date: Thu May 12 12:10:12 2005
from: Tom Melvin <t.m.melvin@uea.ac.uk>
subject: Re: Program
to: Kurt Nicolussi <kurt.nicolussi@uibk.ac.at>

   Kurt,
   Thanks
   Tom
   At 12:03 12/05/2005, you wrote:

     Dear Tom,

     I need to separate modern / sub-fossil trees to test RCS curves. Have you got a list or
     separate raw data files for each (with pith) as I should use the same trees as you use.

     You should have got three different files - historical samples, subfossil material,
     living trees - with the pith-offset information. If you still have these files you can
     produce different RCS curves.

     The RCS curve for the Youngest trees (ALPS -63) goes up instead of down! The climate
     signal is sufficient to change the slope of the RCS curve when all trees come from
     roughly the same period. Your separation of modern trees for an RCS curve will require
     the use of Signal Free methods to remove this distortion. I should be able to
     demonstrate this on your trees when I have a separate "modern" RCS curve.

     There are only few trees with an age between 40 und 63 years - that can also be by
     chance. And in reality the "modern trees" (=living trees) are from the last 400 years,
     many of them started growth during the Little Ice Age. So, the "modern RCS" curve
     obtained are from a wide climate range. For me the effects of coring position and the
     effects of systematic sample bias - the trees with long series are slow growing trees -
     are bigger

     One possibility is to use many RCS curves to standardise producing series of tree
     indices with means roughly 1.0 and then to use the relative magnitudes of the RCS curves
     (to each other) to "reinstate" overall growth rate of each individual tree before
     averaging to get a chronology. I will try this on your trees as well.

     Sounds interesting.
     Kurt
