Assessing heat and freshwater changes in the Southern Ocean using satellite-derived steric height

Cocks, Jennifer; Silvano, Alessandro; Marzocchi, Alice; Dragomir, Oana; Schifano, Noémie; Hogg, Anna E.; Naveira Garabato, Alberto C.

doi:https://doi.org/10.5194/egusphere-2023-3050

Preprints

https://doi.org/10.5194/egusphere-2023-3050

Preprints

22 Dec 2023

| 22 Dec 2023

Assessing heat and freshwater changes in the Southern Ocean using satellite-derived steric height

Jennifer Cocks, Alessandro Silvano, Alice Marzocchi, Oana Dragomir, Noémie Schifano, Anna E. Hogg, and Alberto C. Naveira Garabato

Abstract. The Southern Ocean plays a central role in regulating the global overturing circulation, ventilating the deep ocean, and driving sea level rise by delivering heat to Antarctic ice shelves. Understanding heat and freshwater content in this region is key to monitoring these global processes and identifying multiyear changes; however, in-situ observations are limited, and often do not offer the spatial or temporal consistency needed to study long-term variability. Perturbations in steric height can reveal changes in oceanic heat and freshwater content inasmuch as they impact the density of the water column. Here, we show for the first time that the monthly steric height anomaly of the Southern Ocean south of 50° S can be assessed using altimetry and GRACE gravimetry data from 2002 to 2018. Steric height anomalies are validated against in-situ Argo float and CTD data from tagged elephant seals. We find good agreement in the ice-free ocean and parts of the seasonal ice zone, but that the uncertainty of steric height increases on the Antarctic continental shelf and within the permanent ice zone due to leakage error and anti-aliasing in GRACE. The open-ocean steric height anomalies exhibit spatio-temporally coherent patterns that: (i) capture the expected seasonal cycle of low (high) steric height in winter (summer); and (ii) reflect interannual anomalies in surface heat and freshwater content and wind forcing associated with positive and negative phases of the two major modes of Southern Hemisphere climate variability (the El Niño – Southern Oscillation and Southern Annular Mode).

Received: 18 Dec 2023 – Discussion started: 22 Dec 2023

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.

Download & links

Preprint (PDF, 2439 KB)

Supplement (2307 KB)

Download & links

Jennifer Cocks, Alessandro Silvano, Alice Marzocchi, Oana Dragomir, Noémie Schifano, Anna E. Hogg, and Alberto C. Naveira Garabato

Status: final response (author comments only)

RC1:
'Comment on egusphere-2023-3050', Anonymous Referee #1, 11 Feb 2024
This manuscript presents an analysis of steric height changes in the Southern Ocean in the last 20 years based on altimetry data. One major enabler of this study is the use of altimetry data in sea-ice covered areas, which have been made possible only very recently. Steric height anomalies are computed from the difference between Dynamic Ocean Topography (DOT) and barystatic height (BH) anomalies. Steric height estimates are validated using in situ data. Then, seasonal and interannual variability are presented, and correlations with climate indices are presented.
While the study is useful and would eventually deserve to be published, it first requires some revisions.

Major comment:
The title appears broadly misleading judging by the content of results. Heat and freshwater changes are not estimated, so their contribution to steric height changes is discussed only in a speculative way. This is fine, but the title should reflect core results, not discussion.

A more quantitative estimate of errors would be useful. You are merging information from two very different products. Altimetry has a better spatial resolution; GRACE has some problem of “leakage” (contamination by continental surfaces); temporal resolution between the two products might also differ. What would be the combined uncertainty on steric height estimates and is this expected to be sufficient to capture different scales of variability?

There is a problem with the smoothing applied on temporal curves in Fig. 4, 6 and 7. Judging by eye, the rolling mean is not centered, so it appears lagged by half the smoothing window (i.e. 6 months). This need to be checked, and I apologize if my eye is inaccurate.

The potential of the presented dataset appears under-used in the study. Why not add a map of steric height anomaly RMS? EOF of steric height changes would also be useful. A regression of steric height changes to SAM and SOI would help demonstrate a potential statistical link, together with some quantification of significance. At the moment, we struggle to see any clear novelty in the analysis, despite the obvious value of the dataset and potential of the method.

Minor comments:
l. 10: “The Southern Ocean *circulation* plays …”

l. 27: Reference Haumann et al 2023 absent from reference list. Check all references.

l. 29: “The South Atlantic [..] Oceans *waters* …”

l. 31: “extreme and unpredictable”, vague and potentially misleading. The weather might become more extreme, not sure about the climate. Not sure about the future of climate predictability either.

l. 172: I do not find the agreement between SHA and GHPA particularly striking in Fig. 3e. Fig. 3b seems more convincing. Can you be more specific/quantitative?

We politely suggest the authors to read and incorporate the citation Kolbe et al. 2023 in their work (a paper that went unnoticed as so many other studies from the Covid era), as they will find much information about steric height variability in the Southern Indian Ocean during a period overlapping their period of interest, that may be compared with their results. Notably, the relative importance of heat and freshwater changes is discussed in detail.

There is too much information in the Appendix. Appendix A is useful to interpret the main results. Appendix B could be shortened and put back into the main text. The validation of GRACE against ocean bottom pressure is useful and could be part of the main text. Fig. S4 should also be in the main manuscript. Section 13.5 has nothing to do in an Appendix.

References:
Kolbe, Marlen, Fabien Roquet, Etienne Pauthenet, and David Nerini. “Impact of Thermohaline Variability on Sea Level Changes in the Southern Ocean.” Journal of Geophysical Research: Oceans 126, no. 9 (2021): e2021JC017381. https://doi.org/10.1029/2021JC017381.
Citation: https://doi.org/10.5194/egusphere-2023-3050-RC1
- AC1:
  'Reply on RC1', Jennifer Cocks, 20 Jun 2024
  We thank the reviewer for their thorough assessment and helpful suggestions, which we will address in our revised manuscript.
  Major comment:
  The title appears broadly misleading judging by the content of results. Heat and freshwater changes are not estimated, so their contribution to steric height changes is discussed only in a speculative way. This is fine, but the title should reflect core results, not discussion.
  
  We agree with this and will revise the title to reflect the method and results only (e.g. Variability of Southern Ocean steric height assessed from satellite measurements of sea level and gravitational anomaly)
  A more quantitative estimate of errors would be useful. You are merging information from two very different products. Altimetry has a better spatial resolution; GRACE has some problem of “leakage” (contamination by continental surfaces); temporal resolution between the two products might also differ. What would be the combined uncertainty on steric height estimates and is this expected to be sufficient to capture different scales of variability?
  
  We agree with this in principle, however, the uncertainty of GRACE in this area is poorly understood. Global and regional estimates of GRACE uncertainty are available in the literature, but it is likely not appropriate to use those here, as it is thought to vary dramatically depending on bathymetry, proximity to land, presence of ice shelves and short-term barotropic mass movement. This is discussed in depth in the Appendix and we identify what further work would be required to improve this. For this study, we constrain our results to locations that a) are outside of the regions identified as very uncertain in the appendix, and b) show good agreement in the comparison of steric height against in-situ data, providing confidence that the variability is being captured.
  We do not plan to include uncertainty calculations (as this is a major endeavour on its own, and thereby beyond the scope of this paper), but will restructure the manuscript so that this is clearer; details of GRACE uncertainties will be moved from the appendices to the main text, as suggested by the reviewer in the minor comments.
  There is a problem with the smoothing applied on temporal curves in Fig. 4, 6 and 7. Judging by eye, the rolling mean is not centered, so it appears lagged by half the smoothing window (i.e. 6 months). This need to be checked, and I apologize if my eye is inaccurate.
  
  You are correct, this is because the smoothed data has been plotted against the incorrect time stamps, lagged by 6 months as you say. We will correct this.
  The potential of the presented dataset appears under-used in the study. Why not add a map of steric height anomaly RMS? EOF of steric height changes would also be useful. A regression of steric height changes to SAM and SOI would help demonstrate a potential statistical link, together with some quantification of significance. At the moment, we struggle to see any clear novelty in the analysis, despite the obvious value of the dataset and potential of the method.
  
  We do not show regressions of the steric height against the climate indices as the composite is more illustrative of the response. We acknowledge that more statistical analysis would be useful, therefore we will investigate RMS and EOF analyses for inclusion in the revised manuscript.
  
  Minor comments:
  10: “The Southern Ocean *circulation* plays …”
  
  27: Reference Haumann et al 2023 absent from reference list. Check all references.
  
  29: “The South Atlantic [..] Oceans *waters* …”
  
  31: “extreme and unpredictable”, vague and potentially misleading. The weather might become more extreme, not sure about the climate. Not sure about the future of climate predictability either.
  
  Noted, will change and re-word these
  172: I do not find the agreement between SHA and GHPA particularly striking in Fig. 3e. Fig. 3b seems more convincing. Can you be more specific/quantitative?
  
  Yes, we will consider other ways of displaying this figure (e.g. different rolling mean windows) or will comment on the similarity quantitatively, highlighting why the pattern shown is significant in this context.
  We politely suggest the authors to read and incorporate the citation Kolbe et al. 2023 in their work (a paper that went unnoticed as so many other studies from the Covid era), as they will find much information about steric height variability in the Southern Indian Ocean during a period overlapping their period of interest, that may be compared with their results. Notably, the relative importance of heat and freshwater changes is discussed in detail.
  
  Thank you for the suggestion, we hadn’t come across this paper but agree it looks useful for comparison and theoretical understanding. So we will consider this in the revised manuscript.
  There is too much information in the Appendix. Appendix A is useful to interpret the main results. Appendix B could be shortened and put back into the main text. The validation of GRACE against ocean bottom pressure is useful and could be part of the main text. Fig. S4 should also be in the main manuscript. Section 13.5 has nothing to do in an Appendix.
  
  Noted, will restructure as suggested.
  
  Citation: https://doi.org/10.5194/egusphere-2023-3050-AC1
RC2:
'Comment on egusphere-2023-3050', Anonymous Referee #2, 22 Apr 2024

This manuscript combines almost 20-years of altimeter and gravity satellite data to investigate steric height anomalies in the Southern Ocean south of 50^oS. The steric height anomalies are used to investigate mean, seasonal, and regional variability and correlation with SAM and ENSO. While the study is of potential interest to the community the methods used, presentation and interpretations of results are not always clearly presented. Thus, diminishing the potential impact of the main results. I suggest the manuscript requires a major review before it may be acceptable for publication.
Rather than provide line by line comments and suggestions, this review will provide overarching manuscript comments.

Major Comments.
I agree with Anonymous Referee #1 that title is not appropriate. The manuscript does not assess the Southern Ocean heat and freshwater changes, it does estimate the steric height changes.
I suggest the authors revise how information is written in the manuscript. It is important that clear and concise language that correctly describes the applied data methods is used. For example, but not limited to, section 2.4 explains how in-situ data are used to calculate geopotential height anomaly. It is unclear what is the value of the statement “ in which there are multiple pressure level” when the depth criteria are used to determine what profiles are used (Pmax > 500 dbar and Pmin<25 dbar and quality flag 1). Rather than beginning the sentence with “The maximum depth of 500 dbar…..”, perhaps it is better to begin with “ We calculate GHA relative to 500 dbar, where 500 dbar is sufficiently deep……
Another example, in figure 3 b the correlation coefficient is mostly greater than 0.25 but there are region where the correlation coefficient is less than 0 indicating strong negative correlation, and regions of 0 correlation coefficient (no relationship between SHA and GHPA). The white color in figure 3b may not actually show regions of no relationship but rather region with less than 6 in situ profiles. Here it is important to use a mask for regions where the Pearson correlation coefficient was not calculated. Similarly, in 3c the color bar axis should have a minimum value of 6 and a mask applied to regions with less that 6 profiles. While these are rather small correction, the accurate display of data is important. Finally, what is the value of figure 4d and 4e. A reader may ask why you show a region with anomalous in situ observations. Also, for 4d there appear to be data that in water depth less than 500 m.

Lines 176 to 180 need to be revised. How you validate the SHA is the sea-ice zone is important. What is meant by “ open Southern Ocean”? 60^oS to the seasonal sea-ice zone?
While the interannual SHA variability shows a potentially interesting signal, it is somewhat complicated, and the analysis undertaken needs further investigation and refinement. This more detailed consideration of the causes of the variability would enable the authors to reach much more conclusive findings.
I would suggest that the authors consider only the SHA and the relationship with the large-scale SAM and ENSO forcing for this current manuscript and continue to investigate the interannual and seasonal signals for inclusion in another manuscript.

Citation: https://doi.org/10.5194/egusphere-2023-3050-RC2
- AC2:
  'Reply on RC2', Jennifer Cocks, 20 Jun 2024
  We thank the reviewer for their constructive comments and helpful suggestions, which we will address in our revised manuscript.
  Major Comments.
  I agree with Anonymous Referee #1 that title is not appropriate. The manuscript does not assess the Southern Ocean heat and freshwater changes, it does estimate the steric height changes.
  
  We agree with this and will revise the title to reflect the method and results only (e.g. Variability of Southern Ocean steric height assessed from satellite measurements of sea level and gravitational anomaly).
  I suggest the authors revise how information is written in the manuscript. It is important that clear and concise language that correctly describes the applied data methods is used. For example, but not limited to, section 2.4 explains how in-situ data are used to calculate geopotential height anomaly. It is unclear what is the value of the statement “ in which there are multiple pressure level” when the depth criteria are used to determine what profiles are used (Pmax > 500 dbar and Pmin<25 dbar and quality flag 1). Rather than beginning the sentence with “The maximum depth of 500 dbar…..”, perhaps it is better to begin with “ We calculate GHA relative to 500 dbar, where 500 dbar is sufficiently deep……
  
  Noted. We will modify the example sentence and perform another thorough review with this in mind.
  Another example, in figure 3 b the correlation coefficient is mostly greater than 0.25 but there are region where the correlation coefficient is less than 0 indicating strong negative correlation, and regions of 0 correlation coefficient (no relationship between SHA and GHPA).
  
  We interpret Figure 3b qualitatively only in the manuscript but will improve this interpretation using figures from the data and some basic statistics. In response to the first reviewer comment, we also acknowledge that more statistical analysis would be useful and we will investigate RMS and EOF analyses for the revised manuscript.
  The white color in figure 3b may not actually show regions of no relationship but rather region with less than 6 in situ profiles. Here it is important to use a mask for regions where the Pearson correlation coefficient was not calculated.
  
  Grid squares with fewer than 6 profiles are clear, but appear white due to the white background. The colour of no relationship, or ‘0’ correlation is grey and is visible when plotted on a white background, however we acknowledge that this is not clear, therefore we will include a mask in the revised version.
  Similarly, in 3c the color bar axis should have a minimum value of 6 and a mask applied to regions with less that 6 profiles. While these are rather small correction, the accurate display of data is important.
  
  We are not sure that this will improve visualisation, but we will test this and consider it for the revised manuscript.
  Finally, what is the value of figure 4d and 4e. A reader may ask why you show a region with anomalous in situ observations. Also, for 4d there appear to be data that in water depth less than 500 m.
  
  We think the reviewer might mean 3d and 3e. The purpose of 3e is to demonstrate the agreement between the satellite and in-situ data in a region where we have temporally abundant in-situ profiles. The correlations we derive on large scales (i.e. whole Southern Ocean, Figure 3b) vary in quality between grid-cells based on the temporal availability of in-situ profiles. For example, some grid squares have only a few profiles collected in summer, meaning that the correlation shown may not capture the year-round relationship. Figure 3d was intended to provide a visual aid to show the spatial distribution of the profiles over the Kerguelen Islands, and demonstrate good coverage of the region (i.e. all profiles are not just taken from a single point, so the data is more comparable to the satellite data which is the average over that grid square). We will filter out any data points which have been plotted in water depth less than 500m.
  Lines 176 to 180 need to be revised. How you validate the SHA is the sea-ice zone is important. What is meant by “ open Southern Ocean”? 60oS to the seasonal sea-ice zone?
  
  We will revise these lines to improve clarity of the extent of the zones we describe, referring to the sea ice zones labelled in Figure 1.
  While the interannual SHA variability shows a potentially interesting signal, it is somewhat complicated, and the analysis undertaken needs further investigation and refinement. This more detailed consideration of the causes of the variability would enable the authors to reach much more conclusive findings.
  
  I would suggest that the authors consider only the SHA and the relationship with the large-scale SAM and ENSO forcing for this current manuscript and continue to investigate the interannual and seasonal signals for inclusion in another manuscript.
  
  Thank you for the suggestion. During the restructuring we will consider this suggestion and consider making the manuscript more methods-focussed, using the SAM and ENSO comparisons, and possibly also the seasonal signals, as further validation.
  
  Citation: https://doi.org/10.5194/egusphere-2023-3050-AC2

Jennifer Cocks, Alessandro Silvano, Alice Marzocchi, Oana Dragomir, Noémie Schifano, Anna E. Hogg, and Alberto C. Naveira Garabato

Supplement

https://doi.org/10.5194/egusphere-2023-3050-supplement

Jennifer Cocks, Alessandro Silvano, Alice Marzocchi, Oana Dragomir, Noémie Schifano, Anna E. Hogg, and Alberto C. Naveira Garabato

Viewed

Total article views: 582 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
366	158	58	582	80	25	25

HTML: 366
PDF: 158
XML: 58
Total: 582
Supplement: 80
BibTeX: 25
EndNote: 25

Views and downloads (calculated since 22 Dec 2023)

Month	HTML	PDF	XML	Total
Dec 2023	69	23	9	101
Jan 2024	39	16	5	60
Feb 2024	46	15	4	65
Mar 2024	19	18	1	38
Apr 2024	50	21	5	76
May 2024	15	10	0	25
Jun 2024	54	14	9	77
Jul 2024	24	14	10	48
Aug 2024	13	3	9	25
Sep 2024	12	7	4	23
Oct 2024	17	9	2	28
Nov 2024	8	8	0	16

Cumulative views and downloads (calculated since 22 Dec 2023)

Month	HTML	PDF	XML	Total
Dec 2023	69	23	9	101
Jan 2024	39	16	5	60
Feb 2024	46	15	4	65
Mar 2024	19	18	1	38
Apr 2024	50	21	5	76
May 2024	15	10	0	25
Jun 2024	54	14	9	77
Jul 2024	24	14	10	48
Aug 2024	13	3	9	25
Sep 2024	12	7	4	23
Oct 2024	17	9	2	28
Nov 2024	8	8	0	16

Viewed (geographical distribution)

Total article views: 585 (including HTML, PDF, and XML) Thereof 585 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 19 Nov 2024

Short summary

Heat and freshwater fluxes in the Southern Ocean mediate global ocean circulation and abyssal ventilation. These fluxes manifest as changes in steric height: sea level anomalies from changes in ocean density. We compute the steric height anomaly of the Southern Ocean using satellite data and validate it against in-situ observations. We analyse interannual patterns, drawing links to climate variability, and discuss the effectiveness of the method, highlighting issues and suggesting improvements.


Total:	0
HTML:	0
PDF:	0
XML:	0