Monday, November 24, 2008

Second Mann hockeystick produced entirely from a small, atypical and corrupt sub-group of proxies

Ever since his original "hockeystick" graph of the global temperature trend was discredited, Michael Mann has been trying to justify himself by producing a second, less challengeable hockeystick graph. His latest effort is dissected in the very thorough and detailed article briefly excerpted below. Excerpts below are the abstract plus mention of what the proxies central to Mann's new hockeystick consist of. By using data known to be corrupted, Mann has certainly done nothing to redeem his reputation.

by Willis Eschenbach
Abstract: A new method is proposed for determining if a group of datasets contain a signal in common. The method, which I call Correlation Distribution Analysis (CDA), is shown to be able to detect common signals down to a signal:noise ratio of 1:10. In addition, the method reveals how much of the common signal is contained by each proxy. I applied the method to the Mann et al. 2008 (hereinafter M2008) proxies. I analysed all (N=95) of the M008 proxies which contain data from 1001 to 1980. These contain a clear hockeystick shaped signal. CDA shows that the hockeystick shape is entirely due to Tiljander proxies plus high-altitude southwestern US "stripbark" pines (bristlecones, foxtails, etc). When these are removed, the hockeystick shape disappears entirely.

Mystery solved. The three in red at the top, and one further down, are all the Tiljander lake sediment series, which are known to be corrupted. Once we remove the four Tiljander proxies, it is obvious that the whole edifice is built on a few closely related high-elevation, moisture limited pine trees located in the southwestern US. These tree rings make up no less than 19 of the 21 remaining top proxies after Tiljander is removed. In other words, the bristlecones are back and with a vengeance.

I guess the deal is that no self-respecting paleoclimate reconstruction would be complete without the bristlecone pines (PILO), which make up no less than 12 of the remaining top 21 (after Tiljander is removed). In addition we have the bristlecone's cousins, the limber pine (PIFL) and the foxtail pine (PIBA). All of these records contain are from similar ecosystems and contain similar signals. The overwhelming majority were collected by Graybill. His work has been called into serious question by LInah Abadneh's thesis, wherein she was unable to replicate his results.

If I ran the zoo, I'd throw out all of those high altitude pine tree ring records. They are known to have problems, their use has been recommended against, and the principal investigator's work is under a cloud. I would omit them.

More here (See the original for links, graphics etc.)


Posted by John Ray. For a daily critique of Leftist activities, see DISSECTING LEFTISM. For a daily survey of Australian politics, see AUSTRALIAN POLITICS Also, don't forget your daily roundup of pro-environment but anti-Greenie news and commentary at GREENIE WATCH . Email me (John Ray) here

No comments:

Post a Comment