Hauge Braaten et al. present a combination of thermometers (Mg/Ca and D47) applied to benthic foraminifera from 2 sites (in the Pacific Ocean and in the Atlantic Ocean) to reconstruct the bottom water interaction between the 2 oceans during the mid Piacenzian. This article presents a significant amount of measurements and high quality conclusions. It deserves to be published. I have a few suggestions, mostly on salinity reconstructions and minor comments.

Regarding the salinity reconstruction, I would suggest being more critical of your methodology. The relationship between d18Osw and salinity changes over time. Assuming that the relationship during the mid Piacenzian is similar to today's is probably wrong. Although the authors mentioned it, they assume that their reconstructions of salinity is correct and I suggest being more careful in interpreting your data. Additionally, the authors should add a discussion of the uncertainties in thier estimated salinity. Is a change of 2 PSU, using this methodology, really significant? (Line 377). Same question for a salinity difference of 0.2 between the Atlantic Ocean and the Pacific Ocean (Line 391). Finally, did the authors take into account the global changes during this period (based on sea level change for example) in order to obtain the local salinity signal?

The preservation and contamination tests presented in supplementary material are not mentioned in the text. I found it confusing to only read that the foraminifera are well preserved without any justification, until I read the supplementary material. I would suggest, at least, mentionning the supplementary material.

Lines 209, 254 and 255: do the authors have uncertainties at 1 or 2 sigma?

Line 281: I would mention that the author are still talking about Mg/Ca derived temperature in this paragraph.

Line 295: MIS not MI2

Figure 2 and 3: It would be great if the authors could also add the uncertainties on the Mg/Ca temperatures

Lines 395-396: Can the authors expand on this sentence, please?

Line 410: What are the authors referring to? Model, data? References are missing

Lines 458-462: How did the authors conclude on the ice volume proportion of thier d18Osw parameters? How did they remove the salinity part of the d18Osw, as explaining previously?

In this manuscript, Braaten et al. present a multi-proxy record of deep ocean temperature and d18O of seawater over the mid-Pliocene warm period and the M2 glaciation in the North Atlantic and Equatorial Pacific. I found the manuscript well written, and they presented compelling evidence that the Atlantic and the Pacific were quite different in temperature and d18O of seawater. The implications of their records are that the characteristics of deep-water formation were very different during the Pliocene with the northern component water being very warm and very salty.

I have two moderate requests for the authors.

• In the supplemental, the authors present cross plots of the El/Ca to investigate possible contamination. Could you also please include the El/Ca and Mg/Ca with time on the x-axis?  Are there temporal trends in the Mg/Ca data that could be explained by contamination?  The reconstructed temperatures here and Jakob et al., 2020 are quite high so it would be helpful to know if there are temporal patterns.
• The authors did a nice job of propagating the error on the clumped isotope measurements, but the Mg/Ca did not have the same consideration. Could you please estimate the error on the temperature and d18Osw calculations? I usually use PSUSolver by Thirumalai et al 2016 (doi:10.1002/2016PA002970) which you can find here: https://www.mathworks.com/matlabcentral/fileexchange/59565-paleo-seawater-uncertainty-solver

Minor comment:

Figure 2, (d) panel labeled for eNd

The authors of this study present Mg\Ca and clumped isotope bottom water temperature records for the Late Pliocene from Site 849 (3346m) in the Pacific, and U1308 (3426m) in the Atlantic. They find that the Atlantic site was warmer (>4C as compared to ~1C warmer in the modern Ocean) and saltier than the Pacific site during this time. The authors attribute this to a different ocean circulation regime under which there is a more limited water mass exchange between the deep Atlantic and Pacific basins. Based on the amplitude of the cooling observed at both sites during MIS M2 they suggest that the benthic delta18O changes associated with this cold stage were mostly driven by changes in the deep ocean temperature rather than ice volume.

Overall, I found the new records exciting and very insightful and the manuscript rigorous and well structured. As a modeler I am not in the position to provide a critical assessment of the analysis methods but can offer an assessment of the dynamical interpretation. While the authors are careful to be vague as to mechanism in the abstract and conclusions, I am not completely convinced of the tentative explanation for the larger difference in water mass properties between the two basins stated in the discussion section (Ln 427-429). Given the circumpolar nature of the Southern Ocean and that already ~80% of modern deep water is upwelled in the Southern Ocean it is not clear how this would be enhanced in the Pliocene and lead to a reduction in the inter basin exchange of deep water. It would be good if the authors could point to a study with ocean simulations that reproduce the proposed type of circulation regime change. On the other hand, Pliocene simulations with north Pacific deep water formation result in fresher North Pacific deep waters relative to preindustrial (Burls et al., 2017, Fig. 5) and reduced warming relative to the Atlantic (Burls et al., 2017, Fig. 4). So it seems both scenarios should be considered in the discussion.

The backing out of salinity estimates is a nice part of the manuscript but as one of the other reviewers mentions some more details are needed explaining how the ice sheet contribution was handled. This should not affect the basin gradient though. The limitations/robustness of assuming modern relationships should be discussed e.g. see Fig. S7 in Gaskell et al., (2022).

Minor comments:

Ln 405-406:“Slightly less saline” perhaps add modern values to Table 2 so that the reader can assess just how much fresher for the Pacific.

Fig 2d: The incorrect axis label is provided; it should be insolation and presumably units of W/m^2.

Fig 2e: Why isn’t the 5pt running mean shown as in the other panels?

Fig 5c: Modern d13C values are missing and would be helpful for reference.

References Cited:

Burls, N. J., Fedorov, A. V., Sigman, D. M., Jaccard, S. L., Tiedemann, R., & Haug, G. H. (2017). Active Pacific meridional overturning circulation (PMOC) during the warm Pliocene. Science Advances, 3(9), e1700156.

Gaskell, D. E., Huber, M., O’Brien, C. L., Inglis, G. N., Acosta, R. P., Poulsen, C. J., & Hull, P. M. (2022). The latitudinal temperature gradient and its climate dependence as inferred from foraminiferal δ18O over the past 95 million years. Proceedings of the National Academy of Sciences, 119(11), e2111332119