This paper present an improved chronology for the Antarctic EPICA Dome C ice core for the time interval 0-800 kyr. The development of this chronology involved various methods, including linking to existing Greenland and Antarctic ice cores, orbital tuning with δ¹⁸O_atm, δO₂/N₂, and total air content, and employing firn modeling. One of the significant advancements is the improvement of a section around 110 ka BP, where several previous studies have pointed out that the AICC2012 chronology is too young. Additionally, the increase of new gas data (δ¹⁵N, δ¹⁸O, δO₂/N₂, TAC) has greatly improved the precision of orbital tuning and the estimation of the Lock-in depth scenario, reducing overall chronological uncertainty significantly. While assessing whether the oldest part of the AICC2023 chronology has improved from the AICC2012 chronology is challenging, it does provide a reasonable estimate with a larger age uncertainty compare to AICC2012.

The paper is clearly written and convincingly demonstrates the method, including thorough sensitivity studies. The improved chronology for the EDC core is beneficial not only for the ice core community but also for the broader  paleoclimate community. Therefore, I recommend accepting this paper for publication in Climate of the Past after addressing the following comments.

General comments:

1) I am concerned about aligning the EDC δ¹⁸O_atm and the precession variations older than 590 ka BP, although it seems to be a better solution than the previous one. While Extier et al. (2018) suggested that the Heinrich-like events occurring especially during deglaciations delay the response of δ¹⁸O_atm to orbital forcing, Oyabu et al. (2022) showed that the large lags of δ¹⁸O_atm behind 65N summer insolation (~6 kyr) are not always seen during the Heinrich-like events. For example, they showed that a large lag (>6 kyr) was found during the period of less IRD (around the penultimate glacial maximum), while the lag for HE11 during Termination II is a modest value of 4.1 kyr. Therefore, I think it would be valuable to indicate what potential errors may exist, although the authors have already given a safely large uncertainty. For example, what about applying the same approach to well-dated periods such as the last glacial period, and/or the range of time periods where δ¹⁸O_atm -δ¹⁸O_calcite matching was conducted, with relatively small dating uncertainties on speleothems, and comparing each other? This might serve as a test to evaluate the reliability of the methodology, and the readers will be convinced of the reliability of the obtained chronology.

2) Regarding the gas age for the last 60 kyr, there are some age reversals in the AICC2012 chronology. I believe that the AICC2023 chronology has improved as the tie points have been updated and there have been significant progress on the construction of prior LIDs, but please make sure whether the AICC2023 chronology addressed and resolved the issue.

Specific comments:

• The authors probably have already prepared the dataset, and please indicate data availability. New ages and their uncertainties, age markers used in this study including both updated and old ones, posterior of accumulation rate, thinning function and LID, and gas data should be included.
• In Paleochrono, the assignment of uncertainties to the prior scenarios (accumulation rate, thinning function and LID) has a significant impact on the final estimation of age uncertainties. Please indicate what uncertainties the authors assigned.
• I suggest showing not only the posterior LID but also the posterior of the accumulation rate and thinning function, either in the main text or supplement.
• Figure colors: Grey squares in Fig. 4, 7, 8, 9,10, 12, and S4 are too light in color to see.

Lines 23-24: The use of three orbital markers does not necessarily reduce uncertainties.   

Line 29: Is the uncertainty 1 sigma or 2 sigma?

Line 54: EDC -> EPICA Dome C (because it first appears here in the main text).

Line 59: I would specify the boundary conditions.

Line 72: Need reference(s) for 81Kr dating.

Lines 99-100: Oyabu et al. (2022) also used δΟ₂/Ν₂for the Dome Fuji core over the last 207 kyr and they estimated uncertainties as about 250 to 600 years.

Line 115: Change to “air is trapped in enclosed bubbles and diffusivity becomes effectively zero”. The gas diffuses through the ice matrix (e.g., Salamatin et al., 2001, DOI:10.1016/S0022-0248(00)01002-2), so the original description might possibly be misleading.

Line 145: Same as line 29.

Lines 146-147: “(i) discrepancy between δ¹⁸O_atm, δO₂/N₂ and TAC series and their orbital target”. Difficult to understand what the authors meant. Do the authors mention about the inherent dissimilarity in curve shape?

Line 188: Does the synchronization dating method mean orbital tuning in this context?

Line 189: It seems that Δdepth constraints were not included in this study. Although Δdepth may have a small effect on reducing the age uncertainties, it should be important for constraining the ice-gas relationship. It is possible to make Δdepth constraints using δD and CH₄by assuming a bipolar seesaw relationship, and I think it may help to improve dating accuracy.

Lines 200-204: I read the sentences as the authors did not use age intervals but used dated horizons for the last 60 kyr, and I would refer to Fig. S8 here. The following comments are also relevant to Section 3.4. In Fig. S8, tie points for the CH₄ concentrations are placed up to ~115 ka BP. Did the authors utilize tie points as dated horizons for time periods younger than 60 ka BP and as stratigraphic links between the NGRIP core and other Antarctic cores before 60 ka BP? In the AICC2012 chronology, Veres et al. (2013) employed absolute tie points placed at one-meter intervals to closely fit the AICC2012 chronology to GICC05 over the last 60 kyr. Did the authors apply the same approach for the AICC2023 chronology? Also, where does 122 kyr (lines 162 and 594) come from?

Lines 224-225 and line 40 of Supplementary Material: The δO₂/N₂ obtained from ice stored at -50 ˚C appears much less affected by the gas loss than those obtained from ice stored at -20 ˚C. However, I do not agree that the new data is not affected by the gas loss, because it can be clearly seen that there is offset between the values of Extier et al. (2018) and the new data (e.g., 190 – 260 ka BP in Fig. S1). This discrepancy suggests that the new data was affected by the gas loss. The 3-5 mm surface removal probably does not completely remove the ice affected by the gas loss as shown by Oyabu et al. (2021) (https://doi.org/10.5194/tc-15-5529-2021). While the offset may not affect dating in terms of peak positions (need to check), the absolute value is still important if this data is to be used to reconstruct atmospheric oxygen concentration (e.g., Stolper et al., 2016, doi/10.1126/science.aaf5445; Exiter et al., 2018). In such a case, it would be risky to state that no gas loss correction is necessary. I suggest to the authors mention about the δO₂/N₂ data that the new data is slightly affected by the gas loss, although peak positions are not affected.

Line 227: The reported pooled standard deviation for δO₂/N appears to be small by one order of magnitude (Extier et al. (2018) reported it as 0.37‰).  

Line 268: Need reference(s).

Line 270: Move “Capron et al. (2013)” to the end of the sentence.

Figure 1: I would suggest placing minor ticks between major ticks for the age scale (like fig. 9). Figure 1 and Figure 2 have duplicate items, and it would be sufficient to use only Figure 2.

Line 302 Chapter title: This chapter mostly describes what age constraints and background LID scenarios were used. It may be better to change the chapter title (e.g., 3 Age constraints and background scenarios).

Lines 373-375: I would suggest having a little more explanation of how the 3 kyr was derived.

Figure 2, line 380: Change to “Extrema in the compiled filtered δO₂/N₂ dataset (blue plain line in panel a) are identified and….” .  (b) in the figure should be shown a little lower (next to the insolation curve); it appears to point to the compiled δO₂/N₂.

Line 418: “All extrema are not…” should be “Not all extrema are…”.

Figure 3, line 435: What is the bottom line? Maybe the “bottom” should be “horizontal.”   

Lines 469-470 and Figure 5: There are several cases where the vertical lines in the figure appear to point to the slope instead of the extrema of the temporal derivative. In addition, the type and number of lines connecting (b)-(c) and (d)-(e) do not match in several places—for example, around 160 ka BP and 370 ka BP.

Line 482: I would suggest briefly explaining the source of uncertainty for 1.1 – 7.4 kyr.

Figure 5, line 484: Need a reference for the Chinese δ¹⁸O_calcite.

Figure 5, lines 486-487: Reconsider the descriptions. For example, “Tie points represented by blue vertical bars are determined by Extier et al. (2018) and those by black vertical bars are determined by this study. Both are used in the AICC2023 chronology.”

Line 491: Did the author decide to use the δ¹⁸O_atm orbital markers after 590 ka BP rather than after 640 ka BP because the age uncertainties of the speleothem become large?

Line 499: Delete the period after Bazin et al. (2013).

Lines 508-512 and Section 4.2.3: The Matsuyama-Brunhes geomagnetic reversal has recently been dated with high precision from detailed studies of the Chiba composite section (e.g., Haneda et al., 2020; Suganuma et al., 2020), and ¹⁰Be data has also been published (Simon et al., 2019). International Union of Geological Sciences ratified the Chiba composite section as the Global Boundary Stratotype Section and Point for the Chibanian stage and middle Pleistocene subseries of the quaternary system. In addition, the age of the M-B boundary in Lake Sulmona has been suggested to be affected by remagnetization (Evans and Muxworthy, 2018). Therefore, I recommend referring to the age from the Chiba composite section. The authors possibly be able to increase the accuracy of the chronology by including ¹⁰Be matching as an absolute dated horizon (I am not a ¹⁰Be expert, and it is difficult for me to suggest an appropriate matching method between the ice core and the Chiba composite section) or to verify the final chronology with better precision with the age from the Chiba composite section   .

Haneda, Y., Okada, M., Suganuma, Y., & Kitamura, T. (2020). A full sequence of the Matuyama–Brunhes geomagnetic reversal in the Chiba composite section, Central Japan, Progress in Earth and Planetary Science, 7, 44. http://doi.org/10.1186/s40645-020-00354-y

Suganuma, Y., Okada, M., Head, M. J., Kameo, K., Haneda, Y., Hayashi, H., et al. (2021). Formal ratification of the global boundary stratotype section and point (GSSP) for the Chibanian stage and middle pleistocene subseries of the quaternary system: The Chiba section, Japan. Episodes, 44(3), 317–347. http://doi.org/10.18814/epiiugs/2020/020080

Simon, Q., Suganuma, Y., Okada, M., & Haneda, Y. (2019). High-resolution 10Be and paleomagnetic recording of the last polarity reversal in the Chiba composite section: Age and dynamics of the Matuyama–Brunhes transition. Earth and Planetary Science Letters, 519, 92–100. http://doi.org/10.1016/j.epsl.2019.05.004

Evans, M. E., & Muxworthy, A. R. (2018). A re-appraisal of the proposed rapid Matuyama-Brunhes geomagnetic reversal in the Sulmona Basin, Italy. Geophysical Journal International, 213(3), 1744–1750. http://doi.org/10.1093/GJI/GGY111

Section 3.3 and Figure S4: I am curious to see how well the Bréant model reproduced the δ¹⁵Ν based LID. It is difficult to see a similarity from Fig. 7(b) and 7(c), and I would like to see a figure that both LIDs are plotted on the same panel (either in the main text or supplement).

Figure 7, line 535: The data of Bréant et al (2019) is not mentioned.

Section3.4: I recommend uploading the file containing all tie points used in this study, together with the chronology and its uncertainty, to a data repository.

Figure 8: Grey squares, vertical and horizontal ticks and MIS numbers are too light in color to see.

Line 673: Insert “and” between δO₂/N₂ and δ¹⁸O_atm.

Figure 9: Is the uncertainty shown in the figure 1σ? I suggest adding a figure of gas age similar to Fig 9.

Figure 10, line 699~: References for the CH₄ and δD data are necessary.

Lines 749-750: I agree that a smaller sample size generally increases the noise in data. In addition, I suspect that the EDC samples were slightly affected by gas loss, producing some scatters.

Figure 12: The δO₂/N₂ data from Oyabu et al. (2022) is extended to 207 kyr BP. The gray rectangle in the figure is drawn slightly younger from MIS5e.

Figure S4: Modify the label of panel (b) (Backgroun“d”).

Figure S8: Hard to distinguish between blue and black markers for the volcanic matching points. Also, see the comments for lines 200-204.

Review of Bouchet et al., 2023 – AICC2023

Bouchet et al., 2023 present an update to the AICC ten years after the first AICC. The update is focused on the EDC ice core and the older portion of the timescale that is based on orbital tuning. The increased density of measurements of d18Oatm, TAC, dO2/N2 and d15N are welcome and represent a significant improvement.

This manuscript describes a useful update the AICC. I remain confused by the exclusion of all(?) US,  British, New Zealand, and Australian ice cores from the AICC. This is of relatively minor importance to this manuscript given that the different age ranges and the focus here on ages older than 100 ka and almost exclusively on EDC. I will urge that this chronology is names the EAICC – the East Antarctic Ice Core Chronology – given that there are more West Antarctic cores excluded from this chronology than East Antarctic cores that are included.    

The authors describe a large range of atmospheric gas measurements. The improvement in resolution of the many records is impressive. The orbital tuning of these records remains quite challenging and thus requires a myriad of subjective choices to develop both the timescale and the uncertainty. The orbital tuning, and the tuning to speleothem calcite, suffer from a lack of understanding in either cause of the variations in the measured parameter, the orbital parameter to tune to, or both. In particular, both O2/N2 and TAC have no process-driven explanation for why they vary based on the orbit characteristics and the variations are not produced by firn models. While this highlights the need for better understanding, particularly as great effort is going to extracting multiple >1Ma ice cores that reach the 40 ka world, it should not prevent doing the best that can be done with current understanding. And Bouchet et al. do this. They have produced a thoughtful chronology and while the manuscript is dense, it is also clearly written.

There are a couple of areas that stand out as areas of concern:

1) The firn modeling   

This sentence is particularly confusing: “To obtain

576 a coherent scenario, the firn modeling estimates have been adjusted, by standard normalization, to the scale of LID

577 values derived from δ15N data (later referred to as experimental LID).”

This seems to hiding a major limitation in the methodology. If I understand correctly, the authors cannot get the firn model to match the d15N-inferred firn thicknesses, so they just give up on the actual values and instead seek to match the variations. Whether this is due to an inappropriate firn model (Breant) or outdated forcing (the forcing isn’t shown but I suspect the authors are using the classical isotope-temperature scaling that Buizert et al. 2021 showed to be too cold at the LGM). The firn modeling should really be done with multiple models – which is actually relatively easy to do thanks to the Community Firn Model – and with a range of climate forcings. I think the authors efforts would be better served employing other firn models and forcings rather than the impurity scenarios which the author reject.

2) I would like to see an analyses of the thinning function. The EDC AICC2012 thinning function does not decrease monotonically as expected from ice flow modeling (i.e. the input background scenario). If AICC2023 results in a smoother thinning function, this would provide significant support for the methodology.

General comments on Figures

For all figures, the timescale that each parameters is plotted on should be stated explicitly. It gets really confusing when match points are connected with lines which are not vertical but the two parameters are plotted on the same age x-axis.

Vertical lines corresponding the major axes ticks would be really helpful in assessing the alignment of features

It would be really helpful to see the uncertainty assigned to each tie point. Presumably this could be done with a horizontal bar on the match (on the EDC record)

Specific comments

L36 – The introduction could really use subheadings.

L43 – “zipped” I don’t think this is the right translation to English. I’m not sure what you are going for. I think you are trying to say that a large amount of time is stored in a thin amount of ice.

L44 – need to make community possessive > community’s

L44 – “core” not “cores”

L46 – add “the” before surface

L53 – what about Nye?

L95 – I think it’s worth emphasizing that Bender found no causal link between dO2/N2 and insolation and was quite forthright about that

I don’t expect that the authors will agree to incorporate WAIS Divide, but the introduction should have a paragraph that acknowledges the exclusion and points readers to the timescales for these cores that are tied to WAIS Divide as the best ones to use for past ~60 ka.

L118 – “peculiar” I think you mean “particular”

L308 – shouldn’t you reference Tison et al. 2015 here?

More general comments  

L349 – “superior” in English implies “better”. I think “greater than” is better phrasing

L372 – the discarding of “tie points” worries me. Doesn’t this imply that you don’t understand the underlying mechanisms that link the measurements parameter on the target tuning parameter? If you are discarding tie points all together, should the uncertainty for the tie points you keep be increased to respect that the relationship the ties are based on are not stationary?

L396 – The authors should not use “continuous” to describe the discrete gas measurements. These samples are still quite sparse. Instead, the authors should emphasize increase in sample resolution and the reduction in the largest gaps.

Figure 5 – I find the match points between d18O-O2 and speleothem d18O to be unconvincing. What features are being matched and what features aren’t seems arbitrary. Maybe this would be improved by showing the uncertainty

L500 – I’m concerned the 6ka uncertainty is way to small. 6ka seems reasonable for the actual matches, but shifting the tie points based by 5ka based on whether there is a Heinrich-like event is not well founded. This really needs process modeling for support. Since that is outside the scope of the study, I recommend increasing the uncertainty at least 10 ka (5ka since you don’t know what to tune to and 5ka for the murky matches themselves).

L576 – As mentioned above, I don’t really understand what you are doing to get a coherent scenario. Are there other firn models which get better agreement? And what are the climate forcings?

L588 – Why are you not using the tie points to WAIS Divide directly? These ties are well established in Buizert et al. 2018. The WAIS Divide timescale is more accurate than GICC05 as demonstrated by Svensson et al. 2020 who had to shift the dates of GICC05 more than WDC14 for the bipolar matches.