cc: Hans.von.Storch@gkss.de, Eduardo.Zorita@gkss.de
date: Wed Jan  5 15:03:17 2005
from: Keith Briffa <k.briffa@uea.ac.uk>
subject: Re: Mike Mann's claims
to: Hans.von.Storch@gkss.de, mas@dmi.dk

   Dear Hans
   I can only think that these comments relate to my suggestion (made to Mike and others at a
   recent CLIVAR/PAGES meeting in Canada) that we ( ie those interested parties) need to
   organise a well formulated co-operative experiment , where the same data sets can be used
   to run a series of controlled comparisons of different calibration techniques.
   In discussing the results of the your work and earlier attempts to use the ORIGINAL Mann et
   al 1998,1999 techniques (by Eduardo and Irene ) I commented that we could still not be sure
   that the methods were being applied in the same way.
   Mike's more recent spatial transfer function work (Rutherford et al ) - on which Tim and I
   are co-authors , compared a different technique than he originally used , with our approach
   (based on local aggregation of chronologies and simple regression to produce regional
   reconstructions , which were then used in a PC regression to produce a NH series). These
   give similar results as regards NH land summer temperatures. This paper predates your
   Science piece. As we (Tim and I) believe the local calibration and aggregation approach is
   very likely to lead to underestimation of long-timescale  variance in the larger (NH)
   series , we believed it a logical conclusion that Mike's approach did so also. You then
   published your paper demonstrating that his original approach likely does just that.
   However, Mike has apparently shown that his NEW method , when used with pseudo proxies and
   a different climate model does NOT lose the low-frequency. He showed some results briefly
   in Canada that seemed to demonstrate this - but incidentally, DID seem to lose the high
   frequency extreme variability. I have NOT read this paper , which I believe is in review
   (NOT BY ME). Mike did not , however, use his original technique (that you purportedly
   follow in your Science piece, though others (not sure who ) that he is working with might
   be doing this.

   However the new Mann method DOES seem to produce the same NH curve as the original method.
   Hence an apparent contradiction . Hence my belief that the different methods should be used
   , with the same data (and targets) in a much more systematic way,  to see how they compare.
   This experiment should include the use of the same simulated data also.
   Tim and I have commented (in our perspective piece) that the loss of long-timescale
   amplitude in the NH history might be exaggerated by the results of your use of the ECHO G
   simulation - as this may produce too large an amplitude in  the first place (for different
   reasons) and so the question of how great the underestimation in the Mann original NH
   reconstruction is in real life is still a moot point - and hence the real need to set up
   this controlled test using the alternative simulations, methods.
   I have nothing in press other than the Rutherford paper and have said nothing other than
   the above.
   I would like to set up the comparative exercise - as part of SOAP perhaps - but with the
   collaboration
   of Mike and his colleagues if they are willing.
   Hope we can discuss this at the next SOAP meeting
   best wishes
   Keith
   At 18:34 04/01/2005, Hans.von.Storch@gkss.de wrote:

     Dear Martin, dear Keith,
     Mike Mann has made the following claims:
     "You should be aware that a comment is in press in "Science" casting significant doubt
     on the claims of Von Storch and Zorita, and another paper, in review, suggests that
     their conclusions are incorrect, . A 3rd paper, by the Danish Meteorological Institute
     (DMI) group using the same model as von Storch, cannot reproduce the von Stroch results
     (they find much less variability than von Storch), suggesting that their were some
     serious problems with the von Storch simulation as well as with their analysis of the
     simulation results. I would suggest that you get in touch with these individuals (e.g.
     Keith Briffa: k.briffa@uea.ac.uk or Martin Stendel: mas@dmi.dk) for a more balanced view
     of the Von Storch claims."
     May I ask you to verify this statement and, if true, send us a copy of your papers? And,
     possibly, explain your "more balanced views"?
     All the best
     Hans
     Hans von Storch
     Institute for Coastal Research, GKSS Research Center
     Max-Planck-Strasse 1, 21502 GEESTHACHT, Germany
     ph: +49 4152 87 1831, fx: +49 4152 87 2832
     mobile: + 49 171 212 2046
     [1]http://w3g.gkss.de/staff/storch; storch@gkss.de

   --
   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/

