Rapid onset of ocean anoxia shown by high U and low Mo isotope compositions of sapropel S1
Affiliations | Corresponding Author | Cite as | Funding information- Share this article
-
Article views:350Cumulative count of HTML views and PDF downloads.
- Download Citation
- Rights & Permissions
top
Abstract
Figures
Figure 1 Depth profiles for sapropels S1 (top) and S5 (bottom). The S5 data are from Andersen et al. (2018), the S1 Mo data from Azrieli-Tal et al. (2014). Dashed lines indicate the position of the sapropel boundaries as determined from Ba/Al ratios. The TOC profile in S1 indicates post-sapropel oxidation in the upper 4 cm (see SI). A potential post-deposition ventilation event at ∼116 cm (8 ka) is marked with weakly stippled line for S1. Vertical lines in the isotope panels represent approximate seawater compositions. Note the larger scale bars for S5 than S1 for the U (x2) and Mo (x4) concentrations. | Figure 2 δ98Moauth vs. δ238Uauth crossplots for sapropels (main plot), as well as four Proterozoic to Devonian organic-rich shales (Asael et al., 2013; Kendall et al., 2015; Cheng et al., 2020; Kendall et al., 2020). The main plot shows S1 (blue, filled samples 130–116 cm, stippled 116–110 cm), sapropel S5 (orange) and background sediments. Compositions for modern day seawater and average deep sediments in Cariaco Basin and Black Sea Unit I/II (cores 12GGC/70GGC and 32MUC24, respectively; Brüske et al., 2020) are also shown. Correlated linear relations for the sapropels from start towards peak sapropel maxima are shown by the arrows for S1 and S5. The four palaeo-data sets (converted to authigenic compositions – see SI), identify high δ238U values combined with low δ98Mo, and generally similar evolutions to the S1 trend in the main plot (S5 and S1 trends marked by stippled lines). |
Figure 1 | Figure 2 |
top
Introduction
The redox sensitive behaviour and associated isotope fractionations of molybdenum (Mo) and uranium (U) have led to a prominent role for their sedimentary and isotope geochemistry in the evaluation of ancient ocean redox conditions. Both elements can show strong authigenic (seawater-derived) enrichment in marine organic carbon-rich sediments, accompanied by distinctive isotope signatures (e.g., Asael et al., 2013
Asael, D., Tissot, F.L., Reinhard, C.T., Rouxel, O., Dauphas, N., Lyons, T.W., Ponzevera, E., Liorzou, C., Chéron, S. (2013) Coupled molybdenum, iron and uranium stable isotopes as oceanic paleoredox proxies during the Paleoproterozoic Shunga Event. Chemical Geology 362, 193–210.
; Kendall et al., 2015Kendall, B., Komiya, T., Lyons, T.W., Bates, S.M., Gordon, G.W., Romaniello, S.J., Jiang, G., Creaser, R.A., Xiao, S., McFadden, K., Sawaki, Y. (2015) Uranium and molybdenum isotope evidence for an episode of widespread ocean oxygenation during the late Ediacaran Period. Geochimica et Cosmochimica Acta 156, 173–193.
; Brüske et al., 2020Brüske, A., Weyer, S., Zhao, M.-Y., Planavsky, N.J., Wegwerth, A., Neubert, N., Dellwig, O., Lau, K.V., Lyons, T.W. (2020) Correlated molybdenum and uranium isotope signatures in modern anoxic sediments: implications for their use as paleo-redox proxy. Geochimica et Cosmochimica Acta 270, 449–474.
).Uranium isotopes in anoxic and organic carbon-rich marine sediments (expressed as δ238U, parts per thousand deviation from the CRM145 standard) generally show values considerably heavier than contemporary seawater (−0.39 ‰; Andersen et al., 2017
Andersen, M.B., Stirling, C.H., Weyer, S. (2017) Uranium isotope fractionation. Reviews in Mineralogy and Geochemistry 82, 799–850.
). This isotope enrichment is compatible with partial reduction of U+6 to U+4. In modern organic carbon-rich sediments, authigenic δ238U values rarely exceed +0.2 ‰, equivalent to a Δ238Usediment-water of ∼ +0.6 ‰, approximately half the estimated full isotope fractionation accompanying U reduction (Andersen et al., 2017Andersen, M.B., Stirling, C.H., Weyer, S. (2017) Uranium isotope fractionation. Reviews in Mineralogy and Geochemistry 82, 799–850.
). This has been interpreted in terms of a diffusion-reaction-transport model combining U reduction in the semi-closed sediment system, limited by U diffusion from bottom water through pore waters (Andersen et al., 2014Andersen, M.B. Romaniello, S., Vance, D., Little, S.H., Herdman, R., Lyons, T.W. (2014) A modern framework for the interpretation of 238U/235U in studies of ancient ocean redox. Earth and Planetary Science Letters 400, 184–194.
). However, an increasing number of ancient organic carbon-rich sediments show authigenic δ238U values higher than +0.2 ‰, suggesting redox-driven U isotope fractionation closer to that expected in an open system (Δ238U > +0.6 ‰), and pointing towards U reduction outside the sediments (e.g., Brüske et al., 2020Brüske, A., Weyer, S., Zhao, M.-Y., Planavsky, N.J., Wegwerth, A., Neubert, N., Dellwig, O., Lau, K.V., Lyons, T.W. (2020) Correlated molybdenum and uranium isotope signatures in modern anoxic sediments: implications for their use as paleo-redox proxy. Geochimica et Cosmochimica Acta 270, 449–474.
; Cheng et al., 2020Cheng, M., Li, C., Jin, C., Wang, H., Algeo, T.J., Lyons, T.W., Zhang, F. and Anbar, A. (2020) Evidence for high organic carbon export to the early Cambrian seafloor. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.050.
; Kendall et al., 2020Kendall, B., Wang, J., Zheng, W., Romaniello, S.J., Over, D.J., Bennett, Y., Xing, L., Kunert, A., Boyes, C., Liu, J. (2020) Inverse correlation between the molybdenum and uranium isotope compositions of Upper Devonian black shales caused by changes in local depositional conditions rather than global ocean redox variations. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.026.
).In addition, coupled measurements of molybdenum isotope compositions (expressed as 98Mo/95Mo, parts per thousand deviation from NIST 3134 at +0.25; Nägler et al., 2014
Nägler, T.F., Anbar, A.D., Archer, C., Goldberg, T., Gordon, G.W., Greber, N.D., Siebert, C., Sohrin, Y., Vance, D. (2014) Proposal for an international molybdenum isotope measurement standard and data representation. Geostandards. Geoananalytical Research 38, 149–151.
) in these high δ238U organic carbon-rich sediments, generally show δ98Mo significantly lower than the modern seawater value of ∼ +2.3 ‰ (e.g., Cheng et al., 2020Cheng, M., Li, C., Jin, C., Wang, H., Algeo, T.J., Lyons, T.W., Zhang, F. and Anbar, A. (2020) Evidence for high organic carbon export to the early Cambrian seafloor. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.050.
; Kendall et al., 2020Kendall, B., Wang, J., Zheng, W., Romaniello, S.J., Over, D.J., Bennett, Y., Xing, L., Kunert, A., Boyes, C., Liu, J. (2020) Inverse correlation between the molybdenum and uranium isotope compositions of Upper Devonian black shales caused by changes in local depositional conditions rather than global ocean redox variations. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.026.
). Near seawater δ98Mo values are proposed to reflect euxinic water conditions, where dissolved [H2S]aq exceeds a ‘switch point’ (∼11 μM for modern seawater) at which seawater MoO42− becomes irreversibly and near-quantitatively converted to a highly particle reactive tetra-thiomolybdate (MoS42−) species (Erickson and Helz, 2000Erickson, B.E., Helz, G.R. (2000) Molybdenum(VI) speciation in sulfidic waters: Stability and lability of thiomolybdates. Geochimica et Cosmochimica Acta 64, 1149–1158.
; Arnold et al., 2004Arnold, G.L., Anbar, A.D., Barling, J., Lyons, T. (2004) Molybdenum isotope evidence for widespread anoxia in Mid-Proterozoic oceans. Science 304, 87–90.
). Euxinic waters with [H2S]aq below the switch point, may lead to Mo deposition with significantly lower δ98Mo values because Mo removal occurs from isotopically fractionated oxy-thiomolybdate species coexisting with molybdate (Neubert et al., 2008Neubert, N., Nägler, T.F., Böttcher, M.E. (2008) Sulfidity controls molybdenum isotope fractionation in euxinic sediments: evidence from the modern Black Sea. Geology 36, 775–778.
; Kerl et al., 2017Kerl, C.F., Lohmayer, R., Bura-Nakić, E., Vance, D., Planer-Friedrich, B. (2017) Experimental confirmation of isotope fractionation in thiomolybdates using ion chromatography and detection by multi-collector ICP-MS. Anaytical Chemistry 89, 3123–3129.
). An alternative mechanism for the generation of isotopically light Mo relates to the delivery of authigenic Mo to sediments via a Fe-Mn oxide particulate shuttle (Kendall et al., 2017Kendall, B., Dahl, T.W., Anbar, A.D. (2017) The stable isotope geochemistry of molybdenum. Reviews in Mineralogy and Geochemistry 82, 683–732.
).Though this combination of high δ238U and low δ98Mo is observed in a range of ancient sediments, an understanding of the specific biogeochemical and oceanographic settings that cause it is hindered by the lack of well-studied modern analogues. The periodic deposition of organic carbon-rich sediments (sapropels) in the Eastern Mediterranean (EM) sea provide a means to study the behaviour of this coupled redox proxy system. In particular, the study of sapropels of different intensity at different times in the same physical setting allows the identification of signatures driven by different climatic and oceanographic forcing mechanisms.
top
Authigenic U and Mo Isotopes in Sapropels S1 and S5
This study investigates the contrasting U and Mo isotope systematics of EM sapropels S1 from the Holocene and S5 from the last interglacial (abbreviated ‘S1’ and ‘S5’ in the following), sampled in ODP core 967 (SI; Fig. S-1). Sapropel S5 (∼128.3–121.5 ka) shows strong authigenic enrichment of U and Mo, coupled to progressive δ98Moauth increase (+1.2 to +2.3 ‰) and δ238Uauth decrease (+0.10 to −0.15 ‰), from the beginning to the end (Andersen et al., 2018
Andersen, M.B., Matthews, A., Vance, D., Bar-Matthews, M., Archer, C., de Souza, G.F. (2018) A 10-fold decline in the deep Eastern Mediterranean thermohaline overturning circulation during the last interglacial period. Earth and Planetary Science Letters 503, 58–67.
). Modelling of the S5 δ238U indicates a progressive 10 fold decline in the deep EM thermohaline overturning circulation. A study of S1 (10.8–6.1 ka) has shown that it is less strongly developed than S5, with weakly sulfidic bottom water conditions interpreted from low δ98Mo values (Azrieli-Tal et al., 2014Azrieli-Tal, I., Matthews, A., Bar-Matthews, M., Almogi-Labin, A., Vance, D., Archer, C., Teutsch, N. (2014) Evidence from molybdenum and iron isotopes and molybdenum–uranium covariation for sulphidic bottom waters during Eastern Mediterranean sapropel S1 formation. Earth and Planetary Science Letters 393, 231–242.
), although anoxic conditions were regionally prevalent at water depths >1800 m (de Lange et al., 2008de Lange, G.J., Thomson, J., Reitz, A., Slomp, C.P., Speranza Principato, M., Erba, E., Corselli, C. (2008) Synchronous basin-wide formation and redox-controlled preservation of a Mediterranean sapropel. Nature Geoscience 1, 606–610.
).Here, we report additional U isotope measurements for S1, allowing a direct comparison of the U and Mo isotope systematics for S1 and S5 (methods, age models and additional data are summarised in SI). The S1 samples show lower U concentrations (10–40 ppm) than S5 (25–150 ppm), but high δ238U values, reaching +0.35 to +0.5 ‰ in the lower and middle zones at 127–116 cm (Fig. 1). Thus, both the U concentration and δ238U patterns in S1 differ significantly from S5. Calculated authigenic U (Uauth) and Mo (Moauth) concentrations in the S1 profile show that detrital contributions are only significant in the background sediments, whereas authigenic contributions dominate in S1 (85–96 %). High Uauth concentrations are correlated with high δ238Uauth values (> +0.2 ‰) and high Moauth concentrations with low δ98Mo values (Fig. 1). Peak δ238Uauth values (127–116 cm) vary from +0.41 ± 0.06 ‰ to +0.52 ± 0.06 ‰, whereas δ238Uauth values of –0.7 to +0.1 ‰ occur before and after S1 and in the first sample at its start (130 cm), indicating that a significant proportion of the Uauth may be associated with U deposited directly with organic matter, carrying low δ238Uauth (Andersen et al., 2017
Andersen, M.B., Stirling, C.H., Weyer, S. (2017) Uranium isotope fractionation. Reviews in Mineralogy and Geochemistry 82, 799–850.
), in addition to in situ U reduction (see SI for details).top
Controls on U and Mo Isotope Fractionation in S1 and S5
Sapropels S1 and S5 show contrasting behaviour in their δ238U- δ98Mo systematics, which may relate to different climatic and oceanographic mechanisms creating sediment and/or seawater anoxia. Firstly, the timing of peak sapropel development, as expressed in redox sensitive trace metal profiles, represents a key difference between the two sapropels. Whereas redox sulfide sensitive proxies for S1 (S, Fe/Al, V/Al, Mo/Al, As/Al) indicate rapid and early development to maximum intensity at ca. 120 cm, the same proxies show that S5 developed more slowly, only reaching maximum values in the upper sapropel (SI; Fig. S-2). The higher concentrations of redox sensitive trace metals (e.g., U and Mo concentration profiles; Fig. 1) also demonstrate the more reducing character of S5. In contrast, the two sapropels do not show any systematic differences in TOC at peak sapropel conditions (Fig. 1), and they have similar calculated organic carbon accumulation rates (OCAR) of 261 ± 36 mg/cm2·kyr for S1 and 178 ± 48 mg/cm2·kyr for S5 (see SI for details).
The δ98Moauth of S1 shows a significant early drop towards the low values of ∼ −0.8 ‰ at peak sapropel conditions, followed by a rise to ∼0 ‰ (Fig. 1). The low S1 δ98Mo values have been interpreted to represent Mo uptake from isotopically fractionated oxy-thiomolybdate species at bottom water sulfide levels below that at which Mo is effectively and near-quantitatively converted to tetra-thiomolybdate (Azrieli-Tal et al., 2014
Azrieli-Tal, I., Matthews, A., Bar-Matthews, M., Almogi-Labin, A., Vance, D., Archer, C., Teutsch, N. (2014) Evidence from molybdenum and iron isotopes and molybdenum–uranium covariation for sulphidic bottom waters during Eastern Mediterranean sapropel S1 formation. Earth and Planetary Science Letters 393, 231–242.
). The production of thiomolybdates requires sulfate reduction, driven by biological productivity and long deep water replacement times. Non-quantitative conversion of Mo to thiomolybdate requires low sulfide and sequestration of more sulfidised species to the sediment, e.g., by colloidal FeMo(VI)S4 (Helz and Vorlicek, 2019Helz, G.R., Vorlicek, T.P. (2019) Precipitation of molybdenum from euxinic waters and the role of organic matter. Chemical Geology 509, 178–193.
). Low sulfide levels and non-quantitative conversion of molybdate to thiomolybdates can result in isotopically light sedimentary Mo without invoking a Fe-Mn shuttle, via non-quantitative Mo removal from bottom waters, in a similar way as suggested for a range of modern semi-closed basins also showing authigenic δ98Mo values below seawater (Bura-Nakic et al., 2018Bura-Nakić, E., Andersen, M.B., Archer, D., de Souza, G.F., Marguš, M., Vance, D. (2018) Coupled Mo-U abundances and isotopes in a small marine euxinic basin: Constraints on processes in euxinic basins. Geochimica et Cosmochimica Acta 222, 212–222.
). In S5, with more intense sulfate reduction resulting from protracted thermohaline overturning slowdown and high OCAR, the δ98Moauth are close to the seawater values (Andersen et al., 2018Andersen, M.B., Matthews, A., Vance, D., Bar-Matthews, M., Archer, C., de Souza, G.F. (2018) A 10-fold decline in the deep Eastern Mediterranean thermohaline overturning circulation during the last interglacial period. Earth and Planetary Science Letters 503, 58–67.
).In contrast to S5, S1 is characterised by high δ238U, higher than expected from reduction of seawater-derived U in a semi-closed sediment-porewater system (Andersen et al., 2014
Andersen, M.B. Romaniello, S., Vance, D., Little, S.H., Herdman, R., Lyons, T.W. (2014) A modern framework for the interpretation of 238U/235U in studies of ancient ocean redox. Earth and Planetary Science Letters 400, 184–194.
). Post-sapropel oxidative burndown leading to diagenetic redistribution of previously deposited, isotopically fractionated U downwards, could provide a mechanism for creating high δ238U. However, the expression of the oxidative burndown in S1 is distinct and sharp for all the redox sensitive metals. There is no evidence for downwards U diffusion and reprecipitation (e.g., a gradient in U/Al and δ238U with depth – see SI for details). Alternatively, U isotope reactive-transport modelling has suggested that high δ238U could be a consequence of low sedimentation rates and high productivity in reducing sediments (Lau et al., 2020Lau, K.V., Lyons, T.W., Maher, K. (2020) Uranium reduction and isotopic fractionation in reducing sediments: Insights from reactive transport modelling. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.021.
). However, the similar sedimentation rates for S1 and S5, also make this an unlikely process given their very different δ238U signatures (Fig. 1). Direct U reduction in the water column, followed by scavenging and transport to the sediments, is another mechanism that could lead to high δ238U (e.g., Brüske et al., 2020Brüske, A., Weyer, S., Zhao, M.-Y., Planavsky, N.J., Wegwerth, A., Neubert, N., Dellwig, O., Lau, K.V., Lyons, T.W. (2020) Correlated molybdenum and uranium isotope signatures in modern anoxic sediments: implications for their use as paleo-redox proxy. Geochimica et Cosmochimica Acta 270, 449–474.
; Kendall et al., 2020Kendall, B., Wang, J., Zheng, W., Romaniello, S.J., Over, D.J., Bennett, Y., Xing, L., Kunert, A., Boyes, C., Liu, J. (2020) Inverse correlation between the molybdenum and uranium isotope compositions of Upper Devonian black shales caused by changes in local depositional conditions rather than global ocean redox variations. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.026.
). But the stronger uptake of redox sensitive metals in S5 than S1 (Fig. 1) would likely favour high δ238U in S5, not S1.Another model for providing a greater expression of the Δ238Usediment–water involves U reduction promoted by an organic-rich floccule layer overlying the consolidated sediment (Andersen et al., 2017
Andersen, M.B., Stirling, C.H., Weyer, S. (2017) Uranium isotope fractionation. Reviews in Mineralogy and Geochemistry 82, 799–850.
; Cheng et al., 2020Cheng, M., Li, C., Jin, C., Wang, H., Algeo, T.J., Lyons, T.W., Zhang, F. and Anbar, A. (2020) Evidence for high organic carbon export to the early Cambrian seafloor. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.050.
). The formation of an organic-rich floccule layer is likely to occur in settings with high biological productivity and organic carbon flux. However, the combined Mo-U data for S1 require a further condition: light Mo isotopes require the partial conversion of Mo to thiomolybdate, and therefore mildly sulfidic levels. Redox sensitive element data for S1 (Figs. 1, S-2) make a strong case for rapid early development of euxinia, which could promote deposition of such an organic-rich floccule layer. High resolution simulations of intermediate and deep water circulation suggest that enhanced Nile discharge could have triggered rapid shutdown of EM overturning circulation (Vadsaria et al., 2019Vadsaria, T., Ramstein, G., Dutay, J.-C., Li, L., Ayache, M., Richon, C. (2019) Simulating the occurrence of the last sapropel event (S1): Mediterranean basin ocean dynamics simulations using Nd isotopic composition modeling. Paleoceanography and Paleoclimatology 34, 237–251, doi: 10.1029/2019PA003566.
). The early thermohaline circulation slowdown would potentially promote the development of mildly sulfidic waters.top
Outlook: the U-Mo Isotope Proxy Applied to Ancient Sediments
Both S1 and S5 show inverse correlations between δ238Uauth and δ98Moauth (Fig. 2). However, temporally, the trends go in opposite directions. For S5, the end point of the correlation is defined by maximal euxinic conditions at the end of the sapropel. In S1, the end point of the correlation is represented by the early sapropel peak. This difference in temporal progression for S1 and S5 is likely to reflect the fundamentally different modes of Mo and U incorporation into the sediments. Whilst the S5 data likely records progressive slowdown and increasingly euxinic conditions throughout the entirety of sapropel formation (Andersen et al., 2018
Andersen, M.B., Matthews, A., Vance, D., Bar-Matthews, M., Archer, C., de Souza, G.F. (2018) A 10-fold decline in the deep Eastern Mediterranean thermohaline overturning circulation during the last interglacial period. Earth and Planetary Science Letters 503, 58–67.
), S1 seems to require rapid short-lived euxinia caused by early thermohaline slowdown. These two scenarios provide a potential framework for the interpretation of palaeo-redox evolution. Using the sapropel data, it is possible to delineate three general behaviours (Fig. 2). The first (1) represents mildly euxinic sediments typical of S1, with U and Mo uptake associated with an organic floccule layer. The second (2) is associated with non-quantitative U and Mo uptake in more intensely euxinic conditions and with porewater U reduction. The third (3) involves strongly euxinic sediments moving towards near-quantitative Mo and U uptake, as seen in S5 and in modern restricted basins with variable deep water overturning rates (Fig. 2).Trend (1), seen in S1, is very similar to recent published combined δ238Uauth - δ98Moauth data from Cambrian and Devonian organic carbon-rich sediments (Cheng et al., 2020
Cheng, M., Li, C., Jin, C., Wang, H., Algeo, T.J., Lyons, T.W., Zhang, F. and Anbar, A. (2020) Evidence for high organic carbon export to the early Cambrian seafloor. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.050.
; Kendall et al., 2020Kendall, B., Wang, J., Zheng, W., Romaniello, S.J., Over, D.J., Bennett, Y., Xing, L., Kunert, A., Boyes, C., Liu, J. (2020) Inverse correlation between the molybdenum and uranium isotope compositions of Upper Devonian black shales caused by changes in local depositional conditions rather than global ocean redox variations. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.026.
). These sediments show high δ238U values (up to +0.6 ‰), combined with low δ98Mo (down to −0.1 ‰). Also, Ediacaran organic–rich mudstones (Kendall et al., 2015Kendall, B., Komiya, T., Lyons, T.W., Bates, S.M., Gordon, G.W., Romaniello, S.J., Jiang, G., Creaser, R.A., Xiao, S., McFadden, K., Sawaki, Y. (2015) Uranium and molybdenum isotope evidence for an episode of widespread ocean oxygenation during the late Ediacaran Period. Geochimica et Cosmochimica Acta 156, 173–193.
) and the Palaeoproterozoic Zaonega Formation shales (Asael et al., 2013Asael, D., Tissot, F.L., Reinhard, C.T., Rouxel, O., Dauphas, N., Lyons, T.W., Ponzevera, E., Liorzou, C., Chéron, S. (2013) Coupled molybdenum, iron and uranium stable isotopes as oceanic paleoredox proxies during the Paleoproterozoic Shunga Event. Chemical Geology 362, 193–210.
) show an inverse correlation which parallels the type 1 (S1) trend. Limited Black Sea unit II data (Brüske et al., 2020Brüske, A., Weyer, S., Zhao, M.-Y., Planavsky, N.J., Wegwerth, A., Neubert, N., Dellwig, O., Lau, K.V., Lyons, T.W. (2020) Correlated molybdenum and uranium isotope signatures in modern anoxic sediments: implications for their use as paleo-redox proxy. Geochimica et Cosmochimica Acta 270, 449–474.
) show similarly high δ238Uauth (∼ +0.55 ‰) to S1, while δ98Moauth (+1.23 ‰) is relatively high (Fig. 2). All these localities exhibit similar negatively correlated δ238U- δ98Mo values as seen in S1, but at variable trajectories and with different δ98Mo values. This observation may be partly attributed to variable seawater δ98Mo over time, but it would also depend on the [H2S]aq levels and the specific stability of the oxy-thiomolybdate species in each setting.A key observation is that although high productivity in upwelling-type settings has been suggested to be the critical driver for high δ238U in organic carbon-rich sediments (Cheng et al., 2020
Cheng, M., Li, C., Jin, C., Wang, H., Algeo, T.J., Lyons, T.W., Zhang, F. and Anbar, A. (2020) Evidence for high organic carbon export to the early Cambrian seafloor. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.050.
), this does not fit well with the observations in this study. Although high productivity is an important factor it is not the defining one (e.g., S1 and S5 have similar TOC contents and calculated OCAR, yet show completely contrasting behaviour in the U and Mo isotope systematics). From the sapropel data it may be inferred that the specific conditions that favour combined high δ238Uauth with low δ98Moauth require both a high particulate to water ratio (e.g., the organic floccule layer) and (low) [H2S]aq levels, to aid the reduction of U and formation of oxy-thiomolybdates, followed by Mo and U scavenging by particulate/organic matter to the sediments. In the case of S1 this likely occurs via rapid thermohaline slowdown combined with high productivity and formation of an organic floccule layer. In other settings with high productivity in upwelling zones, the development of [H2S]aq bottom water plumes as observed on the Peru margin (Schunck et al., 2013Schunck, H., Lavik, G., Desai, D. K., Grosskopf, T., Kalvelage, T., Loescher, C.R., Paulmier A., Contreras, S., Siegel, H., Holtappels, M., Rosenstiel, P., Schilhabel, M.B., Graco, M., Schmitz, R.A., Kuypers, M.M.M., LaRoche, J. (2013) Giant hydrogen sulfide plume in the oxygen minimum zone off Peru supports chemolithoautotrophy. PLoS ONE 8, doi: 10.1371/journal.pone.0068661.
), combined with an organic floccule layer, could also provide the right condition for high δ238Uauth with low δ98Moauth in organic carbon-rich sediments. The palaeo-examples discussed here show that conditions that favour these S1 type δ238U- δ98Mo sediment systematics likely occurred across a large part of Earth history. Yet, in inferring palaeo-proxy evolution from U and Mo isotope trends, it is the data systematics from individual settings that are important, as opposed to direct comparison of exact δ238U and δ98Mo compositions between settings. Based on this study, the careful and combined analysis of δ238U- δ98Mo systematics in organic carbon-rich sediments is critical to providing the most robust way of identifying localised isotope effects for both systems and, thus, oceanographic conditions during sediment deposition.top
Acknowledgements
We acknowledge financial support by the Natural Environment Research Council, UK, grant number NE/H023933/1 and Israel Science Foundation grant 1140/12. Olga Berlin and Olga Yoffe of the Geological Survey of Israel are thanked for making the trace and major element analyses. We thank Elvira Bura-Naki´c for discussions. We thank three anonymous reviewers for constructive comments that improved the manuscript.
Editor: Satish Myneni
top
References
Andersen, M.B. Romaniello, S., Vance, D., Little, S.H., Herdman, R., Lyons, T.W. (2014) A modern framework for the interpretation of 238U/235U in studies of ancient ocean redox. Earth and Planetary Science Letters 400, 184–194.
Show in context
This has been interpreted in terms of a diffusion-reaction-transport model combining U reduction in the semi-closed sediment system, limited by U diffusion from bottom water through pore waters (Andersen et al., 2014).
View in article
In contrast to S5, S1 is characterised by high δ238U, higher than expected from reduction of seawater-derived U in a semi-closed sediment-porewater system (Andersen et al., 2014).
View in article
Andersen, M.B., Stirling, C.H., Weyer, S. (2017) Uranium isotope fractionation. Reviews in Mineralogy and Geochemistry 82, 799–850.
Show in context
Uranium isotopes in anoxic and organic carbon-rich marine sediments (expressed as δ238U, parts per thousand deviation from the CRM145 standard) generally show values considerably heavier than contemporary seawater (−0.39 ‰; Andersen et al., 2017).
View in article
In modern organic carbon-rich sediments, authigenic δ238U values rarely exceed +0.2 ‰, equivalent to a Δ238Usediment-water of ∼ +0.6 ‰, approximately half the estimated full isotope fractionation accompanying U reduction (Andersen et al., 2017).
View in article
Peak δ238Uauth values (127–116 cm) vary from +0.41 ± 0.06 ‰ to +0.52 ± 0.06 ‰, whereas δ238Uauth values of –0.7 to +0.1 ‰ occur before and after S1 and in the first sample at its start (130 cm), indicating that a significant proportion of the Uauth may be associated with U deposited directly with organic matter, carrying low δ238Uauth (Andersen et al., 2017), in addition to in situ U reduction (see
View in article
Another model for providing a greater expression of the Δ238Usediment–water involves U reduction promoted by an organic-rich floccule layer overlying the consolidated sediment (Andersen et al., 2017; Cheng et al., 2020).
View in article
Andersen, M.B., Matthews, A., Vance, D., Bar-Matthews, M., Archer, C., de Souza, G.F. (2018) A 10-fold decline in the deep Eastern Mediterranean thermohaline overturning circulation during the last interglacial period. Earth and Planetary Science Letters 503, 58–67.
Show in context
This study investigates the contrasting U and Mo isotope systematics of EM sapropels S1 from the Holocene and S5 from the last interglacial (abbreviated ‘S1’ and ‘S5’ in the following), sampled in ODP core 967 (
View in article
In S5, with more intense sulfate reduction resulting from protracted thermohaline overturning slowdown and high OCAR, the δ98Moauth are close to the seawater values (Andersen et al., 2018).
View in article
Whilst the S5 data likely records progressive slowdown and increasingly euxinic conditions throughout the entirety of sapropel formation (Andersen et al., 2018), S1 seems to require rapid short-lived euxinia caused by early thermohaline slowdown.
View in article
The S5 data are from Andersen et al. (2018), the S1 Mo data from Azrieli-Tal et al. (2014).
View in article
Arnold, G.L., Anbar, A.D., Barling, J., Lyons, T. (2004) Molybdenum isotope evidence for widespread anoxia in Mid-Proterozoic oceans. Science 304, 87–90.
Show in context
Near seawater δ98Mo values are proposed to reflect euxinic water conditions, where dissolved [H2S]aq exceeds a ‘switch point’ (∼11 μM for modern seawater) at which seawater MoO42− becomes irreversibly and near-quantitatively converted to a highly particle reactive tetra-thiomolybdate (MoS42−) species (Erickson and Helz, 2000; Arnold et al., 2004).
View in article
Asael, D., Tissot, F.L., Reinhard, C.T., Rouxel, O., Dauphas, N., Lyons, T.W., Ponzevera, E., Liorzou, C., Chéron, S. (2013) Coupled molybdenum, iron and uranium stable isotopes as oceanic paleoredox proxies during the Paleoproterozoic Shunga Event. Chemical Geology 362, 193–210.
Show in context
Both elements can show strong authigenic (seawater-derived) enrichment in marine organic carbon-rich sediments, accompanied by distinctive isotope signatures (e.g., Asael et al., 2013; Kendall et al., 2015; Brüske et al., 2020).
View in article
Also, Ediacaran organic–rich mudstones (Kendall et al., 2015) and the Palaeoproterozoic Zaonega Formation shales (Asael et al., 2013) show an inverse correlation which parallels the type 1 (S1) trend.
View in article
δ98Moauth vs. δ238Uauth crossplots for sapropels (main plot), as well as four Proterozoic to Devonian organic-rich shales (Asael et al., 2013; Kendall et al., 2015; Cheng et al., 2020; Kendall et al., 2020).
View in article
Azrieli-Tal, I., Matthews, A., Bar-Matthews, M., Almogi-Labin, A., Vance, D., Archer, C., Teutsch, N. (2014) Evidence from molybdenum and iron isotopes and molybdenum–uranium covariation for sulphidic bottom waters during Eastern Mediterranean sapropel S1 formation. Earth and Planetary Science Letters 393, 231–242.
Show in context
A study of S1 (10.8–6.1 ka) has shown that it is less strongly developed than S5, with weakly sulfidic bottom water conditions interpreted from low δ98Mo values (Azrieli-Tal et al., 2014), although anoxic conditions were regionally prevalent at water depths >1800 m (de Lange et al., 2008).
View in article
The low S1 δ98Mo values have been interpreted to represent Mo uptake from isotopically fractionated oxy-thiomolybdate species at bottom water sulfide levels below that at which Mo is effectively and near-quantitatively converted to tetra-thiomolybdate (Azrieli-Tal et al., 2014).
View in article
The S5 data are from Andersen et al. (2018), the S1 Mo data from Azrieli-Tal et al. (2014).
View in article
Brüske, A., Weyer, S., Zhao, M.-Y., Planavsky, N.J., Wegwerth, A., Neubert, N., Dellwig, O., Lau, K.V., Lyons, T.W. (2020) Correlated molybdenum and uranium isotope signatures in modern anoxic sediments: implications for their use as paleo-redox proxy. Geochimica et Cosmochimica Acta 270, 449–474.
Show in context
Both elements can show strong authigenic (seawater-derived) enrichment in marine organic carbon-rich sediments, accompanied by distinctive isotope signatures (e.g., Asael et al., 2013; Kendall et al., 2015; Brüske et al., 2020).
View in article
However, an increasing number of ancient organic carbon-rich sediments show authigenic δ238U values higher than +0.2 ‰, suggesting redox-driven U isotope fractionation closer to that expected in an open system (Δ238U > +0.6 ‰), and pointing towards U reduction outside the sediments (e.g., Brüske et al., 2020; Cheng et al., 2020; Kendall et al., 2020).
View in article
Direct U reduction in the water column, followed by scavenging and transport to the sediments, is another mechanism that could lead to high δ238U (e.g., Brüske et al., 2020; Kendall et al., 2020).
View in article
Limited Black Sea unit II data (Brüske et al., 2020) show similarly high δ238Uauth (∼ +0.55 ‰) to S1, while δ98Moauth (+1.23 ‰) is relatively high (Fig. 2).
View in article
The main plot shows S1 (blue, filled samples 130–116 cm, stippled 116–110 cm), sapropel S5 (orange) and background sediments. Compositions for modern day seawater and average deep sediments in Cariaco Basin and Black Sea Unit I/II (cores 12GGC/70GGC and 32MUC24, respectively; Brüske et al., 2020) are also shown.
View in article
Bura-Nakić, E., Andersen, M.B., Archer, D., de Souza, G.F., Marguš, M., Vance, D. (2018) Coupled Mo-U abundances and isotopes in a small marine euxinic basin: Constraints on processes in euxinic basins. Geochimica et Cosmochimica Acta 222, 212–222.
Show in context
Low sulfide levels and non-quantitative conversion of molybdate to thiomolybdates can result in isotopically light sedimentary Mo without invoking a Fe-Mn shuttle, via non-quantitative Mo removal from bottom waters, in a similar way as suggested for a range of modern semi-closed basins also showing authigenic δ98Mo values below seawater (Bura-Nakic et al., 2018).
View in article
Cheng, M., Li, C., Jin, C., Wang, H., Algeo, T.J., Lyons, T.W., Zhang, F. and Anbar, A. (2020) Evidence for high organic carbon export to the early Cambrian seafloor. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.050.
Show in context
However, an increasing number of ancient organic carbon-rich sediments show authigenic δ238U values higher than +0.2 ‰, suggesting redox-driven U isotope fractionation closer to that expected in an open system (Δ238U > +0.6 ‰), and pointing towards U reduction outside the sediments (e.g., Brüske et al., 2020; Cheng et al., 2020; Kendall et al., 2020).
View in article
In addition, coupled measurements of molybdenum isotope compositions (expressed as 98Mo/95Mo, parts per thousand deviation from NIST 3134 at +0.25; Nägler et al., 2014) in these high δ238U organic carbon-rich sediments, generally show δ98Mo significantly lower than the modern seawater value of ∼ +2.3 ‰ (e.g., Cheng et al., 2020; Kendall et al., 2020).
View in article
Another model for providing a greater expression of the Δ238Usediment–water involves U reduction promoted by an organic-rich floccule layer overlying the consolidated sediment (Andersen et al., 2017; Cheng et al., 2020).
View in article
Trend (1), seen in S1, is very similar to recent published combined δ238Uauth - δ98Moauth data from Cambrian and Devonian organic carbon-rich sediments (Cheng et al., 2020; Kendall et al., 2020).
View in article
A key observation is that although high productivity in upwelling-type settings has been suggested to be the critical driver for high δ238U in organic carbon-rich sediments (Cheng et al., 2020), this does not fit well with the observations in this study. Although high productivity is an important factor it is not the defining one (e.g., S1 and S5 have similar TOC contents and calculated OCAR, yet show completely contrasting behaviour in the U and Mo isotope systematics).
View in article
δ98Moauth vs. δ238Uauth crossplots for sapropels (main plot), as well as four Proterozoic to Devonian organic-rich shales (Asael et al., 2013; Kendall et al., 2015; Cheng et al., 2020; Kendall et al., 2020).
View in article
de Lange, G.J., Thomson, J., Reitz, A., Slomp, C.P., Speranza Principato, M., Erba, E., Corselli, C. (2008) Synchronous basin-wide formation and redox-controlled preservation of a Mediterranean sapropel. Nature Geoscience 1, 606–610.
Show in context
A study of S1 (10.8–6.1 ka) has shown that it is less strongly developed than S5, with weakly sulfidic bottom water conditions interpreted from low δ98Mo values (Azrieli-Tal et al., 2014), although anoxic conditions were regionally prevalent at water depths >1800 m (de Lange et al., 2008).
View in article
Erickson, B.E., Helz, G.R. (2000) Molybdenum(VI) speciation in sulfidic waters: Stability and lability of thiomolybdates. Geochimica et Cosmochimica Acta 64, 1149–1158.
Show in context
Near seawater δ98Mo values are proposed to reflect euxinic water conditions, where dissolved [H2S]aq exceeds a ‘switch point’ (∼11 μM for modern seawater) at which seawater MoO42− becomes irreversibly and near-quantitatively converted to a highly particle reactive tetra-thiomolybdate (MoS42−) species (Erickson and Helz, 2000; Arnold et al., 2004).
View in article
Helz, G.R., Vorlicek, T.P. (2019) Precipitation of molybdenum from euxinic waters and the role of organic matter. Chemical Geology 509, 178–193.
Show in context
Non-quantitative conversion of Mo to thiomolybdate requires low sulfide and sequestration of more sulfidised species to the sediment, e.g., by colloidal FeMo(VI)S4 (Helz and Vorlicek, 2019).
View in article
Kendall, B., Komiya, T., Lyons, T.W., Bates, S.M., Gordon, G.W., Romaniello, S.J., Jiang, G., Creaser, R.A., Xiao, S., McFadden, K., Sawaki, Y. (2015) Uranium and molybdenum isotope evidence for an episode of widespread ocean oxygenation during the late Ediacaran Period. Geochimica et Cosmochimica Acta 156, 173–193.
Show in context
Both elements can show strong authigenic (seawater-derived) enrichment in marine organic carbon-rich sediments, accompanied by distinctive isotope signatures (e.g., Asael et al., 2013; Kendall et al., 2015; Brüske et al., 2020).
View in article
Also, Ediacaran organic–rich mudstones (Kendall et al., 2015) and the Palaeoproterozoic Zaonega Formation shales (Asael et al., 2013) show an inverse correlation which parallels the type 1 (S1) trend.
View in article
δ98Moauth vs. δ238Uauth crossplots for sapropels (main plot), as well as four Proterozoic to Devonian organic-rich shales (Asael et al., 2013; Kendall et al., 2015; Cheng et al., 2020; Kendall et al., 2020).
View in article
Kendall, B., Dahl, T.W., Anbar, A.D. (2017) The stable isotope geochemistry of molybdenum. Reviews in Mineralogy and Geochemistry 82, 683–732.
Show in context
An alternative mechanism for the generation of isotopically light Mo relates to the delivery of authigenic Mo to sediments via a Fe-Mn oxide particulate shuttle (Kendall et al., 2017).
View in article
Kendall, B., Wang, J., Zheng, W., Romaniello, S.J., Over, D.J., Bennett, Y., Xing, L., Kunert, A., Boyes, C., Liu, J. (2020) Inverse correlation between the molybdenum and uranium isotope compositions of Upper Devonian black shales caused by changes in local depositional conditions rather than global ocean redox variations. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.026.
Show in context
However, an increasing number of ancient organic carbon-rich sediments show authigenic δ238U values higher than +0.2 ‰, suggesting redox-driven U isotope fractionation closer to that expected in an open system (Δ238U > +0.6 ‰), and pointing towards U reduction outside the sediments (e.g., Brüske et al., 2020; Cheng et al., 2020; Kendall et al., 2020).
View in article
In addition, coupled measurements of molybdenum isotope compositions (expressed as 98Mo/95Mo, parts per thousand deviation from NIST 3134 at +0.25; Nägler et al., 2014) in these high δ238U organic carbon-rich sediments, generally show δ98Mo significantly lower than the modern seawater value of ∼ +2.3 ‰ (e.g., Cheng et al., 2020; Kendall et al., 2020).
Direct U reduction in the water column, followed by scavenging and transport to the sediments, is another mechanism that could lead to high δ238U (e.g., Brüske et al., 2020; Kendall et al., 2020).
View in article
Trend (1), seen in S1, is very similar to recent published combined δ238Uauth - δ98Moauth data from Cambrian and Devonian organic carbon-rich sediments (Cheng et al., 2020; Kendall et al., 2020).
View in article
δ98Moauth vs. δ238Uauth crossplots for sapropels (main plot), as well as four Proterozoic to Devonian organic-rich shales (Asael et al., 2013; Kendall et al., 2015; Cheng et al., 2020; Kendall et al., 2020).
View in article
Kerl, C.F., Lohmayer, R., Bura-Nakić, E., Vance, D., Planer-Friedrich, B. (2017) Experimental confirmation of isotope fractionation in thiomolybdates using ion chromatography and detection by multi-collector ICP-MS. Anaytical Chemistry 89, 3123–3129.
Show in context
Euxinic waters with [H2S]aq below the switch point, may lead to Mo deposition with significantly lower δ98Mo values because Mo removal occurs from isotopically fractionated oxy-thiomolybdate species coexisting with molybdate (Neubert et al., 2008; Kerl et al., 2017).
View in article
Lau, K.V., Lyons, T.W., Maher, K. (2020) Uranium reduction and isotopic fractionation in reducing sediments: Insights from reactive transport modelling. Geochimica et Cosmochimica Acta, doi: 10.1016/j.gca.2020.01.021.
Show in context
Alternatively, U isotope reactive-transport modelling has suggested that high δ238U could be a consequence of low sedimentation rates and high productivity in reducing sediments (Lau et al., 2020).
View in article
Neubert, N., Nägler, T.F., Böttcher, M.E. (2008) Sulfidity controls molybdenum isotope fractionation in euxinic sediments: evidence from the modern Black Sea. Geology 36, 775–778.
Show in context
Euxinic waters with [H2S]aq below the switch point, may lead to Mo deposition with significantly lower δ98Mo values because Mo removal occurs from isotopically fractionated oxy-thiomolybdate species coexisting with molybdate (Neubert et al., 2008; Kerl et al., 2017).
View in article
Nägler, T.F., Anbar, A.D., Archer, C., Goldberg, T., Gordon, G.W., Greber, N.D., Siebert, C., Sohrin, Y., Vance, D. (2014) Proposal for an international molybdenum isotope measurement standard and data representation. Geostandards. Geoananalytical Research 38, 149–151.
Show in context
In addition, coupled measurements of molybdenum isotope compositions (expressed as 98Mo/95Mo, parts per thousand deviation from NIST 3134 at +0.25; Nägler et al., 2014) in these high δ238U organic carbon-rich sediments, generally show δ98Mo significantly lower than the modern seawater value of ∼ +2.3 ‰ (e.g., Cheng et al., 2020; Kendall et al., 2020).
View in article
Schunck, H., Lavik, G., Desai, D. K., Grosskopf, T., Kalvelage, T., Loescher, C.R., Paulmier A., Contreras, S., Siegel, H., Holtappels, M., Rosenstiel, P., Schilhabel, M.B., Graco, M., Schmitz, R.A., Kuypers, M.M.M., LaRoche, J. (2013) Giant hydrogen sulfide plume in the oxygen minimum zone off Peru supports chemolithoautotrophy. PLoS ONE 8, doi: 10.1371/journal.pone.0068661.
Show in context
In other settings with high productivity in upwelling zones, the development of [H2S]aq bottom water plumes as observed on the Peru margin (Schunck et al., 2013), combined with an organic floccule layer, could also provide the right condition for high δ238Uauth with low δ98Moauth in organic carbon-rich sediments.
View in article
Vadsaria, T., Ramstein, G., Dutay, J.-C., Li, L., Ayache, M., Richon, C. (2019) Simulating the occurrence of the last sapropel event (S1): Mediterranean basin ocean dynamics simulations using Nd isotopic composition modeling. Paleoceanography and Paleoclimatology 34, 237–251, doi: 10.1029/2019PA003566.
Show in context
High resolution simulations of intermediate and deep water circulation suggest that enhanced Nile discharge could have triggered rapid shutdown of EM overturning circulation (Vadsaria et al., 2019).
View in article
top
Supplementary Information
The Supplementary Information includes:
- 1. Methods
- 2. Age Model and Sapropel Event Definitions
- 3. Organic Carbon Sediment Accumulation Rates
- 4. Authigenic U and Mo Concentration and Isotope Estimates
- 5. Authigenic δ238U and δ98Mo Estimates – Estimating the Contribution of U and Mo Directly Associated with Organic Matter
- Figures S-1 to S-4
- Table S-1
- Supplementary Information References
Download Table S-1 (Excel).
Download the Supplementary Information (PDF)