The paper has a clear structure, as it  includes an introduction, a detailed review of relevant datasets, a discussion of methodologies, and a forward-looking conclusions. Each section logically follows from the previous one. Motivations are well stated, even though I would expect a dedicated sub-section named “motivation” inside the introduction. The abstract is concise, but it could be more engaging to summarize key findings and implications, as well as invite the readers to continue the reading. The clarity of the manuscript is adequate, however, in larger sections the readability is a bit lacking. The main point is in section 3: I suggest shortening the description of each subsection and consider summarizing with a table encompassing e.g. data provider, system name/type, key regions surveyed and/or coverage and relevant notes regarding each dataset or a table to collect the dataset grouped by key regions areas. This would allow readers to quickly grasp key distinctions between datasets.

Regarding section 4, I assume it is related to figure 4. I suggest adding the reference to figure subsection at the beginning of each text subsection, e.g. 4.1 Pulse compression, filtering, and image focussing for optimising IRH tracing – fig.4b.  So, in sections like the introduction and dataset descriptions, sentence structures are sometimes complex. I advise shortening some sentences to enhance readability.

Considering the scientific quality, the purpose of the work is clearly articulated, reflected in an adequate methodology, and its achievement compellingly underpinned by the evidence presented with the methods and techniques valid and suitable. In addition, the paper encompasses a robust range of references, except for some minor exceptions described below, the bibliography is sufficient and good.

Technical comments:

Figure2a: maybe consider adding an arrow to highlight the bedrock reflection

Figure4c: typo “differentiation”

Figure7: it is hard to distinguish light green and yellow colors in panel b and c

Figure8: I really like this figure, but may I suggest removing arrows and place the letters close to the area where they are referring to? e.g. put a small (f) close to “layer folding”

Figure9: Even though I appreciate 3D views, this one is difficult to read. I suggest rethinking this figure in 2D, using colours for the third dimension.

L195: typo ”form”

L.232-243: just to stay coherent with the writing style, you could add some references here, as it seems the only paragraph without references.

L.244-264: I suggest shortening this subsection, sticking only to the difference between analogue-digital and coherent-incoherent.

L.413: delete as it is a repetition of the line above

L.457-470: how about to replace the lines with a dotted list of the other institutions? Or another table just for them?

L.488: I think this explanation could be placed in paragraph L244-264.

L.580: I understand that focusing on processing details is not the focus, but when you talk about migration, it is straightforward to ask about velocity estimation, I think that a few lines should be added.

L.588-563: you speak about two approaches, but approach (a) has just one reference, could you provide more references? Otherwise, I would not define it as a “main approach”, just a different one from the commonly used.

L1043: can you say that this is also a priority list?

Final comment: I recommend this manuscript to be published after some minor technical corrections, which are related to editing and enhancing readability,  since the scientific quality of this work is clearly evident.

I review this review article from the viewpoint of a likely user of the product rather than a technical expert in RES or radiostratigraphy. Therefore my comments focus on the accessibility of the article to support a broad user-base for the AntArchitecture data products, rather than precision in the more technical aspects of the work, that may be better assessed by others.   

First of all I enjoyed reading this paper and learned a fair amount. The paper tackles an important topic and has an important message, this being that the development of a knowledge of the internal structure of  Antarctica's ice is important for a better-constrained knowledge of ice dynamics and ice histories, with consequent impacts from that. I have no criticisms with respect to the content, but I think that the message might benefit from some changes to the presentation of the work to have more meaning to a broader audience

Main comments

1) The value of the work outside the immediate community was not always well expressed, and I think it might be good in general to seek to define this further and keep it in focus.

Through the text there tends to be a focus on work done and work to do. For me at least there was a less clear thread on the summative value of all this work. This begins with the title, which is focused on the technical outcome, but not so much its value

Perhaps a change of title might help focus the work. I leave this to the authors but my suggestion would be something like:

Building a radiostratigraphically‐informed age–depth model of Antarctica's ice architecture to resolve cryosphere change. 

Also, the first point of confusion comes in the title...Antarctica's ice sheets are mentioned in plural in the title and early in the abstract but there is no definition of what these ice sheets are. The reader must presume this refers to mean East and West Antarctic Ice sheets but this is not made explicit until much later on in the text.

This could be made explicit, or perhaps in the early text either the singular Antarctic Ice Sheet might be better, or perhaps just 'Antarctic ice' with the sheets to be defined later as the distinction becomes important. 

2) Elements of the article read like a list of achievements completed and key findings, but the synthesis of these into a broader understanding sometimes is lacking.

In particular, for section 3 I saw the value its structure with respect to historical perspective, covering differing (yet co-evolving) equipment and data processing etc, it seems a little writer-focused, rather than reader focused.  Instead of going through by data providers, I think this section might be rewritten focusing on the main 'eras' of airborne RES surveying, each with different value to resolving architecture.

With some transitional work, the main 'eras' seem to be 1) non-GNSS, analogue, incoherent; 2) GNSS, digital, incoherent; 3) GNSS, digital, coherent....so this would be my suggestion for section 3.1, 3.2, 3.3, with 3.4 for ground surveys

In section 5, to connect the applications to the data better, I would recommend at the beginning of each subsection to include a short introduction to the physical premise of the application...that is, how is the phenomenon expected to be observed in RES data....for example in section 5.4 ice flow dynamics, you might say "Moving ice causes originally flat layers to deform through folding, tilting and disruption.  Therefore, deformed isochrons may be analysed to interpret past ice motions"  

3) Section ordering was not ideal for me to follow the concepts

 For me, section 4 I think is logically precedent to section 3 if we consider the need of the reader to understand the means of measuring and defining ice architecture, before we can meaningfully consider the value of surveys done. At least, before section 3, where we learn a lot of details about different radar systems of different providers, we need some brief introduction to the main factors that impact on data quality...this is all made quite clear in section 4.1 to 4.4 though so they could be moved en-bloc. Section 4.5 could be part of section 5

4) The discussion needs a more rigorous defense

Section 6 outlines the importance of the work to develop deeper and broader knowledge of Antarctica's ice.  I felt for this section that while it is a good identification of future directions within the scope of the review, it lacked a framing in the context of targeted decisions for the optimal outcome.

For section 6.1 My key question is what is the (relative) cost vs value vs risk proposition for each of these given technological, logistical and financial constraints? What is the low-hanging fruit and what are the grand challenges? I would appreciate here a table with some information on what resources each is expected to require (e.g. IP, funding, technology, logistics, infrastructure) to achieve a significant improvement, and perhaps some indication of if these requirements can be met in the next decade or so. 

For section 6.2 Deliverable is not the right word for this section. For me, a deliverable is a finite and defined outcome with a specific timeframe. Perhaps outcomes?

In this section it is not very clear to me the mechanisms by which these might be achieved . I think to identify the scale of the collaboration needed (i.e. is it needing a few research groups or a major multinational program), and also look at these through existing collaborative frameworks (are they sufficient?), and also to look at opportunities/needs for broad new collaboration. 

Given the length of the manuscript, to close out , I think we could do with a distinct conclusion, even just 2 paragraphs, to reaffirm the main points of the text.

Minor comments

line 72: what is the basis of these proxy records?

line 104: also sediments are often very limited on the timescales of observation, and further are quite indirect with respect to ice conditions

line 106: which ice sheets East and West?  Here is a good place to introduce the challenge of a composite ice sheet with several distinct parts.

line 107-108 - can you be more specific here?

line 115 - a point to consider here is the extent to which RES-derived architecture can be considered total architecture...perhaps a brief comment that RES can't resolve everything

Figure 1: Axes need to be 3D to match the figure.

line 124 - here again is the plural necessary?..if so we need some definition.

line 160: complex as in complicated or complex as in real and imaginary numbers? perhaps avoid complex

line 194: this active field of research presumably has at least one paper to cite

line 207: instead of for over, perhaps spanning

line 218: 'now commonly employ state of the art' is redundant.  In the absence of specifics, I think improved would do

line 225: 'shallow' RES here refers to 100s m where before and in Fig 2 the 'shallow' sounding reached 2 km. Perhaps a different temrinology could be used to differentiate - perhaps shallow vs intermediate vs deep-penetrating RES

line 249-250 - I think the brackets can go as they include a full sentence

line 251 - 'from before' replace with preceding

line 252 - 'were' needs replaced with a verb - perhaps 'being' or 'becoming'

line 254 - 'acquiring data digitally' > 'digital data acquisition'

line 280 'positional uncertainty' can be singular I think?

line 321 - delete shallow as the depth is defined numerically

line 323 - 'across both west and Eat Antarctica'. I find this confusing. Are some radar setups east and west specific but PASIN is not? Is there a important difference in acquisition for East vs West Antarctica ... if it is just that surveys have been done in multiple places then I would omit this point

line 532 - 'density value ... of the order of several meters'. Unless I misunderstand the use of density this does not make sense

line 549 - Here it might be useful to distinguish semi-automated and fully-automated...if that is the intent

line 560 - My question here is whether IRH brightness necessarily translates to significance for ice conditions. This is addressed later, but perhaps a brief comment here on this link between data and reality is warranted.

Figure 8: I found this figure, while quite nice, but very complex and hard to glean any information from.  Apart from the impression that work has been done i didn't learn so much. Perhaps fewer sub-figures and more annotations would be better?

line 813 - drawing parallels with structural geology, a technical question is whether the details of the fold geometry can indicate the deformation mechanism, or if it is ambiguous.

line 851 - winnowed is not I think the right word...perhaps pinched out or truncated?

line 943 - I don't get the sense here of whether you mean any model might do OK or more that for each circumstance there is an optimal model...this needs rewritten to be clear

line 1158 - this strikes me that the outcome of this approach would more be an upper limit on heat flux but this does not negate the value

line 1226 - This is a key step. The need also for a standardised and automated processing (as far as it possible) is also important so might be added here if relevant. A key goal I think should be to move away from 'decadal' compilations and more towards and ongoing resource...this needs automation