Thomas Stocker

Carbon isotopes in ice cores defy a controversial hypothesis.

Credit: UNIVERSITÄT BERN

Around the start of the Holocene era 11,000 years ago, the levels of carbon dioxide in the atmosphere dipped slightly, only to increase again some 4,500 years later. Scientists have pondered the causes of these slight variations, which do not follow the periodic pattern of CO₂ rises and falls in the past, without finding a definitive explanation.

In 2003, US palaeoclimatologist William Ruddiman suggested that the CO₂ increase might have been a result of forest clearance by early agrarian farmers. “That hypothesis was very thought-provoking and caught the attention of many people. But it took us several years to develop the methods to produce the data,” says Thomas Stocker of the Oeschger Center for Climate Change Research at the University of Bern, Switzerland.

Stocker heads one of ten laboratories that make up the European Project for Ice Coring in Antarctica (EPICA) consortium, launched in 1996 to construct a historical record of atmospheric CO₂ and other greenhouse gases by measuring the amounts of these gases trapped deep in the ice sheets of Antarctica.

Stocker and his colleagues looked specifically at carbon isotope 'signatures' in the ice-sheet CO₂ for clues as to its sources and sinks. For example, because vegetation and soil prefer to take up CO₂ that contains the carbon-12 isotope, rather than carbon-13, a decrease in atmospheric CO₂ accompanied by a relative increase in atmospheric ¹³C signals that the terrestrial biosphere has caused the change. By contrast, CO₂ uptake by the ocean does not show much preference for either isotope.

But ¹³C isotope measurements are not easy to obtain. “Each ice-core sample consists of about 10–40 grams of ice from which you have to release and capture a tiny amount of air,” says Stocker. “You then have to look for variations in ¹³C, which makes up only about 1% of the CO₂.” Perhaps the biggest challenge was finding a way to extract the gas without contaminating it, he says. “We discovered, for example, that CO₂ interacts with the interior wall of the extraction device to produce artefacts.”

The team's perseverance paid off and they obtained the first robust ¹³C record of the early Holocene up to about medieval times. The next step was to interpret the data. “One advantage at our institute is that we have experimental physicists working alongside carbon-cycle modellers,” says Stocker. By comparing models of CO₂ outputs from various processes with actual measurements, the team came up with an explanation for the measured CO₂ levels and ¹³C record (see page 507).

A particular combination of three processes — land-biosphere uptake, carbonate compensation by the ocean, and coral-reef formation — are sufficient to explain the recorded ¹³C levels, says Stocker, which means that no other processes were involved in causing the CO₂ variations. “The hypothesis that early land-use by humans modified CO₂ levels can now be confidently rejected,” he says.

Stocker and his team will continue to look at ¹³C signatures during other time periods to understand the forces that shaped the levels of atmospheric greenhouse gases and their subsequent climate effects. The team is particularly interested in the medieval period: “That is a time of intense human activity, but we cannot yet make any statement on whether such activity had any effect on CO₂ levels because we don't yet have measurements,” says Stocker.