Hughes’ letter to Barton says that NSF issued him an opinion that he was in compliance with all NSF and US government obligations regarding access to data. Why would NSF go out of its way to issue such an opinion letter? I wonder what due diligence that NSF did before issuing the opinion letter. Here are some thoughts.
Hughes’ letter to Barton states:
The U.S. National Science Foundation has stated that I have “complied with the policy guidelines set out by the US government, and the NSF in particular, regarding access to data from publicly funded research” [Email from Program Manager David Verardo 10 August 2004]
In another recent internet posting in response to the Barton inquiry, the President of the University of Arizona and other top university officials posted a memo stating the following:
In December 2003, the NSF declared in writing that the three scientists had fully complied with NSF policies regarding public access to data generated in federally-funded projects.
It’s hard to tell from here whether there were two different letters or whether one of the dates is incorrect. Both dates correspond to dates of my inquiries to NSF. In December 2003, I asked NSF to intervene in my search for MBH source code. They refused. I also copied David Verardo of NSF in a request to Mann for residuals (a very different request not involving source code). To my astonishment, Verardo intervened and said that Mann did not have to – even prior to an actual refusal by Mann.
I can think of no valid reason for a bureaucrat to preempt Mann’s decision – he could have had a change of heart and for some reason be willing to disclose residuals. At the time, I had not raised the issue of unarchived data by Hughes. Verardo’s statement was limited to the position that the source code was personal property — a highly questionable view — see my recent post on this. Perhaps Verardo issued a letter to Hughes around this time, which we do not otherwise know about, or perhaps the President of the University of Arizona has mixed up the dates.
On June 26, 2004, I sent the director of NSF a letter complaining about unarchived data sets from Hughes (and others), referring to unarchived Yakutia data in the case of Hughes. This letter caused a little bit of scurrying by some of the scientists involved to archive data or otherwise respond. On June 29, 2004, Crowley answered my email for the first time and promised to archive his data at WDCP by the end of August (which he didn’t do — I’ll discuss Crowley elsewhere). In July, Mann archived the Corrigendum SI at WDCP; Jacoby archived an updated version of the Sol Dav, Mongolia series and Hughes sent most of his Yakutia data to WDCP (according to information from WDCP), which were posted up in October 2004 – see here.
On July 28, 2004, NSF sent me an email, which was completely inaccurate, saying that:
I understand that their data has been sent directly to you, in some instances, via electronic means. In other cases the data is archived in the publicly accessible World Data Center (WDC) that is managed by the National Oceanic and Atmospheric administration (NOAA) and from which you are free to download information of interest. This facility houses a wealth of paleoclimatic data that is organized by scientific subspecialty involving biological, sedimentological and chemical proxies from terrestrial and marine environments.
Most of the data referred to in my letter was then unarchived (and remains unarchived.) Given that (for example) Hughes’ Yakutia data had been archived between the time of my letter and the time of the response of the administrator, the position of the administrator seems a little disingenuous. If I’d been writing the NSF letter, I’d have been inclined to at least acknowledge that Hughes’ Yakutia data had been archived in the meantime and perhaps even credit my inquiry with prompting some action. Shortly thereafter, on August 10, 2004, (according to the Hughes letter to Barton), Verardo of NSF – the same official who intervened with the Mann residual series – sent him an email saying that he was in full compliance with all NSF and U.S. government archiving requirements.
Examining Hughes’ publication record and grant record in even a cursory manner suggests otherwise. For example, Hughes’ NSF grants are accessible at NSF . These show nearly $700,000 in awards in 1998 and 2002 for Western U.S. tree ring studies: ATM98-09431, Temperature Variability Since AD 1000 in the Western U.S. from Tree Rings; and, ATM02-13962 Natural Spatiotemporal Variability of Climate over the Western United States in the Last Holocene. WDCP has a search function that locates all contributions by contributor and country. Hughes is named as a co-contributor in only 6 western U.S. site chronologies, of which only 2 end in 1998 or 1999. Connie Woodhouse and Peter Brown are co-contributors in all 6 sites. These scientists have received separate funding in which Woodhouse is PI (see ATM97-02951 Expanded Dendroclimatic Reconstruction of Great Plains and Central Rocky Mountains Drought) .
It appears that all 6 western U.S. sites in which Hughes appears as a co-archiver were funded under the grants in which Woodhouse and Brown were co-PIs — and are part of a larger set of about 39 sites archived by Woodhouse and Brown. Some details are at Brown’s web page here. Accordingly, Hughes does not appear to have so far archived any western U.S. sites from the funding of almost $700,000. (His archiving record for North American sites is about on a par with Jacoby – see posts on this). Hughes has also collected measurements from a number of sequoia sites, all sampled prior to 1992, [Hughes and Brown 1992], but has not archived any measurements from sequoia sites at WDCP. It looks to me that some federal funding for the sequoia sites came from the U.S. Forest Service.
I don’t know how the NSF was in a position to warrant to Hughes that he was in compliance with potential federal obligations with respect to his obligations with respect to this data or what due diligence they carried out with respect to the sequoia sites. U.S. government archiving policies propose a very limited grace period for exclusive use, which I discussed here. For example, the guidelines to NSF are as follows:
For those programs in which selected principal investigators have initial periods of exclusive data use, data should be made openly available as soon as they become widely useful. In each case the funding agency should explicitly define the duration of any exclusive use period.
While some period of exclusive use is contemplated, I think that NSF periods of exclusive use are either inappropriate or unenforced. In this case, Hughes has some unarchived measurements from pretty interesting sites. Salzer and Hughes  mentions new ring width measurements taken at bristlecone pine sites at Sheep Mountain, Pearl Peak and San Francisco Peaks AZ in 2002. None of these updates have been archived. These are important sites — Sheep Mountain is the most heavily weighted series in the MBH98 reconstruction. I presume that other high-altitude sites have been measured. Wouldn’t it be interesting to see what’s happened to bristlecone pine growth since 1987? After all, bristlecone growth is supposed to be a uniquely essential proxy for world climate history.
Mann has told us that we have to rely on old data because new data is from “inaccessible” and “expensive” sites. Well, here’s some data that Hughes is simply sitting on. Elsewhere, Hughes has been funded to do studies pertaining to the Sahel, but has not archived any data from this area. Hughes has archived results from 1 site from Jordan, 9 sites in Turkey, 7 sites in India, 1 site from China and 1 site from Great Britain. I have no idea whether this archiving is complete or whether there are unarchived sites. Here is a summary of Hughes’ NSF funding.
In summary, I can see no valid reason why Verardo should have given Hughes a letter stating that he was in compliance with all NSF and U.S. government archiving guidelines (even if he was, which doesn’t appear to be the case). For Hughes and the University of Arizona, he (they) should keep in mind that any unfulfilled obligation does not disappear merely because of an assurance from Verardo. For NSF, it seems impossible that Verardo carried out adequate due diligence to ensure that Hughes had actually complied with all applicable obligations prior to sending out his letter. Time will tell.