date: Thu Sep 25 15:38:04 2003 from: Keith Briffa subject: Re: western US trees data to: "Malcolm Hughes" Malcolm thanks for this - I simply wanted to repeat the analysis of the tree-ring data as was done by MIke (?) to get the 3 PC,s of the western US trees he used in his analysis , and explore the issue of the correction by comparison with the Jacoby data. I would love to be able to discuss this stuff with you and I am very interested in your papers on these data. I don't know what if anything will come out of it - but I would not consider writing anything on it without you. May I come back to you with any problems (data identification/extraction/interpretation etc) I don't suppose you have the exact set of data Mike used ? I presume he had a set of 27 chronologies as a starting point ? Thanks for your help. Keith At 12:54 PM 9/22/03 -0700, you wrote: Dear Keith - sorry to have taken so long to reply. After a somewhat disrupted summer, the beginning of the semester was made all the more messy by my computer getting seriously "wormed". I'll answer your questions as best I can. 1) As you didn't say, I assume that the western chronologies of interest are those used in Mann et al 1999. As the paper says, the first 3 PCs of these were used in that paper. Going back to the original files, I found 27 chronologies from the western US that reached back to AD 1000 of further at the time that work was done. I attach 3 MS Excel files that give some details of them, including the ITRDB identifiers, which is what Mike used to get the chronologies. The Excel files are abstracted from the many huge files Richard Holmes built for me during that seasrch of the ITRDB, so each will contain stuff that may be irrelevant: vchron11000 contains, inter alia, the ITRDB ID, species code, first year, last year, collector's name vchron41000 contains the ITRDB IDm then the first and last years with 5, 10, etc samples vchron81000 contains the ID, etc and then in the following cols: V mn sensitivity W chronology autocorrelation, AE number of series, AG mean correlation of series with chronology AH mean series autocorrelation, AI seores mean length, seires median segment length. Please remember that this set ranges from lower forest border to upper forest border, so that various mixtures from all precip to precip plus temp locally apply. 2) I have no idea of Mike and Phil used - I assume it was the PCs of those in this subset that go back to AD1, but I wasn't involved in any way, so I don't know. 3) As for Gordon's series - the MBH 99 paper says which one was used, but it never passed through my hands - Mike dealt with this. 4) I don't know what alternative standardization methods you had in mind, but you should be aware that it would be completely unjustified to assume that the first measured ring was anywhere near the pith in many of these sites, especially as you go back in time, where the chronologies are based on remnants that have weathered on the inside and the outside. For this, and related, reasons, it would also be completely unjustified to assume any constant, or small, distance in years of the first measured rings from pith. That is, I can see no way of making a remotely reliable estimate of cambial age in the vast majority of these samples. I am sitting on the bones of a manuscript in which I had someone spend several months checking many hundreds of bristlecone and similar cross-sections and cores in our store. They found only a few dozen - less than 10%, were either pith was present, or the innermost ring could reasonbly be described as 'near pith' . I have another manuscript in a slightly better state of preparation where we restandardized many of these series in the following way - identify the long, flat part of the smaple ringwidth curve (i.e. remove the 'grand period of growth', if present) and then fit a straight line of no or negative slope. 5) Of course, I'd be happy to collaborate - what did you have in mind? How are you doing these days? Cheers, Malcolm > > Dear Malcolm > just sending this again in case you did not get it last time (last > month) cheers Keith > > Malcolm > I am exploring the role of the tree-ring data in the various > reconstructions of Northern Hemisphere temperatures , particularly , > as you know, the issue of standardisation of the original data and the > relative influence of specific chronologies on the estimates of > uncertainty relating to the reconstructions. I am anxious to > understand the fine details of the role of the "western US trees " and > the "correction" applied (by comparison with Gordon's Northern tree > line data . Can you point me to the detailed information on what > constitutes the western tree-ring data , as used in Mann et al and > Mann and Jones (in press) , and to where we can pick these up to redo > some of the analyses on them (ditto the Jacoby data). We wish to try > alternative standardisation of both sets and explore the robustness of > the long-term trends etc. Of course I would be happy if yo would > collaborate with us on this. Cheers Keith > > -- > Professor Keith Briffa, > Climatic Research Unit > University of East Anglia > Norwich, NR4 7TJ, U.K. > > Phone: +44-1603-593909 > Fax: +44-1603-507784 > > [1]http://www.cru.uea.ac.uk/cru/people/briffa/ > Malcolm Hughes Professor of Dendrochronology Laboratory of Tree-Ring Research University of Arizona Tucson, AZ 85721 520-621-6470 fax 520-621-8229 -- Professor Keith Briffa, Climatic Research Unit University of East Anglia Norwich, NR4 7TJ, U.K. Phone: +44-1603-593909 Fax: +44-1603-507784 [2]http://www.cru.uea.ac.uk/cru/people/briffa[3]/