The article discuss a new model for the Southern Ocean named BIOPERIANT12. Even if the scientific knowledge of the article is good the introduction needs to be re-written considering that is is now just one paragraph. I would also suggest to put the references in a different color.

-I think the initial condition section needs to be expanded

-In line 122, can you references where the numbers come from ?

-In line 232, do you think if you would run the model for longer than 10 years it would be less stable? 

-In 425, in the article you have used two phytoplankton classes. IS there a reason why you choose only two classes?

Overall the manuscript gives a valuable advance in modelling the Southern Ocean but the author needs to improve the writing before I can endorse.

The manuscript does a good job at describing a relatively high-resolution circulation, sea ice and biogeochemistry model for the Southern Ocean and evaluating its strengths and weaknesses. The authors describe the difficulties of properly representing the complex dynamics of the SO and of validating the results with limited available data often biased towards warmer months, and describe biases found by other studies.

One of my main concerns with the manuscript is the description of its purpose. The conclusion seems to focus on the limitations of the model and on improvements needed, while arguing that it this novel setup can be used for future studies. This section would benefit from a comparison of what this model improves relative to other SO models, and of what type of studies it would be suited for. Even if this document is just to document the model evaluation, a more in-depth discussion of solutions to current biases and mismatches with observations, as well as recommendations for observations would add value to the manuscript.

Some more pointed suggestions are highlighted below.

L53: “run duration which relies on periods of sufficient observations” seems vague

L54-58: Sentence is too long and a bit confusing

L60-61: matters more for simulations of hundreds of years, but does it matter at this scale? If not, why bring this up?

L75-76: “although more recent versions were available” – for which components?

L79: what is the resolution of the eddy-permitting?

L96-97: maybe add some references for how well inputs extracted from ORCA05 are represented in that run?

L103-104: I think you could expand a bit here. What biases were observed? My initial thought was that using a “normal year” would also introduce errors, so why one versus the other?

L116: Is this shown somewhere, to inform users?

L147-148: Maybe a sentence on how mean transport was calculated?

L163: It’s hard to see where 36-43˚S is, would be helpful to add latitudes to the maps

L169: Do you mean higher resolution in the model compared to observations?

L179-183: Sentence is too long

L185: Do you mean monthly mean for the 2000-2009 period?

L200-204: I’m confused about the message on the final sentences. Does the variability of the front position complicate the analysis or improve bgc representation?

Figure 3: grey shadings are all the same, and should specify which satellite

L238: area-weighing of which dataset? WOA?

L239-240: Reasoning is not super clear to me.

Section 3.1.5: could the negative SST bias and positive sub-surface temperature bias be influencing MLD?

L265-267: Need more information on the calculations

L271: What does SP-STPS, SA-STPS and IND-STPS stand for?

L325: would be useful to have a more in-depth description of the data product for readers to interpret the model-data differences

L239-243: Might want to rephrase it. At some point it becomes unclear if you are talking about ICE or the whole domain

L356: Avoid repeating thus?

L363-365: Doesn’t that depend on dFE sources, since it will not stay in the surface long? Are you talking about biases in the initialization files? I would assume this has more to do with model setup than with the choice of model

L381-381: respectively, after the values?

L382: How is the representation of simulated silicate dictated by laboratory experiments? This sentence was a bit confusing to me.

Figure 11: There are substantial differences in the summer pattern that need to be discussed

L396-397: wouldn’t the overestimation be true regardless of biome classification?

Section 3.4.3: How good is the satellite temporal coverage in the different biomes? Is it properly representing intra seasonal to interannual variability? Discussing potential biases in the observation might help the discussion here