Climate sensitivity, or how much the Earth's surface will warm if we double CO2, is one of the major questions in climate science. I take a look at 142 studies of sensitivity over last the two decades in the new @CarbonBrief explainer carbonbrief.org/explainer-how-…
In general there is little evidence that the range of sensitivity has narrowed over time, or that it has gotten higher or lower than our estimates from 20 years ago. 2/5 
Plenty of individual studies show a narrower range, though they are not particularly consistant with eachother. Similarly, some approaches tend to show different sensitivity than others, with instrumental-based analyses in particular giving lower sensitivity estimates. 3/5
However, many scientists have criticized instrumental approaches for using incomplete observational data, not accounting for the changing nature of climate feedbacks over time, and measuring a state of the earth's climate far from equilibrium. 4/5
In particular, there appears to be a pattern of sea surface temperatures and tropical cloud cover over the last few decades that gives an appearance of lower sensitivity in both observations and observationally-forced models but is not necessarily related to sensitivity 5/5
As @DrKateMarvel and colleagues suggest, "the specific internal variability experienced in recent decades provides an unusually low estimate of ECS".
Thanks to @Knutti_ETH for doing the bulk of the work in assembling these studies in his 2017 Nature @NatureGeosci paper: nature.com/articles/ngeo3…
And helpful comments/reviews from @Knutti_ETH, @karmour_uw, @DrKateMarvel, @AndrewDessler, and others.
