February 23, 2009 JCLI-2974 Dear Dr. Grodsky, I am now in receipt of all reviews of the manuscript entitled "Observed subseasonal variability of oceanic barrier and compensated layers" (JCLI-2974) by . While recognizing the significance of your work, both referees saw the need to strengthen physical interpretation of some of your results. Specifically, Referee A suggested additional analysis of observed MLD variability and possibly shortening/eliminating the discussion of tropical Pacific variability. Referee B would like you to explore the role of oceanic processes and atmospheric feedback of BL variability. Looking at Fig. 2a made me wonder if deep BL in the subpolar North Pacific is related to the subsurface temperature maximum layer observed in winter. See recent observational (Ueno and Yasuda 2000, JGR) and modeling (Endo et al. 2004, JPO) studies. The latter study shows that cross-gyre transport is important to maintain this stratification. On the basis of these reviews and my own evaluation, it appears that the manuscript may be acceptable for publication in the Journal of Climate subject to some major and mandatory revision. Acceptance is contingent upon you satisfactorily addressing the concerns and suggestions of the reviewers. Copies of the reviews are enclosed for your information. Please upload your revised manuscript along with a point by point explanation of how you have responded to the comments of each reviewer within three months of today's date. When preparing your revisions, please refer to the enclosed Revision Checklist. Because the AMS is requesting that all revision submission meet final production requirements, following this checklist closely will streamline the processing of your manuscript, should it be accepted at a future date. If you have any questions, please don't hesitate to contact me. Thank you. Sincerely, Shang-Ping Xie Editor for the Journal of Climate ****Referee A, JCLI-2974**** Review of "Observed subseasonal variability of oceanic barrier and compensated layers" (JCLI-2974) by Hailong Liu, Semyon A. Grodsky and James A. Carton. I recommend that major revisions be made to the manuscript in order to see it published. This manuscript contains some significant and original analysis of observed interannual to decadal variations in barrier layer and compensated layer thickness in select regions of the global ocean. However, the important signals identified in this paper (low frequency BL/CL changes in the North Pacific and eastern Atlantic) are not thoroughly explained, in my opinion. Variations in the tropical Pacific are convincingly explained, but these are not original results. I think major revisions to strengthen the arguments made regarding Pacific/Atlantic change would greatly improve the paper. In both regions, variations in MLD are invoked as important in explaining BL/CL variations, and so the observed variations in MLD should be incorporated in the analysis/discussion. A more in-depth analysis in these two regions could be offset by eliminating the analysis of the tropical Pacific. I've also listed below a number of minor issues with the paper, which taken as a whole, may add up to one major revision. A. Major Points: ================== A.1: The manuscript would be strengthened if the North Pacific trends shown in Figures 5 and 6a,b (lines 275-290) could be explained using observations. Are observations really insufficient to support the arguments made to explain this variability? In particular, I would think that the lag correlation of BL/CL thickness with mixed layer depth (Figure 8d) could and should be done with the observed mixed layer depth dataset, since there is a MLD associated with each and every BL/CL thickness data point. Furthermore, NCEP winds are adequate for the Figure 7 regression, and so might also be adequate to illustrate the ekman pumping/mixing/subduction differences (Figure 9) between two example years (say, 1980 and 1984 of Figure 6b) using World Ocean Database data. I'm not convinced that the points made with CM2.1 coupled simulation data could not actually be made with observational data. This should be attempted, if it hasn't already. Figure 8 and associated discussion suffers from several weaknesses: the scale of simulated BL/CL thickness in Figure 8a is half that used in Figure 2 (although this discrepancy is not even mentioned in the text!), March metrics are being compared to JFM metrics, the time series of CL thickness does not show the observed trend (one would not necessarily expect it to... although perhaps a historically-forced ocean simulation might?), and panel e of Figure 8 is not mentioned anywhere in the text. A.2: I recommend that the examination of CL variability in the Atlantic be substantially reworked. The discussion of figures 6c and 10 is inconsistent with the demonstration in Figure 5 of a long-term trend in Atlantic BL/CL thickness (see points B.18 and B.19). The statements in the abstract and conclusion which link CL variability to NAO variability are inconsistent with the statement that the CL variability is more year-to-year than decadal. These inconsistencies are not adequately reconciled in the paper. I don't understand why the authors attribute the lack of a trend in Figure 6c to insufficient sampling (line 337), since the northeast Atlantic is certainly one of the most well-sampled regions in the World Ocean Database. Consequently, I don't understand why the analysis of Figure 10 must be restricted to 10 years. Again, the weak conclusion that variability of the CL thickness is likely related to mixed layer depth variability could be strengthened by actually quantifying the observed correlation of these fields. I'd be very interested to see a maps of the correlations of March MLD and CL thickness, and March bulk Turner angle and CL thickness. Could the authors quantify the relative importance of salinity advection vs. surface fluxes on CL thickness in this region? Perhaps the box region is too large for a simple "one size fits all" account of subseasonal variations. A deeper analysis would generate more substantial conclusions than those given in lines 403-410. A.3: The analysis of BL variability in the western tropical Pacific box (Fig. 11, lines 356-363) is not worth inclusion. The findings are essentially the same as in Ando and McPhaden (1997). The analysis has the feel of being tacked on the end, with more fields plotted in Figure 11 than there are sentences devoted to the figure. Statements about the role of salt advection in modulating BL thickness in the central tropical Pacific (lines 361-363 and again 383-385) are unsupported by any analysis. The second paragraph of the "Conclusions" which deals with tropical BL variability could be kept as a short summary/commentary, but only if appropriate references are made. B. Minor Points: ================== B.1: "The first dataset" refers to what precisely? (line 132). B.2: This study relies on the same dataset and methods as Carton et al (2008). In that paper, it is revealed that only 25% of casts include both the temperature and salinity required for computation of MLD using the density criterion. Therefore, computation of BL/CL thickness is restricted to the subset of casts with salinity. Please clarify how much this restricts the amount of hydrographic data used in the present analysis. Also, comment on how this affects the time distribution of data and to what extent the sampling issues effect Figure 5. B.3: I find Figure 1 to be an interesting plot, and it deserves more significant discussion than the authors give it. Why is there a sharp increase in barrier layer thickness at Tu_b<-45deg? I'd like to know why the relationship isn't tighter, as I would expect. In particular, how can there be a barrier layer when Tu_b is greater than, say, 67.5deg, and delta(T)>0 and delta(S)>0? What is providing the stratification that makes MLD