I’ve now looked through his post carefully and, beneath Schmidt’s fulminations, did not find any rebuttal to any points actually made in my post, as I’ll discuss in detail below. Much of Schmidt’s post fulminates against my criticism of inadequate disclosure of adverse results. This is a large topic in itself that provides a context to the Yamal controversy, but the exposition of this context is lengthy and, in my opinion, the Yamal issues are sufficiently discrete that they can be considered on their own, as I shall do in this post.
Schmidt criticized my estimate of a Yamal-Urals regional chronology from the FOI list as merely “insta-reconstruction”, saying:
if any actual scientist had produced such a poorly explained, unvalidated, uncalibrated, reconstruction with no error bars or bootstrapping or demonstrations of common signals etc., McIntyre would have been (rightly) scornful.
While my “insta-reconstruction” was only produced for a blog post, I submit that it was better documented than many tree-ring chronologies in common use, including the three chronologies of Briffa 2000 link. that have been very popular in multiproxy reconstructions. Indeed, if “actual” scientists had provided equivalent detail for the tree-ring chronologies in common use, much of the controversy of recent years would have been avoided.
Even though this was merely a blog post, I provided all measurement data as used, together with code that, in a turnkey method, uploads the data as used and produces the chronology and graphics. The code shows the precise calculation for a an interested critic. The blog post included a graph of core counts, together with the computation. In contrast, no measurement data for the three chronologies of Briffa 2000 (Taimyr, Tornetrask, Yamal) was archived. Nor was it archived for Briffa et al 2008 until I convinced the journal to require it. New measurement data for Esper et al 2009, which addresses Siberia, is not archived. (I recently requested that it be archived, but have received no acknowledgement from either Esper or Hantemirov.)
Nor is it common practice for dendros to show “error bars, bootstrapping or demonstration of common signals” in typical publication of tree ring chronologies. I think that such practices would indeed enhance the articles, but it’s not something that I’ve taken issue with in dendro publications. (Merely getting dendro data has been hard enough.) None of Schmidt’s desiderata appear in Briffa 2000 or for other important tree ring chronologies (e.g. the Graybill stripbark chronologies that are such an important part of IPCC multiproxy studies.) Calculation of error bars for tree ring chronologies seems very much to be desired, but, in my opinion, to do so would require a major improvement in the statistical theory of chronology construction, something this is long overdue.
One Point of Agreement
In any controversy, it’s always a good idea to try to isolate any points of agreement and work out from there. There are relatively few points in Schmidt’s post that I agree with, but it is common ground between Schmidt and I that CRU’s claim that the “purpose” of Briffa et al 2008 was merely to reprocess the Hantemirov and Shiyatov dataset was a “mis-statement”.
In their June 17 submission to Muir Russell, CRU had stated:
McKitrick is implying that we considered and deliberately excluded data from our Yamal chronology. The data that he is referring to were never considered at the time because the purpose of the work reported in Briffa (2000) and Briffa et al. (2008) was to reprocess the existing dataset of Hantemirov and Shiyatov (2002).
In my post on Yamal, I had stated that the purpose of Briffa 2000 was to reprocess the Hantemirov data, but this “purpose” no longer applied in 2008.
One of the purposes of Briffa (2000) was clearly to demonstrate the effect of RCS methodology on the Hantemirov and Shiyatov 2002 dataset. I have no objection to CRU claiming this “purpose” for Briffa (2000). But, by 2008, this was no longer their “purpose”. Indeed, one doubts whether the editors of Phil Trans B would have accepted a 2008 paper with such a mundane purpose. The actual “purpose” of Briffa et al 2008 is stated quite clearly and was entirely different: it introduced and discussed “regional” chronologies.
Amidst other hyperventilating editorial comment (on which I’ll not comment right now), Schmidt conceded that CRU had committed a “mis-statement” in saying that the “purpose” of Briffa et al 2008 was to reprocess Hantemirov data – a point on which I can agree.
… using the reference to the 2008 paper is a little contradictory to the paragraphs above which were much more explicit about the background and purpose of the 2008 paper. However, to take a slight mis-statement in a single sentence, when copious other information was being provided in the same submission, and accusing people of deliberate deception is a huge overreach. Were they trying to deceive only the people who hadn’t read the previous page? It makes no sense at all. Instead, McIntyre conflates the situation at the time of the 2000 paper with the very different situation around 2008 in order to paint a imaginary picture of perfidy.
I do not agree that this was merely a “slight mis-statement in a single sentence” as the same argument had been made in their March 1 submission to Muir Russell, a submission which did not include accompanying “copious” information evidencing the error to the Muir Russell panel as a “slight mis-statement”. It also speaks poorly that CRU would make a “mis-statement” to Muir Russell on an issue so central to Climategate.
But in any event, let’s take some comfort that Schmidt and I at least agree on this point: the “purpose” of Briffa et al 2008 was not the reprocessing of the Hantemirov dataset, but the development and presentation of regional chronologies.
An Easily Resolved Point
In addition to the above point of agreement, there is one point that can be objectively resolved: did CRU calculate a Yamal-Urals regional chronology in 2006? I believe that the evidence is unequivocal that such a calculation was made; Schmidt believes otherwise, arguing that my belief rests only on a “very strained” interpretation of a Climategate email.Schmidt argued as follows:
McIntyre then quotes an email from Osborn sent in 2006 in support of his claim that the reconstructions were finished at that point, but that is again a very strained reading. Osborn only lists the areas (and grid boxes) in which regional reconstructions might be attempted since “most of the trees lie within those boxes”. It makes no statement whatsoever about the work having already been done.
Schmidt relied on this exegesis of the email to rationalize CRU’s statement to Muir Russell that they had “intended” to do a Yamal-Urals regional chronology, but decided against it because they felt that it could not be “finished in time”. Schmidt outlines his position as follows:
Of course, regional reconstructions are a definite goal of the dendro-climatology community and Briffa and colleagues have been working on these for years. Some of those results were published in Briffa et al (2008) as part of a special issue on the boreal forest and global change. Special issues come with deadlines, and as explained in a submission to the Muir Russell inquiry, a regional Yamal reconstruction putting together multiple sources of tree ring data was indeed ‘considered’ but wasn’t finished in time
Between these [two other reconstructions] we had intended to explore an integrated Polar Urals/Yamal larch series but it was felt that this work could not be completed in time and Briffa made the decision to reprocess the Yamal ring-width data to hand, using improved standardization techniques, and include this series in the submitted paper [Briffa et al., 2008].
So, Briffa et al did consider a regional reconstruction and are indeed working on it for publication, and it didn’t get into the 2008 paper due to time constraints.
I think that the last sentence would ought to read: Clear? No!
Schmidt’s argument that the 2006 calculation was never shall-we-say consummated has been picked up by a number of commenters on the RC thread, who’ve thrown down the gauntlet demanding that I prove the existence of the 2006 chronology. An early commenter on the RC thread endorsed Schmidt as follows:
To me, McI’s (as well as anyone who follows him) accusations have to show that Gavin is wrong here. Otherwise, his entire argument crumbles into nothing. … I have to ask anyone defending McIntyre to prove this. Otherwise, he should withdraw his accusations, which amount to scientific malfeasance, if true.
Subsequent RC commenters pile onto the same theme, with Schmidt complaining inline against Jeff Id that it is impossible to prove a negative:
No one can ever prove that they didn’t do a calculation, and ever more insistent demands that they must, are pointless
On the other hand, it is possible in this case to prove that CRU calculated a Yamal-Urals regional chronology in 2006 because the University of East Anglia has admitted it.
They’ve admitted it both implicitly and explicitly. Implicitly: if they didn’t hold the chronology, they would have claimed that exemption in the FOI. However, they made no such claim, instead arguing other exemptions. Explicitly: East Anglia confirmed in response to my FOI request as follows:
In investigating your appeal a further search has been undertaken and it has become clear that the first time series out of the 1001 bootstraps is actually a single nonbootstrap) chronology.
Tim Osborn of CRU showed up as a commenter on the RC thread. In my opinion, he should have set the record straight – that CRU had done the regional chronology questioned by Schmidt. Unfortunately he failed to do so.
Regardless, this is a point that Schmidt should simply concede: CRU did such a calculation in 2006. End of story.
Of course, making this admission has a knock-on effect: CRU’s claim to Muir Russell – that they had “intended” to do a Yamal-Urals regional chronology, but had been unable to finish it – cannot be accepted at face value. This ought to require Schmidt to re-think this aspect of his post, though I doubt that he will do so.
In the (unlikely?) event that Schmidt concedes that the calculation of the Yamal-Urals regional chronology was consummated in 2006, I presume that his next line of defence will be that the consummation was in-some-still-undisclosed-way unsatisfying. (I presume that these grounds will be argued in the forthcoming CRU article responding to this controversy.}
Even if the consummation of the chronology was unsatisfying, the consideration and discarding of the Yamal-Urals regional chronology was an integral part of the methodology of Briffa et al 2008 and should have been to Phil Trans B as part of the methodology.
For amusement, let’s suppose that the editor and reviewers for Phil Trans B were presented with the statement on the regional chronology that CRU had presented to Muir Russell:
Between these [two other reconstructions] we had intended to explore an integrated Polar Urals/Yamal larch series but it was felt that this work could not be completed in time and [we] made the decision to reprocess the Yamal ring-width data [from Briffa 2000], using improved standardization techniques.
If I were an editor or reviewer confronted with such language, I’d undoubtedly ask the authors to “complete” their analysis before expecting it to be accepted.
Once they considered and discarded the Yamal-Urals regional chronology, they should have disclosed the procedure, with at least the following:
Between these [two other reconstructions] we calculated an integrated Polar Urals/Yamal larch series but decided against using it for the following reasons [….whatever they were]. For reference, the Yamal-Urals regional chronology is shown in the Supplementary Information…
What’s so unreasonable about that?
Similarly, CRU’s submission to Muir Russell should have included straightforward disclosure that they had done a regional chronology calculation in 2006 and decided against using it (for whatever reasons.)
CRU’s failure to provide accurate disclosure, first to Phil Trans B and second to Muir Russell, does not mean that the entire field of climate science is pervaded by “perfidy”. But neither is acceptable, let alone praiseworthy.
The present dispute arises from East Anglia’s refusal to provide the 2006 Yamal-Urals regional chronology in response to my FOI request. East Anglia has argued that they are entitled to refuse this information under the exemptions for incompleteness and intellectual property.
Even if they are ultimately successful in their arguments (and the UEA has lost almost every decision that has been taken to the ICO or Tribunal), surely both East Anglia and the larger climate community should ponder whether this particular fight is worthwhile, particularly at this time. Even if CRU has the right to withhold the data, they are not prevented from voluntarily disclosing it.
Not every battle is worth fighting. If I were managing the file for East Anglia, I’d tell Jones and Briffa: turn the data over voluntarily. Similarly, Schmidt himself realizes that battles over withholding data play poorly outside the most partisan part of the community: why wouldn’t he encourage CRU to voluntarily disclose the data and save themselves the bad publicity.
In addition, arguing that working results from 2006, especially in a controversial area of climate science covered by IPCC, should continue to attract a form of exemption is about as bad a test case as the academic community could imagine. By litigating this case, there is a risk that the wider academic community won’t like the precedent. By disclosing the data voluntarily, no precedent is set.
In a subsequent post, I’ll try to return to other aspects of Schmidt’s rant, including his offence at criticism of failures to disclose adverse data.