Review

This is an interesting work, expanding altimetry products on the sea-ice in time and including additional satellites in the mapping, particularly in regions covered by sea-ice that are usually poorly represented in altimetry products. As the authors show well, the specific processing for the sea-ice covered regions increases the performances of the product, it is thus an important step for future polar science.

Despite the importance of the work and the science and validation being robust, the manuscript could benefit from a thorough revision.

General improvements:

1. Many citations have parenthesis when they should not have (l. 45, 46, 48, 51, …)
2. Please revise all the figures. Too often the labels for the axes, legend, colorbar are not readable because too small. Every letter of a plot should be readable, otherwise it should be removed. And often the images do not look high-resolution.
3. The generated product needs a name. It would help the reader in this manuscript (instead of calling it “our/the product” in the text or just “Product” in the images). Also, it will be beneficial for future works using and referencing it.
4. Check if the tables’ caption has to be below or above the table.
5. Use consistent layout for tables (linewidth of borders)
6. In the methods section, it would be useful to have a recap of all the processing applied to the data as a diagram. What editing, debiasing techniques are applied to each product sequentially, otherwise it’s hard to keep track.
7. The altimetry product created is not characterized enough in terms of capabilities. The minimum information required is at least the effective spatial resolution (which I imagine will change in time and in the different regions due to coverage) and the temporal resolution. This is crucial information to understand what kind of processes can be resolved and studied.

Specific

1. 85 Cryosat-2 not Cryosat
2. Pag 4. The table of the satellite missions should be Table 1
3. 116 Supplementary Table S2. Somewhere in the text or in the Supplementary table it would be interesting to add the distance of the in-situ stations from the coast (in the text at least you can add an average value). This would strengthen your comparison later, when you explain why some of the differences with the altimetry may arise.
4. Fig 2: Please add in the caption the meaning of the red dot of the bottom panels
5. 157 the sentence is missing a verb
6. 174 latex rendering not working
7. 206-211 : this explanation is too quick. Why is the extra-editing needed? What does the threshold remove? Why do you first perform and “extra-editing” on the data (what data?) and then an “iterative editing” on the ocean data (only open ocean, no leads?)? Also add one sentence to briefly describe the “off-nadir hooking” phenomenon. Is this processing the same for all satellites? Please rephrase the full paragraph adding information and clarity.
8. 215: fewer by how much? What is the proportion of data removed? Is there a general statistic about the difference between leads and open-ocean points kept, compared to the initial amount?
9. What is the statistical occurrence of the limitations mentioned in paragraph 225-235 for Envisat, and other satellites if relevant? How many points are correctly classified as leads, how many are incorrectly classified as leads and how many are missing? How does a misinterpretation of a lead influence the results? This is important to judge the robustness of the method
10. 229: what do you mean at least for this example? How the results changes for the other satellites?
11. Fig 3:
    1. Add the units for the time dimension in panel (a) and add the geographical coordinated in panel (c)
    2. Panel b is hard to interpret. Please make bigger. One suggestion could be to keep the horizontal panel (a) at the top of the figure, while (b) and (c) can be displaced in a second row next to each other.
    3. Make legend labels and markers bigger because they are hard to understand
    4. Colorbar label and ticks must have bigger font
12. 257 Why is the additional debiasing step only performed for Envisat?
13. 262 why is Cryosat-2 used as a reference? You say it at l. 234, but should be better explained here.
14. Add reference for ease2 grid: https://doi.org/10.3390/ijgi1010032
15. Section 3.6 rephrase. The interpolation is an important step of your methodology and needs more details. For example, you say that the “the expected variance and noise estimation necessary for the interpolation also come from Auger et al. (2022a) and Veillard et al. (2024)” without mentioning why you need these variables and how do they influence the interpolation. Mention the basics of the DUACS-DT2022 mapping and how this kind of interpolation could influence the results.
16. 277 is this the effective spatial resolution of the product or the distance between grid points? What is the effective temporal resolution? What are the covered latitudes?
17. The comparison between your product and CMEMS is too quick
    1. we need some spectra (at least in the supplementary material). You could do frequency-wavenumber spectra to know exactly what kind of structures the product resolves in space and time. At least in the open-ocean region, to provide more robust results for this comparison.
    2. You should add an extra panel per row with the difference between CMEMS and “your product”, to highlight where the major differences are
18. Section 4.2.1: Frequency-wavenumber spectra of the SLA would be a good way to quantify the scales at which the satellites agree or differ. It is not enough to visually compare the images finding “good agreement in large-scale spatial patterns across missions (Fig. 6b,c and e,g). While discrepancies are evident at smaller spatial scales”. What is large and small scales? It is important to do this to be sure of what you are comparing and to add robustness to the final product. Generally throughout the entire paper, scales need to be quantified better. You can do it monthly, but it would be even better to do it at the highest temporal frequency that you have.
19. Fig 6:
    1. “(b-c/e-g) Comparison of monthly SLA snapshots” it is not the comparison but just the individual snapshots. Right?
    2. Panel (c) should be Cryosat-2
20. 362 at face value? Are you sure this is the right expression you want to use here?
21. Fig 7b: what season is this plot? Is it showing only winter or also other seasons? If other seasons are included, how do you treat the region south of your white contour in spring/summer/autumn? Is it included in the open-ocean distribution?
22. 376-380: can you provide the mathematical formulation of the formal error?
23. 389 (REF)
24. 393: what do you mean by “in addition to Envisat”?
25. Fig 8:
    1. No need to write (top) and (bottom) in the caption if you have separate letters for the panels
    2. Mention panel (d) in the text. Add info on the time scale on which the average is performed. Is it for the full product?
26. 459 Add OSISAF acronym or shortly explain what it is
27. Tab 5: You don’t mention in the text the change of the results when the sea-ice coverage exceeds 10%. How do you explain that most of the Pearson Coefficients increase in the higher sea-ice scenario?

Review for essd-2025-653

 The presented work describes an important dataset generated from observational data for the investigation of polar sea level and geostrophic currents in an environment that is very challenging to monitor. The generation of this data set using satellite altimetry requires particularly careful data processing and the use of special algorithms, for example, to reliably detect open water spots and determine the sea surface height at these locations. This dataset uses observations from four altimeter missions (conventional and delay-doppler). The time period spans from 2003 to 2021. The data was processed with regard to the expected challenges in sea ice areas and interpolated onto a grid with a resolution of 25 km using an optimal interpolation approach.

I have read the work and left general and specific comments. The general comments address the overall context, while the specific questions refer to individual sections, which are marked with line numbers based on the current text version. Due to the large number of comments and many questions as well as open points, I would prefer a significant revision of the entire document but also of the data presented and consider the revisions to be major.

General comments:

A dataset is created based on observations from four altimeter missions. However, consideration should be given to why the dataset was not extended to at least 2024/2025 and why Sentinel-3B was not used. Why were not all missions used up to and including 2021 (i.e. Sentinel-3A)? As a result, the dataset tends to become outdated very quickly or to have limited applications.

Regarding Sentinel‑3, the dedicated sea‑ice thematic data products provided by ESA (LAN_SI) were not utilized, even though they are specifically optimized for sea‑ice applications due to their enhanced waveform binning and associated technical improvements. The manuscript should clarify why these products were excluded from the processing chain. A reprocessing effort that incorporates Sentinel‑3B, as well as the ESA LAN_SI thematic datasets, is strongly recommended to ensure that the dataset fully benefits from the most suitable altimetry products currently (and freely) available. If such an update is not planned, the authors should provide a detailed justification and a thorough discussion of the differences, similarities, and potential advantages or disadvantages of the Sentinel‑3 products used in this study compared with the publicly available ESA LAN_SI sea‑ice–thematic data.

In general, the term “assimilation” is used in the document. From my understanding, assimilation is the process by which observational data (e.g., sensor data, measurements) are systematically incorporated into a running numerical model in order to update its state. This term should only be used in connection with numerical simulations, which is not the case here, as this is a data set based purely on observations. Please rephrase the relevant passages.

The work very often lacks clear scientific statements and precision. In addition, there are many technical shortcomings. Abbreviations are frequently introduced, and satellite mission names are spelled differently. The plots generally have poor print quality (although this may also be related to the creation of the preprint). In addition, the same parameters are often presented in different units (m vs. cm). In addition, the term “along-track” and as a stand-alone description, “along-tracks,” is unclear. Along-track describes, for example, observations along a track, as opposed to gridded data. It describes how the data is available, either as along-track observations/elevation/heights, etc., or as gridded observations/heights. “Along-tracks” does not sound scientific. Sometimes the text contains “along-tracks” or just “along-track”. Please be more precise here.

For the validation, it would be helpful to clarify why the datasets introduced earlier in the manuscript were not included in the comparison. Instead, the study relies on gridded CMEMS data, which is not specifically designed for sea‑ice applications. Providing some context for this choice would strengthen the overall transparency and interpretation of the results.

Geostrophic velocity components are provided in the data product. These variables are not mentioned or presented in any greater detail in the explanations. It is not clear why these were calculated or included at all.

The paper is very difficult to read, as some facts are often repeated or not clearly separated, e.g., possible causes for deviations or differences between datasets used in the validation. Explanations of how the altimetry observations are edited are mixed with descriptions about imaging systems such as MERIS, which are more part of the validation. A validation with a recent independent altimetry dataset appears without any context. I would suggest a restructuring.

At the same time, some parts of the paper could be shortened considerably. I don't think it's necessary to explain the entire story of satellite altimetry data processing. There is already a lot of literature on this subject that can be cited.

Specific comments:

Line 20: “under sea ice cover”: You don’t observe SLAs under the sea ice. You detect leads and interpolate the SLAs to retrieve information under the sea ice. Here, it appears as if satellite altimetry could be used to look beneath the sea ice. Please rephrase.

Line 30: The mention of “coverage in 2023” feels a bit misleading, since the dataset presented here extends to 2021. You may want to rephrase or clarify this to avoid creating confusion for the reader.

Line 32: “characterized” à investigated?

Line 34: “is” à are

Line 35: This relationship is confusing. SLAs refer to a mathematical reference surface and not to a physical one. Geostrophic currents are derived based on a physical reference (i.e. geoid).

Line 37: How can satellite altimetry be used to provide insights into the ocean vertical structure, in particular by using SLAs? I think the sentence must be rephrased.

Line 37: Same as the comment above: How would you obtain information about structural ocean changes if only the sea surface can be spotted. Satellite altimetry is a geometric observation technique in which the sum of the effects acting on the surface height is always observed.

Line 46: “[…] from space.” This sentence does not fit to the sentence in Line 40 “However, this technique has remained [..]”  There must be a logical connection, perhaps through additional information about what needs to be done to enable satellite altimetry to be used in the ice-covered ocean.

Line 46: “open-water hole”, a lead is not a hole. This does not sound scientific.

Line 47: “extremely reflective”, what does this mean? This is not scientific wording.

Line 51: Check the brackets.

Line 52: MSS is Mean Sea Surface not Mean Sea State à here it is correct.

Line 54: There is also important work from Rose, S.K.; Andersen, O.B.; Passaro, M.; Ludwigsen, C.A.; Schwatke, C. Arctic Ocean Sea Level Record from the Complete Radar Altimetry Era: 1991–2018. Remote Sens. 2019, 11, 1672. https://doi.org/10.3390/rs11141672

and

Doglioni, F., Ricker, R., Rabe, B., Barth, A., Troupin, C., and Kanzow, T.: Sea surface height anomaly and geostrophic current velocity from altimetry measurements over the Arctic Ocean (2011–2020), Earth Syst. Sci. Data, 15, 225–263, https://doi.org/10.5194/essd-15-225-2023, 2023.

The authors created datasets that include altimeter observations from nearly 30 years. Please add this references.

Line 55: from the missions ERS-1/-2 and Envisat to quantify …

Line 56: utilized laser altimeter observations from ICESat to map the Arctic dynamic ocean topography

Line 54-60: Please check the paragraph for more missing literature/references

Line 70: Please comment on the fact why you haven’t used Auger et al. 2022b and Auger et al., 2023 and Veillard et al., 2024 as a comparison dataset for your validation activities. What is your innovation in context to these references? This should be somewhere highlighted. Why should users use your dataset and not a dataset presented in the references?

Line 96: The Ka-Band does not allow for a higher sampling rate. AltiKa allows for a higher sampling rate. Please rephrase.

Line 96: 8 km à diameter?

Line 99: SARIn is not entirely used on land ice. Please check the CryoSat-2 acquisition mode maps.

Line 102: If there is a long repeat orbit, the tracks are spatially quite narrow and dense, indicating a high spatial but low temporal resolution. Please rephrase.

Line 103: I don’t understand the sentence with “contain artefacts”. Why is a long repeat orbit responsible for unphysical meridional stripes? Please also check the brackets behind Auger et al. 2022.

Line 105-106: What is the name of the dataset from the French space agency? Which data is used here? What do you mean by “specifically designed”?

Figure 1: Why does Sentinel-3A suddenly stops and is not available after 30.06.2021. Why don’t you use Sentinel-3B to densify your observations (see general comments)?

Section 2.2: As far as I understand you also interpolate via DUACS, but you also compare a DUACS processed dataset? Can you please explain the differences here and why CMEMS is an independent dataset?

Line 121: What do you mean by 100 days globally and 60 days at high latitudes? What defines high latitude against the background that “globally” includes also high latitudes. Why do the days change?

Line 123: Where are these locations? Maybe you can link Figure 9?

Line 129: What do you mean by “key geophysical parameters” All physical parameters are important variables. Please rephrase. You write parameter(s), but list only SSH.

Figure 2 caption: Mean Sea State à Mean Sea Surface

Figure 2 caption: “Range” is not a reference height. SSH is the difference between the height of the satellite above a reference surface and the distance between the phase center of the altimeter and the sea surface. Please be more precise here and rephrase.

Section 3.1: Since the article describes a data set in the ice-covered ocean and represents an advanced data set, it is not necessary to explain the entire processing of the altimetry data. There are numerous publications that do this. In my opinion, section 3.1 can be shortened considerably.

Line 147 – 148: Do you mean LRM data? What does that mean in the case of Sentinel-3A, since Sentinel-3A observes exclusively in SAR mode? I'm lost here. Why MLE4 for SAR missions? Could you please clarify.

Line 151: SSH (abbreviation should be enough)

Line 159: signal à effects?

Line 161: MSS à Mean Sea Surface

Line 165: What is the name of the MSS used?

Line 167: Without direct context, you provide information about the ADT. Why? Is it ensured that the SLA and MDT refer to the same MSS? Please provide further details.

Line 172: Now you use a different MSS, are you also computing different SLA? Please explain.

Line 173: Which geoid are you using here: GOCO06S or GOCO06C? The geoid model is not cited.

Line 174: LaTex issue?

Geostrophic equation: Why do you need a vertical unit vector? What does it contain? You provide only horizontal velocities. Please explain and rephrase.

Line 180: “in this study fly above sea ice” This is not scientific wording. Please rephrase.

Line 180: Why should altimeters not work above sea ice? Please rephrase the sentence.

Line 181: get à receive

Line 182: “reflects the radar wave extremely well” This is not scientific wording. Please rephrase. What does “extremely well” mean in terms of physical behavior?

Line 184: See comment Line 182.

Line 186: Why is it important to know how the classifier is validated? It would be more interesting to know briefly how the classifier works since this is an important step in the sea ice altimetry processing chain. Please add some more info here.

Line 197: Why is there an offset? It's written above there should be no offset? Can you explain this sentence a bit more.

Table 2: Is this MSS used for generating ADT heights or SLAs? In the caption it’s written for SLA, but what MSS are you using for generating ADT heights?

Line 202-203: A reference would support this hypothesis. Otherwise, some more explanations or additional tests are needed. Please comment.

Line 204: Please add a reference (no reliable SSB…)

Line 206 – 211: I don't fully understand this procedure. How do you deal with the very small number of lead observations compared to a large number of observations from the open ocean? How do you deal with along-track data that consists only of lead observations? It would help a lot to restructure this text passage.

Line 215: You refer to Fig.3b?

Line 216 – 218: “Being too conservative […] sea level data à It seems the second part of the sentence is missing. One would expect “and being too optimistic… “Please rephrase or add the counterpart of the sentence.

Line 218: “sea level data” in this areas. Please add.

Line 218: “ the example along-track” à you mean the example pass? Something is missing here.

Line 216- 222: In general, I don’t understand the entire paragraph. I have many questions here.

It starts with “As in all filtering” and ends with “highest uncertainties”. What is the “all filtering” and what are the highest uncertainties? What can we expect here? Concerning along-tracks. It’s along-track data. Along-tracks is not scientific wording. Please change this. What are the repetitive patterns you are talking about? Maybe you can rephrase this text passage.

Line 224: Why do you switch the topic to MERIS? What does this have to do with data editing? I see this part more under data or briefly touched upon with a reference to how MERIS works. What is the resolution of MERIS compared to the spatial extend of the leads you want to detect? Can MERIS be used as a reliable comparison source for this task here or can it be that you detect leads which cannot be spotted with MERIS. Please discuss this.

Line 228: What are “colder” colors? Please rephrase.

Line 232: Drinkwater et al. 1991 is an important reference. But, I’m doubting this can be 1:1 used in this context to derive height offsets. Maybe there are some further references?

Line 232: Point à Observation?

Line 234: You write can be removed, but I guess it is removed?

Line 235: Why cannot it be removed in the cross-track direction? Please add more information.

Line 237-238: Is it like this? I think this sentence is very confrontational. How can you justify using the most recent methods? In the field of classification, for example, there are several other, even newer algorithms that can be used, but only one, Poisson et al., 2018, is listed.

I would suggest rephrasing this last sentence.

Figure 3 caption: The caption is really long and explains MERIS. I think this text should be in the main document and not in the caption.

Figure 3 caption: “Trajectory of the along-track shown in A.” What and where is A? After “along-track” something is missing. Maybe observations?

Line 256: What is this “debiasing step” and why is it depending on the peakiness. The term peakiness was never introduced.

Line 263: What is the mean state value in this context? How is it computed? Is it the same like the mean inter-satellite offsets?

Line 271: What are the "input correlation scale files". A brief summary would be very useful, since this is a very important point.

Line 279: There is no mission included with a native temporal sampling of 10 days. See Table 2. Please explain.

Line 280: How are the daily files interpolated (linear, nearest neighbor etc.)?

Line 283: SLA

Line 284: CMEMS

Line 286: I don’t get the argument including Jason-2. It seems CMEMS is better, because they included Jason-2 to their processing chain? Wouldn't it make sense to include Jason-2 in your data set as well?

Line 294: reveals à reveal

Line 304: Do you mean Fig.4 e?

Line 307: Could you give some more information? How is the climatology computed and what is the Armitage et al., 2018 -180 cm MDT contour?

Line 310: Check brackets of Auger et al.,2022b. What was described in Auger et al.,2022b? Could you please add some short sentences highlighting the main findings.

Figure 4: In general, the printing quality of the figures is poor. Maybe this is related to the fact of a preprint. But it is hard to spot details. The term “product” is somewhat unfortunate, as CMEMS is also a product for example. A separate term would be preferable for better understanding.

Figure 5 caption: See comment Line 307.

Figure 6: In Figure 1 the application of CryoSat-2 data starts in 2010-07. Why is the time axis limited to end-of 2010. The transition of Envisat nominal phase to Envisat extended phase would be quite interesting, since the orbit of Envisat changes dramatically.

Line 347: Which DUACS dataset? I thought the optimal interpolation was also done by DUACS? What are you comparing here? CMEMS vs. your product?

Line 345:end: RMSEs à RMSE values

Figure 7 caption: “The with…” à remove with

Line 380: SLA

Line 382: “Panel a” à Fig8. a?

Line 389: REF??

Line 390: “and its orbit results in non-uniform sampling with temporal gaps“ à What do you mean by this?

Line 390-392: Please add a reference

Line 445-446: “correlation of 0.52”. I wouldn’t say this is “well captured”. It’s a correlation of around 50%, which is more moderate.

Figure 9 caption: The caption is very long. Maybe it can be shorten or the plots shown are limited to the highlights?

Section 4.2.5: Why is this comparison product (Dragomir et al., 2024) not mentioned in the introduction? This section seems somewhat out of context, as a very detailed comparison with CMEMS was made previously. This could somehow be combined, as this is also a comparison with a recent altimetry product. This is a completely independent validation, which is particularly interesting and should be given more attention.

Line 476: SLA

Line 480: SSB

Line 485: “some mesoscale activity”. What do you mean by “some”?

Line 487: What is the DUACS reference product? CMEMS?

Line 490: Delete “then”

Line 491: What exactly do you mean by “independent” You use gridding strategies from DUACS (optimal interpolation). Please comment.

Line 492: What do you mean by “global processing strategy”? You don’t process globally.

Line 494: comparable to what?

Line 496: Add a new line after Southern Ocean.

Line 501: Once again, it is very vague to compare correlations based on two completely different regions and data sets.

Line 508: “:”?

Line 516: “considered” as “a”

Line 521: RMSE

Line 527: a “,” is missing

Line 528: Which other observing systems?