Dear Valeria Luciani,

I have read the manuscript several times, as well as the reviews and responses.

There are three basic problems with the current version: (A) General presentation; (B) Stable isotope interpretations; and (C) Carbonate dissolution and effects. I elaborate on these below, and both referees, as well as I, offer numerous suggestions.

Some of the issues have been amended in the revised manuscript submission. However, to fully address these three problems, the manuscript needs major revisions.

It should be stated that there is no question regarding the quality and quantity of work. The problems almost entirely lie within the scope of presentation and interpretation. As such, there is little doubt that the present work can be made into a really nice paper. This will, however, take some effort, as well as some rethinking.

Sincerely,

Gerald Dickens

***

Referee Commentary

Wade gives a short but generally favorable review. I give 3 elaborations to her review that needs.

(W1) As pointed out Wade, this is a rare data set. However, this is not fully emphasized in the manuscript.

(W2) While it is generally a good idea to run replicates for stable isotopes (indeed most measurements), I do not think analytical error is an appreciable problem in the stable isotope data set, especially for d13C.

(W3) The numbering/labeling of the CIEs needs to change (see also Comment B2, below).

(W4) The plot of stable isotopes (old Figure 2) is difficult to read. I agree, and this is for two reasons: the curves lie on top of one another, and individual data points are not obvious.

Speijer gives a fairly long and critical review. His two major criticisms concern stable isotope diagenesis and foraminifera assemblage preservation. In my opinion, both comments are valid, and have to be addressed more rigorously. Some answers to these criticisms lie in the response; however, these need to be incorporated into the final manuscript not deflected away. Further questions and answers regarding these topics come with a broader view of the site and data. I discuss these matters further, below.

Editor Commentary
(With apologies, Line numbers below refer to those in the initial submission).

Basic Comments:

(1) Place references in chronological order as per CoP final papers.

(2) Check capitalization and italics for genus and species throughout the manuscript.

Main problem A: General presentation
The manuscript would improve markedly with better presentation. Right now, it is not an easy manuscript to review. This is for multiple reasons.

(A1) There are several long paragraphs that introduce multiple concepts. These should be split. Examples are given below.

(A2) The lithological units are not clearly identified in old Figure 2 nor explained in the text. This is important, especially when it comes to fully addressing issues raised in the discussion.

(A3) The paragraph on Lines 232-240 is mostly discussion of d18O and belongs in the discussion (section 5) not in the results (section 4). However, there does need to be a paragraph that actually describes the d18O results. This does not exist right now.

(A4) The presentation of the results is awkward because some results are given in height, some are given relative to polarity chrons, some are given relative to CIEs, and some are given relative to foraminiferal biozones. Furthermore, there are times (e.g., Line 171) where a result (e.g., F value) is placed into context of another result (d13C) before the second result is presented! This leads to some very confusing reading. In general, I think all results should be given in height, although okay to also note in relationship to time, when in logical order.

(A5) The present paragraph on Lines 299-315 needs rewriting and expansion. This is because presentation on the d18O and d13C are jumbled together, and the general trends and short-term excursions are further mixed.

As a suggestion, and after a paragraph on the theory of diagenesis, have one paragraph each on:
1/ Long-term trends in d18O
2/ Short-term shifts in d18O
3/ Long-term trends in d13C
4/ Short-term shifts in d13C

The main point here is to make sure things flow logically.

(A6) From the Title through the Methods, it is never crystal clear what is being done and why. This is because the Site 1051 data comes along for the ride. (The Site 1051 data of the manuscript is not presented in the title, nor is it set-up very well in the Introduction and Methods).

(A7) Perhaps most crucially, the ordering and construction of figures needs improvement. It is difficult to fully follow presentation and discussion with the current figures, their order, and call-outs (or lack thereof).

Here is one suggestion:

Figure 1. This should be the left side of current Figure 5. This is because it is crucial to the Introduction (so should not come at the end). However, on this figure, I would absolutely add key planktonic foraminifera events (especially the Morozovellids) with “ticks/arrows” as well as various foraminiferal biozone schemes. This, of course, would all be on a common and robust time scale.

I think the broad relationship between polarity chron boundaries, stable isotope records and foraminiferal biozones (at least at low latitudes) is established across the early Paleogene. Within this framework, the definition of EECO (and the problem with this definition) is also fairly clear. EECO should be marked clearly on this figure. Broadly, it might go from about 53 to 49 Ma.

Figure 2. This should be a global map, perhaps one centered at 50 Ma (=nominally the core of EECO), showing various locations with early Eocene marine carbonate records, including Possagno and Site 1051. A point here is that our current understanding of Eocene stable isotopes and planktonic foraminifera does not center on these two locations alone; rather, the current work is chasing problems within a broader context.

Figure 3 (old Figure 1): I do not see the point of the second panel for Site 1051, as this site is well documented. Instead, other panels should be a good map of the Possagno section, and pictures of important features relevant to the discussion. Agnini et al. (2006) show a photograph of the overall outcrop, but important details are not obvious.

(A8) It is not totally clear how these results might link to changes at higher latitudes. I point this out, because it is relevant to the big picture story. For example, at least in New Zealand, there appears to be diachroneity in occurrences of Morozovellids relative to low-latitude locations (papers by Hornibrook, Hancock, etc.). This is probably worth a paragraph in the discussion.

Main problem B: Stable isotope interpretations
The stable isotope records need further discussion in light of diagenesis and correlation.

(B1) The oxygen isotopes of indurated limestone are almost always modified by diagenesis. This is because the water/rock ratio is very high for oxygen in marine sediments (i.e., much of the oxygen is in the pore water rather than in CaCO3), and because stable oxygen isotope fractionation is highly sensitive to temperature. Thus, during recrystallization (from ooze to chalk to limestone), which generally happens several hundred meters below the seafloor and along a geotherm, the d18O gets reset to lower values (e.g., Matter et al., Init Repts. DSDP 32, 1975; Hudson, J. Geol. Soc. London, 1977; Schrag et al., GCA, 1993).

However, as the ooze/chalk/limestone transitions are often sharp, at least in deep-sea sections, it means that d18O variations over short distances might in some cases reflect changes in the original carbonate, which might in turn reflect surface water temperature, surface water d18O, or both. Of course, this assumes that bulk carbonate was formed in the surface water. On this topic, one should recognize that bulk carbonate d18O in modern sediment does not clearly reflect surface water temperature.

These sorts of arguments need to be clearly discussed for the Possagno section. Basically, the d18O variations could reflect changes in surface water temperature, especially in an ice-free world. However, this discussion needs better comparison to the benthic d18O record and, ideally, other bulk carbonate d18O records. Certainly, an interesting comparison can be made with the composite benthic foraminifera d18O record (old Figure 5). For example, if the Possagno d18O record was faithfully recording surface water temperatures, then the -1.5 per mil values for EECO would be around 17 °C. So, is this reasonable? I think it would imply, in any case, that diagenesis (if via seafloor burial) must have occurred very shallow, otherwise the values would be much more negative. For reference, the bulk carbonate d18O values at the start of EECO at Mead Stream are generally -3 to -4 per mil (Slotnick et al., 2012, Supplemental Information).

(B2) Unlike for oxygen, d13C records of bulk carbonate usually give very nice, correlative records, even in indurated limestone (above references). This is because the water/rock ratio is very low for carbon in marine sediments (i.e., almost all the carbon is in the CaCO3), and because carbon isotope fractionation is relatively insensitive to temperature.

The problem, here, however, is that d13C record at Possagno does not correlate very well to other records, when one really looks in detail, at least with available presentation (Comment A). In particular, note as shown in old Figure 5, as well as in several other records (e.g., Slotnick et al., 2012; Dickens and Backman, News. Strat.), the K/X/ETM-3 event corresponds to the absolute low in d13C for the Eocene. We also know this event occurs near the top of C24n (above references and Galeotti et al., EPSL, 2010). There seemingly is no prominent d13C excursion at this horizon at Possagno.

It should be emphasized here that there is a similar issue with the d13C records at sites on Demerara Rise (Sexton et al., 2006; Kirtland-Turner et al., 2014). The prominent d13C excursion across K/X/ETM-3, found at multiple other sites (above, plus Cramer et al., 2003) is mostly missing, or greatly condensed. Instead, there are other CIEs that do not show prominently in available stable isotope records, albeit there are very few for the late early Eocene through middle Eocene.

In fact, like Possagno, the records from Demerara Rise seemingly show a high degree of variance in d13C. It would be great to show that some of these are likely coincident, but this is absent.

Circa 2105, the entire situation and labeling of early Eocene CIEs is sort of a mess. This is because it is by no means clear which CIEs in which records are correlated and thus perhaps representing changes in ocean chemistry. For example, Payros et al. (Paleoceanography, 2012) specifically looked for d13C changes near the late/early Eocene boundary as reported by Sexton et al. (2006) in a Spanish section, and found only one significant one, about half way through planktonic foraminifera biozone E7b. One might suggest this CIE is also found at Possagno (C21n-CIE1), but then there is something wrong with polarity chron labeling at one of the sites, because it occurs within the top of C21r in the Spanish section, according to Payros et al..

I am okay with tentatively labeling d13C excursions at Possagno, as long as all the above is made very clear. Of course, it would be great to try and link the CIEs at Demerara Rise and Possagno, as suggested in the text, because the background stratigraphy at both locations seems reasonable. However, in my experience, to do this properly would take considerable time.

Main problem C: Carbonate dissolution and effects
Do the foraminiferal assemblages primary represent original surface water properties or the impact of differential dissolution? There seems to be a general tendency for planktonic foraminifera specialists to acknowledge both processes, but then to try and dismiss differential dissolution, as if this would somehow diminish the impact of their research. This remains the case in the present manuscript.

I think, for bathyal sections in the early to middle Eocene, when carbonate saturation horizons were likely very dynamic, both concepts are often in play. This should be embraced. Certainly, as a chemist, differential dissolution makes the planktonic foraminifera assemblages even more interesting.

(C1) I think the manuscript needs a more open view on this problem. Clearly, dissolution affects the composition of planktonic foraminifera assemblages, and there are times of major carbonate dissolution in the deep sea during the early Eocene. One might also add Nguyen et al., Marine Micropaleo., 2009 to the references.

This is important because, as known by the authors, it appears that, at least some experimental work, there is in order of resistance, acarininds (most), morozovellids, subbotinids (least). However, I think equally important to note (as presently written) that it may be more complicated than this, and depend on species.

(C2) There is some confusion regarding long-term carbon cycling and dissolution (Lines 351-360). A slow increase in pCO2 should NOT cause ocean acidification and carbonate dissolution; this only should only occur during rapid increases in pCO2 (e.g., the PETM, our present-day and near future).

(C3) By closing ones minds on the impact of dissolution, I think the manuscript misses a truly golden opportunity regarding our understanding of the early Eocene.

The interval dominated by acarininids in the Possagno section seems to closely correspond to a “discoaster acme” (Agnini et al., 2006). However, this is not discussed in the text. The importance here is that “discoasters” are also realtively resistant to carbonate dissolution. Now, one might argue for a brief time of high warmth and greater water column stratification. One might also look at the stratigraphy of the Possagno section and realize that this interval must be very condensed (Agnini et al., 2006). Although not discussed in the text, this is consistent with dissolution, and might also explain why the K/X event and L CIEs are not clear-cut.

More interestingly, though, multiple deep-sea sites show an interval of extreme carbonate dissolution in the middle early Eocene, more specifically about equivalent to the top of T. orthostylus and the rapid rise in d13C (e.g., Slotnick et al., CoP, 2015). For reasons unclear, the lyscoline and CCD appear to have been at their minimum water depths across the entire Cenozoic during this time interval (Pälike et al., Science, 2012). I do not think coincidence that the interval characterized by acarininids (and discoasters) in the Possagno section is centered at the top of T. orthostylus (Agnini et al., 2006).

***
Specific Comments:

Line 113. Break paragraph, so new paragraph (on details of the section) begins with “The basal …”

Lines 113-119: Somewhere in here (and on old Figure 2), the lithology has to be explained better. This becomes important when discussing the results (see also Comments A and B).

Line 119: There is also very good calcareous nannofossil stratigraphy for the Possagno section (Agnini et al., 2006). Needs to be mentioned.

Line 119: There needs to be something here (probably 2 sentences) defining the interval of EECO at the Possagna section. How is it placed?

Line 120: Break paragraph, so new paragraph (which concerns Site 1051) begins with “The Blake Nose …”

Lines 135-137: This sentence is “out-of-order” (obviously, one cannot count the forams before describing the samples and extracting them from these samples, which are the next few sentences …). In fact, it seemingly should be part of a short paragraph after the paragraph on isotopes (i.e., Line 163). This is because this is how the results are presented.

Lines 148-157: These two paragraphs should be switched, given the ordering of sections in the results.

Line 156: Should define the fragamentation index. It might be noted that there problems in cases of extreme dissolution, as fragments also begin to dissolve (e.g., Leon-Rodriguez and Dickens, P3, 2010).

Lines 196-204: This information can be placed into table. Right now, it is awkward, including also because they are listed in opposite age to the flow of surrounding text.

To make all very easy for any reader, add two columns to Table S1. One would be the “CIE”, and one would be “Position”. Then shade the rows for samples that characterize the CIEs.

Lines 209-210: This is not correct and needs rewriting (see Comment B1).

Lines 216-219: This sentence is awkward. Rewrite.

Lines 221-231: Place into the table .See above comment.

Line 232: Indurated.

Lines 232-233: This is almost always the case (see Comment B1)

Line 244: Change “modifcations” to “variations”.

Lines 264-270: Not sure what are paragraphs here. In general, one does not want short 1-sentence paragraphs but one does want to start paragraphs with good short sentences that give the scope of the paragraph.

Lines 282-284: I am not sure what this means, as the Cramer et al. record at Site 1051 stops before the K/X event.

Line 289: Awkward wording.

Line 291: The radiolarians are not shown in Figure 3 (or anywhere that I can find).

Line 298: There needs to be a paragraph that clearly explains the theoretical problem of diagenesis (see Comment B).

Lines 316-319: Reword this. There remains little if any direct evidence that CO2 DROVE the EECO or the hyperthermal events. This is especially true for the latter. Clearly, however, changes in temperature appear somehow coupled to variations in CO2.

Lines 320-335: Rewrite this (see Comment B2).

Line 336: This is awkward as written. If the CIEs are NOT synchronous, they are NOT meaningful, at least from any stratigraphic or carbon cycle perspective.

Line 351: Break paragraph, so new paragraph (on details of the section) begins with “The high pCO2 …”

Lines 351-360: This (new) paragraph needs rethinking and rewriting. Furthermore, I think it needs to come after the discussion of dissolution next paragraph (see Comment C).

Line 362: Missing reference. I think should be Nguyen et al., Marine Micropaleo., 2009.

Lines 375-380: I do not fully follow this reasoning. It looks to me (old Figure 3) that several of the acarininid “pulses” align with d13C excursions and increased fragmentation.

Lines 385-389: This paragraph seems partly out-of order with previous paragraphs. I think some rewriting is needed through this section, although see also Comment C.

Line 393: Somewhere in here it has to be made clear that there is assumption that these interpretations come with the caveat that the assemblages are minimally impacted by dissolution (see comment C).

Line 401: Indent.

Line 401: This is awkward because K/X/ETM3 is not obvious in the record (see Comment B2).

Line 421: Change “did evolve” to “evolved”

Line 427: Should be “widespread”

Line 461: Avoid “crisis”

Lines 461-464: This is a long and awkward sentence. Plus, it is not clear where the text is going. Rewrite.

Line 469: Break paragraph, so new paragraph (on details of EECO) begins with “If the …”

Line 482-490: Need to consider comment C2.

Lines 543-551: See Comment B2.

Old Figure 2: I think it should be “Height” not “Thickness” (and in other figures, too).

Old Figure 4: I think it should be “Depth” not “Thickness”.

Table S1: The precision is incorrect for much of the stable isotope data.

Table S2: The assemblage data should be tabulated.

Dear Dr. Luciani,

Two reviewers have commented on your revised manuscript. Both indicate that your work will be of significant value to the community. One of the reviewers still has significant criticism, which can partly be dealt with by changing or adding text. Other points of criticism needs the generation of extra data, which I do not think desirable for the present paper, but rather for future work.
Nevertheless, I advise to incorporate the points of criticism raised by both reviewers in the revised manuscript so that the reader receives a complete overview of the potential uncertainty. I would consider this to represent moderate revision.

I'm looking forward to receiving a revised version of the manuscript.

Sincerely,

Appy Sluijs
Editor

Dear Dr. Luciani,

Based on the views of the reviewers and myself, I am happy to offer publication your paper ‘Major perturbations in the global carbon cycle and photosymbiont-bearing planktic foraminifera during the early Eocene’, pending minor revisions. Your manuscript has much improved and you adequately deal with the reviewers’ comments. Some of the issues they raise will remain but in my opinion, these issues are laid out appropriately in the manuscript and will be for the community to judge.

I performed a final read of the manuscript and spotted a small number of issues that I ask you to correct. The most notable issues regard the unclear taxonomical concepts (also touched upon by reviewer Speijer) and errors in paleolatitude estimates.

I am looking forward to receiving your revised manuscript.

Sincerely,

Appy Sluijs
Editor


*Abstract
Please make sure terminology is consistent throughout the manuscript. In line 20, you explicitly term the genera Morozovella and Acarinina, while in lines 32 and 35 you use acarininids and morozovellids. This suggests to the reader that you are talking about two different things, which I believe is not the case.
It has been a while since I checked the International Code of Zoological Nomenclature, which I believe underlies the foraminiferal classification, but I presume the –ids end must mean that you refer to specimens that are ‘like’ Morozovella and Acarinina; a statement of uncertainty. Therefore, with reviewer Speijer, I would prefer if you discuss the genera sensu stricto throughout the manuscript if you are certain that all specimens can be classified as such.

50-51. Probably fair to cite the first paper presenting a high latitude EECO temperature record here (Bijl et al. 2009)
59. ELMO should be spelled Elmo as it is not an acronym but a given name
69. Cite paper that coined the term PCIM for consistency
113-114: explicit reference to genera: use Morozovella and Acarinina; the terms morozovellids and acarininids are not the names of these genera.
121-122 and 421-422. To some of our colleagues, the terms ‘well-preserved’ and ‘recrystallized’ cannot be used for the same foraminifera because recrystallized means that the biogenic calcite has been replaced by diagenetic calcite. I presume ‘well-preserved’ here means that the preservation is sufficient for taxonomic analysis. Please rephrase.
192-193. Probably fair to cite the first paper that did this (Bijl et al. 2009). Note also that the termination of EECO is potentially not easy to recognize globally because it might well be a phenomenon specific for the southern ocean and deep ocean, related to ocean circulation and gateways (Pearson et al., 2007; Bijl et al. 2013).
226. discoaster; I presume this should be capitalized and in italics
233. I quickly checked the website paleolatitude.org (Van Hinsbergen et al. 2015 PLOS One), which gives a paleolatitude of 30 °N at 50 Ma ago for Site 577 using the default reference frame. This work is most consistent with the current insights in plate tectonic evolution. I presume the 15°N estimate comes from the ODSN database, as does Figure 2. Unfortunately, ODSN is heavily outdated. Please revise accordingly for all treated sites (although Possagno might be complicated; see Van Hinsbergen et al. 2015). Consequently, I would also recommend changing the map in figure 2.
405. ‘is-exists’ please fix this; probably a track-changes thing
440. Fix ‘track-changes’ mistakes
456. This now suggests that morozovellids is something else than the genus Morozovella as it also might include Morozovelloides. Now I am confused. Please specify early on in the manuscript what you refer to as you discuss Morozovella and morozovellids. At present it is very confusing, even for a trained micropaleontologist like myself.
486. Fix ‘track-changes’ mistakes
504. ‘need thought’. Awkward phrasing so please revise.
537. Fix ‘track-changes’ mistakes
611. Fix ‘track-changes’ mistakes
662. What is ‘tepid’?
831. Bohaty and Zachos (2003) misdate the event at 41.5; corrected in their 2009 paper.
845. Delete: ‘during’