An amateur meteorologist in Canada has embarrassed Nasa scientists into admitting that some of the data they used to show significant recent increases in global warming is flawed.

http://www.guardian.co.uk/environment/2007/aug/16/1

Blogger gets hot and bothered over Nasa's climate data error

· Amateur discovers 1934, not 1998, is hottest year
· Climatologists insist trend is still towards warming

• Richard Luscombe in Miami
• The Guardian
• Thursday August 16 2007

An amateur meteorologist in Canada has embarrassed Nasa scientists into admitting that some of the data they used to show significant recent increases in global warming is flawed.

As a result of Stephen McIntyre's calculations, climatologists at the Goddard Institute of Space Science in New York now accept that 1934 was historically the United States' hottest year since records began, not 1998 as they had claimed. It also turns out that five of the 10 warmest years on record in the US occurred before 1939, and only one is from the 21st century, raising questions over the statistics used in Al Gore's environmental film An Inconvenient Truth to highlight the faster pace of climate change.

"They have managed it rather poorly," said Mr McIntyre, a prolific internet blogger from Toronto who pointed out the gaffe to Nasa in an email. He noticed that temperature deviation readings from numerous weather-recording stations around the US showed sudden and inexplicable leaps after 2000. He says the agency refused to share with him the complex methodology it uses to calculate trends from the data, then quietly changed statistics to rewrite history without explanation.

Mr McIntyre said: "I come from a background where you have to announce bad results. They might not like the fact they made a small embarrassing error but if it was me I'd have announced the results and put the best spin on it that I could. I would not have left myself open to the suggestion that I was not being forthcoming."

Climate researchers at the Goddard centre, meanwhile, say Mr McIntyre is making a mountain out of a molehill and that the differences in the recalculated temperatures, hundredths of one degree, are so insignificant as to have no impact on the overall trend towards global warming.

The cause of the error, they say, was a switch to a new data-collection system in 2000 and a faulty assumption that the old and new methods matched, which last week led to a recalculation of the figures.

Now 1934 is the hottest year on record in the US at an average of 1.25C higher than normal; 1998 is second at 1.23C, and 1921 in third place at 1.15C. Under the old system, 1998 was the hottest at 1.24C above normal, with 1934 at 1.23C. 2006, newly relegated to fourth place, was also at 1.23C.

Dr Gavin Schmidt, a climate change expert at GISS and author of the website realclimate.org, said: "The idea that Nasa is faking things or that it hasn't got warmer at all is nonsense. There were some minor rearrangements in the various rankings [but] ... the longer term US averages have not changed."

"The sum total of this change? A couple of hundredths of degrees in the US rankings and no change in anything that could be considered climatically important, specifically long-term trends."

He also noted that the error related only to figures from the US, covering just 2% of the Earth's surface, so could not be applied globally.