Are standard estimates of 'climate sensitivity' too low?

Future rises in temperature depend on two separate numbers. First, how much CO2 and other greenhouse gases are added to the atmosphere and, second, how much the climate is likely to vary in response to increases in the levels of these gases in the atmosphere. A new paper from Kirsten Zickfeld and others looks carefully at the opinions of fourteen leading climate scientists on the latter of these two important figures. (1)  The conclusions suggest that the standard view may be too optimistic. The IPCC’s most recent report (in 2007) provided an estimate of what is often called the ‘climate sensitivity’, the guess – and it is likely more than a guess -  at how fast temperature is likely to change as CO2 rises. Assessment Report 4 concluded that temperatures were ‘likely’ to rise between 1.5 and 4.5 degrees as a long-term consequence of a doubling of pre-industrial levels of carbon dioxide. (‘Likely’ in the IPCC’s language means a probability of between 66% and 90%).  More precisely, the 2007 IPCC document suggested that the median estimate of ‘climate sensitivity’ is about 3 degrees, exactly half way between 1.5 and 4.5 degrees.

This 3 degree figure has assumed a central importance in the discussions of policymakers. We can make reasonably accurate guesses about the tonnage of fossil fuels we can burn before we double the 280 parts per million of CO2 in the atmosphere. So scientists and others can then use the ‘climate sensitivity’ figure to estimate the impact of various scenarios for cutting the growth of emissions on the likely temperature change in the future.

Bodies such as the UK’s Committee on Climate Change have given a crucial role to the 3 degree number. The Committee’s extremely impressive first report in late 2008 carefully calculated how much global emissions needed to decline by 2050 for the most likely temperature increase to be about 2 degrees by the end of the century and it relied on the key figure for ‘climate sensitivity’. It also used the 3 degree number as a crucial input in its calculation of what emissions could be permitted if the world is to have a less than 1% chance of exceeding an extremely dangerous 4 degree rise. The Committee’s world-leading carbon budgets are largely reliant on the reasonableness of the 3 degree estimate. If the figure is too low, then world emissions would have to be cut even faster if we are to avoid temperature rising more than the Committee’s targets.

Zickfeld’s paper suggests that climate scientists now believe that the 3 degree figure  is too low.  The IPCC’s 2007 report used seven scientific papers to provide a consensus figure for climate sensitivity. Only three of this reports suggested a median figure of 3 degrees or above. The lowest was below 2 degrees. Zickfeld says that the current estimate of the experts interviewed for the paper is now not 3 degrees but nearly 3.5 degrees.  No-one in the survey believed that climate sensitivity was less than 3 degrees. Interestingly, four of the fourteen scientists had participated in a similar survey in 1995 and all their estimates of ‘climate sensitivity’ had risen, in one case by a very substantial amount. The increase from 3 to 3.5 degrees may seem a small change, but it actually suggests a substantial upward revision in the expected global response to increased levels of climate changing gases.

Perhaps even more importantly, the climate experts suggested that they were no more confident about the accuracy of their predictions than the IPCC was in its assessment of pre-2007 research. The money and time going into climate prediction isn’t yet giving scientists a sense that uncertainty about the speed of global warming  is improving. Moreover, the interviewees were pessimistic that we would know much more even in twenty years time. These conclusions are very worrying: not only are some of the world’s top climate scientists increasing their estimate of ‘climate sensitivity’, they are also no more certain about the distribution of probabilities than they were. As an aside, it remains extremely difficult to convey to policymakers that the width of the distributions of probability of temperature change matters as much as the central estimate.

We now have a possibility that a cherished figure – 3 degrees as the central estimate of ‘climate sensitivity’ – is too low. What are the implications? Let’s take just the UK Climate Change Committee’s target of assuring that the risk of a more than 4 degree rise by 2100 is less than 1%.  It produced its recommendation that UK emissions should peak by 2030 and then fall at 3% a year in order to achieve this result. (Global emissions, not just the UK, will have to fall by about 50% by 2050 in the same package). If the true ‘climate sensitivity’ is 3.5 degrees, my rough calculations suggest that even if the globe meets the targets the chance of a 4 degree temperature rise by 2100 is about 3%, not the less than 1% that the Committee targets. In order to reduce the risk back down to 1% - which already seems an unacceptably high figure to me – the rate of decline in UK emissions from 2016 needs to be about 4%, not the 3% currently specified.

It is an unhappy truth that the news about the climate always seems to be worsening. The Climate Change Committee would do us all a service if it now assessed whether it needs to revise its projections in the light of higher expectations of temperature rises from future greenhouse gas emissions.  

(1)    Kirsten Zickfeld, M. Granger Morgan, David J. Frame and David W. Keith, Expert judgments about transient climate response to alternative future trajectories of radiative forcing, Proceedings of the National Academy of Sciences, online edition, 28th June 2010.