cc: 'Michael Grabner' , t.m.melvin@uea.ac.uk, Reinhard Boehm , 'Phil Jones' , 'Maurizio Maugeri' , 'Michele Brunetti' , jan.esper@wsl.ch, 'Ulf Buentgen' date: Thu Dec 14 14:37:45 2006 from: Keith Briffa subject: Re: AW: A plot to: David Frank , Kurt Nicolussi Hi David and others The resilience of the tree-ring information , I agree , seems only to be enhanced by the multiple data set comparison. The issue of the specific "band limited" calibration is an important one here , in as much as the different data sets will require different optimal scaling (calibrations) , and the reconstructions should be considered along with their appropriate uncertainty bands. Your remarks on the density , support our ideas regarding the possibility (or even desirability ) of using "band specific calibrations" , as we discussed in the paper by Tim and myself (resurrecting the original idea by Joel). It is desirable to show the separate band reconstructions (and verification performance and regression coefficients) . Having said all this , it remains likely that difference between temperature and tree indices is pervasive . I was interested also to see that in a previous message ( as copied by Kurt) that your group is working on putting all the long Alpine temperature sensitive tree-ring data together - we ( Tom and I with Kurt and Michael) were also working towards this (hopefully with the benefit of the data your group has published) as originally outlined in the ALP-IMP plans, and I wonder what the precise plans you have ? We would not like to work at cross purposes. Cheers Keith At 12:24 14/12/2006, David Frank wrote: Dear Kurt (and all others). Thanks for the nice figures. I can only agree with your demonstration and point that a combination of all suitable data should produce a more robust estimate for past temperature trends. It is more and more apparent that any record which we consider a temperature proxy underestimates the early instrumental warm season warmth. The general tendencies displayed by the newer datasets that you show, seem to be consistent with some comparisons between the early instrumental records and other previously described tree-ring recons. However, in response to Reinhard's question to the tree-ringers, I could easily say there could be a whole variety of reasons why the tree-ring data contain more low-frequency variability than they should. The troubling part is that we can, and have, put out lots of hypotheses why these records all tend to "undershoot" the early instrumental data. From your graphs (and other quicker comparisons that i have done), it appears that Ulf's LADE-MXD record slightly underestimates the recent warming trend in the last 20 or so years in comparison to most other records (and also the instrumental data). During the earlier periods it seems to generally fall in the middle of the crowd and also captures the higher-frequency variability in the inst. records very well over a 240 year period. It seems like an advantage to be able to see how as many independent records as possible lie on the spaghetti plate. Perhaps, Keith or Tom have some helpful insights... Any thoughts on biological autocorrelation(esp. for MXD data) and detrending issues? best wishes, David Quoting Kurt Nicolussi : Dear Reinhard et al., here some plots (attached file) based on slightly different chronologies from Alps - the well know Büntgen et al. larch MXD, the Tyrol spruce MXD, the Pinus cembra TRW and a new Larch chrono (region of the Tyrol, combination of living trees, hist. and subfoss. material - RCS, power transformation) - the first two plots show the four series, single years and about 20 year smoothed, the other show some comparisons between the combined 4 chrono's and temperature data - especially the last plot indicates that the residuals are much better for the combined record. best regards Kurt Jan Esper wrote: Dear Reinhard et al., this is a fascinating discussion and enjoyed very much looking at the files you sent earlier. I just wanted to add that it would be great if you could wait a bit more until Dave came up with some first ideas on optimally combining all the long-term tree-ring data (that might not take too long anymore). I am absolutely convinced that we will produce an improved record including all the new tree-ring data, and that this record will include useful error estimates which might serve as an agrument to do more or less adjustments to the early instrumental data. I am also pretty sure that the combined record will consider certain frequency bands from certain datasets and parameters. Best wishes --Jan At 14:04 Uhr +0100 13.12.2006, Reinhard Boehm wrote: Dear Phil, Maurizio and Michele, Please apologise me not taking part actively enough in our discussion at the moment. The reason is, that since the beginning of last week I am confronted with a really incredible "hype" of the media which eats up all my time. The reason was a half-page message of our press-manager to the APA (Austrian press Agency) about the final ALP-IMP report. Since then I have done not much more than talking and writing about climate change topics, most of it not in relation to the project but about this year's warm autumn, these weeks winter tourism problems in the Alps and so on. So please do not believe I'm not anymore interested in our topic, I follow all your mails and I only want to tell you that I am also more tending to believe that may first version, to fit the early period exactly to the TR-series, may be somehow exaggerated. So Phil's last proposal, to adjust the JJAS by a bit less than I did, seems to make sense also to me. And I also think that something like a Zero-adjustment for winter would be the best solution. The only thing we should consider would be how to describe our arguments for doing so. I would also be interested about the opinion of the treering group about that: As you see, we "instrumentalists" have now come to a point short before deciding on a definite set of monthly adjustments for the early instrumental series which we think should be somewhat less than the total offset versus the TR-series. Do you have arguments to support this? Do you have ideas why TR-series tended to systematically towards a cold bias in these years? Best regards Reinhard -----Ursprüngliche Nachricht----- Von: Phil Jones [[1]mailto:p.jones@uea.ac.uk] Gesendet: Mittwoch, 13. Dezember 2006 12:32 An: Maurizio Maugeri; Reinhard Boehm Cc: Michele Brunetti Betreff: A plot Thanks Michele ! The first of the two plots is the one I'm talking about. Dear All, Apologies for filling your boxes. Here is a plot. This is for average JJA (daily) temps for 1961-90 (red) and 1772-1820 (black). This is all daily T. The middle lines are the averages, the outer solid lines are the 1st and 9th deciles (10the and 90th percentiles) and the dotted lines are the absolute extremes for Tmean. A plot like this for Milan or somewhere else in Northern Italy would be interesting. This implies to us that CET is OK. This makes it harder to change your summers that much - well not as much as NITA would imply. Cheers Phil Prof. Phil Jones Climatic Research Unit Telephone +44 (0) 1603 592090 School of Environmental Sciences Fax +44 (0) 1603 507784 University of East Anglia Norwich Email p.jones@uea.ac.uk NR4 7TJ UK ---------------------------------------------------------------------------- -- Ao.Univ.Prof. Dr. Kurt Nicolussi Tree-ring Group / Institute of Geography University of Innsbruck Innrain 52 A-6020 Innsbruck Tel +43 512 507 5673 Fax +43 512 507 2806 -- David Frank Eidg. Forschungsanstalt für Wald, Schnee und Landschaft WSL Paläo-Klimatologie Zürcherstrasse 111 CH-8903 Birmensdorf +41 44 7392 282 +41 44 7392 215 david.frank@wsl.ch [2]http://www.wsl.ch ----------------------------------------------------------------------------------- This message was sent using IMP ([3]http://horde.org/imp/) at WSL ([4]http://www.wsl.ch). -- Professor Keith Briffa, Climatic Research Unit University of East Anglia Norwich, NR4 7TJ, U.K. Phone: +44-1603-593909 Fax: +44-1603-507784 [5]http://www.cru.uea.ac.uk/cru/people/briffa/