Climate Audit

Two of Boulton’s close associates at the University of Edinburgh (see Bishop Hill here – scientists hired as professors in 2007 while Boulton was a very senior professor in the same department – are Climategate correspondents.

In the searchable anelegantchaos version, Hegerl gets 41 hits; Gabi 29 hits and Crowley (her husband, also an Edinburgh professor) gets 125 hits. Both are coauthors of the hockey stick used in the Royal Society of Edinburgh Policy Paper of Dec 2009 (the “Boulton Hockey Stick”).

In light of Boulton’s apparent resolve to flout Russell’s desire to avoid even an appearance of prejudice, the roles of Hegerl and Crowley in Climategate correspondence assume additional interest. Today I’ll report on one such incident – one that directly concerns M&M participation at the NAS panel in 2006, one that involves the IPCC and one that illustrates Team attitudes to data archiving.

The data issues are sharpened by the fact that until a few days ago Hegerl hadn’t archived the proxies used in the Hegerl reconstruction – the reconstruction used in the Royal Society of Edinburgh Dec 2009 briefing paper on Copenhagen.

This incident starts in fall 2005, beginning with my peer review of the AR4 First Order Draft. Briffa, as IPCC Lead Author, included a discussion of the Hegerl reconstruction, citing “Hegerl et al, submitted”. This was an early version of what later became their Journal of Climate paper (now shown as submitted in January 2006, rather than fall 2005 as stated in the IPCC draft.) The First Draft stated:

The reconstruction is based on 12 proxy records, some of which are area averages of multiple records (Figure 1, data sources see Supplementary Information).

This being the Team, needless to say, no Supplementary Information was available at the IPCC website where the unpublished article was online. In my capacity of IPCC peer reviewer, I asked IPCC to provide supporting data for the Hegerl reconstruction. Subsequent events are a long story, which I’ve told elsewhere. To shorten a long story, the IPCC refused to obtain the supporting data and referred me to Hegerl. I asked Hegerl; she refused. This attempt to obtain data, together with a related inquiry to Rosanne D’Arrigo (see prior accounts for details) resulted in complaints from the authors to IPCC and Susan Solomon, WG1 Chair, warning me that I would be expelled as an IPCC peer reviewer if I asked any other authors of unpublished papers cited by IPCC for data in still unpublished studies. Strange but true.

Because information on the proxies (even what proxies had been used), in my First Order Draft review (6-1357), I stated:

Hegerl et al, submitted do not provide any information on proxies and should not be used [Stephen McIntyre]

IPCC Lead Authors (presumably Briffa here) are required to reply and their replies are supposedly considered by Review Editors – in this case, John Mitchell, Boulton’s co-presenter on Oct 29, 2009. Briffa replied with an untrue answer that presumably convinced Mitchell:

Rejected – this paper does provide information on proxies

Author Responses were not available until after publication of AR4 and I accordingly had no contemporary knowledge of Briffa’s untrue answer.

Six months later, Ross and I met Hegerl (and others) at the NAS panel presentation day (March 2, 2006). There was a pleasant reception afterwards and I did my best to be cheerful.

Climategate Letters show another side to the event.

That evening, Hegerl telephoned Francis Zweiers in Canada with an urgent message for him to contact IPCC Chapter Author Jonathan Overpeck. Hegerl was extremely worried that we would insist on having “access to supporting data” [imagine the gall!] , said that this was “making her nervous” and that she wanted to be sure that Overpeck was comfortable using her reconstruction knowing that we wanted to see her supporting data (1141393414.txt). Zweiers message to Overpeck:

McIntyre and McKittrick were there, and seem to have left Gabi with the strong impression that they will be insisting on having access to supporting data, etc., used to build reconstructions. Gabi says that this is making her nervous, wants to make sure that you are aware of the status of her reconstruction, and wants to be sure that you are comfortable with continuing to use it in Ch 6. She says that if you feel it necessaryto exclude her reconstruction from your SOD of Ch 6, you should do so.

In her message, Hegerl explained the problem arising from check-kiting (also a problem for Wahl and Ammann.) There were two companion papers: the reconstruction was described in a still pending submission to J Climate that didn’t seem likely to be accepted in time; the reconstruction had already been used in a companion detection article that had already been accepted in Nature (“rigorous” Nature reviewers apparently unconcerned about the check-kiting use of a reconstruction still under review). Zweiers:

The reconstruction is used in her Nature paper on sensitivity, which has been accepted, but the Nature paper does not describe the reconstruction or the supporting data in any detail. There is a paper under review at J. Climate that does do that (which is cited in the Nature paper), but unfortunately, an editorial decision is still pending.

Zweiers suggested that Overpeck call Hegerl that night. Overpeck doesn’t seem to have called Hegerl, but, in the wee hours of March 3, Overpeck forwarded Hegerl’s worry about looming demands for supporting data to Briffa, Eystein Jansen and (apparently) WG1 Chair Solomon. Overpeck expressed his view (1141393414.txt) that he didn’t think that “IPCC has to provide anything beyond the report”, mentioning his recollection that WG1 Chair Susan Solomon had made the same point already. After reassuring the others that Hegerl would not have to provide supporting data, Overpeck sarcastically said “M&M can get Congress to ask the FBI to secret Gabi away forever for doing her science the accepted way” – the “accepted way” here apparently being without having to provide supporting data. Overpeck:

Hi guys – great timing here for this message from Francis, and I don’t think we can (or should) do anything. It seems Gabi’s recon is in press, and that’s the way it is. I suspect Gabi’s J Clim paper will come out before the TOD too, but since it’s in press in Nature, it’s published.

I don’t think the IPCC has to provide anything beyond the report – in fact, I’m almost sure Susan made this point to me/a bigger group already. I’ll cc this to her, just so she know’s what might be coming, but I think we’re fine. M&M can get Congress to ask the FBI to secret Gabi away forever for doing her science the accepted way. Seriously, it’s up to her to make things available as appropriate.

Of course, I could be too sleep-deprived too. Am I correct in my assessment? I don’t feel like calling Gabi at 2am (her time) to discuss making changes (e.g., to text, let along figs) that it’s too late to make anyhow. I’ll respond to Francis after I hear from you.

Anyhow, I’m just about to send the full SOD text back to Norway for final minor editing. It looks good.

Best, peck

A few hours later, Briffa replied to Overpeck, Jansen and Solomon (1141393414.txt) that they should keep the Hegerl reconstruction even if the article that actually described the proxies and reconstruction remained unpublished, since it had been used in the Nature paper:

Let us stay with Gabi as it is in Figure etc. and as you say in the Nature paper anyway.

The AR4 Second Order Draft came out a month later and stated:

Hegerl et al., (in press), used a mixture of 14 regional series, of which only three were not made up from tree-ring data (a Greenland ice oxygen isotope record and two composite series, from China and Europe, including a mixture of instrumental, documentary and other data). Many of these are common to the earlier reconstructions. However, these series were combined and scaled using a regression approach (total least squares) intended to prevent the loss of low-frequency variance inherent in other regression approaches. The reconstruction produced lies close to the centre of the range defined by the other reconstructions.

In keeping with the sleight of hand discussed by Briffa, Overpeck, Jansen and Solomon (and presumably Hegerl), they switched the citation from the still unaccepted Journal of Climate submission – the one that (however inadequately) described the proxies in the reconstruction, substituting a citation to the Nature paper on detection, that merely applied the check-kited reconstruction and which did not contain the information on the actual proxies recapped in the Second Order Draft text quoted above.

Needless to say, I noticed the switch that they had plotted in the Climategate exchange. As a reviewer, I stated:

The version of Hegerl et al “accepted” has been switched and the proxy reconstruction presented in chapter 6 relies on their submission to J Climate, which had not been accepted as of April 2006, rather than their Nature article. The articles were switched at the WG1 website between drafts. The Nature article does not provide details mentioned in the Second Order Draft. Non-compliance with WG1 publication deadlines, especially in favor of publications by IPCC lead authors and their associates, is unfair to other authors who might also have sought waivers from published guidelines. [Stephen McIntyre (Reviewer’s comment ID #: 309-120)]

Briffa replied unresponsively as follows:

Papers cited are within the guidlines for in press papers

The paper that was cited was “within the guidelines” – but it didn’t describe the proxies or support the statement in the section.

Later on, I raised a similar point at a different line in the report:

The Nature article of Hegerl et al does not describe the network – either where the proxies are or the total least squares method. The article describing these things is their J Climate article which missed the deadlines. It shouldn’t be used. [Stephen McIntyre (Reviewer’s comment ID #: 309-81)]

This elicited the following curious response – one that looks even more curious in light of the Climategate letters:

Rejected – while the reviewer is entirely correct about the description of the work, the reconstruction itself is in the ‘published’ literature and is considered salient to the discussion of ‘true’ amplitude of past changes – after long consideration it was decided to keep it in the chapter, though it is recognised that the work will be subject to later close scrutiny.

And at line 30:53, I tried once again:

There has been a bait-and-switch in the Hegerl et al submission. The article in which Hegerl et al describe their network – and the one at the WG1 website for the First Order Draft – is their submission to Journal of Climate, which was described in their Nature article as merely being “submitted”. The article at the WG1 website for the Second Order Draft is their Nature article which has been submitted, but which does not provide the information described in this section – which derives from the Journal of Climate submission. Since the Journal of Climate has not met IPCC deadlines, TSU should have removed all references to it in February. In any event, all references to it should be deleted now. [Stephen McIntyre (Reviewer’s comment ID #: 309-94)]

This time, Briffa replied incorrectly that the “new” paper “provides the necessary information” (it didn’t):

Rejected – the new Hegerl et al (2006) paper was accepted in time for new inclusion deadline and provides the necessary information.

I observed that (despite the continued refusal to provide a list of proxies), it was my surmise that Hegerl used Mann’s PC1, Briffa 2000 sites and the Yang composites and would be prey to the same vulnerabilities as the other composites using the same series:

From the map in the article at the First Order Draft, I presume that Hegerl et al used all the Briffa 2000 sites ; Mann’s PC1 and the Yang composite (with tropical ice cores). Conclusions from it as to relative medieval-modern levels will be vulnerable to the same factors as the other studies. [Stephen McIntyre (Reviewer’s comment ID #: 309-95)]

IPCC didn’t care:

Noted – no revision to text necessary.

Hegerl’s Journal of Climate article was re-submitted in June 2006, this time providing some of the information on what proxies were used – the information that I had been denied as a reviewer. (This version was unavailable to IPCC Second Draft reviewers.)

Meanwhile, the NAS panel presentations mentioned above had taken place. The problems with Mann’s PC1 and stripbark bristlecones had been vigorously raised by us at the presentations.

Instead of reporting that she used the controversial Mann PC1 (which we now know for sure that she used in its Mann and Jones 2003 manifestation), Hegerl said that she used an “RCS processed treering composite”:

western U.S.: this time series uses an RCS processed treering composite used in Mann et al. (1999), and kindly provided by Malcolm Hughes, and two sites generated by Lloyd and Graumlich (1997), analyzed by Esper et al. (Boreal and Upper Wright), and provided by E. Cook. The Esper analyses were first averaged. Although there are a number of broad similarities between the Esper and Hughes reconstructions, the correlation is only 0.66. The two composites were averaged.

The problems aired on NAS presentation were reported on in June 2006: the NAS Panel sharply criticized Mann’s PC1 and even recommended that strip bark bristlecones be avoided. Wegman spoke out a month later in even sharper terms.

These results were known to Hegerl and her coauthors. But they had a conundrum: they had check-kited the still unaccepted Hegerl reconstruction in their accepted Nature article. If they adopted the NAS panel recommendations and amended their reconstruction to exclude
Mann’s PC1 and bristlecones, this would cause a problem for their Nature article.

So they did nothing and waited. Needless to say, the supposedly “rigorous” peer review carried out at Andrew Weaver’s Journal of Climate was completely unequal to the challenge of observing that Hegerl’s “RCS processed composite” was the Mann PC1 criticized by the NAS panel and the revised article – Mann PC1 and all – was published in early 2007.

And now the Hegerl reconstruction turns up once again as the Boulton reconstruction – said by Boulton to be “independent” of the other hockey sticks.

There are thousands of people in the world who are not closely to Hockey Team members Crowley and Hegerl, who were not employed for 18 years at the University of East Anglia, who did not have prior views on climate change. Boulton isn’t one of them.