Next Article in Journal
The Impact of SMOS Soil Moisture Data Assimilation within the Operational Global Flood Awareness System (GloFAS)
Next Article in Special Issue
Assessment of AMSR2 Ice Extent and Ice Edge in the Arctic Using IMS
Previous Article in Journal
HTD-Net: A Deep Convolutional Neural Network for Target Detection in Hyperspectral Imagery
Previous Article in Special Issue
Evolution of Backscattering Coefficients of Drifting Multi-Year Sea Ice during End of Melting and Onset of Freeze-up in the Western Beaufort Sea
 
 
Article
Peer-Review Record

Ice Production in Ross Ice Shelf Polynyas during 2017–2018 from Sentinel–1 SAR Images

Remote Sens. 2020, 12(9), 1484; https://doi.org/10.3390/rs12091484
by Liyun Dai 1,2, Hongjie Xie 2,3,*, Stephen F. Ackley 2,3 and Alberto M. Mestas-Nuñez 2,3
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Remote Sens. 2020, 12(9), 1484; https://doi.org/10.3390/rs12091484
Submission received: 21 April 2020 / Revised: 30 April 2020 / Accepted: 5 May 2020 / Published: 7 May 2020
(This article belongs to the Special Issue Remote Sensing in Sea Ice)

Round 1

Reviewer 1 Report

All of my prior comments have been addressed to my satisfaction in this revised version of the manuscript. I have only two minor comments as listed below, but can recommend the paper for publication in Remote Sensing as is.

 

L 73: “Due to the narrow swath width of SMMR, the real repeat time of SMMR was longer than that of SSM/I, SSMI/S and AMSR-E.”

Surely, the longer repeat time of SMMR has more to do with the sensor being operated only every second day due to power constraints of the satellite than the swath width.

L 581 (reference 49): Make sure to edit reference to include subset dates and data access date.

Author Response

All of my prior comments have been addressed to my satisfaction in this revised version of the manuscript. I have only two minor comments as listed below, but can recommend the paper for publication in Remote Sensing as is.

 Re: Thank you very much for your affirmation to our work. We did revisions according to your comments.

L 73: “Due to the narrow swath width of SMMR, the real repeat time of SMMR was longer than that of SSM/I, SSMI/S and AMSR-E.”

Surely, the longer repeat time of SMMR has more to do with the sensor being operated only every second day due to power constraints of the satellite than the swath width.

Re: Thanks for explanation.

“Due to the narrow swath width of SMMR, the real repeat time of SMMR was longer than that of SSM/I, SSMI/S and AMSR-E.”

was revised to

“Due to power constraints, the SMMR was operated every second day; thus, the real repeat time of SMMR was longer than that of SSM/I, SSMI/S and AMSR-E.”

L 581 (reference 49): Make sure to edit reference to include subset dates and data access date.

Re: Thanks for your remind.

Meier, W. N., T. Markus, and J. C. Comiso. 2018. AMSR-E/AMSR2 Unified L3 Daily 12.5 km Brightness Temperatures, Sea Ice Concentration, Motion & Snow Depth Polar Grids, Version 1. [indicate subset used]. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center. doi: https://doi.org/10.5067/RA1MIJOYPK3P. [Data accessed].

was reviesed to

Meier, W. N., T. Markus, and J. C. Comiso. 2018. AMSR-E/AMSR2 Unified L3 Daily 12.5 km Brightness Temperatures, Sea Ice Concentration, Motion & Snow Depth Polar Grids, Version 1. [Brightness temperatures at 36 GHz and 85 GHz from 1 January 2017 to 31 March 2019]. Boulder, Colorado USA. NASA National Snow and Ice Data Center Distributed Active Archive Center. doi: https://doi.org/10.5067/RA1MIJOYPK3P. [accessed on 5 August 2019].

 

Reviewer 2 Report

Dear editor,

The purpose of this study is to estimate ice production in the RISP and MSP using a combination of Sentinel-1 SAR and AMSR2 data from 2017 to 2018. The Sentinel-1 SAR data are used to identify polynya ice area and the AMSR2 data are used to derive thin ice thickness based on an algorithm previously presented by Nihashi et al in 2017b. This paper is an improved version of the one previously submitted by the same authors for publication on MDPI Remote Sensing which needed a major revision.

In general, I found that the authors improved the presentation of their study accepting all the provided suggestions.

The introduction is much more complete now, thus providing a complete picture of the state of the art procedures used for monitoring Ross Sea polynyas. They also clarified methods section including several details about the applied methodologies and their limits (e.g., the time consuming issue, the need of interpretation by experienced scientists, etc.). Figures have been improved too. So results provide interesting and reliable insights (and a valid discussion) about one of the most interesting sites of the Southern Ocean.

Thus I suggest publication after that the following very minor typos have been resolved:

Line 266: Check Italics for “and”.

Line 266: Check font and size.

Line 406-407: Check this sentence.

Best regards

Author Response

In general, I found that the authors improved the presentation of their study accepting all the provided suggestions.

The introduction is much more complete now, thus providing a complete picture of the state of the art procedures used for monitoring Ross Sea polynyas. They also clarified methods section including several details about the applied methodologies and their limits (e.g., the time consuming issue, the need of interpretation by experienced scientists, etc.). Figures have been improved too. So results provide interesting and reliable insights (and a valid discussion) about one of the most interesting sites of the Southern Ocean.

Thus I suggest publication after that the following very minor typos have been resolved:

Re: Thank you very much for your affirmation to our work. We did revisions according to your comments.

Line 266: Check Italics for “and”.

Re: it was corrected.

Line 266: Check font and size.

Re: it was checked and corrected.

Line 406-407: Check this sentence.

Re:“Due to the manually digitizing, the data sets are not easy to be reproduced or automation.”

was revised to

“Due to the manually digitizing, the data sets are not easy to be reproduced or automatically extracted.”

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.


Round 1

Reviewer 1 Report

This work estimates polynya area and sea ice production for two polynyas in the Ross Sea region using Sentinal-1 SAR imagery for two years. Overall the results are clearly presented and the paper is well written. The authors compare and contrast the results of their sea ice production and area measurements with those in the existing literature nicely. I have a few comments and suggestions for minor improvements to the paper, but overall, the topic is very appropriate for publication in Remote Sensing. I recommend publication after minor revisions.

Minor Comments

  1. From my understanding of the methods as discussed in Section 3.1, the Sentinal-1 images are processed manually to identify the extent of polynya area. If so, I suggest the authors more clearly state that this is the case and note any limitations of this data set regarding reproducibility. If the polynya areas are not extracted from the SAR images using an automated process, the resulting areas would not be easily reproducible. Would the data (analyzed polynya images) be made available to the public or on request?
  2. Providing a link (L145 in text) to a general search engine for NASA data is insufficient as a citation for the AMSR2 brightness temperature data. Per the terms of data use, the authors are required to cite the data set authors and provide the DOI for the specific data set used. See the data set documentation at NSIDC (https://nsidc.org/data/AU_SI12/versions/1) where the appropriate citation is provided.
  3. In the discussion (L333), the authors state quite firmly that the ice production area of Tamura & Ohshima (2011) does not represent real ice production area. In light of the differences between data source and methodology for the two studies, this statement is rather extreme. The authors address the limitations and differences between their methodology and that of TO2011 quite nicely earlier in the discussion section and in the conclusions. Therefore, I think the statement in line 333 is unnecessary. The authors should consider removing this statement.

Technical Comments

L47: “continental Antarctic” should be rephrased to make grammatical sense (e.g., continental Antarctic or continent Antarctica)

L51: define RSP; I assume it is Ross Sea Polynya

L57: The phrase regarding clouds should be removed from this sentence and mentioned separately. The ability to measure PMW radiation through cloud cover is independent of the ability to measure without sunlight, which is the topic of the sentence.

L158: typo – completed should be completely

L190-191: The grammar in this sentence is off. I suggest deleting “much”, changing “wind-derive” to wind-driven, and changing “occurred” to occurring.

L210-211: the “H” notation is missing in the denominator (numerator) for the polarization ratios in equation 1 (equation 2).

L322-325: New results are being presented here in the discussion section. These results should be moved to the results.

Table 1: SMMR and SSMI/S data sources are explained or introduced anywhere in the text. I suggest adding a short note about the differences between these and the SSMIs around L66 where you discuss improvements to temporal resolution of PMW data.

Figure 5a, bottom panel (sea ice production plot): What is the value for the ice production in early September 2017? The bar appears to exceed the plot scale. Please annotate the plot similar to the event ~5 July 2017.

Reviewer 2 Report

Review of

Ice Production in Ross Ice Shelf Polynyas during 2017-2018 from Sentinel-1 SAR Images

by Liyun Dai, Hongjie Xie, Stephen Ackley, Alberto Mestas

 

Summary

The authors present a new study with the goal to derive ice production in the Ross Ice Shelf polynyas using Sentinel-1 SAR images in combination with thin-ice thickness from AMSR2 data. Here, the latter represents an existing and already published approach. The novelty in the presented study lies in the use of high-resolution SAR data to identify polynya events and to delineate the edge of the polynya with higher accuracy than with passive microwave data: “The Sentinel-1 Synthetic Aperture Radar dataset with high spatial and temporal resolution provides an unprecedented opportunity to more accurately identify both polynya area and occurrence.”

 

Overall comment and decision

The results and conclusions are not convincing me. All measures and deductions appear very subjective and the paper lacks meaningful comparisons in several instances. The paper uses unusual terms (e.g. “polynya ice area”) without providing appropriate explanations of what they actually mean. It is stated that SAR allows for a more accurate detection of the polynya area, but no meaningful comparison between AMSR2 and SAR polynya area or even an independent additional data set (e.g. VIS data) is presented. The complexities and uncertainties in using SAR backscatter for objective segmentations are not discussed. No information is provided on how the SAR polynya time series was derived (manually, automatically).

In the Discussion section, it is stated that “A new method to calculate ice production in polynyas is presented in this study”. I disagree. The paper in its current from just shows some case studies of SAR backscatter, where polynyas are delineated manually, their temporal evolution is subjectively followed and associated AMSR2 ice thickness are shown. The reader remains with a lot of questions about how the presented final results actual were computed.

I suggest to reject the paper. The authors might want to consider re-defining their research objectives and more focus on quantitative differences between ASMR2 and Sentinel-1 in terms of polynya area, total polynya frequencies and associated consequences for ice production retrievals. I will try to give some detailed reasons and justification for my decision in the following….

 

Detailed comments

 

L33: “temperature”   … of what?;    “surface layer” …. of what?

L37: impacts ;  what is “bio-related”??

L52: Tab.1 : It does not show a large interannual variability, it shows different findings of different studies.

L64: Reference [23]: There are more general papers dealing with TIR ice production retrieval

L68: “identify sea ice from others” ?? What is meant?

L74: “became reality” … strange expression

L76: The term “polynya ice occurrence” is very unusal and needs some explanation: What is the difference to “polynya area”?

L88: Again, “polynya ice area” = polynya area? The used term implies that open water is not present in a polynya or cannot be derived using SAR

L90: “Wind-driven polynya ice” … what’s this?

L106-L119: Paragraph too extensive

L134-L135: The reason for the difference in available images should be explained

L149-L150: Again, what is meant “polynya ice”? Is it the polynya area, or the polynya area minus open water area?

L157: “…we try to determine…”. This is key! How? This definitely needs some more explanation.

L169: “but the polynya and ice production event was well captured.” … How were the gaps filled? Which polynya area was used when SAR data were not available? Crucial information is missing here….

L170: “could clearly be seen” … how are these distinguished from one another? This is a very subjective approach. How does this hold for the entire period processing?

The reader remains clueless about how the polynya area from SAR data is actually derived. Manually, scene by scene, or in some kind of automated approach??

LL186ff: This a paragraph describing some differences in backscatter texture between “wind-wave” (what is meant?) and “wind-driven sea ice”. But it is completely unclear, how this becomes relevant for the polynya area retrieval that is presented. Is “wind-wave” the open water area?

L192: “ice concentration”… Sea-ice concentration is a bad measure in thin-ice areas. See Kwok et al. (2007). “Ross Sea polynyas: Response of ice concentration retrievals to large areas of thin ice”

L193-194: This is all extremely subjective measures and consequences for the derived data are not discussed.

In several places, the authors write about “katabatic wind events”, but no wind data is actually shown. Is it really only katabatic wind causing the polynya to open up?

L232: “The results show that…. “ How can this conclusion be drawn? This is not evidenced by what is shown here.

L249-L251: It is not clear why this is done in the described way. I do not understand the presented reason.

Figure 6: The SAR image in the background is the same for all presented days. So I wonder what is the benefit of the SAR data that this paper is about?

L282: “A new method to calculate ice production in polynyas is presented in this study”. I disagree.

L295: “almost every….” Why only almost? What is the consequence to the final data? What measures were taken to fill the gaps?

L297ff: “In this study, the polynya ice area is only recorded once and the maximum ice thickness from the PMW for the day before the wind-driven event, is used for the ice volume estimation.” It is not understood why this is done and how this effects the final data.

L305: “On the other hand…” Very confusing and not supported by anything that is shown in the paper.

L312: “estimates have large uncertainties”. How about the uncertainties of the data presented here?

L343-L344: In my understanding the TO2011 method DOES identify individual polynya events!

Reviewer 3 Report

The purpose of this study is to estimate ice production in the RISP and MSP using a combination of Sentinel-1 SAR and AMSR2 data from 2017 to 2018. The Sentinel-1 SAR data are used to identify polynya ice area and the AMSR2 data are used to derive thin ice thickness based on an algorithm previously presented by Nihashi et al in 2017b.

Although the topics are promising and the abstract is exciting, in my opinion the paper has several unclear or incomplete reasoning and looks a bit hurried. In general, I found that the present manuscript lacks of care in the presentation of both the background, the methods and the discussion of the achieved scientific results (interesting but sometimes superficially enumerated). Mostly, the authors define their study as a “new method” for retrieving ice production in the study area. To me, it looks much more as a combination of several existing methodologies for providing interesting insights in one of the most interesting sites of the Southern Ocean (that is what I argued from their nice abstract). This aspect make me wonder if they are aware of previous studies on the same topics, e.g. the rich literature of polynya related studies provided by the Italian polar ocean community that is active since late 80s in the Terra Nova Bay coastal polynya (or the Germans in the Weddell Sea and the Arctic polynyas). Then, I cannot understand if this “method” allow (or could allow) scientists to obtain routinely automated ice production information or it strongly depends on human interpretation. Finally, necessary care for formal writing and quality of presentations is missing in some sections; it looks as the manuscript was written by different authors in different sections. I suggest to improve this aspect too. Nonetheless, conclusions could be useful for polar scientists (if properly supported by methodology) and I believe that this study could represent an interesting contribution to the study of the Ross Sea polynyas dynamics and ice production.

So the authors need to clarify several passages and eventually improve the manuscript to meet the high quality standards required for a publication on this journal. I sincerely think that it needs a deep major revision before being suitable for publication on MDPI Remote Sensing.

Line by line specific issues are reported below:

Line 31: The introduction is really incomplete. Citations miss an entire branch of 35 years of polynya related research studies made by Italian polar community, whose Antarctic base actually overlook Terra Nova Bay and its polynya and include a Marine Observatory in the Ross Sea (MORSea). I strongly suggest the authors to improve the state of the art description. Although they focus on a different study area, some analysis/methodologies are similar/comparable or can be improved with previous studies which are not mentioned in this version but could support the authors (including the use of SAR and MODIS images for retrieving coastal polynya areas, their modeling, the role of katabatic winds, the ice thickness estimation from PMW, the ice production estimation from heat fluxes, etc.). A short list of main papers that could interest them follows:

  • Aulicino, G.; Sansiviero, M.; Paul, S.; Cesarano, C.; Fusco, G.;Wadhams, P.; Budillon, G. A New Approach for Monitoring the Terra Nova Bay Polynya through MODIS Ice Surface Temperature Imagery and Its Validation during 2010 and 2011 Winter Seasons. Remote Sens. 2018, 10, 366.

 

  • Sansiviero, M.; Morales Maqueda, M.Á.; Fusco, G.; Aulicino, G.; Flocco, D.; Budillon, G. Modelling Sea ice formation in the Terra Nova Bay polynya. Mar. Syst. 2017, 166, 4–25.

 

  • Parmiggiani, F. Fluctuations of Terra Nova Bay polynya as observed by active (ASAR) and passive (AMSR-E) microwave radiometers. Int. J. Remote Sens. 2006, 27, 2459–2467.

 

  • Parmiggiani, F. Multi-year measurement of Terra Nova Bay winter polynya extents. Eur. Phys. J. Plus 2011, 126.

 

  • Fusco, G.; Budillon, G.; Spezie, G. Surface heat fluxes and thermohaline variability in the Ross Sea and in Terra Nova Bay polynya. Shelf Res. 2009, 29, 1887–1895.

 

  • Fusco, G.; Flocco, D.; Budillon, G.; Spezie, G.; Zambianchi, E. Dynamics and variability of Terra Nova Bay polynya. Mar. Ecol. 2002, 23, 201–209.

 

  • Rusciano, E.; Budillon, G.; Fusco, G.; Spezie, G. Evidence of atmosphere-sea ice ocean coupling in the Terra Nova Bay polynya (Ross Sea–Antarctica). Cont. Shelf Res. 2013, 61–62, 112–124.

 

  • Morelli, S.; Parmiggiani, F. Wind over Terra Nova Bay (Antarctica) during a polynya event: Eta model simulations and satellite microwave observations. Eur. Phys. J. Plus 2013, 128, 135.

 

Lines 35, 38, 40: Check typos, square brackets.

 

Lines 50-51: Please rephrase to clarify the text between commas. I cannot find where RSP is defined.

 

Lines 53-54: Please rephrase this sentence. Maybe better “Table 1 show…” or similar.

 

Lines 59-60: What about 80s? And SMMR? You mentioned in Table 1. Could you explain if/why it is not safe to go back such in the past?

 

Lines 63-64: See also Aulicino et al., 2018 and Paul et al. 2015a and 2015b about the use of MODIS images and the cloud cover limits and corrections.

 

  • Paul, S.; Willmes, S.; Gutjahr, O.; Preußer, A.; Heinemann, G. Spatial feature reconstruction of cloud-covered

areas in daily MODIS composites. Remote Sens. 2015a, 7, 5042–5056.

 

  • Paul, S.; Willmes, S.; Heinemann, G. Long-term coastal-polynya dynamics in the southern Weddell Sea from MODIS thermal-infrared imagery. Cryosphere 2015b, 9, 2027–2041.

 

 

Lines 68-72: Improve references.

Lines 83-85: A few words about the limits of PMW use would be appreciated.

Line 92: I suggest to change the expression “simple” for conclusions.

Line 97: Check typos.

Figure 1: Since you mentioned it, please enlarge the image in order to include the entire Terra Nova Bay (its northern limits are not visible). Red labels cannot be clearly read. Please improve.

Lines 106-108: Improve this sentence.

Line 126: Check typos.

Line 132: Check typos.

Lines 134-135: The temporal resolution does not seem excellent to estimate annual averages. Polynya opening/closing can vary a lot during a single day, much more in 2-3 days. How can you be sure that you are not under estimating the actual processes with such a sub-sample (especially in 2017)?

Lines 137-145: Improve this paragraph.

Lines 154-159: Again, how can you be sure that your average estimates are representative of the actual average (if we only know information each other day at best)? In my opinion, there is a large difference between capturing a polynya event (that can be at initial, medium, or late stage of ice production) and assuming that its extent is a proper value for producing average estimates with a statistical significance. I suggest to include some additional analysis to evaluate the accuracy of your results (especially for 2017 when a limited number of days seem to be covered by satellite information).

Lines 175-177: See previous comment.

Figure 2: Ice thickness cannot be read completely. In my opinion, figure 2f presents an example, it does not “explain”.

Line 187: I cannot see this point. “Identified”: do you mean "distinguished"?

Lines 188-191: I cannot understand completely these sentences. Please, check and eventually improve to help the reader understand your points. Then, in my opinion, a major difference is also represented by the Langmuir streaks that are clearly visible (e.g., in SAR imagery) when katabatic winds push newly formed sea ice offshore. I suggest to mention this difference.

Line 213: Check Italics for “and”.

Lines 220-222: Reference is missing here (you are providing a value obtained from visual observations, a report/paper should be addressed for the readers who were not involved in PIPERS 2017 cruise).

Lines 224-226: You should clarify or address additional details to the following section where you actually try to explain this point (i.e., lines 249-251). Moreover, those assumptions are not supported by a complete reasoning or by references to previous studies. I wonder if you can really misinterpret polynya extent due to its partial closing before the “following” katabatic event. Furthermore, it is still not clear if there is an automated procedure or a statistical criterium for safely distinguishing following events. I cannot find anything about this in the methodology. Please improve and clarify, this is a main point to strengthen the reliability of your results.

Line 228: Results - In general, the results presentation should be improved. Then, the reliability of the presented values strongly depends on the clarification of the methodology issues reported above.

Lines 239-240: How can you argue it, if your temporal coverage is such different between the two years?

Figure 5: Please improve or delete the text in the brackets.

Figure 6: Check singular/plural (left panel/panels?).

Lines 264-266: Improve this important sentence.

Lines 272-274: If available, the authors could include in this study in situ wind observations collected by automatic weather stations to validate their statement. As it is, it looks as you feel safe to retrieve a katabatic wind event from its effects on the sea ice opening and pushing offshore. Except existing literature (however, not mentioned here), how can you be sure that “all” your events are due to katabatic forcing? (I'm sure it is, but you have not argued this point).

Line 282: I don't think this is completely correct. I would say that you applied several existing methodologies to your case study. Of course, it would be a new method if you clearly describe details of this procedure that could be implemented by other scientist for an automated routinely estimation of ice production that go further the present state of the art. In general, your study is absolutely worthy since the area is of huge interest... but the study should be presented and published for what it is (A study about the 2017 and 2018 evolution of the ice production in the study area, since I cannot see any “new methodology” here). The estimation of ice production from multiplying polynya area per its thickness have been implemented in several studies (e.g., see the attached references).

Lines 301-302: From this discussion , it looks as TO2011 is the only existing study about this topic. Please improve.

Lines 302-309: Again, does this imply a human check or can be managed through an automated procedure (eventually how)? Details about this would be appreciated by Remote sensing readers.

Line 338: I don't think that this is a valid conclusion as it is presented. You retrieved 51 polynya opening events in 2017 (presumably due to katabatic wind events), not necessarily 51 katabatic events (except you prove this point). You could provide much more about katabatic winds comparing your RISP and MSP results to independent wind observations (from automatic weather stations, weather models, or inferred from SAR imagery).

Lines 341-342: I still cannot understand if there is an automated procedure and/or a statistical analysis that support these statements. I am still convinced that this study combines existing methods for providing interesting information about RISP and MSP (in this version, it does not represent a new method), and that additional analysis are requested to estimate the accuracy of your results about individual event identification and sea ice production computation.

Lines 345-346: I would be more careful about this statement.

 

I would appreciate to take part to the next steps of the manuscript revisions.

Best regards.

Back to TopTop