cc: john.f.mitchell@metoffice.gov.uk, Eystein Jansen , Jonathan Overpeck date: Fri, 01 Sep 2006 14:01:37 +0200 from: Stefan Rahmstorf subject: Re: Fwd: Re: [Wg1-ar4-ch06] NEW DRAFT FOR LA REVIEW to: Keith Briffa Dear Keith, you disagree with my proposed revision of the paragraph re. the Von Storch papers, but you do not give any reasons or arguments for that. I think there are some good reasons to shorten this discussion and to clarify it, and I would welcome to hear your reasons against it. Firstly, I think your original discussion was too long and complex to understand for non-specialists, and, at this level of detail, not policy-relevant. It took up a disproportionate amount of space for what we can learn from it. Secondly, I don't think we need to cite all those Storch-spinoff papers by Bürger/Cubasch. Most people whose judgement I value (e.g., David Ritson, who I think has no vested interest but a very detailed knowledge of the issue) think these papers are irrelevant at best and misleading at worst (he actually has used stronger wording). You may also have seen that the latest in this series, making similar points, is highly criticised by anonymous reviewers on the open discussion site of the journal Climate of the Past, where one reviewer (this is not the even more scathing review by Mann) recommends rejection of the Bürger/Cubasch paper because of "numerous errors and inaccuracies in the use of statistical concepts and methods". Third, if we cite Von Storch et al. 2004 we need to be very clear that a number of key statements are simply incorrect, which is a fact that is not in dispute and documented in the literature. They implemented the Mann et al. method incorrectly, and it is at least unclear whether in their follow-up paper they have now fixed this (Ritson, who discovered the problem in their original paper in the first place, thinks they still have a problem, the detrending step was not the only one - and certainly in no paper have VS et al. shown any test that verifies their algorithm). Also, they were hiding a major artificial climate drift (which they must have known about, and which makes up half of their climate signal) - it is at least unclear whether you can expect a proxy method based on physical patterns of climate variability to reconstruct an unphysical drift, which has a completely different pattern. I simply think that because of this flaw, we cannot trust or cite any results from this particular ECHO-G run, which also affects several of the Bürger/Cubasch papers using the same data set. Given that the VS04 paper was used in the US Senate and other high-profile fora to discredit IPCC, I think it is imperative that we clarify this and leave our readers in no doubt about the fact that the VS04 results have proven to be incorrect in a major way. I am aware that you authored a favorable Science Perspective on the VS04 paper at the time, but you could not have known of those errors back then, and for a long time I thought myself that it was a valid paper. Therefore, if we state clearly in our chapter what is wrong with it, I do not think this would be a loss of face for you - quite the contrary. I also think you have done a brilliant job on the rest of the very difficult discussion of the past millennium. Best wishes, Stefan -- To reach me directly please use: [1]rahmstorf@ozean-klima.de (My former addresses @pik-potsdam.de are read by my assistant Brigitta.) Stefan Rahmstorf [2]www.ozean-klima.de [3]www.realclimate.org