The consensus from the IPPC Fourth Assessment was that climate sensitivity is most likely 3.0 K, with a 66% probability that it's between 2 and 4.5 K. Now a team from the US and Spain has found a median climate sensitivity of just 2.3 K by including temperature data from the Last Glacial Maximum. What's more, at 1.7–2.6 K, their 66% probability range for climate sensitivity is much tighter.

"Previous studies have suggested that very high climate sensitivities of 10 K or more for a doubling of CO2 may be possible," Andreas Schmittner of Oregon State University, US, told environmentalresearchweb. "Using the data from the Last Glacial Maximum we show such large climate sensitivities are virtually impossible."

Schmittner believes this is good news for a world that's making little progress on cutting carbon emissions. "If the climate sensitivity was very large, serious impacts of climate change would be difficult to avoid," he said. "Our results imply that this is not the case and that there is still time to mitigate the worst consequences of possible future climate change."

Schmittner and colleagues put constraints on their estimate by using data on sea and land temperatures during the Last Glacial Maximum along with climate model simulations.

"It always struck me that the surface ocean was only 2 K colder than today, while the conditions on land were so drastically different with huge ice sheets over North America and Europe, more deserts and large changes in vegetation," said Schmittner. "I had the feeling that these data could constrain the climate sensitivity better than previous studies that used only historical (1800–present) observations."

Schmittner believed that it was important to include land, as well as ocean, surface temperatures from the Last Glacial Maximum; discussions with Peter Clark of Oregon State revealed that Patrick Bartlein of the University of Oregon had recently compiled land temperature data based on pollen analysis. "Reading the literature I became aware that nobody had used such an extensive data set before," said Schmittner.

That said, Schmittner is quick to point out that the study does not include all uncertainties. "For example, uncertainties in the effect of clouds on the Sun's radiation heating the Earth's surface were not considered, but this is a large uncertainty in future projections with complex climate models," he said. "I would like to consider this in the future to infer a more complete probability distribution for climate sensitivity."

The researchers, from Oregon State University, Princeton University, Harvard University, Cornell University, and University of Oregon, all in the US, and Universitat Autonòma de Barcelona in Spain, published their work in Sciencexpress.