Dear Hemmo and colleagues,

You now have received two formal reviews and responded to both.

Schubert gives a lengthy, important and very constructive review. Such reviews are rare, and I strongly encourage you to fully incorporate his commentary -- this will significantly improve the end paper. Referee #2 also gives valuable commentary which will make a better paper.

Schubert raises a great point concerning background conditions. Almost certainly, background conditions (temperature, pCO2, etc.) changed through the early Eocene. This can be inferred from long-term stable isotope records as well as reconstructions of the CCD. Indeed, this very point has been emphasized in several manuscripts (e.g., Leon-Rodriguez and Dickens, P3, 2010; Komar et al., Paleoceanography, 2013; Slotnick et al., CoP, 2015), and is noted in the current manuscript (Page 1861). The assumption of constant background pCO2 across the
Range of hyperthermal events, while simple, is not a good one (Page 1863).

I will be the first to state that the current d13C and CCD records have several problems in terms of construction and interpretation. But a really good manuscript arises from and addresses this realization; the modeling and interpretations should come with at least two end-member cases: one where background pCO2 is high but constant; one where background pCO2 increases over the early Eocene. I will not prescribe the “the poison” on where to set pCO2 for such scenarios, but think reasonable estimates can be made.

In any case, the following two papers almost have to be referenced formally and incorporated:

B.A. Schubert and A.H. Jahren. 2013. Reconciliation of marine and terrestrial carbon isotope excursions based on changing atmospheric CO2 levels. Nature Communications, 4: 1653, doi: 10.1038/ncomms2659.

B.A. Schubert and A.H. Jahren. 2012. The effect of atmospheric CO2 concentration on carbon isotope fractionation in C3 land plants. Geochimica et Cosmochimica Acta, 96: 29-43.

(I think the miscommunication on the date of the Schubert papers results from the fact that there are two different and very relevant papers).

In summary, you have something like 75% of a very good and important paper, but it is saddled with a problematic 25%. So, fix things. I will accept this manuscript for publication, but only after extensive revision that clarifies the modeling and interpretations. I suspect very much, with some simple and entirely reasonable changes in perspective, the CIE across the PETM can be shown to be the big brother and very much related to the other CIEs. This view may be incorrect, but our community does not need any more papers arguing against Occam’s razor on faulty logic. As always, though, I am totally okay with people showing me wrong.

Sincerely,

Gerald (Jerry) Dickens

Additional Editorial Commentary

-- Page 1860 --

Line 5: This is awkward referencing in regards to the text. This is because Kennett and Stott (1991) described the CIE at the PETM but without realizing this related to an injection of carbon; Cramer et al. (2003) documented multiple CIEs, but argued they were unrelated to each other and massive injections of carbon; Zachos et al. (2005) got things right (at least in my mind) but the work only pertains to the PETM; Lourens et al. (2005) also got things correct (at least in my mind), but only documented H.

Line 13: This is awkward because ETM-2 = H-1. It is also awkward because, as noted above, the Cramer et al. (2003) paper argues against the notion of related hyperthermals.

Line 19: Should be Dickens et al. (1995). Again, though, this is awkward. Dickens et al. (1995) discussed mass balance calculations for how to explain the PETM CIE. However, full realization that the PETM and several other hyperthermals were related to carbon injection from a 13C-depleted source(s) has only come from the coupling of d13C records AND seafloor carbonate accumulation records at multiple sites over time.

-- Page 1861 --
Line 1: Should be “… factors that modify C-isotopes …”

Line 13: And also higher pCO2 (e.g., Schubert and Jahren, 2013).

-- Page 1869 --
This needs more serious thought or elaboration. There are several d13C records across the hyperthermals. Things should not hinge on comparisons to Walvis Ridge alone.

Figure 1: Not sure what the base-map colors or gray lines refer to – formation? elevation? Grid makers? Roads?

Figure 2: Should show the available polarity chron information.

Figure 3: As above. This is actually very important. You want to demonstrate that these records are in fact correlated (something for I have no doubt, but there are skeptics). Also, not sure what the two y-axes on the right side refer to.

Dear Hemmo,

I still have very mixed views on this manuscript. The rationale for the work is great, and the data is also great. However, I am not excited by the interpretations. I think embedded somewhere is a flaw in logic. One can realize this readily by looking at figure 2 in Schubert and Jahren (GCA, 2012). What should occur is a smaller amplification of terrestrial CIEs in proportion to increasing carbon input, and this effect would be enhanced with lower background pCO2. This is indeed what you document, but not what is written.

I realize that the Schubert and Jahren papers are controversial, although I am not sure why, because they actually make sense from the view of a chemist. Irrespective, the concept of reduced isotope fractionation with greater pCO2 is an interesting and most probably good one.

I am hugely overcommitted right now, so it will take me awhile to figure out where things are wrong (or of course, where you are right and my thinking is wrong – perhaps an equal probability; I planned on doing this over winter break but other items cropped up). I thus passed the manuscript again to Brian Schubert to see if he could figure out where the logic flaw lies (albeit without my background thoughts). His second review is attached, although I am not sure if he found the problem.

So, my strong recommendation is to tone down the debatable interpretation and instead adjust things to focus on the data, which is really cool and good. I would also very much take on board Brian’s second set of comments.

I will accept this manuscript for publication, because I know you and most of your co-authors, and I know you will want things to be correct and to ensure that this is the case.

Sincerely,

Gerald (Jerry) Dickens


Some additional comments:

The Abstract is really long and contains excess information unrelated to the actual work. Make this one paragraph that emphasizes the actual work. For example, the entire paragraph on Lines 27-31 can be condensed into one sentence as it is “background” and appropriate for an “Introduction” not an “Abstract”.

Line 21: The number of asterisks for the affiliation seems wrong.

Line 79-84: This long sentence is missing something and probably should be split.

Line 269: The use of “progressively” is awkward here as this is clearly not the case in the time domain.

I will accept this version.

Jerry