Please see the attached comment file.

This manuscript primarily has two components, the first is about tracing and dating 16 IRHs across two major East Antarctic regions using newly acquired and some existing data resources. The latter is about demonstrating IRH utility for model calibration using specific 90ka IRH. The authors have put in substantial effort acquiring new radar data, tracing and dating isochrones across two major East Antarctic regions, and demonstrating their utility for the Antarctic modelling community. The datasets being released will be immediately useful, and the modelling insights are valuable. The language is clear and the figures are mostly well made. This work is worthy of publication in The Cryosphere, but after some revisions which are needed to be addressed.

Major comments:

• The main issue is that the manuscript is trying to do two things - present the IRH datasets and demonstrate their utility - but only actually uses 2 of those 16 IRHs in the modelling exercise. So, the reader has to go through detailed descriptions and methods for all 16 IRHs (Sections 2.1-2.2, 3.1, plus Figure 2 with 16 panels!) before getting to the modelling study that only uses the ~90 ka isochrone. This makes the manuscript feel a bit more like a data paper at times. I suggest restructuring the manuscript a bit to improve on this aspect. Below is how I suggest restructuring the paper:

1. Introduction
2. Airborne radar data; 2.1 New data at WSB; 2.1 New data at DML
3. Mapping IRHs; 3.1  Isochrone tracing, dating and validation; 3.1.1 IRH tracing; 3.1.2 IRH uncertainty; 3.1.3 IRH ages and age uncertainty; 3.2 Tracked IRHs; 3.2.1 IRH over WSB; 3.2.2 IRH over DML
4. Data-model comparison; 4.1 The 90 Ka isochrone; 4.2 Model Setup; 4.2 Data–model comparison results
5. Conclusions

The restructured section 2 and 3 cover the data acquisition/tracking/uncertainty/description while Section 4 contains the modelling exercise. Making this distinction clear to readers will help a lot.

• The last paragraph of the introduction, should also be updated according to the restructure and prepare the reader for what is to come.
• Also, add a table (or two tables one for each region) listing the IRHs, age, age uncertainty, dated using which ice core, depth at the core, coverage (km), and maybe highlighting the 90 ka IRH.
• Fig 2 is very busy. And only 2 of the 16 are relevant for modelling section. I suggest splitting it into two figures (2 and 3) one for each region. For each of the regions, make the 90ka subfigure significantly larger than the rest and label the ice cores (similar to figure 1).The coverage of the regions have slightly changed from figure 1, so please include an inset showing the coverage.
• Fig1: The caption clarifies that “..throughout the paper, grid-west, grid-east, and grid-south are defined relative to the map grid orientation” but it is still difficult to follow the text. For e.g. in L268: “central grid-west to grid-east portion”, where is the centre of the WSB grid? Is it the centre of the map as the caption says the directions are relative to the map? Please consider updating the text and/or the figure, to make this read better. Maybe putting in the Lat Lon coordinates in parenthesis in the text will help guide the reader.

• In section 2.1.2, the paper emphasizes the importance of survey design for ice-sheet modeling, noting that DML was specifically designed for modeling applications while WSB was acquired primarily for ice thickness mapping. However, it doesn’t discuss if the newly acquired DML data work better for model constraints than WSB, given the differences. If one looks at the RMSE values, it seems that DML actually performed worse than WSB? Is this all because of the missing mountain in BEDMAP3? I think survey design implications should also be discussed towards the end of the paper.

• L388: The manuscript uses the newly acquired radar data over DML to identify mismatches between modelled and observed isochrones, attributing these discrepancies to inaccurate Bedmap3 bed interpolation. However, this raises two questions: (1) Since the new radar data also includes bed picks, couldn’t the bed elevation mismatch be identified more directly by comparing the new bed data with Bedmap3, rather than inferring it indirectly through IRH-model mismatches? (2) If newly acquired IRH data is being used to constrain the model, shouldn’t the bed elevation boundary conditions also be updated with bed data from the same radar survey? Using new IRH data while retaining outdated/interpolated bed elevations creates an internal inconsistency. I think the primary takeaway from this exercise is that we need more radar data. In most cases, wherever we have such deep IRHs we will also have accompanying bed elevation data.

Minor comments:

• Title: The title is super generic. It also misses out on mentioning the data release.
• L4: where englacial layering is more disrupted: add “typically” or similar after “is”
• L9: “imprint of” should be followed by “on” at some later part of the sentence. After reading the manuscript I understand the intention, but it’s not very easy to grasp with this sentence. Try rephrasing.
• L13: The last sentence of the abstract is unnecessary considering its location. It is describing the nature of the study, which someone who has already read the abstract must have understood. Either move it to an early part of the abstract or remove it or replace it with a more conclusive/forward looking statement.
• L31: Could possibly also refer to Cavitte et al., 2023
• L34: You could also mention that this bias in studies towards the divide is because the ice divides are typically better suited for drilling ice cores. And we need ice cores to use IRH as isochrones. Obvious but worth mentioning.
• L45: “yet to be fully evaluated” So they have been evaluated but not “fully”? Confusing. Please rephrase it.
• L47: remove “our study”. It seems odd here as nothing has been mentioned about the study at this point (except that it lies at the intersections of…).
• L54: Suggest starting with “IRHs offer a ....” for a more clear stance.
• L58: “can complement” sounds weak. Maybe say “can further constrain” instead.
• L60: Better replace this opening with the objectives stated on the L329.
• L102: surface pick “and elevation data”.
• L107: Here by “obtaining” do you mean “accessing” or “acquiring”?
• L108 Replace “been about obtaining information” with something like “aimed towards mapping the ice thickness..”
• L108 This sentence is way too long (4 lines!). I would suggest rephrasing/splitting.
• L113: What is “ice streaming”?

L113: The discussion of what models need is confusing. The terms "non-interpolated bed data" and "englacial layering along specific radar transects" suggest that gridded surveys are inadequate for modelling, but this is misleading as gridded surveys can also provide better along/across flow bed topography and englacial surfaces. Maybe first clarify, what exactly are the requirements for this specific type of modelling? Then how does a gridded isochronal/bed elevation data won’t meet these expectations? How a more targeted survey as per the model requirements could be better. Please clarify this distinction and explain why flow-oriented surveys may be preferred for ice-sheet model calibration, despite gridded surveys offering better spatial interpolation.

• L121: Remove “providing…modelling”. It is implicit considering the context.
• L126: “?” missing reference?
• L129: replace “further” with “additional”
• L131:  remove “both”
• L174: If I understood correctly, the 2% error in the firn correction is representative of the TALDICE ice core site. However with the assumption of spatially invariant firn correction the uncertainty should be higher accounting for the spatial variability in surface density over the 1000s of kms covered by the IRH. The difference in the firn correction values at the EDC and TALDICE core also suggests this point.
• L238: The manuscript compares the horizontal resolution (8 km) to Sutter et al. (2021) but not the vertical resolution (201 levels). Please clarify: 1) How do the 201 vertical levels compare to Sutter et al. (2021)? 2) Given the reported IRH uncertainties, what model resolution (vertical and horizontal) is appropriate for these constraints? 3) At what resolution would the model become too fine-scale for the IRH data to provide effective calibration? This discussion would help justify the resolution choices and guide future modeling strategies.
• L251: Rephrase/remove: “To showcase the usefulness”. The use of “demonstrate the usefulness” in  L228 is still quite fresh.
• L263: Hard to follow on the map. See comment for Fig 1 above.
• L268: rephrase “points to important lessons learned”.
• L272: Maybe expand a bit on how close the match is.

“an important finding”  I also agree, and this important result could be highlighted elsewhere (conclusions) as well. This will also help balance the conclusions section.

• L312: What does the ice core data used to date this couplet tell us about this suspected volcanic eruption?
• L329: This point should be made clear to the reader early on. I suggest moving these objectives to the introduction.
• L342: Split the sentence as it’s too long. First state the two objectives: (1)use of isochrones to constrain the model in off-divide regions   (2) to differentiate between climate and ice-dynamical processes. Then explain how.
• L362: What is an “ice-dynamical region”? Better say, “two ice-dynamically different regions”.
• L367: “will be” -> “are”
• L369: insert “calibration” after “This”.
• L376: “considerably” -> “considerable”
• L378: How do we know its “over tuning” and not “under tuning”? This manuscript says there is a lack of paleo constraints, and we should use isochrones to further constrain/tune the models.
• L380: insert “only” before “relying”
• The interchangeable use of “profile” and “line” is confusing throughout the manuscript.