cc: fiona.smith@metoffice.gov.uk, "Tett, Simon" date: Thu May 12 09:35:28 2005 from: Tim Osborn subject: Re: Palaeo reconstructions to: "Jenkins, Geoff" Thanks for the opportunity to see this, Geoff. I have a few points to raise: (1) There were no years, or indeed any x-axis labelling, on the figure. (2) Perhaps this will be said in a caption, but please make sure that somewhere (caption or figure) it is clear that this is northern hemisphere mean annual temperature, rather than global mean annual temperature. (3) It might also be worth saying that the apparent narrowing of the range in the first 2 centuries isn't because of smaller uncertainties, but rather because of few reconstructions from which to define the "envelope". At 16:41 11/05/2005, Jenkins, Geoff wrote: Are you happy for us to use it in the booklet of slides we are doing? I will credit you as the source. (4) Yes, fine with me. I assume zero is the 61-90 mean for both instrumental and reconstructions. (5) Correct. My only concern is with the very obvious difference between reconstructions and instrumental over the last few decades. I recall Keith mentioning this when he gave us a talk, but I don't remember the reason. If we don't give one then the obvious mismatch will raise unanswered questions and may cast doubt on the reliability of the reconstructions. Does IPCC comment? I suppose another way round would be to simply miss off the last 30 yrs of the reconstructions. Your advice would be appreciated. (6) There are some issues with whether reconstructions capture the post-1980 warming. But in this case, these issues are dominated by effect of there being very few reconstructions past 1980 and thus (as with the first two centuries - point (3) above) the envelope is poorly defined. Indeed only one extends to the end of the envelope (hence the convergence to a point) and this is the borehole record which couldn't resolve any acceleration in warming on time scales of a decade or two anyway! I would suggest cutting off the reconstruction envelope in 1980, after which there are far fewer reconstructions available (and those that are available are based on fewer constituent proxy series). This seems honest because it will be clear that we are not using the reconstructions to "confirm" the recent warming, for which we rely on the instrumental temperatures, but taking it through to 1980 will be far enough to show that while the instrumental series is still just within the reconstruction envelope, it is not central and thus there are possibilities of reconstructions bias which deserve investigation even if they aren't disastrous to our interpretation. I think that's a fair view. Hope this is useful, Best wishes Tim