date: Mon, 3 Jul 2000 14:02:55 +0100 (BST) from: Mat Collins subject: Re: Paper to: Tim Osborn On Mon, 3 Jul 2000, Tim Osborn wrote: > I was having a critical look through the manuscript, to try to anticipate > things that the reviewers might complain about. I came up with two concerns: > > (1) Should we have only used the pre-1900 part of the tree-ring > reconstructions to assess natural variability, to avoid the early 20th > century warming and the tree-ring decline post-1960? It's a question of > balancing record length with avoiding non-natural or non-temperature > changes. I think we could certainly argue against avoiding the early 20th > century warming, as this is unlikely to be only (or even mainly) > anthropogenic. Do we need to add a sentence to pre-empt the referees? I think we can get away with just ignoring the post (say) 1950 data by citing Simons detection and attribution work which says that the early century warming might have been solar. I have so far done the analysis on all the data so perhaps I should go back and do it with just post 1950. This probably wouldn't take too long and I don't think it will change any of the results. The tree-ring decline is the most often asked question when I talk about this work so I suspect the reviewers would pick up on it. I will re-do the analysis and add some text about this. > (2) The model regional averages were computed from all land grid boxes in > each region (I think). The reconstructions were based on the tree sites > shown on the map, calibrated against regional temperatures averaged across > all grid boxes with tree sites in them or with tree sites in an adjacent > box (when the temperature series were short or missing in the same box). > This is fine, I think, for most regions (and NH), but for TIBP and CAS one > might expect the tree-ring-based reconstructions to have more variability > than the model purely because of the small data coverage in the former. Do > you want to do anything about this? The ECCA region is also a worry for > the same reason. I guess this might be an issue. Can we quantify this effect? If it is small in comparison with the residual variance (for example) then I think we can justify our method. In any case we're doing loads better than the sort of people who compare a single ice-core with global temperature. My feeling is that we should just leave it and see if the reviewers are really worried about it. > Let me know what you think about these points. Also, is the wait for a new > front cover for HCTNs over, and is the paper now a technical note? If so, > what number? Finally, let me know if you want copies of the submitted > versions of the two tree-ring papers. It still hasn't appeared as a HCTN. Dunno if the front cover thing is still a problem - it could be that the guy hasn't updated the web page yet. I will try and find out. As for the submitted papers, I wouldn't mind seeing them sometime in the future (perhaps when we next meet up) but there's no hurry. By for now, Mat