Monday, April 9, 2012


cc: Keith Briffa <>,,, Fortunat Joos <>,
date: Tue, 28 Feb 2006 18:32:25 +0100
from: Stefan Rahmstorf <>
subject: Re: latest draft of 2000-year section text
to: Tim Osborn <>

Hi Tim,
my simplistic interpretation as an outside observer of this field is:
VS04 published a high-profile analysis in Science concluding that the performance of the
MBH method is disastrously bad. Subsequently, VS in the media called the MBH result
"nonsense", accused Nature of putting their sales interests above peer review when
publishing MBH, and called the IPCC "stupid" and "irresponsible" for highlighting the
results of MBH. This had *major* political impact - I know this e.g. from EU negotiators
who were confronted with this stuff by their US colleagues.
Then it turns out that they implemented the method incorrectly. If it is done as MBH did,
variance is still somewhat underestimated in the same pseudoproxy test, but only a little,
within the error bars given by MBH and shown by IPCC. Certainly nothing dramatic - one
could conclude that the method works reasonably well but needs improvement. This would have
been a technical discussion with not much political impact.
What VS and their colleagues are doing now, rather than publishing a correction of their
mistake, is saying: "well, but if we add a lot more noise, or use red noise, then the MBH
method is still quite bad..."
The question here is: should our IPCC chapter say something to correct the wrong impression
which had the political impact, namely that the MBH method is disastrously bad? This is not
the same as the legitimate discussion about the real errors in proxy reconstructions, which
accepts that these reconstructions have some errors but are still quite useful, rather than
being "nonsense".
Cheers, Stefan
To reach me directly please use: [1]
(My former addresses are read by my assistant Brigitta.)

Stefan Rahmstorf

No comments:

Post a Comment