"Climate sensitivity is a standard measure of climate change," team leader Gerard Roe told environmentalresearchweb. "It is basically a thought experiment in which you double the amount of carbon dioxide in the atmosphere, wait for the system to adjust and estimate the resulting change in temperature. A system with a high sensitivity has a large warming in response to the carbon dioxide, while one with a low sensitivity has a weak temperature response."

When going through this exercise, climate scientists find that some warming is certain but that the amount of warming is much less certain. Earlier this year, the Intergovernmental Panel on Climate Change (IPCC) concluded that there is a two in three chance of warming between 2–5°C because of increased carbon dioxide levels and a 1 in 3 chance the warming will lie outside this range. There may even be a small chance that it could be up to 8°C or higher.

"Our work provides a simple explanation of what that range of uncertainty is," explains Roe. "Fundamentally, the climate system is an 'amplifier'. If nothing else changed in the climate but the concentration of greenhouse gases, a doubling of carbon dioxide would ultimately lead to a temperature change of about 1.2°C. However, because of internal processes within the climate system – such as changing snow cover and clouds and water vapour in the atmosphere – our best estimate is that the actual warming would be two to four times larger than that."

This means that if the climate system amplifies its response to greenhouse gases, it will also amplify its response to other physical processes. "Uncertainty and sensitivity are inextricably linked,” says Roe. “The two go hand-in-hand."

Roe and colleague Marcia Baker have come up with a simple mathematical description of this amplification and have shown that it describes the behaviour of many previous studies of climate sensitivity well. More importantly, they found that even if we improve our understanding of climate processes, we do not reduce the uncertainty in the range of temperatures predicted.

"That is a bit odd – normally you hope that if you study a problem more, you will become more confident, or in other words, reduce the uncertainty in the answer," says Roe. "This has not happened for climate and it got us thinking. We wanted to ask the question of how uncertainties in the understanding of physical processes translate into the uncertainties in the response of the system."

The new equation will allow climate modellers to obtain meaningful results after running a model just a few times, instead of having to run it thousands of times, adjusting various climate factors each time. It also shows that a relatively small change in climate processes could lead to extreme climate sensitivity. “This means that if we just stabilize emission concentrations today, we may still risk the highest temperature change shown in the models,” adds Roe.

The work was published in Science.