Here is some more analysis based on the Esper measurement data sent to me by Science on March 16, 2005 — this time on the “Athabasca” site, which refers to a location in the Rocky Mountains, Alberta near the Athabasca Glacier. This note is rather a status report, so that I keep track of what I know right now, as I’ll have to return to the matter.
There are a few interesting comments about medieval warmth which I’ve excerpted here (and which contrast somewhat with the site chronologies.) There are also lots of dry details here, but I show one interesting graph at the end. I’ll provide some analysis with a little bit more interesting graphs in the next day or two.
The Athabasca Glacier site has been reported on in several publications, notably Luckman et al.  (L97) and Luckman and Wilson . Although this site is not used in MBH98-99, it is a mainstay of multiproxy studies under different names. As a temperature reconstruction from L97, it is used in Jones et al 1998 (Jasper), Crowley and Lowery 2000, Briffa 2000 and Jones and Mann 2004; one RCS version (Esper) is used in Esper et al 2002 (Athabasca); a second RCS chronology of Luckman and Wilson  is used in Osborn and Briffa 2006 and D’Arrigo et al 2006.
None of the studies provide accurate data citations and sorting out the provenance of the data versions used in the various studies has been non-trivial. I’ve corresponded on this matter with both Rob Wilson and Brian Luckman, who provided cheerful, but incomplete replies, and were unable to provide a detailed reconciliation at present due to other commitments, but did not exclude a more detailed reply on a further occasion.
Samples were originally taken by Schweingruber in 1984. Luckman and University of Western Ontario (UWO) associates have taken more samples up to 2000-1. Luckman sent some samples from these later programs to Schweingruber for MXD measurements (after doing RW measurements at UWO). In the course of MXD measurement, Schweingruber also did RW measurements. Luckman has not archived any measurement data. However, Schweingruber has archived results, not simply from his own sampling, but from Luckman samples sent to Switzerland. (I think that this may explain some puzzling features of the Mackenzie Mountain data set as well. I’ll bet that Szeicz did the same thing as Luckman – measure himself, then send a selection to Schweingruber for MXD measurements. I’ll bet dollars to doughnuts that some of the cores in the Szeicz archive for Campbell Dolomite are measurements of the same cores as in the Schweingruber archive for Campbell Dolomite. Of course there is no codex sorting reconciling the identification numbers.)
Luckman also did not archive chronologies for L97; however, a temperature reconstruction from 1073-1987 is available in a grey version at Briffa’s website here. This series goes 4 years later than the title of the article which refers to "1073-1983". Just recently, Luckman archived the chronologies and reconstruction for LW05 here.
Indications of MWP Warmth
Before I go on to analyze data provenance, here are a few interesting comments about medieval warmth from these two Luckman publications, which need to be kept firmly in mind when one sees the “site chronologies”. L97 stated:
Tree-line is presently advancing upslope but abundant snag material lying on the slope surface indicates that tree-line was formerly higher (figure 3). Radiocarbon dates of 980-1160 14C yr BP from these snags indicated that this phase of higher treelines predated the Little Ice Age.
Luckman and Wilson  stated:
Development of the new long composite RW chronology also allowed cross-dating of a critical Athabasca snag that was previously undated. Sample A78-S2 was identified as Larixlyalli (Y. Being, 2004 pers comm.) and lived between AD960 and 1107 (the previous chronology had only one sample between AD1073 and 1107). This is the only known sample of Larix from Jasper National Park and this is approximately 30 km north of the present range limit of the species, supporting the possibility that conditions were as warm at that time as at present.
Here is one more example of medieval treeline being higher/further north than modern treelines. Other examples discussed here include California, Siberia, Fennoscandia – so it is a widespread phenomenon. L97 also noted two samples of Pinus albicaulis (1138-1315;1179-1383) in MWP times, but did not report any in modern times. LW05 also listed the dating of subfossil stands at Peyto Glacier and Robson Glacier to medieval times – from the 9th to 14th century.
Provenance of Athabasca Data in Esper et al 
The availability of the dataset “ath.rwl” as one of 10 Esper data sets supplied by Science prompted my present inquiry. Esper et al  does not provide a data citation. It shows a minute map with a site denoted “Athabasca” located more or less in Alberta. Its only other information on provenance, in footnote 25, is an acknowledgement to “B. Luckman (Athabasca),” as being one of the persons who “sampled the tree-ring data sets in 14 different regions of the Northern Hemisphere and/or made them available to do this study”.
Esper also cited L97 as one of a number of studies relied on for this study as follows:
Several of the tree-ring collections analyzed here have been described and used previously in individual and large-scale temperature reconstructions and related studies (11, 13–19). Ring widths of trees growing in cold environments usually reflect the influence of warm-season temperatures on growth most strongly. However, in some cases, they also reflect temperatures from the cool-season months before the radial growth season as well (20). Here, we will not explicitly model the temperature signals of the individual tree-ring chronologies, because this has mostly been done already (11, 13–19).
When I searched WDCP on an earlier occasion, I was unable to identify Luckman as a contributor at WDCP. On a previous occasion, I’d stopped investigation at this point. (Luckman contributed chronologies for LW05 in Feb. 2006). However, the Esper data set ath.rwl (available only on March 16, 2006) had a number of clues that enabled what seems to be a definitive identification. The file ath.rwl consisted of 102 cores. The identification codes for the individual sites had Schweingruber-type nomenclature (sites 92 and 1059) and so I checked WDCP for possible Schweingruber series in the area (even though Luckman was acknowledged). In this case, the reconciliation between the acknowledgement to Luckman and the WDCP attribution of the series to Schweingruber seems to result from the sequence discussed above.) Through a little experimenting, I was able to identify candidate Schweingruber sites (cana170w and cana171w) for the Esper file.
Identification codes for all 73 cores at cana170w and all 29 cores at cana171w matched to identification codes of ath.rwl (and vice versa). Table 1 below shows some particulars of these two data sets, together with two other datasets archived at WDCP which are relevant to the topic.
|WDCP ID||Name||Birminsdorf ID||Start||End||Trees||Radii|
|cana170w, cana170x||Athabasca historisch||NA||1072||1991||38||73|
|cana171w, cana171x||Athabasca Glacier||athapcen||1665||1994||29||29|
|cana096, cana096x||Sunwapta Pass||sunw||1608||1983||13||26|
|cana097, cana097x||Peyto Lake||peyto||1634||1983||13||26|
Provenance of L97 Data
Although these datasets were at WDCP and are related to data used in L97, it turns out that they do not match the data sets used in L97. L97 says that chronology was constructed using a “composite of snags from the Icefield site and living spruce trees from the Sunwapta, Icefield, Ancient Forest and Athadome sites” (p. 378). Unusually for a dendro study, L97 Figure 1 provides a site map showing the general location of the Sunwapta, Icefield, Ancient Forest and Athadome sites on a readable scale. (Luckman must have done some geology.) L97 states:
The MXD chronology developed from this snag material covers the interval from 1073-1991 but is poorly replicated after ~1700. Few of the snag records extend into the 19th century and only one living tree (1778-1991) was sampled for MXD at the Icefield site. Therefore the MXD series from the snag material at the Icefield site was combined with the living spruce material from the Sunwapta site to produce the chronology used in this paper. (Figure 4). Note that , although the chronology ends in 1991, there are only two radii from a single tree that postdate 1983.
Elsewhere, L97 (p. 378) mentions that the data from Sunwapta Pass goes from 1608-1983. Although L97 do not provide any data citations, my guess is that L97 used MXD data from Sunwapta (cana096x) together with UWO version of the Sunwapta RW data (the Schweingruber version of which is cana096). The number of radii (26) in cana096x (cana096) is close to the 28 said to have been used in Table 1 of L&W 05, as shown below and the date range of cana096x (1608-1983) is close to the date range reported in LW05 Table 1 (1634-1983). Possibly a couple of cores were not sent to Switzerland for MXD measurements. Although there are a few discrepancies, the association of L97 with some version of cana096 and cana096x seems highly likely.
However, the data set cana091w used in Esper et al  seems highly unlikely to have been used in L97. It matches the information for the Athabasca Glacier data (line 3 in LW05 Table 1), said not to have been used in L97. The date range of cana171w matches (1665-1994) and the number of radii (29) is a near-match to the 28 in LW05 Table 1 line 3.
Some variation of cana170w (cana170x) appears certain to have been used in L97. Several observations in the running text of L97 link the two data sets. L97 says that the Icefield snag material consisted of 38 trees (2 Pinus, 14 Abies, 22 PCEN). The Schweingruber data set cana170w.rwl has 38 unique trees (73 cores). L97 mentions dates for the two Pinus trees (1138-1315; 1179-1383); these can be matched with trees 92110 and 92109 in cana170w, respectively. L97 mentions that only one living tree (1778-1991) was sampled for MXD at the Icefield snag site; this corresponds to tree 92140 in cana179w. So the link between this data set (also used in Esper) and L97 appears conclusive. There are some discrepancies with LW05 Table 1 — which lists 38 radii, as compared to the 73 radii in cana170w. Perhaps some of the snags listed in the 4th line should have been included here. Or perhaps the number of trees was inadvertently used instread of the number of radii. There is another inconsistency with L97 here: cana170w contains a 2nd tree (92141 — two radii) which was very-long lived and which was sampled in 1989.
I’ve plotted up this long-lived core below because it’s pretty interesting. The spike is in 1908-1909. The low values in the 19th century are at 1842-43 (which matches a time of glacier expansion); there were low values from 1680-1720; and low values from 1464-1520.
Returning to LW05 Table 1, one additional identifications appears possible: the Peyto Lake data appears to match cana097 as the date range and number of radii match exactly. However, the Peyto Glacier, Robson Glacier and Hilda data are definitely not archived (Rob Wilson). In addition, UWO ring width measurements were used in LW05 and were not archived. They may relate to Schweingruber measurements but no reconciliation codex exists. (Luckman, pers. comm.) The Peyto Lake MXD (and RW) data match cana097x (cana097), which exactly match in data range and number of radii. The Peyto Glacier MXD (and RW) data and the Robson Glacier MXD (and RW) data are all unarchived (confirmed by Rob Wilson). The Athabasca Glacier MXD data in line 3 is almost certainly cana171x, which has a date range of 1665-1994 and contains 29 radii (29 trees). Wilson said that it was possible that if a tree was weakly correlated under COFECHA, it might have been removed. There is a COFECHA report for cana171w here ftp://ftp.ncdc.noaa.gov/pub/data/paleo/treering/measurements/correlation-stats/cana171w_gap.txt . Core 059021 has a lower correlation than other cores, but it seems to have been alive in 1994 and well-dated. I don’t like this sort of exclusion of data (which wasn’t mentioned in L&W 05).
Luckman and Wilson 2005 Table 1 show on the next line a reference to more sampling of Icefield snag data (yielding 124 cores) going from 1083 to 1898. This was RW only and does not appear to have been archived at WDCP. Based on subsequent information from Luckman, it seems that the RW measurements in cana170w and cana171 were not done at UWO (although UWO has a larger collection of unarchived RW measurements), but was done at Schweingruber’s lab when MXD measurements were done.
The provenance of the Esper data is cana170w and cana171w; the provenance of the data in Luckman et al  was cana170w and cana096. Could one have guessed at the substitution? I don’t see how. Luckman et al. made a temperature reconstruction by linear regression of gridcell temperatures against 4 “predictors” — RW, lagged RW, MXD and lagged MXD. They did not archive the temperature reconstruction or the RW or the MXD chronologies at WDCP, but the temperature reconstruction is archived in a grey version at Briffa’s website covering a period from 1073 to 1987 (although the title of the article mentions the period “AD1073-1983″.
B. H. Luckman, K. R. Briffa, P. D. Jones, F. H. Schweingruber, The Holocene 7, 375 (1997)
Luckman, B.H. and R. Wilson, 2005. Clim Dyn.