January 17, 2018

expert reaction to new research on climate sensitivity

Researchers refine previous estimates of how sensitive the climate is to carbon dioxide by considering the historical variability in global temperature, in a study, published in Nature.

Prof. Gabi Hegerl FRS, Professor of Climate System Science at the University of Edinburgh, said:

“The key result in this new research is that values at the upper and lower end of the IPCC range of climate sensitivity, 1.5 to 4.5 degrees, are less likely than estimated in 2013. This is inferred from a strong correlation between climate sensitivity and climate variability. Such a relationship had been suspected for a long time on theoretical grounds, but this is the first time it has been used to constrain sensitivity.

“Questions still remain of course if there could be compensating errors in the global variability of those models that look realistic, such as high weather noise compensating low coupled ocean atmosphere variability. So the paper may not be the last word, but the result is interesting and also fits well to other lines of evidence on climate sensitivity.

“The result is important for decisions on how to tackle climate change. Having lower probability for very high sensitivity is reassuring, as very high sensitivity would have made it extremely hard to limit climate change according to the Paris targets. This is where the strongest data points are in the study. The study also affirms that climate change won’t be small, so ambitious mitigation will still be needed to meet the targets.”

Prof. Richard Allan, Professor of Climate Science at the University of Reading, said:

“The severity of future climate change depends most on current and future emissions of carbon dioxide and other greenhouse gases from human activities. But it also hinges on how sensitive Earth’s climate is to this heating effect. A highly sensitive, ‘grumpy’ climate will respond more to the pushes and kicks from heating or cooling effects.

“This new paper elegantly refines our best estimate of how sensitive climate is by considering the year to year fluctuations in global temperature. It makes sense that if climate is more sensitive, global temperature varies more from year to year in response to climatic nudges and bumps.  Using observed fluctuations, their results increase confidence that an eventual global warming of about 3 degrees Celsius can be expected every time atmospheric carbon dioxide concentrations doubles.

“This new research can make us even more confident that the best estimate predictions assessed by the Intergovernmental Panel on Climate Change have been about right all along.  Rather than warming being inconsequential or catastrophic, as some have suggested, we can be sure societies are facing a dangerous temperature rise, but one which we still have time to fix. Their conclusions confirm that human caused climate change is a serious concern but if we act now with sustained and substantial cuts in greenhouse gas emissions, societies will still be able to avoid much of the most dangerous climate change predicted by comprehensive computer simulations.

“Their results hinge to some extent upon the skill of computer simulations in representing realistic natural variations in climate yet their diversity of decade to decade ocean fluctuations may not adequately represent the real world. The 32 year satellite records of changes in energy arriving and leaving the top of Earth’s atmosphere are also not yet long enough to aid their technique. Yet, through rigorous physics involving observations and simulations, their results increase confidence in predictions of how much the planet will warm in response to continued pollution from human activity.”

Prof. Piers Forster, Director of the Priestley International Centre for Climate at the University of Leeds, said:

“Climate sensitivity has been a hard nut to crack. For nearly four decades, we atmospheric scientists have had only a rough idea of how much the Earth’s temperature will warm as carbon dioxide levels rise. Thankfully, this paper and other recent research is finally cracking this impossibly hard nut, and these scientists have produced a more accurate estimate of how the planet will respond to increasing CO2 levels.  It confirms that we will see significantly more warming and impacts this century if we don’t increase our ambition to reduce CO2 emissions; but the possibility of 6 degrees or more warming with associated devastating impacts can perhaps begin to be ruled out.”

Supplied by the Australian SMC:

Prof. Steven Sherwood, ARC Laureate Fellow at the ARC Centre for Climate System Science and UNSW Climate Change Research Centre, said:

“This important and interesting new study by Cox et al. addresses how sensitive the Earth’s climate is to human influence.  Climate contrarians say it’s not sensitive, while some models say it is very sensitive; most scientists figure it’s unlikely to be too far from a middle value (around 3C per doubling of CO2).  Middle implies about 4C of warming by 2100 if we continue on a high-emissions trajectory, which would not be good.  Low sensitivity would imply less to worry about, while high sensitivity would be really alarming.

“What this paper argues is that size and duration of natural year-to-year changes in global temperature can tell us the answer.  These natural changes are captured by all climate models, but the study shows that the stronger and more persistent they are, the more the longer-term warming.  If their approach is right it basically means the mainstream view has been correct: there is essentially no chance that contrarians have been right, but also hardly any chance of the worst-case scenario.  Both the lowest and highest previous predictions are contradicted by this new approach.

“The new study has an appealing basis in statistical mechanics, adding credibility compared to previous studies of this type, which pointed more to higher sensitivity.  However  no single study so far is without caveats—for example, the Cox approach mixes up natural variability due to El Nino, decadal variations, volcanic eruptions and air pollutants, and we know that models have different biases with respect to each of these.  There are also theoretical problems with applying their statistical approach in this way even though it seems to work.  So it is not clear whether to put more weight on this study, or the previous ones suggesting even higher sensitivity.  But one thing all the studies agree on is that climate sensitivity is not very low, i.e., it is above 2C per doubling.  As such, the new study adds to the already very strong evidence in favour of the central predictions made over the years by the IPCC, though also offering the good news of further reducing the chance of the worst-case possibilities.”

Supplied by the German SMC:

Dr Joeri Rogelj, Research Scholar Research Program Energy, International Institute for Applied System Analysis (IIASA), Laxenburg, Austria, said:

“ECS estimates are available in the literature from a wide variety of studies, which have tackled this question with diverse methods. This particular study combines a theoretical linkage between the variability in global mean temperature and ECS, with model projections and observations of historical temperatures. Other studies use observational constraints only, future model projections, or perturbations of modelling results.

“IPCC reports rely on the combined available literature to make their assessment of ECS. That means, the information and evidence provided by this study will be included in the large set of studies on ECS. Because different methods have difference strengths and weaknesses, relying on purely one study would not provide us with the most robust assessment of ECS.

“This study reports a narrower range than the range derived from the full uncertainty assessment of ECS of the IPCC. However, this is not surprising. Based on the same model data, earlier studies already found a mean ECS of 3.2°C, with a 90% uncertainty range of +- 1.3°C [1].

“The main limitation of this new study is that they infer their ECS range from an “ensemble of opportunity”, that is, a set of model simulation which were not explicitly designed to span the full range. An analogy would be to ask 50 children to bring a piece of fruit to school. The next day you might have quite some information about the spread and variety of fruits preferred by the children, but this variety will most likely not span all fruits available in the world.

“When asked to what extent a climate sensitivity ECS of 2.8°C would change the currently discussed scenarios for climate change:

Not much. ECS is not a parameter in the most advanced and most complex climate models. It is instead an emergent property due to the interplay of the various mechanisms in the Earth system. In simple climate models, ECS is used as a parameter. Here, the most common used distributions are derived from the assessed ranges of the IPCC [2]. One additional study does not change our understanding of the entire body of literature, particularly not given its limitations.

“No major changes in international climate policy can be deduced from this study. Currently international climate policy is already following a path of five-yearly cycles during which progress to limiting warming to well below 2°C and potentially 1.5°C is assessed in light of the best available science. This study mainly reinforces the importance of this regular process. If at all, this study could suggest that very high ECS is less likely, and that warming at the very high end of the distribution might be lower. That is good news and important as an additional line of evidence for ECS, but given that the center of the distribution barely moves from the earlier assessments (and it is rather the center than the extremes on which international climate policy focusses) this new information won’t result in a landslide in the international climate negotiations.”

[1] Flato, G et al (2013): Evaluation of Climate Models. In Stocker T F et al. (2013): Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge, United Kingdom and New York, NY, USA: Cambridge University Press, 741-866.

https://www.ipcc.ch/pdf/assessment-report/ar5/wg1/drafts/WG1AR5_FOD_Ch09_All_Final.pdf

[2] Rogelj J et al (2012): Global warming under old and new scenarios using IPCC climate sensitivity range estimates. Nature Clim. Change, 2(4):248-253. DOI: 10.1038/nclimate1385

https://www.nature.com/articles/nclimate1385?WT.ec_id=NCLIMATE-201204

Dr Thorsten Mauritsen, Head of Working Group Climate Dynamics, Department of the Atmosphere in the Earth System, Max Planck Institute for Meteorology, Hamburg, Germany, said:

“This study builds on 40 years old idea by the German Klaus Hasselmann (founding director of the Max Planck Institute for Meteorology in Hamburg) that natural variability of the climate system is related to long term climate sensitivity (ECS) [4]: The more sensitive the climate is to CO2, also the longer time it takes for global temperatures to return from unusual conditions or natural fluctuations (large auto-correlation of natural variability). Peter Cox and co-authors derive a relationship between auto-correlation in natural variability and climate sensitivity in climate models, and use the relationship to constrain Earth’s climate sensitivity using observations.

“The applied method is new, and the robustness of the results must now be investigated by independent scientists. At the moment it is difficult for me to imagine how the best ECS estimate by Cox et al. could be substantially biased, by more than say 10-20 percent, but the uncertainty could be larger than that indicated.

“Concerning ECS estimates in the IPCC reports, these assessments must consider all studies, as well as known and unknown unknowns. I therefore doubt that the next IPCC report will provide as narrow an estimate as found by Cox et al., even if the study is likely to have a small impact on the assessed range.

“The impact of an ECS of 2.8 degrees on climate model projections of future climate change will be relatively small. When considering projection scenarios for year 2100, as a rule of thumb, about half the change in ECS is realized as a change in global mean temperature by that time because the oceans act to temporarily dampen warming by taking up most of the excess heat. Thus, if climate models have an average ECS of 3.4 degrees, this new best estimate of 2.8 degrees leads to approximately 10 percent reduction in projected global warming by 2100.

“On shorter timescales, in the next few decades, there will be no important change because here most of the climate change is determined by human emissions of climate pollutants (CO2, CH4, aerosols etc.) and natural variations.”
[4] Hasselmann K (1976): Stochastic climate models. Tellus, 6, 473-485, 1976.

http://onlinelibrary.wiley.com/doi/10.1111/j.2153-3490.1976.tb00696.x/abstract

* ‘Emergent constraint on equilibrium climate sensitivity from global temperature variability’ by Peter Cox et al. published in Nature on Wednesday 17 January 2018. 

Declared interests

Prof. Piers Forster: Wrote a News and Views article which will be published in Nature alongside the research.

No others received.