The mind-boggling coincidence hypothesis
Feb 27, 2014
Bishop Hill

Also hot off the press is a new paper by Gavin Schmidt and colleagues. Doug McNeall reckons I'm not going to like it, but having taken a look (it's open access for registered users of the Nature website), I have to say I think it's lots of fun.

Schmidt and his colleagues are looking at the hiatus in surface temperature rises and considers why the CMIP5 ensemble all got it so wrong. In their new paper they explain that the reason for this is not – as wild-eyed readers at BH might think – that the models are wonky. In fact it's all down to an incredible, incredible coincidence

Here we argue that a combination of factors, by coincidence, conspired to dampen warming trends in the real world after about 1992. CMIP5 model simulations were based on historical estimates of external influences on the climate only to 2000 or 2005, and used scenarios (Representative Concentration Pathways, or RCPs) thereafter4. Any recent improvements in these estimates or updates to the present day were not taken into account in these simulations. Specifically, the influence of volcanic eruptions, aerosols in the atmosphere and solar activity all took unexpected turns over the 2000s. The climate model simulations, effectively, were run with the assumption that conditions were broadly going to continue along established trajectories.

Apparently, if you go back and rework all the forcings, taking into account new data estimates (add half a bottle of post-hoc figures) and 'reanalyses' of old data (add a tablespoon of computer simulation) you can bridge the gap and explain away the pause.

We conclude that use of the latest information on external influences on the climate system and adjusting for internal variability associated with ENSO can almost completely reconcile the trends in global mean surface temperature in CMIP5 models and observations. Nevertheless, attributing climate trends over relatively short periods, such as 10 to 15 years, will always be problematic, and it is inherently unsatisfying to find model–data agreement only with the benefit of hindsight.

So, with the benefit of hindsight, the climate modellers can fit their square peg into a round hole. It wasn't that the models were running too hot, it was just that nature has got it in for climate modellers.


Of course, they still have the problem that the energy budget estimates of TCR are all pointing to much lower climate sensitivity than the GCMs. These studies are, of course, strongly suggestive of the "mind-boggling coincidence" hypothesis being incorrect and the original supposition - that the models are overheated - is right. However, Schmidt and his colleagues make no attempt to address such minutiae, waving them aside, with characteristic bonhomie, as mere speculation:

We see no indication, however, that transient climate response is systematically overestimated in the CMIP5 climate models as has been speculated8, or that decadal variability across the ensemble of models is systematically underestimated, although at least some individual models probably fall short in this respect.

Told you it was fun, didn't I?

Article originally appeared on (
See website for complete article licensing information.