Dear Prof. Soreghan, dear authors,

After the appearance online of the manuscript titled “Paleocene-Eocene age glendonites from the Norwegian Margin - Indicators of cold snaps in the hothouse?” submitted for review in Climate of the Past by Vickers et al., I read the manuscript with interest and noticed some things which the authors may find useful while revising their text in response to the other review comments.

Overall, I found this to be a well written manuscript which details an important study into the somewhat enigmatic occurrence of glendonites in sediments deposited during warm climates. The authors applied a diverse set of methods used to arrive at a very plausible description of the formation pathway of the glendonites and how they are chemically related to volcanic ashes deposited in the basin. This makes this manuscript a very useful contribution to the field of paleoclimatology and carbonate chemistry, even though I am not fully convinced that the “cold snaps” in Eocene climate put forward by the authors are the only viable mechanism to explain what is seen in the data and observations (see comments below).

Below, I first highlight some general remarks I had while reading through the manuscript. These are followed by some more minor (textual) points and elaboration on the general remarks sorted by line number.

General comments

In my opinion, the observations and data the authors put forward do not conclusively support the hypothesis that the Eocene hothouse was punctuated by geologically brief “cold snaps”. The lack of precise timing of the formation of the glendonites leaves room for other explanations, such as the hypothesis that the fast-growing glendonites form out of isotopic equilibrium (and may record lower temperatures than their growing temperatures) or that they grow seasonally (e.g. during the cold season). Unless the authors can convincingly reject these two alternatives using either evidence from the literature or new observations, they should at least name them as alternative explanations. I think it is well possible that there were indeed geologically short cold periods in the Early Eocene in this locality, but it does not seem like the only possible explanation for the data and observations put forward in this manuscript (see details in my comment on lines 441-462).

The PHREEQC modelling is, in my opinion, the weakest part of the manuscript. The outcomes of this model are hardly described and play only a very small role in the discussion. Perhaps the authors can elaborate a bit more on these results and how they add to the discussion in this manuscript.

On my first read-through, I was quite skeptical about the use of pore water chemistry in Eocene sediments to infer something about paleo-conditions. It seems hard to argue that these preserve a remnant of the original pore waters after such a long time. However, I must acknowledge that I am not an expert in this subject and the authors wrote a nuanced discussion about the observed changes in porewater chemistry (lines 367-397), so I am happy to go along with its qualitative interpretation if more expert reviewers agree that this data can be used to make the inferences the authors put forward in the manuscript.

Minor comments

Line 70: “…the conditions under which this was achieved in the laboratory is unlike any natural setting.” should probably read “…the conditions under which this was achieved in the laboratory are unlike any natural setting.”

Line 76-78: On checking the preprint by Jones et al. (2023), most specifically figures 3 and 6, I did not find any biomarker-based temperature reconstructions that yielded results below 10°C. It seems the only datapoints in this compilation that yield colder temperatures are those originating from Vickers et al. (2020), which represent clumped isotope datapoints on glendonites if I’m not mistaken. Unless I have overlooked any temperature reconstructions in Jones et al. (2023) the authors are referring to here, I think this statement should be rephrased. The authors should acknowledge that the cold temperature reconstructions are only found from analyses on the glendonites themselves, and not from other (independent) proxies and archives. This observation has implications for the “cold snap” hypothesis, as it remains possible that the temperatures reconstructed from the glendonites themselves are underestimations (see comments on lines 441-462).

Line 92-94: The train of thought in this sentence is a bit hard to follow for me. Perhaps the authors could briefly explain how these colder temperature reconstructions from glendonites imply changes in circulation or stratification in this basin. I’m sure this has been discussed in one or more of the papers by the authors cited earlier in the Introduction, but for the sake of clarity I would suggest the authors indulge the reader who has not read these contributions in detail by explaining the line of reasoning here and refer to these previous studies for a more detailed discussion.

Section 2.2-2.6: I commend the authors on their detailed explanation of the geochemical measurement procedures.

Section 3.7: It would probably benefit the manuscript if the results of PHREEQC modelling are described in more detail here. As it stands, this section now contains one sentence referring to a figure and supplement, which leaves the reader searching for the information.

Line 263-269: While I understand the decision to keep the terminology of the different carbonate phases observed within the glendonites consistent with previous literature, I think it would be helpful to include a brief description of the crystal habit (e.g. “botryoidal”, “sparry”, etc.) and texture of each phase (or “type”) here in the Results section.

Line 296: Is “blebs” a scientific term in this context? Perhaps the authors can define it here for clarity.

Line 311: A space is needed between “localized” and “increase”.

Line 313: “due to the rapidity of the reaction” This reasoning requires a bit more explanation, I think. Do the authors mean that the faster mineralization rate causes less discrimination against Mg in the crystal structure compared to phase 1A? If so, it would be helpful if the authors supported this claim about the influence of reaction rate with a reference.

Line 339-344: Here I think the authors should acknowledge that this link between methane seepage and ikaite formation (and the appearance of glendonites in the sedimentary record) has also been recorded in geological history (e.g. Morales et al., 2017)

Line 394-397: I find it hard to see the connection between PHREEQC model results and the discussion in this section, but I think that is mostly because the results of the modelling have not been described in the manuscript. It would be easier for the reader to follow the reasoning here if the outcomes of PHREEQC simulations are first described (in Results) and then discussed earlier in the discussion before they are used to support the discussion here.

Lines 399-440: I think the hypothesis for the formation of ikaite in connection to the deposition of the ashes in this section is very plausible.

Lines 441-462: I’m not sure if I am fully convinced that the data presented in this study and the previous studies cited here are conclusive evidence pointing towards “cold snaps” during the Eocene hothouse. In my opinion, there remain two other possibilities that could reconcile the cool temperatures measured in the glendonites using clumped isotope thermometry with the warmer temperatures inferred from biomarkers.

Firstly, it is possible that the transformation of ikaite to glendonite does not take place in (clumped) isotopic equilibrium. Studies using the comparatively new “dual clumped” method highlight that some carbonates (e.g. those in brachiopod shells or corals) are affected by rate-limiting processes such as the hydration of CO2 in the water and the diffusion of DIC to the mineralization site (Davies et al., 2023). If glendonites indeed grow as fast as hypothesized in the previous section (lines 427-433), these processes may also cause kinetic effects in their isotopic composition. In brachiopods, such effects are demonstrated to cause an offset between the D47-based temperature and the actual environmental temperature under which the carbonate mineralizes of up to 10 degrees, enough to potentially explain a large part of the temperature offset cited here (Bajnai et al., 2020; Davies et al., 2023). While I would not suggest the authors add dual clumped measurements of their glendonites to this study (which might be analytically challenging considering their heterogeneity in terms of carbonate phases), I think this caveat should be recognized as part of the discussion here.

Secondly, I think the authors should consider the possibility that the difference between the temperatures recorded in the glendonites and those in the organic proxies can be explained by a difference in the season in which these materials are formed. If the glendonites formed in the winter and the biomarkers represent a summer signal, it is not inconceivable that the former yields temperatures of 1-9 degrees and the latter 20-30 degrees, especially in the higher latitudes. In fact, similar seasonal temperature contrasts are common in modern marine systems like the North Sea (e.g. Van Aken, 2008) and have been observed in seasonal-scale temperature reconstructions from higher mid-latitudes in the same region during past greenhouse periods (e.g. de Winter et al., 2021). In addition, while the season of growth of recrystallization of the glendonites would be hard to constrain, at least there is some evidence that biomarker-based SST reconstructions may be seasonally biased (Jia et al., 2017; Udoh et al., 2022). It seems plausible to me that the ikaites form in the winter season when temperatures drop enough for the mineralization to start, even if the chemical conditions that seed ikaite formation (made possible by the volcanic ashes) are in place earlier in the year. Therefore, I think this hypothesis cannot be disregarded in this discussion.

Finally, the authors make a link between crystal size and growth rates (lines 427-433). This allows room for the argument that relatively large (cm-scale) crystals can grow in relatively short cold periods. However, even considering the fast mineralization rates observed for ikaite, I wonder how the very large (up to meter-scale) in the Fur formation can form in a setting that is otherwise indicative of typical warm Eocene hothouse conditions (e.g. Schultz et al., 2020). Even assuming growth rates in the order of centimeters per year, growing such crystals would still require decades to centuries of cold periods in this area. While this observation seems to favor the “cold snap” hypothesis rather than the seasonal growth I suggested as an alternative above, it still seems hard to explain such prolonged cool periods of which no evidence is present in the local geological record beyond the glendonites. This makes me think that perhaps the thickness or composition of the ash layers across the basin might be a factor influencing the differences in glendonite size between the locations (with the larger glendonites begin found further south). Perhaps the authors could comment on this in the manuscript.

I hope the comments above will be helpful to the authors in improving their manuscript during the review process, and I welcome any replies to my comments above in case I overlooked anything in my reasoning above.

Kind regards,

Niels de Winter

References

Bajnai, D., Guo, W., Spötl, C., Coplen, T. B., Methner, K., Löffler, N., Krsnik, E., Gischler, E., Hansen, M., Henkel, D., Price, G. D., Raddatz, J., Scholz, D., and Fiebig, J.: Dual clumped isotope thermometry resolves kinetic biases in carbonate formation temperatures, Nat Commun, 11, 4005, https://doi.org/10.1038/s41467-020-17501-0, 2020.

Davies, A. J., Brand, U., Tagliavento, M., Bitner, M. A., Bajnai, D., Staudigel, P., Bernecker, M., and Fiebig, J.: Isotopic disequilibrium in brachiopods disentangled with dual clumped isotope thermometry, Geochimica et Cosmochimica Acta, https://doi.org/10.1016/j.gca.2023.08.005, 2023.

van Aken, H. M.: Variability of the water temperature in the western Wadden Sea on tidal to centennial time scales, Journal of Sea Research, 60, 227–234, https://doi.org/10.1016/j.seares.2008.09.001, 2008.

de Winter, N. J., Müller, I. A., Kocken, I. J., Thibault, N., Ullmann, C. V., Farnsworth, A., Lunt, D. J., Claeys, P., and Ziegler, M.: Absolute seasonal temperature estimates from clumped isotopes in bivalve shells suggest warm and variable greenhouse climate, Commun Earth Environ, 2, 1–8, https://doi.org/10.1038/s43247-021-00193-9, 2021.

Jia, G., Wang, X., Guo, W., and Dong, L.: Seasonal distribution of archaeal lipids in surface water and its constraint on their sources and the TEX86 temperature proxy in sediments of the South China Sea, Journal of Geophysical Research: Biogeosciences, 122, 592–606, https://doi.org/10.1002/2016JG003732, 2017.

Morales, C., Rogov, M., Wierzbowski, H., Ershova, V., Suan, G., Adatte, T., Föllmi, K. B., Tegelaar, E., Reichart, G.-J., and De Lange, G. J.: Glendonites track methane seepage in Mesozoic polar seas, Geology, 45, 503–506, 2017.

Schulz, B. P., Vickers, M. L., Huggett, J., Madsen, H., Heilmann-Clausen, C., Friis, H., and Suess, E.: Palaeogene glendonites from Denmark, Bulletin of the Geological Society of Denmark, 68, 23–35, https://doi.org/10.37570/bgsd-2020-68-03, 2020.

Udoh, E. C., Li, L., Chen, M., Dan, S. F., Chen, L., Zhang, J., Jia, G., and He, J.: Distribution characteristics of terrestrial and marine lipid biomarkers in surface sediment and their implication for the provenance and palaeoceanographic application in the northern South China Sea, Marine Geology, 452, 106899, https://doi.org/10.1016/j.margeo.2022.106899, 2022.