March 22, 2005
Peer Review & The Scientific Method
One of the cornerstones of environmental man-made global warming theories is the “MBH98 Northern Hemisphere Climate index” study. The basic findings of the Mann, Bradley & Hughes (MBH) study conclude the the 20th century is (by far) the hottest century on record; using known global temperature indicators such as tree rings and ice core samples. These number have been the basis for global temperature models and public environmental policy.
The primary rebuttal to this study is the ” M&M Critique of the MBH98 Northern Hemisphere Climate index.” McKitrick & McIntyre’s (M&M) basic findings are that the algorithms used by MBH98 produce a “hockey stick” shaped pattern (a pattern showing the 1900′s as the hottest century ever) in 99% of cases… even when the feed data was random noise! This would point to a statistical failure (or outright bias) on the part of the MBH98 algorithms. Using the same base data (i.e. the tree rings, ice core samples, etc..) against their own statistical analysis package, the M&M team gets a very different result. Their results conclude that the hottest century on record is the 15th century, not the 20th.
The debate has gone on for a number of years with strong opinions and heavy politics on both sides. Well things have taken a fairly significant turn recently. The M&M team has decided to open source their analysis and algorithm data. This provides for the much higher level of peer review. In addition, they use the well know (and well respected) R Statistics Package as the basis for their calculations suite. MBH has described their process but refused to open the entire algorithm suite (some select sources have been made available) to full peer review. While a number of third party studies have duplicated the MBH results, in each case the MBH algorithms (in closed source form) where used to produce the results. This leads to some serious questioning of the MBH methodology and therefore the validity of their results.