The UnAustralian

Monday, November 24, 2003
Reconstructing Mann

The accusations of McIntyre and McKitrick appear to have taken over this blog (more due to my lack of spare time than anything else), so I may as well carry on my updating of the story.

Following Josh Halpern's quick investigation into Mann's archive, he has confirmed that Mann's statement is nature "Certain densely sampled regional dendroclimatic data sets have been represented in the network by a smaller number of leading principal components (typically 3-11 depending on the spatial extent of the data set). This form of representation ensures a reasonably homogeneous spatial sampling in the multiproxy network (112 indicators back to 1820)" is correct.

Now, David Ball, is using a small subset of Mann's data to create a reconstruction, which can then be compared against Mann's reconstruction. So far, David states "The result was virtually the same as MBH, with some fractional differences." It's worth while to carry on reading this thread as it details (at level far beyond my understanding) some of the issues involved in a multicomponent analysis.

So far, I suspect, that M1 & M2 have no dirt on Mann, except that he didn't explain his methodology well enough for a non-expert to understand, and that he has a messy archive. These criticisms could be applied to approx. half the scientists who I have worked with.
| 6:32 PM