In this manuscript, Rousseau and colleagues present a small review of millennial scale variability in the North Atlantic, with a particular focus on DO events. They use recurrence plots to determine the main transition times in the past 3 million years, and also to link millennial scale variability in the last glacial period to Bond cycles.

The review section is unexpected in CP, but could make sense as a contribution for this special issue. I have no major comments on that section and will leave this to the editor. On the research side, the use of recurrence plots to identify specific transition is an interesting approach. The advances presented are not very substantial when compared to standard CP papers; in my second major comment I suggest some aspect that could be fleshed out a bit. My main problem with this manuscript is the seemingly arbitrary way in which the transitions in the recurrence plots are determined (see major comment below). I wonder if the results are robust against small changes in parameter selection. I would like to see a sensitivity analysis before I recommend publication of this manuscript.

Major comments:

1. The choice of transition in the recurrence plots is not explained or justified. In line 166 and 167, the authors refer to Eckmann et al., 1987 and Marwan et al., 2013 to determine “sufficiently close”. That is not acceptable, as a publication should include all necessary information to replicate the results. The authors should explain in detail what choices they made to produce the red lines in the recurrence plots. An emblematic example of this problem is Figure 4. Looking at the recurrence plot in figure 4b I see no justification for the line at 32 kaBP, it seems very arbitrary. The same goes for the line at 78 kaBP; between 70 and 78 kaBP there seem to be three more transitions that could reasonably have been drawn. The question arises about the sensitivity of the results to small variations in the parameters of the algorithm chosen to identify transitions. An uncertainty/sensitivity analysis needs to be added for each RP.
2. The sentence in line 305 is unclear. Are the authors defining new GIs based on the recurrence plot? If so how are they defined? If instead they are talking about the GI numbers in Figure 4a, which ones do they mean? There are several numbers in each interval defined by the red lines. I think the authors may have missed an opportunity to make a clear contribution here. This paragraph is the only one of the chapter that appears to be more than a review, and the relationship between GI duration and sea-level is very interesting. A scatter plot of sea-level (or sea-level trend) vs. GI duration would make their point much clearer and add a bit more results to this chapter.

Minor comments:

Lines 12-13: “relatively” used twice in one sentence

Line 16: “constant” is the wrong word here since these are periodic variations. Maybe “regular”?

Line 98: It is unclear what “those” stands for in the second part of the sentence. I imagine it must refer to the shorter periodicities metioned in the first part? Please clarify.

Line 99: I’m not sure “affected” is the right word here. Maybe something like “the frequency of abrupt changes is in part modulated by…”

Line 131-132: As I understand this sentence, it now says that during the late Pliocene the ice sheets over Greenland and Scandinavia were larger than during the Quaternary. That is not the message of the Naafs et al. 2013 paper. Please clarify.

Line 141-143: Yes, but Barker’s record starts at 800 kaBP without any information about the occurrence of millennial scale variability before that. I think it is important to make clear that we don’t know if millennial-scale variability (i.e. DO events) started during the MPR or not.

Line 154-155: This sentence is too vague, as ice sheet extent was very large also during MIS6 and LGM. It also doesn’t convey much important information. I suggest rephrasing it or deleting it.

Line 184: I think it would be helpful to explain in one or two sentences what a “drift topology” is here, with deeper insights being referred to Marwan et al.

Line 186: Please refer to Figure 2a at the end of this sentence already.

Line 196-197: This sentence seems unnecessarily complicated. I suggest “Our analysis further identifies the steps at 0.9 Ma, 1.25 Ma, and 2.75 Ma (with 1.25 step also noticed in the d18O).”

Line 204: The sea-level change increased since the value of the change is not higher

Lines 208-211: That is only true for the glacial maxima. The glacials themselves have all kind of different orbital configurations due to their long duration.

Line 221: You could reference Figure 2 here since it’s the same plot.

Lines 224-225: Not every cold period is a Heinrich Event.

Line 261: In the text above, the “canonical” DOs were those described by Dansgaard.

Line 333: “event” is included in HE. Please provide a definition separating HEs from regular IRD events.

Line 350: “prevailed” may not be the best word choice here. How about “existed”?

Line 354-355: It’s not clear what “these results” refers to here. I’m guessing the authors mean the relationship between sea-level and GI duration? Please clarify.

Line 368-373: This mechanism was already posited by Shaffer et al. in 2004 (https://doi.org/10.1029/2004GL020968)

Figure 1: In paleoclimatic sciences and in this manuscript as well for most of the figures, the “Age” scale on the x-axis increases in values towards the right. I would advise the authors to either flip the figure around to make it consistent with the rest of the figures, or to use “time” instead as an x-axis with negative numbers if you want to keep the present on the right side.

Figures 5 and 6: x-axis has again been reversed, please flip the figure around or use “time” with negative numbers.

Table 2: “Last” is already included “LGM”.

Review of “Abrupt climate changes and the astronomical theory” by Rousseau et al.

Rousseau et al. provide a short review of astronomical theory and abrupt climate change. They plot some selected climate records using recurrence plots, and discuss the findings. They argue that DO oscillations are a type of internal oscillation, and that Bond cycles are formed through interactions with ice sheet volume.

It is unclear what the purpose of the manuscript is. Are the authors providing a review study, or original research? Unfortunately, the paper does not live up to the standards of either type of paper. It is not comprehensive enough for a review paper and does not provide an unbiased overview of relevant work. It further does not provide the kind of novel results and insight that would be the hallmark of a research paper. The two aspects are also not clearly separated. In both sections 2 and 3 one finds RP analysis and historical review mixed together. Given the shortcomings of the manuscript, I think the authors need to resubmit a very different paper for it to be suitable for Climate of the Past.

The introduction (section 1) as currently written bears no relationship at all to the main topic of the paper. It provides a short historical introduction to astronomical theory – given the short length of the section, it is necessarily incomplete. The overview stops in the 1970s, and it does not give the reader an idea of the recent ideas and challenges.

The authors suggest they are interested in the relationship between orbital and millennial-scale climate change. Then why not write an introduction / overview of the literature written on that topic instead? The authors do not acknowledge that there is a long history of such studies; these earlier studies should be evaluated and discussed instead.  The first such study is probably (McManus, Oppo, & Cullen, 1999), who linked DO variability to sea level. More recently, the Dome Fuji community members have presented a detailed study of the link between DO recurrence times and background climate (Kawamura et al., 2017). While these are probably the most important, many other studies should be listed also – my list is by no means complete: (Schulz, Berger, Sarnthein, & Grootes, 1999; Schulz, 2002; Schulz, Paul, & Timmermann, 2002; Sima, Paul, & Schulz, 2004; Buizert & Schmittner, 2015; Lohmann & Ditlevsen, 2018, 2019). Most of these studies are not cited.

The review given of abrupt climate change (mostly sections 2 and 3) are likewise not very  comprehensive or complete. They authors seem mostly interested in highlighting their own contributions. For example, the 2020 and 2021 papers by Bagniewski et al. (the same as the authors on the present paper) are given a detailed description (L255-263), while their method is not even used in the manuscript. Likewise, Boers et al. (2018) (with several of the current authors) is cited extensively throughout, while many seminal / standard papers on DO variability are ignored.

The interpretation of the recurrence diagrams is very subjective. The authors appear to visually identify “steps” in the RP diagrams, that are listed. However, it remains unclear what criteria were used to select these steps. What is the significance of such “steps”? Are these times that the climate system undergoes some transition? From looking at the records, in can just be a period of below-average variability. In most cases the steps from the RP diagrams are not meaningfully evaluated. By looking at the diagrams, it is unclear that I would have picked the same “steps”, adding to the sense of subjectiveness. The RP terminology is further not clearly defined. Terms like “drift topology” are used throughout, but not defined. Doesn’t this simply mean that there is a long-term trend in the underlying dataset? I am unclear what new insights, if any, have been gained using the RP.

Last, the paper has several statements that are either incorrect, or simply not supported by the available evidence. Most of the bullet points in their conclusions fall in the latter category.

For the benefit of revising their manuscript for future submission, I provided some minor suggestions by line number.

Line 17: “these processes varied considerably during the past 2.6 Myr” Where does this claim come from? I don’t think we know

Line 88: that ARE dominant

Line 96: “Recent” perhaps only compared to studies of the orbital theory.

Line 106: the structure of this section is somewhat unclear. The section provides more review-type writing, but also presents the methods used, the results, and their discussion.

Line 108- 112: This section adds little. Consider removing?

L142 to 144: The Barker record is artificial, and not a good reference for the onset of DO variability.  DO-like events have been observed 1.3Ma ago (Birner, Hodell, Tzedakis, & Skinner, 2016).

Line 157: “mere visual inspection”; isn’t that exactly how you evaluate the recurrence plots also? Visually?

Line 184: “recurrence analysis shows a drift topology”. Isn’t this just a fancy way of saying that there is a long-term trend in the data?

Drift topology is not formally defined. What does it mean in this context?

Line 187: What are these 5 steps based on? It seems to be a somewhat arbitrary pick. What are the criteria for selecting a pick? How robust is the number of steps to the selection criteria?

Line 195: Again, what is a drift topology precisely?

L204: we don’t know the CO2 concentrations during this interval very well. Van de Wal is cited as if it were a true reconstruction, which it is not of course.

L230: “This return generally happens in two steps, thus forming DO cycles of variable duration that

does not exceed a millennial time scale (Broecker, 1994; Boers et al., 2018; Boers, 2018).” I don’t know what the authors are trying to state, and why these references are used. The studies by Boers et al. don’t present any original data, and any estimates of DO timescales have been given by earlier authors.

L242: The 1982 Dye 3 core already confirmed the rapid events seen in Camp Century

L255 – 263: I don’t understand the goal of discussing this. The KS analysis is not used in the manuscript, is it?

 L263: “with Southern Hemisphere warmings occurring prior the Northern Hemisphere ones.” A better way to describe their phasing is an integrator / integrand relationship. Also, the Antarctic and Greenland ice cores are not representative of their respective hemispheres of course.

L271: It appears here that the authors confuse the ideas or propagation direction of the climate signal, and the direction of the heat transport. Oceanic heat transport in the N Atlantic is northward, but that does not mean that DO events originate in the Southern Ocean. One of the few studies suggesting a true South-to-north direction is Knorr and Lohmann (2003); most others all suggest N-to-S, despite the direction of heat transport being S-to-N.

L278: There are many good models of DO dynamics, this is a case of self-citation.

L305: ”the length of the GIs appears to be related to the mean sea level.”. Variations on this observation has been made several times by various authors. See my list of suggested papers at the beginning of this paper. Also, I don’t see how or why this is derived from the RP.

L317: The first naming of the Bond cycles comes from this paper (Lehman, 1993), and not from the papers cited.

L350: it goes back to at least 1.3 Ma, and perhaps further (Birner et al., 2016).

L371: How can a 2011 study confirm something suggested in a 2018 paper? In this case the 2011 paper already suggested it.

 L398: “This fairly well-agreed-upon fact leads support to the interpretation of the enhanced millennial variability during glacial times as arising from an internal oscillation of the climate system — as proposed by several authors”. I don’t follow (or agree with) this line of thinking. Yes, the DO events are not forced by orbital variability, but it could still be forced by something else – such as internal ice sheet variability.

Line 410: aren’t abrupt changes identical to these phenomena?

Line 412: none of the material presented provides any evidence for an internal oscillation. While the internal oscillation is certainly the most commonly held view in the field, this is not proven – certainly not by the authors in the present paper.

Line 415: This is again pure speculation. There is no evidence provided that Bond cycles are linked to the NH ice sheet extent. For example climate model experiments would be needed to prove this.

Line 419: It is observed as early as 1.3 Ma, but possibly earlier (Birner et al., 2016)

Line 427: I agree that it modulates their period, but I have not seen any evidence for their amplitude being modulated.

Line 429: I am not sure this is supported by the available evidence. Birner et al. (2016) suggests for MIS 38 and 40 the DO-type variability is indistinguishable from MIS3, despite presumably very different ice sheet sizes in the NH (40ka world vs. 100 ka world). So much for NH ice sheets being important!

References used in this review:

Birner, B., Hodell, D. A., Tzedakis, P. C., & Skinner, L. C. (2016). Similar millennial climate variability on the Iberian margin during two early Pleistocene glacials and MIS 3. Paleoceanography, 31(1), 203-217. doi:10.1002/2015PA002868

Buizert, C., & Schmittner, A. (2015). Southern Ocean control of glacial AMOC stability and Dansgaard-Oeschger interstadial duration. Paleoceanography, 30(12), 2015PA002795. doi:10.1002/2015pa002795

Kawamura, K., Abe-Ouchi, A., Motoyama, H., Ageta, Y., Aoki, S., Azuma, N., . . . Yoshimoto, T. (2017). State dependence of climatic instability over the past 720,000 years from Antarctic ice cores and climate modeling. Science Advances, 3(2), e1600446. doi:10.1126/sciadv.1600446

Lehman, S. J. (1993). Ice sheets, wayward winds and sea change. Nature, 365(6442), 108-110.

Lohmann, J., & Ditlevsen, P. D. (2018). Random and externally controlled occurrences of Dansgaard–Oeschger events. Clim. Past, 14(5), 609-617. doi:10.5194/cp-14-609-2018

Lohmann, J., & Ditlevsen, P. D. (2019). Objective extraction and analysis of statistical features of Dansgaard–Oeschger events. Clim. Past, 15(5), 1771-1792. doi:10.5194/cp-15-1771-2019

McManus, J. F., Oppo, D. W., & Cullen, J. L. (1999). A 0.5-Million-Year Record of Millennial-Scale Climate Variability in the North Atlantic. Science, 283(5404), 971-975. doi:10.1126/science.283.5404.971

Schulz, M. (2002). The tempo of climate change during Dansgaard-Oeschger interstadials and its potential to affect the manifestation of the 1470-year climate cycle. Geophysical Research Letters, 29(1), 2–1-2–4. doi:10.1029/2001gl013277

Schulz, M., Berger, W. H., Sarnthein, M., & Grootes, P. M. (1999). Amplitude variations of 1470‐year climate oscillations during the last 100,000 years linked to fluctuations of continental ice mass. Geophysical Research Letters, 26(22), 3385-3388.

Schulz, M., Paul, A., & Timmermann, A. (2002). Relaxation oscillators in concert: A framework for climate change at millennial timescales during the late Pleistocene. Geophysical Research Letters, 29(24), 2193. doi:10.1029/2002gl016144

Sima, A., Paul, A., & Schulz, M. (2004). The Younger Dryas—an intrinsic feature of late Pleistocene climate change at millennial timescales. Earth and Planetary Science Letters, 222(3), 741-750.