Report

Late-Neoproterozoic Deep-Ocean Oxygenation and the Rise of Animal Life

See allHide authors and affiliations

Science  05 Jan 2007:
Vol. 315, Issue 5808, pp. 92-95
DOI: 10.1126/science.1135013

Abstract

Because animals require oxygen, an increase in late-Neoproterozoic oxygen concentrations has been suggested as a stimulus for their evolution. The iron content of deep-sea sediments shows that the deep ocean was anoxic and ferruginous before and during the Gaskiers glaciation 580 million years ago and that it became oxic afterward. The first known members of the Ediacara biota arose shortly after the Gaskiers glaciation, suggesting a causal link between their evolution and this oxygenation event. A prolonged stable oxic environment may have permitted the emergence of bilateral motile animals some 25 million years later.

Large, architecturally complex life forms first appeared about 575 million years ago (Ma) (1, 2). These lifeforms are epitomized by the Ediacara biota, a globally distributed assemblage of fossil impressions of centimeter- to meter-scale soft-bodied organisms and colonies. The Ediacara biota developed after the Gaskiers glaciation at 580 Ma and abruptly disappeared coincident with the Cambrian explosion of skeletal animals about 35 million years later (3, 4). The Ediacara biota most likely included a mixture of stem-group animals and potentially other extant kingdoms of eukaryotes, along with fossils that may represent “failed experiments” in evolution (3). Fossilized animal embryos (5) dated at 560 to 580 Ma (6) further support the view that animals were an important part of Ediacaran life. Animals have an absolute requirement for oxygen, and it has been widely hypothesized that a late-Neoproterozoic rise in oxygen created an environment permissive for animal evolution (710). Direct evidence for late-Neoproterozoic ocean oxygenation, however, has been lacking. We used several geochemical tracers to show that the deep ocean became oxidized shortly before the first appearance of the Ediacara biota.

In the Avalon Peninsula, Newfoundland, 6 km of late-Neoproterozoic sedimentary rocks represent at least 15 million years of late-Neoproterozoic time (Fig. 1). The sediment sequence begins about 800 m below the glacial deposits and cap carbonate of the Gaskiers Formation dated at 580 Ma (2). The Gaskiers glaciation was the last of the major Neoproterozoic glaciations (4, 11) and is represented on four continents (12), although it was probably not as widespread as the previous Sturtian and Marinoan “snowball” glaciations (11, 12). Ediacara-type fossils are present through most of the post-Gaskiers succession. The first Ediacaran fossils, represented by a low-diversity assemblage of Ivesheadia (“pizza disks”), Thectardis, and species of Charnia (1, 3, 4, 13) (Fig. 2, A and B), were found in the upper Drook Formation within 5 million years of the glacial terminus. These are the earliest large and architecturally complex eukaryote fossils known anywhere in the world. By 565 Ma, the high-density and high-diversity Ediacaran assemblages of Mistaken Point existed (Fig. 2C) (3, 13, 14), and numerous occurrences of Ediacaran disks continue through the Ferme use Formation (Fig. 2D).

Fig. 1.

Location (A and B) and stratigraphic setting (C) of geochemical samples. Location numbers are the same as in (19) (sites 1, 2, and 4 to 17) and (39) (sites 1 to 8, 13, and 14), which contain details on the precise location, access, lithostratigraphy, sedimentology, and fossil taxa for each locality. The supporting online material includes stratigraphic positioning of the samples (26). U/Pb dates are from (2, 40).

Fig. 2.

Ediacaran fossils from Newfoundland. Scale bars represent 5 cm. (A and B) Low-diversity assemblage of fossils from the Drook Formation (575 Ma) at locality 4. (A) Two specimens of the discoid fossil Ivesheadia. (B) Charnia frond. (C) High-diversity Mistaken Point assemblage (565 Ma), exhibiting large fronds (Charniodiscus) along with frondose, bush like, and spindle-shaped rangeomorph fossils at locality 7. (D) Fermeuse assemblage (about 560 Ma) of Aspidella disks at locality 13.

The Ediacaran animals of the Avalon Peninsula lived on the sea floor in deep-water environments well below the photic zone (3, 13, 15). The Mall Bay through Briscal Formations (Fig. 1) accumulated as axial basin floor deposits in water depths exceeding several hundred meters, and possibly extending 1 km or more (3, 16). The Mistaken Point and Trepassey Formations represent a deep-water slope environment (14, 15). Beginning with the Fermeuse Formation, the sequence gives way to mudstones likely deposited as delta front deposits (3, 17). The Neoproterozoic sediments of Newfoundland were deposited at the northern margins of ancient Gondwana, with no evidence for basin restriction, suggesting open access to the global ocean (18, 19).

We used iron extraction techniques to explore ocean redox conditions. In this approach, operationally defined iron extraction protocols (20) are used to partition iron into its highly reactive components and its unreactive phases. Highly reactive iron includes iron oxide, carbonate, and sulfide minerals; this represents the iron that is geochemically and biologically active during early sediment diagenesis (21). By contrast, “unreactive” iron is geochemically inert on early diagenetic time scales. Previous studies have shown that in a broad suite of marine sediments deposited from an oxygen-containing water column, the ratio of highly reactive iron to total iron (FeHR/FeT) is consistently below 0.38, with a modern average (±SD) of 0.26 ± 0.08 (22, 23) and a Phanerozoic average (past 542 million years and excluding the modern) of 0.15 ± 0.06 (23). By contrast, sediments deposited from anoxic water columns may obtain additional reactive iron from iron mineral formation in the water column; in these environments, FeHR/FeT may exceed 0.38.

This is true both if the anoxic water column is sulfidic, such as the modern Black Sea (22) and ancient sulfidic marine water bodies (23, 24), and if it contained dissolved iron, as was the case early in Earth'shistory (25). We also used the concentrations of organic carbon and pyrite sulfur, as well as the isotopic composition of sulfur, to aid our characterization of the depositional environment (26) (table S1).

There is a marked difference in the proportion of highly reactive iron in sediments deposited before and after the Gaskiers glaciation (Fig. 3A). In numerous instances, the Gaskiers diamictite shows FeHR/FeT ratios exceeding 0.38, indicative of anoxic deposition. This is also the case, to a lesser extent, in the upper Mall Bay Formation. In other cases, the FeHR/FeT ratio of Gaskiers and Mall Bay samples is less than 0.38 but more than both the modern and Phanerozoic average ratios (Fig. 3A). In these cases, the extraction results are somewhat equivocal, but given the numerous instances in which FeHR/FeT exceeds 0.38 and the high overall ratios of FeHR/FeT, the results suggest that the water column was anoxic during deposition of the Gaskiers diamictite and probably also the upper Mall Bay (27). There is little sulfide sulfur in these rocks, and most of the reactive iron is bound as iron oxide and iron carbonate (26). Therefore, a sulfidic water column can be ruled out (25), indicating that the water column was most likely ferruginous, containing elevated concentrations of dissolved ferrous iron (Fe2+). Therefore, the deep-ocean chemistry accompanying the Gaskiers glaciation (and possibly also the Mall Bay) appears similar to the water chemistry associated with the earlier “snowball Earth” glaciations, in which banded iron formations accompanied the deposition of Sturtian-aged diamictites (about 700 Ma) (11) and Fe-enriched carbonates were deposited immediately after the Marinoan glaciation (about 630 Ma) (28), both of which are indicative of ferruginous oceanic conditions.

Fig. 3.

The chemistry of late-Neoproterozoic sediments from the Avalon Peninsula, Newfoundland. (A) FeHR/FeT ratios. Solid line represents the 0.38 ratio. Dashed lines represent the ratios for average modern and average Phanerozoic sediments deposited in an oxic water column. (B) Concentrations of reduced sulfur (pyrite) and organic carbon (OC) by weight (wt) %. (C) Isotopic composition of pyrite sulfur. A dashed line at 0 per mil is shown for reference. Also indicated are key dates (as in Fig. 1) and the stratigraphic level of prominent fossil locations.

In sediment accumulated immediately after the Gaskiers glaciation and during a period of time representing more than 15 million years, most FeHR/FeT ratios are much lower than in the underlying rocks and fall between the Phanerozoic and modern average for oxic sediment deposition; nearly all are less than 0.38 (Fig. 3A). These data provide evidence for a long period of stable deep-water oxic marine conditions. Overall, our data point to a pre-Drook anoxic iron-containing ocean giving way to oxic marine conditions after the Gaskiers glaciation.

For the Mall Bay through the Trepassey Formations, organic carbon concentrations are, in general, extremely low and are completely consistent with deposition in oligotrophic deep-water basinal or outer slope marine settings (29). Whereas higher organic carbon concentrations might be expected under anoxic conditions (30), our rather low concentrations in the Mall Bay and Gaskiers Formations are similar to those found (0.1 to 0.3 weight % C) in the Sturtian-aged (about 730 Ma) Rapitan Iron Formation, which also deposited from anoxic Fe-containing waters (31). The somewhat higher concentrations observed in the Fermeuse Formation are consistent with an environment closer to shore of higher sediment deposition rate such as a delta front would offer.

The isotopic composition of sulfur is quite variable but demonstrates patterns consistent with the above scenario for deep-ocean oxygenation, providing further insights into the nature of ocean chemistry. The isotopic composition of sulfide in pre-Gaskiers sediments is consistently greater than zero, indicating relatively small fractionations from seawater sulfate of around 18 ± 10 per mil (Fig. 3C) [the isotopic composition of seawater sulfate from 590 to 560 Ma is well constrained at between 22 and 28 per mil (32)]. This, combined with generally low concentrations of organic carbon and sulfide, is consistent with low rates of sulfate reduction under sulfate-limiting concentrations. Previous modeling (33) suggests that such low fractionations should occur with submillimolar sulfate concentrations. Higher fractionations in the Gaskiers, and particularly in the Drook Formation, demonstrate a change in the sulfur cycle. Indeed, the higher post–Mall Bay fractionations are consistent with an increase in sulfate concentration, which allows the expression of higher fractionations when compared with those produced with low sulfate levels. This pattern of increased fractionations also occurs with sediments deposited in association with and immediately after the Sturtian and Marinoan glaciations (34) and is thus a general feature of Neoproterozoic glacial and postglacial deposits.

To explain these increased fractionations, we suggest that glacial melting increased the nutrient load to the ocean. This stimulated primary production and carbon burial and thus increased atmospheric oxygen levels. Increased oxygen enhanced the oxidative weathering of sulfide to sulfate on the continents, thus increasing the flux of sulfate to the ocean and marine sulfate concentrations. The presence of anoxygenic photosynthetic biomarkers in post-Sturtian and post-Marinoan deposits argues against substantial ocean oxygenation after the earlier Neoproterozoic glaciations (35, 36), but the post-Gaskiers event was sufficient to result in oxygenation of the deep ocean. A return to much lower fractionations in post–Mistaken Point sediments could reflect a return to lower marine sulfate concentrations or, perhaps more likely, a reduction in isotope fractionation as might be expected with higher rates of sediment deposition and increased rates of sulfate reduction in a delta-front environment (33).

Our evidence for deep-water oxygenation in the post-Gaskiers ocean may place some constraints on the minimum level of atmospheric oxygen at this time. We reason as follows: Organic matter produced in the surface ocean consumes oxygen during degradation as it falls through the water column. In the modern ocean, oxygen concentrations reach a minimum at depths of 500 to 1500 m (37). The magnitude of the oxygen deficit is about 40 to 100 μM in the North Atlantic and 100 to 300 μM in the North Pacific (37). If we assume that the ocean structure was similar to its structure today and that the Drook and Briscal Formations were deposited in water depths of 500 to 1500 m, then oxygen deficits would have ranged somewhere between 40 and 300 μM. If we take 40 μMas the most conservative estimate, then at least this much O2 was dissolved in the waters supplying the deep ocean. Probably a bit more oxygen was required, given that the Ediacara biota would have likely needed 10 to 20 μM for their respiration (38), which we add to our minimum estimate of the oxygen content of the water supplying the deep ocean. At present, deep water is formed at high latitudes with an air-saturated O2 concentration of 325 μM. If we require a minimum of 50 μMO2 in this water, then we need to saturate with atmosphere containing greater than 15% of present day oxygen levels. Thus, 15% of present day oxygen levels is a minimum estimate for post-Gaskiers atmospheric O2.

In the Avalon region, this oxygenation was stable and persisted for at least 15 million years. In our preferred scenario, this oxygenation was widespread, marking the first time that oxygen concentrations reached levels permissive for the metabolism of large multicellular heterotrophic eukaryotes. If so, evolution into this new permissive ecology could have been quite rapid (7), resulting in the emergence of the Ediacara biota within 5 million years (1, 2), and recognizable motile animals within another 20 million years (3, 4). In another scenario, the Ediacaran biota populating Avalonia evolved earlier, perhaps before the Gaskiers glaciation, and migrated to Avalon after oxygenation of the local environment. The available data support the first scenario, but further exploration of pre-Gaskiers ocean chemistry and biology will help to elucidate the possibility of the second.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1135013/DC1

Materials and Methods

Table S1

References

References and Notes

View Abstract

Navigate This Article