Review by Daniel Parkes and Ian Candy (Royal Holloway, University of London) of Reorganization of Atlantic waters at sub-polar latitudes linked to deep water overflow in both glacial and interglacial climate states. Holmes et al., 2021

Paper summary

The stated aim of the paper is to investigate abrupt climate events during interglacials/warm climates, with the authors highlighting that these are “generally attributed as a characteristics of glacial climate states”. In order to do this the paper looks at changes in both the surface and deep ocean at site DSDP 610, the southern extremity of the Rockall Trough, during the transition from MIS 11c – MIS 11b. The paper is based on a multi-proxy approach, including; 1) Benthic stable isotope data of 18O and 13C across 82 samples at an 8cm resolution, 2) Lithic abundance (IRD) at 0.5, 1, and 5cm resolution, 3) Faunal counts on forams at 0.5, 2.5, and 5cm resolution for SST reconstruction using MAT and 4) Grain size analysis for bottom flow strength. The papers main conclusions are that at least 2 abrupt climate events (the 1^st lasting 7000 yrs and the 2^nd last 2000 yrs) occurred during the transition from MIS 11c to MIS 11b and that this challenges the idea of abrupt events being exclusive to glacial periods. Furthermore, the authors argue that the offset between proxies indicates that there is a lag between the first slowdown of Wyville Thomson Ridge Overflow Waters (WTOW) and the subsequent cooling of SST of ca 320 years and ca 710 years for the first and second event. In both cases WTOW recover slower than SSTs.

Summary of decision

Whilst the paper is data rich and represents an interesting new record we have a number of significant issues that we feel need to be resolved prior to the paper is accepted for publication. We have flagged this as majot revision as we feel that it needs a re-thinking of some of the underlying concepts but it may be that the changes required can be dealt with more rapidly than this and the changes may be relatively minor. Firstly, we feel that the way that the paper is set up does not really give a true reflection of the papers findings. The paper claims that abrupt climate events are; 1) a characteristic of glacial climate states, and 2) the results of this study (showing abrupt events occurred at the transition between MIS 11c and 11b) question this and our ideas about our concept of warm climate stability. Neither of these claims are strictly true. It has long been known that abrupt climate events occur during interglacials – the 8.2 ka event (along with the pre-boreal oscillation, 4.2 ka event and 2.8 ka event) is a good example of this. Whilst it is likely that none of these were of a magnitude or duration comparable to the Lateglacial Interstadial (i.e. the Younger Dryas) they had transformative effects on ecosystems, surface processes and societies so are still significant. The concept of “warm” climate stability was surely abandoned a long time ago. If the authors are distinguishing between high magnitude (glacial) and low magnitude (interglacial) abrupt events then they need to do so more clearly and, ideally, in a quantified way. Regardless of how abrupt events are defined, the study presented here doesn’t move this argument forward. The events discussed here occurred on the climatic downturn into MIS 11b and, consequently, long after fully interglacial conditions had ceased. They are actually more true of abrupt events under a glacial, or transitional, state in that, as shown in Figure 6, a significant fall in sea level had already been experienced prior to their occurrence. A number of authors have shown that abrupt events occur under fully interglacial conditions during MIS 11c (i.e. Barker et al., 2015; Kandiano et al., 2017). The events described here are more similar to those that occurred during the transition from MIS 5e to 5d that are discussed in the introduction. The occurrence of such events, at interglacial/glacial transitions, are relatively well-known, particularly from the North Atlantic, which slightly detracts from the originality of this study. The paper needs to consider the rationale and significance of this work in much greater detail, in terms of how it is discussed in the abstract, introduction and conclusion.

Secondly, the Rockall trough, as a system, is a hydrographically unique area with cyclonic re-circulation of sediments. Consequently, changes in sediment characteristics within DSDP 610B could reflect variations in the strength of flow within Rockall Trough and not just actual changes in the strength of WTOW. The authors may have considered this but at the moment the manuscript reads as though the complexity of this location is being ignored and overlooked. The paper needs to show a much greater consideration of the complexities of the oceanographic processes that operate at DSDP 610B and explain why the proxies record the role of WTOW and not more local processes.

Thirdly, we are not convinced that the cooling that occurred between 397 and 390 ka should be classified as an “abrupt event”. Not only does the event last for some 7,000 years but it is characterised by relatively slow and protracted cooling. This is relatively clearly seen in the SST data presented here where a decline of some 6^oC occurs progressively over ca 5 ka. The second event that is described is, as the authors acknowledge, is much more typical of an abrupt event (a decline of  >6^oC in ca 0.5 ka) though this is confidently outside the main interglacial phase and thus does not support their conclusions of high magnitude events during interglacial periods. The change in grain size data for the first event is more dramatic and has much in common with the second event, however, this elicits a very different response in SST values and this is not really acknowledged or addressed. It is also quite important for the authors to discuss the discrepancy in the SST data of late MIS 11c between DSDP 610 and M23414. From 403 ka to 398 ka there is an offset of upto 6^oC between the two sites, significantly greater than the modern temperature gradient between these two localities. Again this isn’t discussed but is fundamental to an acceptance of the data and ideas presented here. The difference between the two events need to be discussed and explored in much greater detail, whilst the validity of the SST estimates for DSDP 610 need to be discussed in more detail, particularly with reference to the record from M23414.

Further points to be considered are:

• Line 328 – 334 – It seems odd and out of place to make this Holocene comparison without any accompanying graphs or quoted data to support this.
• Line 396 – 400 – authors argue that C13 data for 980 and U1308 reduce at 398.4 and 399ka respectively. This isn’t that convincing – at 980 values increase during the event / at the onset of the event. U1308 also reflects a gradual reduction. The resolution seems too low to claim that the data depicts WTOW depleting during this event. The authors also do not discuss other causes for this.
• Line 426 – This begins with wider palaeoceanographic context, but this has already been stated in the previous 2 paragraphs. Perhaps the whole discussion needs segmenting better? M23414 has been discussed (indeed in the figures) whilst other sites are being discussed (U1308, 980, MD99-2277) in the previous.

Furthermore there are series of more minor points that need to be considered:

• Figures: Some general points on uniformity of font sizes, writing (610B – this study) or (610B) or (this study) rather than a combination of the 3
  • Figure 1
    • Indicating which branch of ISOW is WTOW would be helpful for the reader, particularly as this is the focus of your study. It may also be helpful to include other labels (DSOW, other ISOW branches, NADW etc but not necessary).
    • The grey site label names on a grey background are something I’d advise to change for legibility
  • Figure 2
    • In some figures you have labelled the data for this study and in figure 2 you have not. I’m assuming those not labelled relate to DSDP 610 in this study?
    • You’ve plotted N.incompta + N.deutertrei % together from Kandiano et al 2007 but you’ve listed this as sub-polar (following Kucera et al 2007); but Deutertrei is a sub-tropical species. You also don’t talk about this in the text so wondered why it is plotted?
    • The two shades of each colour may be difficult for colour
    • Colour blind people so symbols may help with this.
    • The graph feels busy – it might benefit from extending horizontally
  • Figure 5
    • IRD for 983 is in grey in the axis and appears black in the graph
    • The range of font sizes looks untidy
  • Figure 6
    • IRD for 983 is in grey in the axis and appears black in the graph
    • The NPS % from barker et al 2015 was updated in the barker et al 2019 paper. There aren’t any major changes in %, though some minor peaks are smaller in the data you have used (e.g., ~ 394ka)
    • Is the age model for ODP 983 tied to DSDP 610B in any particular way or just placed on a timescale with 2 different models (+ associated uncertainties)? I ask as the LR04 age model for ODP 983 places the increase in NPS % at ~ 395 ka – 389 ka and a second 387-385 ka which is much closer to the authors claims for these events (in both duration and timing). The authors also seem to have tied their core to LR04 so I wonder why (it seems) they have not tied ODP 983 to this.
  • Text: Some general points line by line through the text. The piece in general would benefit from some sub-headings to organise the flow as presently sections seem to overlap considerably.
    • Line 74-75 – would cite Barker et al 2015 “icebergs not the trigger for NA cooling events” (they reference the paper later but they do not cite it here)
    • Line 80-84 – quantifying importance of NADW to AMOC and quoting overall contribution from WTOW would be good
    • Line 90 – Global average temperature difference between MIS 11 and MIS 1 would be good
    • Line 109 – add space between reference and ‘today’.
    • Line 85 – 115. This seems muddled. It seems to be a descriptive piece setting out the conditions during MIS 11 but starts and ends with talk about orbital similarities as a justification for looking at MIS 11. I would have set out the orbital similarities prior to then describing MIS 11.
    • Line 347 – 350 – the authors state that the offset is 9 samples (4.5cm), 320 years. Firstly, from the graph, it doesn’t look like all 9 of these samples have been run so this is confusing wording
    • Line 404 – random ‘o’ in the sentence
    • A good paper to cite on surface waters in the Nordic seas being unusually cold and fresh in MIS 11 https://www.frontiersin.org/articles/10.3389/fmars.2018.00251/full#h7 which is absent in the bibliography
    • The reference list needs to be checked there are a number of typos throughout and some repetition (i.e. McManus et al., 1999 is included twice).

Review of Holmes et al., Reorganization of Atlantic waters at sub-polar latitudes linked to deep water overflow in both glacial and interglacial climate states.

In this paper authors use multiple proxy retrieved from one core (DSDP-610) over a part of the MIS 11 interglacial. They produced new data of oxygen and carbon isotope for benthic foraminifera Uvigerina, IRD counts, planktic foraminifera abundance, x-ray fluorescence and grain size analysis. They used x-ray fluorescence to build the preliminary age model which was further developed using the δ¹⁸O of benthic foraminifera. They used the grain size data to perform an end member analysis, which was then used to infer changes in the flow intensity of the Wyville Thomson Overflow Water. SST reconstruction was done using modern analogue technique from the planktic foraminifera abundance data. The authors then compared their results with cores from the sub-polar North Atlantic.

The authors used their new data to investigate short term variability of both surface and deep circulation within the interglacial to glacial inception. Authors aims to produce new knowledge that would help better understanding present-day weakening of the AMOC and threshold of ocean circulation stability.

General comments:

The authors are investigating the coupling between surface and deep-water, which is very important. However, I wonder why there is not more comparison with surface condition in the Nordic Seas, as the deep-water at site 610 will be mostly influenced by convection in the Nordic Seas. Moreover, there is a growing body of evidence of the uniqueness of the Nordic Seas during MIS 11 and its key role on North Atlantic circulation. It would be interesting to see how these data compared and a discussion about the mechanisms involved. Authors may try to focus on finding mechanistic explanations of the linkages between surface and deep water and go a bit deeper into how this might be relevant to actual climate. What we need to advance the field is a better understanding of the mechanisms behind these critical climatic feedbacks, as we already know very well that interglacials are not stable.

I do not understand the choice of timescale investigated here. I would have been interested to see how these data compare from the peak interglacial and the evolution toward the glacial inception rather than focussing only on the 390-399 ka period. This gives a snapshot without much comparable and not much room for more meaningful interpretation of climate evaluation throughout this important climatic period.

In term of structure, I found the paper rather difficult to follow.

I also noted some results and even interpretation presented in the method section, especially in the grain size and chronology section. It makes it more difficult to clearly understand what the original data from this paper are and what is based on literature. The results section starts with a general statement discussing the results, I suggest trying to refrain interpretating the results within this section and concentrating the discussion and interpretation in the discussion section. As an example, I would suggest presenting the results in in term of the actual proxy (e.g., grain size variation, foraminifera assemblage) and explain how there are used to reconstruct WTOW and temperature in the discussion section only. Therefore, I recommend a major rework on the structure of the method, results, and discussion sections.

Specific comments:

The last sentence of the abstract is misleading in my opinion, first because this paper does not provide evidence that the changes, they observed are of similar magnitude than their glacial counterparts. Secondly, while this paper might add some evidence, the concept of stable interglacial climate, was already challenged and altered in the past (Bauch, Kandiano, Dickson, etc). Finally, while within an interglacial period, most of the data are not within the peak interglacial, and therefore already within a (long) transition phase as depicted by the ice core data, IRD, etc, so it is not very surprising to see this kind of variability.

Introduction: Most of the discussion on deep water is based on the WTOW, please introduce its significance to AMOC, climate, its geometry, etc.

Introduction: I would suggest focusing the introduction on the interesting climate feedback that are investigated within the present study and how the uniqueness of MIS 11 rather than the relatively outdated and mostly settled argument on stable interglacials.

L83: Caesar et al., is a brief communication and not a research article per se, I suggest referring to the original research publications e.g., Rahmstorf et al., 2015, etc.

L230 Whole paragraph: It is unclear to me if this should be in the method or results

L275 Chronology section: I feel most of this should be in the discussion, as the method to acquire the data used here (δ¹⁸O, XRF) were already presented. There is a great deal of interpretation to build the chronology.

L330 whole paragraph; not necessary in the results section

L335: One suggestion is to build the results section similarly to the method section, so the reader can easily spot the original data provided by this study. Beware of interpretation the results within this section, the data should be presented here (XRF, grain size, δ¹⁸O), but the link to what they are used for (SST, WTOW, etc) should normally go into the discussion.

L385: I would suggest adding sub-sections to the discussion to try to better structure the arguments that are built here. I find the discussion very hard to follow as of now and I often find myself wondering what exactly the authors want to communicate. I suggest trying to avoid excessive description of published work and instead really focus on new findings.

L400: The relationship between fresh and cold water and IRD is not as definitive in the Nordic Seas compared to other regions of the North Atlantic (Doherty et Thibodeau, 2018). See work from Kandiano for salinity reconstructions using alkenones.

L485: what would trigger the release of freshwater in a low insolation, cooling climate?

L485: Is this weakening of the AMOC seen elsewhere? Could it rather be a change in the geometry of the AMOC that leads to this change in your proxy? How does it compare to general trend of AMOC during MIS 11 (e.g., Vazquez Riveiros et al., 2013)

L523: Robinson et al., (2017) model of GIS dynamic also supports this timeframe I believe.

L600-605: maybe it would be good to plot the precession signal somewhere so the reader could appreciate the potential link between these events.

L605-610: I am not sure I follow what the authors want to say by: irrespective of magnitude or boundary conditions and on what data/evidence this is based.

L620: Rate, volume…what about the location of the freshwater input?

Figure 2 and 6: the grey bands are not described in the caption (and I believe they are not highlighting the same as in figure 5…)

Figure 6: Panel d, there is two datasets listed, but three lines. I suspect this is what is referred to in the last sentence of the caption, but I then I wonder why only one dataset as running average?

Bauch, H. A. “Interglacial Climates and the Atlantic Meridional Overturning Circulation: Is There an Arctic Controversy?” Quaternary Science Reviews 63 (March 2013): 1–22. https://doi.org/10.1016/j.quascirev.2012.11.023.

Dickson, Alexander J., Christopher J. Beer, Ciara Dempsey, Mark A. Maslin, James A. Bendle, Erin L. McClymont, and Richard D. Pancost. “Oceanic Forcing of the Marine Isotope Stage 11 Interglacial.” Nature Geoscience 2, no. 6 (2009): 428–33. https://doi.org/10.1038/ngeo527.

Doherty, John M., and Benoit Thibodeau. “Cold Water in a Warm World: Investigating the Origin of the Nordic Seas’ Unique Surface Properties during MIS 11.” Frontiers in Marine Science 5, no. AUG (2018): 251–251. https://doi.org/10.3389/fmars.2018.00251.

Kandiano, Evgenia S., Marcel T.J. van der Meer, H. A. Bauch, Jan Helmke, Jaap S.Sinninghe Damste, and Stefan Schouten. “A Cold and Fresh Ocean Surface in the Nordic Seas during MIS 11: Significance for the Future Ocean.” Geophysical Research Letters 43, no. 20 (2016): 10,929-10,937. https://doi.org/10.1002/2016GL070294.

Kandiano, Evgeniya S., and H. A. Bauch. “Phase Relationship and Surface Water Mass Change in the Northeast Atlantic during Marine Isotope Stage 11 (MIS 11).” Quaternary Research 68, no. 3 (November 2007): 445–55. https://doi.org/10.1016/j.yqres.2007.07.009.

Kandiano, Evgenia S., H. A. Bauch, Kirsten Fahl, Jan P. Helmke, Ursula Röhl, Marta Pérez-Folgado, and Isabel Cacho. “The Meridional Temperature Gradient in the Eastern North Atlantic during MIS 11 and Its Link to the Ocean-Atmosphere System.” Palaeogeography, Palaeoclimatology, Palaeoecology 333–334 (May 2012): 24–39. https://doi.org/10.1016/j.palaeo.2012.03.005.

Rahmstorf, Stefan, Jason E. Box, Georg Feulner, Michael E. Mann, Alexander Robinson, Scott Rutherford, and Erik J. Schaffernicht. “Exceptional Twentieth-Century Slowdown in Atlantic Ocean Overturning Circulation.” Nature Climate Change 5, no. 5 (March 2015): 475–80. https://doi.org/10.1038/nclimate2554.

Vázquez Riveiros, Natalia, Claire Waelbroeck, Luke Skinner, Jean-Claude Duplessy, Jerry F. McManus, Evgenia S. Kandiano, and Henning A. Bauch. “The ‘MIS 11 Paradox’ and Ocean Circulation: Role of Millennial Scale Events.” Earth and Planetary Science Letters 371–372 (June 2013): 258–68. https://doi.org/10.1016/j.epsl.2013.03.036.

General comments

The study focuses on paleoceanographic changes in the high latitudes of the North Atlantic during the 403-388 ka interval based on the multi-proxy analysis of the DSDP 610B site located on Feni Drift. By comparing existing data for other sites in the North Atlantic subpolar zone, the paper aims to reconstruct the sequence of events on a spatial scale.  The quality of the multi-proxy data, which have an excellent temporal resolution, and the new information provided clearly deserve a publication in Climate of the Past. The paper nevertheless presents some structural and conceptual problems that require some modifications before publication.

Specific comments

The focus of the paper on the concept of interglacial climate stability is somewhat disturbing because the period in which the authors show circulation changes and the presence of IRD does not correspond to the interglacial period of MIS 11 but rather to the glacial inception. The episode starting at 397 ka is associated with the very end of the MIS 11 interglacial period or even marks the beginning of MIS 11b. The paper should be re-worked to aim to study the reorganization of the circulation in the high latitudes of the North Atlantic during the glacial inception and not to test the stability of warm climates since the detected circulation changes are not occurring during the course of the interglacial period (even if we remain in an interglacial isotopic stage).

Nevertheless, I agree with the authors that even if the ice volume increases, it is still low during the first NDW reduction interval. If we follow the theory put forward by McManus in 1999, the threshold isotopic value of 3.5 per mil is not exceeded during this interval (it will only be exceeded during the 390 ka episode) which indeed suggests that the ice volume would still be too small to drive millennial events related to ocean-ice feedback mechanisms, leading to changes in the strength of the MOC and changes in interhemispheric heat transport. High latitude circulation changes are apparently happening despite the still low ice volume but the manuscript does not highlight that the episode between 397 and 392 ka corresponds to a typical millennial-scale event, i.e. an event widely recorded in the North Atlantic SST (from high latitudes to subtropics) and characterized by bipolar see-saw. The authors may also want to discuss the possibility that the circulation conditions during these 5000 years (that is quite long duration for a millennial event) may reflect orbital variability or processes associated with the glacial inception during the obliquity minimum.

The authors cited the paper of Oliveira et al. (2016) mentioning that SST does not record millennial cooling during the 397-397 ka interval but it would have been interesting to note that centennial events are detected in southwestern Europe by the pollen record of the same site (IODP Site U1385, same paper). 3 rapid climatic events showing forest reduction occured at 396 and 393.5 ka without associated SST changes on the Iberian margin and one at 390 associated SST change. It was noted that the first 2 episodes are of lower amplitudes than the 390 ka event, during which the ice volume is larger and for which the other ODP 980 and 983 sites record iceberg discharge episodes.  The correspondence is striking enough to be cited, potentially suggesting that the discharge events detected by DSDP site 610B are potentially coupled to atmospheric changes affecting mid-latitude climate.

Structure: Almost all the figures are called in the method section even though they show results or even a comparison with data from the literature. It would seem more appropriate to me to call these figures in the result section. In my opinion, the paragraph between lines 228 and 235 corresponds to results.

Chronology: Did you keep the original age models from core ODP 983 and U1308? It is difficult to justify that it is better to correlate with a record from the same region (whose age model is based on Martinson (1987) benthic isotope stack) instead of LR04 and then compare with the benthic records from those sites, one of which based on EDC 3 age model and the other on LR04 and later, mention that there is a good correlation between the records. This procedure is lacking of consistency.

Technical corrections

• Summer insolation and astronomical parameter curves should be added in Figure 5 and/or 6.
• The limit of the isotopic substages should be indicated in at least one figure.
• The correction of isotopic values used to present the ODP 980 & 983 and U1308 sites in Figure 4 should be indicated.
• May be great to add in Figure 6 the curve of NPS % from site DSDP 610 with those of ODP Site 983.
• 328: Beginning the results section with a sentence mentioning the similarity to the Holocene does not seem appropriate.
• 404: “Further our results”.
• 473: Values of CO2 and CH4 are given with a too high precision without considering age uncertainties.