and Saunders had just reached the second hut, at 6 a.m., when an enormous rock avalanche pounded down the eastern face of the mountain. Cowering behind the building as stones bounced all around them, the two climbers would have been well advised at that point to turn tail and descend as quickly as possible. But mountains have strange effects on people's minds, and the two Brits pressed on.

Then, three hours later, the mountain shook once again as a further gigantic rockfall crashed down, this time from the north face. Shortly after, a third rockfall struck-and this time the Hornli ridge itself was giving way. A Swiss mountain guide found himself inches from disaster as the ground began to crumble just in front of him. With no hope of crossing the dangerously unstable zone, the guide radioed for help. For the next four hours two Air Zermatt helicopters ferried stranded climbers off the ridge and back to the main hut. ‘As we climbed slowly down,’ recalled Saunders, ‘the smoking plume of rock dust and the returning helicopters told us of a major rescue taking place.’ Both British climbers, realising they too were trapped, joined the queue of people waiting to be plucked to safety.

Ninety people were rescued that day, and amazingly no lives were lost or injuries reported-a tribute to the professionalism of the Swiss mountain guides and emergency services. The mountain remained closed for days afterwards as experts tried to assess the likelihood of further rockfalls taking place. In fact, falling rocks were not the only hazard in the area: on the same day as the Matterhorn drama was taking place, massive chunks of ice broke off from a glacier above the nearby resort of Grindelwald and plunged into a river, causing a two-metre-high wave to flood down the mountain. Fast-acting police managed to clear the area of holidaymakers just before the mass of rocks and mud washed by.

When he heard about the two near-disasters, the glaciologist Wilfried Haeberli had no doubts about the cause. ‘The Matterhorn relies on permafrost to stay together,’ the Zurich University scientist told reporters. But Switzerland had just been suffering its strongest-ever heatwave. With the fierce summer heat having melted all the winter snow much earlier than usual, the permafrost and glaciers themselves were beginning to melt down. Once that process begins, Haeberli warned, ‘water starts to flow, and large chunks of rock begin to break away from the mountain’.

Most ground in the Alps above about 3,000 metres remains permanently frozen throughout the year, and is anchored, as Haeberli says, by permafrost. But in the summer of 2003 the melt zone reached as high as 4,600 metres-higher than the summit of the Matterhorn, and nearly as high as the top of Mont Blanc, western Europe's highest mountain. And whilst the Matterhorn climbers were lucky to get down safely on 15 July, at least fifty other climbers were less fortunate during that boiling summer-most were killed by falling rocks.

Haeberli, a world expert on permafrost, has since co-written a scientific paper on the impacts of the 2003 hot summer in the Alps. He and colleagues calculated that the thaw experienced during that heatwave outranked anything the mountains had suffered in recent history, and that most rock fall as a result took place during the hottest months of June, July and August.

They also found that the 2003 thaw penetrated up to half a metre deeper into the rock than any thaw during the previous two decades.

Surprisingly, however, the worst rockfalls didn't take place on sunny slopes where the direct heat was strongest, but on shaded northern faces, where the high air temperature penetrated the mountain. Ominously, the study concludes that with one degree of further global warming, more permafrost degradation in the Alps is unavoidable. ‘Widespread rockfall and geotechnical problems with human infrastructure are likely to be recurrent consequences of warming permafrost in rock walls due to predicted climatic changes,’ Haeberli and his colleagues warn. ‘The extreme summer of 2003 and its impact on mountain permafrost may be seen as a first manifestation of these projections.’

As mountain slopes thaw out and fail, whole towns and villages will be at risk of destruction in the Alps and other mountain regions. Some towns, like Pontresina in eastern Switzerland, have already begun building earthen bulwarks to guard against deadly landslides from the melting slopes above. But many more will remain unprotected and unprepared-until the worst happens, bringing death crashing down from above, suddenly and with no warning. Moreover, this won't be the only danger associated with mountains in the warming world: as later chapters show, just as dangerous will be the likelihood of running out of life's most precious resource-water.

Queensland's frogs boil

No one could accuse the Australian authorities of not taking their responsibility to protect the Queensland Wet Tropics rainforest seriously. Visitors must stay on walking tracks at all times. Fuel for stoves must be brought in, as campfires might disturb the delicate nutrient cycle of the forest. Every tuft of moss, leaf and twig is protected: removing living materials is a criminal offence. Dogs and cats are banned, as are soap, toothpaste and sunscreen, in case these chemicals leach into streams and harm aquatic animals. And you're most certainly not allowed to swing from vines in the trees.

There is good reason for this intense conservation focus. Recognised since 1988 as a UNESCO World Heritage Site, seven hundred species of plants are found nowhere else on Earth. The Wet Tropics ecosystem contains many species that are unique relicts from ancient rainforests which once grew on the Gondwanan supercontinent 120 million years ago. Many of the same fern species which are still found there today were once grazed by dinosaurs. Amazing carnivorous plants-like pitcher plants and sundews-poke out from the forest floor. Pythons wind around branches, whilst skinks and geckos scurry across rocks and up tree trunks. Thirteen mammal species-including tree kangaroos and ringtail possums-are also unique to the Wet Tropics region. Overall the area is home to a quarter of Australia's frogs, a third of its freshwater fish, and nearly half of its birds-all on a fraction of 1 per cent of the continent.

Yet there is one threat which the Australian authorities are powerless to prevent-and indeed have actively conspired to ignore. It comes not from feral pigs or cane toads, nor even from a thousand swinging, littering, tooth-brushing humans. This threat comes as the climate which sustains the forests-in some areas with a staggering 8 metres of annual rainfall-gradually begins to warm up. It turns out that the Queensland Wet Tropics rainforest is one of the most sensitive areas on the planet to climate change. Just one degree of warming will have devastating impacts on species diversity and habitats.

The reason lies in the unusual topography of the Wet Tropics area. Unlike the Amazon forest, which covers a huge, flat basin until it rises into the eastern slopes of the Andes, the Queensland rainforest comprises hilly terrain-starting from the white sands fringing the ocean to heights of 1,500 metres or more in places. Many of the species which are unique to the area are found only above certain heights: there's a ring-tailed possum which is only found above 800 metres in altitude, and many birds, reptiles and frogs only live at the tops of mountains. As the climate warms, temperature zones rise up the mountainsides, squeezing these species into diminishing islands of habitat-and eventually leaving them with nowhere to live at all. They, like the polar species in the Arctic, will have been literally pushed off the planet.

Dr Steve Williams, of James Cook University's School of Tropical Biology, has been warning for years about the dangers that even small degrees of climate change pose to the Wet Tropics rainforest. Williams-who leads teams of Earthwatch volunteer helpers on his survey trips-has conducted 652 bird surveys, 546 reptile surveys, 342 frog surveys, and at various times set around 50,000 night traps to catch small mammals. Armed with this voluminous wildlife data, he ran a computer model representation of the area under a changing climate and studied the results to see what happened. Even with just a one-degree rise, the results were dramatic. In particular, 63 of the 65 modelled species lost around a third of their core environment. One species of microhylid frog, which instead of having tadpoles in ponds lays its eggs in moist soil, is predicted to go extinct altogether. With higher degrees of warming, rates of biodiversity loss become increasingly dramatic, adding up, in Williams's words, to ‘an environmental catastrophe of global significance’.

Nor are animal species the only ones affected. A similar modelling study by David Hilbert of the CSIRO Tropical Forest Research Centre concluded that one degree of warming would reduce the area of highland rainforest by half, wiping out the habitat of many of the rare animal species mentioned above. Rainforests as a whole won't disappear from Queensland as long as the region receives high rainfall, but without these precious throwbacks to an ancient supercontinent, today's world will be immeasurably poorer. Moreover, Australia's national government, which refused for over a decade to take global warming seriously, will have failed in its international duty to protect a UNESCO World Heritage Site.

Just a few miles offshore from the sparkling white sands of the Queensland coast lies another threatened World Heritage Site: the Great Barrier Reef. This is the biggest and most pristine of all the world's coral reefs, a massive subsea wall of coral which is the largest natural feature on Earth, stretching more than 2,300 km along the north-east coast of Australia. One of the most spectacular and diverse ecosystems on the planet, the reef is home to 1,500 species of fish, 359 types of hard coral, 175 bird species and more than 30 types of mammal. It is one of the last refuges of the dugongs (sea cows) and hosts six of the world's seven species of threatened marine turtle.

But the oceans around the Great Barrier Reef are warming-as they are all over the planet-threatening to tip this unique ecosystem into irreversible decline. Coral reefs are actually the external skeletons produced by billions of tiny coral polyps, which secrete calcium carbonate into branches, fans and globes. These constituents in turn combine over thousands of years to form a reef. Each polyp contains algae, tiny plants which live in symbiosis with their animal hosts. Both parties benefit-the coral gets the sugars which the algae produce by photosynthesising light (turning it into energy), whilst the algae derive fertility from the polyp's waste products. But this cosy relationship can only continue in the right aquatic conditions: once the corals' thermal tolerance threshold of 30°C is crossed, the algae are expelled, and the ‘bleached’ corals will die unless cooler waters return quickly.

Bleaching is undoubtedly a recent phenomenon, observed around the world's oceans only since about 1980. Scientists have drilled deep into reefs and found no evidence that such episodes have happened during past millennia. But as the oceans have warmed due to the human-enhanced greenhouse effect, bleaching episodes have hit the world's coral reefs with increasing-and devastating-regularity. The first mass bleaching event occurred on the Great Barrier Reef in 1998. Since then things have got steadily worse. In 2002 another mass bleaching event occurred-this time 60-95 per cent of all the reefs surveyed across the marine park were bleached to some extent. A small number of reefs, particularly those close to shore where the waters were hottest, suffered almost total wipeout.

As luck would have it, I was on the Great Barrier Reef in the summer of 2002, visiting the University of Queensland's research station on Heron Island. The place was frighteningly efficient-within minutes of disembarking from the Grahamstown catamaran I had learned to tell the difference between a white-capped noddy tern and a muttonbird, and discovered that Heron Island was actually misnamed: the white birds in question are in fact eastern reef egrets. The place was stunning-‘an aviary surrounded by an aquarium’, as one of the scientists accurately put it. Buff-banded land rails scampered about like domestic chickens, in and out of the research huts. (Two female students had adopted one and named it Sheryl.) Very soon I spotted the man I had really come to see, striding purposefully around the station with his wetsuit peeled off down to the waist. Ove Hoegh-Guldberg was clearly a man happy in or out of the water. One of his favourite stories was about the time he managed to get his finger clamped in the jaws of a giant clam, and had to rip the animal off the seabed in order not to be drowned by the rising tide-only to be scolded on returning to the beach by a marine park official for damaging a protected species.

Hoegh-Guldberg was the author of a landmark 1999 paper which first drew the world's attention to the threat posed to coral reef survival by bleaching. Having determined the thermal tolerance threshold for corals in different parts of the world, he then applied these to a model of rising sea temperatures during the twenty-first century. The results shocked even him. By the 2020s, he discovered, with less than a degree of global warming, the seas will have heated up so much that the 1998 Barrier Reef mass bleaching event would be a ‘normal’ year. Given that it takes 30 years or so for a seriously bleached reef to recover, annual bleaching events will devastate the ecosystem-transforming, as Hoegh-Guldberg wrote in the paper, ‘Great Barrier Reef communities into ones dominated by other organisms (e.g. seaweeds) rather than reef-building corals’. Other coral reef ecosystems-from the Caribbean to Thailand-would be similarly transformed. With the end of the coral reefs, one of the world's great treasure troves of biodiversity will be destroyed for ever.

It was with this grim scenario in mind that we both went for a snorkel on the afternoon I arrived on Heron Island. Splashing through the shallows, we disturbed a huge shoal of pilchards, which turned en masse and darted off further up the shore. Halfa dozen large stingrays flapped lazily further out, where a stronger wind raised enough of a chop to make snorkelling a hazardous experience. Every so often a wave would break over the top of my snorkel, giving me a sudden gulp of salty water. Ove was unfazed, and we trod water for a while as he pointed out affected corals.

‘See that bright blue and red? That's actually bleached. It's ironic you get the best colours when it's bleached.’ Some of the worst-hit was the branching coral, where whole areas were blanched bone-white. In some places just the tips of the underwater antlers were bleached, whilst in others the entire structure had been hit. But only a minority were the healthy brownish colour indicating the symbiotic algae still at work.

‘How likely is it to come back?’ I spluttered, swallowing another wave.

‘So long as it stays cool from now on, most of it will probably come back,’ he replied. ‘But some of it won't, and if temperatures rise again soon much of this will probably die.’

Hoegh-Guldberg's work has been complemented by a more recent study which gives a slightly more optimistic forecast. Work in the Caribbean and Indian Ocean by Andrew Baker and colleagues (later published in Nature) suggests that corals may be more adaptable-and therefore less vulnerable to outright extinction-than previously thought. The scientists studied coral communities that had been bleached in the 1998 event to see how far they had managed to recover, and were surprised to find that the type of symbiotic algae within the coral had changed to a more heat-tolerant version in all the places they surveyed. With a higher thermal stress threshold than before, damaged reefs may be able to survive future warmer seas without dying out completely, the scientists suggested.

But Ove Hoegh-Guldberg doesn't agree. Even with an increase in heat tolerance with different algae, he points out, ocean temperatures are still set to get too hot for most corals to survive. He and his co-authors used the latest models and reef analysis to again project bleaching frequencies in the decades to come-and their results confirmed the earlier pessimistic analysis. Severe bleaching will occur on most of the world's reefs every 3-5 years by the 2030s, and by the 2050s will strike every two years.

A more recent bleaching event, which struck the Caribbean in 2005, also seems to bear this out. That summer saw sea temperatures in the region reach highs never before measured during the entire 20-year satellite record. These were the same high temperatures, in fact, which made 2005's hurricane season so deadly: Hurricane Katrina hit New Orleans in 2005 after travelling over these same unusually warm areas of ocean. And they were temperatures which would be vanishingly unlikely in an atmosphere without today's loading of greenhouse gases. The effect on Caribbean corals was disastrous. According to surveys carried out by scuba divers, 90 per cent of coral bleached in the British Virgin Islands, 80 per cent in the US Virgin Islands, 85 per cent in the Netherlands Antilles, 66 per cent in Trinidad and Tobago and 52 per cent in the French West Indies. Some reefs may recover in years to come, but model predictions suggest that in this region too bleaching events of this magnitude will hit every other year by the middle of the century.

In any case, very few of the world's reefs are in any state to take on the challenges of climate change. Direct human interference-from sewage, overfishing and agricultural run-off-has already reduced coral reefs across the globe to shadows of their former pristine selves. In total 70 per cent of the world's reefs are now either dead or dying. This is a disaster of an almost unimaginable scale for global biodiversity: second only to rainforests in terms of the vibrancy and diversity of life they nurture, coral reefs worldwide shelter and feed a third of all life in the oceans, including 4,000 types of fish.

Heron Island's reef may be under good management, but the same cannot be said for reefs elsewhere in the Pacific. In the same trip as I visited Ove Hoegh-Guldberg, I also snorkelled along Fiji's so-called Coral Coast, at one of the few gaps I could find between the 5-star hotels and luxury resorts which now blight the entire area. Instead of vibrant-coloured reefs, teeming with parrotfish and groupers, I found piles of rubble-the shattered remains of coral-looming bleakly through a murky ocean. None of the sun-bathers crowding onto the beach seemed to mind, but for me the experience was a depressing reality check. Fiji's Coral Coast is no longer vulnerable to climate change, I was forced to conclude, because it is already dead.

Another hot spot of biodiversity-and yet another World Heritage Site threatened by global warming-is the Cape Floristic Region of South Africa. Covering a huge coastal arc inland from Cape Town, it is home to the greatest concentration of higher plant species in the world outside the tropical rainforests. Its inauspicious rocky soils and arid Mediterranean climate support 9,000 different plants, more than 6,000 of which are found nowhere else on the planet. The most iconic plants in the region are the proteas. The king protea, with its massive sunlike flower head, entirely deserves its designated title as South Africa's national flower. The region is far from pristine, however-lions and rhinoceroses once roamed these hillsides, where now vineyards and rooibos tea plantations encroach on the last wild areas.

According to a team of researchers based at South Africa's National Biodiversity Institute, just small changes in climate could have a devastating impact on the remaining strongholds of the proteas and other endemic species. Using the UK Hadley Centre's model for climate changes in the region by 2020, the scientific team concluded that up to a third of protea species would become threatened or endangered, whilst four would become completely extinct.

In North America too, one degree of climate change could push a threatened species over the brink to extinction-and this one is cute and furry According to WWF, pikas-small, hamsterlike creatures with rounded ears and bushy whiskers-are the first mammal to be endangered by climate change. Pikas live in broken rock on high mountains in the western US and south-western Canada, and are notable not just for being cute and furry, but for their agricultural activities: these small relatives of the common rabbit cut, sun-dry and then store vegetation for winter use in characteristic ‘haypiles’ on top of rocks. (As a charismatic species, pikas have acquired quite a cult following: check out www.pikaworks.com for everything from pika music to pika mouse mats.)

However, as the climate warms, pikas-timid beasts, which never stray more than a kilometre from their nests-are set to become increasingly isolated in ever-smaller geographical islands as temperature zones migrate upwards towards the summits. Already local extinctions have been documented at sites in the United States. As the ecologist and pika enthusiast Dr Erik Beever puts it: ‘We're witnessing some of the first contemporary examples of global warming apparently contributing to the local extinction of an American mammal at sites across an entire eco-region.’

It has become something of a cliché to talk about the ‘canary in the coal mine’ when discussing climate impacts on the natural world-but one group of animals more than any other exemplifies this point: the amphibians. With their moist skins and early lives in water, frogs, salamanders and toads are particularly vulnerable to changes in their environment. Indeed, an amphibian-the Costa Rican golden toad-is often cited as the first known case of a global warming extinction.

Once the ‘jewel in the crown’ of Costa Rica's Monteverde Cloud Forest (to paraphrase the scientist and author Tim Flannery), this Day-Glo orange amphibian was observed in its hundreds back in 1987, gathered around pools in the forest in preparation for mating. But there were already signs of danger: the amphibian expert Marty Crump, who witnessed this last golden toad mating frenzy, also watched the resulting eggs get left behind as the forest pools dried. Only twenty-nine tadpoles lasted out the week, whilst 43,500 eggs were left desiccated and rotting. The following year Crump found only a single, solitary male, and a year later, in 1989, the same male was back once more. That day, 15 May 1989, was the last time anyone saw a golden toad. The species was eventually listed as extinct in 2004. The cause of death seems to have been the general lifting of the mist that nourishes the forest with moist cloud droplets: as the air surrounding the mountains warmed, the cloud base simply rose too high above the forest, allowing the golden toad's spawning pools to dry out.

This memorable animal may be the first, but it is no longer the only amphibian to have gone extinct because of rising temperatures: frog populations have crashed all around the tropics, with more than 100 out of 110 tropical American harlequin frog species having disappeared-even in seemingly pristine forests far away from direct human disturbance. No one knows exactly why: some biologists blame the chytrid fungal pathogen, which is invading new areas and may be causing sudden population crashes. Others blame mystery diseases which are so far undiscovered and unidentified. But experts are largely agreed on one thing: rising temperatures are central to the extinction epidemic, either by helping the new diseases spread, or by stressing amphibian populations and making them more susceptible to die-offs. In this particular murder scene, the weapon may still be in dispute but the overall culprit is clear.

Nowhere, it seems, is safe. One degree of global warming will have severe impacts in some of the world's most unique environments, adding to the biodiversity crisis which is now well under way for reasons unrelated to our changing climate. Pushed out to the margins and isolated in smaller and smaller pockets of natural habitat by ever-expanding zones of human influence, vulnerable wild species will find it impossible to adapt to rapidly changing temperatures by migrating or altering their behaviour.

Whilst coral reefs have a vital role in protecting coastlines from storms and nurturing fisheries, no one can reasonably claim that pikas, proteas and harlequin frogs are essential for global economic prosperity. Their value is intrinsic, not financial. But the world will still be a much poorer place once they're gone.

Hurricane warning in the South Atlantic

With all the headlines about hurricanes hitting the United States, there is one storm that stands out above all others in the way that it caught the scientific community off guard. It wasn't Katrina, which devastated New Orleans and killed over a thousand people. It wasn't Rita, another Category 5 monster which reflooded parts of the city only a month after Katrina struck. Nor was it Hurricane Wilma, which bombed in a single day from being a minor tropical storm into the strongest hurricane ever recorded in the Atlantic basin. No, the storm that really made the forecasters scratch their heads occurred a year earlier, in 2004. And it hit in a part of the world that isn't supposed to experience hurricanes. It was called Catarina, and it struck the coast of Brazil.

Received scientific wisdom holds that hurricanes can only form where sea surface temperatures top 26.5°C. As well as warm oceans, tropical storms also need low ‘wind shear’: cross-winds at high altitude which can cut the vortex of a developing storm in half. These conditions, as any weatherman can tell you, only occur in the North Atlantic tropics. Not a single South Atlantic hurricane had ever been documented-that is, before March 2004. Indeed, when a strange swirl of clouds began to form off the Brazilian coast on 20 March 2004, local meteorologists couldn't quite believe their eyes. So unheard of was a South Atlantic hurricane that many of them were still refusing to employ the term ‘hurricane’ when Catarina-complete with 95 mph winds and torrential rain-swept ashore near the town of Torres, damaging 30,000 houses and killing several people. Many of those who suffered, because they also refused to believe that hurricanes were possible in Brazil, had neglected to take shelter as the storm barrelled towards the shore.

In the inevitable meteorological post-mortem, it did indeed look as if the storm was simply a freak, a once-in-a-lifetime experience for those who suffered it. What was strange was that sea temperatures had not been unusually high when it began to gather strength. Instead, what really gave Catarina a boost was a very rare combination of other atmospheric factors which meant that the storm's vortex experienced very little of the deadly wind shear which normally precludes hurricane formation in the South Atlantic. It's a complex picture, but raises an obvious question: will global warming, as well as making the seas warmer and therefore more likely to spark hurricanes, lead more regularly to the conditions which allow tropical cyclones to gather strength in new areas like the South Atlantic?

Two Australia-based meteorologists, Alexandre Bernardes Pezza and Ian Simmonds, addressed this question in their forensic dissection of Catarina which was published in August 2005 by the journal Geophysical Research Letters. Their conclusion was tentative, but contained an alarming prospect: it did indeed seem as if the warming atmosphere would favour the conditions which allowed Catarina to form in such an unusual place. ‘Therefore,’ they wrote, ‘there is evidence to suggest that Catarina could be linked to climate change in the southern hemisphere circulation, and other possible future South Atlantic hurricanes could be more likely to occur under global warming conditions.’

Given that if just 0.8°C of global warming so far has already allowed one hurricane to form, a further degree of global warming could make storms in this vulnerable region much more likely in the future. Not only will Brazilians have to batten down the hatches-and perhaps evacuate large areas of their heavily populated coastline-more often, but hurricane-forecasting services will need to be extended to a whole new oceanic basin.

The hurricane season in the following year, 2005, also contained a surprise, which suggests that Brazil is not the only area which will have to keep an eye out for tropical cyclones in our globally warmed future. On 9 October 2005 a new tropical storm appeared about five hundred miles south-east of the Azores, in the East Atlantic, and rapidly gained strength to hurricane status as it drifted past Portugal's Madeira Islands. Hurricane Vince luckily weakened before it made landfall near Huelva in Spain, but set a new record as the first tropical cyclone ever to affect Europe.

Again, conventional wisdom has it that tropical storms can only form over warmer waters thousands of kilometres to the southwest of Iberia. At the time of writing, tropical meteorologists have yet to dissect the unusual combination of factors responsible for Vince, but again the implication is clear: as global warming accelerates, past experience about areas of hurricane formation is not necessarily a reliable guide to the future. Many more hurricane forecasters may be left scratching their heads before they finally admit that not just Brazil but now Europe is already vulnerable to these terrifying storms.

Indeed, there is now evidence for how this might happen: a paper published in July 2007 by Spanish and German climatolo-gists, looking at simulated storms in a computer model, suggests that the whole of the Mediterranean may soon come into the firing line as sea temperatures climb to levels able to spark off genuine tropical cyclones-in a region which has never seen them before. The greatest number of virtual cyclones appeared in the hottest part of the Mediterranean, between Italy and Libya, and once formed, the powerful storms lasted for a week or more.

One computer-generated hurricane formed in the eastern part of the Mediterranean, and then wandered westward all the way to the southern coast of France-much to the amazement of the watching scientists. Another storm formed a tight, symmetrical eye of torrential rain, just as real tropical cyclones do. The idea that once-placid coastlines from Spain to Cyprus could be at risk of landfalling hurricanes in a globally-warmed future has to be one of the most striking projections ever to come from the climate modelling world.

But already real-world evidence is fast emerging that hurricane characteristics are changing as the world's oceans warm up. One of the grandfathers of tropical cyclone physics, Massachusetts Institute of Technology's Kerry Emanuel, recently published a paper in Nature that caused its own academic storm. In contrast to the usual view that global warming is still too small a signal to have a measurable impact on tropical cyclones, Emanuel looked again at the data and concluded that storms were in fact getting more intense and lasting for longer, due in large part to rising sea temperatures driven by global warming. The storm intensity index hadn't simply gone up by a few percentage points over the last 30 years, either, it had actually doubled-a far greater increase than either theory or modelling had predicted.

Emanuel's data and methods have since been challenged in an academic discussion too technical to analyse here, but it is worth noting that his conclusions are supported by a second piece of work, this time published in Science by a team of experts based at the Georgia Institute of Technology in Atlanta. Analysing much of the same storm data-collected by aircraft, satellites and ships over the past three decades-this scientific team identified a large increase in the number and proportion of those hurricanes reaching the strongest categories 4 and 5, despite an overall decrease in the number of cyclones.

Like Emanuel, the team were looking at data from the Pacific as well as the Atlantic Ocean in order to build up a global picture. And like him (though using a different statistical measure), they found a near-doubling in the number of the strongest storms between 1970 and 2004. The Georgia team concluded that the upswing in Category 4 and 5 hurricanes is unlikely to be the result of natural climate cycles, but instead is probably connected to rising temperatures in the tropical oceans.

A year later, after the record-breaking 2005 Atlantic hurricane season-which left 1,000 people dead, 1 million homeless, and caused $200 billion of damage-two leading climatologists tried to settle the argument about whether or not global warming had contributed to the run of disastrous storms. Noting that the warm sea temperatures measured that year-the highest ever-undoubtedly contributed to the ferocity of Katrina, Wilma, Rita and the other 2005 storms, Kevin Trenberth and Dennis Shea used complex maths to figure out how much of the Atlantic warming signal was due to global warming, and how much to natural cycles. Their conclusion should be a wake-up call to us all: at least half of the extra warmth had come from human-caused global warming. As so many people suspected at the time, Katrina was only partly a natural disaster.

Sinking atolls

I hate to put this so bluntly, but in all probability nothing can save the Pacific island of Tuvalu. Like a slowly boiling kettle, the oceanic system has very long response time to changing conditions, and the seas will go on slowly rising for centuries even if all greenhouse gas emissions stopped tomorrow. With Tuvalu already experiencing regular flooding events due to past sea level rise-as I documented in High Tide-this extra rise in the world's oceans will sound the death knell for this fascinating and lively island society.

Tuvalu, with only 9,000 inhabitants, is actually one of the smallest of the five atoll nations which will shortly cease to exist. The others are Tuvalu's sister atoll group Kiribati, with 78,000 population, the Marshall Islands, with 58,000 people, tiny Tokelau (2,000 people; a dependent territory of New Zealand) and the Maldives, the largest and most densely populated of all the island groups, with 269,000 inhabitants. Together with people displaced from coastal areas of other non-atoll islands, this already totals about half a million people who-suddenly divorced from their cultures and their origins-will need to find new homes. New Zealand has hesitantly offered to take a small number of Tuvaluans, but no other nations have yet stepped up to offer themselves as places of refuge, least of all those rich countries who have done most to cause the problem in the first place.

Unless spurred on by a major hurricane or storm surge, the end for atoll countries will not be rapid or cathartically dramatic. Instead it will be death by a thousand cuts, an incremental dimin-ishment of each nation's ability to support itself, as young people lose confidence in the future and old people sink back into comforting dreams of the past. Each bit of beach lost, each vegetable garden invaded by salt water, each undercut coconut tree which topples into the waves will add to the inevitable toll. Decades before the last bit of coral disappears under the sea, community services will decline, children will emigrate, schools will close, and the fabric of a nation will begin to unravel.

Bear in mind that as future chapters of this book unfold, unacknowledged and mostly forgotten, atoll nations will be submerging-bit by bit.