I read with interest the article entitled “Subduction plate interface shear stress associated with rapid subduction at deep slow earthquake depths: example from the Sanbagawa belt, Southwest Japan” by Koyama and co-authors.

Using quartz-rich samples the authors provide new data on shear stress at depths of 17-27 km during subduction and discuss those in the regional framework, with a few implications on subduction-zone modelling.

Even if I think that the article may be a relevant and novel contribution for this topic, I think that major revisions are needed.

Below I state my main comments, supported by all the specific ones in the attached PDF file (so please check that file carefully).

All the study is conducted on quartz microstructures, as also stated “we focused on the microstructure of quartz-rich metamorphic rocks—quartz is the main component of the rocks we collected and its deformation stress is assumed to be representative of the region”, and those results are extrapolated to be relevant for the unit. However, as presented in the geological setting (methods section), the unit is composed by several rock types with different compositions. Additionally, the quartz-rich metasediments are composed of abundant mica. Even if a brief part of the discussion considers this point, I think that it needs much more attention. In this respect, the authors should also expand their citations, using relevant literature discussing deformation mechanisms in quartz-rich metasediments in subduction in other subduction zones, such as Trepmann & Seybold, 2019 (https://doi.org/10.1016/j.gsf.2018.05.002), Condit et al., 2022 (https://doi.org/10.1029/2021GC010194), Tulley et al., 2020 (DOI: 10.1126/sciadv.aba1529), Giuntoli et al. 2022 ( https://doi.org/10.1029/2022JB024265). Among others, those articles' results and implications should be discussed. In particular, the role of phyllosilicates in the bulk deformation needs much more attention, also expanding on phase mixing between quartz and phengite (see for example Hunter et al., 2016 http://dx.doi.org/10.1016/j.jsg.2015.12.005). The discussion should be expanded in this regard. And what about the role of ultramafic rocks? These are not discussed, yet present in the unit. Along this line of thoughts, I think that the discussion needs to be expanded considering the results obtained by similar studies conducted on other orogens (differential strain rates and deformation mechanisms related to deep slow earthquakes). Regarding the latter point, the discussion states only “Therefore, the estimated stress may represent the initial conditions from which slow earthquakes in the same domain nucleated”.

More geological context is needed (see specific comments in the attached PDF), in particular for the reader to picture the relation between the different rock types and the relation between minerals marking the fabrics. Additionally, could you add a figure with field photos (e.g. where these samples were collected, main structures,..).

Finally, as EBSD was performed, please also show pole figures for the <a> axis for all analysed samples and EBSD maps, such as grain size maps, KAM maps, IPF maps. This is to improve documentation and to support your interpretation of deformation mechanisms and grain size used for piezometry.

I hope to have provided constructive comments to make the newer version of the manuscript stronger.

Bests

The authors present data on absolute maximum shear stress based on estimated differential stress in warm subduction zones. To do so they measured quartz grain sizes and applied various piezometers to obtain the differential stress. Stresses obtained are homogeneous over the region and thus suggest that the shear stress in subduction zones is independent of pressure and temperature.

Overall, this manuscript is nicely written and I enjoyed reading it. The data presented are novel and relevant for subduction zone modelling as well as the understanding of slow earthquakes. However, the manuscript would benefit from expanding the discussion prior to publication.

I hope these comments help to further improve the manuscript.

General comments:

The manuscript carefully distinguishes various ductile deformation stages and then focuses on the main deformation stage. This main stage, however, represents early (and in one case late) exhumation. My concern is to what extent shear stress during exhumation can be applied to rapid subduction as the title suggests.

The authors focus only on the quartz-rich regions. As outlined in the geological setting, the rocks are highly heterogeneous. The authors shortly address the fact that ultramafic bodies are minor and can be neglected. Even if so, figure 3 clearly shows that the quartz shists are not the major lithology and that they are intercalated by pelitic and mafic shists. Such heterogeneities can cause stress concentration and result in larger scale stress gradients. Expanding the discussion in this direction as well as discussing relevant literature is needed.

Furthermore, heterogeneities also occur on a micro scale. The piezometers were applied to quartz-only domains. The authors argue that the presence of sheet silicates inhibits grain growth and might cause wrong estimates on differential stresses. The authors argue further that sheet silicates do not form a network. However, in figure 6a it seems the sheet silicates form a continuous layer. Again, such heterogeneities can cause stress gradients. It would be interesting to see how much variation in shear stress is obtained between quartz-only domains and more heterogenous domains. And if significant these uncertainties should be included into the discussion. Knowing that additional measurements need time and effort, I think the manuscript would already benefit if these points were addressed theoretically in the discussion.

Minor comments:

• Line 70: “Shear stress is equal to half the differential stress.” Only the maximum shear stress is equal to half the differential stress. Indeed, on line 37 the author write that shear stress is used for absolute maximum shear stress. I would suggest to strictly write maximum shear stress. The data presented are estimates on the maximum shear stress and for the discussion it is crucial to use accurate terms.
• Figure 1: the unit boundary of the smaller eclogite units is hardly distinguishable from small ultramafic bodies. I suggest using different colors for the boundary and the ultramafic bodies. (Actually, the color for ultramafic bodies in figure 3 is different)
• Figure 2c: Can you add PT values here? Or otherwise plot the ductile deformation stages in 2a.
• Figure 4: Can you also add pole figures?
• Table 3: The Cr+Ho data are a based on a corrected version of the Cross et al. piezometer after Holyoke et al. This is only mentioned in the discussion part. Please add some details also in the method section for better understanding of the present table.