The Economist On Climate Sense & Sensitivity

  • Date: 13/03/14
  • J.L.P., The Economist

CLIMATE science is never settled. For several years, debate has raged around “the pause”—that is, around the causes and significance of the slowdown in global surface temperatures since 1998. The past few weeks have seen some useful explanations for it—indeed, arguably too many, see here.

But no sooner have advances occurred in one part of the field than disagreements break out in another. In this case, the area of dispute is how much the climate warms up in response to increases in carbon-dioxide concentrations.

Carbon dioxide absorbs infra-red radiation at a fixed rate. For each doubling of CO2 levels you get roughly 1°C of warming. But the Earth is warming up faster than that because the climate system has lots of feedback loops. The warming due to the carbon dioxide influences other things, such as the amount of water vapour and the number and type of clouds; those can lead to further warming, or to countervailing cooling. Add in all these effects and the amount of warming expected for a doubling of carbon dioxide—a figure known as climate sensitivity—will be more than that 1°C.

The question of how much more, though, is vexed. Last year’s assessment of climate science by the Intergovernmental Panel on Climate Change estimated the sensitivity at somewhere between 1.5°C and 4.5°C. If in reality the figure is at the low end of that range, climate change in the 21st century would still be unprecedented in human history, if not necessarily catastrophic. At the top end things would tend towards Old Testament levels of badness.

Deciding which end of the spectrum looks more plausible is hard. First because climate sensitivity estimates are derived from computer models which may not correctly capture the various feedbacks on which the result depends. And second because the thing the models are measuring is the “equilibrium” sensitivity—the theoretical warming that you would get if you doubled the carbon dioxide level overnight and then waited for a century or more to see how much warming resulted. This is not much use to policymakers.

Enter a shorter-term, and hence more useful, measure of sensitivity called the transient climate response, or TCR. This is the temperature rise you get if you double carbon-dioxide levels by increasing them at a rate of 1% a year for 70 years. This figure is lower than the equilibrium sensitivity because it does not take account of the time it takes to warm up the ocean, which would lead the world to keep warming long after the 70 years were up. Most estimates put the TCR at around 1.5°C, with a range of 1-2°C. So far, this measure has been less controversial than the equilibrium one, because it can be studied in part using direct observations.

But controversy is growing. A new study by Drew Shindell of NASA’s Goddard Institute for Space Studies argues that most TCR estimates are too low. The TCR estimated with the help of computer models is sensitive to the cooling effect of aerosols—reflective particles in the atmosphere, some of which are natural, some of which are man-made pollutants. Mr Shindell says that new observations from satellites (and other data) suggest the modelling of the interaction between aerosols and clouds in a particular set of climate models, called CMIP5, is wrong and that it exaggerates cooling. He argues that you need to adjust the models to correct for this bias. With his adjustments, the TCR comes out to be 1.7°C not 1.5°C and that the lower end of the sensitivity range (below 1.3°C) appears very unlikely.

Dr Shindell was a co-author of a paper by Alexander Otto of Oxford University and others published in Nature Geoscience last year (here) which had suggested the climate models were producing TCRs that were too hig—so he is arguing against his own previous work. But that is what scientists are supposed to do, if the information changes.

But has the information really changed that much? Another new study, by Nic Lewis (also a co-author on the Otto et al paper) and Marcel Crok—neither of whom have academic affiliations—also trawls over recent research and argues that the new data justify lower sensitivity estimates: 1.75°C for equilibrium sensitivity and 1.3-1.4°C for the TCR. Some of this work was written about in The Economist last year. Messrs Lewis and Crok put a lot of weight on the significance of the recent pause in temperature observations, which has lowered the rate at which the world is actually warming. If warming is in fact slower that the climate models in CMIP5 can account for, then the climate models are a poor guide to the TCR.

Full story