The paper by Klose et al. address the multi-millennial sea level commitment of the Antarctic ice sheet using two different ice sheet models forced by a set of four coupled climate simulations from the CMIP6 initiative. An anomaly method is used to create the climate forcing based on regional atmospheric models MAR and RACMO on which GCM future climate is added. To estimate the long-term sea level commitment corresponding to different level of global mean atmospheric warming, the simulations are branched off at different moments with the next two centuries and climatic conditions are maintained constant at their branching-off level.

The scientific content of the manuscript is good and very interesting. This is something that is needed for different reasons: testing the physics of ice sheet models and parameterizations, pushing to obtain ling-term multi-centennial climate forcing etc… I really liked the manuscript. However I can feel that this is perhaps the first or on of the first article written by the first author here: the writing of the manuscript needs some substantial work to be clear an readable. Information are some times provided in a very messy way, spread out in different sub sections etc…In addition I feel that the description of the results is sometimes approximative and also messy amongst the two ice sheet models. The authors should consider describing everything in depth with one model and then describing the discrepancies with the other model. The discussion (from Uncertainties to boundary conditions) is also chaotic and does not allow the reader to really appreciate the real advance of the work. I below provide some generic comments, but most of the specific comments can be found in the attached commented pdf version of the main manuscript.

General comments:

The description of the result is too “descriptive” and many times, no real explanation is provided for some observed behavior, or really little. Some of them are explained further int eh discussion, some others not at all. The results are sometimes described in a very approximative way. Sentences are sometime useless because not bringing any substantial info. In general try to group the info related to one topic or one model together. Right now, the reader needs to jump from one paragraph to another to synthesise all the info about one process or one model.

I generally find the climate analysis a bit weak, given that the paper looks at sea level commitment. I would have expected a bit more climate analysis to really show the relationship between the different steps of the retreats and the competition between atmospheric warming induced melting and basal melting from oceanic warming. For example Figure S3 to my opinion should be inserted within the main manuscript and with two additional panels showing atmospheric warming and oceanic warming evolution through time. Although it is a bit complicated for the oceanic warming since it depends very much on the sector of Antarctica. 

We never see one figure of climate forcing and this is instead very important since the climate forcing here plays a critical role in all the results. Thus I expect to see a bit more of climate in terms of figures and forcing description. That will allow the authors and the reader to better described and understand the results.

There is no real description of the outcomes of the initialization. There is only one sentence stating that ice sheet models reproduce correctly the AIS geometry- Which is not true, since PISM is far from having the GL in the right place, especially for the big ice shelves that are extensively then discussed in the rest of the manuscript. As stated in the discussion, PISM large sensitivity and large retreat of the WAIS is likely due to the already retreated grounding line. PISM performance is not very good in general because the final elevation differs quite a lot from the observed one and the discrepancies fall in the range of what is observed int terms of elevation changes by satellites. In addition there is no description of the historical run at all. This is also an important part to be added.

To really appreciate the full description of the results, you also need to introduce a map, as first Figure of the manuscript show the different drainage basins, ice shelves and related names. You only show this in Fig3a, and honestly, it is so hard to understand even when printed.

The figures are too dense, e.g. Fig 2., it would benefit from separating each ice sheet model simulations in different frames to better appreciate the difference and understand them.

The section about “Intra and inter-models uncertainties” is not useful in its present form. It would be better to divide it in several sections. Eg: “Initialization” (actually once again, the info about the impact of initialization are spread out through the section and is thus a bit messy), “model physics”, etc… it would help organising a bit this part.

Specific comments: see the attached pdf.

Best,

Florence