Review of Harris Stuart et al.: Towards an understanding of the controls on δO₂/N₂ variability in ice core records

This paper presents the relationship between environmental conditions of the ice sheet surface and O₂/N₂ fractionation in combination with O₂/N₂ records from 14 ice cores and results from a snowpack model. The relationship between O₂/N₂ and local summer solstice insolation is well known, while the physical mechanism to create the insolation signal is poorly understood. The paper provides new O₂/N₂ data from Greenland and Antarctica. It qualitatively demonstrates the role of accumulation rate, surface temperature, and solar radiation on surface snow properties with the snowpack model, which contributes to understanding the mechanism of close-off O₂/N₂ fractionation. Overall, the subject is appropriate to The Cryosphere. However, there are several points which require some major revision. It appears that there are several areas with insufficient explanations throughout the manuscript.

General comments:

1. The analysis investigating the relationship between O₂/N₂ and SSI, A, or T (section 3.1) and the sensitivity experiments with the snowpack model (section 3.2) provide new results that deserve a thorough examination. In the current manuscript, the model results are well discussed, while the discussion connecting the observation of O₂/N₂ and the model results (i.e., surface conditions) is weak. The relationship between grain size and O₂/N₂ is not well described. I understand that the focus of this study is not to investigate how the surface snow conditions affect the densification processes/metamorphisms in the deeper firn, but readers may expect that the discussion would address the link between surface snow metamorphism and subsequent firn properties and how they are related to O₂/N₂ Section 4.1 contains a recitation of the results and gives an example of Greenland, but the main argument of this section is unclear. Section 4.2 is almost entirely a review of previous studies. In Section 4.3, I expect the author to develop their discussion here. The model results are discussed, but the arguments connecting to O₂/N₂ are unclear. In the revised manuscript, I expect to add some discussion connecting the results of Sections 3.1 and 3.2, for example, by introducing arguments from previous studies such as Fujita et al. (2009) and Hutterli et al. (2009). See also my specific comments.

2. I am not satisfied with the analysis of the link between O₂/N₂ and water isotope ratios of the EDC core, which is one of the bases for the discussion linking the O₂/N₂ fractionation mechanism with the model results. First, the authors found positive correlations between δD and O₂/N₂ in the EDC core, but they are all separated by relatively short time periods where the amplitudes of the 20 kyr cycle of δD are large (interglacials). Therefore, it seems to be not surprising to find correlations within those specific periods because O₂/N₂ has a strong precession component. The absence of a positive correlation in MIS11 may be because this interglacial period has a length of 2 precession cycles. If there is indeed a correlation between O₂/N₂ and T and/or A, a positive correlation would be expected even in MIS11. What would happen if correlations were examined over longer continuous periods, including periods with a smaller amplitude of 20 kyr cycle, such as MIS6? The validation may be possible by using data from the South Pole and Dome Fuji cores.

Second, if the authors want to investigate whether the 1000-year scale variations shown in grey shadings in Fig. 4 are related to the O₂/N₂ signal, it is necessary to remove orbital variations with appropriate methods. For this purpose, I suggest applying the low-pass filter used in orbital tuning to both δD and O₂/N₂ and extracting residuals from low-pass filtered curves, or applying a high-pass filter to δD and O₂/N₂ to cur off the insolation frequency. The current approach can introduce artificial variations due to the potential mismatch between the insolation and ice core ages. If the AICC2012 age scale is perfect in terms of absolute age, the long-term variation in Fig. 4b is real. However, as Extier et al. (2018) pointed out, the AICC2012 chronology tends to deviate from the U-Th chronology, especially during deglaciations. This raises the question about the accuracy of this chronology. Therefore, the orbital-scale variation shown in Fig. 4b may just reflect the phase difference between the AICC2012 chronology and insolation. Using the most recent AICC2023 chronology aligned with the U-Th chronology over the last ~600 kyr (Bouchet et al., 2023), or the DF2021 chronology (Oyabu et al., 2022) synchronized with the local summer insolation, could potentially alter the appearance of the residuals in O₂/N₂.



Therefore, I suggest to conduct the following analyses, and add the results and corresponding discussion as needed.

1. Apply a low-pass filter to both δD and O₂/N₂ and extract residuals from low-pass filtered curves to eliminate the potential discrepancy between the ice core and insolation ages. Alternatively, apply a high-pass filter to cut off the insolation signal.
2. Use the AICC2023 chronology to examine SSI-residuals.
3. Use the Dome Fuji O₂/N₂ and water isotope data, and use the DF2021 chronology to examine SSI-residuals.
4. If applicable, use the South Pole core O₂/N₂ and water isotope data, and use the SP19 chronology to examine SSI-residuals.

Specific comments:

Title: I suggest changing the title to be more specific. This paper focuses mainly on inland Antarctica (Dome C), although the Greenland records are used, and discusses the relationship between O₂/N₂ and the surface environments. Thus, I suggest including the idea of “the relationship between O₂/N₂ and the ice sheet surface environment” and “inland Antarctica” in the title.

Line 24: Please consider adding two more references: Lipenkov et al. (2011) for the Vostok chronology and Oyabu et al. (2022) for the DF chronology. Both used O₂/N₂ for orbital dating.

Around line 60: Please consider adding a hypothesis by Kawamura et al. (2007) that the absence of a climatic signal may result from the cancellation of temperature and accumulation effects on O₂/N₂.

Line 100: I suggest also providing a brief explanation of the methodology at Scripps in a similar manner as that for LSCE.

Line 103: “An over view of …”  the same information is written in Section 2.1.

Section 2.2.1: It is not clear which samples were measured at LSCE and which at Scripps. If only the GISP2 samples were measured at Scripps, it would be better to switch the order and write it at the end.

“measured on the 10-collector Thermo Delta V Plus” is not necessary for all samples. This information should be shown at once in the first paragraph of Section 2.2.

Line 163-165: The brittle zone is a zone of poor ice-core quality and not necessarily consistent with the bubble-clathrate hydrate transition zone (BCTZ) (Neff, 2014). Thus, I suggest using “BCTZ” (or a similar term) as in the supplement text. If you did not find the BCTZ information in some cores and thus employed the brittle zone of Neff (2014), you should write it in the text. Also, “the air in the gas phase has a very different composition to that in the clathrate hydrates” should be “the air in the bubble has a very different composition to that in the clathrate hydrates” (e.g., Ikeda-Fukazawa et al., 2001).

Line 166: O₂/N₂ is not always increased in the BCTZ (e.g., Oyabu et al., 2021).

Lines 178-183: Regarding the data rejection criteria, it is unclear what was rejected with and without gas loss correction, and what was employed with and without gas loss correction. It seems that some of the data you utilized was affected by the gas loss but was included in the analysis because the gas loss correction worked. Also, the data with correction should be written as “corrected” since it is not unaffected by gas loss.

I would like to see a figure that displays all O₂/N₂ data, including rejected and employed data with and without gas loss corrections, with different colors or symbols.

Lines 279-280: I didn’t understand what the authors meant.

Line 288: “1000 year averaged SSI”  Did you average for the last 1000 years?

Line 292: “bias toward cold, low-accumulation conditions” is unclear. Does it mean that A and T are averaged between the LGM and the present and do not include past interglacials that were warmer than the present?

Lines 301-302: I am curious whether this regression would be better if you used SSI for the same time period as the O₂/N₂ data. Have you confirmed before?

Lines 315-317: I don’t think these descriptions are necessary here. It is obvious that this paper does not use O₂/N₂ for orbital dating.

Line 317: If all EDC O₂/N₂ data were published in Bouchet et al. (2023), I think it would no longer be new in this paper.

Section 2.3: What is the depth range or thickness of each layer considered in this model? Also, the authors mention that the maximum number of layers was increased from 50 to 80; was this a new modification made specifically for this study?

Line 323:325: As I pointed out above, this long-term variability may just reflect the phase difference between the AICC2012 chronology and insolation.

Lines 342-343, Figure 5a and 5b: What factors contributed to the increased SSA for the top 10 cm and the subsequent dips in both SSA and density around the 10 cm depth in the model results?

Lines 371-373: “These opposing influences of accumulation rate and temperature on snow properties at first appears to contradict the observation in Figure 3”

I agree that the opposing influences appear to contradict, while this is consistent with the hypothesis by Kawamura et al. (2007) (cancellation of temperature and accumulation effects on O₂/N₂). How about mentioning this consistency?

Line 373-375: I didn’t understand the sentence. “most evident is the sensitivity of grain size to accumulation rate” Why can you say that?

Line 386: It is not clear why the decrease in mean density and increase in density variability with A max and T min is “surprising”. Some more explanation is needed.

Line 389: I suggest inserting “increase in” or similar words between “with” and “SSI”.

Lines 421-423: “The observations from….” The sentence is unclear. Please clarify.

Lines 423-425: This section is for discussion and I don't think this statement fits here.

Lines 469-470: I would suggest to delete the sentence “Our findings…” and move the contents of lines 470-472 to the last paragraph of section 4.3.1. If you would keep the sentence here, more words are needed (it is unclear what “our findings” and “a density-dependent grain size mechanism” refer to).

Lines 473-474: Not necessary. This sentence is a repetition of the sentence in lines 468-469.

Lines 489-492: Hard to understand. “leading to bulk ice with decreased δO₂/N₂”, but the sentence before this phrase alone does not yet clarify the causal relationship. Your analysis of the ice core data shows that O₂/N₂ decreases as SSI increases (anti-correlation), and your model results show that grain size decreases as SSI increases. This would mean that there should be an anti-correlation between grain size and O₂/N₂. In addition, there seems to be a lack of explanation why/how the increased grain size depletes O₂/N₂. You may consider adding discussion, drawing arguments from previous studies as described in the Introduction section, to explain why a decreased (increased) O₂/N₂ is associated with a larger (smaller) grain size. One idea may be bring the discussion of Calonne et al. (2022) and Gregory et al. (2014), which appeared in lines 470-473, to here.

Lines 498: What is the link between your findings and residence time in the LIZ? The model results show only near the ice sheet surface, and there seems to be no discussion of how the results relate to the O₂/N₂ fractionation in the LIZ (deep firn).

Lines 519-520: What does “the variability– and bulk mean – differences” refer to? I didn’t understand what the authors meant.

Lines 556-567: Need to explain why elongated pores lead to a greater fractionation of O₂/N₂.

Lines 583-585 (We argue that the…): I don’t see this argument in Discussion. This is the conclusion section and not a good way to introduce a new argument. The argument should be addressed in the Discussion section.

Table 3: A max of the Dome Fuji core seems to be too large (even larger than at EDML). The accumulation rate of the Dome Fuji core over the last 720 ka can be found at NOAA Paleo Data Search.

Technical corrections:

Line 18: “LID” only appears here, but “LIZ” appears without abbreviation (e.g., line 69).

Line 41: “Tomoko Ikeda-Fukazawa and Hondoh, 2004” is “Ikeda-Fukazawa et al., 2004”.

Line 154: Add “slope” after “Chemical”

Line 277 and 3rd line of the Fig. 2 caption: ‰.m².W^-1 Remove periods.

Line 297: Figure 2b may be 3b.

Line 300: panels (a) and (b) may be panels (b) and (c).

Line 301: Large residuals in “Figure 2a” should be in “Figure 3a”. I suggest replacing “residual” with another term, such as deviation from the regression line.

Line 314: EPICA Dome C is already shortened in Line 82.

Line 363: Figure 6c and 6d may be 6a and 6b.

Line 388: 50cm. ,  Remove period and space.

Line 417: (Suwa and Bender, 2008b) -> Suwa and Bender (2008b)

Table 1: Brittle zone should be bubble-clathrate transition zone (BCTZ) or a similar term.

Figure 5 caption: Density (a) and SSA (b)

References:

Bouchet, M., Landais, A., Grisart, A., Parrenin, F., Prié, F., Jacob, R., Fourré, E., Capron, E., Raynaud, D., Lipenkov, V. Y., Loutre, M.-F., Extier, T., Svensson, A., Legrain, E., Martinerie, P., Leuenberger, M., Jiang, W., Ritterbusch, F., Lu, Z.-T., and Yang, G.-M.: The Antarctic Ice Core Chronology 2023 (AICC2023) chronological framework and associated timescale for the European Project for Ice Coring in Antarctica (EPICA) Dome C ice core, Clim. Past, 19, 2257–2286, https://doi.org/10.5194/cp-19-2257-2023, 2023.

Extier, T., Landais, A., Bréant, C., Prié, F., Bazin, L., Dreyfus, G., Roche, D. M., and Leuenberger, M.: On the use of δ18Oatm for ice core dating, Quat. Sci. Rev., 185, 244-257, https://doi.org/10.1016/j.quascirev.2018.02.008, 2018.

Fujita, S., Okuyama, J., Hori, A., and Hondoh, T.: Metamorphism of stratified firn at Dome Fuji, Antarctica: A mechanism for local insolation modulation of gas transport conditions during bubble close off, J. Geophys. Res., 114, https://doi.org/10.1029/2008JF001143, 2009.

Hutterli, M., Schneebeli, M., Freitag, J., Kipfstuhl, J., and Röthlisberger, R.: Impact of local insolation on snow metamorphism and ice core records, Physics of Ice Core Records II : Papers collected after the 2nd International Workshop on Physics of Ice Core Records, held in Sapporo, Japan, 2-6 February 2007. Edited by Takeo Hondoh, 2009.2009.

Ikeda-Fukazawa, T., Hondoh, T., Fukumura, T., Fukazawa, H., and Mae, S.: Variation in N2/O2 ratio of occluded air in Dome Fuji antarctic ice, J. Geophys. Res., 106, 17799-17810, https://doi.org/10.1029/2000JD000104, 2001.

Ikeda-Fukazawa, T., Kawamura, K., and Hondoh, T. (2004) Mechanism of Molecular Diffusion in Ice Crystals, Molecular Simulation, 30:13-15, 973-979, DOI: 10.1080/08927020410001709307

Kawamura, K., Parrenin, F., Lisiecki, L., Uemura, R., Vimeux, F., Severinghaus, J. P., Hutterli, M. A., Nakazawa, T., Aoki, S., Jouzel, J., Raymo, M. E., Matsumoto, K., Nakata, H., Motoyama, H., Fujita, S., Goto-Azuma, K., Fujii, Y., and Watanabe, O.: Northern Hemisphere forcing of climatic cycles in Antarctica over the past 360,000years, Nature, 448, 912-916, https://doi.org/10.1038/nature06015, 2007.

Lipenkov, V. Y., Raynaud, D., and Loutre, M. F.: On the potential of coupling air content and O2/N2 from trapped airfor establishing an ice core chronology tuned on local insolation, Quat. Sci. Rev., 30, 3280-3289, https://doi.org/10.1016/j.quascirev.2011.07.013, 2011.

Neff, P. D.: A review of the brittle ice zone in polar ice cores, Ann. Glaciol., 55, 72-82, 10.3189/2014AoG68A023, 2014.

Oyabu, I., Kawamura, K., Buizert, C., Parrenin, F., Orsi, A., Kitamura, K., Aoki, S., and Nakazawa, T.: The Dome Fuji ice core DF2021 chronology (0-207 kyr BP), Quat. Sci. Rev., 294, https://doi.org/10.1016/j.quascirev.2022.107754, 2022.

Review for Harris Stuart et al. 2023: Towards an understanding of the controls on δO₂/N₂ variability in ice core records by Jochen Schmitt, Bern

The paper, led by Romilly Harris Stuart, aims to improve our process understanding of a key ice core parameter - the O2/N2 ratio -  that is used to orbitally date Antarctic ice core records. Decades ago it was discovered that the O2/N2 ratio resembles the orbitally-controlled solar insolation at the drill site. Already at that time, it was speculated how the upper firn layer is modulated by the amount of sunlight during summer. To affect the archived O2/N2 ratio in the bubbles, firn surface properties need to travel through the firn column to influence gas-specific (size-dependent) gas loss processes during the pore closure at the bottom of the firn. Over the years, a large number of studies suggested ideas to explain the observed O2/N2 variations, but we still lack an overall process understanding. While insolation apparently contributes a large fraction of the measured O2/N2 ratio, local temperature and accumulation rate modulate the orbital signal and lead to noise and uncertainty in the orbital tuning. At this target, Harris Stuart et al. place their study, which consists two approaches. Their study contributes to an important question relevant to the readers of The Cryosphere. Mostly, the paper is written clearly and provides the right depth of information and the figures are well-crafted and provide a visual support for the text. Overall, I support the publication of this study after minor revisions.

Their first approach is to apply an existing snowpack model to see if and to what extent differences in the solar radiation lead to changes in firn properties that might explain the observed O2/N2 ratios. Since the snowpack model was originally designed for alpine firn, applying it to low accumulation sites in Antarctica sets limitations. The authors became well aware of several limitations of their model (1-dimension, no wind compaction, merging box design) and thus interpreted their results with care. I got the impression that they used the model as much as it was possible for this study and then realized that no further insight could be obtained with this setup and that the model would need a significant improvement to capture the situation of low accumulation sites.

Their second approach is data-based and it was certainly a large effort to collect and screen all available O2/N2 records. The screening and data evaluation of the different cores and measurement campaigns is an important step and it would be crucial to provide a figure or two to allow the reader to see and understand the underlying problems of that step. Since it likely took a long time to collect all the records it would be helpful for others and the next generations of scientists to have easy access to these data sets and their meta information. So please spend some hours (perhaps more realistically days) of your time to bring all these data sets to a public database (both the already published and the new data). The analyses done on these 14 selected ice core sites conclude that factors other than insolation (accumulation rate and local temperature) have a sizable effect on the observed O2/N2 records and set limits to the precision and accurate of orbital tuning. This is a valuable outcome, but I feel that - in an ideal world with more time and resources - more can be done to disentangle the interplay between accumulation rate and temperature. As for the length of the diffusive firn column (i.e. d15N-N2), it might be the location on an accumulation vs temperature plot that determines if the firn column gets longer or shorter, or if the grain size within the first meter of firn increases or decreases. Since the temperature and accumulation rates are either known from present-day conditions or are output parameters of models (e.g. can be derived from delta age etc.), the team of this study might want to look a bit deeper into the interplay of temperature and accumulation rate in modulating O2/N2 ratios.    

I also wondered if more process understanding can be gained from analyzing the O2/N2 data from firn air studies. At least there should be some O2/N2 data from some drill sites available. The authors mention several times that one modulating factor of the O2/N2 imprint in the archived air bubbles is the degree by which the O2-enriched air that was expelled by the closing pores is advected upwards or diluted in firn. In other words, the O2/N2 fractionation during pore closure is only seen if it happens in an open system, ie. if the O2-rich air is removed from that layer. See e.g. lines 433 – 438. Perhaps using a full firn model that allows the simulation of permeability in the deep firn could help here?          

Further suggestions and technical comments:

Line 3: “trapped bubbles”. I guess you want to say that the air in the bubble is sealed off from the open pore space; you can just say bubble since bubbles are closed anyway.

Line 4: write “… N2 molecules in extracted ice core air relative to the modern atmosphere - ”

Line 6: write “…and show a new additional link…” delete: “, in addition to the influence of the summer solstice insolation”

Line 8: “… forcings modulate snow physical properties near the surface ”

Line 10: “a mechanisms..”

Line 16: firn…unconsolidated snow? Firn is the consolidated snow  

Line 18:  rewrite “become sealed off from the firn air to form bubbles within the ice.

Line 18: “lock-in depth (LID)” actually you never use LID throughout the paper while you often use lock-in zone.

Line 22:  komma after sites?

Line 21 to 28: perhaps restructure this a bit. Essentially you describe two different kinds of dating approaches. O2/N2 and TAC are due to local effects of the firn column, thus these parameters are highly site-specific. On the other hand, d18O of O2 is a globally mixed atmospheric gas parameter that is not site-specific, and all ice cores yield the same record. Thus it can be used to wiggle-match different records but also relate the record to a certain orbital parameter.

Would be good to mention these two different approaches

Ideally, you could mention that d18O2 is used to date the gas phase of the ice core while O2N2 is an ice age parameter

Line 28:  delete “trapped within the ice” so it gets a bit more general

Line 30/31: delete “vice versa” 

Line 31:  you could delete “numerous” as you already name quite a few sites…

Line 34/35: you might rewrite this to convey that the modification due to insolation happens at the snow surface but the process that effectively alters the archived O2N2 ratio happens at the depth where the pores close off 

Line 37: (COD) is just used twice …just write it out in both cases 

Line 40: replace ; with :

Line 41:  Why cite also her first name Tomoko? 

Line 48: Why “They”? you refer to Bender (2002) so technically just Michael Bender although he acknowledges at the end of his paper that he profited a lot from the discussion with many giants in this field

Line 85: WAISD would be a new abbreviation, commonly used is WD or WAIS   

Line 102 Table 1 (and other tables):  for better visibility please align numbers in columns on the right side, e.g. Table 3 in Petrenko et al. 2016 http://dx.doi.org/10.1016/j.gca.2016.01.004

Line Table 1: If possible and available please also add other site characteristics to this table, e.g. close-off depth or ice age at close-off depth (delta age) they might be useful as well   

Line 157: “gas loss during coring”, can you explain a bit more here?  

Line 163: Note that the brittle zone does not always correspond to the BCTZ, while for most of the ice cores, this is the case. I guess some ice cores have a technically defined brittle zone while they do not have the conditions to form clathrates at a certain set of depth or temperature; thus, without this coexistence of clathrate and bubbles, there shouldn’t be a strong fractionation. Perhaps elaborate shortly on that.

Line 165: O2/N2 measurements within the brittle (or BCTZ) ice are not per se unreliable; it requires a post-coring gas loss, so the fractionated air in the bubbles escapes and thus induces scattered results. Also, small sample sizes resolve individual layers of bubbes vs clathrates

Line 167: see above, does Berkner have a chlathrate zone? Perhaps this explains good data within the brittle zone.  

Line 176: post-coring gas loss to differentiate between the gas loss happening during pore closure in the ice sheet  

Line 214: you mention the black carbon content. How sensitive is the model to the black carbon? What about a similar effect of mineral dust during glacial times (OK, mineral dust is mostly light quartz but there are also darker particles…)  

Line 277: are the dots after the permil and the m2 correct?  

Line 282: “integrated summer insolation”: can say a few words on the difference between integrated summer insolation and SSI and why you use SSI?

Line 290: Figure 2 caption:   you don't need to say that Dome C is plotted in dark-blue and Dome F in mid-blue because you can indentify each panel with their name already. Please add a), b), c)  as you do in Fig. 3

Line 290: Figure 2 caption:  why do you use r2 here, while in Fig. 3a, you use r for the same type of plot? perhaps always use r (as r2 can be calculated from that)

Line 312: Figure 3:   I very much like your colour scheme, but here, it would also help to provide more visual hints to distinguish between some sites, e.g. LD and BI have quite similar colours (same for NEEM and WAIS). You could additionally use squares and diamonds.

Line 328: Table 3:  The 5 EDML samples (596 – 860) are from the brittle zone. Are there no other samples measured at EDML, why just in the brittle zone?

Line 331: Fig. 4: Would the residuals look different if it would be plotted on the AICC2023?  

Line Figure. 4 caption:  the respectively structure always requires the reader to go to the end of the sentence while the classical way "Panel d shows the correlation" is often quicker to access

Line 354 Figure 5d: since there is no overlap between Jan and Jul, you could put both distributions into a single panel

Line 381 Figure 7:  it is not easy to see the difference between the faded line and the max line, perhaps increase the thickness of the line or use dashed lines etc

Line 399:  the long list of references affects a bit the readability ....not sure if you need all the references here in the discussion section, perhaps write e.g. and two refs

Line 406: Fig. 3c, I guess you mean Fig. 3a showing as well O2/N2 vs SSI while Fig. 3c shows temperature. Where is the slope for Fig. 3a to compare it with the slopes of Fig. 2?

Line 438-440: I am not so sure if this argument holds that the O2/N2 signal would then be on the gas age scale. Still it happens in the lock-in zone due to a process that was imprinted originally at the surface.

Line 490: I am puzzled a bit about the term bulk ice…

Line 490: “The opposite – a decrease” not sure if this sentence describing the opposite effect is necessary I guess the sensitivity of grain size for a given density works in both directions

Line 493: yes, temperature and accumulation rate do generally covary, but they are not super tightly correlated, and there are sites that are above or below the expected line for the temperature–accumulation relationship. Perhaps you can derive some useful information from the deviations from this temperature-accumulation relation, i.e. a site that has too little accumulation rate for a given temperature. A scatter plot showing all sites with their accumulation vs temperature might help to identify sites that deviate from others in Figure 3. This requires O2/N2 data for the present-day conditions for accumulation and surface temperature that are likely more accurate than the reconstructed values based on modelling via water isotopes.

Line 500: this sentence is a bit unclear to me

Line 545: could you spend a few words on how the local SSI at EDC is linked to local accumulation rates since this is a larger-scale weather phenomenon and involves low-pressure systems entering the continent, etc.? Perhaps elaborate a bit on that?

Line 575: “local climate (accumulation rate” I understand what you mean but accumulation rate might be largely determined by the circulation patterns in the Southern Ocean region.

Line 576: I guess this statement also holds for AICC2023 (although there seems to be a small circularity hidden into that because the age scale is constructed using the O2/N2 orbital tuning)

Line 582: this sentence misses some words…support the idea…