It was a small change, made quietly two weeks ago on the Web site of the Goddard Institute for Space Studies. But it could have big implications.
Al Gore claimed in his 2006 crockumentary "An Inconvenient Truth" that nine of the 10 hottest years in history have been in the last decade, with 1998 the warmest year on record.
Not so, says the GISS, which is affiliated with the National Aeronautics and Space Administration and Columbia University, and is headed by Dr. James Hansen, scientific godfather of global warming alarmism. According to the GISS, the hottest years ever in the United States were, in order: 1934, 1998, 1921, 2006, 1931, 1999, 1953, 1990, 1938 and 1939.
Only one year in the last five (2006, fourth) is on this list, and only three in the last 10, compared with four in the 1930s.
The National Climatic Data Center of the National Oceanic and Atmospheric Administration also publishes annual data on U.S. temperatures. Its estimates for the last decade are higher than GISS's. Data collected from NASA weather satellites do not show the warming trend GISS and NCDC do. (Satellites indicated a warming of only a third of a degree Fahrenheit between 1979 and 1999.)
The adjustments in annual temperatures from the GISS's original data to the corrected version are quite small (temperatures from the year 2000 forward were reduced by about 0.15 degrees Celsius), but the rhetorical implications are great. It sounds so much more alarming to say: "Nine of the 10 hottest years on record were in the last decade" than to say: "One of the last five years was almost as warm as 1934."
It is interesting to note how the GISS was made aware of its error. The GISS data are based on temperature readings collected at surface stations throughout the United States. California weatherman Anthony Watts suspected (correctly, as it turned out) the readings at some of these stations were showing more warming than had actually occurred, either because the area around the station had become more urban (asphalt and concrete reflect more heat than grass and dirt do), or because there was a heat source close to the station.
A surface station in Detroit Lakes, Minn., showed a big jump in annual mean temperature in the year 2000. Mr. Watts figured this was because two air conditioning vents were installed near the station that year, blowing hot air on the sensor. He expounded upon his theory on his Web log (Watts Up With That?) and promptly got shot down. Several readers noted much of the spike occurred in winter, when the air conditioning units weren't running.
Canadian mathematician Steve McIntyre, who reads Mr. Watts' blog, had another theory. Perhaps there was a Y2K bug in the software GISS used. He reverse engineered the GISS data, and discovered an error that went far beyond one surface station in Detroit Lakes. Data collected after 1999 wasn't being adjusted to account for the time of day when readings were taken.
Without disclosing how it arrived at its conclusions, the Goddard Institute quietly acknowledged that Mr. McIntyre was right.
The United States is only 2 percent of the world's land mass. It's possible the rest of the world's been getting hotter in the last few years, even if the United States hasn't. But as Lorne Gunter of Canada's National Post noted, we only have surface temperature readings for half the world today. Prior to World War II, we had readings for less than a quarter of it.
Many of those readings are suspect. Earlier this month, British mathematician Douglas Keenan accused Dr. Wei Chyung Wang, upon whom the United Nations' Intergovernmental Panel on Climate Change relied for data on China, of research fraud.
Mr. Keenan was able to cry foul on the dubious data because the lead author of the paper to which Dr. Wang contributed his figures was a British scientist, Dr. Phil Jones of East Anglia University. Under British law, those who conduct publicly funded research must disclose the data and methodology on which they based their conclusions.
No such law exists in the United States. Though publicly funded, neither the scientists at the Goddard Institute nor the National Climatic Data Center disclose how they arrive at conclusions so they can be independently verified. They ought to.
As the GISS was quietly acknowledging its error, Newsweek magazine, with exquisitely bad timing, declared in an Aug. 13 cover story that the debate on global warming was over.
"The story was a wonderful read, marred only by its being fundamentally misleading," wrote Newsweek contributing editor Robert Samuelson in the following issue.