NASA Data Goof Fuels Global Warming Skepticism

The blogger, Stephen McIntyre, reported the discrepancy to NASA scientists on Aug. 4, and they had corrected the data by the following Tuesday, thanking McIntyre via email for pointing out the error. They updated the data on the NASA GISS Web site with an acknowledgement to McIntyre, who has blogged about the issue on ClimateAudit.org.

The discrepancy occurred because the source of U.S. temperature data was changed between 1999 and 2000. It was thought that the data were matched across the two sources, but there turned out to be a very subtle mismatch between the temperature records.

Mismatched record

To fix the mismatch after McIntyre brought it to light, the NASA scientists took an area where the two records overlapped and adjusted one to match the other.

The adjustments that were made to the data were fairly small—the net effect was to reduce the U.S. temperature anomalies for the period between 2000 and 2006 by 0.15 degrees Celsius, on average, noted climatologist Gavin Schmidt of NASA Goddard (who is not directly involved with the temperature analyses) in the climate science blog he co-runs, RealClimate.org.

However, some blogging skeptics now claim that these adjustments cast doubt on the conclusions of warming temperatures, because some of the anomalies were reduced such that certain years might not have been as warm as was originally thought. Numerous comments have been posted on RealClimate and global warming skeptic sites debating the import of the adjustments.

"Do you think we can now be “99% certain” that 1934 was the warmest year in the last 1000 years, or are we still 99% certain that it was 1998. How certain are we of ANYTHING that [climate scientists] say, now?", wrote one poster on RealClimate.

While the reductions changed the absolute numbers of temperature anomalies for certain years, climatologists aren't looking at the values for particular years, Schmidt explained.

"We're only interested in the trends," he said, and the amount of change in the trend of warming U.S. temperatures is larger than the 0.15 degree Celsius adjustment.

Furthermore, Schmidt notes, because the United States makes up a small portion of the Earth's total surface area, the adjustment had little effect on the global temperature data--the key measure when looking for trends in warming temperatures.

"This particular issue didn't change the global mean except to the third decimal place," Schmidt said, as compared to the 0.1 to 0.2 degrees Celsius of warming observed per decade over the globe.

'A very special year'

One result of the adjustment that was widely touted in the blogosphere as a refutation of global warming was the replacement of 1998 as the warmest year on record in the GISS data with 1934. Schmidt says this replacement is largely irrelevant to the global warming discussion because that year was something of an oddity.

"1934 was a very special year," Schmidt said, because it stood alone in the years around it as unusually warm and was only warm in the United States. In contrast, recent years have been steadily warmer all over the globe--and that trend is unlikely to reverse course.

"Unless we have a big volcano [volcanoes spurt aerosols, which have a cooling effect], we're not going to see another cold year for a long time," Schmidt told LiveScience.

Schmidt also noted that with the numerous other lines of evidence that show that global warming is taking place, including melting sea ice, changing migratory patterns among animals and rising sea levels, small adjustments like that made to the GISS U.S. temperature series are insignificant.

"There's not going to be any statistical problem that's going make [global warming] go away," he said.

• Video: Goldilocks and the Greenhouse
• Global Warming: How Do Scientists Know They're Not Wrong?
• Top 10 Surprising Results of Global Warming