Molybdenum isotope constraints on the temporal development of sulfidic conditions during Mediterranean sapropel intervals
Affiliations | Corresponding Author | Cite as | Funding information
Published by the European Association of Geochemistry
under Creative Commons License CC BY-NC-ND 4.0
Keywords: sapropels, Mediterranean, anoxia, euxinia, Mo-isotopes, molybdenum
- Share this article
Article views:404Cumulative count of HTML views and PDF downloads.
- Download Citation
- Rights & Permissions
top
Abstract
Figures
 Figure 1 Location of piston core 64PE406-E1 (this study), the LC21, and ODP 968 cores used for the age model, and ODP 967 studied by Andersen et al. (2018). |  Figure 2 δ98Mo and Mo concentration data for core 64PE406-E1. The bottom panel reflects indicators of monsoon-driven freshwater input. LC21 core ‘residuals’ (blue) show the δ18OG. ruber values up to 150 ka BP after correction for a seawater effect (Grant et al., 2016). The black line shows the modelled North African run off (CLIMBER-2; Ziegler et al., 2010). XRF core scanning data by Hennekam et al. (2020). The uncertainties (2 s.d.) on the Mo isotope data are smaller than the symbol sizes. |  Figure 3 Mo isotope systematics. (a) Surface sediment (0–1 cm depth) data from different water depths (as data labels) in the modern Black Sea (Neubert et al., 2008; Brüske et al., 2020). (b) Sedimentary δ98Mo values as a function of aqueous H2S concentrations following calculations by Matthews et al. (2017) for different relative particle affinities and seawater δ98Mo. The 2.86 ‰ value is the average of deep Black Sea water compositions (Nägler et al., 2011). δ98Moseawater can evolve to higher values in strongly restricted basins, such that sediments formed from an evolved water mass might show higher δ98Mo values. The grey circles represent constraints on [H2S] from δ98Mo that agree with all four models and are therefore used in Figure 3c to constrain [H2S] conservatively. Sediment data from Neubert et al. (2008). (c) Framework for the semi-quantitative reconstruction of [H2S]. |  Figure 4 Close up of data presented in Figure 2 for sapropel intervals, including cross plots with age labels. The orange arrows highlight similarities in temporal patterns between S5 and S7. The bottom right plot shows a schematic interpretation of temporal patterns. |
| Figure 1 | Figure 2 | Figure 3 | Figure 4 |
top
Introduction
Reconstructions of past deoxygenation events from the sedimentary record are crucial in improving projections of future deoxygenation and associated perturbations to biogeochemical cycles. Organic-rich sapropel layers in the Mediterranean reflect some of the largest scale basin deoxygenation events in recent Earth history. The setting allows the study of repeated intervals of basin deoxygenation of various intensities, at high resolution and within a well established stratigraphic framework.
The timing of Mediterranean sapropel deposition is controlled by maxima in Northern hemisphere summer insolation during precession minima, which lead to higher monsoon intensity over the North African continent (e.g., Ziegler et al., 2010
Ziegler, M., Tuenter, E., Lourens, L.J. (2010) The precession phase of the boreal summer monsoon as viewed from the eastern Mediterranean (ODP Site 968). Quaternary Science Reviews 29, 1481–1490.
; Grant et al., 2016Grant, K., Grimm, R., Mikolajewicz, U., Marino, G., Ziegler, M., Rohling, E. (2016) The timing of Mediterranean sapropel deposition relative to insolation, sea-level and African monsoon changes. Quaternary Science Reviews 140, 125–141.
). Greater freshwater input resulted in nutrient-rich conditions, enhanced water column stratification, and a decrease in local deep water formation, which ultimately caused the drawdown of dissolved oxygen at depth and the deposition of sapropels (e.g., Rohling et al., 2015Rohling, E., Marino, G., Grant, K. (2015) Mediterranean climate and oceanography, and the periodic development of anoxic events (sapropels). Earth-Science Reviews 143, 62–97.
). As the environmental boundary conditions during these events are relatively well established, sapropels present excellent analogues for intermittent anoxic and sulfidic conditions in past ocean basins (e.g., Dahl et al., 2019Dahl, T.W., Siggaard-Andersen, M.L., Schovsbo, N.H., Persson, D.O., Husted, S., Hougård, I.W., Dickson, A.J., Kjær, K., Nielsen, A.T. (2019) Brief oxygenation events in locally anoxic oceans during the Cambrian solves the animal breathing paradox. Scientific Reports 9, 1–9.
) as well as targets for modelling efforts aiming to address large scale basin deoxygenation, provided that redox conditions are quantitatively constrained.Molybdenum (Mo) concentrations of marine sediments are well established as a tracer for local marine redox conditions, while its isotopes have more recently emerged as a valuable proxy to estimate redox conditions on local to global spatial scales (e.g., Scott and Lyons, 2012
Scott, C., Lyons, T.W. (2012) Contrasting molybdenum cycling and isotopic properties in euxinic versus non-euxinic sediments and sedimentary rocks: Refining the paleoproxies. Chemical Geology 324, 19–27.
; Kendall et al., 2017Kendall, B., Dahl, T., Anbar, A.D. (2017) The stable isotope geochemistry of molybdenum. Reviews in Mineralogy and Geochemistry 82, 683–732.
and references therein). In oxic waters, dissolved Mo is mostly present as the molybdate ion (MoO42−), which behaves conservatively. However, in the presence of H2S, conversion of molybdate to particle-reactive thiomolybdate species (MoO4−xSx2−) results in the enhanced export and burial of Mo (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.
). High sedimentary Mo enrichments can thus be taken to reflect sulfidic water column conditions (e.g., Scott and Lyons, 2012Scott, C., Lyons, T.W. (2012) Contrasting molybdenum cycling and isotopic properties in euxinic versus non-euxinic sediments and sedimentary rocks: Refining the paleoproxies. Chemical Geology 324, 19–27.
).The Mo isotope composition of organic-rich sediments is generally lighter than seawater (2.34 ‰; Nakagawa et al., 2012
Nakagawa, Y., Takano, S., Firdaus, M.L., Norisuye, K., Hirata, T., Vance, D., Sohrin, Y. (2012) The molybdenum isotopic composition of the modern ocean. Geochemical Journal 46, 131–141.
), with reported δ98Mo values as low as ∼ −0.7 ‰ (e.g., Andersen et al., 2020Andersen, M.B., Matthews, A., Bar-Matthews, M., Vance, D. (2020) Rapid onset of ocean anoxia shown by high U and low Mo isotope compositions of sapropel S1. Geochemical Perspectives Letters 15, 10–14.
; Brüske et al., 2020Brüske, A., Weyer, S., Zhao, M.-Y., Planavsky, N., Wegwerth, A., Neubert, N., Dellwig, O., Lau, K., Lyons, T. (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.
). Low δ98Mo values in mildly euxinic environments can be attributed to isotope fractionation associated with the incomplete conversion of molybdate to thiomolybdate species, with the more sulfidised species becoming progressively isotopically lighter (Tossell, 2005Tossell, J. (2005) Calculating the partitioning of the isotopes of Mo between oxidic and sulfidic species in aqueous solution. Geochimica et Cosmochimica Acta 69, 2981–2993.
; 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 chromatographic separation and detection by multicollector ICPMS. Analytical Chemistry 89, 3123–3129.
). Large isotopic differences between sediments and seawater for such sediments can be modelled by the kinetically controlled preferential scavenging of the more sulfidised thiomolybdate species (Dahl et al., 2010Dahl, T.W., Anbar, A.D., Gordon, G.W., Rosing, M.T., Frei, R., Canfield, D.E. (2010) The behavior of molybdenum and its isotopes across the chemocline and in the sediments of sulfidic Lake Cadagno, Switzerland. Geochimica et Cosmochimica Acta 74, 144–163.
; Matthews et al., 2017Matthews, A., Azrieli-Tal, I., Benkovitz, A., Bar-Matthews, M., Vance, D., Poulton, S.W., Teutsch, N., Almogi-Labin, A., Archer, C. (2017) Anoxic development of sapropel S1 in the Nile Fan inferred from redox sensitive proxies, Fe speciation, Fe and Mo isotopes. Chemical Geology 475, 24–39.
). Sedimentary Mo isotope values only approach the original seawater composition in strongly restricted and euxinic basins (e.g., Neubert et al., 2008Neubert, N., Nägler, T.F., Böttcher, M.E. (2008) Sulfidity controls molybdenum isotope fractionation into euxinic sediments: Evidence from the modern Black Sea. Geology 36, 775–778.
; Brüske et al., 2020Brüske, A., Weyer, S., Zhao, M.-Y., Planavsky, N., Wegwerth, A., Neubert, N., Dellwig, O., Lau, K., Lyons, T. (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.
). Mo isotope data have been published previously only for sapropels S1 and S5 and show highly contrasting values. High δ98Mo (up to ∼2.3 ‰) values have been found for S5, reflecting strongly euxinic conditions with long deep water renewal times (Andersen et al., 2018Andersen, M.B., Matthews, A., Vance, D., Bar-Matthews, M., Archer, C., de Souza, G. (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.
). By contrast, low δ98Mo values (mostly <1 ‰) have been found for sapropel S1, indicating on average lower bottom water H2S concentrations (Reitz et al., 2007Reitz, A., Wille, M., Nägler, T.F., de Lange, G.J. (2007) Atypical Mo isotope signatures in eastern Mediterranean sediments. Chemical Geology 245, 1–8.
; 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.
; Matthews et al., 2017Matthews, A., Azrieli-Tal, I., Benkovitz, A., Bar-Matthews, M., Vance, D., Poulton, S.W., Teutsch, N., Almogi-Labin, A., Archer, C. (2017) Anoxic development of sapropel S1 in the Nile Fan inferred from redox sensitive proxies, Fe speciation, Fe and Mo isotopes. Chemical Geology 475, 24–39.
; Andersen et al., 2020Andersen, M.B., Matthews, A., Bar-Matthews, M., Vance, D. (2020) Rapid onset of ocean anoxia shown by high U and low Mo isotope compositions of sapropel S1. Geochemical Perspectives Letters 15, 10–14.
).The Eastern Mediterranean is punctuated by at least 6 further deoxygenation events during the last ∼300 kyr, but basin redox conditions have not been quantitatively constrained for these periods. Here, we provide semi-quantitative constraints on bottom water H2S concentrations for these recent short lived (<104 years) deoxygenation events in the Eastern Mediterranean Sea during deposition of sapropels S3 to S9, that reveal a range of different redox conditions (see Supplementary Information for a description of the methodology and results). High resolution Mo concentration (mm scale) and isotope (cm scale) data allow a detailed assessment of the temporal evolution of the dissolved Mo pool on kyr time scales, which is used to reconstruct the development of deep water redox conditions and residence times.
top
Discussion
Inter-Sapropel Comparisons. Low δ98Mo values and Mo concentrations for sapropels S3, S4, and S9 of core 64PE406-E1 (Fig. 1; Supplementary Information) are similar to those previously published values for sapropel S1, suggesting that these sediments were deposited under mildly euxinic conditions ([H2S] < 11 μM) (Fig. 2; Azrieli-Tall 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.
; Matthews et al., 2017Matthews, A., Azrieli-Tal, I., Benkovitz, A., Bar-Matthews, M., Vance, D., Poulton, S.W., Teutsch, N., Almogi-Labin, A., Archer, C. (2017) Anoxic development of sapropel S1 in the Nile Fan inferred from redox sensitive proxies, Fe speciation, Fe and Mo isotopes. Chemical Geology 475, 24–39.
; Andersen et al., 2020Andersen, M.B., Matthews, A., Bar-Matthews, M., Vance, D. (2020) Rapid onset of ocean anoxia shown by high U and low Mo isotope compositions of sapropel S1. Geochemical Perspectives Letters 15, 10–14.
). Alternative models for low δ98Mo values for these sediments, such as deposition associated with Fe or Mn (oxyhydr)oxides (e.g., Reitz et al., 2007Reitz, A., Wille, M., Nägler, T.F., de Lange, G.J. (2007) Atypical Mo isotope signatures in eastern Mediterranean sediments. Chemical Geology 245, 1–8.
) are not supported by the Mn, Mo, and U concentration data as these data do not show evidence for a strong Fe-Mn particulate shuttle (Supplementary Information; Figs. S-3 to S-5). High δ98Mo values and Mo concentrations for S5 and S7 can mostly be explained by smaller fractionations and more complete removal from seawater in more euxinic conditions. These same intervals are thought to have experienced the highest monsoon intensity over the North African continent in the last 300 kyr, suggesting a direct link between monsoon strength and degree of basin deoxygenation (Fig. 2; Ziegler et al., 2010Ziegler, M., Tuenter, E., Lourens, L.J. (2010) The precession phase of the boreal summer monsoon as viewed from the eastern Mediterranean (ODP Site 968). Quaternary Science Reviews 29, 1481–1490.
; Grant et al., 2016Grant, K., Grimm, R., Mikolajewicz, U., Marino, G., Ziegler, M., Rohling, E. (2016) The timing of Mediterranean sapropel deposition relative to insolation, sea-level and African monsoon changes. Quaternary Science Reviews 140, 125–141.
).Figure 1 Location of piston core 64PE406-E1 (this study), the LC21, and ODP 968 cores used for the age model, and ODP 967 studied by Andersen et al. (2018)
Andersen, M.B., Matthews, A., Vance, D., Bar-Matthews, M., Archer, C., de Souza, G. (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.
.Figure 2 δ98Mo and Mo concentration data for core 64PE406-E1. The bottom panel reflects indicators of monsoon-driven freshwater input. LC21 core ‘residuals’ (blue) show the δ18OG. ruber values up to 150 ka BP after correction for a seawater effect (Grant et al., 2016
Grant, K., Grimm, R., Mikolajewicz, U., Marino, G., Ziegler, M., Rohling, E. (2016) The timing of Mediterranean sapropel deposition relative to insolation, sea-level and African monsoon changes. Quaternary Science Reviews 140, 125–141.
). The black line shows the modelled North African run off (CLIMBER-2; Ziegler et al., 2010Ziegler, M., Tuenter, E., Lourens, L.J. (2010) The precession phase of the boreal summer monsoon as viewed from the eastern Mediterranean (ODP Site 968). Quaternary Science Reviews 29, 1481–1490.
). XRF core scanning data by Hennekam et al. (2020)Hennekam, R., van der Bolt, B., van Nes, E.H., de Lange, G.J., Scheffer, M., Reichart, G.-J. (2020) Early-warning signals for marine anoxic events. Geophysical Research Letters 47, e2020GL089183.
. The uncertainties (2 s.d.) on the Mo isotope data are smaller than the symbol sizes.Interpretative Framework for Mo Concentrations and Isotopes. Combining Black Sea surface sediment data and a mechanistic understanding of Mo isotope systematics derived from kinetic models (Matthews et al., 2017
Matthews, A., Azrieli-Tal, I., Benkovitz, A., Bar-Matthews, M., Vance, D., Poulton, S.W., Teutsch, N., Almogi-Labin, A., Archer, C. (2017) Anoxic development of sapropel S1 in the Nile Fan inferred from redox sensitive proxies, Fe speciation, Fe and Mo isotopes. Chemical Geology 475, 24–39.
) may facilitate a more detailed interpretation of sedimentary Mo signals. Surface sediment data from different Black Sea water depths highlights systematic patterns that reflect variations in dissolved Mo and H2S concentrations (Fig. 3a). A generally good agreement between Black Sea sediment data and models based on higher particle affinities for the more sulfidised thiomolybdate species (Matthews et al., 2017Matthews, A., Azrieli-Tal, I., Benkovitz, A., Bar-Matthews, M., Vance, D., Poulton, S.W., Teutsch, N., Almogi-Labin, A., Archer, C. (2017) Anoxic development of sapropel S1 in the Nile Fan inferred from redox sensitive proxies, Fe speciation, Fe and Mo isotopes. Chemical Geology 475, 24–39.
) implies that δ98Mo values may be used to constrain past bottom water H2S concentrations (Fig. 3b). This relationship between sedimentary δ98Mo values and [H2S] depends on (1) Mo speciation as a function of H2S, (2) fractionation factors between the different Mo species, (3) the isotopic composition of seawater, and (4) particle affinities of the different thiomolybdate species. Currently, the largest uncertainty in this approach is associated with particle affinities because these are not constrained by direct measurements. Particle affinities for different thiomolybdate species can be described by partition coefficients, i.e. Ki = [MoO4−iSi2−]particulate/[MoO4−iSi2−]dissolved, with i being the number of S atoms (Dahl et al., 2010Dahl, T.W., Anbar, A.D., Gordon, G.W., Rosing, M.T., Frei, R., Canfield, D.E. (2010) The behavior of molybdenum and its isotopes across the chemocline and in the sediments of sulfidic Lake Cadagno, Switzerland. Geochimica et Cosmochimica Acta 74, 144–163.
). Previously, much higher particle affinities for the more sulfidised species have been proposed (K0, K1, K2, K3, K4 = 0, 0, 1, 25, 100) but these values seem to overestimate the fractionation between seawater and sediment for intermediate H2S concentrations (Fig 3b; Dahl et al., 2010Dahl, T.W., Anbar, A.D., Gordon, G.W., Rosing, M.T., Frei, R., Canfield, D.E. (2010) The behavior of molybdenum and its isotopes across the chemocline and in the sediments of sulfidic Lake Cadagno, Switzerland. Geochimica et Cosmochimica Acta 74, 144–163.
). Tuning the particle affinities to the available Black Sea data (0.0001, 0.1, 5, 25, 25) improves the general fit, although the relative particle affinities remain poorly constrained. Therefore, here we only conservatively estimate [H2S], by using Mo isotopes to constrain upper or lower limits of past [H2S] that are valid for a range of different particle affinities (Fig. 3b, c).Figure 3 Mo isotope systematics. (a) Surface sediment (0–1 cm depth) data from different water depths (as data labels) in the modern Black Sea (Neubert et al., 2008
Neubert, N., Nägler, T.F., Böttcher, M.E. (2008) Sulfidity controls molybdenum isotope fractionation into euxinic sediments: Evidence from the modern Black Sea. Geology 36, 775–778.
; Brüske et al., 2020Brüske, A., Weyer, S., Zhao, M.-Y., Planavsky, N., Wegwerth, A., Neubert, N., Dellwig, O., Lau, K., Lyons, T. (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.
). (b) Sedimentary δ98Mo values as a function of aqueous H2S concentrations following calculations by Matthews et al. (2017)Matthews, A., Azrieli-Tal, I., Benkovitz, A., Bar-Matthews, M., Vance, D., Poulton, S.W., Teutsch, N., Almogi-Labin, A., Archer, C. (2017) Anoxic development of sapropel S1 in the Nile Fan inferred from redox sensitive proxies, Fe speciation, Fe and Mo isotopes. Chemical Geology 475, 24–39.
for different relative particle affinities and seawater δ98Mo. The 2.86 ‰ value is the average of deep Black Sea water compositions (Nägler et al., 2011Nägler, T., Neubert, N., Böttcher, M., Dellwig, O., Schnetger, B. (2011) Molybdenum isotope fractionation in pelagic euxinia: evidence from the modern Black and Baltic Seas. Chemical Geology 289, 1–11.
). δ98Moseawater can evolve to higher values in strongly restricted basins, such that sediments formed from an evolved water mass might show higher δ98Mo values. The grey circles represent constraints on [H2S] from δ98Mo that agree with all four models and are therefore used in Figure 3c to constrain [H2S] conservatively. Sediment data from Neubert et al. (2008)Neubert, N., Nägler, T.F., Böttcher, M.E. (2008) Sulfidity controls molybdenum isotope fractionation into euxinic sediments: Evidence from the modern Black Sea. Geology 36, 775–778.
. (c) Framework for the semi-quantitative reconstruction of [H2S].Semi-Quantitative Reconstructions of Sapropel Euxinia. We distinguish between four different levels of euxinia, based on the calculations (Matthews et al., 2017
Matthews, A., Azrieli-Tal, I., Benkovitz, A., Bar-Matthews, M., Vance, D., Poulton, S.W., Teutsch, N., Almogi-Labin, A., Archer, C. (2017) Anoxic development of sapropel S1 in the Nile Fan inferred from redox sensitive proxies, Fe speciation, Fe and Mo isotopes. Chemical Geology 475, 24–39.
) fitted to the Black Sea data (Fig. 3b) with additional constraints from the Mo concentration data (Fig. 3c). Low sedimentary δ98Mo values (< −0.1 ‰) are taken as evidence for mildly euxinic conditions (0 < [H2S] < 11 μmol/L). Conversely, for high δ98Mo values (>1.8 ‰), the H2S to δ98Mo relationship flattens such that δ98Mo values can only roughly constrain past H2S concentrations. These values (δ98Mo > 1.8 ‰) are accordingly taken to reflect dominantly highly sulfidic deep water conditions, with prevailing H2S concentrations of at least 17 μmol/L, but potentially much higher. δ98Mo values between −0.1 and 1.8 ‰ are less diagnostic by themselves, but can be interpreted considering the Mo concentrations of organic-rich sediments based on the limits for predominantly non-euxinic (<25 ppm), at least temporary euxinic (25–100 ppm), and mostly euxinic (>100 ppm) conditions (Scott and Lyons, 2012Scott, C., Lyons, T.W. (2012) Contrasting molybdenum cycling and isotopic properties in euxinic versus non-euxinic sediments and sedimentary rocks: Refining the paleoproxies. Chemical Geology 324, 19–27.
). Therefore, reasonably high δ98Mo values (>1 ‰) with concentrations of >25 ppm, and δ98Mo values of at least −0.1 ‰ coupled to high (>100 ppm) Mo concentrations, are both considered to have been deposited in ‘moderately euxinic’ conditions, where conditions were dominantly sulfidic but with H2S concentrations low enough to still allow fractionation between sediments and seawater.Based on this framework, during deposition of sapropels S5 and S7 more sulfidic water column conditions (H2S > 17 μmol/L) developed. This is in line with the study of Benkovitz et al. (2020)
Benkovitz, A., Matthews, A., Teutsch, N., Poulton, S.W., Bar-Matthews, M., Almogi-Labin, A. (2020) Tracing water column euxinia in Eastern Mediterranean Sapropels S5 and S7. Chemical Geology, 119627.
, who infer euxinic conditions for sapropels S5 and S7 based on Fe isotope and speciation data. By contrast, moderately euxinic bottom water conditions developed during sapropel S6, whereas during sapropels S3, S4, S8, and S9 mildly euxinic conditions were reached (0 < H2S < 11 μmol/L).Temporal Patterns in Mo Concentrations and Isotopes. Similar temporal patterns in [Mo] and δ98Mo are observed for different sapropels, particularly for sapropels S5, S6, S7, and S8 (Fig. 4). This general pattern is similar to that defined by Black Sea surface sediments from different water depths (Fig. 3a), which implies that the sapropel S5-S8 data reflect a progressive increase in [H2S] and evolution of the dissolved Mo pool. The relatively rapid changes in δ98Mo observed here (>2 ‰ in ∼2 kyr for S5) show that reconstructions of global redox changes on longer time scales from δ98Mo values in organic-rich sediments come with considerable uncertainty.
Figure 4 Close up of data presented in Figure 2 for sapropel intervals, including cross plots with age labels. The orange arrows highlight similarities in temporal patterns between S5 and S7. The bottom right plot shows a schematic interpretation of temporal patterns.
Mo Drawdown and Deep Water Renewal Times. The lower Mo concentrations and higher δ98Mo values found for the latter third of sapropels S5 and S7 is in line with progressive drawdown of dissolved Mo under highly sulfidic deep waters with long deep water renewal times, similar to conditions in the modern Black Sea (Nägler et al., 2011
Nägler, T., Neubert, N., Böttcher, M., Dellwig, O., Schnetger, B. (2011) Molybdenum isotope fractionation in pelagic euxinia: evidence from the modern Black and Baltic Seas. Chemical Geology 289, 1–11.
). Previous estimates of deep water residence times for sapropel S5 of 1030 +820/−520 years approach that of the modern Black Sea (Andersen et al., 2018Andersen, M.B., Matthews, A., Vance, D., Bar-Matthews, M., Archer, C., de Souza, G. (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.
). Similarities in δ98Mo and Mo concentrations for sapropel S5 and S7 imply that deep water renewal times for S7 were comparable to that of S5. These results highlight that conditions remained strongly reducing throughout the sapropel, including towards the end of sapropel deposition, despite gradually lower Mo concentrations. The other studied sapropels likely featured lower overall drawdown of dissolved Mo, implying shorter deep water residence times and/or rates of Mo drawdown.top
Conclusion
High resolution Mo isotope and concentration data from different Mediterranean sapropels reveal consistent temporal patterns. With the use of models with a kinetic control on the different thiomolybdate species and by comparison to Black Sea data, these patterns can be interpreted to reflect increasingly more sulfidic conditions during the early stages of sapropel formation. The differences between sapropels seem to be driven primarily by changes in monsoon strength over North Africa, and associated run off of freshwater into the Mediterranean Sea. As such, these data highlight an under-appreciated ability of Mo isotope data to quantitatively constrain basin redox changes over several kyr. The highly variable Mo isotope values on relatively short (103 years) time scales show that Mo isotopes should be used with great care when reconstructing global ocean redox conditions from black shales.
top
Acknowledgements
We thank the captain and crew of the R/V Pelagia cruise 64PE406, carried out under NESSC programme and financially supported by the Netherlands Ministry of OCW, grant number 024.002.001. We are thankful to B. van der Wagt, P. Laan, M. Klaver, and A. Dickson for analytical support. RH is supported through NWO grant ALW OP.2015.113 awarded to G.-J.R. Constructive comments by Tais Dahl and two anonymous reviewers have helped us to improve our manuscript.
Editor: Sophie Opfergelt
top
References
Andersen, M.B., Matthews, A., Vance, D., Bar-Matthews, M., Archer, C., de Souza, G. (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 contextHigh δ98Mo (up to ∼2.3 ‰) values have been found for S5, reflecting strongly euxinic conditions with long deep water renewal times (Andersen et al., 2018).
View in article
Location of piston core 64PE406-E1 (this study), the LC21, and ODP 968 cores used for the age model, and ODP 967 studied by Andersen et al. (2018).
View in article
Previous estimates of deep water residence times for sapropel S5 of 1030 +820/−520 years approach that of the modern Black Sea (Andersen et al., 2018).
View in article
Andersen, M.B., Matthews, A., Bar-Matthews, M., Vance, D. (2020) Rapid onset of ocean anoxia shown by high U and low Mo isotope compositions of sapropel S1. Geochemical Perspectives Letters 15, 10–14.
Show in contextThe Mo isotope composition of organic-rich sediments is generally lighter than seawater (2.34 ‰; Nakagawa et al., 2012), with reported δ98Mo values as low as ∼ −0.7 ‰ (e.g., Andersen et al., 2020; Brüske et al., 2020).
View in article
By contrast, low δ98Mo values (mostly <1 ‰) have been found for sapropel S1, indicating on average lower bottom water H2S concentrations (Reitz et al., 2007; Azrieli-Tal et al., 2014; Matthews et al., 2017; Andersen et al., 2020).
View in article
Low δ98Mo values and Mo concentrations for sapropels S3, S4, and S9 of core 64PE406-E1 (Fig. 1; Supplementary Information) are similar to those previously published values for sapropel S1, suggesting that these sediments were deposited under mildly euxinic conditions ([H2S] < 11 μM) (Fig. 2; Azrieli-Tall et al., 2014; Matthews et al., 2017; Andersen 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 contextBy contrast, low δ98Mo values (mostly <1 ‰) have been found for sapropel S1, indicating on average lower bottom water H2S concentrations (Reitz et al., 2007; Azrieli-Tal et al., 2014; Matthews et al., 2017; Andersen et al., 2020).
View in article
Low δ98Mo values and Mo concentrations for sapropels S3, S4, and S9 of core 64PE406-E1 (Fig. 1; Supplementary Information) are similar to those previously published values for sapropel S1, suggesting that these sediments were deposited under mildly euxinic conditions ([H2S] < 11 μM) (Fig. 2; Azrieli-Tall et al., 2014; Matthews et al., 2017; Andersen et al., 2020).
View in article
Benkovitz, A., Matthews, A., Teutsch, N., Poulton, S.W., Bar-Matthews, M., Almogi-Labin, A. (2020) Tracing water column euxinia in Eastern Mediterranean Sapropels S5 and S7. Chemical Geology, 119627.
Show in contextThis is in line with the study of Benkovitz et al. (2020), who infer euxinic conditions for sapropels S5 and S7 based on Fe isotope and speciation data.
View in article
Brüske, A., Weyer, S., Zhao, M.-Y., Planavsky, N., Wegwerth, A., Neubert, N., Dellwig, O., Lau, K., Lyons, T. (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 contextet al., 2012), with reported δ98Mo values as low as ∼ −0.7 ‰ (e.g., Andersen et al., 2020; Brüske et al., 2020).
View in article
Sedimentary Mo isotope values only approach the original seawater composition in strongly restricted and euxinic basins (e.g., Neubert et al., 2008; Brüske et al., 2020).
View in article
Mo isotope systematics. (a) Surface sediment (0–1 cm depth) data from different water depths (as data labels) in the modern Black Sea (Neubert et al., 2008; Brüske et al., 2020).
View in article
Dahl, T.W., Anbar, A.D., Gordon, G.W., Rosing, M.T., Frei, R., Canfield, D.E. (2010) The behavior of molybdenum and its isotopes across the chemocline and in the sediments of sulfidic Lake Cadagno, Switzerland. Geochimica et Cosmochimica Acta 74, 144–163.
Show in contextLarge isotopic differences between sediments and seawater for such sediments can be modelled by the kinetically controlled preferential scavenging of the more sulfidised thiomolybdate species (Dahl et al., 2010; Matthews et al., 2017).
View in article
Particle affinities for different thiomolybdate species can be described by partition coefficients, i.e. Ki = [MoO4−iSi2−]particulate/MoO4−iSi2−]dissolved, with i being the number of S atoms (Dahl et al., 2010).
View in article
Previously, much higher particle affinities for the more sulfidised species have been proposed (K0, K1, K2, K3, K4 = 0, 0, 1, 25, 100) but these values seem to overestimate the fractionation between seawater and sediment for intermediate H2S concentrations (Fig 3b; Dahl et al., 2010).
View in article
Dahl, T.W., Siggaard-Andersen, M.L., Schovsbo, N.H., Persson, D.O., Husted, S., Hougård, I.W., Dickson, A.J., Kjær, K., Nielsen, A.T. (2019) Brief oxygenation events in locally anoxic oceans during the Cambrian solves the animal breathing paradox. Scientific Reports 9, 1–9.
Show in contextAs the environmental boundary conditions during these events are relatively well established, sapropels present excellent analogues for intermittent anoxic and sulfidic conditions in past ocean basins (e.g., Dahl et al., 2019) as well as targets for modelling efforts aiming to address large scale basin deoxygenation, provided that redox conditions are quantitatively constrained.
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 contextHowever, in the presence of H2S, conversion of molybdate to particle-reactive thiomolybdate species (MoO4−xSx2−) results in the enhanced export and burial of Mo (Erickson and Helz, 2000).
View in article
Grant, K., Grimm, R., Mikolajewicz, U., Marino, G., Ziegler, M., Rohling, E. (2016) The timing of Mediterranean sapropel deposition relative to insolation, sea-level and African monsoon changes. Quaternary Science Reviews 140, 125–141.
Show in contextThe timing of Mediterranean sapropel deposition is controlled by maxima in Northern hemisphere summer insolation during precession minima, which lead to higher monsoon intensity over the North African continent (e.g., Ziegler et al., 2010; Grant et al., 2016).
View in article
LC21 core ‘residuals’ (blue) show the δ18OG. ruber values up to 150 ka BP after correction for a seawater effect (Grant et al., 2016).
View in article
These same intervals are thought to have experienced the highest monsoon intensity over the North African continent in the last 300 kyr, suggesting a direct link between monsoon strength and degree of basin deoxygenation (Fig. 2; Ziegler et al., 2010; Grant et al., 2016).
View in article
Hennekam, R., van der Bolt, B., van Nes, E.H., de Lange, G.J., Scheffer, M., Reichart, G.-J. (2020) Early-warning signals for marine anoxic events. Geophysical Research Letters 47, e2020GL089183.
Show in contextThe black line shows the modelled North African run off (CLIMBER-2; Ziegler et al., 2010). XRF core scanning data by Hennekam et al. (2020).
View in article
Kendall, B., Dahl, T., Anbar, A.D. (2017) The stable isotope geochemistry of molybdenum. Reviews in Mineralogy and Geochemistry 82, 683–732.
Show in contextMolybdenum (Mo) concentrations of marine sediments are well established as a tracer for local marine redox conditions, while its isotopes have more recently emerged as a valuable proxy to estimate redox conditions on local to global spatial scales (e.g., Scott and Lyons, 2012; Kendall et al., 2017 and references therein).
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 chromatographic separation and detection by multicollector ICPMS. Analytical Chemistry 89, 3123–3129.
Show in contextLow δ98Mo values in mildly euxinic environments can be attributed to isotope fractionation associated with the incomplete conversion of molybdate to thiomolybdate species, with the more sulfidised species becoming progressively isotopically lighter (Tossell, 2005; Kerl et al., 2017).
View in article
Matthews, A., Azrieli-Tal, I., Benkovitz, A., Bar-Matthews, M., Vance, D., Poulton, S.W., Teutsch, N., Almogi-Labin, A., Archer, C. (2017) Anoxic development of sapropel S1 in the Nile Fan inferred from redox sensitive proxies, Fe speciation, Fe and Mo isotopes. Chemical Geology 475, 24–39.
Show in contextLarge isotopic differences between sediments and seawater for such sediments can be modelled by the kinetically controlled preferential scavenging of the more sulfidised thiomolybdate species (Dahl et al., 2010; Matthews et al., 2017).
View in article
Low δ98Mo values and Mo concentrations for sapropels S3, S4, and S9 of core 64PE406-E1 (Fig. 1; Supplementary Information) are similar to those previously published values for sapropel S1, suggesting that these sediments were deposited under mildly euxinic conditions ([H2S] < 11 μM) (Fig. 2; Azrieli-Tall et al., 2014; Matthews et al., 2017; Andersen et al., 2020).
View in article
By contrast, low δ98Mo values (mostly <1 ‰) have been found for sapropel S1, indicating on average lower bottom water H2S concentrations (Reitz et al., 2007; Azrieli-Tal et al., 2014; Matthews et al., 2017; Andersen et al., 2020).
View in article
Combining Black Sea surface sediment data and a mechanistic understanding of Mo isotope systematics derived from kinetic models (Matthews et al., 2017) may facilitate a more detailed interpretation of sedimentary Mo signals.
View in article
A generally good agreement between Black Sea sediment data and models based on higher particle affinities for the more sulfidised thiomolybdate species (Matthews et al., 2017) implies that δ98Mo values may be used to constrain past bottom water H2S concentrations (Fig. 3b).
View in article
(b) Sedimentary δ98Mo values as a function of aqueous H2S concentrations following calculations by Matthews et al. (2017) for different relative particle affinities and seawater δ98Mo.
View in article
We distinguish between four different levels of euxinia, based on the calculations (Matthews et al., 2017) fitted to the Black Sea data (Fig. 3b) with additional constraints from the Mo concentration data (Fig. 3c).
View in article
Nägler, T., Neubert, N., Böttcher, M., Dellwig, O., Schnetger, B. (2011) Molybdenum isotope fractionation in pelagic euxinia: evidence from the modern Black and Baltic Seas. Chemical Geology 289, 1–11.
Show in contextThe 2.86 ‰ value is the average of deep Black Sea water compositions (Nägler et al., 2011).
View in article
The lower Mo concentrations and higher δ98Mo values found for the latter third of sapropels S5 and S7 is in line with progressive drawdown of dissolved Mo under highly sulfidic deep waters with long deep water renewal times, similar to conditions in the modern Black Sea (Nägler et al., 2011).
View in article
Nakagawa, Y., Takano, S., Firdaus, M.L., Norisuye, K., Hirata, T., Vance, D., Sohrin, Y. (2012) The molybdenum isotopic composition of the modern ocean. Geochemical Journal 46, 131–141.
Show in contextThe Mo isotope composition of organic-rich sediments is generally lighter than seawater (2.34 ‰; Nakagawa et al., 2012), with reported δ98Mo values as low as ∼ −0.7 ‰ (e.g., Andersen et al., 2020; Brüske et al., 2020).
View in article
Neubert, N., Nägler, T.F., Böttcher, M.E. (2008) Sulfidity controls molybdenum isotope fractionation into euxinic sediments: Evidence from the modern Black Sea. Geology 36, 775–778.
Show in contextSedimentary Mo isotope values only approach the original seawater composition in strongly restricted and euxinic basins (e.g., Neubert et al., 2008; Brüske et al., 2020).
View in article
Mo isotope systematics. (a) Surface sediment (0–1 cm depth) data from different water depths (as data labels) in the modern Black Sea (Neubert et al., 2008; Brüske et al., 2020).
View in article
Sediment data from Neubert et al. (2008). (c) Framework for the semi-quantitative reconstruction of [H2S]
View in article
Reitz, A., Wille, M., Nägler, T.F., de Lange, G.J. (2007) Atypical Mo isotope signatures in eastern Mediterranean sediments. Chemical Geology 245, 1–8.
Show in contextBy contrast, low δ98Mo values (mostly <1 ‰) have been found for sapropel S1, indicating on average lower bottom water H2S concentrations (Reitz et al., 2007; Azrieli-Tal et al., 2014; Matthews et al., 2017; Andersen et al., 2020).
View in article
Alternative models for low δ98Mo values for these sediments, such as deposition associated with Fe or Mn (oxyhydr)oxides (e.g., Reitz et al., 2007) are not supported by the Mn, Mo, and U concentration data as these data do not show evidence for a strong Fe-Mn particulate shuttle (Supplementary Information; Figs. S-3 to S-5).
View in article
Rohling, E., Marino, G., Grant, K. (2015) Mediterranean climate and oceanography, and the periodic development of anoxic events (sapropels). Earth-Science Reviews 143, 62–97.
Show in contextGreater freshwater input resulted in nutrient-rich conditions, enhanced water column stratification, and a decrease in local deep water formation, which ultimately caused the drawdown of dissolved oxygen at depth and the deposition of sapropels (e.g., Rohling et al., 2015).
View in article
Scott, C., Lyons, T.W. (2012) Contrasting molybdenum cycling and isotopic properties in euxinic versus non-euxinic sediments and sedimentary rocks: Refining the paleoproxies. Chemical Geology 324, 19–27.
Show in contextMolybdenum (Mo) concentrations of marine sediments are well established as a tracer for local marine redox conditions, while its isotopes have more recently emerged as a valuable proxy to estimate redox conditions on local to global spatial scales (e.g., Scott and Lyons, 2012; Kendall et al., 2017 and references therein).
View in article
High sedimentary Mo enrichments can thus be taken to reflect sulfidic water column conditions (e.g., Scott and Lyons, 2012).
View in article
δ98Mo values between −0.1 and 1.8 ‰ are less diagnostic by themselves, but can be interpreted considering the Mo concentrations of organic-rich sediments based on the limits for predominantly non-euxinic (<25 ppm), at least temporary euxinic (25–100 ppm), and mostly euxinic (>100 ppm) conditions (Scott and Lyons, 2012).
View in article
Tossell, J. (2005) Calculating the partitioning of the isotopes of Mo between oxidic and sulfidic species in aqueous solution. Geochimica et Cosmochimica Acta 69, 2981–2993.
Show in contextLow δ98Mo values in mildly euxinic environments can be attributed to isotope fractionation associated with the incomplete conversion of molybdate to thiomolybdate species, with the more sulfidised species becoming progressively isotopically lighter (Tossell, 2005; Kerl et al., 2017).
View in article
Ziegler, M., Tuenter, E., Lourens, L.J. (2010) The precession phase of the boreal summer monsoon as viewed from the eastern Mediterranean (ODP Site 968). Quaternary Science Reviews 29, 1481–1490.
Show in contextThe timing of Mediterranean sapropel deposition is controlled by maxima in Northern hemisphere summer insolation during precession minima, which lead to higher monsoon intensity over the North African continent (e.g., Ziegler et al., 2010; Grant et al., 2016).
View in article
These same intervals are thought to have experienced the highest monsoon intensity over the North African continent in the last 300 kyr, suggesting a direct link between monsoon strength and degree of basin deoxygenation (Fig. 2; Ziegler et al., 2010; Grant et al., 2016).
View in article
The black line shows the modelled North African run off (CLIMBER-2; Ziegler et al., 2010). XRF core scanning data by Hennekam et al. (2020).
View in article
top
Supplementary Information
The Supplementary Information includes:
- Material and Methods
- Results
- Comparison with Other Sapropel S5 Data
- Alternative Models for Low δ98Mo Values
- Supplementary Information References
Download Data Table (Excel).
Download the Supplementary Information (PDF).
Figures
Figure 1 Location of piston core 64PE406-E1 (this study), the LC21, and ODP 968 cores used for the age model, and ODP 967 studied by Andersen et al. (2018)
Andersen, M.B., Matthews, A., Vance, D., Bar-Matthews, M., Archer, C., de Souza, G. (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.
.
Figure 2 δ98Mo and Mo concentration data for core 64PE406-E1. The bottom panel reflects indicators of monsoon-driven freshwater input. LC21 core ‘residuals’ (blue) show the δ18OG. ruber values up to 150 ka BP after correction for a seawater effect (Grant et al., 2016
Grant, K., Grimm, R., Mikolajewicz, U., Marino, G., Ziegler, M., Rohling, E. (2016) The timing of Mediterranean sapropel deposition relative to insolation, sea-level and African monsoon changes. Quaternary Science Reviews 140, 125–141.
). The black line shows the modelled North African run off (CLIMBER-2; Ziegler et al., 2010Ziegler, M., Tuenter, E., Lourens, L.J. (2010) The precession phase of the boreal summer monsoon as viewed from the eastern Mediterranean (ODP Site 968). Quaternary Science Reviews 29, 1481–1490.
). XRF core scanning data by Hennekam et al. (2020)Hennekam, R., van der Bolt, B., van Nes, E.H., de Lange, G.J., Scheffer, M., Reichart, G.-J. (2020) Early-warning signals for marine anoxic events. Geophysical Research Letters 47, e2020GL089183.
. The uncertainties (2 s.d.) on the Mo isotope data are smaller than the symbol sizes.
Figure 3 Mo isotope systematics. (a) Surface sediment (0–1 cm depth) data from different water depths (as data labels) in the modern Black Sea (Neubert et al., 2008
Neubert, N., Nägler, T.F., Böttcher, M.E. (2008) Sulfidity controls molybdenum isotope fractionation into euxinic sediments: Evidence from the modern Black Sea. Geology 36, 775–778.
; Brüske et al., 2020Brüske, A., Weyer, S., Zhao, M.-Y., Planavsky, N., Wegwerth, A., Neubert, N., Dellwig, O., Lau, K., Lyons, T. (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.
). (b) Sedimentary δ98Mo values as a function of aqueous H2S concentrations following calculations by Matthews et al. (2017)Matthews, A., Azrieli-Tal, I., Benkovitz, A., Bar-Matthews, M., Vance, D., Poulton, S.W., Teutsch, N., Almogi-Labin, A., Archer, C. (2017) Anoxic development of sapropel S1 in the Nile Fan inferred from redox sensitive proxies, Fe speciation, Fe and Mo isotopes. Chemical Geology 475, 24–39.
for different relative particle affinities and seawater δ98Mo. The 2.86 ‰ value is the average of deep Black Sea water compositions (Nägler et al., 2011Nägler, T., Neubert, N., Böttcher, M., Dellwig, O., Schnetger, B. (2011) Molybdenum isotope fractionation in pelagic euxinia: evidence from the modern Black and Baltic Seas. Chemical Geology 289, 1–11.
). δ98Moseawater can evolve to higher values in strongly restricted basins, such that sediments formed from an evolved water mass might show higher δ98Mo values. The grey circles represent constraints on [H2S] from δ98Mo that agree with all four models and are therefore used in Figure 3c to constrain [H2S] conservatively. Sediment data from Neubert et al. (2008)Neubert, N., Nägler, T.F., Böttcher, M.E. (2008) Sulfidity controls molybdenum isotope fractionation into euxinic sediments: Evidence from the modern Black Sea. Geology 36, 775–778.
. (c) Framework for the semi-quantitative reconstruction of [H2S].
Figure 4 Close up of data presented in Figure 2 for sapropel intervals, including cross plots with age labels. The orange arrows highlight similarities in temporal patterns between S5 and S7. The bottom right plot shows a schematic interpretation of temporal patterns.