General comment

This study uses a regional atmospheric model to investigate the impact of extreme ENSO events on the surface mass balance (SMB) of the Antarctic ice sheet. The detailed SMB change in many subregions and catchments is described quantitatively about the simulated results. While such information is of great value, the robustness of the conclusion is rather limited due to the smallness of the sample size (as the authors are well aware). More specifically, the conclusion drawn from the composite analysis of only three samples (for El Nino) is not highly convincing. Such a weakness of statistical power can be augmented by revealing the mechanism: why each ENSO event has a differing impact. Is the SMB difference between cases due to the difference in ENSO itself (and its teleconnection) or the effect of other internal variabilities near Antarctica? Neither the effect of other internal variabilities nor the difference of each ENSO event is analyzed in depth. I think the analysis needs to be strengthened to provide more robust insight into interpreting the results.

Specific comment

1. Abstract, L.8-9: It is unclear how much of the observed anomaly is due to the ENSO during 2015/16. Other modes of variability can affect the SMB, and thus the result may not necessarily represent the ENSO response alone. Therefore, I am not sure if the statement of L.11-12 has its evidence within the manuscript.
2. Conclusion, L.413-414: I did not understand that “SMB changes differ greatly between the ENSO events” on the one hand and “numerous catchments exhibit similar SMB responses when comparing the impacts of extreme and moderate ENSO events” on the other. How are they consistent? Does this indicate that the observed feature does not represent ENSO response? It is misleading to conclude that the SMB response (signal) to extreme and moderate ENSO events is similar if noise and noise are compared (or if the signals are embedded by noise due to the small sample size).
3. Without the presented mechanism, it is unclear whether the differences between ENSO events are due to the diversity of ENSO itself or whether they are affected by other internal modes of variability. It would be helpful to present wave trains from the tropics to the Antarctic so that the link to each ENSO event becomes more visible. Then, whether the observed difference in the SMB anomaly arises from the difference in ENSO influences or the local variability may become clearer. There might be other ways to distinguish.
4. In Fig. 3, the SLP anomaly appears different for each extreme ENSO. It would be helpful to show moisture fluxes and explain why the SMB anomaly in Enderby Land is the same for all three cases despite different SLP responses. It was only explained by noting that the location is in the center of the action.
5. It is unclear what the benefit of using the regional model is. If the authors draw Figs 3 and 4 from the ERA5 dataset, are they very different? The authors should state the advantages of using the regional model. In particular, the resolutions of both models are not so much different (0.25 degrees vs. 27 km), and the SMB seems to be controlled primarily by moisture transport specified by the boundary conditions.
6. Figures 3 (and 4): Please stress the meaning of statistical significance here. How should one interpret the statistical significance of a single event when each ENSO-induced SMB is so different (including signs in some places) from the others? Or note that the anomaly is significantly different from the baseline but it does not necessarily represent the ENSO response (alone).

Macha et al., 2024 investigates the effect of extreme ENSO events on the Antarctic surface mass balance using a RCM (RACMO). They found that Antarctic surface mass balance generally does not vary significantly during most extreme ENSO events, though regional differences exist especially over Enderby land (increase in snowfall and SMB); or during the extreme 2015/15 El Niño events.



Overall, I find the text well supported by the results where the authors also use robust statistical tests. The figures are clear, and personally I find figure 5 very well done. In this type of study, it's difficult to balance between an overly detailed analysis by region, which risks losing readers, or something too simple. I sometimes had a little trouble remembering the most important points, even if the balance seems good. I would encourage the authors to summarize the main results at the end of each section to help the reader follow the flow of the manuscript; (and if the results were more coherent for each event, to make a summary graph or table).

Although the authors point this out, I think that the too-small number of events qualifying as extreme reduces the robustness of the analysis. Perhaps it would be possible to use a wider range based on SST distribution percentiles, by varying the threshold defining an extreme event. It might also be the opportunity of using something more statistically-robust than arbitrary of +2 and then -1.5 “because -2 is too limiting for El-Nina events”.

I also had to reread several times to understand why the manuscript concentrates essentially on the SON period, while all the other periods are somewhat left to the reader's analysis. Even in the supplement, I'd suggest that the authors add a few explanatory lines on what they think is important for the other periods.

Finally, I have little to reproach this study. The method seems robust to me, the paper remains quite understandable and it's a good piece of work. The only rather annoying limitation comes more from the lack of events, and I would really encourage the authors to try to increase the samples in one way or another; while taking advantage of the review to further improve the clarity of the paper.

Best regards,

C. Kittel

PS: there are two missing spaces L47 and 49.