Humphreys et al., characterize biogeochemical (dissolved) properties in sink holes, recently discovered in the Caribbean. The data is high quality and they make use of it to infer local processes, how sink holes may have evolved and for what purposes their study may be useful in the future. Their study is novel, interesting, and well written. I enjoyed reviewing this paper and have only few comments (mostly related to readability and clarifications).  

General comment:

• I got confused quite often with when data referred to the sinkhole and when did it refer to the acid lake. For example, T_Si seems to refer to the acid lake but it is introduced in the methods to refer to the sinkhole. If possible, this should be carefully assessed in the revised version and cleaned up.

Comments along the text:

Introduction:

The introduction was well written and referenced, and was informative for a person who has no specific knowledge on sinkholes like me.

Methods:

The calculation of AOU should be described at least briefly.

Where was the optical O2 sensor mounted? Outside the vessel at the surface? Or on the CTD? This was confusing and also did not become obvious from Fig. 1c. Please clarify.

I would abbreviate dissolved inorganic carbon as DIC as it seems most widely used. Or otherwise C_t, following Dickson et al. (2007). But I don’t have strong feelings about this.   

Since the constants by Sulpis et al. are relatively new, it would be useful to provide a comment why Humphreys et al. use those the most for their calculations.

I did not fully understand how the authors calculated the preformed values. When I look into the supplement then the S or sigma values in the sinkhole are a range. But for P_pref, I believe they need a distinct value. So what value would that be within the range? Is it the corresponding S or sigma value at the surface of the sinkhole? Please clarify and perhaps amend supplementary figures accordingly.

The description of the delta T_C components is not sufficient to understand what happened here (at least not for me). What is R_N, R_P, etc.? I assume R_N is the molar C/N ratio? This needs to be clarified. Also, for R_P there is only one number which is understandable if it is C/P but I would add the 1 anyways to avoid confusion.

Related to this but more generally: I wonder how much sense it makes to use average Anderson and Sarmiento stoichiometry for this when we know that stoichiometry systematically varies across latitudes (or with nutrient concentration), according to Martiny et al., 2013/2014. It would perhaps be helpful to accommodate for this discussion. Shouldn’t make much of a difference because T_C(OM) is low but it may raise some eye-brows.

Line 164: This is unclear. Do you mean that delta TC(R) was delta TC-(deltaTC(OM)+deltaTC(CO3)) ? Perhaps just make a new equation to avoid confusion.

Line 165: what is the “stoichiometric coefficient for OM-associated silicate remineralisation (RSi)”?? Please try to be less cryptic and more descriptive. You lost me here until line 170. Please explain why you did, what you did here and wat the goal was before you provide how you did it.

Line 194: Perhaps remind the reader what the connection depth was.

Line 195: I am not sure about this statement. Do the authors mean that denser waters from deeper ocean regions cannot reach to the top of the platform because the surface of the platform is too shallow? If so, can the authors exclude that water may be pushed upwards somehow (e.g. upwelling) and then overflow into the sinkholes? I would agree that something like this seems unlikely but not sure if they can exclude this.

Line 197: check the unit for O2.

Line 198: I don’t understand this sentence. Please explain why each physical tracer corresponded to a “significantly different” off-platform water depth.

Line 205: I assume the authors have also considered that the corrosive conditions in the acid lake may have increased dissolution/chemical weathering which may have increased salinity? Can this thought be easily dismissed or is it worth mentioning it? The +0.45 in S cannot come from CaCO3 dissolution, I guess, but could it be something else? Do you have data on major ion composition? (I am not sure if this is at all relevant but your data made me wonder so it may be worth throwing a sentence at this).

Line 215: It is interesting that TA is comparatively little increased relative to DIC. Does this mean that the water inside the acid lake has a short residence time because it gets enriched with CO2 but doesn’t have enough time to stay in there to dissolve CaCO3 to its full potential?

Line 220: This should be moved to the methods, as I was wondering where the 1/3 of each was justified in equation 3. Or at least say that the reason for this will be given in secion XXX.

Fig. 2e. Why is delta T_si (R) only shown for sinkhole N, which has the acid lake (or is it?). The data representation is a bit confusing.

Line 286: I agree that the sinkholes could be useful natural analogues but I don’t think the authors have provided particularly useful examples. The analogy to Precambrian conditions does not really make sense because that would mean that the acid lake is isolated for very, very long. But this does not seem to be the case as the authors state themselves. Thus, I don’t think it has value as an analogue for that. The acid lake certainly has value for Ocean Acidification but care must be taken due to low O2, which may restrict its value. I think the acid lake constitutes an excellent natural analogue for the deliberate sequestration of CO2 into the ocean, either linked to “Direct Air Capture with carbon capture & storage (DACCS)”, “Bio-energy with carbon capture & storage (BECCS)”, or electrochemical splitting of H2O into NaOH and HCl (the “SEA MATE” approach from Eisaman et al), where the HCl could be sequestered in the deep ocean. I believe that the sinkholes have more values for this than what the authors mentioned so far. It may be worth considering this, potentially even in the abstract where they mention this also.

Line 299: “Likely”? Or is “may have” more appropriate?

The Figures are generally really nice.

Fig. 1. Why are is one sinkhole (e.g. sinkhole S) at different locations? This is confusing.

Fig. 2. Subplot b: Fig. 1 shows multiple sinkhole locations for sinkhole S, so what do the blue triangles in the case of sinkhole S refer to? (Same question for sinkhole N and E).

Table 1: Why are DIC and TA in mmol/kg when they are usually reported in µmol/kg? Not really important but distracted me a bit while reading.

Submarine sinkholes on carbonate platforms occur worldwide and have been recently discovered on Luymes Bank, Carribean Sea. Humphrey et al. study four different sinkhole complexes on Luymes bank – one of those sinkholes contains dense, acidic and oxygen-poor waters ('acid lake'). The authors study the formation of the sinkholes and of the 'acid lake' and decipher the physical and biogeochemical conditions and processes that sustain its occurrence using bathymetry and water column data.

I found the manuscript overall very well written and very well described. I felt confident that the authors have carried out their study carefully and thoroughly. Therefore, I only have minor comments.

Minor comments:

I agree that sinkholes are useful 'natural laboratories' but I think that the connection the authors draw between the acid lake and Precambrian times is not really applicable as the time and spatial scales are very different – the authors themselves write that the bubble plume is likely only short-lived. Furthermore, many processes on Luymes Bank (e.g. difference in AOU) have not been studied in detail, and thus, the “Impact and wider context” should be addressed more carefully.

Line 52: Is there more data/knowledge on the occurrence and composition of hydrothermal seeps on Luymes Bank/Saba Bank, potentially also with regards to the host rock composition?

Line 92: Please extent on the calculation of AOU and for what it is used.

Line 94: Is T_c = DIC? I find the abbreviation T_c confusing and I would recommend to re-name it to DIC.

Line 110: Please introduce pH_T.

Line 110: Maybe my knowledge is too limited but I am wondering what the difference/relation is between pH_T and pH_NBS? How do you convert these?

2.3.3 Nutrients: Please provide additional information on the data quality.

Line 144: I have troubles understanding what is meant here with “preformed properties upon departure from the open ocean” – are those the off-platform seawater conditions right before the water enters the sinkhole? Please consider to rephrase.

Line 147: Please define/indicate, which P_pref values are used for this calculation as Fig. S1-S2 only show the ranges. Maybe the P_pref values can be added to Table 1 or indicated more clearly in Fig. S1‑S2.

Line 158: I would encourage the authors to give more detail on why the values have been set like this. Did Anderson and Sarmiento work in the same area, also shouldn’t Refield be mentioned here? It took me quite long to understand that R_N is actually C/N and R_P = C/P. Why is “R” used here as this is otherwise introduced to represent the residual fractions? This is quite confusing.

Line 191: Why are the depth profiles for θ and σ not shown?

Line 199: “In the acid lake, each physical tracer…” maybe better “the (conc.) range of each physical tracer…”?

Figure S1-S2: How did you determine such low NO2- concentrations? What is the LOQ of your method? Also for NH3 and Si?

It would be useful to have the abbreviations of the variables (such as A_T, T_C…) defined in the figure captions.

The study by Humphreys et all is a geochemical investigation of water masses in several sinkholes on the Saba Bank.  The authors seek to identify the process that control the chemistry of the waters and relate those to the formation of the sinkhole.  There is a lot going on with respect to where the waters originate, how they are altered during their migration through the platform and how waters move from one sinkhole to the other. Multibeam surveys are used to identify gas seeps and density layers. Sensors characterize the physicochemical properties of the water masses, and the analysis of total inorganic carbon, alkalinity, nutrients and silica are carefully integrated to attribute the sources of carbon and dissolution dynamics. Beyond being able to explain how material is transported and altered locally, the key contribution of this study is that they describe a novel mechanism for sinkhole creation that does not involve mixing or meteoric and seawater or sea-level oscillations. In this case, a geologic source of CO2 is entering the bottom of one of the sinkholes that creates a corrosive solution. The evidence in support of that is very convincing, and their speculation that this may have formed other sinkholes in the area reasonable.  To what extent this process is important on a larger scale is unknown. I also appreciate the significance of finding a body of low pH water that may be used an analog of a more acidic ocean.  Indeed this might be a great natural laboratory for such studies.

The paper is well written and thoughtful.  With only a few suggested edits and comments, this paper is ready to go. One general recommendation is that calling the deeper portion of Sinkhole N an “acid lake” may be an exaggeration. While a pH of 6.4 is technically acidic, the acidity is very mild.

Specific Comments and Edits:

Line 14: The sentence about how sinkholes are thought to form does not include the importance of sea-level low stands.  A slight edit to incorporate the effect of that process would be helpful.

Line 16: I would have “effects” to “products.”  I don’t think effects accumulate.  Products do.  Sounds better to me at least.

Line 17: Just say the studies provide insight…aren’t all insights new?  It sounds less grandiose.

Line 157: Consider adding the OM respiration equation so the stoichiometric ratios are more palpable.

Line 185: Express the volume of the sinkhole as m3 rather than km3.  I think most people can visualize that better.

I do not think much more than that needs to be said.  This is an excellent paper that I think a lot of people will enjoy reading.  Thank you for the opportunity to review it.