A Late Archean Sulfidic Sea Stimulated by Early Oxidative Weathering of the Continents

See allHide authors and affiliations

Science  30 Oct 2009:
Vol. 326, Issue 5953, pp. 713-716
DOI: 10.1126/science.1176711


Iron speciation data for the late Archean Mount McRae Shale provide evidence for a euxinic (anoxic and sulfidic) water column 2.5 billion years ago. Sulfur isotope data compiled from the same stratigraphic section suggest that euxinic conditions were stimulated by an increase in oceanic sulfate concentrations resulting from weathering of continental sulfide minerals exposed to an atmosphere with trace amounts of photosynthetically produced oxygen. Variability in local organic matter flux likely confined euxinic conditions to midportions of the water column on the basin margin. These findings indicate that euxinic conditions may have been common on a variety of spatial and temporal scales both before and immediately after the Paleoproterozoic rise in atmospheric oxygen, hinting at previously unexplored texture and variability in deep ocean chemistry during Earth’s early history.

The first two billion years of Earth’s history were characterized by little to no free atmospheric oxygen (1, 2). A large body of evidence points to a sharp rise in the concentration of atmospheric O2 during the Paleoproterozoic between 2.45 and 2.32 billion years ago (Ga) (13), but the history of deep ocean oxygenation is less well-known. The deposition of banded iron formations (BIF) during the Archean and early Proterozoic (~3.8 to 1.8 Ga) has been taken to imply that deep ocean water masses were anoxic and rich in dissolved ferrous iron (Fe2+) derived from high-temperature weathering of seafloor basalt under low oceanic sulfate (SO42–) concentrations (4, 5). Reducing and iron-rich (ferruginous) deep ocean conditions are thought to have persisted for most of Earth’s early history, although a relative paucity of BIF between 2.4 and 2.0 Ga (6) has rendered deep ocean chemistry during this period obscure. In any case, the cessation of BIF deposition at ~1.8 Ga is generally linked to the accumulation of oxygen in the atmosphere through the eventual removal of Fe2+ from the ocean either as ferric (hydr)oxides (7) or as pyrite in euxinic basins (8). A corollary of the latter model is that oxidative delivery of sulfate to the ocean was not sufficient to remove reactive iron through microbial sulfide production before ~1.8 Ga. However, recent studies of the late Archean Mount McRae Shale suggest that oxidative sulfur cycling may have preceded the Paleoproterozoic rise in atmospheric oxygen (9) and that conditions sufficient to authigenically enrich molybdenum (Mo) in marine sediments existed at ~2.5 Ga (10). On the modern Earth, substantial enrichment of Mo into sediments occurs after the conversion of soluble molybdate (MoO42–) to particle-reactive thiomolybdates (MoO4-xSx2–) in stable sulfidic environments (11), indicating that the Mo enrichments seen in the Mount McRae Shale may have resulted from the development of a euxinic water column in association with increased oxidative transport of crustal sulfur as SO42–.

To examine the possibility of euxinia during the late Archean, we analyzed iron mineral speciation in the Mount McRae Shale (12). The distribution of iron among different biogeochemically labile mineral phases (“highly reactive iron”) can reveal local redox conditions (13, 14). Highly reactive iron (FeHR) is defined as the sum of pyrite iron (FePY) and iron in phases that are reactive to hydrogen sulfide (H2S) on short diagenetic time scales, such as ferric oxides (Feox), magnetite (Femag), and iron present as carbonate (Fecarb). In modern sediments from oxic continental margins and the deep sea, FeHR makes up 6 to 38% of the total sedimentary iron (FeT) (i.e., FeHR/FeT = 0.06 to 0.38); an average FeHR/FeT ratio of 0.26 ± 0.08 defines the modern siliciclastic baseline (13). Values for FeHR/FeT that are elevated above this siliciclastic background suggest reactive iron input that is decoupled from detrital sources, an indication of iron transport and scavenging within an anoxic water column (15). We also look toward total iron enrichments (expressed as FeT/Al ratios) as an indicator of water column anoxia (16, 17).

If FeHR/FeT and FeT/Al data provide evidence for anoxia, the ratio FePY/FeHR can be used to distinguish between anoxic but nonsulfidic conditions and anoxic water columns containing free H2S (euxinic). This approach is based on the simple premise that under anoxic conditions dissolved Fe2+ and dissolved H2S cannot coexist in abundance in solution because of the insolubility of iron sulfide phases, and therefore high values for FePY/FeHR indicate H2S-dominated water column chemistry. For confirmation, we also measured degree of pyritization (DOP) as a conservative indicator of iron-limited pyrite formation and euxinia (12, 17). The distribution of highly reactive Fe species in the Mount McRae Shale is shown in Fig. 1, along with FeT/Al, bulk molybdenum (Mo), and organic carbon (TOC) concentrations from (10). We focus here on the pyritic and organic-rich lower shale interval (LSI) and upper shale interval (USI). Ferric oxides make up a small proportion of FeHR for the entire sequence analyzed here, indicating water column and/or pore fluid conditions that were reducing with respect to iron (Fig. 1). Values for FeHR/FeT and FeT/Al are elevated throughout, suggesting that the entire sequence was deposited beneath an anoxic water column. In a few instances, FePY/FeHR values in the LSI approach a threshold (FePY/FeHR ≥ 0.8) interpreted to reflect euxinia when paired with evidence for anoxic deposition (14, 18); however, the average FePY/FeHR for this unit (0.55 ± 0.20) suggests a predominance of ferruginous conditions. Variations in FeHR within the LSI are governed by differences in Fecarb rather than FePY (fig. S2). These data are consistent with sulfate reduction and pyrite formation within or beneath an anoxic water column, but with reactive Fe in excess of dissolved H2S such that H2S did not persist in the pore fluids or water column during LSI deposition.

Fig. 1

(A to G) Stratigraphic profiles for iron speciation data from the ADBP-9 core. Squares, diamonds, and circles represent the LSI, siderite-facies BIF, and USI, respectively. The striped box in (A) represents the range of FeHR/FeT values seen in modern oxic continental margin and deep-sea sediments (13). The dotted line in (A) represents the mean FeHR/FeT value (0.26) for normal (oxic) marine settings (13). The striped box in (B) represents FePY/FeHR values that are above 0.8. Euxinia is implied when both of these thresholds are exceeded and FeT/Al values exceed 0.5. The dark line in (B) is the best fit through DOP values (not shown). The two dotted lines in (E) reflect average bulk Mo enrichments for the Archean [3 parts per million (ppm)] and Proterozoic (18 ppm) (21). Data for (E) and (F) from (10).

The USI shows pronounced enrichment in FeHR, indicating extensive reactive Fe scavenging beneath an anoxic water column (Fig. 1). Values for FeT/Al, although lower than those seen in the underlying siderite-facies, remain elevated. In contrast to the LSI, FePY/FeHR values are persistently high (0.85 ± 0.17), as is DOP (0.78 ± 0.23). A strong linear relation between FeHR and FePY for the USI (fig. S2) demonstrates that variations in the amount of FeHR are governed by differences in FePY content and that FeHR is all but completely pyritized. This combination of parameters (elevated values for FeHR/FeT, FeT/Al, FePY/FeHR, and DOP) indicates that the water column was euxinic for a substantial portion of USI deposition.

To examine whether euxinia occurred in association with a transient or secular change in the oxidative transport of MoO42– and SO42– (19), we turn to the sulfur isotope composition of syn-genetic and early diagenetic pyrite from deep-water facies (shales and BIF) of the Neoarchean-Paleoproterozoic (2.7 to 2.45 Ga) Hamersley Basin (Fig. 2). Neoarchean samples below the USI, including those from the LSI and the siderite-facies BIF directly beneath the USI, show large Δ33S values and positive covariation between Δ33S and δ34S (Fig. 2). This pattern has been hypothesized to reflect a primary atmospheric array in the isotopic composition of elemental sulfur aerosols (20). The corresponding linearity and large positive ∆33S anomalies of these data suggest a tight isotopic coupling between atmospherically derived reduced sulfur species and sedimentary pyrite formation and also indicate that the transfer and mixing mechanisms that contributed to the signal ultimately preserved in the sediments were similar on at least a basinal scale and through large periods of Archean time (9, 20).

Fig. 2

Sulfur isotope data for deep-water Hamersley Basin pyrite samples spanning 2.7 to 2.45 Ga, displayed as δ34S versus δ33S (A) and δ34S versus ∆33S (B). Pre-USI data are from the Jeerinah Formation and lower Mount McRae Shale (20, 27); the LSI and siderite-facies BIF beneath the USI (9); and the Marra Mamba BIF (27), which was deposited between the Jeerinah Formation and the Mount McRae Shale. USI/Brockman BIF (BrIF) data are from the USI (9) and the overlying BrIF (27, 28). The line labeled “MDF” in (A) is the mass-dependent fractionation line, defined as δ33S = 0.515 × δ34S (29). The gray box in (B) represents the range of ∆33S values attainable by mass-dependent processes (30, 31).

The sulfur isotope composition of pyrite in the USI and the overlying Brockman BIF shows a different distribution (Fig. 2). Values for Δ33S are attenuated during euxinic deposition, with the largest positive Δ33S values in the USI found in intervals that are transitional with the siderite-facies BIF unit below or the overlying carbonate unit. The linear array that characterizes the data before deposition of the USI is no longer evident, and a linear regression through the USI/Brockman data is closely aligned with the mass-dependent fractionation array in δ34S-δ33S space. This shift is accompanied by predominantly small negative Δ33S values and relatively depleted δ34S values within the USI, followed by subdued variability in Δ33S and a wide spread in δ34S values [from –5 per mil (‰) to +35‰] in the overlying Brockman BIF. We interpret this isotopic shift to reflect increased SO42– availability during deposition of the USI and Brockman BIF accompanied by mixing of photolytically produced sulfur and isotopically normal crustal sulfur oxidatively mobilized under an atmosphere that remained O2-poor (12). A transient or secular increase in the oxidative transport of MoO42– and SO42– during USI deposition is also supported by the contrasting strong non–mass-dependent (NMD) signal (20) and essential lack of Mo enrichment (21) preserved in pyritic shales of the Jeerinah Formation underlying the Mount McRae—analogous to the signals seen in the LSI and the siderite-facies BIF beneath the USI. The persistence of distinct NMD anomalies, despite the overall shift in isotopic arrays, requires the formation and burial of sulfur with NMD isotope composition throughout this period. Ground-level atmospheric O2 concentrations of less than 2 parts per million by volume (ppmv) (i.e., below 10−5 the present atmospheric level) are therefore implied (22), and concentrations throughout most of the troposphere may have been substantially lower than this (22, 23). This assertion is also supported by Δ33S/Δ36S relationships (9).

Combined, the high-resolution Fe speciation, Mo enrichment, and sulfur isotope data for the Mount McRae Shale indicate the development of euxinia during deposition of the USI and that these conditions were contemporaneous with a change in sedimentary sulfur isotope systematics. However, the stratigraphic position of the USI between two BIFs, coupled with FeT/Al ratios that are persistently and substantially elevated above crustal values (Fig. 1), suggest that hydrothermal iron fluxes to the deep basin were important at this time. Our interpretation therefore implies a water column structure that would allow for both the accumulation of dissolved H2S and the subsequent or coeval deposition of voluminous BIF.

To reconcile these observations, we postulate locally enhanced microbial H2S production, stimulated by organic matter (OM) delivery and facilitated by an increased flux of dissolved SO42– to the basin. Local loading of OM would have fueled vigorous sulfate reduction along the basin margin, resulting in an oxidant minimum zone in which dissolved H2S accumulated and quantitatively removed dissolved Fe2+ from the water column (Fig. 3). Euxinia would have expanded or contracted periodically as a function of the balance between reactive Fe input and OM flux, with the possibility of dissolved H2S transiently accumulating on a basin scale or receding beneath the sediment-water interface. This lateral redox structure is similar to the basin-scale lithofacies framework hypothesized for contemporaneous strata from the South African Transvaal basin (24), indicating that such conditions may have been common during this period.

Fig. 3

Schematic representation of the Hamersley Basin during the deposition of the upper Mount McRae Shale (USI). Oxidative delivery of SO42– and MoO42–, combined with a high local organic matter flux, resulted in the accumulation of free H2S in the water column in excess of dissolved Fe2+ (euxinia), supporting authigenic Mo enrichment. Atmospheric O2 concentrations below 2 ppmv could have driven the enhanced oxidative weathering recorded in the USI but would still have allowed for SO2 photolysis and the preservation of NMD sulfur isotope anomalies (12). Atmospheric photochemistry simplified from (32).

Although OM delivery was the proximate cause of euxinia, we propose that it was the increased availability of SO42– attendant to oxidative weathering that ultimately allowed microbial H2S production to overwhelm reactive Fe, at least locally, during USI deposition. Elevated total sulfur concentrations in this interval, coincident with increased TOC and high FeT values (9, 10), also point to increasing availability of water column SO42– such that microbial sulfate reduction was able to keep pace with substantial OM flux and relatively high reactive Fe availability. It is possible that mid-water column euxinia existed subsequent to USI deposition, with the stratigraphic transition to Brockman BIF recording a change in water depth rather than a temporal change in basin chemistry.

Our findings suggest that weak oxidative forcing could have stimulated the development of euxinia 50 to 100 million years before the Paleoproterozoic rise in atmospheric oxygen and that stable and persistent euxinia could have developed at least locally, and perhaps on a much larger scale, even within BIF-forming basins. Sulfur isotope data indicate that the weathering flux of SO42– to the ocean increased substantially after the rise in atmospheric oxygen between 2.45 and 2.32 Ga (8, 25). The lack of BIF between 2.4 and 2.0 Ga may therefore reflect the frequent or sustained development of euxinia within Paleoproterozoic basins (1), presaging the possibly widespread and protracted development of similar oceanographic conditions hypothesized previously for the Mesoproterozoic (~1.8 to 1.0 Ga) (8). Constraints on deep ocean redox during this intervening period are sparse, but existing data intimate that euxinic deep basins were much more common than ferruginous ones between 2.4 and 2.0 Ga (21).

More generally, we argue that deposition of BIFs represented episodic pulses of reducing power from Earth’s interior (6) rather than persistent deep-water conditions. Significant spatial variability in water column chemistry is indicated for intervals of BIF deposition, with intervening periods throughout the Archean and Paleoproterozoic during which at least portions of the water column may have been euxinic. Vacillation between euxinic and ferruginous conditions would have favored the early evolution and ecological expansion of a variety of anoxygenic photosynthetic metabolisms in pelagic environments. Expressions of biological oxygen production (such as those seen in the upper Mount McRae and Brockman BIF) would then have varied with the extent to which episodic or sustained pulses of reductants from the Earth’s interior would have buffered photosynthetic oxygen, contributing to the protracted nature of Earth surface oxygenation during the Archean and Proterozoic (26).

Supporting Online Material

Materials and Methods

Figs. S1 to S3

Table S1


References and Notes

  1. Materials and methods are available as supporting material on Science Online.
  2. Interpretation of this measurement follows the same rationale as that for FeHR/FeT (i.e., enrichments above the average FeT/Al ratio for continental crust of ~0.5 imply transport and scavenging of iron under anoxic conditions), but FeT/Al is immune to concerns regarding authigenic iron-silicate formation or metamorphic repartitioning of reactive iron phases into poorly reactive silicate mineralogies.
  3. Because Mo enrichments require both an oceanic Mo reservoir and the accumulation of free H2S, it is possible that the metal enrichments recorded in the USI point only to the development of euxinia rather than a temporally constrained increase in the flux of MoO42– and SO42– to the Hamersley Basin during USI deposition.
  4. The NASA Astrobiology Institute and Exobiology Program and the NSF Geobiology and Low Temperature Geochemistry Program provided financial support. The authors thank B. Gill, S. Severmann, N. Planavsky, M. Claire, J. Kaufman, and R. Buick for helpful discussions, and G. Arnold for handling of core material.
View Abstract

Navigate This Article