This is more than a few klicks outside my bailiwick, but maybe somebody out there can run with this. From a reader:
My PhD is in physics/nonlinear dynamics from Duke; my field first characterized and observed the “Butterfly Effect” with the nonlinear coupled differential equations in Ed Lorentz’s weather models.
Since a lot of the computer analysis code was leaked with the CRU files, I wonder if it would be possible to stuff pseudo-random data into the required data structures and reproduce AGW with their analysis routines? If an AGW signal is present in random soup or in data with a long-term artificially-enforced decline bias, then the whole AGW-thing really does collapse.
Dr. Bob [withhold, please, as I may one day wish to rejoin academia…]
Update: Rand Simberg writes:
McIntyre reportedly fed random data into Mann’s model a few years ago,
and the output was a hockey stick.
I put together a simplified pseudocode of the model.
Update II: PJM’s Charlie Martin writes:
Jonah, pass Dr Bob the following URLS
the short answer is that he’s right; there have been many
reconstructions of the Mann, Wahl and Ammann results using noise with
various spectral characteristics. (If you’re not careful I’ll explain
Update III: And another reader (affiliated with a free market think tank*):
Subject: AGW–random noise, tech stocks produce hockey stick
Regarding your post in The Corner about whether AGW may survive the crisis: I doubt it, as the temperature data themselves have been manipulated.
To see the McIntyre experiment alluded to by the reader, talk a look at this page from Climate Audit. The problems with the Mann data handling are pretty well summed up in the abstract to the M&M paper here.
For an accessible discussion of why modern temperature measurements are likely biased, see the Chiefio blog, here. Basically those who decree where temperatures should be measured have decided that high mountains are too expensive to measure and that all temperatures will be measured at the beach. In the US, and in South America. For Russia, they just copied September data 2009 data into October 2009 and called it warming.
[Name and affiliation withheld]
* For the record, I’d bet this person wouldn’t mind being identified, but since I don’t have explicit permission, I’m going with the name withheld business. That’s my policy. Unless you say you’re ok being identified, I don’t (and often, even if you say you are okay with it, I don’t).