From: Michael Mann To: Phil Jones Subject: Re: Past Millennia Climate Variability - Review Paper Date: Thu, 13 Mar 2008 08:58:49 -0400 Reply-to: mann@psu.edu Hi Phil, Sorry, one other point. In item #4 below, the point that is being made, as shown (and discussed) elsewhere, applies both to the MBH method and the the canonical regression method (the latter is demonstrated in experiments by Wahl and Ammann not shown but referred to elsewhere in the text). So to be accurate and fair, the sentence in question on page 50 really has to be rephrased as follows: Examinations of this kind are shown in Figures 3a,b (and parallel experiments not shown) demonstrating that, at least for the truncated-EOF CFR method used by MBH98 (employing inverse regression) and the canonical regression method that has been widely used by many other paleoclimate researchers, there is some degree of sensitivity to the climatological information available in calibration. I realize there are many co-authors on the paper that have used the canonical regression method before, so perhaps there is pressure to focus the criticism on the MBH method. But that is simply not fair, as the other analyses by Wahl and Ammann not shown clearly demonstrates this applies to canonical regression as well--we can debate the relative sensitivity of the two methods, but it is similar. This is an absolutely essential issue from my point of view, and I'm afraid I cannot sign my name to this paper w/out this revision. I'm sure you understand--thanks for your help, mike Michael Mann wrote: Phil, Looks mostly fine to me now. I'm in Belgium (w/ the Louvain crowd) and only intermittent internet access, so will be difficult to provide much more feedback than the below. I hope that is ok? Here are my remaining minor comments: 1) the author list is a bit front-loaded w/ CRU folks. You should certainly be the first author, but the remaining order makes this paper look more like a "CRU" effort than a "Wengen" effort, and perhaps that will have an unintended impact on the way the paper is received by the broader community. I was also wondering how I ended up so far down the list :( I think I was one of the first to provide a substantive contribution to the paper. Was my contribution really so minor compared to those others? The mechanism behind the author list is unclear, partially alphabetical (towards the end), but partly not. You are of course the best judge of peoples' relative contributions, and if the current author order indeed represents that according to your judgment, then I'm fine w/ that. Just thought I'd check though. 2) page 45, 2nd paragraph, should substitute "(e.g. Shindell et al, 2001; Collins et al 2002)" for "Collins et al 2002" 3) page 48, 2nd paragraph, 3rd sentence, should substitute "RegEM (implemented with TTLS as described by Mann et al 2007) for "RegEM". 4) page 50, bottom paragraph, first sentence: I think that the use of "crucially" here is unnecessarily inflammatory and overly dramatic. This word can be removed without any detriment to the point being made, don't you think? 5) page 51, 2nd paragraph, logic does not properly follow in certain places as currently phrased (a frequent problem w/ Eugene's writing unfortunately!): a. sentence beginning at end of line 9 of paragraph, should be rephrased as follows: Mann et al. (2005) used pseudo-proxy experiments that apparently showed that this method did not underestimate the amplitude of the reconstructed NH temperature anomalies: however, Smerdon and Kaplan (2007) show that this may have been a false positive result arising from differences between the implementation of the RegEM algorithm in the pseudo-proxy experiments and in the real-proxy reconstructions which leads to a sensitivity of the pseudoproxy results to the calibration period used (also noted by Lee et al., 2008). b. the sentence following the one above should be rephrased: Mann et al. (2007; cf. their Figs. 3-4) demonstrate that a variant of the RegEM method that uses TTLS, rather than ridge regression produces an NH temperature reconstruction whose amplitude fidelity does not exhibit the calibration interval dependence of the previous implementation by Mann et al 2005, and yields reconstructions that do not suffer from amplitude loss for a wide range of signal-to-noise ratios and noise spectra (though Lee et al., 2008, suggest that an appropriately implemented ridge regression can also produce good results). c. the sentence following the one above should be rephrased: With TTLS as implemented by Mann et al (2007), RegEM performs without amplitude loss in model-based tests (versions without trend removal), including using the high-amplitude ECHO-G model output utilized by Bürger et al. (2006), von Storch et al. (2006), and Küttel et al. (2007) to examine truncated-EOF methods. 6) page 52, 1st paragraph, 7th line, the reference ot "the MBH reconstruction" is erroneous, because the tests have nothing to do w/ the MBH reconstruction per se, only--potentially-the MBH method under certain circumstances. In fact, Mann et al (2007) [and Wahl and Amman(2007)] both show that the actual amplitude loss realized in the MBH reconstruction in reality is probably quite small. This very point is made at the top of page 53! So the reference to "the MBH reconstruction" needs to be eliminated here. It is already clear by context what this is actually referring to (idealized experiments using both the MBH and canonical applied to surrogate proxy networks). 7) Re, Caspar--well he seems to be in his "non-responsible" phase right now, hasn't replied to my messages either. Will keep on trying, let me know if any of the above needs further elaboration. we're travelling for the weekend but will still have intermittent email access, mike Phil Jones wrote: Dear All, Attached is the penultimate draft of the Wengen paper. If you have time can you look through this. If you've not much time, can you look through your sections and the intro/conclusions. I hope we in CRU have got all your comments in. We have been through them all - including Gene's which came last night and Francis' the night before. WE URGENTLY NEED CASPAR TO REPOND. Can Gene, Mike and anyone who can get Caspar to respond to emails tell him that there are a few questions in this draft we need him to respond to. We need better versions of Figure 3, plus there are some flagged points in Sections 3 and 4. Juerg - is Figure 5 OK. If not resend separately - don't embed as this screwed up last time. Plan A is for us to submit this to The Holocene next Wednesday. So we need by then, from each of you a quick email to say you've got this and any comments by next Monday - March 17. Submission will be March 19. There is no Plan B. With the Feb 20 email, there were no responses from Peck, Eystein and Nick. If we don't hear from you three by next week, we will remove you from the author list! If anyone knows if any of these three are in the field please let me know? Things to check: 1. Everybody happy with the author order. The idea here was the three us in CRU, the main authors of the sections in section order, then others in alphabetical order. 2. If you have time also look at sections 2.5 and 2.6. Issue here is - is there enough there. Thanks to Juerg for some of these sections. 3. There are a couple of refs (Juerg) we need - Buntgen et al. and D'Arrigo et al. Next week, we (CRU) will be working on the alterations- using IPCC rules. These are - if you want a change justify it, and if you say this is unbalanced, or just European, or emphasizes Lee et al. (2008), then gives us the additional text to make alterations. We've left a few comments in where these sorts of comments were made last time. There will be time to make alterations while The Holocene reviews it. It will also be better to read it later when there is time after submission. I've not read this version yet, so apologies if there are any pieces of poor English. I will be reading again this weekend. Finally, Thorsten, if you think I've missed anybody off this email, forward and let me know. Juerg needs to send on to the others within Bern. Cheers Phil Prof. Phil Jones Climatic Research Unit Telephone +44 (0) 1603 592090 School of Environmental Sciences Fax +44 (0) 1603 507784 University of East Anglia Norwich Email [1]p.jones@uea.ac.uk NR4 7TJ UK ---------------------------------------------------------------------------- -- Michael E. Mann Associate Professor Director, Earth System Science Center (ESSC) Department of Meteorology Phone: (814) 863-4075 503 Walker Building FAX: (814) 865-3663 The Pennsylvania State University email: [2]mann@psu.edu University Park, PA 16802-5013 [3]http://www.met.psu.edu/dept/faculty/mann.htm -- Michael E. Mann Associate Professor Director, Earth System Science Center (ESSC) Department of Meteorology Phone: (814) 863-4075 503 Walker Building FAX: (814) 865-3663 The Pennsylvania State University email: [4]mann@psu.edu University Park, PA 16802-5013 [5]http://www.met.psu.edu/dept/faculty/mann.htm References 1. mailto:p.jones@uea.ac.uk 2. mailto:mann@psu.edu 3. http://www.met.psu.edu/dept/faculty/mann.htm 4. mailto:mann@psu.edu 5. http://www.met.psu.edu/dept/faculty/mann.htm