Not waving but drowning

A couple of journalists seem to have shown some interest in Nic Lewis's critique about the UKCP09 climate projections. This may explain why the Met Office has suddenly issued a response a week after Nic's report came out:

Today an article by the GWPF think-tank looks at one element of this – the UK’s official climate projections, known as UKCP09, which were produced by the Met Office.

It claims the Met Office climate model used to make those projections, HadCM3, contains an error and that, because of this error, the projections overestimate warming. The GWPF’s article, however, accepts that the claims of an error have not been substantiated.

UKCP09 used a sophisticated method that used both model projections and observations to provide a range of potential future warming which attempts to take in the uncertainties in model parameters. The GWPF article fails to note that UKCP09 also used information from many other climate models, and that the projections were independently reviewed prior to publication.

Ultimately there is nothing in the GWPF article which undermines UKCP09 or the way climate models, including the Met Office’s HadCM3, project future temperature changes.

This is an astonishingly lightweight argument. Perceptive readers will note that it doesn't actually address the substantive criticisms made in Nic's report and my briefing. There is just a lot of handwaving going on. Or perhaps the bigwigs at the Met Office are not waving but drowning.

Meanwhile, Nic emails the text of the comment he has left under the Met Office blog post (currently in moderation).

I didn't claim that the HadCM3 model contained an error (and nor, so far as I can see, does the GWPF document claim it did, contrary to what is stated). I wrote that, since the observational data strongly contraindicate aerosol forcing being highly negative, "whatever the actual level of ECS [climate sensitivity], HadCM3-derived ECS estimates are bound to be high." This indicates that the HadCM3 study was unsuitable for use in generating projections that are, supposedly, constrained by observations.

The Met Office claimed in their July report The recent pause in global warming(3) that, in the Harris et al 2013 study underlying UKCP09 "uncertainty in the response of the climate system to CO2 forcing is comprehensively sampled". That is clearly incorrect.

The statement "The GWPF’s article, however, accepts that the claims of an error have not been substantiated." is misleading since it suggests that the claims referred to in the article (those in my critique of he Met Office report) have been refuted. They most certainly have not been refuted. The Met Office has not yet even made a substantive response to my claims.

The statement says "UKCP09 used a sophisticated method that used both model projections and observations to provide a range of potential future warming which attempts to take in the uncertainties in model parameters." deserves comment. The method was indeed sophisticated, and there were attempts to take in uncertainties about model parameters. I said as much in my critique of the Harris et al 2013 study. But sophisticated does not imply valid, and attempts are not always successful. And in fact the taking into account in the study, within limits, of uncertainties in model parameters did not enable the HadCM3 model to simulate a climate system in which both climate sensitivity and aerosol forcing were modest, as suggested by several recent observational studies.

The statement that "the projections were independently reviewed prior to publication" is irrelevant. It is doubtful whether any reviewer would have spotted the problem that I have identified, which involved some detailed analysis and modelling work.

The use of information from other climate models that is referred to is also irrelevant, since use of that information did not overcome the fundamental problem – the inability of the HadCM3 model to simulate a climate system in which both climate sensitivity and aerosol forcing are modest.

The Met Office have not provided any justification for the unsupported claim that "Ultimately there is nothing in the GWPF article which undermines UKCP09 or the way climate models, including the Met Office’s HadCM3, project future temperature changes."

Update on Sep 23, 2013 by Bishop Hill

The first of the news stories is live, in the shape of Fiona Macrae at the Mail.

The Met Office’s global warming predictions are flawed and could result in millions of pounds being squandered, it is claimed.

A report for a think tank led by former Tory chancellor Lord Lawson says a computer programme behind figures that shape climate change policy is biased in favour of higher temperatures.

Large sums of public and private sector money could be ‘malinvested’ in everything from wind farms to heat-proof road surfaces as a result, it claims.

Bob Ward's response is hilarious:

This is a political stunt by Lord Lawson’s Global Warming Policy Foundation to try to distract attention from the IPCC report.

Comments

[NW]

It appears to me that whatever accuracy they do achieve in short term forecasting over that which is simply "same as yesterday" is obtained from traditional observation of weather patterns rather than models and supercomputers.

This of course only works when the weather patterns are moving in predictable ways; there were periods this summer when this was identifiably not happening and consequently UKMO clearly didn't have much clue what was going to happen, resulting in some extremely "wide" forecasts which really didn't say anything much of use.