In this submission, the authors tackle the subject of opal sedimentation and siliceous microplankton productivity increase across the Eocene-Oligocene transition. The manuscript is well-written and presents a balanced discussion, considering and exploring a broad spectrum of viewpoints on cooling at the EOT. An important addition of this study to the existing literature is calculating absolute concentrations of not only diatoms but also radiolarians. In general, I recommend publication with some revisions, minor to major, including additional methods. Detailed comments are listed below. 

General comments:

When talking about bulk opal measurements, authors predominantly refer only to diatoms and radiolarians. On occasion, sponge spicules often comprise a large proportion of biogenic silica in Eocene sediments. Considering their overall much larger size, even smaller abundance could contribute to equal BSi value. Further silicoflagellates,  ebridians, and archaeomonds can reach higher concentrations as well. I think it's worth including it, either in the introduction or the discussion in case bulk opal measurements cannot be explained only by diatoms and radiolarians (with the current representations of the graphs it is difficult for the reader to determine). 

Fig. 2.d. It is confusing and difficult to correlate when bulk opal accumulation and diatom accumulation are overlapping and on different scales. I would suggest changing it to the same scale or presenting it on separate graphs, as I can imagine the diatom ranges would completely flatten on the same scale. Potentially making a graph where diatoms and radiolarians are on the same scale and bulk opal accumulation rates are e.g., above would be easier and clearer to analyze. 

Fig. 2. I would suggest making a line or shaded area across the graph to mark the E/O. It would be easier to read the results. 

Fig. 3. As mentioned above radiolarian accumulation rates could be shown in one figure with diatoms. It is difficult to compare across figures. Also please be consistent with figure labeling and presentation. In Figure 2 diatom accumulation rate on the graph is written without the unit, in Figure 3 radiolarian accumulation rate is written with the unit. 

Fig. 4. Likewise I suggest marking the E/O across the graph. 

In line 418 authors write "...further questioning the assumption that these two metrics, diversity and abundance, are directly linked, or that observed diversity is primarily controlled by preservation." I indeed agree, that diversity is not necessarily controlled by abundance, or as authors point it is a complex feedback loop. However, preservation is important to consider. It's often easy to observe whether assemblage is well/poorly preserved when observing how the frustule presents itself on light and scanning electron microscope. Lack of lightly silicified genera, which are characteristic of Eocene such as Asteropmhalus or small genera could indicate higher dissolution rates. I think overall it is always worth making general notes on the preservation state. Potentially authors can go back to slides and make general observations about whether high diversity, high abundance samples have better frustule preservation.  

Methods comments:

Was the sieved fraction (below 10 um) checked for the diatoms? Diatoms in genera eg. Actinoptychus, Paralia can be as small as 6-8 um and still easily distinguished under the light microscope. Overall I doubt omitting this fraction caused a large bias in the data, however, sieved fraction when looking at the abundances should always be checked. 

My biggest comment would be about how the authors calculated absolute abundances. The formula used by authors is indeed used broadly in the literature, especially for paleo studies. Another method, which in my subjective opinion, is more accurate and less biased for such calculations is divinylbenzene (DVB) microspheres (Battarbee and Kneen, 1982). In these calculations, only dry sediment weight and microsphere concentrations are needed, which would introduce fewer potential errors than the authors' calculations. It might be beneficial to look at a few samples (e.g. 5) and use microspheres to establish concentrations and compare the results. This method is more used in freshwater studies on diatoms, however, it is a good chance to introduce it to the paleo world as well.     

Technical comments

Line eg. 51, 76, 80: missing comma in the citation. Please check the whole manuscript for this type of correction.

In line 52, the authors write "...poorly constrained timing of the SO gateway...", the reader can gather SO corresponds to the Southern Ocean however for clarity in line 49, the authors should write "...focuses particularly on the deepening of Southern Ocean (here and after SO) gateways."

Line 54: ca (circa) please correct to ca.

Line 67: Please correct CO2 to CO₂.

Line 235: Be consistent with the usage of hyphens and en dash. For age ranges always use en dash. Please check and correct where necessary through the text.

Lines 238–240: Please avoid one-sentence paragraphs. Try to incorporate in previous.

I enjoyed reading the manuscript. Best of luck with the corrections. 

Thank you for giving me the opportunity to review this manuscript. As pointed out by the referee 1, the manuscript is overall well written and pleasant to read. It focusses on new diatoms and radiolarian data as evidence for increased carbon export preceding the Eocene Oligocene transition. Overall, I recommend this manuscript for publication after minor to major revisions. I might sound a bit harsh in the comments below, but I did enjoy reading this manuscript, and I think it is worthy of publication after revisions. As you'll see, most of my comments are minor.

Comments:

-Overall, although nicely written, the manuscript is quite long, and I would suggest to try to shorten it by being more concise.

-Abstract: I feel like you talk a lot about radiolaria here but maybe under-used them in the main text?

-L36: "ocean cooling" maybe add "by X °C"?

-L39: missing dot after reference.

-L98 to L102: this has already been developed in the introduction earlier, avoid repetition. Overall the introduction is quite (too) long in my opinion.

-L102 and throughout the text: the reference to Özen et al., submitted is not accepted and not available to read. Lots of sentences rely on this reference which is yet to be accepted, so I am not sure about the validity of using it yet.

-Methods: please add the latitudes and longitudes of investigated core sites, especially as you investigated records of various locations, it is important to quickly find this information. If possible also add the paleolocation?

-Regarding this, it would be nice to have latitudes and longitudes available on figure 1 too. Maybe this figure can be use to show current opal export too for example, instead of just depth?

-L175: is "ab" the absolute abundance?

-L190: again reference not available

-L191: Capital T after dot

-L198: Maybe define MAR?

-Fig2: maybe connect the points too for panel C? Like in D and E?

Not sure what to suggest but panel D contains lots of graphs and is not easy to read.

-L216: use of "fluctuate" is a bit strange

-L228-229: last sentence sounds a bit strange

-L232: Taking age models into account, and so on, I'm not sure about "coincided", as we have a 0.5 to 1 ma gap while we could expect a faster / synchronous response?

-L235: "corresponds" Maybe use "is concomitant" or something similar? To me, "corresponds" induces a causal effect.

-L238-240: just a few more words on that?

-L242: So radiolarian responded 0.5 earlier than diatoms? Of course taking sample resolution into account.

-L231-236: So one peak corresponds to a warming and the other one to a cooling?

-L252: "This reorganization": remove "is"

-L255 and throughout this paragraph and in the manuscript too: One of the biggest issues of this manuscript for me. Although not critical: I feel like I'm guided to see what the authors want me to see because the trends in their records is not so obvious. I can see want they say for the bioBa, but not always for the diatom records that are new here, and looking at figure 2, I can see two diatom records that hardly vary, and two that vary a lot, one of which stay high after the transistion, and the other one that increase, decrease, and so on, so the trend is not very clear here in my opinion. I think this should be discussed more and try to not oversell what we can actually see.

-Fig. 3 and in the text: Can we really talk about peaks here? I guess yes, but for diatoms, peaks are orders of magnitude higher than low values, while for radiolarian, it is "only" 2 or 3 times higher.

-Fig. 3: Keep in mind that resolution is relatively low, so a bit difficult to draw conclusions like "peaks", it could also be isolated values. Any margin of error for this?

-Maybe find a way to fuse Fig 2 and 3 for a direct comparison? Also maybe draw a line or something at the E/O transition?

-Fig.4: Again, lots of data in this manuscript seems to rely on unpublished data, yet to be accepted and not available.

-L284: add color after ODP689.

-Fig.4: Maybe put the side captions on all panels?

-Discussion: Again I question the certainty with which conclusions can be drawn with this "Low" resolution. maybe try to be a bit more moderate?

-L299-303: I agree with this, but I am not sure all of this is truly visible in your data without being guided to see it.

-L310: "environmental factors": such as?

-L323: "circum-Atlantic": Atlantic? Antarctic?

-L337-340: More productivity is usually less dissolved O2 that is consumed, so why is it the opposite here?

-L359: missing "is" between abundance and particularly?

-L372: "a striking pattern" again, with this resolution, I would be more moderate. Plus, avoid this repetition at the end of paragraph.

-L405: remove "simply"

-L449-450: I don't understand this sentence, what are primary sediment names?.

-L456-457: avoid long sentence with phrasing like "as a natural consequence of the fact that". I can be just replaced by "as". Overall try to be more concise, the manuscript is quite long as a natural consequence of the fact that there are lot of phrasing like this.

-L470: I'm not sure what is "enhanced" here? Please rephrase.

-Last paragraph of the conclusion: It sounds a bit like: " This is what we propose for now until we find something more consistent to say" I now this is not what you mean so I recommend rephrasing it.

Review of Ozen et al.,

The current manuscript attempts to assess the role of diatoms during the E/O and Antarctic Ice Sheet establishment. In particular, the authors want to understand the overall contribution of diatoms to the total opal productivity before and after the E/O and compare it with radiolarians. They also want to show that diversity is a key component for high productivity. The objectives of the manuscript are of high interest, and the methodology is good, thus this manuscript can be a good contribution to the journal.

However, it is not at a mature stage, as there are numerous concerns. I thus suggest major revisions. I have highlighted all my main concerns and other specific concerns below.

Main concerns:

• The introduction is not focused on the problem the authors want to address in the paper. The introduction is too long, and while the numerous subsections are good, not everything is needed to discuss the authors' main focus: the quantification of diatom and radiolarian contribution to biogenic silica productivity during the E/O. I would recommend making it shorter and presenting only the essential information needed to discuss what the authors want to clarify, rather than creating a kind of review of the E/O boundary in Antarctica.
• Material: Please provide a lithological column showing different lithologies and hiatuses well illustrated on the column. Then, please also briefly describe the key lithologies and major lithological changes, if there are any.
• There is no description or explanation about the depth-age model used at each site. Because the authors discuss temporal disparities of diatom productivity, I request that the authors add a paragraph clearly explaining the age model used, along with its potential error margins.
• Overall, I think the approach is good, but the authors do not consider silicoflagellates at all in the manuscript. For instance, in the Pacific, like the Japan Sea, they were also important primary producers during the Middle Miocene, similar to diatoms. So why not consider them as well? If they are not significant in the Southern Ocean, please state this; if they are significant but you did not assess them, please revise the manuscript by focusing specifically on the radiolarian and diatom contributions rather than claiming to cover all siliceous plankton.
• 3.0 Results: Please be more quantitative and not only describe the timing but also indicate average values, minimal and maximal values with the standard deviation, and the total number of samples (N).
• I am not sure that sections 3.1 and 3.2, which mix results with discussion, are needed here. I suggest going straight to the discussion to avoid redundancy.
• The discussion is overall very redundant, with similar topics being discussed in different paragraphs, making the end of the manuscript difficult to read. I don't think there are that many things that can be discussed with the new data proposed in this study. I suggest making only 2 parts: 4.1 "Diatom and Radiolarian Productivity vs. opal records" and 4.2 "Diatom diversity and Productivity: A cause/effect relationship." Then delete all the small subchapters that are very similar to each other.

Specific comments:

L.39: please add a dot after "Lear et al. 2008."

L.41-45: I think there is a third way, and it is the tectonics of Antarctica and the evolution of the Ross Sea rifting as well, which modulate Antarctic topography and thus affect the volume of ice-sheet able to be carried and therefore may affect climate changes. see:

• Wilson, D. S., & Luyendyk, B. P. (2009). West Antarctic paleotopography estimated at the Eocene-Oligocene climate transition. Geophysical Research Letters, 36, 4 PP. https://doi.org/200910.1029/2009GL039297

• Wilson, D. S., Pollard, D., DeConto, R. M., Jamieson, S. S. R., & Luyendyk, B. P. (2013). Initiation of the West Antarctic Ice Sheet and estimates of total Antarctic ice volume in the earliest Oligocene. Geophysical Research Letters, 40(16), 4305–4309. https://doi.org/10.1002/grl.50797

L.59-60: Please explain why you think E/O pCO2 reconstructions are not well constrained and tell us what is the key factor hampering these reconstructions.

L.58-64: I understand the authors' meaning, but I would like them to explain and relate climatic dynamics and Southern Ocean oceanography in more detail. Which kinds of climatic events would enhance shallow water mixing? What type of topography would enhance current circulation? Why would pro-ACC strength enhance primary productivity? I think we have many questions in mind reading the current text, so I would suggest revising this part in more detail.

L.67-72: I agree, but it is also very simplistic to attribute such a change in diatom productivity uniquely as well. This is why I would recommend carefully considering the tectonics and rifting of Antarctica with the reference I suggested, for highlighting the influence of geography on current paths and topography, which influence the potential volume of ice sheet that can be held by Antarctica and thus potentially impact pCO2. Rather than trying to explain that the E/O event is 100% caused by silica plankton blooms, I believe it is much fairer to propose clarifying the real contribution of silica plankton blooms to the E/O events among other events, including tectonic ones. 

L.85-90: I think this paper's objectives should be related to what is highlighted here and not try to overestimate the role of diatoms during the E/O. Diatoms bloomed because a substantial change occurred or a threshold was exceeded in other factors controlling Antarctica, but it was not diatom rise that caused the big changes. I think they are one of the consequences of other factors much more related to paleogeography and topography.

L.135-139: This sentence needs to be revised as its meaning is hard to understand.

1. 165-175: Sample preparation and accumulation rates: For radiolarians, and considering their weight (particularly those from the Eocene), is a sample weight of less than 1g suitable for conducting quantitative analysis? A too small weight could increase the margin of error in measurements, and it might not provide enough material for robust statistical analysis. Then, counting only a small area of the slide to get absolute abundances is theoretically good, but it relies on the assumption that your slide is perfectly homogeneous. How homogeneous are your slides? I have tried many ways to get absolute abundances, but I found that Itaki et al.'s (2018 IODP 346 Data Report) method, which involves mounting specific Q-slides using micropipette and scanning the whole Q-slide, to be the most accurate at that time. 

1. 252-257: to not fit with Ba proxy data also happened in the Japan Sea for the late Miocene, where Matsuzaki et al. (2022, Scientific Reports) data did not really fit with Zhai et al.'s 2021 Paleoceanography Ba data.

L.289: What do you mean by "mode of opal productivity pattern"? Please explain.

L.296-297: Please explain what happens to radiolarians in more detail.

L.305-312: Can this lag be associated with age model uncertainty? Because you did not provide information about that, it is hard to judge and thus I ask you to add this explanation. Actually, I assume there are low sedimentation rates, and usually the Eocene has a ±0.5 Ma error for each biostratigraphic datum used which can easily give a 1 Ma offset…

L.346: I don't see any part of the discussion addressing radiolarian contribution except here as evidence. As you discuss radiolarians later, I would suggest focusing only on diatoms here.