Review of Pickup et al. “Cold lenses in the Amundsen Sea: Impacts of sea ice formation on subsurface pH and carbon” submitted to Ocean Science.

Brief summary

This study investigates subsurface "cold lenses" found beneath the Dotson Ice Shelf region of the Amundsen Sea polynya, Antarctica. These lenses are distinct pockets of cold (Θ < –1.7 °C), salty, and dense water located at 240–500 m depth. High-resolution ocean glider measurements (temperature, salinity, dissolved oxygen, pH, and dissolved inorganic carbon, DIC) reveal that these lenses are colder, more saline, and denser than the overlying Winter Water (WW), but fresher and less dense than underlying modified Circumpolar Deep Water (mCDW). They exhibit slightly higher dissolved oxygen, lower pH, and elevated DIC, indicating intense surface cooling and brine rejection during sea ice formation. Two formation mechanisms are proposed: (1) formation in shallow coastal polynya regions (e.g., Martin Peninsula), where strong cooling and brine rejection drive dense water downslope to its neutral buoyancy depth (~400 m), and (2) local deep convection ("convective chimneys") during winter, with subsequent subsurface trapping. Seal tag data support high surface heat loss and deep mixed layer formation, suggesting sea ice production rates of ~3 cm/day.Ten lenses were identified, ranging from ~5 to 25 km in horizontal extent, and are likely recurring features formed annually. These lenses potentially reduce subsurface heat content, possibly limiting heat transport beneath ice shelves and acting as a barrier to basal melting if advected under the Dotson Ice Shelf. Furthermore, they may provide a mechanism for transporting carbon-rich water deeper than typical WW, influencing regional carbon budgets. Overall, the lenses highlight the importance of shallow shelf processes in shaping Antarctic subsurface water properties and carbon dynamics.

The paper is overall well written and presented. The topic is highly relevand and add to the few studies of ice-ocean interaction. Before I can recommend the paper for publication I would like the authors to take into account my specific comments below.   

Specific comments:

Figure 2: In order for the reader to easily follow what is shown, I suggest that you spell out the names of the water types e.g. AASW, mCDW, WW, Lens. Can be difficult to follow for scientists that is not local to the area. Suggest to change the depth bar so the cold water lences between 200-500 m has a distinct color. Then it will be easier to localise the cold lenses in the T-S space.

You have indicated a mCDW-glacial meltwater mixing line. Could that be similar to the “Gade-line”? In a T–S diagram below an infinit ice cover, a melt line with an observed slope of 2.5 °C per salinity unit corresponds to the Gade slope.

Reference: Gade, H. G. Melting of ice in sea water: a primitive model with application to the Antarctic ice shelf and icebergs. J. Phys. Oceanogr. 9, 189–198 (1979).

Line 211: During sea ice formation, lighter oxygen isotopes are favoured in the ice and heavier isotopes remain in the water, lowering the δ18O of the water. This is not correct.

The fractionation effect during freezing is relatively small but tends to favour the 18O isotope in the ice compared to the residual liquid water. Sea ice typically has a δ18O value close to that of the source seawater, with a slight enrichment (more positive δ18O). In contrast, meteoric ice (ice formed from precipitation) is strongly depleted in 18O (more negative δ18O) compared to seawater. See Moore et al. (2017) Fractionation of hydrogen and oxygen in artificial sea ice… Cold regions Science and Technology 142:93-99.

Line 219: The more negative δ18O within the lenses than in WW suggests more intense sea ice formation as the water has gotten more isotopically light. This is confusing. Do you mean:

The more depleted δ18O values within the lenses than in WW suggest a more intense sea ice formation (as the sea ice brine is more depleted in δ18O)…?

Line 207-245: I suggest to rewrite this section where you specify if an isotope is “enriched” or “depleted”.

Line 254: The mCDW contribution would also have a low O2 concentration, high δ18O, high DIC concentration and low pH content. This should be changed to: The mCDW contribution would also have a low O2 concentration, less depleted δ18O values, high DIC concentration and low pH content.

Line 308: same comment as in line 207-245.

Figure 8: Very simple figure. Do not think it is needed. If decided to keep, I would suggest you to change the left horizontal arrow to follow more the bedrock and point out just above the “depth of lenses”. You could also consider to add the information on 18O, DIC, O2 and pH to the conseptual figure.

Overall this is a well written, scientifically rigorous study focused on observations of subsurface cold water lenses and a selection of their physical and biogeochemical signatures in the vicinity of the Dotson Ice shelf and the Dotson-Getz trough and surrounding environs. The observations presented here warrant prompt publication as they are incredibly difficult to capture and scientifically novel. 

Specific comments are below:

Introduction: There is a new Nature Reviews article on Antarctic coastal polynya’s that may be worth referencing https://www.nature.com/articles/s43017-024-00634-x 

A paper by Couto et al., 2017 https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2017JC012840 similarly used gliders to track subsurface eddy features (with very different water mass characteristics) on the Western Antarctic Peninsula. It would be a nice reference to highlight and compare methodologies.

Methods:

Line 72 - 80:  How significant were the observed salinity spikes prior to removal? What  approximate vertical resolutions were the data collected at at the 5s intervals? Was CTD data used from both downcast or upcast, or just the upcast data in conjunction with the slow upcast sampling for the pH sensor? I am surprised there’s a thermal lag issue for such slow vertical speeds on upcasts for pH sampling and wonder if upcast only data is considered if the Garau method is necessary.  How meaningful are the corrections to the final results of the paper?

Line 85: 3km is quite far and 6 profiles are not very many. I don’t have an issue with the offset, but I suspect it would be helpful to say that this correction is small relative to the scale of the measured oxygen differences between the lens features of interest and surrounding waters.

Results: 

lines 140 - 142: I found this section confusing. I understand what you’re going for referring to two temperature minima, but for the full dataset there’s really only one minimum, consider rephrasing to clarify or highlighting that it is the minimum of temperature on either side of a salinity value in the first sentence. Furthermore, including a box or marker on Figure 2 of what ‘minima’ you are referring to would be helpful.

Line 142: You refer to the melting-freezing line here, but in the figure it only says ‘freezing line.’ I recommend consistency for clarity.

Figure 3: I recommend plotting the start and end locations on the map so it’s easier to reference the figures on the left, which are in distance traveled. The light gray tracks are very difficult to make-out.

Discussion:

Line 264: The reference formatting looks incorrect.