Tracing North Atlantic volcanism and seaway connectivity across the Paleocene&ndash;Eocene Thermal Maximum (PETM)

Jones, Morgan T.; Stokke, Ella W.; Rooney, Alan D.; Frieling, Joost; Pogge von Strandmann, Philip A. E.; Wilson, David J.; Svensen, Henrik H.; Planke, Sverre; Adatte, Thierry; Thibault, Nicolas R.; Vickers, Madeleine L.; Mather, Tamsin A.; Tegner, Christian; Zuchuat, Valentin; Schultz, Bo P.

doi:https://doi.org/10.5194/egusphere-2023-36

Preprints

https://doi.org/10.5194/egusphere-2023-36

Preprints

16 Jan 2023

| 16 Jan 2023

Tracing North Atlantic volcanism and seaway connectivity across the Paleocene–Eocene Thermal Maximum (PETM)

Morgan T. Jones, Ella W. Stokke, Alan D. Rooney, Joost Frieling, Philip A. E. Pogge von Strandmann, David J. Wilson, Henrik H. Svensen, Sverre Planke, Thierry Adatte, Nicolas R. Thibault, Madeleine L. Vickers, Tamsin A. Mather, Christian Tegner, Valentin Zuchuat, and Bo P. Schultz

Abstract. There is a temporal correlation between the peak activity of the North Atlantic Igneous Province (NAIP) and the Paleocene–Eocene Thermal Maximum (PETM), suggesting that the NAIP may have initiated and/or prolonged this extreme warming event. However, corroborating a causal relationship is hampered by a scarcity of expanded sedimentary records that contain both climatic and volcanic proxies. One locality hosting such a record is Fur Island in Denmark, where an expanded pre- to post-PETM succession containing hundreds of NAIP ash layers is exceptionally well preserved. We compiled a range of environmental proxies, including mercury (Hg) anomalies, paleotemperature proxies, and lithium (Li) and osmium (Os) isotopes, to trace NAIP activity, hydrological changes, weathering, and seawater connectivity across this interval. Volcanic proxies suggest that NAIP activity was elevated before the PETM and appears to have peaked during the body of the δ¹³C excursion, but decreased considerably during the PETM recovery. This suggests that the acme in NAIP activity, dominated by flood basalt volcanism and thermogenic degassing from contact metamorphism, was likely confined to just ~200 kyr (ca. 56.0–55.8 Ma). The hundreds of thick basaltic ashes in the post-PETM strata likely represent a change from effusive to explosive activity, rather than an increase in NAIP activity. Detrital δ⁷Li values and clay abundances suggest that volcanic ash production increased basaltic reactive surface area, likely enhancing silicate weathering and atmospheric carbon sequestration in the early Eocene. Signals in lipid biomarkers and Os isotopes, traditionally used to trace paleotemperature and weathering changes, are used here to track seaway connectivity. These proxies indicate that the North Sea was rapidly cut off from the North Atlantic in under 12 kyr during the PETM recovery due to NAIP thermal uplift. Our findings reinforce the hypothesis that the emplacement of the NAIP had a profound and complex impact on Paleocene–Eocene climate, both directly through volcanic and thermogenic degassing, and indirectly by driving regional uplift and changing seaway connectivity.

Received: 10 Jan 2023 – Discussion started: 16 Jan 2023

Download & links

Preprint (PDF, 3289 KB)

Supplement (152 KB)

Morgan T. Jones et al.

Status: final response (author comments only)

CC1:
'Comment on egusphere-2023-36', Nina M. Papadomanolaki, 22 Feb 2023

The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2023/egusphere-2023-36/egusphere-2023-36-CC1-supplement.pdf

Citation: https://doi.org/10.5194/egusphere-2023-36-CC1
- AC2: 'Reply on CC1', Morgan Jones, 10 May 2023
  
  This is a duplicate of reviewer comment 1 (RC1), which we have addressed there.
  
  Citation: https://doi.org/10.5194/egusphere-2023-36-AC2
RC1:
'Comment on egusphere-2023-36', Nina Papadomanolaki, 24 Feb 2023
Review for Tracing North Atlantic volcanism and seaway connectivity across the Paleocene-Eocene Thermal Maximum (PETM) by Jones et al.
General comment:
I really enjoyed reading this manuscript, it’s thorough and very well-written. I believe it provides key insights into the timing and nature of NAIP activity during the PETM which may allow us to move forward in our understanding of the forcing behind this event. The proxies used complement each other well and may, hopefully, also serve as a basis for future studies into other LIP-related events. I appreciate the clear outlining of limitations and uncertainties, which make their final conclusions more robust. At this stage, my main ‘issue’ is that it can be a difficult paper to understand for non-specialists, due to the nature of the used proxies. I do believe it’s an important paper for a wide audience and I think some small adjustments could already make it more accessible.
Specific comments:
Lithium: I am not an expert on the Li system, but I know that basalt weathering also contributes to the Li pool. Would it be possible to use Li in any way as an additional indicator for volcanism (together with Os). Does your interpretation of δ⁷Li in the paper take this into account or would it change if part of the signal is affected by basalt availability/emplacement?

Seaway connectivity: Does the timing/progression of North Sea restriction such as you find it show any links to changes in North Sea oxygenation (from Schoon et al., 2015)? Overall, it seems to me oxygen concentrations recover (or atleast deox. becomes less severe) in the Fur Formation. Could you add a few words on this, perhaps in section 5.2 where you list previously published evidence for the restriction?

PETM C sources/sinks: I personally would like to read more specifically if these conclusions can also inform us a bit about when different types of emissions (CO₂/CH₄) may have occurred and how the balance between source/sink shifted in this time interval (e.g. evidence for volcanism also during recovery and after PETM, coincident with climate recovery). I understand this might be a bit beyond the scope of the paper, but as the introduction goes into quite a bit of detail regarding sources/sinks, a few sentences on this would definitely be of interest.

Line 752: Would your conclusions be affected by the longer estimate (~170kyr) of Zeebe and Lourens (2019)? Is there a specific reason for choosing the shorter duration?
Technical corrections:
It would be helpful if the order in which proxies are described/shown is kept the same throughout the paper (e.g. first Hg, then osmium, etc.).

A short description of the clays you discuss, and their interpretation, is missing from section 3.4 Clay mineralogy

Lines 65-68: reference?

Lines 83-85: This sentence gives the impression that constraining volumes/fluxes is the goal of the paper, so it may need to be reformulated to avoid this.

Figure 1: the color for subaerial volcanism looks more pink in the figure, while in the legend it appears more orange. Check if it’s indeed the same.

Figure 1: A bigger contrast in colors between subaerial volcanism and volcanic centers would also be welcome.

Figure 3: What does the red space in the d13C panel indicate? It’s also not listed in the caption.

Figure 3: As I mentioned above, the number and complexity of the proxies that have been used (esp. Hg, Li and Os) make this paper somewhat difficult to read for non-experts. I was wondering if mechanism indicators could be added to this figure? Something like an arrow with more volcanism/more weathering on either side. Otherwise, a table with the proxies and their interpretations could also work and give space for a bit more nuance than such an arrow might. If there’s space within the paper, I think it would help a lot.

Lines 702-703: What does ‘waning’ mean in this case? Less overall explosive volcanism (and shift to more effusive) or a shift from more explosive to less explosive volcanism (e.g. because of less hydromagmatic interactions?)? Could this be clarified?
Citation: https://doi.org/10.5194/egusphere-2023-36-RC1
- AC1: 'Reply on RC1', Morgan Jones, 10 May 2023
  
  I really enjoyed reading this manuscript, it’s thorough and very well-written. I believe it
  
  provides key insights into the timing and nature of NAIP activity during the PETM which may
  
  allow us to move forward in our understanding of the forcing behind this event. The proxies
  
  used complement each other well and may, hopefully, also serve as a basis for future studies
  
  into other LIP-related events. I appreciate the clear outlining of limitations and uncertainties,
  
  which make their final conclusions more robust. At this stage, my main ‘issue’ is that it can be
  
  a difficult paper to understand for non-specialists, due to the nature of the used proxies. I do
  
  believe it’s an important paper for a wide audience and I think some small adjustments could
  
  already make it more accessible.
  We thank the reviewer for their positive feedback and thorough review. We have addressed all of their points below (in italic).
  Specific comments:
  
  Lithium: I am not an expert on the Li system, but I know that basalt weathering also
  
  contributes to the Li pool. Would it be possible to use Li in any way as an additional indicator
  
  for volcanism (together with Os). Does your interpretation of δ7Li in the paper take this into
  
  account or would it change if part of the signal is affected by basalt availability/emplacement?
  As it is a stable isotope system, the lithium isotopic signature in the sediments is primarily affected by clay formation and dissolution. Therefore, the emplacement of easily-weathered fresh volcanic material is likely to drive siliciclastic δ⁷Li values to lower values, which indeed seems to be the case with sections of this dataset.
  
  The large negative δ⁷Li excursion at the PETM onset is likely to be a response to extreme warming, which would have increased chemical weathering through an enhanced hydrological cycle (see Pogge von Strandmann et al., 2021). However, the slight radiogenic shift in Os isotopes during the earliest phases of the PETM suggests that increased weathering of volcanic material was not the cause of the initial δ⁷Li excursion.
  
  In contrast, the post-PETM δ⁷Li variations appear to covary with Os isotopes, with low δ⁷Li values coinciding with unradiogenic Os isotopes, which supports the weathering of fresh volcanic material as a primary driver of the observed changes in these proxies. Both of these isotope systems are likely to be sensitive to changes between effusive and explosive activity (even if the latter is volumetrically less significant) because the formation of ash and scoria leads to an increase in the particle surface area by several orders of magnitude compared to lavas. Therefore, the large but gradual shift in both Li and Os isotopes towards the positive ash series may reflect a shift from effusive- to explosive-dominated volcanism.
  Seaway connectivity: Does the timing/progression of North Sea restriction such as you find it
  
  show any links to changes in North Sea oxygenation (from Schoon et al., 2015)? Overall, it
  
  seems to me oxygen concentrations recover (or at least deox. becomes less severe) in the Fur
  
  Formation. Could you add a few words on this, perhaps in section 5.2 where you list previously
  
  published evidence for the restriction?
  It is certainly plausible that the restriction of the North Sea had an effect on marine anoxia. However, the Schoon et al (2015) redox record does not continue into the post-PETM strata. Stokke et al. (2021) show that S, Mo, and U contents all decrease during the PETM recovery, suggesting a reduction in anoxic conditions. However, it is difficult to separate the regional effects of the North Sea isolation from the global oceans, and the recovery from marine anoxic conditions induced by the PETM.
  PETM C sources/sinks: I personally would like to read more specifically if these conclusions
  
  can also inform us a bit about when different types of emissions (CO2/CH4) may have occurred
  
  and how the balance between source/sink shifted in this time interval (e.g. evidence for
  
  volcanism also during recovery and after PETM, coincident with climate recovery). I
  
  understand this might be a bit beyond the scope of the paper, but as the introduction goes
  
  into quite a bit of detail regarding sources/sinks, a few sentences on this would definitely be
  
  of interest.
  There are a few papers in preparation that will deal with this issue more directly, based on material from the recent IODP Expedition 396 on the Norwegian continental margin. In short, it is not straightforward to draw any clear conclusions regarding changes in C balance based on the currently available data, so we would prefer not to go into too much detail here. A couple of sentences can be added to the conclusions to briefly touch on this subject if required.
  Line 752: Would your conclusions be affected by the longer estimate (~170kyr) of Zeebe and
  
  Lourens (2019)? Is there a specific reason for choosing the shorter duration?
  Although a longer PETM body duration will decrease our estimates of mass accumulation rates by ~40%, the sedimentation rates compared to pre-PETM and post-PETM strata are still an order of magnitude higher, which implies that our conclusions would not be affected.
  
  We chose a PETM onset age of 55.93 Ma (Westerhold et al., 2017) as this cyclostratigraphic age appears to be the better fit with existing geochronological data. In particular, the precise U-Pb age of 55.785 ± 0.034 Ma from a bentonite within the PETM carbon isotope excursion (CIE) from the Longyearbyen section in Svalbard (Charles et al., 2011) appears to be incompatible with an onset age of 56.01 ± 0.05 Ma and a PETM body duration of 170 ± 30 kyr (Zeebe and Lourens, 2019). It is a little difficult to discern where in the CIE the bentonite is in the Svalbard strata, as there is no obvious inflection between the body and recovery phases. However, the ash layer is still within the zone of elevated concentrations of Apectodinium Augustum (Charles et al., 2011), which suggests that it is indeed part of the CIE body or early in the recovery. Therefore, the 145 kyr difference between this bentonite age and the cyclostratigraphic age of Westerhold et al. (2017) is a better fit than the 225 kyr difference between the Zeebe & Lourens (2019) onset age and this syn-PETM ash layer.
  Technical corrections:
  
  1. It would be helpful if the order in which proxies are described/shown is kept the same
  
  throughout the paper (e.g. first Hg, then osmium, etc.).
  We carefully revisited the manuscript structure to see if there were ways to improve clarity for the reader. The proxies are described in the same order in the Methods, Results, and Figures, while in the Discussion these proxies are applied to specific time intervals and events (such as seaway connectivity).
  2. A short description of the clays you discuss, and their interpretation, is missing from
  
  section 3.4 Clay mineralogy
  We did not go into detail describing the clays because this was presented in Stokke et al. (2021), but we can add a brief synthesis of those findings if required.
  3. Lines 65-68: reference?
  We have added a reference to Storey et al. (2007a).
  4. Lines 83-85: This sentence gives the impression that constraining volumes/fluxes is the
  
  goal of the paper, so it may need to be reformulated to avoid this.
  We have reformulated this sentence to:
  
  “Constraining the timing and style of NAIP activity is critical to understanding the volumes and fluxes of each potential carbon source, in order to determine their roles in the initiation and long duration of the PETM.”
  5. Figure 1: the colour for subaerial volcanism looks more pink in the figure, while in the
  
  legend it appears more orange. Check if it’s indeed the same.
  There is indeed a mismatch between the key and the figure. We have edited the figure to standardise the colours.
  6. Figure 1: A bigger contrast in colours between subaerial volcanism and volcanic centres
  
  would also be welcome.
  A good suggestion, this was combined with the edit above to increase the contrast.
  7. Figure 3: What does the red space in the d13C panel indicate? It’s also not listed in the
  
  caption.
  It shows the PETM carbon isotope excursion, as labelled by the column just to the left. We have added the following text to the Figure caption to improve clarity.
  
  “The δ¹³C_org data are from previous studies (Jones et al., 2019a; Schoon et al., 2013), with the red infill denoting the PETM CIE.”
  8. Figure 3: As I mentioned above, the number and complexity of the proxies that have been used (esp. Hg, Li and Os) make this paper somewhat difficult to read for non-experts. I was wondering if mechanism indicators could be added to this figure? Something like an arrow with more volcanism/more weathering on either side. Otherwise, a table with the proxies and their interpretations could also work and give space for a bit more nuance than such an arrow might. If there’s space within the paper, I think it would help a lot.
  Arrows have been added to the Li and Os isotope systems to improve clarity for non-experts in Figure 3, as Li isotope changes are dominated by clay formation whereas Os isotopes are dominated by mixing between unradiogenic (mantle) and radiogenic (continental) end members. On the other hand, Hg is quite a complex system without distinct end-member behaviour, and isolated Hg anomalies do not always indicate an increase in Hg input (such as increased volcanism). We cover this interpretation in detail in the text, so we believe that adding arrows for Li and Os isotopes, but not Hg, is the most appropriate way forward.
  9. Lines 702-703: What does ‘waning’ mean in this case? Less overall explosive volcanism (and shift to more effusive) or a shift from more explosive to less explosive volcanism (e.g. because of less hydromagmatic interactions?)? Could this be clarified?
  We intended to convey that there is very little evidence of explosive volcanism during the PETM body at Fur, which means that either the magnitude of explosive eruptions decreased (thereby not transporting ash as far as Denmark) and/or there was a decrease in explosive activity (so that fewer ash producing events occurred). We have improved the text to avoid confusion:
  “The PETM onset at Fur is bounded by ash SK2 and the much thinner SK3 and SK4 layers that are the last ash occurrences for >19 m of strata (Figure 3), suggesting that explosive volcanism either decreased in magnitude to not transport ash as far, or that there was a period of explosive volcanic quiescence.”
  An important point to consider is that the presence/absence of ash layers does not necessarily inform on whether this could be a switch to more effusive activity. The NAIP is large enough that effusive and explosive activity could be occurring in different parts of the province at the same time. That said, the high Hg deposition rates at Fur during the PETM, coupled with the field evidence of >5 km of flood basalts erupted in East Greenland and radiometrically dated to within the PETM interval, provides strong evidence that intense effusive activity did indeed occur during the PETM.
  
  Citation: https://doi.org/10.5194/egusphere-2023-36-AC1
RC2:
'Comment on egusphere-2023-36', Marcus Gutjahr, 27 Mar 2023

In this manuscript under discussion Morgan Jones and co-authors present a comprehensive set of geochemical and isotopic data in order to reconstruct regional and global aspects of NAIP activity, hydrological changes, weathering, and seawater connectivity across the PETM in an outstanding extended sedimentary succession on Fur Island in Denmark. I particularly like the combined multi-proxy study of proxies for volcanic activity (e.g. Hg/TOC, Hg anomalies, Os isotopes), temperature proxies, and chemical weathering indicators (Li and Os isotopes). The authors discuss all these geochemical sedimentary parameters on an outstanding sedimentary section in relative proximity to the North Atlantic Igneous Province. It is arguably a long manuscript, maybe here and there the discussion could be a little shorter, but overall this work is well prepared, very well written and appears quite polished. I have no major comments, but several moderate and minor, which I raise as presented in the manuscript.
Lines 78-81:
Here the authors state: “Yet, this high-volume carbon release scenario might be at odds with the extremely enhanced organic carbon burial rates for the PETM, a carbon sink would rapidly drive exogenic δ13C to positive values if not balanced by a heavily 12C-enriched source…”
I find this statement a little puzzling. In our mentioned study (Gutjahr et a., 2017, Nature) we clearly showed that – despite our modelled very high carbon emission rates over the CIE – enhanced organic carbon burial following the peak CIE is required for our intermediate complexity model to track the marine d13C evolution as recorded in post-CIE planktic foraminifera. In other words, without enhanced organic carbon burial, our geochemical data could not be brought into agreement with the cGENIE model output. Therefore our most realistic carbon release budget was on the order of 12,200 Gt C. I agree, however, that the required total budget of additional organic carbon burial still is, and will be, a matter of debate for some time.
90-92:
Gernon et al. (2022, Nature Geoscience) recently alternatively suggested release of mantle-derived carbon from the subcontinental lithospheric mantle with much higher CO2 concentrations of 8% or more.
Figure 2B:
The d18O and d13C data shown here are not from this section, right? Could the authors please make this clearer in the figure? I initially thought these would be local stable isotope records.
260:
Do the authors have any idea towards the origin of this glauconite-rich silty horizon at the base of the CIE? Presuming these are authigenic in origin, what conditions would have been needed to allow formation of this glauconite layer?
Section 3.3.:
What about total procedural blank levels for Li, and did the authors report any secondary Li isotope standard results?
461:
Is fcren’ hence a qualitative proxy for warm and saline waters?
Figures 3, 5, 6, 7 and throughout the text:
Please do not use Os(i) as axis title for a 187Os/188Os isotopic composition. Better use 187Os/188Os(i) (all with super-/sub-scripts respectively). Same goes for its usage throughout the manuscript. Just using Os(i) creates the wrong impression that we are dealing with an elemental proxy.
Section 4.2:
I find the evolution in Hg content (both elemental and relative to TOC) striking in that no peak is seen at the interval with most abundant ash layers. This could suggest that the type of volcanism (sub-marine vs sub-aerial) may have quite some impact on Hg abundance. And the general pattern of Hg abundance makes it appear like a very general proxy for volcanic activity, but I may be wrong. The authors discuss the Hg evolution in section 5.1., which is appreciated. I’d also be interested to know whether the almost anti-correlated ash layer abundance vs. Hg concentration peaks simply track these different styles of volcanism that were encountered during the emplacement of the NAIP. Could the authors expand a bit more on this feature? Is there a good understanding in the PETM NAIP literature as to the importance of sub-marine as opposed to sub-aerial volcanism, or transitions from one phase to another? This certainly ought to have an impact on geochemical records such as presented here. This is already slightly touched upon in section 5.1 but could be expanded.
621:
“Phreatomagmatic” is probably a more accurate term than “hydro-magmatic”.
674-677:
Do unradiogenic Os isotope ratios really only reflect elevated basalt weathering, or could this also be partially controlled by direct Os release to seawater during sub-marine volcanism or hydrothermal activity? The PETM core section with these unradiogenic Os isotope compositions looks remarkably constant compared with the sections above and below.
727-731:
With regard to the compositional change in Os isotopes during the PETM section, the authors mention the potential importance of thermogenic release of Os during contact metamorphism. Above in the same paragraph the authors correctly discuss the importance of weathering feedbacks controlling the Os isotopic shift. This part here with the thermogenic contribution seems a bit desperate. It may play a role, yet is not really required, is it?
Section 5.2.1
Good discussion. But before comparing seawater Os isotope data from different ocean basins with regard to their comparability I consider it mandatory to introduce to what extent these literature sourced compositions indeed reflect an original past seawater signature that is not altered by detrital contributions in the sediment. Recovering a past seawater Os isotope signal from marine sediments may not be trivial at all, therefore such a preliminary assessment must be done. If sediments were completely digested to extract its Os isotopic signal these always contain a detrital component.
Section 5.3
The Li isotopic results were probably a little sobering for the authors. When I first saw the detrital d7Li record in Pogge von Strandmann et al. (2021, Science Advances) I was quite excited to see such apparent dramatic climatic Li isotopic signatures being preserved in shallow marine sediments over the PETM CIE. Now being able to see the bouncing of Li isotopic compositions over the entire section makes the initial changes look less significant and (playing Devil’s advocate) in the worst case coincidental. The discussion here is fine, but I have the impression we still need to understand the underlying controls in generating lithogenic Li isotopes better.
---
Marcus Gutjahr

Citation: https://doi.org/10.5194/egusphere-2023-36-RC2
- AC3: 'Reply on RC2', Morgan Jones, 10 May 2023
  
  In this manuscript under discussion Morgan Jones and co-authors present a comprehensive set of geochemical and isotopic data in order to reconstruct regional and global aspects of NAIP activity, hydrological changes, weathering, and seawater connectivity across the PETM in an outstanding extended sedimentary succession on Fur Island in Denmark. I particularly like the combined multi-proxy study of proxies for volcanic activity (e.g. Hg/TOC, Hg anomalies, Os isotopes), temperature proxies, and chemical weathering indicators (Li and Os isotopes). The authors discuss all these geochemical sedimentary parameters on an outstanding sedimentary section in relative proximity to the North Atlantic Igneous Province. It is arguably a long manuscript, maybe here and there the discussion could be a little shorter, but overall this work is well prepared, very well written and appears quite polished. I have no major comments, but several moderate and minor, which I raise as presented in the manuscript.
  
  We thank the reviewer for their thorough and fair assessment of the manuscript. We have addressed all of their points below (in italic).
  Lines 78-81:
  
  Here the authors state: “Yet, this high-volume carbon release scenario might be at odds with the extremely enhanced organic carbon burial rates for the PETM, a carbon sink would rapidly drive exogenic δ13C to positive values if not balanced by a heavily 12C-enriched source…”
  
  I find this statement a little puzzling. In our mentioned study (Gutjahr et a., 2017, Nature) we clearly showed that – despite our modelled very high carbon emission rates over the CIE – enhanced organic carbon burial following the peak CIE is required for our intermediate complexity model to track the marine d13C evolution as recorded in post-CIE planktic foraminifera. In other words, without enhanced organic carbon burial, our geochemical data could not be brought into agreement with the cGENIE model output. Therefore our most realistic carbon release budget was on the order of 12,200 Gt C. I agree, however, that the required total budget of additional organic carbon burial still is, and will be, a matter of debate for some time.
  We thank the reviewer for this comment, as it highlighted that our original statement was unclear. Clearly, we agree with the reviewer that a ¹³C-depleted C sink (orgC) is required to match the shape of the CIE recovery. We meant to convey that, as highlighted in a recent study from Papadomanolaki et al. (2022), there is a remaining disagreement between model and proxy data regarding the timing and magnitude of orgC burial. Most modelling studies require a reduction in organic carbon burial during the CIE body to sustain the extreme δ¹³C conditions, whereas the opposite is observed in the geological record (orgC burial appears focused and elevated during the CIE body; John et al., 2008, Kaya et al., 2022). The study of Gutjahr et al. (2017) is a seminal paper and advances our understanding of carbon cycle dynamics during the PETM, but some issues remain. For example, their modelled CIE body is shorter than estimated from sedimentary sections, and the organic carbon burial starts 30 kyr after, not coincident with, the CIE onset. Based on the reviewer’s comments, we have rephrased this section to the following:
  
  “However, most carbon cycle model scenarios appear to be at odds with the extremely enhanced organic carbon burial rates during much of the PETM (Kaya et al. 2022, John et al. 2008). A recent modelling study demonstrated that a large organic carbon sink would rapidly drive exogenic δ¹³C to positive values unless the impact of organic carbon burial was reduced during the CIE body (Papadomanolaki et al., 2022). More often, a scenario is chosen that focuses the impact of organic carbon burial to the later parts of the CIE to match the CIE recovery (Bowen & Zachos, 2010; Bowen, 2013; Gutjahr et al., 2017; Papadomanolaki et al., 2022). However, scenarios with reduced organic carbon burial during the initial phases of the CIE remain in conflict with field observations (e.g. John et al. 2008, Kaya et al. 2022) and complicate direct interpretation of the CIE purely on grounds of the source δ¹³C signature.”
  
  90-92:
  
  Gernon et al. (2022, Nature Geoscience) recently alternatively suggested release of mantle-derived carbon from the subcontinental lithospheric mantle with much higher CO2 concentrations of 8% or more.
  
  There are critical geochronological and geochemical issues with this study (briefly outlined below) that impact the conclusions and viability of the proposed concept of extreme sub-crustal carbon release coincident with the PETM. Discussing these issues is outside the scope of our present manuscript and hence we prefer not to cite this work.
  
  The geochronology and the placement of the PETM in the DSDP Site 555 locality (Rockall Trough) is complicated, but the decision not to include a δ¹³C record due to thermal alteration effects is not borne out by the site report (Backman et al., 1984) that states that the mature organic component is reworked and not heated in situ. Moreover, the cited sedimentation rates of 50 cm kyr^-1 that form the basis for the PETM duration in the Rockall section do not match the site report’s estimate of 9.5 cm kyr^-1 with a maximum of 26 cm kyr^-1 in the section. As such, the PETM interval is based on biostratigraphic observations of a single dinoflagellate cyst marker species in a single sample (site report), so the duration of this critical interval of the succession at Site 555 cannot currently be constrained with confidence.
  
  The presented ages of the East Greenland and Faroe Islands basalts have not been corrected to the most recent 28.201 Ma Fish Canyon Tuff calibration (Kuiper et al., 2008). Recalculating these ages gives 56.78 ± 0.25 Ma for the base of the Milne Land Fm in East Greenland (from 56.1 ± 0.4 Ma), which places the chemical heterogeneities observed in the basalts much earlier than the PETM interval. The ages of the Faroe Islands lavas are heavily debated, with studies arguing that the PETM interval could be above or below the hiatus. A recalibrated Ar/Ar age of 55.57 ± 0.35 Ma from the Middle basalt series in the Faroe Islands (Storey et al., 2007b) gives a post-PETM age for these lavas. This result suggests that the post-hiatus, high-Ti basalts observed in East Greenland and the Faroes may not be synchronous. This implies, with the current best estimates, that the analysed materials span at least 1 Myr, not ~200 kyr as proposed.
  
  The authors choose a pre-eruptive CO₂ concentration of 2 wt% for flood basalt eruptions in their model, citing Self et al. (2005), despite this cited paper stating “…0.5 wt% [is] a reasonable but possibly high value for pre-eruptive CO₂ concentration [in flood basalt eruptions]”. The Monte Carlo calculations assume concentrations ranging from 1–8 wt%, all in excess of this value. There is no convincing geochemical evidence from the northeast Atlantic margin that currently supports such elevated CO₂ concentrations.
  
  Figure 2B:
  
  The d18O and d13C data shown here are not from this section, right? Could the authors please make this clearer in the figure? I initially thought these would be local stable isotope records.
  
  We have added the headers "Global isotopic records" and "Regional stratigraphy" to figure 2B to remove ambiguity.
  
  260:
  
  Do the authors have any idea towards the origin of this glauconite-rich silty horizon at the base of the CIE? Presuming these are authigenic in origin, what conditions would have been needed to allow formation of this glauconite layer?
  
  The formation of the glauconite-rich horizon is mainly authigenic in origin, suggesting extended exposure at the sea floor and indicative of slow sediment accumulation rates or a hiatus in deposition. We have added the following sentence to the manuscript:
  
  “The transition from the Holmehus/Østerrende Formation to the Stolleklint Clay is marked by a possible hiatus of unknown duration and a glauconite-rich silty horizon (Heilmann-Clausen, 1995; Schmitz et al., 2004), which is comprised of mainly authigenic grains and is interpreted as evidence of very low sedimentation rates (Schoon et al., 2015). Above this glauconite-rich horizon, there is no clear evidence of any breaks in sedimentation until the top of the Fur Formation (Heilmann-Clausen et al., 1985; 2014; Stokke et al., 2020a).”
  Section 3.3.:
  
  What about total procedural blank levels for Li, and did the authors report any secondary Li isotope standard results?
  
  The Li concentration is high in shales, so blank levels are not an issue. We have added the following text to section 3.3, where secondary standards are also now listed:
  
  “The measured δ⁷Li values of secondary standards at this facility are Seawater: 31.17 ± 0.38‰ (n=43); USGS BCR-2: 2.57 ± 0.30‰ (n=11); USGS SGR-1b: 3.82 ± 0.28‰ (n=9). The total procedural blank is ≤ 0.003 ng Li (Pogge von Strandmann et al., 2019).”
  
  461:
  
  Is fcren’ hence a qualitative proxy for warm and saline waters?
  
  It could potentially be used in such a way, but with the currently available data, we cannot make such definitive statements. Further dedicated investigations of the pros and cons of this proxy are required, particularly the underlying mechanisms, but are beyond the scope of this paper. This is why we use it as a supporting argument to the Os isotopes for investigating seaway connectivity, rather than as a standalone feature.
  
  Figures 3, 5, 6, 7 and throughout the text:
  
  Please do not use Os(i) as axis title for a 187Os/188Os isotopic composition. Better use 187Os/188Os(i) (all with super-/sub-scripts respectively). Same goes for its usage throughout the manuscript. Just using Os(i) creates the wrong impression that we are dealing with an elemental proxy.
  
  We have replaced Os_i with ¹⁸⁷Os/¹⁸⁸Os_(i) in figures 3, 5, 6, and 7, and throughout the text.
  
  Section 4.2:
  
  I find the evolution in Hg content (both elemental and relative to TOC) striking in that no peak is seen at the interval with most abundant ash layers. This could suggest that the type of volcanism (sub-marine vs sub-aerial) may have quite some impact on Hg abundance. And the general pattern of Hg abundance makes it appear like a very general proxy for volcanic activity, but I may be wrong. The authors discuss the Hg evolution in section 5.1., which is appreciated. I’d also be interested to know whether the almost anti-correlated ash layer abundance vs. Hg concentration peaks simply track these different styles of volcanism that were encountered during the emplacement of the NAIP. Could the authors expand a bit more on this feature? Is there a good understanding in the PETM NAIP literature as to the importance of sub-marine as opposed to sub-aerial volcanism, or transitions from one phase to another? This certainly ought to have an impact on geochemical records such as presented here. This is already slightly touched upon in section 5.1 but could be expanded.
  
  We agree with the reviewer that the Hg proxy is sometimes challenging, and that it is striking that there is not a significant enrichment associated with the positive ash series. The increase in Hg anomalies with increasing proximity to the NAIP was covered in detail by Jones et al. (2019), but we have added to section 5.1 in this manuscript to draw the reader’s attention to how this proxy may have been affected by changing volcanism/magmatism across the PETM. The subaerial vs submarine emissions appear particularly important for the atmospheric dispersal of Hg, so that the Hg signal from submarine LIPs has a geographically limited impact (Percival et al. 2018). Furthermore, there is compelling evidence from modern systems that passive submarine degassing of Hg significantly reduces the Hg flux to the atmosphere, and may restrict Hg deposition in enclosed marine environments to within a few kms from the source in passive degassing scenarios (Tomiyasu et al., 2007). This finding suggests that submarine emissions are going to be heavily restricted in terms of their geographical distribution, and that the explosivity of the eruption or hydrothermal vent will determine whether a far-field signal is recorded. We have added the following text to section 5.1:
  
  “The regional distribution of Hg emissions would be heavily affected by whether Hg degassing is subaerial or submarine (Jones et al., 2019a; Percival et al., 2018). Passive submarine degassing around modern volcanic systems can limit Hg deposition to just 10’s km from the source in enclosed, shallow marine environments (e.g. Tomiyasu et al., 2007), so the depth and explosivity of submarine emissions will have a major impact on subsequent Hg dispersal.”
  621:
  
  “Phreatomagmatic” is probably a more accurate term than “hydro-magmatic”.
  
  We have changed it to phreatomagmatic.
  
  674-677:
  
  Do unradiogenic Os isotope ratios really only reflect elevated basalt weathering, or could this also be partially controlled by direct Os release to seawater during sub-marine volcanism or hydrothermal activity? The PETM core section with these unradiogenic Os isotope compositions looks remarkably constant compared with the sections above and below.
  
  We have changed the text to include all mantle-derived sources:
  
  “These values imply that mantle sources were already a major component of global Os fluxes, including basalt weathering from ongoing NAIP activity and the earlier tropical emplacement of the Deccan Traps at 66.5–65 Ma (Schoene et al., 2019).”
  727-731:
  
  With regard to the compositional change in Os isotopes during the PETM section, the authors mention the potential importance of thermogenic release of Os during contact metamorphism. Above in the same paragraph the authors correctly discuss the importance of weathering feedbacks controlling the Os isotopic shift. This part here with the thermogenic contribution seems a bit desperate. It may play a role, yet is not really required, is it?
  
  The reviewer is correct that it is not required to explain the observed Os isotopic shift, so we have deleted the second part of the sentence and removed the reference to Dubin (2015). The sentence now reads:
  
  “The shift to radiogenic ¹⁸⁷Os/¹⁸⁸Os values suggests that an increase in submarine volcanism is unlikely during the onset.”
  
  Section 5.2.1
  
  Good discussion. But before comparing seawater Os isotope data from different ocean basins with regard to their comparability I consider it mandatory to introduce to what extent these literature sourced compositions indeed reflect an original past seawater signature that is not altered by detrital contributions in the sediment. Recovering a past seawater Os isotope signal from marine sediments may not be trivial at all, therefore such a preliminary assessment must be done. If sediments were completely digested to extract its Os isotopic signal these always contain a detrital component.
  
  We thank the reviewer for pointing out this omission in our discussion. We have added an introductory paragraph at the beginning of Section 5.2.1 addressing this important point:
  
  “There are now several ¹⁸⁷Os/¹⁸⁸Os data sets from numerous global localities that can be used to assess the extent of ocean connectivity during the latest Paleocene and early Eocene (Figure 7). The methodology of Os retrieval has evolved through time. Older PETM studies used inverse aqua regia for sample digestion (e.g. Dickson et al., 2015; Ravizza et al., 2001; Schmitz et al., 2004; Wieczorek et al., 2013), while more recent analyses used chromic acid to preferentially liberate hydrogenous Os (Liu et al., 2019; this study). The inverse aqua regia digestion method is more aggressive, potentially leading to contamination from detrital Os in silicate minerals (e.g. Kendall et al., 2004). However, the existing ¹⁸⁷Os/¹⁸⁸Os data across the PETM are largely homogeneous across the North Atlantic, Indian, and Tethys Oceans (Dickson et al., 2015; Liu et al., 2019; Ravizza et al., 2001; Schmitz et al., 2004). This global homogeneity suggests that any detrital contamination is minimal, and that the open marine Os residence time exceeded the ocean mixing time.”
  Section 5.3
  
  The Li isotopic results were probably a little sobering for the authors. When I first saw the detrital d7Li record in Pogge von Strandmann et al. (2021, Science Advances) I was quite excited to see such apparent dramatic climatic Li isotopic signatures being preserved in shallow marine sediments over the PETM CIE. Now being able to see the bouncing of Li isotopic compositions over the entire section makes the initial changes look less significant and (playing Devil’s advocate) in the worst case coincidental. The discussion here is fine, but I have the impression we still need to understand the underlying controls in generating lithogenic Li isotopes better.
  
  On the contrary, we found the lithium isotopic results exciting! The study of Pogge von Strandmann et al. (2021) highlighted the perturbation of ocean Li isotopes using three carbonate localities, as well as large Li isotope excursions from two siliciclastic, shallow marine localities. These epicontinental sea sections are affected by numerous processes, so the clear signal of increased weathering at the start of the PETM is a key finding. The continued perturbation to the Li isotopic system into the Eocene is not that surprising, given the changes to regional uplift, subsidence, eustatic sea level change, flood basalt volcanism, hydrothermal venting, and explosive eruptions. An important consideration is the duration of these disturbances. The δ⁷Li excursion of -4‰ at the PETM onset occurred in <10 kyr, while the δ⁷Li excursion of -4‰ encompassing the start of the positive ash series took place over an estimated 160 kyr. This latter timescale suggests a more gradual response to changing weathering conditions in the vicinity of the eastern North Sea catchment, such as an increase in ash-forming, explosive eruptions that increased silicate weathering through enhanced particle surface area. We also note that perhaps we should not find rapid Li isotope perturbations in detrital sections surprising, given their local nature. It suggests that they are significantly more sensitive to weathering changes over short timescales than the much more buffered marine carbonate Li isotope records, which therefore makes them a very interesting archive to explore. To emphasise this distinction, we have added the estimated durations of the δ⁷Li excursions to the first paragraph of section 5.3.
  
  Citation: https://doi.org/10.5194/egusphere-2023-36-AC3
CC2:
'Comment on egusphere-2023-36 (Jones et al.) by Gernon et al.', Thomas Gernon, 22 May 2023

Comment on: “Tracing North Atlantic volcanism and seaway connectivity across the Paleocene–Eocene Thermal Maximum (PETM)” in EGUsphere
Following the recent public comments made by Jones et al. (in review) regarding our recent paper (Gernon et al., Nat. Geosci., 2022; https://doi.org/10.1038/s41561-022-00967-6), we feel compelled to respond, as we have identified some key errors and misrepresentations of our work in their comments. Unfortunately, the authors did not communicate their specific concerns directly to us prior to posting or alert us to their public comments on our study. Our primary concern here is to ensure the accurate representation of our work. In this Comment (attached; Comment_Preprint egusphere-2023-36_TG.pdf), we address the specific text we take issue with, presented by Jones and co-authors in response to a review by Marcus Gutjahr, whilst also providing some important clarifications.

Citation: https://doi.org/10.5194/egusphere-2023-36-CC2
- AC4: 'Reply on CC2', Morgan Jones, 24 May 2023
  
  We thank the Climate of the Past editors for reopening the discussion process to give Tom Gernon and colleagues a chance to defend their work. Tom reached out to me privately when our manuscript first appeared online asking why they had not been cited, and I replied that we would prefer not to cite than to criticise, and that we could discuss our perceived issues with their work if he wanted. That email never received a response. When the specific request from the second reviewer to cite Gernon et al. (2022) was made, we then felt that we needed to justify our reasons for omitting this reference, hence our response which included some discussion of the paper in question. We appreciate the time taken by Tom and his colleagues to respond here. We will briefly address some of the points raised on the geochronological control in their comment at the three localities below. However, we also believe that continuing a deep discussion of the merits of another paper is beyond the scope of the review process of this manuscript.
  
  There remains a lack of (bio)stratigraphic control at Site 555: biostratigraphic datums are available for very few samples and are not supported by carbon isotope data. We believe that δ¹³C data, in combination with up-to-date biostratigraphic data, are required for pinpointing the onset and duration of the PETM.
  For the flood basalt sequences in the Faroe Islands and Greenland, we used the ages that were corrected by the extensive review paper conducted by Wilkinson et al. (2017) in our reply to Marcus Gutjahr. The full NAG-TEC Geochronological Database and Data Evaluation are freely available to download in their supplementary material (https://doi.org/10.6084/m9.figshare.c.3554472). The authors did not discuss the extensive and thus far unresolved literature debate as to whether the PETM is above or below the hiatus in the Faroe Islands flood basalts in their paper. As we mentioned in our reply to Marcus Gutjahr, a recalibrated Ar/Ar age of 55.57 ± 0.35 Ma (uncorrected 54.9 ± 0.35 Ma) from the Middle basalt series in the Faroe Islands (Storey et al., 2007b) gives a post-PETM age for these lavas. The correction by Gernon et al. of an Ar/Ar age from ~56.1 Ma (in the published version of their work) to ~56.47 Ma (in the here posted comment) from the Milne Land Formation in Greenland is welcome, and while different to the corrected age in Wilkinson et al (2017) (56.79 ± 0.25 Ma, 1σ), appears to correlate reasonably well with the regional stratigraphy. However, this suggests that the post-hiatus, high Ti basalts observed in East Greenland and the Faroes may not be synchronous.
  We do not agree with the assertions in the comment here that “this recalculated age is indistinguishable (0.36 ± 0.59 Ma) from the age determined by Storey et al. (2007)”, and “On face value this age presented by Jones et al. is 0.44 ± 0.35 Ma younger than the PETM, which is again not inconsistent with Gernon et al.” as the validity of their hypothesized model hinges on good stratigraphical and chronological control. Considering the PETM is a 200 kyr-long event, a 360 kyr correction (let alone 690 kyr as suggested by the Wilkinson et al. 2017 recalibrated data) to the age of the base of the Milne Land Formation is a significant deviation from their published work. Therefore, we remain unconvinced that these sparse geochronological data are robust enough to conclusively identify the PETM, or pinpoint its onset and duration at the key Site 555 and the two flood basalt localities. Respectfully, we stand by our decision not to cite this paper, but we look forward to seeing future testing and development of those ideas.
  
  Citation: https://doi.org/10.5194/egusphere-2023-36-AC4

Morgan T. Jones et al.

Supplement

https://doi.org/10.5194/egusphere-2023-36-supplement

Morgan T. Jones et al.

Viewed

Total article views: 998 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	Supplement	BibTeX	EndNote
696	274	28	998	34	7	13

HTML: 696
PDF: 274
XML: 28
Total: 998
Supplement: 34
BibTeX: 7
EndNote: 13

Views and downloads (calculated since 16 Jan 2023)

Month	HTML	PDF	XML	Total
Jan 2023	202	113	6	321
Feb 2023	155	59	6	220
Mar 2023	108	30	5	143
Apr 2023	48	27	0	75
May 2023	179	45	10	234
Jun 2023	4	1	5

Cumulative views and downloads (calculated since 16 Jan 2023)

Month	HTML	PDF	XML	Total
Jan 2023	202	113	6	321
Feb 2023	155	59	6	220
Mar 2023	108	30	5	143
Apr 2023	48	27	0	75
May 2023	179	45	10	234
Jun 2023	4	1	5

Viewed (geographical distribution)

Total article views: 993 (including HTML, PDF, and XML) Thereof 993 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 02 Jun 2023

Short summary

There are periods in Earth’s history when huge volumes of magma are erupted at the Earth’s surface. The gases released from volcanic eruptions and from sediments heated by the magma are believed to have caused severe climate changes in the geological past. We use a variety of volcanic and climatic tracers to assess how the North Atlantic Igneous Province (56–54 million years ago) affected the oceans and atmosphere during a period of extreme global warming.


Total:	0
HTML:	0
PDF:	0
XML:	0