A drop in carbon dioxide appears to be the driving force that led to the Antarctic ice sheet's formation, according to a recent study led by scientists at Yale and Purdue universities of molecules from ancient algae found in deep-sea core samples.

The key role of the greenhouse gas in one of the biggest climate events in Earth's history supports carbon dioxide's importance in past climate change and implicates it as a significant force in present and future climate.

The team pinpointed a threshold for low levels of carbon dioxide below which an ice sheet forms in the South Pole, but how much the greenhouse gas must increase before the ice sheet melts - which is the relevant question for the future - remains a mystery.

Matthew Huber, a professor of earth and atmospheric sciences at Purdue, said roughly a 40 percent decrease in carbon dioxide occurred prior to and during the rapid formation of a mile-thick ice sheet over the Antarctic approximately 34 million years ago.

A paper detailing the results was published Thursday (Dec. 1) in the journal Science.

"The evidence falls in line with what we would expect if carbon dioxide is the main dial that governs global climate; if we crank it up or down there are dramatic changes," Huber said. "We went from a warm world without ice to a cooler world with an ice sheet overnight, in geologic terms, because of fluctuations in carbon dioxide levels."

For 100 million years prior to the cooling, which occurred at the end of the Eocene epoch, Earth was warm and wet. Mammals and even reptiles and amphibians inhabited the North and South poles, which then had subtropical climates.

Then, over a span of about 100,000 years, temperatures fell dramatically, many species of animals became extinct, ice covered Antarctica and sea levels fell as the Oligocene epoch began.

Mark Pagani, the Yale geochemist who led the study, said polar ice sheets and sea ice exert a strong control on modern climate, influencing the global circulation of warm and cold air masses, precipitation patterns and wind strengths, and regulating global and regional temperature variability.

"The onset of Antarctic ice is the mother of all climate 'tipping points,'" he said. "Recognizing the primary role carbon dioxide change played in altering global climate is a fundamentally important observation."

There has been much scientific discussion about this sudden cooling, but until now there has not been much evidence and solid data to tell what happened, Huber said.

The team found the tipping point in atmospheric carbon dioxide levels for cooling that initiates ice sheet formation is about 600 parts per million. Prior to the levels dropping this low, it was too warm for the ice sheet to form. At the Earth's current level of around 390 parts per million, the environment is such that an ice sheet remains, but carbon dioxide levels and temperatures are increasing.

The world will likely reach levels between 550 and 1,000 parts per million by 2100. Melting an ice sheet is a different process than its initiation, and it is not known what level would cause the ice sheet to melt away completely, Huber said.

"The system is not linear and there may be a different threshold for melting the ice sheet, but if we continue on our current path of warming we will eventually reach that tipping point," he said. "Of course after we cross that threshold it will still take many thousands of years to melt an ice sheet."

What drove the rise and fall in carbon dioxide levels during the Eocene and Oligocene is not known.

The team studied geochemical remnants of ancient algae from seabed cores collected by drilling in deep-ocean sediments and crusts as part of the National Science Foundation's Integrated Ocean Drilling program. The biochemical molecules present in algae vary depending on the temperature, nutrients and amount of dissolved carbon dioxide present in the ocean water.

These molecules are well preserved even after many millions of years and can be used to reconstruct the key environmental variables at the time, including carbon dioxide levels in the atmosphere, Pagani said.

Samples from two sites in the tropical Atlantic Ocean were the main focus of this study because this area was stable at that point in Earth's history and had little upwelling, which brings carbon dioxide from the ocean floor to the surface and could skew measurements of atmospheric carbon dioxide, Huber said.

In re-evaluating previous estimates of atmospheric carbon dioxide levels using deep-sea core samples, the team found that continuous data from a stable area of the ocean is necessary for accurate results.

Data generated from a mix of sites throughout the world's oceans caused inaccuracies due to variations in the nutrients present in different locations.

This explained conflicting results from earlier papers based on the deep-sea samples that suggested carbon dioxide increased during the formation of the ice sheet, he said.

Constraints on temperature and nutrient concentrations were achieved through modeling of past circulation, temperature and nutrient distributions performed by Huber and Willem Sijp at the University of New South Wales in Australia.

The collaboration built on Huber's previous work using the National Center for Atmospheric Research Community Climate System Model 3, one of the same models used to predict future climates, and used the UVic Earth System Climate Model developed at the University of Victoria, British Columbia.

"The models got it just about right and provided results that matched the information obtained from the core samples," he said.

"This was an important validation of the models. If they are able to produce results that match the past, then we can have more confidence in their ability to predict future scenarios."

In addition to Huber, Pagani and Sijp, paper co-authors include Zhonghui Liu of the University of Hong Kong, Steven Bohaty of the University of Southampton in England, Jorijntje Henderiks of Uppsala University in Sweden, Srinath Krishnan of Yale, and Robert DeConto of the University of Massachusetts-Amherst.

The National Science Foundation, Natural Environment Research Council, Royal Swedish Academy and Yale Department of Geology funded this work.

In 2004 the team used evidence from deep-sea core samples to challenge the longstanding theory that the ice sheet developed because of a shift from warm to cool ocean currents millions of years ago.

The team found that a cold current, not the warm one that had been theorized, was flowing past the Antarctic coast for millions of years before the ice sheet developed. Huber next plans to investigate the impact of an ice sheet on climate.

"It seems that the polar ice sheet shaped our modern climate, but we don't have much hard data on the specifics of how," he said. "It is important to know by how much it cools the planet and how much warmer the planet would get without an ice sheet."