The IPCC's computer models have proved just as wrong in predicting global temperatures as the Met Office has been in forecasting those mild winters and heatwave summers, says Christopher Booker.

Sir John Houghton
© unknownSir John Houghton, former head of the Met Office, edited the first three IPCC reports between 1990 and 2001

Everyone has enjoyed the discomfiture of the Met Office, caught out over its April forecast that we were in for a "barbecue summer" - not least because this is the third year running that our weathermen have got their predictions for both summer and winter hopelessly wrong. In 2007 and 2008 they forecast that summers would be warmer and drier, and winters milder than average - just before temperatures plunged and the heavens opened, deluging us with abnormal rain or snow according to season.

One cause of the blunders that have made the Met Office a laughing stock is less widely appreciated, however. It is that the multi-million pound computer it uses to assist its short-term forecasting for Britain is also one of the four main official sources of data used by the UN's Intergovernmental Panel on Climate Change (IPCC) to predict global warming. In this respect the IPCC's computer models have proved just as wrong in predicting global temperatures as the Met Office has been in forecasting those mild winters and heatwave summers.

Back in 1990, Mrs Thatcher, temporarily under the spell of the prophets of runaway global warming, authorised lavish funding for the then-head of the Met Office, Sir John Houghton, to set up its Hadley Centre in Exeter, as a "world-class centre for research into climate change". It was linked to the Climate Research Unit (CRU) at the University of East Anglia, to create a record of global temperatures based on surface weather stations across the world, a data set known as HadCrut. Sir John himself played a key role at the top of the new IPCC as chairman of its scientific working group.

Sir John was a fervent believer in the theory that the cause of global warming is man-made CO2, and the HadCrut computer models, run by his CRU ally Professor Phil Jones, were programmed accordingly. Sir John (and the Hadley Centre) continued to play a central part in the running of the IPCC, selecting many of the contributors to its reports that were the main driver of global warming alarm. He and Prof Jones were also prominent champions of the IPCC's notorious "hockey stick" graph, which rewrote climate history by suggesting that global temperatures had suddenly shot up in the late 20th century to easily their highest level in history.

In recent years, however, the whole theory has come under increasing fire because, as CO2 levels continue to rise, temperatures have failed to follow suit as the IPCC's computer models predicted they should. Part of the reason why the Met Office has made such a mess of its forecasts for Britain is that they are based on the same models which failed to predict the declining trend in world temperatures since 2001.

In recent months, in fact, a curious little drama has been unfolding over attempts by Steve McIntyre, a Canadian statistical expert, to get the Met Office and the CRU to divulge the computer data on which they base their temperature record. Mr McIntyre was not only the chief demolisher of the "hockey stick", showing how it was based on a seriously skewed computer model, but later exposed the "adjustments" which had skewed the other official record of surface temperatures, run by Dr James Hansen of the Goddard Institute for Space Studies. (The two other official sources of temperature data are based on satellite measurements.)

When Mr McIntyre made Freedom of Information requests to see the data used to construct the HadCrut record (as he has chronicled on his ClimateAudit blog) he was given an almighty brush-off, the Met Office saying that this information was strictly confidential and that to release it would damage Britain's "international relations" with all the countries that supplied it.

The idea that temperature records might be a state secret seems strange enough, but when the policies of governments across the world are based on that data it becomes odder still that no outsider should be allowed to see it. Weirdest of all, however, is the Met Office's claim that to release the data would "damage the trust that scientists have in those scientists who happen to be employed in the public sector".

Doesn't the Met Office realise that trust in it has already been damaged enough by its batty predictions of "barbecue summers"? If it wants to restore that trust, it should first come clean about its data, and then reprogramme its computer to give us forecasts that are not skewed by its obsession with global warming - which is not happening.