Sunday, September 7, 2008

A new hockeystick?

This has been lifted with permission from Jon Ray's Greenie Watch. RK

After years of headscratching, Michael Mann thinks he has finally devised a new set of statistical procedures that will revive his discredited "hockeystick" graph. An email about it from John A [] below:

The BBC in their traditional position of custodian of climate science orthodoxy, have announced that Michael Mann has produced yet another Hockey Stick:
A new study by climate scientists behind the controversial 1998 "hockey stick" graph suggests their earlier analysis was broadly correct. Michael Mann's team analysed data for the last 2,000 years, and concluded that Northern Hemisphere temperatures now are "anomalously warm".
Yep, all the omens are good.
In their latest study, Dr Mann's group collated more than 1,200 proxy records - the majority from the Northern Hemisphere - and used different statistical methods to analyse their cumulative message. We used two different methods that are quite complementary in the assumptions they make about data, so that provides a test of the sensitivity of data to the methods used," he told BBC News. "We also made use of a far wider network of proxy data than previously available. "Ten years ago, the availability of data became quite sparse by the time you got back to 1,000 AD, and what we had then was weighted towards tree-ring data; but now you can go back 1,300 years without using tree-ring data at all and still get a verifiable conclusion."
Of course Dr Mann made this claim of robustness to the removal of dendroclimatic records last time, which turned out to be a flat out lie: The Hockey Stick shape disappeared when the bristlecone pines of Western Colorado were removed as Mann himself knows because he tested for their removal and then failed to report that salient and inconvenient truth (ie the CENSORED directories).
The same basic pattern emerged when tree-ring data - whose reliability has been questioned - was excluded from the analysis. "I think that having this extra data and using more methods to analyse it makes the conclusions more robust," commented Gabi Hegerl from the University of Edinburgh, UK, who was not involved in the research.
Yep. Of course Gabi Hegerl was involved in making her own proxy reconstruction of past climate using Michael Mann's same flawed method and incorporating Mann's PC1 as a proxy within her own limited set of proxies (where of course, it dominated the result). See here for this. Not exactly an unbiased observer, is she?

As you follow these links, follow the comments...they are as enlightening as the articles. Two of the things I have learned about modeling are the terms "smoothing" and "proxy data" and that is where the snakes "lie" in the grass. If decisions are to be made as a result of "modeling" we are in real trouble. Science's holy grail is grant money...not truth. RK

No comments: