Our Climate Numbers Are a Big Old Mess

Interesting notes from the Wall Street Journal - we can't be sure what exactly is happening to the average global temperature, in part because the variances are so small, and in part because the measurement technology keeps changing:

Our Climate Numbers Are a Big Old Mess - WSJ.com:
. . . "The earth's paltry warming trend, 0.31 degrees Fahrenheit per decade since the mid-1970s, isn't enough to scare people into poverty. And even that 0.31 degree figure is suspect.

For years, records from surface thermometers showed a global warming trend beginning in the late 1970s. But temperatures sensed by satellites and weather balloons displayed no concurrent warming.

These records have been revised a number of times, and I examined the two major revisions of these three records. They are the surface record from the United Nations' Intergovernmental Panel on Climate Change (IPCC), the satellite-sensed temperatures originally published by University of Alabama's John Christy, and the weather-balloon records originally published by James Angell of the U.S. Commerce Department.

The two revisions of the IPCC surface record each successively lowered temperatures in the 1950s and the 1960s. The result? Obviously more warming – from largely the same data.

The balloon temperatures got a similar treatment. While these originally showed no warming since the late 1970s, inclusion of all the data beginning in 1958 resulted in a slight warming trend. In 2003, some tropical balloon data, largely from poor countries, were removed because their records seemed to vary too much from year to year. This change also resulted in an increased warming trend. Another check for quality control in 2005 created further warming, doubling the initial overall rate.

Then it was discovered that our orbiting satellites have a few faults. The sensors don't last very long and are continually being supplanted by replacement orbiters. The instruments are calibrated against each other, so if one is off, so is the whole record. Frank Wentz, a consulting atmospheric scientist from California, discovered that the satellites also drift a bit in their orbits, which induces additional bias in their readings. The net result? A warming trend appears where before there was none.

There have been six major revisions in the warming figures in recent years, all in the same direction. So it's like flipping a coin six times and getting tails each time. The chance of that occurring is 0.016, or less than one in 50. That doesn't mean that these revisions are all hooey, but the probability that they would all go in one direction on the merits is pretty darned small."