More and more likely that double CO2 means <2°C: New study

The results of a new approach to calculating the effect of CO₂ - using empirical observations - suggest it has a lower impact on the climate than previously thought, and its effects are being over-estimated by the IPCC.

Publishing in the American Meterological Society's Journal of Climate, a new paper called An improved, objective Bayesian, approach for applying optimal fingerprint techniques to estimate climate sensitivity, Nicholas Lewis applies objective Bayesian techniques and uses more up-to-date observational data to derive his conclusions.

Estimates of ECS (Equilibirum Climate Sensitivity, defined by the IPCC here) produce a range of temperatures that we can expect from a doubling of atmospheric CO₂ concentrations over pre-industrial levels. They're derived in three ways: from complex climate models (GCMs) which largely ignore observational evidence, simpler models constrained by observations, and direct observation.

The Lewis paper takes the second approach. Aerosols, both natural and manmade, have a potential negative feedback, with volcanic eruptions proving to have a dramatic short-term effect. Lewis uses more up-to-date estimates of aerosol forcing (from volcanos, for example) drawn from the observational data, rather than other papers, and compares global temperatures over short decadal periods, to remove the effect of volcanic eruptions. He also uses a technique called objective Bayesian analysis. Conventional subjective Bayesian analysis relies on highly subjective uniform priors, aka deliberatively-informative "expert" priors, as parameters.

The results don't diverge completely from the IPCC estimates, but they are significantly lower. Lewis finds mode and median climate sensitivity of 1.6°C, with 90 per cent confidence in a range of 1.2°C to 2.2°C. This compares to the IPCC's 2°C-4.5°C 'likely range'. Lewis also has a crack at updating the Forest 06 study [pdf] of climate sensitivity using the new techniques and data - from which emerges an ECS range of 1.1°C-2.2°C.

What gives, then?

Very few people disagree with the basic fact that the greenhouse gas CO₂ warms the climate, but without some kind of positive feedback mechanism, it doesn't add very much: around 1°C-1.2°C per doubling of CO₂. (See this discussion on no-feedback sensitivity). The global warming "crisis" emerged from a belief that small rises in CO₂ concentrations result in large knock-on effects, or strong positive feedbacks. These remain conjectural, as the forcings and feedbacks are poorly understood. Just how much of an effect does a rise in CO₂ have - a little, or a lot? Hence the importance of new and better studies in the area of climate science dealing with "attribution".

Lewis finds that in recent years neither the global temperature nor ocean heat uptake have changed very much, while CO₂ concentrations have continued to rise. Therefore, the climate sensitivity must be lower.

How insensitive

Lewis' is the latest of a clutch of recent papers, all pointing to a lower CO₂ sensitivity range than assumed by activists and alarmists - which were used to influence climate policies which favour immediate reductions in carbon dioxide over less costly longer-term measures.

"The average warming rate of 34 CMIP5 IPCC models is greater than observations, suggesting models are too sensitive to CO2," climate scientist John Christy wrote in testimony to Congress last year. Assertions of high climate sensitivity contained "overstated confidence" he added.

In 2011 Schmittner posited 2.3°C, as we reported here. The previous year NASA studies suggested 1.64°C for a doubling. Terje Berntsen at the University of Oslo has suggested only 1.2°C-2.9°C with 1.9°C as a mean. Another Bayesian analysis by Magne Aldrin [source - discussion] posited 1.2°C to 3.5°C at 90 per cent confidence.

The differences may not sound like a lot. But the global mean temperature has risen by around a degree since 1850, and the current CO₂ concentration is rising by a mere 2 ppm (parts per million) annually - it's around 390ppm at present. Any changes will be slow, giving policy makers and technologists a long time to prepare, and develop economical low-carbon alternatives.

The objection to short-term measures is that renewable sources of energy are not currently economically viable. As the firmly warmist Economist noted, if a mere 2°C rise over a long time scale is likely, then "perhaps the world should seek to adjust to (rather than stop) the greenhouse-gas splurge. There is no point buying earthquake insurance if you do not live in an earthquake zone."

Lewis told the Reg via email that the AMS study differs from earlier publicized work, with the strength of aerosol forcing and ocean effective vertical diffusivity (which controls the time-evolution of ocean heat uptake) now inferred from a range of observational data rather than taken from other studies.

"This means that a fairly complex (but calibrated) climate model has to be used to estimate the correspondence between the three climate system parameters involved and each of the various observational data variables. Moreover, the statistical issues involved in deriving probability density functions (PDFs) for the parameters are complex - that is the main area in which my paper improves on previous methods, as used for most of the climate sensitivity PDFs given in IPCC AR4," he adds.

Lewis previewed his paper here and added further comments here. You can find it at the AMS here. ®