Are global climate change measurements accurate?

Answer 1

Annual temperatures are approximately accurate to +/- 0.05°C for the period since 1951. They were about four times as uncertain during the 1850s when instrumental measurements began, with the accuracy improving gradually between 1860 and 1950, except for temporary deteriorations during data-sparse, wartime intervals. Global averages are now given to a precision of three decimal places.

Answer 2

Over the past two decades, three independent groups have used varying statistical methods and yet arrived at nearly identical conclusion that the planet's surface, on average, has warmed about 0.75 degree since the beginning of the twentieth century. This should be conclusive evidence that their measurements are accurate, but Richard Muller, a Physics Professor and longtime critic of climate studies, believed otherwise. In particular, he believed that insufficient correction had been allowed for those measuring devices that were originally located in rural areas but were now surrounded by suburbs, since urbanisation tends to result in local temperature rises.

Professor Muller launched the Berkeley Earth Surface Temperature project at the University of California, Berkeley, to address what he called "the legitimate concerns" of sceptics who believe global warming is exaggerated. But Professor Muller unexpectedly told a congressional hearing that the work of the three principal groups that have analysed the temperature trends underlying climate science is "excellent ... We see a global warming trend that is very similar to that previously reported by the other groups." Professor Muller acknowledged that previous studies accounted for the urban effect. "Did such poor station quality exaggerate the estimates of global warming?" he asked in his written testimony. "We've studied this issue, and our preliminary answer is no."

The work of climate scientists has passed the most rigorous examination, with a finding that it is excellent work and that due allowance has been made for likely variables.