First of all, I would suggest a different paper title to reflect the scope of the ESSD, e.g. "A new digital elevation model (DEM) dataset of the entire Antarctica continent derived from ICESat-2"

This manuscript provides an Antarctic DEM data set based on NASA’s new generation of ICESat-2 altimeter. The authors applied the spatiotemporal fitting method, so the data set covers both the ice sheet and ice shelves. This is not the first manuscript to tackle the DEM data set for the Antarctic. Nevertheless, the authors have demonstrated their product and evaluated it using the OIB and GNSS data under various surface conditions. As far as I understood, the ICESat-2-derived Antarctic DEM is not available. Considering the high-resolution and accurate measurements of ICESat-2, I expect this dataset could be valuable for Antarctic glacier research. For this reason, I would like to see this paper to be published.

In general, this paper is well written, and the structure is clear and easy to follow. It is an interesting topic and is worth to be published as ICESat-2 provides elevation measurements in much higher spatial-temporal resolutions. One weak point I can tell is that the authors provide just a 1-year data set. On the other hand, think of the entire Antarctic domain, one year data set is already quite comprehensive, in particular, authors have claimed that they can provide annual data sets in s sustainable way meaning the data set can be accumulated on an annual base.

I think the conclusion is rather short, I would like to see recommendations that authors could point out, e.g., a number of potential applications applying this data set or the forthcoming new DEM data sets generated by the methodology authors have applied. This improvement would strengthen this data paper.

Some specific comments I hope authors may find useful:

Abstract: “Antarctic digital elevation models (DEMs) data sets are essential,,,”; “human fieldwork”, is there any nonhuman fieldwork?

Introduction: P3, L80: Do you apply any other quality control criteria than what you have mentioned here?

P3, L84: “Although the signal energies of strong and weak beams are different, all six beams provide centimetre-scale elevation measurements, and the biases of two beams in one pair are less than 2 cm (Brunt et al., 2019) and 5 cm (Shen et al., 2021) for flat and steep surfaces. Thus, the effect of elevations estimated from weak beams is negligible” Not very clear text, please explain more in detail.

P4, L92: “Icessn” ?

P4, L95: What do you mean by ‘the effect of interannual changes’ here?

P6, L137:  this model-fitting method has been used in other papers (e.g., Slater et al., 2018). They have produced multi-annual data, while in this paper you have made just one year of data.  Can you point out the differentiation between your work and theirs, e.g., does the length of data processing matter?

Additionally, considering the higher coverage and spatial resolution of ICESat-2, applying a fitting model to ICESat-2 will resolve its finer observations which are not obtained by other satellite altimeters. Can you try to make use of all ICESat-2 data and apply the kriging interpolation directly, in this way you may obtain a more detailed and accurate elevation map, due to the higher resolution and accurate measurements of ICESat-2? The Authors should clearly state why this estimation method is suitable for ICESat-2 data.

P6, L144: I would like to see a figure for elevation change rate (a5), which can be used to evaluate the method performance. In addition, I have some concerns if one year of data is enough to estimate a reliable elevation change. Could you please provide the elevation change rate map (a5) to see if the method makes sense?

P8, L185- 195:  please explain clearly among those resolution numbers, What exactly number you have finally applied and why?

P9, L218:  why do you use this method, why don’t you resample the OIB to the DEM data and calculate the difference and its statistics?

Table 3:  your uncertainty map shows values of < 2m, in this Table an SD of 15 m can be found, which means that the uncertainty map may not represent this, can you explain this? Additionally, the predicted uncertainty (i.e., uncertainty map) with the actual uncertainty (comparison with OIB and GNSS data) should be compared and discussed.

Figure 7:  I found some negative values in your DEM map in the boundary of ice shelves, can you explain them?

P13, L268: Can you prove more evidence here to clarify why ice sheet elevations are more accurate than those estimated for ice shelves.

Figure 9: I noticed that OIB elevations are near> 0 while your DEM has some elevations even less than -200 m, can you explain this?

Table 5: why the number of used OIB measurement points in this table is different from that in Table 4.

Table 5: here you compared the other DEM and your DEM to OIB data, as the same OIB data were used, but the timestamps of DEMs are different, do you consider this effect or how to reduce it?

P19, L385: Again, please provide an elevation change rate map to evaluate the elevation estimation performance.

Please make a revision of the manuscript accordingly, I recommend this manuscript be considered as an ESSD publication after a revision.

Regards,

Review:

A new digital elevation model of Antarctica derived from ICESat-2

Shen et.al. 2022

The study presents a new elevation model of Antarctica based on one-year of ICESat-2 observations. The authors provide a specific time-stamped DEM with a final pixel size of 500m, following the same approach as presented by Slater et.al. (2017).

The new DEM is validated against OIB and GNSS data and compared to existing Antarctic DEMs. Results show an improved accuracy compared to DEMs based on Radar altimetry but with less accuracy than DEMs based on Radar interferometry or Stereo-Photogrammetry.

In general, it is an interesting project and worth to be published  as ICESat2 provides precise point information with high accuracy and good coverage.  This large data base should be used to generate a gridded data product of high quality which is easily accessible and to be used in different applications. The authors did this approach in a comprehensible way. The paper was already submitted to TC and underwent a review process. Most of points of my former review were answered and corrected.

The paper is well structured, methods are explained and figures are of high quality. The validation against OIB and GNSS is clear and shows at least in numbers an improved DEM compared to other Altimeter based DEMs.

Data is accesible via the data link

Based on the paper I was now curious to have a look to the DEM itself.

However, after looking at the data I'm a bit dissapointed and have some question marks with respect of data quality and usability of the dataset. Attached are two screenshots of the DEM underlain by its hillshade. It can be seen that there are a lots of artefacts visible, even over the flat lake Vostok. Elevation differences at those erroneous pixels are in the order of meters to tenth of meters. In addition a grid like structure is visible in the hillshade or in a roughness image created from the DEM itself. Furthermore the uncertainty map makes absolutly no sense to me. One can see tracks with uncertainties of 0.01m and between those tracks the values jump to 30 / 50m or more. 

I'm wondering why the erroneous pixels are not seen in the statistics of the validation. I think the authors should re-think their methodology in respect to outlier detection as well as the uncertainty . Where are the outliers come from? Why do you have such large jumps in the uncertainty map.

Based on the data set itself I cannot recommend a publication  at the current stage as to my opinion this data set is not useful because there are too many erroneous pixles all over Antarctica.

 

 

This manuscript generates a 500 m resolution DEM of Antarctica based on the ICESat-2 data from November 2018 to November 2019 using a spatio-temporal fitting approach. The authors validated the DEM using IceBridge airborne altimetry data and GNSS ground measurements, and also compared it with six other published Antarctic DEMs. Although the results show that the accuracy of this DEM is very superior, I am doubtful about this result.

General Comments:

It is a good attempt to build an Antarctic DEM dataset based on ICESat-2 altimetry data, and Antarctic DEM is important for the study of Antarctic ice sheet changes. ICESat-2 satellite can indeed provide very high resolution and high accuracy ice sheet elevation data and has great potential to be a reliable data source for building Antarctic DEM. However, unfortunately, the dataset was not completely utilized by this manuscript.

Other specific comments：

First of all, there is no innovation in the method study as the spatio-temporal method was referred from Slater et al. (2018). On the one hand, although Slater et al. used this method to build a superior performance Antarctic DEM based on data from the radar altimetry satellite CryoSat-2, the authors' transposition of this method to the data processing of the laser altimetry satellite ICESat-2 may create unknown uncertainties. On the other hand, I consider that using the altimetry data with only a time span of 1 year cannot show the priority of the spatio-temporal fitting model, and can cause fitting errors due to the limited data density and spatial distribution. In fact, this is also reflected in the manuscript, where only 46% of the grids in the 500m resolution DEM claimed by the authors are directly generated by fitting sampling points within the 500m grids, with other gaps either obtained by resampling the grids at low resolution or by kriging interpolation.

Secondly, I have serious doubts about the reliability of this Antarctic DEM dataset, and although the authors use some measured data to validate it, I do not think this validation method is reliable. Although these variations are neglected in areas with small elevation changes in the interior of the ice cap, I do not agree that it is reasonable to use OIB and GNSS data with large time differences to assess the DEM accuracy. In addition, it is not representative of the accuracy of the whole DEM, limited by the amount and distribution of the validation data. Of course, this is due to the limitation of obtaining large-scale field measurement data. However, it cannot be arbitrarily claimed that the accuracy of DEM under such validation conditions is better than the results of other scholars.

In addition, from the perspective of manuscript writing, this manuscript is well structured and the language is more fluent, but there are some places where the expression is not very clear and there are also a large number of obvious typographical errors. For example, Fig. 1b and Fig. 1c are not seen in Fig. 1, but they appear in lines 105 and 108, respectively; in line 305, the description of the comparison of other DEMs should be discussed in Section 4, which seems very confusing here; in Table 6. it should be 'Number of used GNSS measurement points' instead of 'Number of used OIB measurement points', etc.

In fact, I have also seen this manuscript in the discussion forum of The Cryosphere last year, and this does not seem to be too much changed from the previous manuscript. Moreover, another article by the authors using the same approach applied to Greenland has been published in ESSD (Fan et al. 2022), and the two manuscripts are similar in approach and writing style, and I do not think it is worth publishing a similar work again.

In conclusion, I think the manuscript has no innovations in the DEM generation method and the dataset is not reliable, and its validation data are not enough to support the authors' conclusion, so it is not recommended for publication.

Reference：

Fan, Y., Ke, C.-Q., and Shen, X.: A new Greenland digital elevation model derived from ICESat-2 during 2018–2019, Earth Syst. Sci. Data, 14, 781–794, https://doi.org/10.5194/essd-14-781-2022, 2022.

Slater, T., Shepherd, A., McMillan, M., Muir, A., Gilbert, L., Hogg, A. E., Konrad, H., and Parrinello, T.: A new digital elevation model of Antarctica derived from CryoSat-2 altimetry, The Cryosphere, 12, 1551–1562, https://doi.org/10.5194/tc-12-1551-2018, 2018.