Estimation of Arctic Sea Ice Thickness from Chinese HY-2B Radar Altimetry Data
Ruibo Lei
Round 1
Reviewer 1 Report
This manuscript first used the reprocessed HY-2B radar altimeter L1 data to retrieve the Arctic SIT, and calibrated with the AWI CS-2 SIT products. The study compared the HY-2B calibrated SIT estimates with the Goddard Space Flight Centre (GSFC) CS-2, Jet Propulsion Laboratory (JPL) and GSFC ICESat-2 (IS-2) SIT products and airborne campaign data (Operation Ice Bridge) and ice draft data from the Beaufort Gyre Exploration Project (BGEP). I consider the data from HY2B and this submission is a good contribution to the community, and will help us better know sea ice response to climate change. But I do have the following comments that in my opinion that should be addressed first.
General comments
G0: The study computed the sea surface height by averaging the three lowest modified relative surface height points per 25 km. How do you determine that the average of lowest three points per 25km segment is sea surface height? Why not the average of lowest five points per 25km segment or other number? Is this method associated with the higher raw SIT retrieved by HY-2B, as shown in figure 6 and 7?
G1: The differences between HY-2B and AWI CS-2 SIT are roughly eliminated by a simple linear regression method due to different satellite measurement modes and different retrieval algorithms. I don't think it can be corrected simply by linear regression. Because it obscures the real reason for the difference. Does the corrected linear regression equation generalize to other months? Or will they rely on the AWI CS-2 SIT products? In addition, the study did not discuss in further detail including the aspects that are causing some of these large discrepancies between the derived HY-2B SIT and AWI CS-2 SIT. We encourage the authors to analyze the reasons for the differences, but not merely states that they are.
G2: The description of the geophysical corrections data sets is not clear enough. The website where the data were obtained and the description of the calculation method are not clear. What formula is used to calculate the geophysical corrections terms? Or by interpolating the model data? Nearest neighbor or linear interpolation?
G3: According to Table 2 and 4, why is the RMSE of HY-2B compared with other SIT products after calibration larger than before calibration in the thick SIT ranges (3-6 m)? At the same time, why is the RMSE of HY-2B compared with OIB SIT after calibration larger than before calibration (Figure 14)? But the RMSE of HY-2B compared with BGEP SIT after calibration lower than before calibration (Figure 16).
G4: How do you calculate the uncertainty if you use the linear regression to calibrate the sea ice thickness?
G5: The authors mainly compared the HY-2B retrieval with CS-2 and IS-2. However, Sentinel-3 might be a better comparison, because it has the same inclination angle as HY-2B, and works on Ku-band (although they are of delay-Doppler type). Is it better to compare against Sentinel-3 retrievals? This is only a suggestion. The authors can decide whether this is an option or not.
G6: The HY-2B satellite is carrying a dual frequency Ku (13.575 GHz) and C bands (5.25 GHz) radar altimeter. The ground footprint diameter of Ku and C bands are 1.9 km and 10 km, respectively. The differences between Ku band and C band are mainly reflected in footprint size, waveform and radar penetration factor. Do you think the C band can be used to retrieve Arctic sea ice thickness? What should you pay attention to?
Specific comments
Line 128: add “used” after “we also”.
Line 129: use“SIT”instead of“SIT”.
Line 130: add space between “IS-2” and “products”.
Line 179/184: The purposes of removing the sliding average of 25 km and the average of 3 lowest points per 25 km should be stated. This is not commonly applied for the lead/floe retrieval methodologies, and as such can add to the reasons why the results differ.
Line 188: The wave propagation factors (0.22 or 0.25) depending on snow density and thus varying, even in the AWI product. It is based on slower wave propagation, given by (), where c is the speed of light and cs is the speed of light through snow. Here, , where depends on the snow density (which you vary when using Mallet et al. (2020)). If so, this should be corrected.
Line 214-215: The manuscript mentioned that the method of the average of 3 lowest points per 25 km can overestimate HY-2B sea ice freeboard, but it didn’t give the reason of the differences.
Line 217: The HY-2B SIT is different with the AWI CS-2, resulting from different measurement mode, different retrieval algorithm, different MSS model and so on. The study directly used the linear regression to calibrate the discrepancies between the derived HY-2B SIT and AWI CS-2 SIT, which maybe ignore the real reason for the discrepancies. We encourage the authors to discuss the reason for the differences.
Line 385: How to compare the SIT estimates between BGEP and HY-2B and CS-2? How is BGEP gridded (Figure 16)?
Comments for author File: 
Comments.pdf
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
Review on “ Estimation of Arctic Sea Ice Thickness from Chinese HY-2B Ra-2 dar Altimetry Data ” by Jiang et al.,
Arctic sea ice thickness is one of the most important sea ice geophysical parameters, indicating climate change, and has an important feedback effect on the changes of Arctic climate and marine environment. It is also one of the most-difficult sea ice parameters to be obtained by satellite remote sensing. Therefore, it is worthwhile to use the observation data from a new satellite remote sensing payload to retrieve the sea ice thickness and compare it with the existing satellite remote sensing products, submarine mooring ice draft data and aerial remote sensing observation data to evaluate the accuracy of the new data products. However, there are still some problems in data analysis and expression at present, so I recommend that the paper should be further revised before publication.
General comments:
1) The scientific significance of developing new sea ice thickness products (study motivation) is unclear, and the final evaluation results show that the new products do not show higher accuracy compared to the existing satellite remote sensing products.
2) Relative error is also very important for evaluating the products of sea ice thickness. It is suggested to increase the analysis of relative error.
3) Compared to the AWI CryoSat-2 ice thickness products, it is better to consider establishing different correction models based on the data from different months, because the correction model may depend on the ice thickness and the physical characteristics of snow and sea ice.
4) For the evaluation of sea ice thickness products, the observation data of the sea ice mass balance buoy is also very important. MOSAIC provides the observation data of the growth process of sea ice with different thickness at the grid scale (20-30km) from Oct 2019 to Apr 2020, which is conducive to verifying the ability of satellite remote sensing products to estimate the growth rate of sea ice (Lei et al., 2022).
Lei R, Cheng B, Hoppmann M, Zhang F, Zuo G, Hutchings J K, Lin L, Lan M, Wang H, Regnery J, Krumpen T, Haapala J, Rabe B, Perovich D K, Nicolaus M. 2022. Seasonality and timing of sea ice mass balance and heat fluxes in the Arctic transpolar drift during 2019–2020, Elementa: Science of the Anthropocene, 10 (1), doi: https://doi.org/10.1525/.elementa.2021.000089, 2022.
Special commentsï¼›
1) Introduction: The study significance of sea ice and sea ice thickness should focus on Arctic climate, Arctic ocean marine environment, but not global scale climate system.
2) Line 39: continuous hemispherical SIT information: In fact, the data is not continuous.
3) Snow depth, snow and ice density: These are important parameters for retrieving sea ice thickness from satellite altimetry data. Are the selected values of these parameters consistent for the satellite remote sensing products you compared?
4) Figure 17: The comparison of growth rate is also very important
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper as a potential merit in presenting the HY-2B altimeter a possible source for ice thickness data at lower latitudes. However, to improve the paper, I suggest a rewrite of the abstract and motivation, as well as adding some more concrete conclusions.
It is unclear what the reprocessing, which is mentioned several times in the paper, includes. It seems that merely the surface heights are processed into SIT, which is presented quite well.
The selected path from HY-2B surface heights to SIT, calibration of this with CS2 and comparison to other satellite products plus OIB and sonar measurements is presented mostly in orderly and straightforward manner. However, the main conclusion is lacking.
The concept of calibration needs some clarification. A linear regression with AWI CS2 SIT is clear and straightforward solution, but conclusions of the result and justification for the method are lacking. Why is AWI CS2 selected for the calibration? Does this contribute to the SIT retrieval method presented?
It is not clear why the RMSE and correlation coefficient for the HY-2B calibrated SIT estimates and JPL IS-2 SIT products are mentioned in the abstract, emphasizing them over others.
More detailed notions below.
Section 2.
More information of instrument characteristics needed, namely surface resolution: footprint size and geometry. Are all comparisons made for monthly averages projected on a 25 km EASE grid? Consider the differences in footprint and temporal scales.
Parameters for formulas should be more diligently written, e.g. row 173: mean sea surface.
line 93: SIRAL stands for SAR Interferometric Radar Altimeter
160: explain the parameters for the range and geophysical error corrections.
173: mean sea surface h_mss
175: a reference is needed for the DTU 21 MSS
178: Elaborate.
192: What is snow density here?
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
Review of the Manuscript titled “Estimation of Arctic Sea Ice Thickness from Chinese HY-2B Radar Altimetry Data.”
In general, the Manuscript demonstrates significant amount of work performed by the authors. Methodologically, the estimation of freeboard would possibly reduce the bias. The authors promise to address the usage of freeboard in future. In my opinion, the Manuscript can be interested to the readers as a preliminary work. I have only two minor comments:
Minor typos: almost all references don’t have space (e.g., on line 128, page 4 “products[21]” it has to be “products [21]”).
The sentence on line 128 says: In this study, we also GSFC and JPL IS-2 SIT products to validate the HY-2B SITes estimates. Probably work “used” is missed
Sincerely,
Reviewer
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
Authors have revised the MS according to my comments. One small suggestion:
For the result of Fig. 14, the accuracy of before cal is better than that of after cal. Authors think that the linear calibration model is not suitable for all SIT range. You can give another result of 10points of FIg. 14 (0-4m) to prove your opinion.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 2 Report
The paper has been revised accordingly, and can be considered for publication after further revision of the language.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
