If there was a golden age for humans on the Earth - a Garden of Eden that flowed with milk and honey - it was the high point of the Holocene, the era that followed the end of the last ice age. From around 8,000 to 5,500 years ago, the world was as warm as it is today, but there appear to have been few strong hurricanes and few disruptive El Niños; and it was a world in which the regions occupied today by great deserts in Asia, the Americas and Africa were much wetter than they are now.
Optimists suggest that such conditions might be what awaits us in a greenhouse world. There are celestial reasons why that might not happen, but that era, and its abrupt ending, may still offer important lessons about our future climate in the 21st century.
No place on Earth exemplifies the fall from this climatically blessed state better than the Sahara. The world's largest desert was not always so arid. Where seas of sand now shimmer, there were once vast lakes, swamps and rivers. Lake Chad, which today covers a paltry few hundred square kilometres, was then a vast inland sea, dubbed by scientists Lake Megachad. It was the size of France, Spain, Germany and the UK put together. Today the lake evaporates in the desert sun, but then it overflowed its inland basin and, at different times, drained south via Nigeria into the Atlantic or east down a vast wadi to the Nile.
The difference is that back then the Sahara had assured rains. The whole of North Africa was watered by a monsoon system rather like the one that keeps much of Asia wet today. Rain-bearing winds penetrated deep into the interior. From Senegal on the shores of the Atlantic to the Horn of Africa in the east, and from the shores of the Mediterranean in the north to the threshold of the central African rainforest, vast rivers flowed for thousands of kilometres. Along their banks were swamps and forests.
Beneath the Algerian desert, archaeologists have found the remains of wadis that once drained 1,000 kilometres from the Ahaggar Mountains into the Mediterranean. And in waterless southern Libya, archaeologists are finding the bones of crocodiles and hippos, elephants and antelope. If there was a vestige of true desert at the heart of North Africa, it was very much smaller than the desert is today.
And there were people: shepherds and fishers and hunters, and some of the earliest known fields of grains. Archaeologists digging in the sands of northern Chad, the dustiest place on Earth, have found human settlements around the shore of the ancient Lake Megachad. Paintings in caves deep in the desert depict the lives of the inhabitants of the verdant Sahara of the Holocene.
There are other, more practical remains. Rocks beneath the Sahara contain the largest underground reservoir of fresh water in the world. They were mostly filled from leaking wadis in the early Holocene. Some desert settlements today tap these waters at oases. Colonel Gaddafi has constructed pumps and a huge pipeline network to take this water from beneath southern Libya to his coastal farmers. He calls the network his Great Man-made River, though it is a feeble imitation of the real rivers that once ran here.
The wet Sahara and the era known more generally as the African Humid Period began around 13,000 years ago as the ice age abated, and, except for the Younger Dryas hiatus, lasted to the end of the "golden age" of the African Humid Period. It coincided with a time when the Earth's tilt known as the precession ensured that the sun was blazing down on the Sahara with full intensity in summer. The land cooked and convective air currents were strong. As the warm air rose, wet air was drawn in from over the Atlantic to replace it. The process was the same one that creates today's monsoon rain system in Asia. Meanwhile, the monsoon rains were recycled by the rich vegetation across North Africa. Rather as in the Amazon today, the rain nurtured lush vegetation that ensured that much of it evaporated back into the air. The continually moistened winds took rain to the heart of the Sahara.
But the African Humid Period came to an end - known as the Fall - suddenly. In the space of a century, the Saharan rivers emptied, swamps dried up and the monsoon rain clouds were replaced by clouds of wind-blown sand. The climate system had crossed a threshold that triggered massive change. Why?
First, the sun moved. Or rather, the precession continued its stately progress and gradually took away the extremely favourable conditions for Saharan rains. And as summer solar heating lessened, the warm air rose a little less and the monsoon winds from the ocean didn't penetrate as far inland some years. The process went on without any appreciable effect on rainfall in most of the Sahara for more than 3,000 years. The vegetation feedback ensured that, most years, the rain kept falling.
But at some point, the feedback began to falter. Perhaps there was a chance variation in rainfall that dried out the bush for a year or two. The sun was no longer strong enough to make good and revive the rains. Suddenly, what had been a feedback that kept the Sahara watered became a feedback that dried it out. The system had passed a threshold, and it never recovered. The green Sahara had become a brown Sahara. The North African monsoon rains had died.
Not everybody agrees that the vegetation feedback was the only trigger for the drying of the Sahara. There may have been some influence from one of the geologist Gerard Bond's solar pulses. Bond, formerly of the Lamont-Doherty Earth Observatory at Columbia University, New York, argued that regular pulses in solar activity drive cycles of climate change. But climate models show that in all probability this flip in the Saharan climate was extremely sudden.
Martin Claussen at the Potsdam Institute for Climate Impact Research in Germany has played out this tragedy in detail in his model. He turns time forwards and backwards, recreates the subtle orbital changes and fine-tunes the vegetation feedbacks. More or less whatever he does to mimic the conditions of 5,500 years ago, the result is the same. The system flips abruptly, turning bush to desert and seas of water to seas of sand.
Other researchers have replicated his findings. Peter deMenocal at Lamont-Doherty calculates that the system flipped when solar radiation in the Sahara crossed a threshold of 470 watts per square metre. Jon Foley of the University of Wisconsin found that a reduction in Holocene summer sun sufficient to reduce temperatures by just 0.4C would have cut rainfall across the Sahara by a quarter, and by much more in the furthest interior. He says that once a region such as the Sahara becomes dry and brown it requires exceptional rains to trigger a regreening. Beyond a certain point - such as that reached 5,500 years ago - virtually no amount of extra rain is likely to be enough. Lack of vegetation "acts to lock in and reinforce the drought".
The people of the Sahara couldn't have known if the droughts were permanent. But as the desert asserted control, and waterways dried up, they had to leave. Lakeside settlements near the Sudanese border in Egypt were all abandoned at about the same time.
One was Nabta, famous as the site of the world's earliest known stone structures with an astronomical purpose. They predate Stonehenge by 1,000 years. The key stones point to where the sun would have set at the summer solstice 6,000 years ago. Nobody can be sure what the structures' precise purpose was, but it is intriguing to suppose that they were used in an attempt to track the celestial changes that were disrupting the rains.
It may have been from such places that the myths of past golden ages, and of the Garden of Eden, emerged. The people who departed from the Sahara would have taken their memories of a golden past. Biblical scholars have calculated mankind's expulsion from Eden at around 6,000 years ago, when kingdoms across the Sahara would have been collapsing.
But the Eden need not have been in the Sahara: similar stories were played out elsewhere. Arabia dried out at the same time, leaving behind a huge underground reservoir of water not much smaller than that beneath the Sahara. Claussen calculates that the desertification of Arabia could have been caused by the same combination of gradual orbital change and a dramatic vegetation feedback.
The evidence is as yet sketchy, but dramatic drying of the Sahara and Arabia appears to coincide with other climate changes round the world. In the Pacific Ocean, El Niño appeared to switch into a more active mode at around this time. There were cold periods from the Andes to the European Alps. In both cases, glaciers advanced strongly; often they are only returning to their former positions today.
In the Austrian Tyrol, one victim of the advance was the "ice man" Otzi, whose freeze-dried remains emerged in 1991. In Ireland, a 7,000-year temperature record held in tree rings shows cold times that included the coldest summer in the entire record.
All this is intriguing because, unlike previous great climatic events of the ice ages, there is little evidence that the primary action had much to do with the polar regions. It seems to have been an abrupt change formed in the tropics and with its major impacts there, and only ripples beyond.
What does this say about the future of the Sahara? Could warming in the 21st century trigger a greener, wetter Sahara? The idea has with plenty of adherents. Reindert Haarsma, a climate modeller at the Royal Netherlands Meteorological Institute, says the Sahara could be destined for a 50 per cent increase in rainfall - enough to trigger a "golden age" in which crocodiles float through swamps where today locusts swarm.
Claussen, whose model first stimulated the idea, is more sceptical. He points out that the orbital situation today is very different, so summer solar radiation is not great enough to create a revived African monsoon. DeMenocal says solar radiation is 4 per cent lower in the Sahara than it was when the Holocene flip occurred. But on the other hand, he admits, much higher levels of CO2 in the air might compensate for this by stimulating an earlier recovery of Saharan vegetation.
Optimists point out that there is currently something of a revival going on in Saharan rains - albeit from the depths of the droughts that afflicted the region in the 1970s and 1980s. It hasn't happened everywhere, and some places have since slipped back.
But, according to Chris Reij of the Free University in Amsterdam, improved farming methods, such as digging terraces and holding water on the land, may have encouraged a modest greening of parts of the Sahara, and the resulting vegetation feedback could be one reason for the revived rains. But it would be a big step to predict from that a reversion to "Eden" days.
While some in the Sahara may conceivably be able to look forward to greener, wetter times, the prognosis for many other arid regions round the world is not so good. The big fear from the American West to northern China and Southern Africa to the Mediterranean is of a 21st century dominated by longer and fiercer droughts.
Again, history is the first guide. DeMenocal has been looking at the history of droughts and civilisation in the Americas, and finds strong evidence of droughts much longer than any known in modern times. "Vast regions of North America witnessed several such periods during the last millennium, with devastating cultural consequences," he says. "These mega-droughts can persist for a century or more."
The six-year Dust Bowl of the 1930s, which caused mass migrations westwards, was "pale by comparison" with its predecessors. Droughts in the 19th century devastated many Native Americans as well as their bison. At the end of the 16th century, a 22-year drought destroyed an early English colony at Roanoke in Virginia. It became known as the Lost Colony after all its inhabitants disappeared between their arrival in 1587 and the return of a supply ship four years later. And tree rings show there was near permanent drought from AD 900 to 1300 west of the Mississippi and through Central America that destroyed the Mayan and Anasazi civilisations.
DeMenocal concludes that complex, organised societies can get by in short droughts. They have stocks of food and water and know how to trade. But few can deal with megadroughts. If hunger doesn't get them, the strife caused by trying to survive does.
The signs are that worsening droughts are again becoming the norm in regions that have suffered megadroughts in the past. In the American West, the biggest river, the Colorado, is a shadow of its former self. Early in the 20th century, the average flow was 16 cubic kilometres a year. In 2002, it fell to just 3.7 cubic kilometres - worse even than the Dust Bowl years of the mid-1930s.
In Central Asia, the Afghan war of 2002 was fought against a backdrop of drought as debilitating as any Taliban tyranny. The Hamoun wetland covering 4,000 sq km on Afghanistan's border with Iran has for millennia been a place of refuge for people from both countries. But that year it turned to salt flats.
Richard Seager at Lamont-Doherty says that there is a long-standing correlation between drought in the western US and in South America, parts of Europe and Central Asia. And that is a pattern we see reasserting itself in the 21st century as the Arizona desert creeps north, Southern Europe increasingly resembles North Africa and Central Asia takes on the appearance of Iraq or the Arabian peninsula.
Kevin Trenberth of the National Center for Atmospheric Research in Boulder, Colorado, reports that the percentage of the Earth's land area stricken by serious drought has more than doubled in 30 years. In the 1970s, less than 15 per cent of the land was drought-stricken, but by the first years of 21st century it had risen to 30 per cent.
That seems to be a common view. Mark Cane, a specialist in Pacific weather at Lamont-Doherty, says: "The medieval warm period a thousand years ago was a very small forcing compared with what is going on with global warming now. But it was still strong enough to cause a 300- to 400-year drought in the western US. That could be an analogue for what will happening under anthropogenic warming. If the mechanisms we think work hold true, then we'll get big droughts in the West again."
Many believe that El Niño and the pattern of ocean temperatures in the Pacific are heavily implicated in the historic megadroughts, perhaps as part of a global reorganisation of climate systems linked to Gerard Bond's pulses. And this should set modern alarm bells ringing, says Ed Cook, a leading tree-ring expert from Lamont-Doherty.
"Any trend towards warmer temperatures [over the tropical Pacific] could lead to a serious long-term increase in aridity over western North America." Martin Hoerling of the National Oceanic and Atmospheric Administration in Boulder, Colorado thinks that such a process is already underway. He blames the increasingly drought-prone nature of the tropics on a persistent ocean warming in the Pacific. The pattern of dryness is beginning to look less like a local, short-term aberration and more like a long-term trend, he says, and predicts that global warming "may be a harbinger of future severe and extensive droughts".
It won't happen everywhere, of course. Climate models predict that a warmer world will, on average, have more moisture in the atmosphere and that, in general, the wet places will get wetter and the dry places drier. They predict that areas of uplift, where rising air will trigger storm clouds and abundant rain, will see the uplift become more intense. But the areas of sinking air, which are the traditional deserts of the world, will see more intense sinking and drying.
In many parts of the world, this "hyperweather" is likely to set competing forces against each other. Stronger storms will blow off the oceans and, in places, monsoon-type rains may begin to restart. But the rain-bearing winds will often by confronted by intensifying arid zones of descending air in the continental interiors. It is not obvious which force will win, and where.
Will the Sahara desert expand and intensify as the drought theorists argue? Or will North Africa be reclaimed by a revived African monsoon? Megadrought or Garden of Eden? Nobody can answer that question yet. Perhaps the greatest likelihood is that in many places, from the Sahara to the American West and Arabia, there will be more and longer droughts, interspersed with brief but devastating floods.
SEE-SAW ACROSS THE OCEAN - how the Sahara desert greens the Amazon
Two of the world's largest and most fragile ecosystems face each other across the Atlantic. On one side is the Amazon rainforest; on the other the Sahara. The Sahara is rainless and largely empty of vegetation. The Amazon is one of the wettest places on Earth, and its most biologically diverse. But these two opposites are not so far apart. For one thing, the physical gap is surprisingly small. Near the equator, the two are less than half as far apart as London and New York.
Many believe the two areas have a surprising symbiosis. Their fates may be intertwined in a rather unexpected way - and one that could have important consequences in the near future.
The key to the symbiosis lies in a region called Bodele in northern Chad. Few people go here. It is littered with unexploded bombs and land mines left behind during Libya's invasion during the 1980s. And it is by some way the dustiest place on Earth. Satellite images show year-round dust storms raging across Bodele and entering the atmospheric circulation. According to Richard Washington of Oxford University, two-fifths of the dust in the atmosphere comes from the Sahara, and of that half comes from Bodele.
Some of this dust stays local. But much of it is carried west across the desert wastes of Niger, Mali and Mauritania before heading out over the Atlantic. The red dust clouds can grow three kilometres high as they approach America. They cause spectacular sunrises over Miami, before falling in the rains of the Caribbean and the Amazon. And there have been a lot of good sunrises in recent decades. The amount of dust crossing the Atlantic grew five-fold between the wet 1960s and the dry 1980s.
The Sahara dust has a series of unexpected effects on the Americas. According to hurricane forecasters in Florida, during dry, dusty years in the Sahara there are fewer hurricanes on the other side of the Atlantic. It seems that dust in the air interrupts the updraughts that fuel the storms. Equally surprisingly, desert bacteria caught up in the winds are being blamed for bringing new diseases to Caribbean coral reefs and triggering asthma among Caribbean children. And there is a third important link. Saharan dust storms carry huge amounts of minerals and organic matter that enrich soils widely in the Americas. Bodele dust seems especially valuable. Its dunes are the dried-out remains of the bed of the vast Lake Megachad, which covered central Sahara until its abrupt demise 5,500 years ago.
Most of the dunes are not made of sand or broken rock. They are the remains of trillions of diatomites, microscopic freshwater creatures that once lived in the lake. These fragments are light enough to blow freely in the wind. And they make great fertiliser.
If Bodele had any rain, the diatomites would make rich farmland. Instead, Chad's loss is the Americas' gain, says Hans Joachim Schellnhuber, a German physicist turned Earth system scientist, who as director of Britain's Tyndall Climate Centre in Norwich has made a study of the unlikely connection. "The Sahara fertilises the Amazon rainforest. This process has been going on for thousands of years."
The two habitats are on a kind of see-saw, he says. When the Sahara is dry, as it has been for much of the past 25 years, its dust crosses the Atlantic in huge quantities and fertilises the Amazon. When the Sahara is wet, the dust storms subside and the Amazon goes hungry. That the Sahara seems to have only two basic modes, wet and dry, suggests that there may be two distinct modes in the Amazon, too. The last big change in the Sahara came 5,500 years ago when the region lurched from wet to dry, probably within decades. As yet we know little about how the Amazon changed at that time. But if Schellnhuber is right, the Sahara's loss at that time may have been the Amazon's gain.
In the 21st century, the see-saw could be on the move again. There are hints that the Sahara may become wetter. And if the wetting turns to greening, and the vegetation feedback kicks in, the whole of North Africa could change dramatically. That would be good news for the Sahara. But perhaps bad news for the Amazon, which seems to be close to its own tipping point as the climate dries and rainforests give up their carbon. Could a wetter Sahara be the final nail in the Amazon coffin? Schellnhuber believes so.
These are edited extracts from The Last Generation by Fred Pearce, published by Transworld at £12.99. To order it for the special price of £11.50, including p&p, call Independent Books Direct on 08700 798 897