Kolås and colleagues provide a detailed analysis of the oceanographic characteristics of the polar front in the northwestern Barents Sea (BS). The paper uses a comprehensive observational dataset (a subset of the companion paper submitted to JRL, as I understand it) obtained as part of the Nansen Legacy project with traditional and autonomous platforms (gliders), which is undeniably more detailed in terms of spatio-temporal variability, but also more complete in terms of physical variables than any previous dataset. To this end, the authors make a very valuable and serious contribution to the community. In addition, the writing and overall form are excellent and well streamlined, making life really easy for the reader.

This study is rather technical (which is fine!) and I find that the authors could discuss implications slightly more. The BS has long been thought a "hot spot" for marine productivity.... Due to strong vertical and/or cross-frontal mixing. However, this study confirms that the front is a place of "moderate turbulent mixing" (mainly at the surface and at the bottom) and that subsurface mixing occurs more along the isopycnals. And in fact, to the best of my knowledge, the front was never demonstrated as a significantly more productive area than the rest of the BS. The area studied is the only northern gateway to the western Barents Sea. As the front extends to the east, we might expect the same effect to occur here in the West. Interestingly, this is not the case seasonally, and from one year to another. The implications, both in terms of biogeochemical tracer and heat, are that the PF front acts more as a barrier to the AW domain (rather than a mixing machine) and to the ongoing "Atlantification". To go beyond, the AW have to subduct, transporting heat, salt and… carbon along isopycnals... which could eventually be sequestered further in the Nansen Basin.

Another interesting aspect of this study is that it provides the first very interesting evidence of the baroclinic structure of FP. This structure could favor baroclinic instabilities. So I can see why the authors took the time to study eddies (although they were created elsewhere). These vortices could temporarily tilt the isopycnals even further and provide additional mixing. It's fascinating that this physical feature has been studied for decades and still holds so many mysteries.

I found no major problems with the manuscript, which I recommend for direct publication. I recommend only very minor edits/additions for which (I think) it is not necessary to send the manuscript to the reviewers again:

• Please specify somewhere that altimetry provides only surface geostrophic velocities. No need extra work but you could provide existing (very few) studies that evaluated those products for quality control in the region.
• Figure 2: please provide brief information in the caption about transects A, D, F so that the reader so remains in the blue until Figure 9.
• Figures 3 and 4: curious arrangement… panel labels at the bottom, colorbars squeezed in the middle. It would help to pop out the colorbars. Panels are un-scaled, you can notice through the y-axis which should be the same everywhere. This is okay and probably complex to fix, but I recommend to improve the visual. I would just use the same “frame” and leave blank where there is no data (no obligation, you probably tried already, just my suggestion).
• Line 327, just say add “following” section, it helps to grasp the nice flow of the article. Nice transitions.
• Figure 6: I recommend to draw some box, or arrow to describe the eddy position, structure, etc… because from isopycnals it looks like 2 eddies which are merging and it creates a bit of confusion when reading.
• In the discussion I would have liked a word on the intensity of the density gradient.
• I would also like a comparison of magnitude of the dissipation rate with previous estimates and/or close-by regions, besides Fer and Drinkwater (2014).
• If the structure of the PF is baroclinic, then the use of altimetry must be at least questioned.

Review of “The Polar Front in the northwestern Barents Sea: structure, variability and mixing” by Kolås et al.

The authors provide a nice description of the structure of the Polar Front in the Barents Sea based on extensive ship board data (a fall and a winter cruise) as well as data from several gliders. Even though this region has been extensively studied, the study provides some novel aspects of the dynamics of the front, in particular that it is baroclinic at the sill, unlike its barotropic nature encountered further south. The paper is well-written, easy to read and does not suffer from any major issues in the methodology or interpretations. On several occasions during the read I however asked myself “So what?”. Where possible it might be nice for the authors to clarify the motivation for what they are doing or what the implications of their findings could be. That being said, the baroclinic nature as well as the level of mixing at the front (weaker than at the surface and at the bottom, but quite large for the mid-water column!) are important additions to the literature that are well placed in “Ocean Science”. Consequently, I can recommend this manuscript for publication after minor revisions; the minor points below do not warrant a second round at the reviewers.

Minor points:

Fig. 1a This figure does not provide the important information. Most of the topographic features are not labelled and the isobaths are very hard to understand/interpret. The sill is not label or marked even though it is the key location of the paper. Fig. 11 of the authors’ JGR preprint is much clearer and I’m wondering why the authors don’t use a modified version of that figure here, in particular the labeling and the water depth as a color scale and not just as contour lines.

L143 Do you mean “southward wind” or “northerly winds”?

L199 It might be nice to explicitly give the equation for C_D as a function of sea ice concentration rather than only referring to the 2005 paper.

L214 You give a reason, but that still leaves the question as to why you decided to do that.

L221 How are L_x and L_z estimated?

L237 Is this recalculation of salinity from the sorted density profiles common oceanographic practice? Then please cite examples from the literature. Otherwise, it is a worthwhile methodological advancement that should be motivated, justified, and also advertised a bit more prominently than only with this single sentence.

L280 Does this bias your estimation of EKE at a grid point when sea ice is present (only) at certain periods of time (which might e.g. be high, or low, EKE time periods)?

L318 Why do you not calculate the gradients from each transect directly and then average the gradients to substitute the numbers currently given in L319. Note that averaged T/S will smooth (among others due to differences in the horizontal location of the strongest gradients) the gradients substantially compared to what is presumably present in each of the individual sections.

L323-330 These lines are repetitive. Consider “This reversal is discussed in the next section.” And then “Simultaneously, …”

L336-339 I can’t quite follow this Eulerian vs. Lagrangian view.

Fig. 6 caption “specified in the lower left corner” I can’t see it.

Fig. 7 caption Consider rephrasing “at B5 in fall (left) and at B7 in winter (right)”.

Fig. 8 right part of figure: delta time = 1 day is a different amount of centimeters on the printed page for the upper panel (October 2020) than for the lower panel (February 2021). Consider making it equal.

Fig. 9 caption “in Figure 2” (a space is missing in front of the “2”).

L432 “We expect the contribution”

L440 “in mid-October”

L442 “between the averaging box and the position”

L456 There are other possible explanations (L 454 “may be related to”). There might be a sea ice related bias (see comment L280). There might be an uneven distribution of events driven by external (non-climate change related) interannual variability in the 2 decades. E.g. (I’m just making up numbers/causal relations for point of illustration) EKE could be high during high NAO phases. In the first decade there were 3 years with high NAO and in the second decade there were 6 years with high NAO even though there is no long-term trend in the frequency of high NAO events.

L501 “scientists” “Haakon cruise”