X- and Ku-Band SAR Backscattering Signatures of Snow-Covered Lake Ice and Sea Ice

Round 1

Reviewer 1 Report

The authors investigated relationships between the snow depth on lake and sea ice and SAR backscattering at the X- and Ku-bands. In situ measured snow depth, ice thickness, ice salinity, SWE, snow density was used.  The goal is to see whether snow depth estimation is possible with the backscattering data. It is a very valuable exploration and reported with detailed data analysis. I recommend a minor revision after some modification.

1.      The linear regressions are all based on single variable, why not try multiple regression?

2.      The coefficient of determination should have the 2 upper case.

3.       Table3, why some dates are in bold? What does the number in blue mean? There is no corresponding notation.

4.      L475: “…show R2’s than …” R2’s, maybe the word “lower” is missing.

5.      Figure 7 and Figure 9 b, legends are missing

6.      L709：“…It was found to be the most consistent temperature interval with the time series of the TSX data…”I do not find the curve of TSX in Fig 9b. From Figure 9a, I also do not observe the corresponding trend  from the blue curve. So to be clear, recommend the add the line of TSX backscatter in Fig 9b.

Author Response

Response: We thank the reviewer for the constructive comments to our paper.

1. The linear regressions are all based on single variable, why not try multiple regression?

Response 1: The purpose in this article was to concentrate on snow depth to see if it is possible to make some conclusions about snow depth by using backscatter data. Thus, multiple regressions are not used in the data analyses. Further, using multiple snow parameters, like also snow density and snow grain size, in the regression would not allow estimate any single snow parameter from the backscatter data. Furthermore, only limited data on other snow parameters than depth was available; a snowpit with detailed stratigraphy, for example, was typically available from one location only. This was also added to the text of the article. (highlighted in lines 783-787 in re-submitted file)

2. The coefficient of determination should have the 2 upper case.

Response 2: The coefficient of determination is the same as R2=r², where r is the coefficient of correlation. This explanation was added to the text. (lines 373-374)

3. Table3, why some dates are in bold? What does the number in blue mean? There is no corresponding notation.

Response 3: Explanation of bold and blue text is added to the Table 1 text. “The missions that are used for analysis in this study are shown with blue color. The dates for in situ data where data is covering lake ice area and the dates for SnowSAR flight data in missions 2, 4 and 5, where data is covering lake ice area and have in situ data are marked in bold.” (lines 158-160)

4. L475: “…show R2’s than …” R2’s, maybe the word “lower” is missing.

Response 4: Word “lower” added to the text.(line 376)

5. Figure 7 and Figure 9 b, legends are missing

Response 5: Added s°[dB] to Figure 6 (previously Figure 7) and [°C] to Figure 8b (previously Figure 9b).

6. L709：“…It was found to be the most consistent temperature interval with the time series of the TSX data…”I do not find the curve of TSX in Fig 9b. From Figure 9a, I also do not observe the corresponding trend from the blue curve. So to be clear, recommend the add the line of TSX backscatter in Fig 9b.

Response 6: TSX data curve (at IMB location)) is added to Figure 8b (previously Figure 9 b). TSX data curves for both Lake Orajärvi and IMB location are now shown in Figure 8a (previously Figure 9a).

Reviewer 2 Report

The manuscript assessed is well written with relevant findings. However, there is some comments regarding the structure and design of the first section (Introduction).

1-The section 1 (introduction) extensively discusses on previous studies on snow depth and snow water equivalent (SWE) estimation on sea ice and lake ice using the MR and SAR data and how sea ice and lake ice properties and σ are typically related (from 84 to 184). This part should be a subtitle of section 4 (Discussion).  Or if this section remains in the introduction, it should be summarized, because this is not a review article.

2- Section 1.1 (Study overview) should be grouped at the end of the introduction. There's no need to separate it.

3-There is an error with quote 22 on line 94. 22 should be in the same brackets as 23 and 24.

Author Response

Response: We thank the reviewer for his/her constructive comments. We have made changes in the Introduction as proposed.

1-The section 1 (introduction) extensively discusses on previous studies on snow depth and snow water equivalent (SWE) estimation on sea ice and lake ice using the MR and SAR data and how sea ice and lake ice properties and σ are typically related (from 84 to 184). This part should be a subtitle of section 4 (Discussion). Or if this section remains in the introduction, it should be summarized, because this is not a review article.

Response 1: Discussion on previous studies were moved under Section 4 Discussion / 4.1 Previous studies. We also added Section 4.2 Discussion of this study.

2- Section 1.1 (Study overview) should be grouped at the end of the introduction. There's no need to separate it.

Response 2: Section heading ”1.1 Overview of the study” was taken away from the end of Introduction.

3-There is an error with quote 22 on line 94. 22 should be in the same brackets as 23 and 24.

Response 3: The error was corrected.

Reviewer 3 Report

In this paper, the authors investigated X- and Ku-band SAR backscattering signatures of snow-covered lake ice and sea ice, and assessed the feasibility of retrieval of snow characteristics using X- and Ku-band dual-polarisation over lake ice or sea ice. This work is interesting and meaningful for estimating snow depth or SWE on lake or sea ice. Overall, this paper belongs to the analysis of experimental results, and there is less algorithm innovation. There are some questions the authors may answer or consider before this manuscript goes for further consideration.

1.  P8-Line 324: It is recommended to use a diagram to illustrate the approximate location of the in-situ samples .

2. P8-Line 328：The authors can choose a larger averaging window size for the sea ice and lake ice, such as 7 by 7 pixels

3. P12 - Fig3: It suggests to list the coefficient of determination (R2) between snow depth and the meanσ0 at the four channels.

4. P23-Line 729: In Fig. 9. It is recommended that the vertical axis in figure 9(b) should be  illustrated.

5. Since the observed coefficients of determination were found to be weak or moderate based on the linear regression analysis, the form of nonlinear fitting should be considered based on multiple factors.

6. P29-Line 897: In Fig. A3. It is recommended to indicate the information of the horizontal and vertical axes in the graph.

Author Response

Response: We thank the reviewer for taking the time to review our paper and for the constructive comments provided. The changes are highlighted in the re-submitted file.

1. P8-Line 324: It is recommended to use a diagram to illustrate the approximate location of the in-situ samples.

Response 1: It is shown more clearly in appendix images for lake ice. Also added a picture for sea ice in situ. In addition, added to the text below the figure: “(see also Figure A2 in Appendix A for other missions). e) SnowSAR KuVV imagery with in-situ measurement transect on sea ice.” (highlighted in line 137-138)

2. P8-Line 328：The authors can choose a larger averaging window size for the sea ice and lake ice, such as 7 by 7 pixels

Response 2: We tested different window sizes and they had very little influence on the results. It is noted  that larger window would lead to larger spatial averaging of backscattering coefficient data, and therefore, we would like to keep the 5 by 5 pixel window. (The text of the paper was also improved). (lines 238-240)

3. P12 - Fig3: It suggests to list the coefficient of determination (R2) between snow depth and the meanσ0 at the four channels.

Response 3: The list is shown in Table 4 Added to the figure (now Figure 2) texts: “The values for the coefficient of determination (R2) are listed in Table 3.” The same is made for the case of sea ice (now Figure 4, previously Figure 5): “The values for the coefficient of determination (R2) are listed in Table 6.” (lines 446-447)

4. P23-Line 729: In Fig. 9. It is recommended that the vertical axis in figure 9(b) should be illustrated.

Response 4: Added to Figure 8b (Previously Figure 9b), Snow depth [cm] Ice thickness [cm] to y-axes.

5. Since the observed coefficients of determination were found to be weak or moderate based on the linear regression analysis, the form of nonlinear fitting should be considered based on multiple factors.

Response 5: Nonlinear fitting was also tried and multiple factors could be considered although in this article we concentrated on investigating if it would be possible to estimate snow depth by using only linear fit to backscatter data. This was added to the text of the article. (lines 780-781)

6. P29-Line 897: In Fig. A3. It is recommended to indicate the information of the horizontal and vertical axes in the graph.

Response 6: Added [km] to Figure A3 axes.

Reviewer 4 Report

I congratulate the authors for their study. This study is a good for examining snow and ice masses with radar backscattering signature. The study revealed that radar data still poses major challenges for areas covered with snow and ice. These were also supported by regression values.

In general, it is a well-designed study. Only one minor issue I mentioned below can be corrected.

-Figure 2- Scale is not used in any of the maps. It would be more appropriate to use a scale. A more effective map could have been made rather than showing map A on Google earth.

Author Response

Response: We thank the reviewer for evaluating our work, and for the provided recommendation.

-Figure 2- Scale is not used in any of the maps. It would be more appropriate to use a scale. A more effective map could have been made rather than showing map A on Google earth.

Response 1: Figure 2 (now Figure 1) was revised. We now used a QGIS image with scale and [m] added to axes, and a scale added in the last image.

Round 2

Reviewer 3 Report

All my concerns have been addressed, and the paper can be accepted as a publication.