In some cases, RC non-linking to climateaudit is mere pettiness, but, in this particular case, they cite information on Loehle that was initially made available at climateaudit. In this CA post, we discussed discussed the provenance of Loehle proxies and requested that Loehle provide his proxies as used – which he provided. The numbering in this article (which differs from the original article) is the numbering used in the real climate article, which they have explained to have derived from the proxy version supplied to them by Loehle, which is presumably the same version that we had requested. Loehle’s article did not include data citations for the Loehle versions. Exact data citations are provided in the CA post here ; Gavin uses exactly the same data citations and has not suggested that Loehle sent him the data citations. While the data citations could be developed independently, in this case, I doubt that Gavin can honestly say that he did not incorporate CA information on Loehle proxies.
For reference, plagiarism (Wikipedia) includes:
… incorporating material from someone else’s written or creative work, in whole or in part, into one’s own without adequate acknowledgement.
Having said that, there are useful analyses of individual proxies – a type of analysis that I, for one, welcome and believe to be relevant, including some caveats on individual proxies that have not been previously raised here. The post is a “climate audit” type of post and shows that Gavin can be a pretty good “climate auditor” when he turns his mind to it.
However, given that 9 of the 11 Moberg low-frequency proxies are used in Loehle 2007, presumably most of these criticisms were equally applicable to the prior use of these low-frequency proxies in Moberg et al 2005 or for that matter in Juckes et al 2007.
My take on Loehle 2007 has been (and I hope that this has been understood) that it is really a variation on Moberg and it’s pretty hard for me to see a rational basis on which Moberg is qualified for inclusion in spaghetti charts while Loehle isn’t. If you go through the RC critique of Loehle, my impression is that virtually every criticism can be leveled equally fairly against Moberg – raising the question as to why RC is only now raising these issues.
Some nits are pointed out in Loehle methodology. I haven’t checked the correctness of these points. And I definitely endorse the idea of realclimate (or anyone else) checking for defects in data handling and reporting. However, they would be a little more credible if they dealt with the many beams in their own eye, such as, for example, the incorrect geographic locations of the Mann et al 2007 precipitation series.
While I am mostly in agreement with their proxy comments, I am not in agreement with their views on multivariate methodology. I don’t have time to discuss this today, but Mann’s present RegEM is not an obvious panacea. It’s hard for a statistical method to be sufficiently bad as to be “wrong”, but Mann has accomplished this twice with the MBH data set: first with the MBH98 PCA-regression combination with its erroneous PCA method; more recently, with the Rutherford et al 2005 RegEM method (the code for which has now been expunged from the record). The new Mann RegEM method gets the same results as these two erroneous methods (a bristlecone-pine shaped Hockey Stick). Is the new method “right”? Readers should recognize that all that is done in these long-winded statistical efforts is choose weights for the individual proxies. The new Mann method does not report the weights assigned to bristlecones, but you can be sure that it is large.
Their comments on multivariate methodology appear weak to me, but the comments on individual proxies are well worth reading. But what’s sauce for the goose is sauce for the gander and surely apply equally or even more so to Moberg.