General Comments

In this paper, Hättig et al. generate a paired surface-benthic ocean temperature / hydrogen / oxygen isotope record from the Chilean Margin spanning the last 40 kyr. Their new data are similar to those from a neighboring site and they also show that the mismatch in planktic-benthic seawater isotope change is consistent to that observed in the handful of other extant LGM-modern alkenone hydrogen isotope records. This multi-proxy approach is creative and demonstrates the potential of the alkenone δ²H proxy. This manuscript also shows a good awareness of prior work and current technical and analytical limitations. I believe it is well within the purview of Climate of the Past and therefore recommend it for publication following extensive but relatively minor revisions.

My first suggestion is a careful grammatical overhaul; I am sure I did not identify every single instance where the language and logical flow could be improved and I would encourage the authors to actively seek out opportunities to do to so rather than merely addressing those listed below.

My other main recommendation would be a more extensive treatment of the uncertainties. This work involves several transfer functions (i.e., δ¹⁸O to δ²H, δ¹⁸O to salinity, foraminiferal δ¹⁸O to seawater δ¹⁸O, alkenone δ²H to seawater δ²H), calibration error, and general unknowns (benthic paleotemperatures); the authors do a good job of bringing these to our attention, but I believe a slightly more quantitative assessment of how these compounded uncertainties (roughly calculated) affect their final conclusions would be beneficial. (E.g., does this mean that the last deglaciation is the smallest change we can reasonably do this type of analysis for? Is the surface-depth isotope offset within error?)

Specific Comments

• Lines 52-56: My understanding was that the alkenone-seawater hydrogen isotope fractionation’s salinity-dependence was observable in culture (Schouten 2006 and your other references) but not always in sediment (Weiss et al., 2019; Mitsunaga et al., 2022). I think you can say this and acknowledge that effect of salinity on alkenone-water fractionation is not settled.
• Line 73: I think you should put a sentence around here on why you are using Site 1235; explicitly call out the benefits of having the nearby Site 1234 δ²H_C37 and temperature records, then get into why your site and approach are different (which you do: different depths, etc.). Also, in general, especially in your abstract and introduction, you can do a better job saying why your study is so important and why you needed to make these measurements so close to an existing δ²H_C37 A sentence or two here or there could make a big difference.
• Lines 94-97: Were planktic foraminifera unavailable? That could be another means of reconstructing surface water δ¹⁸ And then you could avoid the δ¹⁸O- δ²H transfer function, which is an additional source of error…
• Lines 128-130: How different were the δ²H_C37:3 and δ²H_C37:2 values? I.e., was there any C_37:3-C_37:2 (inter-alkenone) fractionation? I think you should mention it in your results; I know it is not the main thrust of this paper, but some people (Sachs, D’Andrea) have observed an offset while others (Weiss, Mitsunaga) have not, and this could be of interest to the biochemists studying coccolithophore cell water dynamics. At the very least, if there is no (statistically significant) offset, mentioning that legitimizes your choice to report combined δ²H_C37
• Lines 137-150: I think an obvious question is why bother with converting δ²H_C37 to δ²H_SW to salinity if there are existing δ²H_C37-salinity calibrations, and I think you need to address this, at least briefly. What could be said is that the direct salinity calibrations are worse than the δ²H_C37 to δ²H_SW ones in core-tops (Mitsunaga, Weiss). Of course, then you have to explain why δ²H_C37-salinity is worse than δ²H_C37-δ²H_SW if the δ²H_SW-salinity relationship is so tight…
• Lines 195-208: I get that you were unable to measure benthic paleotemperatures, but that does introduce some holes in your argument. It is good that you have the Schrag et al. bottom-water temperatures, because otherwise I would be skeptical that planktic and benthic warming would be the same. Is there any way you could show how much the uncertainty in your calculations matters? I.e., if you are off by X ºC, how much does it matter to your calculations of benthic δ¹⁸O_SW change? And then how does that error propagate to δ²H_SW calculations?
• Lines 249-259: Makes sense – if you assume the same LGM-modern temperature change, you get similar changes in benthic δ¹⁸O as global. If you talk about temperature uncertainties earlier, this might be a good time to discuss what you discovered. What if deglacial benthic temperature change (and therefore δ²H_SW change) was more / less than expected based on Schrag et al. numbers? Does it come close to affecting your argument that planktic δ²H_SW change was much greater than benthic?
• Line 261-281: I think you need to qualify this a bit. It is a very interesting trend if true, but since there are only seven records—including the anomalous Mozambique channel Kasper et al. one—from mostly low to mid-latitudes, as you acknowledge, I am not sure the data is there to do so quite yet. Presumably, there are other Mozambique channel-type sites out there, we just have not discovered them yet. I think you need a sentence or two at the start of this section acknowledging the small number of extant records. But as long as the trend is real…
• Lines 268-269: I think it is believable that the surface ocean would have freshened first, but I think it is intriguing that these changes would not have propagated to the deep ocean within several thousand years given the mean mixing time of seawater. Do you have any other sources that could support this?
• Lines 274-275. I think you could come right out and say that the LGM was drier than the modern. A source or three here could really strengthen your argument and / or I think you need to flesh out the mechanisms by which a drier atmosphere could change the salinity- δ²H relationship, which then more directly affects the surface than the depths. (That last part makes perfect sense to me.)
• Table 1: I think you need to explain in the text or table caption somewhere why the Mozambique channel Kasper record is excluded from your calculations (i.e., changing riverine inputs rather than seawater δ²H change). I would also use “source” instead of “paper” for the leftmost column.

Technical Corrections

• Line 2: [Here, and subsequently] I do not have strong feelings about this, but I have been told that delta notation—i.e., “δ²H” or “δ¹⁸O”—is an adjective and therefore that it always needs to be followed by a noun, i.e., “δ²H_C37 values,” “this δ²H_C37 record,” “our new δ²H_C37 data,” etc.
• Line 2: [Here, and subsequently] I believe that the number of carbons should be subscripted, i.e., “C₃₇.”
• Line 12: “Shift” is overused throughout the abstract, try to find synonyms.
• Line 12: The last / most recent deglaciation?
• Line 16: Separate “waterline” into two words.
• Line 21: “Physiochemical”
• Line 26: “Local climate regime” is vague; do you mean temperature?
• Line 28: “Low oxygen and hydrogen isotopic values” is vague, could be referring to D/H ratios, etc. If you mean “δ²H or δ¹⁸O values,” just say so. (You do this in lines 30-31 anyway.)
• Line 36: The “f” in “δ¹⁸O_foram” is sometimes capitalized, sometimes not. I would go with lowercase, but just be consistent. You also re-abbreviate this (Line 95) elsewhere; you only need to once.
• Line 36: [Here, and elsewhere?] You already introduced an abbreviation for the oxygen isotopic composition of seawater (δ¹⁸O_sw), might as well use it.
• Lines 43-45: Also LGM porewater δ¹⁸O_sw measurements; see Adkins & Schrag 2001.
• Line 45: “To derive” should be “deriving.”
• Line 61: What is the expected deglacial salinity change?
• Line 70: “Alkenone” should be “alkenones.”
• Figure 1: The location of the word “Chile” is misplaced. I get not wanting to put it near where Chile actually is, because it’s a bit crowded there, but maybe that means it is unnecessary. There are also several typos in the caption.
• Line 85: [Here, and elsewhere] I would mostly use “site” instead of “core” in this section since you are referring to the physical locations of the cores rather than the cores themselves. I would use “core” if you were discussing the characteristics of the core (i.e., length, lithology, etc.).
• Line 86: Remove “located.”
• Line 88: Add a comma between “upwelling” and “stimulating.”
• Line 96: [Here, and elsewhere] be consistent with using a comma after “i.e.” or “e.g.” I believe the difference is British versus American English; see which the journal follows.
• Lines 96-97: This is a run-on sentence.
• Line 106: Misplaced parenthesis.
• Lines 107-108: “Ka” should be “kyr.” (Ka [or Ma] specifically means years ago, where here I think you just mean x cm per x thousand years.) Also, you do not need the periods between “cm” and “kyr.”
• Line 111: Check if you need the hyphens here. I am not sure that you do.
• Line 114: Is the standard deviation referring to NBS-19 or the internal standard or both?
• Line 115: Both might be acceptable, but I have more commonly seen “per mille.” Also, I do not think you need to define the ‰ symbol, but if you do, you should do so the first time you use it, not here.
• Line 133: [Here, and elsewhere] to clarify, “online” refers to the value measured in the lab that day and “offline” refers to the published / “correct” values (via Schimmelmann or others)?
• Lines 134-135: [Here, and elsewhere] be consistent with spacing between mathematical symbols (±, <, ‰, etc.). Check journal-specific conventions.
• Line 144: “RSME” should be “RMSE.”
• Line 145: “Waterline” should be two words.
• Line 162: Italicize “in situ,” but I cannot tell what it adds here. (Do you mean from sediment versus from the water column or culture?) Try “estimated using the benthic genera Cibicidoidies and Planulina.”
• Line 163: Not sure these are recent. “Subsequent,” maybe? Also, phrased a little awkwardly (“other species’ values are projected onto the Cibicidoidies and Planulina function?”).
• Lines 251-252: I am wondering if you could use separate abbreviations for surface and deep water δ²H_SW values and change – δ²H_SWp and δ²H_SWb? I am sure you can think of something better and / or do not be afraid to use synonyms for “surface” and “deep” (“planktic” and “benthic,” etc.) too.
• Line 262: “Remarkable” should be “remarkably,” although I would probably just take it out.
• Line 265: “Waters” typo.
• Line 269: “Salinity-δ²H” instead of “salinity-²H?”
• Line 285: “Std” = standard deviation?

General Comments

Hättig et al. present an interesting new collection of paired surface ocean δ²H_C37 and benthic δ¹⁸O_FORAM records from the Chilean Margin for reconstructing changes in ocean isotopic composition ultimately speaking to water mass and/or salinity variability from the LGM.

The data show good agreement with a previously published record from the same locality, and the treatment of sample and data analyses are consistent with those accepted in the literature. The authors’ new approach/application of hydrogen isotopic composition of alkenones for reconstructing the hydrogen isotopic composition of surface water (and then ultimately of salinity) is unique and interesting and certainly warrants publication. I believe this body of work is suited to Climate of the Past, following careful revision.

My overall recommendation to the authors would be to carefully consider reviewing the paper for clarity and detail. There are many instances of grammatical errors/sentences that are difficult to follow. In particular, the discussion and even more so the conclusion is not as well organized as it could be. The conclusion should be a thoughtful overview of the major findings of the work and as it is written it is relatively difficult to follow. I strongly encourage the authors to review the organization of this section and to also improve on the overall portrayal of the major conclusions from this work.

Next, if the authors could be clearer and more explicit about the errors that they are reporting throughout, it would a more powerful reflection of the data. There are of course assumptions that must be made throughout; however, I also wonder if the authors would consider reporting the error in the transfer functions used (i.e., Equations 1 through 8).

Because much of the discussion and conclusions deal with the offset found between surface and benthic estimations/reconstructions of δ²H_sw , and because the reconstruction specifically of benthic  δ²H_sw  relies on temperature, which the authors do not model, I wonder if the authors considered running (and then reporting on) an experiment where they vary the bottom temperature estimates for the benthic reconstruction of δ²H_sw to determine a conceivable (appropriate) range (and to demonstrate the sensitivity) of this particular offset (Lines 195-208). It is difficult to follow that the authors assume temperature is consistent at depth when they report a major shift in isotopic/salinity signatures between surface and benthic environments. Whether this can be resolved here or not, I do believe the authors should spend some time in their discussion on this discrepancy. Are there no planktonic foramifera δ¹⁸O data to use for reconstructing surface salinity?

Finally, I wonder why the authors did not test the δ²H_C37 – salinity calibration models for reconstructing salinity at this site (at the very least plotting their new δ²H_C37 data alongside previously published data). The authors mention this calibration in the literature; however, the paper overall might be made more powerful and rigorous by including this evaluation.

Specific Comments:

Line 9/the overarching argument that the work presented illustrates a ‘regional’ not ‘local’ phenomena/record: I suppose it is relatively subjective, however, it is my experience that a regional phenomenon (especially in the vast ocean environment) would require the authors to have evaluated core samples taken from a much larger range of latitude/locations along the Chile margin. In my experience, 12km distance between two cores is showing consistency at one specific local, where the two cores would be treated essentially as replicate records to illustrate a robust finding. I suggest the authors reconsider whether their work presented is truly extending a local record to a regional one with just two cores so close in proximity. Although this does not detract from the importance of their record, nor does it negate the impact two core records has on improving the results.

Throughout the manuscript, I would also suggest the authors be very clear and careful about describing isotopic trends/directionality. Whenever a change is indicated in isotopic composition through time, it would be very helpful to the reader to know in which direction the change occurred (i.e., the seawater has becoming enriched in 𝛿¹⁸O for example). This is an important detail that will aid the reader throughout.

Figure 2 and Table 1: The figure and table headings here are a little bit difficult to follow. I suggest (as was done nicely in Figure 1) the authors start the headings with a more general/overarching statement of what the figure/table are describing/illustrating before describing specific aspects of the figure/table. Ensure that the figure/table titles can be read independently of the main text and be fully understood.

Technical  Corrections:

Line 2: subscript on C37 (C₃₇)

Line 9 and throughout the work: please determine whether you will spell the word ‘seawater’ as one word (as in my experience it is often spelled) or as two words ‘sea water’. There are many instances of both iterations throughout the manuscript, and this should be cleaned up before publication.

Line 12: It is important I think to indicate that the second core reported on in this work is indeed a previously reported on/published record  - this is good to indicate even in the abstract here and then also throughout.

Line 16: Make it clear here that “this suggests a shift of ca. 5 per mil in benthic/deep water”.

Line 21: again be clear about how you spell seawater – perhaps here the parentheses are not necessary and just spell out ‘seawater’.

Line 22: the phrasing ‘which could improve’ is a bit weak, perhaps just ‘which improve our understanding’…

Line 28: Be sure the use of the colon “:” throughout is necessary. For instance this sentence could instead read: “… isotopes of seawater, where during glacial periods large amounts of water….”

Line 28: “Low oxygen and hydrogen isotope values” is not clear, do you mean more negative? It may be helpful to include/define the equations for 𝛿¹⁸O and 𝛿²H here. Since 𝛿²H especially is a relatively ‘new’ proxy for seawater salinity, defining the hydrogen isotopes explicitly could be informative.

Line 29: You might consider removing the word “the” in this sentence: “….ice on land, leaving seawater more enriched in heavier deuterium isotopes.”

Line 30: Awkward wording choice to describe ‘leading the formation of …’. Perhaps something along the lines of: “…the decrease in global ice volume releases low density fresh water with relatively low…” Also, consider throughout the manuscript when discussing isotopic values whether you are being consistent in your description of ‘low vs high’ or ‘heavier vs light’ or ‘enriched’ etc. isotopic signatures. Choosing one way to describe the directionality of the isotopes is much clearer.

Line 31: seawater

Line 32: “… have a close relationship…” is perhaps not the most technical way to describe this, perhaps something more like “… are correlated with…” or “…are tightly coupled with…”

Line 33/35: seawater (please check this throughout, I won’t call it out in my further comments)

Line 36: I was always taught to refer to the “isotopic” composition of something… so perhaps “… is a function of the oxygen isotopic composition of ambient seawater…”

Line 41: “… or clumped carbon isotopes (∆₄₇)…” for those unfamiliar with ∆₄₇.

Line 41: “In some cases, independent temperature proxies based on organic…” The authors might consider including the word ‘independent’ here to indicate they are proxies not coming from forams themselves but other sources.

Line 43: “…records and modelling…” is a bit awkward wording. Please consider re-evaluating this sentence structure.

Line 46/47: “… relationship between 𝛿¹⁸O_sw and salinity.” The ‘salt content’ is strange wording.

Line 47: Perhaps the last sentence would be clearer if written something like: “However, it is uncertain whether this relationship holds through time.”

Line 50: Again, please be clear on the direction of the isotopic change.

Line 51: Perhaps define C_37:2 and C_37:3 here because you do not define this ‘shorthand’ earlier. Since you are opting to combine the signals, it is important to describe to the reader what these two parameters are.

Line 57: What 𝛿²H signature are you referring to here? The algal 𝛿²H response to changes in salinity? I believe here is where the authors describe the range in salinity reconstructions possible when using culture vs SPOM vs core top 𝛿²HC37 – salinity calibration models. I do think it would be a powerful discussion to include a small discussion on what the data in this new core would show if a 𝛿²H_C37-salinity calibration model was used to reconstruct salinity directly from 𝛿²H_C37. Especially given the authors argue that the dominant species of alkenone producers in this region are likely to be E. hux

Line 60/61: This sentence is difficult to follow, larger than expected based on what?

Line 72: The authors mention that the 𝛿¹⁸O – salinity relationship may not hold through time (line 47), to use 𝛿²H_sw to reconstruct salinity, do the authors bot also have to assume this relationship holds through time? Or at least indicate here that we also do not know if this holds through time?

Line 73: Again, I would suggest the authors reconsider whether this truly speaks to regional versus local variability.

Line 81/82: This sentence is difficult to follow in the figure legend. Please consider re-phrasing.

Line 86: “… from a previously collected deeper neighboring ODP…” Be sure to give credit/indicate this was a previously collected core.

Line 90: Transports cold saline waters where? Isotopically depleted water can potentially be added where?

Line 91: The average salinity in the area was defined how? Based on a quarter grid perhaps?

Line 97: I think you are missing a word in this sentence. This should maybe read: “… of <2 ka, and a total of 27 samples were…” Also, you might consider noting how ODP 1234 was sampled (the same way?).

Line 101: I think you meant to use a comma not a period to indicate 41,000 years.

Line 103: Perhaps the word ‘with’ should be replaced with ‘at’ here.

Line 113: Please define the average differences here – is this a standard error or standard deviation etc.?

Line 114: The standard deviation of what is within 0.1 per mil?

Line 119: Indicate the model number or instrument name for the ASE.

Line 143: RSME should be RMSE

Line 146: Are you referring to the measured salinity in the modern surface open-ocean?

Line 149: Again, are you referring to surface ocean data?

Line 154: Is there a good citation to include here for detailing that the oxygen isotopic composition of forams depends mainly on temperature and the isotopic composition of seawater?

Line 155: “…calcite sources…” is likely not the best/most correct term. Perhaps, calcite species or polymorphs?

Line 159: “… temperature proxies and then obtain salinity…”

Line 161: “… (2011), where the 𝛿¹⁸O- paleotemperature…”

Line 162: in-situ should be italicized.

Line 166: “…at the time of Eq. (4)…” is awkward wording, please considering clarifying.

Line 167: Please define this equation, where is it from? As it stands it is not well introduced/prefaced.

Line 169 (Equation 6): perhaps the authors might define earlier how temperature is solved for in this equation typically? Which proxy(ies) is/are used to resolve temperature.

Line 190: “The oxygen isotopic signatures of benthic Uvigernia in ODP core 1235  range between 4.15 …”

Line 192: 𝛿¹⁸O_foram

Line 192: Perhaps change the word “till” to “to”.

Line 193: This sentence structure needs work. “…showing an LGM to 1 ka shift of…” is difficult for a reader to follow. Again, indicate the direction of the shift.

Line 195: This sentence structure could be improved.

Line 201: What is the temperature change reported in ODP 1234 – if it is consistent, please report the value(s).

Line 204: It is a little bit difficult to follow when the authors are referring to surface and benthic water isotope values – is there a clearer way to indicate benthic vs surface throughout?

Line 206: Please indicate the directionality of the isotopic shift.

Line 218: ‘…globally surprisingly similar in magnitude,” is not very clear wording.

Figure 2: General introductory statement first. “Alkenone hydrogen isotopic ratios measured in ODP 1235….”. What does 5pt average refer to? Where are the black circles on the figure indicating 𝛿¹⁸O of the benthic foram from ODP 1234? What are the seawater temperature data reconstructed from? TEX UK37?

Line 246: Where was de Bar et al.’s work? Indicate this to the reader.

Line 262: I am not sure if it is commonplace to capitalize ‘Recent’?

Line 269: salinity – ²H should be salinity- 𝛿²H_sw ?

Line 271: “… of a dominating evaporative regime on the slope of the isotope…” and on which isotope are the authors referring? Both?

Line 273: “Surface waters may be more sensitive to changes in this…”

Line 274: salinity – ²H should be salinity- 𝛿²H_sw ?

Line 274: “… for surface waters, due to, for example…”

Line 275-276: This sentence is very confusing to follow, please consider elaborating on the message here.

Table 1: Please consider adding a general statement to describe what this table is detailing. I would also consider being explicit about why the Kasper data are not included in your discussion.

Line 297-298: Please consider adjusting the “Globally distributed…” sentence as it is difficult to follow. Subscript C37 (C₃₇).