General Comment:

This is a very well written manuscript about an improved and high resolution configuration of a coupled Ocean-Sea ice-Ice shelf model over the southern pole. The authors clearly explain the configuration of the model, and thoroughly evaluate several physical aspects against observational datasets.

Specific Comments:

1. Sentence at Lines 183-185. An appropriate reference supporting this statement would be helpful.

2. In Lines 262-264, the authors indicate that deeper than 5500m, they obtain the initial conditions by downward extrapolation, but the exact method is not described. Please clarify how this extrapolation is performed and what the effects of this choice are in the circulation of the model.

3. For Figure 5 and paragraph starting at Line 392. The map shows the three sectors, but the exact longitudes are not defined. Please specify the longitude ranges for each sector.

4. Paragraph starting at Line 849. A possible explanation for the weak positive correlation with AVISO observations for surface MKE might be the large discrepancy between the resolution of SOSIM and AVISO, a byproduct of the interpolation from the low resolution of AVISO to match the higher resolution of SOSIM. One would expect that the features would be broader in AVISO than in SOSIM, leading to areas with higher values in AVISO than SOSIM (the area below the 1:1 line in Figure 16c). And the intrinsically higher resolution of SOSIM will lead to larger values of MKE in areas with low and moderate values of MKE in AVISO (area above the 1:1 line in the same figure).

5. Section 3.7. Sea Ice Production (SIP) (for March-October) in the Atlantic sector (at 45W and between 60-65S) is smaller than observation (HU) (around -1 m/yr, Figure 25b). At the same time, the Sea Ice Thickness (SIT) is considerably larger in the same area in September (+0.2-0.4 m, Figure 24e) (albeit a bit smaller in February at between 0 and -0.2 m). How are these things reconciled?

6. Paragraph starting at Line 1245 and Figure 26d. How exactly were the statistical metrics (e.g., correlation and RMSE) calculated in this case of the area-integrated ISMRF within each sector over the continental shelf?

7. Paragraphs starting at Line 1267 and Line 812: Regarding the significant underestimation of ACC transport. Some possible explanations are given in the manuscript. Are there additional diagnostics that can be used to narrow down even more the possible causes?

Technical Comments:

1.Lines 614-615: This sentence is somewhat difficult to follow due to the “and while”. Please consider revising the sentence to improve the clarity.

This manuscript presents a valuable new circum-Antarctic coupled ocean–sea ice–ice shelf configuration based on MITgcm and documents it through an extensive evaluation against a wide range of observational and reanalysis products. The configuration itself is a substantial technical achievement, combining relatively fine horizontal resolution, explicit ice-shelf cavities, and a long hindcast forced by ERA5 over 1979–2022. I also appreciate the authors’ transparency in discussing remaining biases and limitations. Overall, I see this work as a strong contribution that will likely be of broad interest to the Southern Ocean and Antarctic modeling communities.

In my view, the manuscript is already in good shape scientifically, but it would benefit from a to help clarify the model’s novelty, better contextualize some of the remaining biases, and strengthen the discussion of where the configuration is especially robust versus where caution is still warranted. 

I want to emphasize that one of the best features of the manuscript is its transparency. The authors do not oversell the model and openly discuss important remaining issues such as the warm abyssal drift, sea-ice seasonal bias, and underestimated total ice-shelf melt. That gives the paper credibility.

Rather than changing the underlying results, I think the manuscript would benefit mainly from reorganizing some of this discussion so that readers can more easily distinguish:

• where the model performs particularly well,

• where biases are moderate but acceptable for many applications,

• where caution is needed.

A summary figure or table listing strengths, limitations, likely causes, and implications for model use would be extremely helpful.

Moreover, the use of repeated 1979 forcing for spin-up is a reasonable practical choice, especially given the desire to begin the hindcast consistently in 1979. At the same time, the manuscript shows that some domain-integrated drifts remain, especially in deeper waters and in the early evolution of kinetic energy. I do not see this as undermining the value of the model, but I do think the paper would benefit from a slightly fuller discussion of what this means for interpretation. In particular, it would help readers if the authors could state more explicitly that:

• the deep ocean remains less equilibrated than the shelf regions,

• some large-scale integral diagnostics should be interpreted cautiously,

• shelf-focused applications may be more robust than basin-scale abyssal trend analyses.

That would strengthen the manuscript considerably.

A third point is the topography workflow. The geometry preparation appears to have required careful expert judgment, especially in handling remapping artifacts and problematic isolated cells. That is completely understandable in a configuration of this complexity. Still, because this paper will likely serve as a reference for future users, it would be valuable to document these steps a bit more fully, perhaps in the supplement. A brief workflow description, a mask of edited cells, or a short note on the scale of the edits would likely be enough. This would make an already useful paper even more reusable by others.

A final comment concerns the underestimated ice-shelf melt. The manuscript’s treatment of ice-shelf melt is appropriately cautious, and the underestimation relative to observational products is clearly shown, with the omission of tides identified as one likely explanation. I would encourage the authors to frame this more explicitly as a priority avenue for future model development. This would leave readers with a clearer sense not only of the current model performance, but also of the most promising next steps.

Overall, I consider this manuscript to be a strong and worthwhile contribution. The configuration is impressive, the evaluation is extensive, and the manuscript is strengthened by the authors’ transparent discussion of remaining biases. The revisions I suggest are primarily aimed at sharpening the interpretation and improving the utility of the paper as a reference for the wider community.