As an aside, your response to the post just above: “The issue of retention of principal components is key and McIntyre has never properly addressed it.”, is a bit off the mark. It seems pretty clear that she was asking if MBH had ever addressed their criteria for retention of PCs, not McIntyre. After all, it is their data and their methods, not McIntyre’s. It is their place to address the issue, if anyone is going to.

[DC: It is addressed in MBH98, as well as Wahl and Ammann. But, as far as I know, it's not mentioned in any of the three M&M papers, nor in Wegman et al. Thus, this is a significant error in the M&M analysis and "correction" of MBH98.]

But back to what I had started to say. I read your side-by-side comparison and had to laugh. As far as it goes, I agree with you that Wegman should have identified the source of a great deal of his information. If he did not, then shame on him. But that is the only error I see here. It certainly does not call into question the quality of his conclusions.

[DC: Some of Wegman's findings flow from mistaken assertions in this section, as I have explained previously. ]

You make a big deal out of Wegman’s adding the bits about CO2, but apparently you don’t see fit to mention that what he states about it in those additions you so deride happens to be true. So his adding those statements is not an error or distortion, it is simply a matter of including additional information in his report that is very much relevant to the issue. In at least one place Wegman clearly cites his source for those comments about CO2. If you have an issue with their inclusion, you will first have to refute the source before you have any legitimate cause to berate Wegman for including them.

And what is your problem with Wegman’s statement that there are “many confounding factors”? The fact that there are, in fact, many confounding factors is self-evident from the very material presented. I count at least 9 confounding factors in Bradley’s own paragraph. So what’s your problem? You don’t think 9 or more qualifies as “many”? That’s merely your interpretation of English usage, and does not qualify as a judgment of the science involved.

[DC: The problem is we don't know the provenance of any of these assertions. Material from Bradley has been mixed up with apparent errors and distortions of unknown provenance. You can twist and turn it as you like, but if one author wants to criticize another, it should be done explicitly and the sources and reasoning given. That's called scholarship.

No one even knows (other than whoever drafted this section) if Wegman got the Bradley material directly from Bradley, or some intermediate source. That's a huge problem. ]

In fact his emphasis on confounding factors is absolutely relevant to the point. There ARE confounding factors. Many of them. That is fact. If those confounding factors are not somehow adjusted for, then how well measurements of the tree rings correlate with actual temperature is little more than a GUESS, and likely not a good one. I strongly suspect that there is not enough existing information about those factors to allow real compensation for them; which would mean the tree proxy data amounts to just that: guesses. I could be wrong, but so far I am not aware of anybody coming forth to explain those compensations.

You make a big deal about Wegman adding the word “relatively” to one passage. Again, I had to laugh. In fact Bradley SHOULD HAVE used the word “relatively” there. The temperature component is still subject to other confounding factors; the only way that the temperature component will be “strong” instead of “relatively strong” is, again, if those confounding factors can be factored (no pun intended) out.

You make an issue of Wegman leaving out the part about regional data being used to reduce the effects of confounding factors. There was nothing improper about that. The temperature proxy data from the MBH98 and MBH99 data were drawn from relatively few samples; to the best of my knowledge there was no “regional data” available from which to make such adjustments. Therefore the whole topic is irrelevant to the issue at hand, and it was quite proper of Wegman to leave it out.

[DC: You've misunderstood the point here. Wegman's exposition glosses over the difference between an individual tree-ring proxy sample, and a series based on a collection of samples from one site or site cluster. Again, how did this happen? What source was Wegman working from for this passage?]

In these particular circumstances I question the existence of the kind of equation mentioned by Bradley to describe variability, and even if there were one, I would question its veracity. In order for an equation to function as Bradley describes, the confounding factors of the interval instrumentally measured would have to be the same as those of the period the paleo tree rings were formed.

But there is a stumbling block interfering with that idea: a central part of the MBH claim is that the conditions are NOT the same now as they were then. This leaves them in something of a quandary: simple logic dictates that if their hypothesis is correct, then they have no way of knowing whether an equation they use for calibrating the paleo data is correct.

Therefore Wegman’s discussion of confounding factors is not specious. Rather, it is an important issue that to date has not properly been addressed.

Have those confounding factors actually been addressed somewhere without my knowing about it? Possible. But if so, I would be interested in knowing where that information is so I could read about it.

[DC:A good tree-ring website is here. That's a much better starting place for knowledge, than Wegman's hopelessly muddled exposition. Good luck!]