What DECC knew
Aug 5, 2015
Bishop Hill in Climate: Sceptics, Climate: Statistics, DECC

Greenpeace have been doing some rather odd FOI work in recent months. It seems they have decided to investigate the series of parliamentary questions that Lord Donoughue put to to DECC ministers about the Met Office's statistical reasoning. Readers will recall that these questions were formulated with Doug Keenan's advice, were aimed at determining how the Met Office justified its claim that recent temperature rises were statistically significant, and that the eventual result, after months of non-answers from the Met Office, was that they effectively withdrew the claim.

The documents Greenpeace have made public are very interesting but I'm not sure that our environmentalist friends have considered exactly what it is they have got.

It does rather come across as if the DECC team wanted to "move on". In Document 4, the briefing ahead of the meeting between Keenan, Donoughue and the DECC team of Baroness Verma, David Mackay and David Warrilow, officials list their objectives for the meeting as being:

It would appear then that responding to the statistical concerns raised was not a priority. In fact, the only discussion of these ahead of the meeting was the following:

The argument is based around the comparison of a simple linear 'line-of-best-fit?) model and a more complicated statistical model, favoured by Doug Keenan. A simple linear model shows a statistically significant increase in global temperatures, using Keenan's more complex model does not. It is asserted that claims of statistical significance in the global temperature record are the Met Office's best evidence for concern around global warming.

Thus these Parliamentary Questions appear to be an attempt to undermine the general recognition by the scientific community that the rise in global surface temperatures over the last century is significant. REDACTED

So the DECC team seem internally quite clear that Keenan's calculations are correct (and recall that his question was only to give the calculation!). It is perhaps unsurprising therefore that when the meeting took place (Document 5 has the minutes) they didn't want to talk about any of the points Keenan and Donoghue were making about statistics and time series. So we see Baroness Verma trying to move the conversation onto the wider question of the use of statistics in climate work, David Mackay wanting to talk about physical models, and then Verma wondering whether they should talk about the use of statistics in DECC's 2050 calculator. Any attempt to discuss the point at issue seems to have been squashed.

So why the prevarication? The clue is in the extract above. The bit about "the best evidence for concern" about global warming is not something I can recall Keenan ever having said. As far as I know, he examined the Met Office's claims on statistical significance because the Met Office had stated them as important evidence in the global warming debate. This would appear to be a bit of "sexing up" of the facts for the benefit of ministers ahead of the meeting.

Why would they do that? Well the answer is in the next sentence, when they tell ministers that the parliamentary questions were "an attempt to undermine the general recognition by the scientific community that the rise in global surface temperatures over the last century is significant". So while the statistics showed that the warming was not significant, there was a "general recognition" among scientists that it was. And they must have been telling ministers this in the full knowledge that the "simple linear model" is recognised by statisticians on all sides as being wholly inadequate therefore seems wholly culpable. This is astonishingly culpable.

I think we can see that DECC officials were unconcerned about the science. What actually concerned them was that an honest response might give "ammunition to the sceptics" (as someone once said). Seen in this light, what Greenpeace has revealed is as damning an indictment of the integrity of DECC officials as you could ever hope to find.

Update on Aug 6, 2015 by Registered CommenterBishop Hill

Lord Donoughue emails:

There were so many Questions because they refused to answer my repeated request for observational evidence. And they finally admitted their statistical model was useless.
They never responded to my request at the meeting for observational evidence, though they promised to. Because they dont have any evidence. Just forecasting models which depend on the assumptions built in - and which have been wrong in forecasting significant warming this century.
Doug Keenan's thoughts are in the comments thread.
Article originally appeared on (http://www.bishop-hill.net/).
See website for complete article licensing information.