General Comments

Frieling et al., present records of carbon isotope fractionation from the resting cysts of dinoflagellates to investigate their utility in reconstructing ancient atmospheric CO2. This record of core-top material advances earlier work based on laboratory cultures (and based on sound theoretical basis) and so brings the community closer to confidence that this proxy may work in environmental settings. They show there are differences in carbon isotope fractionation between different species, emphasising the importance of single-species records, and show greater 13-C depletion in their core-top samples compared to cultured, motile organisms. The paper is interesting and makes an important contribution, but some of the analysis is unsatisfactory due to uncertainty about the age of the individual cysts in the “core-top” samples (detailed below). Therefore without a thorough treatment of that uncertainty (which is currently lacking) it’s difficult to know whether this proxy has utility. There are certainly hints that it does, but unfortunately this paper does not yet demonstrate that compellingly.

Specific Comments

The problem with using core-top samples is the substantial increase in atmospheric CO2 since the industrial revolution. As the authors note, it is highly uncertain whether the cysts are from the last week, the last year, decades or even centuries ago. The uncertainty around the contemporaneous CO2 is potentially very large. The “rough correction” to 1850 isn’t really a correction at all, but an assumption which is not well supported, at best highly uncertain, and not really dealt with satisfactorily in the later analysis. The best approach (although expensive) would be to 14-C date some of these samples to see when this material actually dates to. The cheaper, and for this present study, more plausible approach would be to propagate through what is a really quite large uncertainty and see whether the conclusions still hold. Lines 147-8 state that “With the exception of pCO2, we hence assume all parameters (SST, SSS, nutrients) to be constant over the period the core top samples represent.” A fundamental problem here is that the authors have little information (or at least present little data) about how long a period of time the core top samples do in fact represent. I’m not sure that the approach taken to this, systematically removing the most 13-C depleted samples is appropriate. Whilst it is certainly plausible that these individuals represent modern samples, the evidence is fairly circumstantial, and they could represent another confounding variable. What is the impact on the analysis if these samples are not removed?

Ie L 213-4 “We assume these assemblages are representative of ocean conditions prior to the massive increase in anthropogenic carbon emissions.” This is very sizable assumption which whilst plausible is not currently supported by very much evidence.

It appears from Figure 5a that no uncertainty at all has been applied to the assumed CO2 value – is this correct? This is not a fair assumption given the uncertain age of each cyst. In fact a plain reading of Figure 5a suggests that, rather than supporting a function between CO2 and ep apart from at <240 uatm CO2, ep is effectively constant, and only slightly higher above 310. Why therefore has 240 uatm been emphasised?

Whilst the data presented here are interesting and important, the analysis at present is not sufficient to support the conclusions drawn robustly.

Technical corrections

1. 22 use of “significantly” if this is meant in the statistical sense, please add p and n values, else reword.
2. 24 ibid “significant”
3. 40-42 This is a slightly eccentric choice of papers to cite here. At a minimum add an “e.g.” but better to make it clear why these papers or a more comprehensive survey of the pCO2 proxy literature.
4. 43 “However, many of the organic compounds used for CO2 reconstructions are not related to a single species, genus or even group.” A fairly sweeping statement here not supported by any references. Which records and compounds are you referring to?
5. 54 “extremely long-ranging” in space or time? Please be specific and it time list age range.
6. 70 Should “cyst species” by cyst-forming species?
7. 90 “Using standard palynological techniques” Please provide a citation.
8. 94 “ultraclean water” what is this? Ie quote a specific measure such as resistivity if reverse-osmosis teqnique has been used
9. 95 “milliQ” is a brand name not a type of water. Please revise.

L104-5 is the 0.3-0.4 permil number precision or accuracy? How has accuracy been determined.

L343-4 “Badger, 2003, 2021;” These are two difference Badgers. Check BG style but likely need to include initials (lots, because they share first first name initial too).

L 385. I’m not sure this is sufficient to meet the journal data policy. Pangaea doi should be available at publication. 

The work by Frieling and colleagues is strong framework and a much-needed study that will open a new opportunity for applications of organic microfossil 13C analysis. Like single species foram analyses (the benchmark for modern carbon and oxygen isotope studies) single or several organic microfossil 13C limits the breadth of sources to sedimentary organic matter and limits the degrees of freedom in a highly advantageous way. This study is the gateway to the deeper geologic record that will allow broad application of the dinocyst proxy to ancient carbon cycle studies. The questions below are meant to enhance the discussion, but the work, as it is, stands on its own as it is presented.

From a methodological perspective I appreciate the details provided here. Controlling for size and process length is great approach but do you see relationships between d¹³C_cyst and cyst size?

To what degree do you feel that the time averaging affected your data? Do you have access to any 14C dates of the surface sediments? From here you could potentially model the expected range of 13C values of DIC accounting for Suess Effect. More details in the manuscript on your rough correction would be helpful.

What do you think is the background blank source? Is it from atmospheric aerosols that adhere to all surfaces regardless of precautions or is it from within the nickel plate? (Does the nickel plate show scoring from the laser?). Regardless, the approach to signal size to noise, considerations of the blank and other corrections seem reasonable. These consideration are important not only for your study and approach but for the future potential of this kind of analysis for sample return from Mars and elsewhere.

Line 280: From this discussion I think I favor your argument that intercyst variability reflects individual differences. One can envision that individual cells or cysts have significantly different 13C values owing to the randomness of cellular growth, changes in microenvironments of growth that also affect DIC and CO2 13C. Add in the time averaging from core top sample collection it is not a surprise that you see large variance. In fact, I would be worried if you did not. Your suggestion of controlling for size, as much as one is able, is a good idea.

Line 280: For standards have you considered dissolving a standard material like caffeine in water and allowing it to dry onto a surface and analyzing that (you could spray it or something). At the very least here you could assume that the starting composition is isotopically uniform. I supposed 13C differences could arise from the drying process, but it may be better than PEF.

Line 300: Have you investigated the compositional differences between cyst and motile cells? I am familiar with the references you report on this issue but what specifically are the differences? What proportion of the carbon from the cell transferred to the cyst? Is this known?