Comment Britain's Met Office has come under fire for two pieces of crystal-ball gazing involving global temperature and British rainfall. On Christmas Eve, the Met's temperature prediction for the UK was quietly revised downwards, and only merited a press release this week after physics blog Tallbloke's Talkshop noticed the change.
According to the Met's Richard Betts, an IPCC lead author and head of the Met's Climate Impacts team, the new projection the result of new climate models, with different inputs.
The new temperature prediction is 20 per cent lower than the previous estimate, with a mean deviation of 0.43°C above the 1971 to 2000 average over the next five years. If it holds true, then global temperatures will have experienced a 20-year standstill, with no statistically significant warming. The Met didn't predict, as the BBC erroneously reported, a 0.43C increase in global temperature over the next five years.
comparison of old and new temperature anomaly
In a comment at the Tallbloke science blog, Betts explained that the new cutting-edge models incorporated, "a whole new setup for simulating the fluid dynamics of the atmosphere", and are initialised using more recent atmospheric data. The models also make more detailed shorter-term projections, so it should be easier to evaluate how they are performing, Betts added.
However, MP Graham Stringer said the Met Office should have been more upfront.
"By putting out the information on Christmas Eve they were just burying bad news – that they have got their climate change forecast wrong," said Stringer.
A twenty year period without statistically significant warming doesn't falsify the theory that manmade industrial emissions are the key driver in climate change - the oceans may be storing energy that isn't yet manifest in higher atmospheric temperatures. But it certainly wasn't in the script, which raises questions over the validity of the models on which policy decisions have been made, as the BBC's Paul Hudson points out here. 11 of the last 12 annual Met forecasts erred on the warm side - so new and better models should be widely welcomed.
Another claim by the Met has also drawn fire - as the criticism directly addresses the validity of the Met's science, rather than its communications strategy. Last week the Met made a widely-reported claim that Britain is experiencing more frequent extreme rainfall. Statistical analysis of rainfall records by the Met Office claimed to show days of heavy rainfall had become more common in England since 1960. "The apparent trend mirrors increases in extreme rain seen in other parts of the world," wrote the BBC's Roger Harrabin.
For Channel 4 news, the Met's statistical press release was apocalyptic. There were "clear signals of wetter weather emerging", apparently.
But the claim has puzzled some observers, not least because last March the Met was predicting a continuing drought for the UK, advising last March that:
"The probability that UK precipitation for April-May-June will fall into the driest of our five categories is 20-25% whilst the probability that it will fall into the wettest of our five categories is 10-15%".
When the data is examined, statistician Doug Keenan finds, anything but a clear trend is apparent:
Annual rainfall (mm) England and Wales 1766-2012
In correspondence with the Met's chief scientist Julia Slingo, Keenan points out time series analysis has not been performed, and adds that it is not possible to draw the conclusions the Met does, based on recent data. The Met failed to consider autoregression, it told Lord Donoghue in a written answer to a Parliament question. Keenan concludes:
"Was the Met Office aware that there is no valid evidence for an increase when it issued its news release? Either it knew, and thus acted fraudulently in issuing the release, or it was misleading when it told Lord Donoughue that it has the requisite expertise. (I suspect the latter.)
"Taxpayers pay for the Met Office to provide expertise with climatic analysis. They are not getting anything close to what they pay for."
You can find the Keenan-Slingo correspondence here.
There's no argument that the UK had a very wet 2012, just as it had a very dry 2011. UK rainfall for April-June turned out to be 176%, 94% and 203% of normal. The Met would make four more monthly predictions a likelihood of drier-than-normal weather for the UK, all subsequent months were wetter than normal (September 117%, October 101%, November 111%, and December 150%). As one climate blogger noted dryly:
"It is very kind of Julia [Slingo] to tell us now that she knew all along it was likely to be wetter. It is just a pity, though, that she forgot to tell us at the time."