by Judith Curry
Presentations are now available from the WCRP Workshop on Earth’s Climate Sensitivity.
The website for the Workshop is [here]. Several articles on the Workshop have been written:
- Gavin Schmidt at RealClimate: Reflections on Ringberg
- Roz Pidcock at CarbonBrief
Gavin Schmidt writes:
There were two major themes that emerged across a lot of the discussions: the stability of the basic ‘energy balance’ equation that defines the sensitivity; and the challenge of estimating cloud feedbacks from process-based understanding. The connection occurs because the clouds are the cause of the biggest variation in sensitivity across GCMs.
Roz Pidcock’s article discusses some broader topics related to climate sensitivity, including the Workshop. Links to new publications that are discussed:
- Equilibrium climate sensitivity in light of observations over the warming hiatus, by Johansson, O’Niell, Tebaldi & Haggstrom [link]
- Clouds, circulation and climate sensitivity, by Bony, Stevens, Frierson, Jakob, Kageyama, Pincus, Shepherd, Sherwood, Siebesma, Sobel, Watanable, Webb [link]
The punchline of Pidcock’s article is evident from the title: Climate sensitivity is unlikely to be less than 2C, say scientists.
Re Bjorn Stevens’ paper on aerosol forcing (discussed in a post by Nic Lewis). Media Matters reports: Climate Scientist: No, My Study is Not a Death Blow to Global Warming Hysteria, based on letter written by Bjorn Stevens. The title of Bjorn Stevens’ talk at Ringberg sums up his take on the subject: Some (not yet entirely convincing) reasons why 2<ECS<3.5.
Andrew Montford tweets: “Alarmism has taken a beating” is equivalent to BS’s “highest estimates look less likely”.
So, am I buying the conclusion that climate sensitivity is unlikely to be less than 2C? In a word, NO.
Here are some interesting items I was able to glean from the Ringberg presentations. Note, only a few of the presentations really ‘worked’ as stand-alone documents; I may have very well missed some important points made at the meeting since they weren’t clear in the ppt presentations. A list of publications is provided on the Ringberg web site; presumably more can be gleaned from these publications.
Evaluating the methods
Hegerl et al.: What observed climate change can and can’t tell about transient and equilibrium sensitivity.
Punchline: A good overview presentation. Natural variability influences ECS and TCR estimates. No robust contradiction between observational and model estimates, given uncertainties. Observed climate change provides good lower limits, but cannot reliably set an upper limit particularly for ECS until we address nonlinearities and structural model problems.
Andrews, Webb and Gregory: Feedbacks, their inconstancy and dependence on SST patterns
Punchline: Feedbacks are very sensitive to warming patterns: can get ECS values from 1.3 to 5C just by changing the warming pattern in a climate model. May expelling differences between observational and climate model estimates.
Armour: Robust increase in effective climate sensitivity with transient warming
Punchline: CMIP5 models show a robust increase in climate sensitivity over time. Estimates of equilibrium climate sensitivity based on observed global energy constraints should be viewed as a lower bound.
Golaz and Zhao: Tuning the indirect effect; engineering the climate sensitivity: What should modelers do with these newly found powers?
Punchline: To some extent, magnitude of aerosol indirect effect is tunable, and climate sensitivity can be engineered. Can we build modes that represent the 20th century with different combinations of forcing and sensitivity? Eg. high sensitivity, strong indirect effect; low sensitivity weak indirect effect.
Gregory et al: The inconstancy of the climate feedback parameter
Punchline: Ocean heat uptake efficiency declines with time. TCR in the historical period may underestimate future CO2 forced warming. Climate sensitivity to volcanic forcing may be smaller than for CO2 forcing, and the sensitivity may be smaller for larger forcing. Volcanic forcing may be overestimated in magnitude by the AR5 forcing because of cloud adjustments.
Nic Lewis: Pitfalls in climat sensitivity estimation
Punchline: Problems – overstrong aerosols, bad priors, AMO influence. Most CMIP5 models have excessive TCR and ECS.
Oceans
Church et al.: Ocean heat uptake during the Argo period
Punchline: The Argo Array provides high-quality, global coverage to 2000m, from about 2006. Beware of biases from historical data bases. Interannual fluctuations in upper 500 m; deeper warming trend. Heat uptake in mid-latitudes, particularly the southern hemisphere. The majority of the heat uptake occurs south of 20S, with about the same amount of heat in the upper 500 m and 500-2000 m. The majority of the heat is stored south of 20S.
Fasullo: Understanding sea level as a constraint on climate variability and sensitivity.
Punchline: Sea level reconstructions suggest that the Grand Hiatus (1945-1975) was not forced. Interpreting climate sensitivity from the instrumental record depends critically on separating forced changes from internal variability. Assumptions regarding internal variability are key, model dependent, and challenging to validate.
Kosaka et al. Earth’s energy budget in the presence of internal variability
Punchline: During hiatus, net incoming energy decreases instead of accelerates (the traditional energy equation predicts the latter). Observed ocean heat content tendency is consistent with the internal variability hypothesis of the hiatus.
Latif et al.: The challenge of validating climate models
Interesting points: If anything an enhanced equatorial SST gradient and Walker circulation has been observed during the past 50 years. Southern Ocean internal centennial variability impacts global surface air temperature and Antarctic sea ice. The surface Southern Ocean cools, while the abyssal Southern Ocean warms.
Clouds
Bony: Do climate models overestimate cloud feedbacks?
Interesting points: There is evidence for an Iris effect . . . but not for a negative cloud feedback associated with it (so far). Could a change in convective organization with temperature affect this feedback? Remains to be investigated.
Bengtsson: What are the best temperature data to use in estimating climate sensitivity?
Interesting points: The variance in cloud feedbacks fc is almost twice as large as the net feedback fnet. This suggests that the other feedbacks compensate for variability in fc. Tuning climate sensitivity to lie within the observed spread across models is a sufficient explanation for the origin of compensation between fc and the other feedbacks. There is no reason to assume that climate sensitivity is a ‘constant of nature’ as it depends fully on the physical behavior of the climate system and changes in feedback structure.
JC note: there were several additional presentations that addressed clouds at the process level.
Structural uncertainty
I was particularly hopeful for some fundamental papers that addressed structural uncertainty of the methods used to determine sensitivity. Several presentations mention structural uncertainty, but I didn’t get much out of any of them:
- Tamsin Edwards: Multi palaeo archive constraints on climate sensitivity.
- Crucifix: Ideas and definitions from the NPG literature.
JC reflections
I was hopeful for insights into 3 topics:
- lower bounds for ECS/TCR
- upper bounds for ECS/TCR
- ideas for new structural models for ECS/TCR determination
It seems like the community at the workshop is resisting ECS<2 (other than of course Nic Lewis). Not much at all was said about inferences of upper bounds, although the ‘consensus’ seems to be creeping downwards (to 3.5 or 4.0, from and AR5 value of 4.5). Hegerl correctly states that we cannot reliably set an upper limit particularly for ECS until we address nonlinearities and structural model problems.
Re structural uncertainty, it is good to see the community acknowledging the confounding factor of natural internal variability, inconstancy of the feedback parameter, inconstancy of ocean heat uptake, sensitivity of feedbacks to spatial warming patterns, aerosols as a tuning factor for sensitivity. In light of these structural uncertainties, I find it inappropriate to state any conclusion regarding climate sensitivity with anything greater than ‘low confidence’.
The ocean presentations were quite interesting and all of them work reasonably well as stand alone presentations.
Bottom line is that I don’t see anything here that makes me want to question my recent assessment on this topic: Climate sensitivity – lopping off the fat tail.
Filed under: Sensitivity & feedbacks