The rise and fall of a hockey stick

by Lars Kamél, March, 2005

In my article about pseudo science in climate research, I mention the study by Mann, Bradley and Hughes from 1999 (MBH99). Their reconstruction of the last thousand years of Northern Hemisphere climate looked like a hockey stick. It is no wonder that their result, so widely used by IPCC, was nick named the Hockey Stick. Actually, this study was the second part of a study, whose first part was published in 1998 (MBH98) and dealt with the period 1400-2000. It came as no surprise to me when several new investigations showed that the Hockey Stick analyses was full of faults and errors. In 2003, I considered the immediate acceptance of the Hockey Stick by IPCC as the scientific truth to be pseudo scientific. Soon thereafter, it turned out that the Hockey Stick itself may well be an example of pseudo science.

The two heroes in this plot are the two Canadians Steve McIntyre and Ross McKitrick. Neither of them is a climate scientist for profession. In the infected science of climate research, this may be an advantage rather than a disadvantage. In any case, I could give you many examples where "amateurs" have made scientific and/or technical breakthroughs. These "amateurs" include William Herschel, who was neither a telescope maker nor an astronomer for profession, but still built the best telescopes of his time and was the first since Ancient times to discover a new planet (Uranus). Another "amateur" was John Harrison, a carpenter who made improvements to clocks to such a degree that he solved the longitude problem (the problem of calculating the longitude when aboard a ship far out to sea). Anyone claiming discoveries are worthless because they are made by "amateurs" is an ignorant person who knows nothing about the history of science and technology.

In three papers (MM03, MM05a,MM05b), McIntyre and McKitrick have showed that MBH98 contained "collation errors, unjustifiable truncation or extrapolation of source data, obsolete data, geographical location errors, incorrect calculation of principal components and other quality control defects", to cite the abstract of the first of these paper. The method used in MBH98 actually "mines" for hockey sticks. If any of the data series in the analyses has this kind of shape, it will dominate the final result. This is shown in MM05a and MM05b, where it is also shown that the hockey stick shape is almost entirely due to an inclusion of tree rings from Bristlecone pine. This is an extremely long-lived species of pine, growing in Western USA. Its growth shows a drastic increase in the 20^th century, giving the tree ring data from these trees a hockey stick shape. This increase in growth is, however, not a result of global, or even local, warming. Most likely, it is a result of a fertilization effect because of the increasing amount of CO₂ in the atmosphere.

Another nail in the coffin to bury the Hockey Stick in, was delivered by von Storch et al. (2004), in a paper which shows that the method used in MBH98 and MBH99 will take away much of the effects of natural climate variability in the data. Left is a false impression of an almost stable climate for most of the period, up to the time when the industrializing of the world became significant.

The question about the reliability of the Hockey Stick has almost turned into a war. McIntyre and McKitrick have put up web sites to give additional support to their claims. Mann, together with a few of his colleagues, has put up a web site to defend his faulty research. The latter site invites comments and questions from the general public. I know several cases where critical comments and "difficult" questions have been censored by the web masters of that site.

McIntyre and McKitrick have not only found problems with the Hockey Stick. They have also found problems with climate reconstructions in general, and especially how data and programs are dealt with by many researchers. Data and programs are often kept hidden from public view by those who made the investigations. At least the data should be made available, for different reasons. One reason is that most of the research has been paid for by public funds, and thus the public has the right to investigate the data. Another reason is replicability. Experiments, observations, and calculations should always be possible to replicate. If they cannot be replicated, it is not science, but pseudo science. That is a simple fact and one of the differences between real and pseudo science. But if data are hidden and analyses algorithms not thoroughly explained, it is not even possible to try to replicate the results. If data and algorithms are not disclosed, the investigation should not be taken seriously, because then it is not possible to check it.

I think there could be no doubt that the Hockey Stick is broken. It was, at best, the result of bad science and a programming error. But you don't need to take my word onfor it. Make your own opinion. I am biased. I never liked the Hockey Stick, I never believed in it, and I always thought something must be wrong with the research behind it.

References