It is a well-known fact that gangster Al Capone was eventually jailed not for murder or extortion but for tax evasion. In other words, it was not detectives that finally nailed him, but accountants. I am beginning to think that in an odd parallel, it will not be physicists that expose the flawed foundations upon which global warming alarmism is built, but statisticians.
Steve McIntyre, whose fantastic blog Climate Audit is a must-read for anyone on either side of the debate, has long taken this approach. By digging through the detail of Michael Mann’s famous Hockey Stick he revealed, among other abuses of scientific norms, that Mann had both misused statistical techniques and ignored inconvenient statistical results.
Then of course, there was the time Dr Phil Jones admitted he was unable to plot a simple graph with a trend line. Let’s not forget that Dr Jones is Director of Climate Research at the University of East Anglia, having authored a number of studies claiming to show how uniquely dangerous the warming between the 1970s and 1998 was. Really, it’s the sort of thing people get taught in a two-day introductory course in Excel.
Recently there has been a flurry of statistical activity. Ross McKitrick and Tim Vogelsang have published a paper (with explanatory discussion here) that demonstrates a fundamental flaw in the much-vaunted computer models that have provided so many alarming forecasts of climate doom. Basically, the models can generate a linear trend in the climate, but not the step changes which are often observed in temperature records. So not only do they over-estimate future warming, but the way they do so is different to the way nature actually works. It’s another very poor showing from the models but we should be used to that by now.
It is worth noting that in writing the paper, Vogelsang created some novel statistical techniques which he then proceeded to test thoroughly before applying them to the models. This sort of rigour is sadly lacking among ‘mainstream’ climate scientists.
Another paper hot off the press is by Nic Lewis (again, with discussion here), examining some of the statistics used in the IPCC 4th Assessment Report in 2007. His paper demonstrates that the techniques used to assess the probable range of climate sensitivity (in effect, how much warming would occur for a given increase in atmospheric CO2) were flawed, and that some of the higher estimates for this sensitivity are not valid. Once again an IPCC publication, supposedly the distillation of state-of-the-art climatology, is found to contain fundamental errors.
It’s sad to see any branch of science undermined by so much sloppy work, but the climatologists are reaping the whirlwind they themselves have sown. They refuse to bother to learn the correct application of statistics in the same way Capone thought filing tax returns was beneath him, and I’m increasingly of the opinion that it will lead to their ultimate undoing.
PS: For particularly excitable readers, please note that my use of the Al Capone analogy is for illustrative purposes only, and I am not suggesting that anyone who supports the IPCC ‘consensus’ is involved in criminal activity. Except of course Peter Gleick.