"Southern Ocean polynyas and dense water formation in a high-reslution, coupled Earth System Model," by Jeong et al.

Over the past decade, general circulation models (GCMs) have significantly improved in terms of resolution and parameterizations of sub-grid scale processes. However, the bottom-water formation on the Antarctic shelves remains a key challenge due to small and local formation sites related to coastal polynyas. The study by Jeong et al. contributes to the study of coastal polynyas and the formation of dense/bottom water in GCMs. The authors use a high-resolution coupled Energy Exascale Earth System Model (E3SM-HR) with a horizontal resolution of ~8 km over the Antarctic continental shelves to study dense water formation south of 60 ^oS and compare the results with a lower-resolution version of the same model and available gridded datasets based on observations and reanalysis. They find that the increased resolution improves the representation of coastal polynyas. However, the associated bottom-water production is too weak and does not produce sufficiently dense bottom water to compare with observations.

The manuscript is well written and includes high-quality figures to support the main results. The methodology is sound and builds on state-of-the-art coupled Earth System Model (ESM) development. However, the discussion section is too brief and offers limited new insight into the main topic of missing dense-water formation in the coastal polynyas in GCMs and ESMs. I summarize key aspects the authors should address to improve the manuscript and its contribution to the ESM development near the Antarctic shelves:

• Increasing the resolution of the E3SM improved the representation of coastal polynyas, but the associated dense-water formation was too weak and did not produce dense water with similar characteristics as the observations. What is your advice to the ESM community to further improve the polynya representation?
• You show that the winds over the continental shelves are too strong, resulting in a too strong Antarctic Slope Front which prevents shelf-ocean water exchange. What means can you do to get a better representation of the winds over the Antarctic shelves? Would it be possible to run the E3SM-HR with an idealized wind forcing near Antarctica to test the response to strong and weak wind scenarios? Including such a test in the manuscript would strongly increase the relevance of this study to the modeling community.
• You describe plans to include landfast ice to improve the model further. How will the representation of landfast ice improve the model if the winds remain unchanged?
• The Southern Ocean is an area of upwelling. How well is the E3SM-HR representing the water mass characteristics in the Southern Ocean, and how are these characteristics affecting the formation of open-ocean polynyas?
• The E3SM-HR does not include ice shelf cavities where dense water interacts with the ice shelf base to form the densest versions of shelf water (Ice shelf water). How are the missing ice shelves and the presence of ISW affecting the model performance?
• The modeled sea ice is much thinner than the observations in OOP years over the whole WS. How is this affecting the dense water formation and the likelihood of increasing the lifespan of an OOP?
• Are there any caveats in comparing the model with the SOSE database? Could it make sense to also compare the transects in Figures 7-9 with CTD sections?

Minor comments:

Line 159: It is very hard to see the “relatively higher latent heat flux…in East-Antarctica” in Fig1a.

Line 256: It is very hard to see the easterly winds in figure 6b.

Southern Ocean polynyas and dense water formation in a high-resolution, coupled Earth System Model 

Hyein Jeong et al.

The authors investigate dense water and polynya formation in two versions of the Energy Exascale Earth System Model. The manuscript is clearly structured and the results are well presented. Some of the arguments don't seem to be supported by the material and the results need to be expanded upon.

Main comments here:

• Could you give recommendations for development? What are the key improvements needed here (easterlies and Ekman transport)?
• You mention the impact of the overly strong polar easterlies and associated Ekman transport throughout the paper and abstract. However - you do not seem to directly calculate the Ekman transport. Can you do this and compare to observations?
• The mean-state open ocean stratification is important for the formation of open ocean polynya in models. What does the open ocean stratification look like in these models compared to observations? Is it overly weak (explaining the convection in HR model) or overly strong (explaining no convection in the LR model)?
• You mention that the model is fully coupled. The ocean-ice interaction is very important for DSW and presumably the coastal polynya development. Please include a discussion of this.

Minor comments:

Line 18: Are you referring to katabatic winds? If so, please introduce term as you use it again later.

Line 20: Are coastal polynya important for other aspects of the earth system e.g., marine biology or biogeochemical cycles? 

Line 20: You may introduce that these are areas of high sea ice production. Also, coastal polynya may form due to oceanic currents.

Line 46: Could add citation to On the Role of the Antarctic Slope Front on the Occurrence of the Weddell Sea Polynya under Climate Change.

Line 55: You may introduce the two types (coastal and open) of polynya in paragraph one.

Line 134: You can have open ocean convection that doesn’t form polynya in models (e.g., Dufour et al 2017 and Lockwood et al 2020).

Line 181: Moved westward with an average velocity of  0.013 m s−1 (Gordon 1978, 1982).

Line 193: Episodic open ocean deep convection events in the Weddell Sea have been linked to anomalies in the Southern Annular Mode index (Gordon et al. 2007; Cheon et al. 2014; Francis et al.  2019; Campbell et al. 2019; Cheon and Gordon 2019). Have you considered the representation the SAM in these models?

Line 209: You can have convection and dense water formation without polynya formation (see Dufour et al. 2015 and Lockwood et al. 2020). Please check if the LR model is infact creating dense water and convection, just without polynya formation. Convection can be calculated via. the mixed layer depth (see de Lavenge et al. 2015)..

Lines 258: Although you're correctly taking the transects from the respective areas following Thompson et al 2018 - the Western and Eastern Weddell Sea transects seem very close together. I'd like to see how this model holds in another region of Fresh Shelf (e.g., along the Ross sea). Also, can the LR model capture the hydrography?

Line 295: Could you produce depth average velocities around the full Antarctic ?

Line 325: Have you considered comparing to very high resolution simulations like the MITGCM or LCM-4320?

Figure 6. c-d The vectors are difficult to see in the figures.