see attached PDF

The authors used a fully coupled MPI-ESM to study soil hydro-thermodynamics in permafrost regions. They examined the effects of soil hydro-thermal and depth on soil thermodynamics and permafrost distribution by comparing the default model setup (REF) and two extreme hydrological setups (WET and DRY), as well as the soil discretization setup (5 and 11 layers with approximately 10 m and 18 layers with 1391 m). The results suggest that the wide range of permafrost extent in CMIP6 is associated with soil depth and hydrological conditions. This well-designed paper falls within the scope of The Cryosphere and may be of interest to readers in the permafrost and modeling communities.

The manuscript is rather lengthy and could be written more concisely. There are also 12 figures that could be eliminated. Additionally, some sections, especially the introduction and methods, could be rearranged for clarity. For example, the simulation setups are introduced in the introduction, beginning at L85. This information could be moved to the methods section. The methods section also contains several model abbreviations, such as MPI-ESM1.2, JSBACH3.2, JSBACH-HTCp, and MPIESM-PePE. This section should be simplified to avoid confusion. If the study used MPIESM-PePE with the JSBACH-HTCp land component, then MPI-ESM1.2 and JSBACH3.2 could be introduced first, and followed by an explanation of the improved features of JSBACH-HTCp and MPIESM-PePE together with the experimental setup. Alternatively, a table could be added with the names and brief descriptions of the versions used for the simulations. Furthermore, the results and discussion sections sometimes contain a lot of values that could be summarized in a table, for example, the section 3.2.

L150: "Standard physics" sounds vague.

L258, L300: If the snow schemes differ between the WET/DRY (JSBACH-HTCp) and REF (JSBACH3.2) simulations, this should be mentioned in the methods section. Currently, it is unclear what differs between the WET/DRY and REF simulations. It should clearly describe how they differ, not only in the hydrological scheme (extreme wet and dry), but also in other schemes, such as the presence of an organic layer and water phase changes.

L281: Increased soil depths (5 and 11 layers versus 18 layers) reduce the warming trend. However, REF18 appears warmer than REF5 and REF11, which may be due to their greater responsiveness to a cooler scenario (SSP1) compared to REF18. It would be better to describe that the deeper layers (18 layers) are less responsive to temperature change.

L290: For the SSP5-8.5 scenario? Figures 4d and 4f show only temporal changes.

L347 (Fig. 7): What does "18L value is out of bounds" mean? Are there no data beyond this range? For the 18L simulations, can the maximum ALT be 9.98 m even though the soil layers exist deeper than that?

L396: What could cause these spatial variations?

L404 (Fig. 10): The REF and WET simulations show a similar range, while the DRY simulations show slightly less permafrost extent, closer to the ESApCCIv3 value. Does this mean that the DRY simulations better represent the real world? If the REF and WET simulations overestimate the extent of permafrost, does that imply that current models (e.g., the REF simulation setup) are wetter than the real world?

L455: Again, what could be causing these spatial variations?

Dear authors, please find my comments in the attachment.