While we’re being nostalgic, a couple of years ago (late 2023 into 2024) we did a series called #ECS in the Real World in which we looked at estimates of Equilibrium Climate Sensitivity, the absolute amount of warming in degrees Celsius to be expected from a relative doubling of atmospheric CO2 that were based on, of all weird things, real-world data rather than computer simulations. It’s a topic we had explored even earlier in our 2020 video The Climate Sensitivity Question. Briefly, for many decades climate scientists used computer models to estimate the sensitivity of the climate to rising CO2 levels. They usually pegged it somewhere between 1.5C and 4.5C with the “best guess” (emphasis on “guess” and definitely not “best”) at 3C. This range was first proposed in 1979 and didn’t change for nearly 40 years. But in the past 6 years climate scientists have found their pleasing models getting hotter, with ECS at the top end getting closer to 6C even though over the past 20 years a series of published estimates based on disappointing historical data showed that ECS was more likely between 1.3C and 2.6C. In its last report the IPCC ignored the hottest models and proposed a range between 2.6C and 3.9C with a best guess of 3.1C, based on what were then the most recent empirical studies. So they dropped the top end but boosted the bottom end and goosed their central estimate up a bit. But now another real world estimate has dropped from Oxford University and while it’s a bit higher than some of the ones in our #ECS series, it’s still only 2.6C. The IPCC just can’t catch a break. But they sure can invent one.

The new paper is based on a rather complex statistical model using data on temperatures, ocean heat content, greenhouse gases, volcanic activity, Arctic sea ice and a bunch of other variables from 1860 to 2023, all stirred together and run through something called a Vector Error Correction Model. It’s all Greek to us, and probably to most climate scientists too. But the fact that, with all that data, the best estimate turns out to be right at the low end of the IPCC range adds to the evidence that the IPCC keeps trying to claim a higher ECS than is justified by mere evidence. And it’s no defence for them to say that the answer falls in the range they offered. If it was a reasonable range the ECS estimates would fall on both sides of the median. But they keep coming in below it, and sometimes below the bottom end of the range.

Just like with RCP8.5, eventually the IPCC will have to start getting rid of its most biased models and start focusing only on the plausible ones. Plausible emission scenarios run through models that match real world ECS will mean no more climate panic and maybe, finally, a reasonable approach to thinking about the issue. If we ever get there, you can bet the IPCC will be the last to know. And the last to admit that they had acted irresponsibly in making the initial bad guess, and even more irresponsibly in sticking to it so stubbornly.