An amateur meteorologist from Toronto has embarrassed NASA scientists by catching an error in recent climate-change data. The resulting flap has led to accusations and finger-pointing over whether NASA's error was genuine.

Thanks to blogger Stephen McIntyre's calculations, climatologists at NASA's Goddard Institute of Space Science (GISS) in New York now concede that 1934 was the hottest year in U.S. history, and not 1998 as it previously reported.

McIntyre first emailed NASA on Aug. 4 saying he'd discovered an error in its climate change data from 1999-onwards. He noticed inexplicable jumps in temperature around that time and, as it turns out, NASA's temperature readings have all been too high since then, to a maximum of 0.15 C.

NASA has since adjusted its numbers. They now show that 1998 is the U.S.'s second-hottest year and that five of the 10 warmest years on record in the U.S. date from before 1939. The worldwide numbers remain unchanged, with 1998 and 2005 tied as the hottest year on record.

McIntyre's findings did more than just embarrass NASA, though. They have triggered a storm of criticism from climate-change skeptics who accuse NASA of inflating the numbers to promote global hysteria over the environment and others accusing NASA of trying to keep the error quiet.

Meanwhile, climatologists at GISS say their critics are making a mountain out of a molehill and that the differences in the recalculated temperatures are so insignificant that they have no impact on the overall trend towards global warming.

They attributed the error to a new data collection system established in 2000 that used different methods than the previous one.