Fred Singer was here on Monday and gave a rather uninspired talk criticizing global warming, etc. He did show an example of the mis-use of statistics in the greenhouse attribution debate and it was one of Wigley's papers. It was the one in which Tom tried to show that the autocorrelation function of instrumental temperatures was far greater than the acf of temperatures from unforced OAGCM models, therefore "proving" that greenhouse gases were forcing instrumental temperatures. It was a pathetically poor paper that had Mark Cane, Yochanan Kushnir, Upmanu Lall, Balaji Rajagoplan (all good maths/stats people), and me just shaking our collective heads wondering what the f*ck Wigley was trying to do. Needless to say, Singer quite easily showed how hopelessly flawed and ridiculous the analysis was, and everyone agreed with him for once.
The North Atlantic Seesaw
9 hours ago