While I still argue for additional refinement of text prior to publication, I appreciate the many additions of text (e.g., more detail on methodology) that the author has provided in this iteration; the transparency and reproducibility of the second draft of the manuscript has greatly improved. However, I still have concerns regarding the (claims of the) methodology and lack of quantitative information. (I apologize in advance for my repetitive nature but I tried to detail a concern every time I had it as the text progressed.)
*Note: the line numbers used here refer to those in https://www.ocean-sci-discuss.net/os-2017-57-AC2-supplement.pdf (which are not numbered consecutively between pages). Pages and paragraphs used here refer to those in os-2017-57-AC2.pdf
///////////////////
Figures
1. All figure axes (except Figure 3) are still too small for clear readability.
2. Thank you for updating the analysis to include most recent 2016 data. Heads up: now there are differing time periods 1993-2015 and 1993-2016 between Figures 2-4.
Lines 184-186
The author states, ‘we develop a new method to study variability in the position of the fronts in the Southern Ocean’ but no quantitative results on front variability, other than N-S shifts, are presented. Percentages are provided in lines 488-539 but this pertains to the shapes of the enhanced CKE envelopes and not the front location(s) found within the envelopes. For example, what is the standard deviation in the 8 distinct CKE half-power point values (‘front latitudes’) at each longitude (i.e., variability in the ‘front’ itself)?
Lines 190-191
This is the most accurate statement of the study: ’we track the positions of these envelopes from 1993 until 2016 to quantify if the envelopes have shifted…’. Given that these envelopes can span more than 5 degrees latitude N-S, and that multiple (traditional) fronts can exist within one envelope, I take away that this study tracks shifts in a mean ACC position rather than individual fronts themselves, as claimed. The fact that the community is accustomed to thinking of fronts in a particular way, it is important that the author makes their definition of a ‘front’ as clear as possible.
Line 214
Describe these ‘envelopes’ more. I appreciate Figures 3 and 4 but this is at one specific pass in the Indian sector. How do these envelopes vary by region, etc.? What is the average meridional width/extent, etc.? What is the standard deviation in the width of the envelope at each longitude?
Lines 348-369
Please comment on the reasoning behind the choice to bias the methodology to the Orsi study, in particular.
Line 376
“The mean CKE profile pictured in Figure 3 has multiple local maxima, most likely associated with the narrow jets that surround the front.” Can the author comment on how the half-power point method finds one front in this latitude band while the Orsi study finds one instance of the SACCF, two instances of the PF and one instance of the SAF (= 4 fronts) at this longitude/swath. This might suggest that the CKE method presented here can alias a front to the mean latitude of any front activity/presence (within the envelope) and not be particularly representative of a physically realistic front.
Lines 375-386
While I appreciate the author’s attempt to provide more information for this particular pass (via a modified figure to show variability and states the temporal range in number of defined maxima), this is for one distinct pass of the Indian sector. I conclude that no substantial quantitative information has been provided in the results in this iteration. Is there really no other quantitative information that the author can provide for this work? For example, mean and standard deviation in number of maxima or width of envelope, etc. One pass surely cannot represent the full Southern Ocean.
Lines 375-376, re: Figure 3
It would be helpful for clarity to add to the end of this sentence: “The mean CKE profile pictured in Figure 3 has multiple local maxima, most likely associated with variability of the narrow jets that surround the front (defined here as the location of the half-power point).”
Lines 377-378
I assume these percentages are pulled from the Chapman (2017) study - could the author place their work into context? - otherwise this feels out of place/unnecessary.
Lines 379-380
What might this suggest about the front field?
Lines 389-390
So again, this study does not find individual fronts (even if speaking in a mean sense) but a mean latitude of frontal activity (especially as these envelopes can sometimes span ~6 deg. latitude, which is much greater than the 2-3 degree standard deviation cited).
Lines 390-420
Please be clear on how the half-power point is distinct from just a sort of ‘mean latitude’ of the latitude band encompassing any ‘ACC activity’ (the full enhanced CKE envelope being the ‘ACC activity’ region) — and therefore the distinction between tracking actual fronts and not just a sort of ‘mean ACC position’ like Gille 2014 (if in fact this study is more like the latter, fine, this just needs to be clearly articulated to the reader).
page C14, paragraph 2
1. As hoped, the addition of Figures 6 and 7 have not yet alleviated my previous concerns about actual front identification. This stems from most of my other comments here on actual frontal detection vs. a mean ACC position.
2. I don’t believe the addition of Figure R2 to the manuscript is necessary. As I understand it, the Chambers ‘front’ would then be defined as the half-power point (not plotted here but) located to the south of the other studies’ fronts (between 50 and 51S; same as red dot in Figure 3)? If thinking in terms of actual named fronts, the Chambers front detection method would find an equivalent PF at this particular location, not the SAF. And when comparing to Figure 7, this appears to be the mean latitude of the colored variability area.
page C3, paragraph 3
Thank you for testing this. I don’t think that including the one sentence you provided as a response will take up too much space in the manuscript and will only benefit transparency: ‘We find that mean front positions are not sensitive to choice of x-year periods.’
Figure 6
1. I understand that the author added Figure 6 to boost the argument for/robustness of the presented methodology (given that the mean half-power point lies near previous climatologies) but it does not necessarily confirm that the methodology is actually locating fronts at a given time. As in the first review, I still feel strongly that the text of the manuscript misrepresents/misstates what is actually being performed.
2. It is very difficult to decipher individual fronts within the same color grouping (particularly the orange and blue groupings), font sizes are too small throughout, and the overall quality/resolution of the figure is low.
3. Since the author claims that envelopes are being tracked, doesn’t this suggest that some sort of envelope should also be indicated around your mean CKE dots as well? Single dots hide the fact that there is a known width (meridional extent N-S where CKE exceeds 200 units) that is unique to each location and time step (e.g., something like the spread in the colors plotted in Figure 7). It would help quantitatively to provide either a mean or standard deviation in the N-S width of the envelopes surrounding the mean CKE latitudes. And then going further, any detected shifts could be interpreted in light of this underlying/inherent envelope width.
4. Please comment on the result that most of the CKE method findings would appear to be the SAF, as revealed in Figure 6. I know I asked this previously but this same line of thought came up again this round. And I honestly don’t know what to make of the author’s response on page C15, second to last P. If anything, this confirms that this study is detecting an ACC feature but with no information on which feature unless you relate/compare to previous studies/climatologies.
Figure 7
Very difficult to decipher colored dots; perhaps make this a 4-panel figure with x-y plots allowing for zooming in on the frontal area (e.g., limiting the latitude range of each)?
Line 546
I disagree with the implication/tone of this sentence. The fact that the various studies’ climatologies don’t lie on top of one another is not that the calculations that went into finding the fronts are uncertain, but is a result of different methodologies used, which includes different time periods and different data sets. Each front analysis should be treated and interpreted based on its own methodology. A caveat is different from an uncertainty.
Line 547
Please replace the period with a colon (prior to ‘For instance,…’) to make clear that the author is discussing the spread in the various estimates across the different studies.
Lines 552-553
This sentence is muddy. Does the author mean, “Many front location estimates, defined as the half-power points of enhanced CKE, are found within the spread of the PF or SAF across multiple studies.” ?
Lines 554-557
I’m not sure this comparison is fair and appropriate. This study has not claimed to find the PF in particular, and in fact, most half-power points seem to lie along the SAF (as determined by using previous climatologies to compare qualitatively); we have not been provided any information on the standard deviation of the SAF. As a reader, I feel misguided here.
Line 558
‘Probably a better method…’ ??? The author presents the methodology but then states that a different method presented by another study (in this case, Chapman, 2017) is superior. Please address.
Lines 565-567
Since this is the only discrepancy that the author points out, it warrants further discussion. Please comment. Personally, I’m curious whether this is an artifact of the half power-point methodology - it is curious that the half-power point is found between two regions to the north and south where Chapman would ‘probably’ find a feature (colored in this figure) and likely matches the north and south bounds of an ‘envelope’ identified by the author. Has the author isolated this particular region/swath and looked at the features of this envelope (like Figure 3)? Is the half-power point aliasing the front between two major peaks here?
Lines 568-571
Please update/remove casual/informal language like ‘the good comparison’ here.
Lines 635-636
I’m not sure this claim can be made. The author makes no distinction between the PF and SAF throughout the manuscript but here, focuses on shifts in this study’s SAF specifically. Please address.
Figure 8
This study tracks the half-power points of the envelope of enhanced CKE. Therefore, they inherently have an underlying width. If we had information on the width of the CKE envelopes (i.e., ACC activity), we might also be able to confirm whether these shifts are significant given the widths (i.e., does the confidence interval associated with the change in the single point exceed the width and variability of the underlying envelope?). I guess that follows from whether the author could provide a sort of error statistic that relates the half-power point to the associated width of each envelope. |
- Printer-friendly version
- Supplement
