date: Tue, 30 Oct 2007 16:22:38 +0000 from: Gerard van der Schrier subject: [Fwd: Your Submission] to: Tim Osborn , Keith Briffa Hi Tim & Keith, Received the comments of two reviewers to the Clim. Dyn. paper last week. Haven't had the opportunity to look at is closely - but it doesn't look too good....... Cheers, Gerard Ref.: Ms. No. CLIDY-D-07-00069 Detecting trends in meridional heat transport in the North Atlantic using variations in sea surface height Climate Dynamics Dear Dr. van der Schrier, The reviewers' comments on your work have now been received. The comments are appended below. You will see that they are advising against publication of your work in its present form. Therefore I must reject it. However, I am ready to consider a new manuscript describing your work, substantially revised following the referees' comments. This manuscript should be submitted as an entirely new submission. Thank you for giving us the opportunity to consider your work. Yours sincerely Jean-Claude Duplessy Executive Editor Climate Dynamics Reviewers' comments: Reviewer #1: The authors are tackling a timely topic of how to deduce MHT changes from SSH using an optimal detection approach. The method is applied on a long control simulation of the HadCM3 coupled model from which the SSH response function to MHT changes and the estimate of the internal variability are derived. As the bottom line the authors state that changes need to be at least 0.024PW/yr to be deemed statistically different from zero. Of course this number means that MHT would decrease/increase 25% in 10 years from the mean MHT at 30N (=0.96PW as stated on page 4) before the method could find it significantly imprinted to SSH. Somehow this sort of large and rapid change in MHT should manifest itself with more confidence in SSH ... While the method appears straightforward, the authors fail to disclose salient information how it was actually applied. Since the main conclusion is that the basic method (which was presented) is not as reliable as a variant of the method, why not use the variant method to start with or at least present it in the same manuscript ? I am not convinced that the manuscript is publishable until the variant of the optimal fingerprinting method is presented. I don't want to see multiple manuscripts that present only partial and inadequate solutions when potentially better methods are available. Also at its present form, the manuscript is rather lightweight. Further comments: Page 5: discussion related to noise: the k highest variance EOFs of the control run, EOFs of what field ??? And using what time period ? Page 6: 2.2.1 regression of MHT and SSH, what time period ? A 1000 year segment ? Or was it split into smaller segments ? And how exactly was the response pattern constructed ? Page 6: Composite analysis, was the response pattern different from Fig. 1 ? Could be included to gain some information what the differences are ? Was the regression pattern chosen because it performed the best in the validation period ? Page 7: Estimation of climate noise issue: Here I got totally lost : 'SSH contains the response to MHT'. Is this a bit too late in the manuscript to worry about it ? Pages 9-10: 3.2: I don't understand how the noise pattern was derived for the altimetric SSH. Pages 10-12: Discussion: discussion of the drought index seems irrelevant except it exposes the same overlapping influences as MHT on SSH and NAO wind forcing (local Ekman pumping and Rossby waves) on SSH (the noise part). Reviewer 2:This article investigates the potential change (trend) in oceanic meridional heat transport (MHT) in the North Atlantic using altimetry-based sea surface heights (SSH) and runs from the HadCM3 climate model. Detecting MHT trends is important issue in the context of current climate change. Published results based on direct measurements are so far inconclusive, reported slowing down of the Atlantic meridional overturning circulation being barely above measurement uncertainty. I am not expert in climate modelling; Thus my comments will remain general (although I have a few detailed comments about the figures). General comments 1. A 10-year record of SSH data is quite short to detect trends. Why only use 10 years while nearly 15 years are now available (from the Topex and Jason-1 record)? 2. Some studies (e.g., Latif et al., 2005) have shown that the meridional overturning circulation is subject to strong multidecadal variability related to NAO low-frequency variability, suggesting that anthropogenic weakening may be hard to detect. Again a 10-year record may be too short. 3. Several studies have shown that regional trends in SSH (or sea level) are mostly due to the combined trends in thermal expansion and halosteric effects (see the recently published IPCC 4th Assessment Report, chapter 5; Bindoff et al., 2007). Thus instead of a short altimetry-based SSH record, why not use observed steric trends over the last 40-50 years? 4. Another possibility would be to use outputs (SSH and/or steric sea level) from OGCMs runs with data assimilation which provide 'data' for the last ~40 years (e.g., the SODA reanalyse; Carton et al, 2005). The global mean sea level trend is possibly unrealistic but the spatial trend patterns agree well with observed trends based on satellite altimetry (over their overlapping time span). I suggest the authors provide more convincing arguments for the use of such a short SSH record. Other comments - The section on model adjustments (sections 2.2.2 and 2.2.3) are hard to understand for a non-expert in climate modelling : for example, the differences between the 3 cases shown in Fig.3. By the way, if these figures are considered s essential by the authors, I suggest to combine them into a single one. - Fig.4 shows considerable low-frequency variability; But this is not discussed. Why? Cannot distinguish between simulated and 'observed' curves. Which is which? - What is the time unit in Fig.4? years? - The figures need legends on the X and Y axes - The figure captions are not enough informative. Should be expanded. -- ---------------------------------------------------------- Gerard van der Schrier Royal Netherlands Meteorological Institute (KNMI) dept. KS/KA PO Box 201 3730 AE De Bilt The Netherlands schrier@knmi.nl +31-30-2206597 www.knmi.nl/~schrier ----------------------------------------------------------