Science has recently weighed in with an editorial in which the editor of Science, Donald Kennedy, stated that he is “outraged” by the Barton Committee inquiring into processes for due diligence and disclosure in connection with science being applied for large-scale public policy. I thought that people might be interested in an account of my experience with Science in trying to obtain underlying data from Lonnie Thompson’s ice core studies, regularly published in Science.
Recent Science Statements
First, here are recent statements from Science in connection with the Barton Committee. The publisher of Science sent a letter to Barton about a month ago. The press release is here; the letter is here; I posted a short comment here. The publisher of Science stated:
My colleagues and I would be pleased to discuss these matters with you and your staff should you so desire.
I hope that the Barton Committee takes them up on their offer as their viewpoint on due diligence and disclosure from a journal aspect would be invaluable.
The recent Kennedy editorial is here . I will link to a better URL if I identify one. Kennedy concludes:
There are ways of avoiding both the harassment and the precedent. Chairman Boehlert could take charge of matters, because this debate belongs with the real science committee. If hearings are necessary, they can be held. If independent and objective information is needed, the Congressional Research Service could help. Better still is the time-tested way of reaching scientifically sound conclusions: scientific experiment, analysis, debate, and review. A letter* to Chairman Barton from Science’s publisher, the American Association for the Advancement of Science, points that out in prose more tactful and elegant than I can presently manage. As for me, I’m just the editor–and I’m outraged at this episode, in which science becomes politics by other means.
Since Science as a publication is committed to “scientific experiment, analysis, debate, and review”, I thought that readers might be interested in how this works itself in a particular case where they are involved as the journal of record.
Lonnie Thompson’s publication of record has generally been Science, although he has published review articles and analyses relying on the original Science publications quite widely. Major ice cores published by thompson in Science include: in South America: Quelccaya, Sajama and Huascaran; in the Himalayas: Dunde, Guliya, Dasuopu; in Africa, Kilimanjaro. Purugangri (2000) in the Himalayas is still unpublished.
The Dunde ice core was drilled in 1987. It was used in MBH98 and is widely used in multiproxy studies. In late 2003, when I was trying to verify Mann’s dataset against archived versions, I noticed that there was nothing archived on Dunde. (There is an incomplete Quelccaya archive at WDCP for the 1983 drilling, and very incomplete archives at WDCP for Sajama and Huascaran, but there was then nothing archived at WDCP on Dunde, Guliya or Dasuopu.).
In late 2003, I wrote the Thompsons, asking politely for information on Dunde, but got blown off. My side of the correspondence is here ; I can’t presently locate the answers, but you can tell that they were non-responsive by my side. In June 2004, in a letter to NSF about some non-compliant scientists, I listed Thompson as non-compliant in his archiving responsibilities. Again, I got blown off, with NSF saying that the information was either archived or already sent to me privately, both obviously false. I’ve discussed NSF before.
Now for a little detour. Thompson had published a review article in Climatic Change. I had had some contact with Climatic Change in connection with a submission by Mann et al. criticizing our 2003 article, in which I acted as a reviewer. In my capacity as a reviewer, I requested the source code and supporting data e.g. R2 statistics, a digital version of the 15th century step etc., which Mann had refused to provide. This provoked great consternation at Climatic Change, with Stephen Schneider, the editor, (who was an engaging and interesting correspondent), saying that this was the first such request that he had received in 28 years of editing the magazine – which says something about how people do “peer review”. Climatic Change. Climatic Change then convened their editorial board and decided that Mann would not be required to produce his source code, but would be asked to provide supporting calculations. Mann refused to do so. I pointed out to Schneider that the authors were in breach of the newly articulated policy. The article was never published.
However, by this time, Mann had managed to get in a dig at us in Jones and Mann , which cited the Climatic Change submission. He did not submit an erratum, when Climatic Change rejected the submission. Here’s the connection of this detour to Thompson. Shortly after the above exchange in connection with Mann, I wrote to Climatic Change, asking for Thompson’s data pertaining to his then recently pulished Climatic Change article under their new policy. Thompson complied in the narrowest and most technical way. The Figure in Climatic Change showed the decadal dO18 averages for the various drill cores. So rather than archiving all measurements (including potentially relevant things like Cl, SO4, particulates), Thompson archived only the decadal dO18 information at his website back to 1000, exactly as shown in the Climatic Change figure, and not one gram more information than that. Later in 2004, he archived the same information at WDCP – the very first archiving of Dunde, Guliya and Dasuopu information by Thompson at WDCP. So I perhaps had a very small impact on this situation.
Prior to this, there had been grey versions of the Dunde data floating around (one version in the MBH98 dataset, another version in the Yang et al dataset) and the new version was inconsistent with the two prior versions. You really need to go back to the original sample data to see how the various versions reconcile. Since this is a very important data set in the multiproxy studies, it’s actually of quite a bit of significance. I posted up a note here on the differing Dunde versions. Science Correspondence In Feb. 2005, there was a rush of publicity surrounding the recent publication of our GRL article and I was asked for an interview by Richard Kerr of Science. Following that interview, I sought to involve Science in trying to get Thompson’s data. To date, I’ve had no luck (although, uncharacteristically for my correspondence with officials, I left a ball in my court for a few months as you will see below). Here’s my exchange:
Feb. 6, 2005 McIntyre to Kerr Dear Richard, perhaps you could help me with something. Does Science have a policy requiring contributing authors to archive (or otherwise make available) data reported in Science. I have been trying for nearly 18 months to get data published in Science by Lonnie Thomspon on the Dunde, Guliya and Dasuopu ice cores, all staples of multiproxy studies. Last year, in response to my inquiries to Climatic Change, Thompson archived 10-year average values for Dunde, Guliya and Dasuopu ice cores, corresponding to Figure 5 in their Climatic Change article( see http://bprc.mps.ohio-state.edu/Icecore/Climatic-change-2003-Fig5-table.XLS). This was slightly helpful but fell far short of proper archiving. This is especially so because this data version is inconsistent with different “grey” versions of Dunde previously circulating and used in Mann et al.  and Yang et al as shown in the two figures below.
Figure 1. Dunde: Yang version versus smoothed version of Thompson  version.
The next figure compares a smoothed version of the annual Dunde data used in MBH98 (11-year smoothing) with the CC 2003 decadal version, again with obvious differences.
Figure 2. Dunde – Thompson 2003 version versus smoothed MBH98 version.
There has never been any explanation of the differences, which are visually quite material. In order to effect a reconciliation, one needs to see the original data by sample (which was taken 17 years ago), together with an explanation of the differences. Since Thompson originally published in Science, perhaps Science could take some initiative in getting him to archive the original data and methods. Secondly, Cook, Woodhouse and others published an article on American aridity, calculated from tree rings, referring to 835 sites. While there is an SI to this study, the study does not contain a listing of sites (together with ITRDB codes.) I am very anxious to compare the network in Cook et al  with the network in MBH98. I have sent a couple of emails to Ed Cook. On prior occasions, he has either failed to reply to emails or failed to provide requested data. Obviously, with my recent publicity, he is not very willing to provide this information, but it seems to me that the Science may have some obligations in this respect. If there is anything that you can do (including forwarding this to the person responsible), I would appreciate it. Thanks, Steve McIntyre
Feb. 2005 Kerr to McIntyre
Dear Steve, Interesting questions. I’m only a news writer here, so I would point you to the following two links in our Advice to Contributors pages. Requirements for archiving data seem to focus on the biological and genomic, or at least those areas with prominent public archives. Climatic and geophysical data archives exist, but Science doesn’t seem to notice. Also, even the biological requirements are fairly new. http://www.sciencemag.org/feature/contribinfo/prep/gen_info.shtml#datadep http://www.sciencemag.org/feature/contribinfo/prep/prep_online.shtml#categories You might direct any questions to our online editor, Stewart Wills: swills AT aas.org Good luck. Dick
Feb 15, 2005 McIntyre to Wills
Dear Stewart, I am writing at the suggestion below of Richard Kerr who interviewed me about 10 days ago. In the email below [see email above] , I requested data pertaining to paleoclimatic studies published in Science, where I have been unable to obtain data from the authors and requested the assistance of Science as the publishing journal. Thanks for your attention, Steve McIntyre
Feb 15, 2005 Wills to McIntyre
Subject: Re: Fw: Thompson et al (2003) Dear Dr. McIntyre: Dick Kerr is correct that our policy on deposition of data to public repositories has tended to focus on the life sciences. However, one of our published conditions of acceptance is that “any reasonable request for materials, methods, or data necessary to verify the conclusions of the experiments reported must be honored,” which would seem to cover the situation you discuss herein. If I am understanding your request > correctly, you are looking for the data underlying the three Science papers by Thompson et al. from 1997, 1998, and 2000, which undergird a later (2003) study by Thompson et al. in a different journal. I am copying this to Brooks Hanson, the deputy editor for physical sciences here at Science, who is really more qualified to respond to the issues raised here. Cordially yours, Stewart Wills Stewart Wills, Ph.D. Online Editor, Science American Association for the Advancement of Science 1200 New York Avenue, NW Washington, D.C. 20005 USA
Feb. 15, 2005 McIntyre to Wills
Dear Dr. Wills, Thanks for the prompt and courteous reply. With respect to the Thompson data, the problem is that the version of the Thompson data published in 2003 in Climatic Change on a decadal basis is inconsistent with other versions (also published in smoothed forms). Accordingly, there is little point in merely providing the plot-points for previous alternative smoothed version, as Thompson would almost certainly do. What is needed is a digital record of the sample data together with a Readme reconciling the different versions. This applies to the Dunde, Guliya and Dasuopu data, all of which was published in Science. Since these sites have been widely applied in multiproxy studies and were drilled many years ago, I think that comprehensive archiving of all drill data is long overdue. I have also requested that Climatic Change take initiative in this matter, but they have either not done so or have been unsuccessful. I also requested a listing of the 835 sites used in Cook et al , including ITRDB codes, and would appreciate consideration of this matter as well. Thanks for your consideration, Steve McIntyre
March 17, 2005 Hanson to McIntyre
Dear Dr. McIntyre: Dr. Wills passed me your inquiry, and I’m looking into it further. Can you please provide a bit more detailed request of what you would like with respect to the data for these papers and what specifically you have already requested from the authors (Thompson and Cook), then I’ll be happy to look into it further. We can look into requesting data, but would not ask the authors for or require further interpretations of the data at this point (e.g., your note on a readme file) or for data published in other journals. Sincerely, Brooks Hanson Deputy Editor, Physical Sciences.
Apr. 20, 2005 Hanson to McIntyre
Dear Dr. McIntyre: I have not heard back from you, and am thus resending this note. Please let me know if you are still considering this as an open matter. Sincerely, Brooks Hanson
Apr. 20, 2005 McIntyre to Hanson Thanks for the reminder. I had partly drafted a reply and got sidetracked on other matters. I’ll be right back at you on this. Thanks, Steve McIntyre
July 15, 2005 McIntyre to Hanson
Sorry not to respond earlier. I failed to do so at the time and just noticed my oversight. I presume that you will have taken some initiative in the mean time. In the event that you have not done so, here are some thoughts. As far as I’m concerned, Thompson et al. should archive at the World Data Center for Paleoclimatology an authoritative, organized and complete archive for each drill-hole of all their samples, including whatever measurements were taken for each sample, as well as the log for each hole. Science has been the primary outlet for the publication of their results. However, for example, Thompson et al. had never archived any results for Dunde, Dasuopu or Guliya (all published in Science) until last year when they made a limited archive of decadally averaged results after I pressed Climatic Change, where they published decadally-averaged information on dO18. However, this is far from being a complete archive. Their archiving responsibilities under the purview of Science extend to every ice core published in Science. They have archived information from Kilimanjaro, but not for samples. Since the authors have carried out age adjustments on their ice cores based on pattern-matching, it is highly pertinent to have the entire corpus of samples in order to validate their proposed sample matching. Given the authors concern that the glaciers themselves may be receding, it is particularly vital that they archive this data in a permanent archive such as WDCP. Regards, Steve McIntyre
Aug. 26, 2005
Dear Dr. Hanson, Any progress with this inquiry originally made on Feb. 6? Regards, Steve McIntyre
I realize that there was a ball in my court for a period of time. Usually, I’m pretty good about not letting this happen, but I did here. Having acknowledged that, the underlying issue is that Science does not seem to either have policies that require authors to archive data or administration practices that ensure that their policies are applied. Since NSF then relies ( a reliance which seems to me to be an abdication of their own separate responsibilities) on journals like Science, with either inadequate policy or inadequate administration, there’s a knock-on effect.