Satellite-based analysis of ocean-surface stress across the ice-free and ice-covered polar oceans

Liu, Chao; Yu, Lisan

doi:https://doi.org/10.5194/essd-17-4159-2025

Articles | Volume 17, issue 8

https://doi.org/10.5194/essd-17-4159-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/essd-17-4159-2025

© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 17, issue 8

Data description paper

|

27 Aug 2025

Data description paper |

| 27 Aug 2025

Satellite-based analysis of ocean-surface stress across the ice-free and ice-covered polar oceans

Chao Liu and Lisan Yu

Publisher's note: the correspondence author has been changed post-publication because the former corresponding author was a postdoc on the project and left the project. He is no longer involved in this ongoing project.

Download

Final revised paper (published on 27 Aug 2025)
Preprint (discussion started on 06 Feb 2025)

Interactive discussion

Status: closed

RC1:
'Comment on essd-2025-14', Anonymous Referee #1, 19 Apr 2025
Revised Review of Satellite-based Analysis of Ocean-Surface Stress across the Ice-free and Ice-covered Polar Oceans by Liu and Yu (2025)
The authors present an analysis of satellite-derived ocean-surface stress across the ice-free and ice-covered polar oceans. They utilize a bulk parameterization approach, combining multiple satellite datasets, to estimate ocean-surface stress and Ekman transport. The study provides valuable insights into the spatial and temporal variability of these key drivers of polar climate dynamics. However, I have several concerns regarding the methodology, data processing, and presentation, which must be adequately addressed before the manuscript is suitable for publication.
General Comments:
The authors need to update their references to include some relevant papers especially when it relates to relevant dataset (CPOM extended dataset was recently used in Lin et al (2023)), parameterisations (Brenner et al 2021, Lupkes et al 2015), modelling work that can server as motivation (Tsamados et al 2014, Sterlin et al 2023).

The validation with ITP observations is a valuable component, but its limitations need to be discussed in more detail. In particular more discussion is needed and this section needs rewriting to better account for the impact of time and space sampling in the input datasets on the derived ocean surface stress.

The authors perform a sensitivity analysis to value of drag and gaussian spatial filters but do not discuss errors due to model parameterisation (i.e. impact of ue vs ug vs nothing), impact of uncertainty of reanalysis, impact of ocean (and atmospheric) stratification (see papers by Lupkes and Brenner for example), impact of time and length scale over which stresses are calculated.

The authors could consider extending the analysis period to include more recent available data (CPOM or CLS) and by removing their current constraint due to the reanalysis they use. Please justify why this reanalysis is really so good to justify loosing 5+ years of analysis.

The authors use open access satellite products but fail to share their code! Share code / Github

References:

Lin, Peigen, Robert S. Pickart, Harry Heorton, Michel Tsamados, Motoyo Itoh, and Takashi Kikuchi. "Recent state transition of the Arctic Ocean’s Beaufort Gyre." Nature Geoscience 16, no. 6 (2023): 485-491.
Brenner, Samuel, Luc Rainville, Jim Thomson, Sylvia Cole, and Craig Lee. "Comparing observations and parameterizations of ice‐ocean drag through an annual cycle across the Beaufort Sea." Journal of Geophysical Research: Oceans 126, no. 4 (2021): e2020JC016977.
Lüpkes, Christof, and Vladimir M. Gryanik. "A stability‐dependent parametrization of transfer coefficients for momentum and heat over polar sea ice to be used in climate models." Journal of Geophysical Research: Atmospheres 120, no. 2 (2015): 552-581.
Tsamados, Michel, Daniel L. Feltham, David Schroeder, Daniela Flocco, Sinead L. Farrell, Nathan Kurtz, Seymour W. Laxon, and Sheldon Bacon. "Impact of variable atmospheric and oceanic form drag on simulations of Arctic sea ice." Journal of Physical Oceanography 44, no. 5 (2014): 1329-1353.
Sterlin, Jean, Michel Tsamados, Thierry Fichefet, François Massonnet, and Gaia Barbic. "Effects of sea ice form drag on the polar oceans in the NEMO-LIM3 global ocean–sea ice model." Ocean Modelling 184 (2023): 102227.
Specific Comments:
L6 EASE2? clarify if so

L15 clarify why you need both CPOM and CLS?

L19 clarify / correlation of what

L21 not showed in this paper?

L40 ice-ocean governor is Menenghello citation?

L47 add references here i.e Lin et al 2022 but also more references in introduction for motivation / modelling work

L51 Seems a waste to limit yourselves to 2018 due to reanalysis

L77 cite recent papers (including modelling)

L92 discuss impact on ocean surface stress of time sampling i.e. daily vs 3d vs monthly

L92 discuss why this reanalysis so important

L117 add references to paper Lin that used more extended CPOM extension reference

L126 is it worth limiting to that reanalisis then?

Figure 2 why gaussian filtering to reconcile with CPOM?

L144 why not just use CLS if better? You criticise CPOM product but you fail to discuss all issues in these CLS and CPOM product when it comes to continuity of SSH at the ice-ocean boundary.

L186 clarify

L190 not so clear on the map. Shouldn’t the maps be monthly?

Figure 4 mean is noisy so what about monthly / daily products. Is that mean over all winter and all months?

Figure 5 what is the total region used? Not clear what is aaw

L277 cite recent papers and modelling ones

L312 what about error from time sampling, model physics parameterization

Figure 9 is based on two errors only not sure I buy this. Also not clear how the range of tau_w chose 10^-3 to 10^-2.

L340 test impact on derived stress

L366 shouldn't you compare at the time averageds used in stress calculations? i.e. daily

Figure 11 is u really daily?

Figure 12 representation issue here clearly seen (space or time or both?)

Figure 14 obvious representativity issue (space, time or both?). I would redo all this section on monthly products and discuss better

Figure 15 what about plot for daily timescales

L433 share code!
Citation: https://doi.org/10.5194/essd-2025-14-RC1
- AC1: 'Reply on RC1', Chao Liu, 17 May 2025
  
  We thank the reviewer for their time and for providing positive feedback and constructive comments on our manuscript. We have carefully revised the manuscript accordingly and incorporated all the proposed changes. The attached PDF file provides a detailed response addressing each of the reviewer’s concerns.
  
  Citation: https://doi.org/10.5194/essd-2025-14-AC1
RC2:
'Comment on essd-2025-14', Anonymous Referee #2, 21 Apr 2025

This is a well written manuscript and the analysis seems legitimate and thoughtful. The first thing that came to me after reading the manuscript is why the new dataset of ocean surface stress is needed? Since there are many assumptions have been made in order to calculate the ocean-surface stress. It will be great if the authors can explain the motivation and reasoning that this data is needed for the community. Second, the value of Ekman depth used in the calculation is not explained, eqn. 4 & 5. The Ekman depth is needed in order to calculate the vertical Ekman velocity. What are the values of Ekman depth that you used? Third, for the scatterplots of ITP vs satellite (Fig. 12 & 13), how the ITP data is processed to be comparable to the daily satellite product? Did you do time average on ITP data to daily data or interpolate the satellite daily product to the ITP profile time? From the number of samples, it seems the author interpolated the daily satellite data to the ITP data (profile) time. If so, the scatter plots are meaningless. The comparison is wrong, since the satellite temporal resolution is under sampled. You can not draw conclusions from these scatterplots. Also, the 30-day low pass resulted in too few data points (Fig. 14, 15 & Table 3). I suggest to do a time-average ITP data to daily or weekly and , and redo the analysis.
Other small comments
Line 40: "ice-ocean governor," → "ice-ocean governor",
Line 60: Figure 1(a) caption: there is no dashed magenta has shown. Is the September ice extent boundaries out of the scope of the plot? The Solid line near the Novaya Zemlya has a portion of dashed line. Is this supposed to be solid?
Line 84 It is unclear to me how the De: Ekman layer depth is determined in the calculation. Can you explain?
Acronym stands for?
Line 98 acronym OAFLux2?
Line 106 acronym SSM/I, AMSR-E, AVHRR, MODIS, QuikSCAT?
Line 107 acronym IABP?
Line 111 acronym CLS/PML?
Line 120 acronym SMMR DMSP, NSIDC?
Line 139 "Noting the 25 km resolution could introduce uncertainties near the 15% sea ice concentration boundary." --> How did you arrive this conclusion?
Line 152 "where W represents the local vertical Ekman velocity W_e" → W and W_e were used interchangeably. Can you decide which one to use and keep it consistently?
Line 227 "in the Indian Ocean sector and the southeast Pacific" → Can you label it on Fig 6b?
Line 232 of 50°S (Figure 6d), → Can you label the latitude and longitude on Fig 6?
Line 293 Baffin Bay, the Chukchi Sea, and north of Fram Strait → Can you label these places in Figure 8?.
Line 309 What do you mean “ full eight-year/six-year period” period?
Lien 324: the Mendeleev Ridge. → please label it on the Figures
Line 327-328 Weddell and Ross Seas, Antarctic Peninsula and western Ross Sea, Enderby Land and the Amundsen Sea → please label these places on the Figures.
Line 381 What do you mean “paired” observations?
Line 385 “ITP-80 pair shows better agreement in the zonal component after 200 days.” → I can't tell if this is true from the figure 12. How did you arrive this conclusion?

Citation: https://doi.org/10.5194/essd-2025-14-RC2
- AC2: 'Reply on RC2', Chao Liu, 17 May 2025
  
  We thank the reviewer for their time and for providing positive feedback and constructive comments on our manuscript. We have carefully revised the manuscript accordingly and incorporated all the proposed changes. The attached PDF file provides a detailed response addressing each of the reviewer’s concerns.
  
  Citation: https://doi.org/10.5194/essd-2025-14-AC2

Peer review completion

AR: Author's response | RR: Referee report | ED: Editor decision | EF: Editorial file upload

AR by Chao Liu on behalf of the Authors (17 May 2025) Author's response Author's tracked changes Manuscript

ED: Reconsider after major revisions (23 May 2025) by Baptiste Vandecrux

Dear Dr. Chu & Dr. Liu,

Thank you for responding to the reviewers' comments.

I can see that reviewer one expressed the same frustration as I did concerning the limited temporal coverage of your dataset. I think this also contributes to reviewer two wondering about the usefulness of such a dataset. I also do not understand your justification, mentioning OAFlux2 as a limitation, since that dataset is available up to this day (according your Table 1).

If I understand correctly, you derive ocean surface stress from readily available datasets, which provides a faster way of calculating it compared to using computationally expensive ocean-atmosphere models. Of course, this comes at the cost of several, sometimes simplifying, assumptions. I do see value in your approach, provided that the coverage and release rate of your dataset are on par with current state-of-the-art datasets providing ocean surface stress. I therefore assumed (it's a bit outside my expertise) that your dataset’s temporal or spatial resolution, or the level of detail it could offer for process understanding, would justify the limited temporal coverage. But there is no description in the text, or example, of the possibilities your dataset offers that were not possible before with state-of-the-art datasets. Both reviewers also seemed to miss this proof of added value.

Additionally, if the point of your manuscript is to show that it is possible to derive ocean surface stress from an outdated or unmaintained dataset while waiting for an update to be released, then your study is a proof of concept, not a dataset that people will be willing to use. ESSD also advocates against the fragmentation of research: it is not acceptable to publish your analysis on the old data now and then publish the updated analysis later with the new input fields.

I strongly suggest you either extend your dataset to its maximal potential span or find an alternative source for the time-limited forcing datasets , so that your dataset covers more recent years. Alternatively, if you are in communication with your different input data providers (I see from your Table 1 that the wind data would not be the only limiting dataset) and an update is planned within the next year, then I would gladly give you the opportunity to update your manuscript after that future release.

I hope you understand my concerns.

Sincerely,
Baptiste Vandecrux

Hide

AR by Chao Liu on behalf of the Authors (30 May 2025) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (04 Jun 2025) by Baptiste Vandecrux

RR by Anonymous Referee #2 (13 Jun 2025)

ED: Publish as is (13 Jun 2025) by Baptiste Vandecrux

AR by Chao Liu on behalf of the Authors (14 Jun 2025) Manuscript

Download

Please read the editorial note first before accessing the article.

Article (13220 KB)
Full-text XML

Short summary

A daily dataset for ocean-surface stress is synthesized for both ice-covered and ice-free Arctic and Antarctic areas. It is based on satellite data on ocean winds, ice movement, and sea surface height. Sensitivity analyses address uncertainties, including variations in sea level products and ice–water drag. The dataset's accuracy is validated against in situ measurements, showing moderate to good agreement on monthly and longer timescales.