Almost 3 million years ago a previously moist and warm east Africa, which was covered with tropical woodlands and green landscapes, was transformed into a dry, grass-covered savannah. It had been assumed1 that this climatic change originated in the north Atlantic. Now Mark Cane from the Lamont-Doherty Earth Observatory in Palisades, New York, and Peter Molnar of the Massachusetts Institute of Technology argue that it could have been caused by the closure of the Indonesian seaway.

Given that the human family tree diversified at around the same time, this new hypothesis throws a new light on the factors that may have determined our ancestors’ evolution.

The most popular explanation for the drying out of east Africa has been the coincident onset of the first ice age. Huge ice sheets in the Northern Hemisphere led to low surface temperatures in the north Atlantic. According to this argument, less water evaporating from a cold north Atlantic led to less rain over Africa.

I am convinced that past climate oscillations originate in the tropics.Mark Cane Lamont-Doherty Earth Observatory

Cane and Molnar propose an alternative explanation. Australia and New Guinea are on a tectonic plate that moves north at about 70 km every million years. Three to four million years ago, the researchers calculate, the displacement and uplift of the Indonesian islands caused a fundamental redirection of ocean currents that shifted the connection between the Pacific and Indian oceans north. Instead of warm water from the south Pacific, cooler water from the northern Pacific flowed through the seaway. Cooling the surface of the Indian Ocean would have initiated a different wind system over east Africa, bringing less rain.

So far, the theory is an addition to our present understanding of climate change, rather than a contradiction. But Cane and Molnar go further. They assert that before the northward displacement of the Indonesian Seaway, the distribution of water masses in the Pacific Ocean was quite different from that of today.

The researchers suggest that before the changes in Indonesia, the Pacific Ocean would have been in a state of 'permanent El Niño'. Nowadays, large amounts of heat are transported from the tropics to high latitudes in El Niño years. New Guinea's northward displacement would have ended this heat transport, triggering an ice age.

So according to Cane and Molnar, glacial cycles are a result of tectonic changes between the Indian and Pacific oceans.

Previously, the closing of the Middle American Seaway has been the explanation for the start of the ice age. A huge gap between North and South America once allowed water to flow freely between the Atlantic and Pacific oceans. But tectonic changes closed this ocean gateway about 3 million years ago, lifting up the Isthmus of Panama, the thin sliver of land that connects North and South America.

The closing of the Straight of Panama initiated the Gulf Stream in its present form. Before the isthmus existed, part of the tropical water mass had been diverted into the Pacific Ocean through the Americas. After that route closed, the warm current flowed along the North American coast. A change of such a magnitude would have had important consequences for climate in the north Atlantic.

The first doubts about this hypothesis appeared in 1982, when Lloyd Keigwin of Woods Hole Oceanographic Institution in Massachusetts found evidence that the closing of the Isthmus of Panama influenced ocean circulation more than a million years before the onset of the new climate regime.

Sixteen years later, Gerald Haug of ETH Zuerich and Ralf Tiedemann of the GEOMAR research center in Kiel, Germany, investigated the paradox further, with more accurate and more abundant data. "Our analysis supports Lloyd Keigwin’s data -- 4.2 million years ago at the latest, the division of water masses between the Atlantic and Pacific oceans in the Straight of Panama was complete," says Haug. So why did glaciation not start until 3 million years ago?

"To build up an ice sheet, you need cold temperatures and moisture. The evolution of the Isthmus of Panama provided the necessary moisture 4.2 million years ago at the latest, but to trigger glaciation you would also need the temperatures to fall," Haug explains.

One important factor for global temperature change is solar radiation, which oscillates regularly in cycles of about 20,000, 41,000 and 100,000 years through variations in the Earth’s orbit around the sun. The combination of these cycles did not allow cool Northern Hemisphere summers -- essential for an ice-sheet buildup -- for a long time. Between 4.5 and 3.1 million years ago, the proto-ice sheet that formed in winter would have melted during the warm summers.

But Cane does not believe this explanation. "A stronger Gulf Stream since the closing of the Panama seaway does not only transport moisture to the north Atlantic, but also heat. You will not get an ice sheet in this way. I am convinced that past climate oscillations originate in the tropics, and not in the north Atlantic."

Cane and Molnar's proposed relationship between glacial cycles, changes in African climate, displacement of the Indonesian Seaway and the closure of the Straight of Panama also invites a re-evaluation of the idea that several new species may have emerged in response to the aridification of East Africa.

A changing climate drives different populations apart and brings them together again.Richard Potts Smithsonian Institute

Thanks to a series of spectacular new fossil finds -- culminating earlier this year in Kenyanthropus platyops, the first clear evidence of a second species of human ancestor that existed between 4 and 3 million years ago -- the theory of climatic influence on early human evolution is not quite as simple now as it was a few years ago. "We see a strong widening of the human family tree between about 4 and 2 million years ago. But I do not believe that there was a magic point in time when all this happened," says Daniel Liebermann, an anthropologist at the George Washington University in Washington DC.

The onset of cycles of glacial and interglacial climates not only dried Africa out but also led to more variation in global climate. Richard Potts, director of the Human Origins Program at the Smithsonian Institute in Washington DC, believes that the extent of variability is most important for human evolution. "A changing climate drives different populations apart and brings them together again. This could have facilitated speciation," he says.

Why, asks Potts, after the human family tree diverged into several species, has only Homo sapiens survived to the present? He suggests that the excellent adaptability of our species -- for example in contrast to the cold-adapted Neanderthals -- secured our survival throughout the climatic rollercoaster of the past.

Clearly, there is a lot left to discover about prehistoric climate change. Too many significant changes at roughly the same time make it hard for researchers to distinguish causes and effects. And perhaps it is not pure chance that the north Atlantic region -- long thought to be the key to climate variability -- is the home to a large proportion of climate researchers.