Broken Hockey Stick
Mathematical Prestidigitation and Global Warming
The latter part of last year saw the mass media going into a frenzy over the announcement of a plot of global temperature that became known as the "Hockey Stick" from its shape, which showed steady conditions for a millennium and then a sudden rise beginning around a hundred years ago, which of course had to be caused by the rise of industry and coal-burning. As with all Global Warming scares, the predictions are products of computer programs, not anything actually being observed in the real world. And two Canadian scientists, Stephen McIntyre and Ross McKitrick, have discovered that the program used to generate this latest doomsday offering (which contradicts every other reconstruction of past climate that I've seen) is (surprise) fundamentally flawed. The media have made no great rush that I'm aware of to correct any wrong impressions they may have created (another surprise).
Ed Zahurak pointed me to a piece by Robert Muller in MIT's Technology Review showing that whereas the analysis purported to apply standard statistical techniques to find the dominant features of a set of 70 climatic records, the procedure actually employed had the effect of emphasizing any data that happened to exhibit such an upturning pattern, and to suppress any that did not. To demonstrate this inbuilt property, McIntyre and McKitrick supplied the program with randomized test data devoid of any trend. Lo and behold, out popped a hockey stick.
The two researchers considered these findings important enough to submit to Nature, but their paper was rejected. In frustration, they have posted full details, including editoria exchanges and review comments, on their own web site. No doubt they'll now be criticized for failing to have published in a peer-reviewed journal. Whether the original claim was a result of deliberate data massaging or sloppy science, and how much of the response to attempts to publicize it might have been due to pressures of political correctness is left as an exercise for the reader to decide.
While Muller does a great job in reporting this, he apparently buys into the global warming threat nevertheless, which I'm skeptical of -- to put it mildly. Nobody disputes that carbon dioxide has increased significantly over the past hundred years (from a level that was exceptionally low compared to that recorded through most of Earth's history); but practically all of that has taken place since 1950, whereas the warming that it's supposed to have caused happened before 1940, after which came 30 years of pronounced cooling to the extent that the early '70s resounded with hysteria about an imminent ice age. That warming seems to have been the tail end of three centuries of recovery from the period known as the "Little Ice Age," around the 1600s, which followed the medieval "Climatic Optimum," when Greenland was indeed green and the Vikings had settlements there. The world was even warmer around 4,000 years ago, when southern England was subtropical -- as everyone knows, part of the global deterioration brought about by bronze-age industrialization.
See also earlier postings. For a good technical summary, Global Warming: A Guide to the Science, by Sallie Baliunas, Arthur B. Robinson, and Zachary W. Robinson.