Climate change.

Good news at last?

The climate may not be as sensitive to carbon dioxide as previously believed.

CLIMATE science is famously complicated, but one useful number to keep in mind is "climate sensitivity". This measures the amount of warming that can eventually be expected to follow a doubling in the atmospheric concentration of carbon dioxide. The Intergovernmental Panel on Climate Change, in its most recent summary of the science behind its predictions, published in 2007, estimated that, in present conditions, a doubling of CO2 would cause warming of about 3°C, with uncertainty of about a degree and a half in either direction. But it also says there is a small probability that the true number is much higher. Some recent studies have suggested that it could be as high as 10°C.

If that were true, disaster beckons. But a paper published in this week's Science, by Andreas Schmittner of Oregon State University, suggests it is not. In Dr Schmittner's analysis, the climate is less sensitive to carbon dioxide than was feared.

Existing studies of climate sensitivity mostly rely on data gathered from weather stations, which go back to roughly 1850. Dr Schmittner takes a different approach. His data come from the peak of the most recent ice age, between 19,000 and 23,000 years ago. His group is not the first to use such data (ice cores, fossils, marine sediments and the like) to probe the climate's sensitivity to carbon dioxide. But their paper is the most thorough. Previous attempts had considered only small regions of the globe. He has compiled enough information to make a credible stab at recreating the climate of the entire planet.

The result offers that rarest of things in climate science-a bit of good news. The group's most likely figure for climate sensitivity is 2.3°C, which is more than half a degree lower than the consensus figure, with a 66% probability that it lies between 1.7° and 2.6°C. More importantly, these results suggest an upper limit for climate sensitivity of around 3.2°C.

Before you take the SUV out for a celebratory spin, though, it is worth bearing in mind that this is only one study, and, like all such, it has its flaws. The computer model used is of only middling sophistication, Dr Schmittner admits. That may be one reason for the narrow range of his team's results. And although the study's geographical coverage is the most comprehensive so far for work of this type, there are still blank areas-notably in Australia, Central Asia, South America and the northern Pacific Ocean. Moreover, some sceptics complain about the way ancient data of this type were used to construct a different but related piece of climate science: the so-called hockey-stick model, which suggests that temperatures have risen suddenly since the beginning of the industrial revolution. It will be interesting to see if such sceptics are willing to be equally sceptical about ancient data when they support their point of view.