Validation and Application of Satellite-Derived Sea Surface Temperature Gradients in the Bering Strait and Bering Sea

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The Arctic is one of the most important and sensitive regions for climate change. Yet, it is also difficult to measure because of extreme weather, as well as ice conditions. In this manuscript, the authors directly compared four data sets from the Group for High Resolution Sea Surface Temperature (GHRSST) with a NASA Saildrone deployment along the Alaskan Coast and the Bering Sea and Bering Strait. The results are sound. The manuscript is clear and well written. I recommend minor revision for publication.

 

 

It is from Table 1 that both CMC and OSTIA have comparable performance. However, only OSTIA is highlighted in the abstract and conclusion.

There are two recent papers might be useful for discussion about northward shift in the southern ice edge after 2018.

 

Figure 19. The gap between 64.5-67.5 N is connected with straight lines, which may give readers a false impression. I recommend let this gap with blank or connect with dashed.

Line 48. Remove the duplicated right parenthesis.

 

 

Dai, A., Luo, D., Song, M. et al. Arctic amplification is caused by sea-ice loss under increasing CO₂. Nat Commun 10, 121 (2019). https://doi.org/10.1038/s41467-018-07954-9

Gaopeng Xu, M. Cameron Rencurrel, Ping Chang, Xiaoqing Liu, Gokhan Danabasoglu, Stephen G. Yeager, Michael Steele, Wilbert Weijer, Yuchen Li, Nan Rosenbloom, Frederic Castruccio and Qiuying Zhang. High-resolution modelling identifies the Bering Strait’s role in amplified Arctic warming. Nature Climate Change, 2024, 14(6):615, DOI: 10.1038/s41558-024-02008-z

Author Response

The Arctic is one of the most important and sensitive regions for climate change. Yet, it is also difficult to measure because of extreme weather, as well as ice conditions. In this manuscript, the authors directly compared four data sets from the Group for High Resolution Sea Surface Temperature (GHRSST) with a NASA Saildrone deployment along the Alaskan Coast and the Bering Sea and Bering Strait. The results are sound. The manuscript is clear and well written. I recommend minor revision for publication.

We thank the reviewer for the positive comments and the review. We responded to the individual comments below.

It is from Table 1 that both CMC and OSTIA have comparable performance. However, only OSTIA is highlighted in the abstract and conclusion.

Thank you for the comment.

We decided to use OSTIA because it did have the highest correlation. The goal was to simply use OSTIA as exemplary for the application. Certainly, as mentioned in the paper, the results should be considered as a starting point for future work, which needs to include further comparisons between SST products in the Arctic.

There are two recent papers might be useful for discussion about northward shift in the southern ice edge after 2018.

Figure 19. The gap between 64.5-67.5 N is connected with straight lines, which may give readers a false impression. I recommend let this gap with blank or connect with dashed.

Thank you for the recommendation. Instead of using a dashed line we have added the following statement to the Figure caption: A gap in the SIC concentration exists between 65°N to 68°N which is reflected by straight line.

Line 48. Remove the duplicated right parenthesis.

Done

Dai, A., Luo, D., Song, M. et al. Arctic amplification is caused by sea-ice loss under increasing CO₂. Nat Commun 10, 121 (2019). https://doi.org/10.1038/s41467-018-07954-9

Gaopeng Xu, M. Cameron Rencurrel, Ping Chang, Xiaoqing Liu, Gokhan Danabasoglu, Stephen G. Yeager, Michael Steele, Wilbert Weijer, Yuchen Li, Nan Rosenbloom, Frederic Castruccio and Qiuying Zhang. High-resolution modelling identifies the Bering Strait’s role in amplified Arctic warming. Nature Climate Change, 2024, 14(6):615, DOI: 10.1038/s41558-024-02008-z

Thank you for the recommendation. I have added the following to the discussion section after Figure 19. I also added references.

 

The northward shift of the southern ice edge can also be related to some critical issues related to changes in the Arctic.

 

[19] relates Arctic Amplification (AA) to sea ice loss. The results in this paper showing the northward movement of the southern ice edge indicate a relationship to the Arctic Amplification (AA). The (AA) is related to increasing greenhouse gases. Additionally, [19] states the enhanced greenhouse warming is seen north of 67°N, consistent with the results found in this work and the northern movement of the southern ice edge.

 

[20] using model simulations showed that oceanic heat transport (OHT) through the Bering Strait had a more significant impact on Arctic warming than previously thought. A key result was also that the increased OHT was dependent on the resolution of the model. This is motivation for future and further analysis of the different SST products in the Arctic. Thus, the results presented here provide a starting point for future applications of SST in Arctic.

 

 

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

Review on the manuscript

Validation and Application of Satellite Derived Sea Surface 2 Temperature Gradients in the Bering Strait and Bering Sea

 

This work compares four data sets, sea surface temperature (SST) from Group for High 16 Resolution Sea Surface Temperature (GHRSST), the Remote Sensing Systems Microwave Infrared Optimally Interpolated (MWIR) Product, Canadian Meteorological Center (CMC) product, the Daily Optimally Interpolated Product (DOISST), and the Operational Sea Surface Temperature and Ice Analysis (OSTIA) Product. All those datasets are well-known and frequently used for different purposes by oceanographers. I see in first time the attempt of direct comparison of these datasets and therefore, I think this paper deserves attention and eventually publications. However, my first objection is that I am slightly disappointed that this work is limited only by the consideration of Arctic zone nearby Bering Strait and Bering Sea. As I understand this is because the authors use the saildrones (special devices) as the original and true data. But the comparison (maybe with another scheme) can be done for entire globe. However, even for Arctic zone it is important and valuable work.

My second objection is focused on the methodology. To my mind it is presented very unclearly.  As I understand from the text, firstly the authors calculate the observed gradient of SST taken from the saildrones and only then recalculated these gradients into grid. I think it should be explained explicitly and more carefully. Without this explanation I cannot estimated any possible errors and range of these errors.

Finally, I would like to see the results not only by figs and tables but with confidence bounds, they should indicate which data are preferable even with comparison with saildrones which are taken as the true (we confidence them with the probability 1).  May be it would be better to introduce a prior distribution and the statistically test the fitting of data with this distribution. This is only the proposal, I do not insist to fulfil it, but to my mind it would be more strongly

Since I am not English native person, I cannot estimate the style and grammar. To my mind the paper is written (except methodology part) good enough, clearly. I recommend its publication after making the amendments I did.

Author Response

Validation and Application of Satellite Derived Sea Surface Temperature Gradients in the Bering Strait and Bering Sea

 

This work compares four data sets, sea surface temperature (SST) from Group for High 16 Resolution Sea Surface Temperature (GHRSST), the Remote Sensing Systems Microwave Infrared Optimally Interpolated (MWIR) Product, Canadian Meteorological Center (CMC) product, the Daily Optimally Interpolated Product (DOISST), and the Operational Sea Surface Temperature and Ice Analysis (OSTIA) Product. All those datasets are well-known and frequently used for different purposes by oceanographers. I see in first time the attempt of direct comparison of these datasets and therefore, I think this paper deserves attention and eventually publications. However, my first objection is that I am slightly disappointed that this work is limited only by the consideration of Arctic zone nearby Bering Strait and Bering Sea. As I understand this is because the authors use the saildrones (special devices) as the original and true data. But the comparison (maybe with another scheme) can be done for entire globe. However, even for Arctic zone it is important and valuable work.

Thank you for your encouraging words and comments. With respect to the decision to focus on the Bering Straits and Bering Sea off Alaska, it was exactly as you mentioned, the capability to do validation with Saildrone. Validation of gradients has occurred in other areas of the world’s ocean, such coastal regions, and the Mediterranean Sea. To our knowledge this is the first study of SST gradients in the Arctic with a scientific application. Thus, the validation with Saildrone is a critical part for confidence in the results. Comparisons with Saildrone have been done already for the California Coast and other areas of the world’s ocean. Thus, we made the decision to focus the study on the Bering Strait and Bering Sea.

My second objection is focused on the methodology. To my mind it is presented very unclearly. As I understand from the text, firstly the authors calculate the observed gradient of SST taken from the saildrones and only then recalculated these gradients into grid. I think it should be explained explicitly and more carefully. Without this explanation I cannot estimated any possible errors and range of these errors.

Thank you for your recommendations. They are very appreciated. We have made significant changes to the methods. We hope it makes things clearer. Additionally, we have added a workflow diagram that summarizes the steps and order in which the methodology was implemented. Please see below.

2.2. Methodology

Step one in validating satellite SST using Saildrone was to collocate the satellite derived SST products with the Saildrone deployment, using the following steps:

1. Smooth the Saildrone 1 minute sampling to the daily time scales of satellite data.
2. Derive daily SST gradients from the daily Saildrone smoothed product.
3. Derive SST gradients from the four satellite products based on the finite different approach.
4. Collocate satellite derived SST gradient to the daily smoothed SST gradients along the Saildrone deployment. The method used was a nearest neighbor approach where for a given day, the satellite derived SST pixel closest to the Saildrone daily average for that day was chosen.
5. Spatial gradients for all data sets were computed along the Saildrone track.
6. Linear fits were applied to the time series of the satellite derived SST gradient maps to examine possible trends.

 For step 1, the following equation was applied:

 

SST_sail (x,y)=                      (1)

 

where SST_sail (x,y) is the SST derived from the Saildrone at a longitude “x” and latitude  “y”  after smoothing over the daily time steps “N.” The index “i” is simply indicating the  time step along the Saildrone deployment track.

 

 For step 2, the gradients are then derived at derived at the daily smoothed locations along the Saildrone track.

 

The spatial gradients are computed such that:

 

SST_sailgradx (x_i,y_j) = [SST_sail (x_i+1,y_j)- SST_sail (x_i-1,y)]/Δx        (2)

SST_sailgrady (x_i,y_j) = [SST_sail (x_i,y_j+1)- SST_sail (x_i,y_j-1)]/Δy        (3)

 

where SST_sailgradx (x_i,y_j) and SST_sailgrady (x_i,y_j) are the x (longitude) and y (latitude) components of the gradient at the specified longitude and latitude along the Saildrone deployment track. The distances Δx and Δy are the distances in kilometers between the pixels at the specified longitude/latitude locations. The magnitude of the gradient at location (x_i,y_j) along the Saildrone deployment track can then be calculated  as:

 

SST_sailgrad (x_i,y_j)=        (4)

 

where SST_sailgrad (x_i,y_j) is the magnitude of Saildrone SST gradient at position (x_i,y_j) along the Saildrone track.

 

     

For step 3, gradients were derived for the satellite MWIR, CMC, DOISST, and OSTIA for each of the daily maps covering the period of the Saildrone deployment. The equation used was a simple finite difference approach such that:

 SST_gradx(x_i,y_j) = [SST(x_i+1,y_j) - SST(x_i-1,y_j)]/ Δx                   (5)

 SST_grady(x_i,y_j) = [SST(x_i,y_j+1) - SST(x_i,y_j-1)]/ Δy              (6)

 

   The magnitude is then defined as:

 

    SST_sailgrad (x_i,y_j)=     (7)

 

   Step 4 was then defined as the collocation of values from equation (6) with values from

                          Equation (4).

 

The workflow diagram below summaries the above steps in the methodology used for collocation the satellite derived SSTs from the four products with the Saildrone deployment.

 

               

 

Finally, I would like to see the results not only by figs and tables but with confidence bounds, they should indicate which data are preferable even with comparison with saildrones which are taken as the true (we confidence them with the probability 1). May be it would be better to introduce a prior distribution and the statistically test the fitting of data with this distribution. This is only the proposal, I do not insist to fulfil it, but to my mind it would be more strongly

Thank you for the recommendation. To better examine the statistical relationships between satellite products and Saildrone we have added scatter plots. To maintain the flow of the manuscript and the focus on statistics derived from the Saildrone comparison we added the scatter plots as an appendix. The following appendix was added to show elucidate more clearly the relationship between Saildrone and the four remote sensing products. A goal of the manuscript was not to conclude which product is best for the Arctic, but to motivate future work that address that address that question.

 

Appendix

 

Figure A1: Shows the scatter plots for a) Saildrone versus MWIR, b) Saildrone versus CMC, c) Saildrone versus OSTIA, and d) Saildrone versus DOISST.

The scatter plots show (statistics summarized in table in manuscript)  that OSTIA performs best at reproducing the magnitude of Saildrone gradients. The slope is closest to one at 1.01 with a bias of 0.  DOISST has the minimal slope (0.45), indicative most likely of the smoothness of the product. Extreme values in MWIR and CMC are associated with high latitudes > 68°N and most likely associated with issues of the ice mask. This needs to be the focus of future work. The correlation statistics, standard deviations, signal to noise rations, along with the slopes were a primary factor in deciding to apply the OSTIA product for application to understanding the possible movement in the southern ice edge.

 

Since I am not English native person, I cannot estimate the style and grammar. To my mind the paper is written (except methodology part) good enough, clearly. I recommend its publication after making the amendments I did.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Abstract: Authors should prefer to have more results presented in the Abstract section.

 

Introduction: Authors should rewrite the introduction section. It is almost the same as the first paragraph abstract. Additionally, the importance of the subject studied should be explained in the introduction section. Current study is directly mentioned here. In addition, past studies have not been adequately examined and the originality of the study has not been well expressed.

 

Methodology: The methodology section is not well planned and well explained. The authors briefly listed the process steps. I recommend you write this section in more detail and create a workflow.

 

Results: The Results section is well written and the results are presented in sufficient quantity.

 

Line 231: This is not a map. North arrow, scale and legend are missing. Also the resolution is not good.

 

The resolution is not sufficient in all figures. scales cannot be read.

 

Discussion: This chapter is well written, but the authors need more in-depth discussion.

 

Conclusion: Adequate and complete.

Author Response

Comments and Suggestions for Authors

Abstract: Authors should prefer to have more results presented in the Abstract section.

 

Thank you for the recommendation. The following sentences were added to the abstract:

Scatter plots presented in the appendix indicate that OSTIA had the slope closest to one and thus reproducing the magnitudes of the Saildrone gradients. Both OSTIA and CMC had the highest correlations of 0.79 and 0.81 respectively.

 

This indicates that the shift of the Southern ice edge is not happening gradually but has dramatically increase over the last decade.

 

 

Introduction: Authors should rewrite the introduction section. It is almost the same as the first paragraph abstract. Additionally, the importance of the subject studied should be explained in the introduction section. Current study is directly mentioned here. In addition, past studies have not been adequately examined and the originality of the study has not been well expressed.

 

Thank you for the comments. The following paragraph was added to the Introduction. The authors feel they have covered the work focused in the Arctic on gradients. The primary work (which is cited) was by Castro et al.. That is described in the Discussion section.

 

“The focus on gradients will be two-fold, examination of both SST and Sea Ice gradients. The rationale for this can be seen in multiple ways. Freshwater fluxes from rivers can also cause changes in SST, impacting gradients. Additionally, changes in SST will also impact sea-ice formation. Of course, another motivating factor for focusing on gradients is the importance of air-sea coupling. A major goal of the work is to examine the possible movement of the southern ice edge over the last 20 years. For this purpose, gradients of sea ice concentration are crucial. Overall, to our best knowledge, there has not been extensive research on the topics of gradients in the Bering Strait and Bering Sea. A review of the work will be highlighted as the results are presented in the Results and Discussion sections.”

 

 

Methodology: The methodology section is not well planned and well explained. The authors briefly listed the process steps. I recommend you write this section in more detail and create a workflow.

 

We thank the reviewer for reviewer for the helpful comments. The following rewrite of the methodology has been done. We have added a simple workflow diagram.

 

 

2.2. Methodology

Step one in validating satellite SST using Saildrone was to collocate the satellite derived SST products with the Saildrone deployment, using the following steps:

• Smooth the Saildrone 1-minute sampling to the daily time scales of satellite data.
• Derive daily SST gradients from the daily Saildrone smoothed product.
• Derive SST gradients from the four satellite products based on the finite different approach.
• Collocate satellite derived SST gradient to the daily smoothed SST gradients along the Saildrone deployment. The method used was a nearest neighbor approach where for a given day, the satellite derived SST pixel closest to the Saildrone daily average for that day was chosen.
• Spatial gradients for all data sets were computed along the Saildrone track.
• Linear fits were applied to the time series of the satellite derived SST gradient maps to examine trends.

1.  

     For step 1, the following equation was applied:

SST_sail (x,y)=                      (1)

    where SST_sail (x,y) is the SST derived from the Saildrone at a longitude “x” and latitude           “y”  after smoothing over the daily time steps “N.” The index “i” is simply indicating the            time step along the Saildrone deployment track.

 

     For step 2 the gradients are then derived at derived at the daily smoothed locations along the Saildrone track.

The spatial gradients are  computed such that:

SST_sailgradx (x_i,y_j) = [SST_sail (x_i+1,y_j)- SST_sail (x_i-1,y)]/Δx        (2)

SST_sailgrady (x_i,y_j) = [SST_sail (x_i,y_j+1)- SST_sail (x_i,y_j-1)]/Δy        (3)

 

where SST_sailgradx (x_i,y_j) and SST_sailgrady (x_i,y_j) are the x (longitude) and y (latitude) components of the gradient at the specified longitude and latitude along the Saildrone deployment track. The distances Δx and Δy are the distances in kilometers between the pixels at the specified longitude/latitude locations. The magnitude of the gradient at location (x_i,y_j) along the Saildrone deployment track can then be calculated  as:

 

SST_sailgrad (x_i,y_j)=        (4)

 

where SST_sailgrad (x_i,y_j) is the magnitude of Saildrone SST gradient at position (x_i,y_j) along the Saildrone track.

 

     

For step 3, gradients were derived for the satellite MWIR, CMC, DOISST, and OSTIA for each of the daily maps covering the period of the Saildrone deployment. The equation used was a simple finite difference approach such that:

                       SST_gradx(x_i,y_j) = [SST(x_i+1,y_j) - SST(x_i-1,y_j)]/ Δx                   (5)

                       SST_grady(x_i,y_j) = [SST(x_i,y_j+1) - SST(x_i,y_j-1)]/ Δy              (6)

                       The magnitude is then defined as:

                           SST_sailgrad (x_i,y_j)=     (7)

                          Step 4 was then defined as the collocation of values from equation (6) with values from

                          Equation (4).

      The workflow diagram below summarizes the above steps in the methodology used to collate the satellite derived SSTs from the four products with the Saildrove deployment.

 

               

 

Results: The Results section is well written, and the results are presented in sufficient quantity.

 

Line 231: This is not a map. North arrow, scale and legend are missing. Also, the resolution is not good.

 

Thank you for the comment. We feel the map shows the study area well. The map was taken directly from Wikipedia. The main issue was that we wanted to show the study region along with the major river outflows for the region. We have added the words, North, South, East, and West for those that might not be familiar with the area.

 

The resolution is not sufficient in all figures. scales cannot be read.

 

Thank you for your comment.

 

I am vision impaired. All the figures were created with font sizes that can be read. Because of my vision impairment, we also used different figures on the page to make them easier to read. For example, we placed Figures 5,6,7,8 on separate pages for this reason. I have gone through the figures, even with my poor vision and find the resolution adequate and readable. I have used this resolution and size in many papers. The following figure replaced Figure 1:

 

 

Discussion: This chapter is well written, but the authors need more in-depth discussion.

Thank you for your comments.

 

The following was added at the end of the discussion. The two references were also added.

 

 The northward shift of the southern ice edge can also be related to some critical issues related to changes in the Arctic.

 

[19] relates Arctic Amplification (AA) to sea ice loss. The results in this paper showing the northward movement of the southern ice edge indicate a relationship to the Arctic Amplification (AA). The (AA) is related to increasing greenhouse gases.  Additionally, [19] states that enhanced greenhouse warming is seen north of 67°N, consistent with the results found in this work and the northern movement of the southern ice edge.

 

 [20] using model simulations showed that oceanic heat transport (OHT) through the Bering Strait had a more significant impact on Arctic warming than previously thought. A key result was also that the increased OHT was dependent on the resolution of the model. This is the motivation for future and further analysis of the different SST products in the Arctic.  Thus the results presented here provide a starting point for future applications of SST in the Arc. .  

 

 

Conclusion: Adequate and complete.

 

 

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

Necessary corrections have been made. The article can be published in its current form.

Author Response

I do not seen any comments from round 2. The one comment was to add workflow as a figure which has been done.