date: Fri Sep 23 10:55:40 2005
from: Tim Osborn <t.osborn@uea.ac.uk>
subject: Re: Fwd: McIntyre Jones and Briffa
to: Keith Briffa <k.briffa@uea.ac.uk>, p.jones@uea.ac.uk, t.m.melvin@uea.ac.uk

   David and I were just doing a quick search for "briffa error" on google and it came up with
   this page:
   [1]http://www.climateaudit.org/?p=159
   with discussion of modern sample bias and a link to Tom's PhD thesis, plus quotes from it,
   etc.  They may well have got some of the ideas in the abstract below from Tom's thesis!
   Hopefully nobody will preempt your papers on standardisation and bias.
   Cheers
   Tim
   At 13:19 22/09/2005, Keith Briffa wrote:

     Keith,
     found this posting on McIntyre´s blog. Something in the making, just to warn you if you
     did not see it yet.
     Cheers
     Eystein
     9/16/2005
     U.S. climate change Science Program Workshop
     I've had an abstract accepted for the U.S. Climate Change Science Program (CCSP)
     Workshop, "Climate Science in Support of Decisionmaking," to be held November 14-16,
     2005. My abstract is entitled: "More on Hockey Sticks: the Case of Jones et al [1998]".
     The workshop is described here. Here is my abstract (which moves on from MBH):
     More on Hockey Sticks: the Case of Jones et al [1998]
     Abstract. Multiproxy studies purporting to show 20th century uniqueness have been
     applied by policymakers, but they have received remarkably little independent critical
     analysis. Jones et al. [1998] is a prominent multi-proxy study used by IPCC [2001] and
     others to affirm the hockey stick shaped temperature reconstruction of Mann et al.
     [1998]. However, the reconstruction of Jones et al. [1998] is based on only 3-4 proxies
     in the controversial Medieval Warm Period, including non-arms-length studies by Briffa
     et al. [1992] and Briffa et al [1995]. We show that the Polar Urals data set in Briffa
     et al [1992] fails to meet a variety of quality control standards, both in replication
     and crossdating. The conclusion of Briffa et al. [1995] that 1032 was the "coldest year"
     of the millennium proves to be based on inadequate replication of only 3 tree ring
     cores, of which at least 2 are almost certainly incorrectly crossdated. We show that an
     ad hoc adjustment to the Tornetrask data set in Briffa et al [1992] cannot be justified.
     The individual and combined impact of defects in the Polar Urals data set and Tornetrask
     adjustments on the reconstruction of Jones et al [1998] is substantial and can be seen
     to have the effect of modifying what would otherwise indicate a pronounced Medieval Warm
     Period in the proxy reconstruction. Inhomogeneity problems in the Polar Urals and
     Tornetrask data sets, pertaining to altitude, minimum girth bias and pith centering bias
     will also be discussed.