Dear Dr Coulon,

Thank you for submitting your replies to the reviewer comments. The online discussion has now been closed, and no further comments can be posted. Both reviewers raised minor comments and questions, which I think you have answered in full. You have also provided good evidence of how these comments will be incorporated in a new version of the manuscript. In particular, the expansion of Appendix A with a discussion on mesh resolution (reviewer #1 and #2) and additional justification for the use of spatially constant viscoelastic bed properties (reviewer #2) will benefit the reader. I invite you to submit a revised versions of your manuscript at your earliest convenience, taking all reviewer comments into account. It would be great if you could also upload a version with tracked changes to expedite further review.

All the best,
Jan De Rydt - editor

Dear Dr Coulon and co-authors,

Thank you for providing an updated manuscript with tracked changes in reply to the reviewer comments. Both reviewers requested minor changes, in particular 1) a justification for the use of 16km grid resolution and some evidence to show accuracy of this approach (R#1 and R#2), and 2) some additional discussion about the possible implications of a constant mantle viscosity (#R2). Whilst I think the second point and all other minor comments have been addressed in the revised manuscript, I was hoping to see a more in-depth discussion on the first point. You argue that the model can reliably simulate grounding line movement based on results for the MISMIP+ experiment. However, these simulations were done with 1, 2 and 4km resolution, which is a much higher resolution than the 16km used here, and results are obtained for a highly idealized geometry. Cornford et al. (2016) argue very strongly that convergence tests should be done on a case by case basis, and ideally I would like to see something done along those lines for your Antarctic setup. Would you be able to do a convergence test for 1 ensemble member of your Antarctica-wide setup with, e.g. 4, 8 and 16km resolution and running for a few 100 years? I appreciate this might be computationally prohibitive, in which case you might be able to demonstrate mesh convergence in a regional setup, e.g. for the Amundsen Sea, which I believe is available in KORI? If that’s also unfeasible, is it possible to include results from simulations with 8km (and 16km?) in the MISOMIP+ plot, with the understanding that this might be very crude given the 80km width of the domain? At the very least, please refer to Cornford et al. (2016) and Reese et al. (2018) in your discussions to raise awareness of the potential pitfalls of using coarse resolution and the grounding line flux condition.

All the best,
Jan De Rydt - editor

References
-------------
Cornford, S., Martin, D., Lee, V., Payne, A., & Ng, E. (2016). Adaptive mesh refinement versus subgrid friction interpolation in simulations of Antarctic ice dynamics. Annals of Glaciology, 57(73), 1-9. doi:10.1017/aog.2016.13

Reese, R., Winkelmann, R., and Gudmundsson, G. H.: Grounding-line flux formula applied as a flux condition in numerical simulations fails for buttressed Antarctic ice streams, The Cryosphere, 12, 3229–3242, https://doi.org/10.5194/tc-12-3229-2018, 2018

Dear Dr Coulon and co-authors,

Thank you for expanding your discussion on the sensitivity of your results to the model resolution, and for including a comparison between simulations with 8km and 16km resolution in Fig S13. I think it makes the important point that the 25% (or 0.2m) difference in SLR by 2300 is well within the spread due to other uncertainties such as the climate forcing scenario and process representation (as shown in your Fig 2). This is a valuable addition to the manuscript, which I think addresses comments by both reviewers on the impact of model resolution on your results, at least in a way that is achievable within the scope and the timeline of this manuscript. I'm therefore happy to accept your paper in its current form for publication in The Cryosphere.

Thank you very much for your timely handling of the review process and I would also like to extend my thanks to both reviewers for their constructive contributions.

All the best,
Jan De Rydt