A new paper undermines totally inaccurate claims in the DOE report.
Readers will recall the litany of cherry picks, misrepresentations and evasions in the CWG DOE report that we discussed multiple times last year. In particular, we highlighted an egregiously bad presentation by Steve Koonin at UT Austin. In that presentation, there was a figure presented (around 29:00) providing (it seemed) evidence that CMIP6 climate models were missing out on a massive warming of the tropical lower stratosphere and that the models had (quote) “a fundamental misunderstanding … of how the atmosphere warms”.
This was not just an overstatement based on a biased comparison, but rather, it was totally wrong. Notably, while the DOE report made similar claims, it did not include this figure – which has not been published anywhere else either.
The new paper (Santer et al., 2026) demonstrates the errors clearly. First, they use the CMIP6 data for the 2020-2025 tropical MSU/SSU diagnostics (similar to the comparisons that we update every year) and compare that data to the satellite temperature data at multiple reported levels (LT, TT, LS, and SSU1-3) (note that TT is TMT corrected for the lower-stratospheric influence).
The recent cooling of the stratosphere is clearly visible in both the models and the observations (very consistently), as is the warming in the troposphere, and in both cases there is a significant offset from the control run trends (indicating that there is a clear forced trend). The only level where there isn’t a clear departure from the control run trends, is the lower stratosphere, but there the observations and models (historical and control) are still indistinguishable.
Very clearly, the observed trends in the tropical lower stratosphere are within ±0.1ºC/decade and the signals are expected to be small compared to the noise. On the other hand, the observed trends shown in Koonin’s (unpublished) figure are on average ~0.27ºC/dec (range of 0.1 to 0.33ºC/dec). These two calculations cannot both be correct. And indeed, John Christy confirmed (in the paper’s acknowledgements) that the data in Koonin’s graph was just wrong (though it’s not exactly obvious how).
Thus Koonin’s claims are not only based on cherry-picks but also on fundamentally incorrect comparisons.
The point here (and in the new paper) is not merely to correct the record, but rather to illustrate why assessments are done the way they are. Relying on half-baked bespoke calculations that have not been peer-reviewed (or even sanity-checked) is just a recipe for embarrassment. Assessments (like the NCA or the IPCC reports) have got better over the years at trying to minimise this kind of error and are now quite good at it (by relying only on published references and extensive peer review at all stages). The cos-playing being done by the CWG simply recapitulates all the pitfalls that led to the modern assessment process in the first place, making the case (again) that these efforts have to be serious to avoid wasting everyone’s time.
References
- B.D. Santer, S. Solomon, D.W.J. Thompson, Q. Fu, and A.E. Dessler, "Misrepresentation of Human Climate Fingerprint in DOE Report", Earth's Future, vol. 14, 2026. http://dx.doi.org/10.1029/2026EF008831
Comment Policy:Please note that if your comment repeats a point you have already made, or is abusive, or is the nth comment you have posted in a very short amount of time, please reflect on the whether you are using your time online to maximum efficiency. Thanks.