Brian Hoskins, then and now
Dec 22, 2013
Bishop Hill in Climate: Models

The question of how fair a representation the Royal Meteorological Society has given of the reliability of climate models has been discussed in a couple of posts now and readers seem to have concluded that if you wade through the obfuscation they are in fact only giving weak support to climate models. I don't think anyone has made the case that they have given a clear and balanced account though. In my opinion, by speaking of “a confidence in the models’ suitability for their application in detection and attribution studies and for quantitative future predictions and projections” the authors strongly imply that such a confidence in fact exists.

I thought it might be interesting to see what the Science-y Grantham Institute at Imperial (as opposed to the eco-warrior bit at LSE) had to say on the subject. The author team behind the submission was headed by Brian Hoskins, who has been quite vocal on the subject of climate models. Famously, he described them as "lousy" and "terrible" in an interview with the Economist in 2010.

Here's what he has to tell politicians on the subject at the end of 2013:

The IPCC AR5 WG1 dedicates one whole chapter of its report (Chapter 9) to the evaluation of climate models.  The summary to this chapter gives a fair assessment of the improvements made by climate modellers since the last Assessment report (AR4) as well as the remaining challenges.  For example, the AR5 WG1 TS suggests that some models may be too responsive to the effects of greenhouse gases.  We do not yet fully and quantitatively understand the slowdown in surface warming (see our answer to Q9 for further comment).

We agree however, with the key conclusion of the authors of Chapter 9 that, “Climate and Earth System models are based on physical principles, and they reproduce many important aspects of observed climate. Both aspects contribute to our confidence in the models’ suitability for their application in detection and attribution studies (Chapter 10) and for quantitative future predictions and projections.”

This does not imply bind acceptance of the results from the current models: continual questioning of the results and improvements in the models is crucial.

The similarity in approach to the Royal Meteorological Society is striking: there is exactly the same ambiguity, allowing the authors insinuate that the scientific community has confidence that the models are suitable for quantitative prediction while giving them plausible deniability in future. However, as you can see, the Grantham statement is at least surrounded by caveats about the unknowns of the climate system, so it's a better summary of the state of play than the weathermen's.

That said, you really don't get a sense of "lousy" and "terrible", do you?

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