Low-oxygen waters limited habitable space for early animals

Abstract

The oceans at the start of the Neoproterozoic Era (1,000-541 million years ago, Ma) were dominantly anoxic, but may have become progressively oxygenated, coincident with the rise of animal life. However, the control that oxygen exerted on the development of early animal ecosystems remains unclear, as previous research has focussed on the identification of fully anoxic or oxic conditions, rather than intermediate redox levels. Here we report anomalous cerium enrichments preserved in carbonate rocks across bathymetric basin transects from nine localities of the Nama Group, Namibia (∼550-541 Ma). In combination with Fe-based redox proxies, these data suggest that low-oxygen conditions occurred in a narrow zone between well-oxygenated surface waters and fully anoxic deep waters. Although abundant in well-oxygenated environments, early skeletal animals did not occupy oxygen impoverished regions of the shelf, demonstrating that oxygen availability (probably >10 μM) was a key requirement for the development of early animal-based ecosystems.

ARTICLE

Received 18 Nov 2015 | Accepted 3 Aug 2016 | Published 23 Sep 2016

DOI: 10.1038/ncomms12818 OPEN

Low-oxygen waters limited habitable space for early animals

R. Tostevin1,w, R.A. Wood2, G.A. Shields1, S.W. Poulton3, R. Guilbaud4, F. Bowyer2, A.M. Penny2, T. He1, A. Curtis2, K.H. Hoffmann5 & M.O. Clarkson2,6

The oceans at the start of the Neoproterozoic Era (1,000541 million years ago, Ma) were dominantly anoxic, but may have become progressively oxygenated, coincident with the rise of animal life. However, the control that oxygen exerted on the development of early animal ecosystems remains unclear, as previous research has focussed on the identication of fully anoxic or oxic conditions, rather than intermediate redox levels. Here we report anomalous cerium enrichments preserved in carbonate rocks across bathymetric basin transects from nine localities of the Nama Group, Namibia (B550541 Ma). In combination with Fe-based redox proxies, these data suggest that low-oxygen conditions occurred in a narrow zone between well-oxygenated surface waters and fully anoxic deep waters. Although abundant in well-oxygenated environments, early skeletal animals did not occupy oxygen impoverished regions of the shelf, demonstrating that oxygen availability (probably 410 mM) was a key requirement for the development of early animal-based ecosystems.

1 Department of Earth Sciences, University College London, Gower Street, London WC1E 6BT, UK. 2 School of GeoSciences, The University of Edinburgh, James Hutton Road, Edinburgh EH9 3FE, UK. 3 School of Earth and Environment, University of Leeds, Leeds LS2 9JT, UK. 4 Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge CB2 3EQ, UK. 5 Geological Survey of Namibia, Private Bag 13297, Windhoek, Namibia.

6 Department of Chemistry, University of Otago, Dunedin 9054, New Zealand. w Present address: Department of Earth Sciences, University of Oxford, Oxford OX1 3AN, UK. Correspondence and requests for materials should be addressed to R.T. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 7:12818 | DOI: 10.1038/ncomms12818 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12818

Geochemical proxies based on Fe-S-C and trace metal systematics have been widely used to reconstruct the progressive oxygenation of the oceans during the

Neoproterozoic and Cambrian17. Accumulating evidence indicates that the deep oceans were dominantly anoxic and ferruginous (Fe containing) throughout most of the Precambrian, with euxinic (suldic) mid-depth waters prevalent along continental margins from B1.8 to 1.0 billion years ago (Ga)1,6,8. From B1.0 to 0.58 Ga, however, euxinic mid-depth waters became less prevalent and ferruginous conditions expanded, with oxic conditions still largely restricted to surface waters1,8,9. The oxygenation of the deeper marine realm was both protracted and spatially heterogeneous, with some marine basins recording persistent deep-water oxygenation from B580 Ma, whereas regional anoxia remained a feature of some deeper shelf environments into the Cambrian, B520 Ma (refs 4,5,7,10) and beyond.

The course of Neoproterozoic oxygenation, and cause and effect associations with the appearance of animals, remains controversial4,1113. Although modern soft-bodied sponge-grade animals may tolerate oxygen concentrations as low as1.2510 mM14, new innovations in the late Ediacaran, such as motility15, the rise of predation and skeletonization1618, are all hypothesized to have required higher levels of oxygen19. However, the oxygen demands of early animals are unconstrained and observations from modern biota cannot necessarily be applied to early animals of unknown afnity. Furthermore, although soft-bodied and skeletal Ediacaran fauna dominantly occur in sediments interpreted to have been deposited from oxic waters, fossil occurrences have also been reported in sediments characterized by anoxic geochemical signals5,20. In the latter case, this may be because some early complex organisms were able to colonize habitats during eeting periods of oxia (such short-lived oxygenation is difcult to detect by geochemical proxies that tend to integrate relatively long periods of time). In both of the above

cases, however, there is uncertainty as to whether early animal evolution occurred under fully oxygenated conditions or whether intermediate redox conditions were more prevalent, which by extension suggests that the oxygen requirements of more complex organisms were lower3,14. An in-depth understanding of these links is currently hampered by the inability of most redox proxies to distinguish between fully oxygenated and intermediate redox states, including nitrogenous or manganous conditions, which may overlap with low concentrations of oxygen21,22. Indeed, it is possible that oxic horizons identied through Fe and trace element geochemistry may in fact have formed under low-oxygen conditions (but not fully anoxic), at levels insufcient to support diverse skeletal animal communities.

In oxic environments, Ce(III) is oxidized to insoluble Ce(IV) and preferentially scavenged relative to the rest of the rare earth elements and yttrium, REY23. The standard reduction potential of Ce(IV) ( 1.61V) is closer to Mn(IV) ( 1.23V) than Fe(III)

( 0.77V) and Ce oxidation is catalysed on the surface of Mn

(oxyhydr)oxides24. Therefore, the redox cycling of Ce in seawater is closely related to Mn(II)/Mn(IV) transformations, which occur at a higher redox potential than the Fe(II)/Fe(III) couple, and hence Mn cycling is more sensitive to intermediate redox conditions2326. Ce anomalies (CeSN=Ce SN) are calculated here based on relative enrichments or depletions in shale-normalized

Ce ([Ce]SN) compared with neighbouring non-redox sensitive REY:

CeSN=Ce SN

1

Owing to the accumulation of Ce(IV) on the surface of Mn (oxyhydr)oxides, oxic seawater becomes Ce depleted and exhibits a negative Ce anomaly (o0.9)23. These Mn (oxyhydr)oxides may be buried intact in sediments beneath oxic bottom waters, or may dissolve in the water column if they encounter low-oxygen waters,

Ce/Ce*

1 2

Pr SN2=Nd SN

Ce SN

O2(M)

10

0

0

Figure 1 | Schematic representation of redox zones and associated geochemical signals. Generalized redox conditions across the Zaris Basin,Nama Group, during a highstand systems tract. Positive Ce anomalies form as Mn (oxyhydr)oxides dissolve in the manganous zone and Fe enrichments form under anoxic ferruginous conditions. Ten micromole is an estimate of O2 concentrations in the manganous zone, but overlying well-oxygenated waters probably contained higher O2 concentrations. Representative REY patterns, including positive Ce anomalies (magnitude in brackets), are shown for the Omkyk section in the Nama Group, alongside manganous zones from two modern environments25,27 (modern water column data plotted as [REY] 106

for easy comparison with sedimentary [REY]).

2 NATURE COMMUNICATIONS | 7:12818 | DOI: 10.1038/ncomms12818 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12818 ARTICLE

releasing excess Ce. Therefore, waters beneath the Mn(IV)/Mn(II) redoxcline commonly exhibit a positive Ce anomaly (41.3)2527.

Positive Ce anomalies have been recorded alongside Mn enrichments in some modern waters, including Lake Vanda, Antarctica (CeSN=Ce SN up to 2.3, Fig. 1)25, in anoxic brines in the eastern Mediterranean (CeSN=Ce SN up to 2.43, Fig. 1)27 and in the deep-marine Cariaco Basin (CeSN=Ce SN up to 1.21)26.

Water column REY and associated Ce anomalies are thought to be preserved in non-skeletal carbonate rocks without fractionation28. Carbonate-bound REY are relatively robust to diagenetic alteration28 and dolomitization29, but any alteration of the Ce anomaly can be identied using non-redox sensitive REY anomalies, such as the Y/Ho ratio, which would also be altered away from seawater patterns30,31. Sequential dissolution methods enable REY in the carbonate phase to be isolated, preventing contributions from sedimentary (oxyhydr)oxides or clays31, which would carry a non-seawater signature.

In the present study, we measured REY in 259 carbonate rocks of the Nama Group from nine sites across two basins. The majority of samples are very pure calcites with low siliciclastic components, but where samples are partially dolomitized they have been treated differently during leaching31. The resulting REY data have been screened for traditional seawater features (Y/Ho ratios 467) and samples with evidence for diagenetic alteration or contributions from non-carbonate phases have been excluded from the presented Ce/Ce* data (see Methods). We additionally use redox interpretations based on published Fe speciation data for these carbonate samples5. Fe speciation distinguishes anoxic from oxic water column conditions through enrichments in highly reactive Fe (FeHR) relative to total Fe (FeT)1,32. Anoxic enrichments in FeHR occur due to the water column formation of either pyrite under euxinic conditions32 or non-suldized FeHR minerals (such as Fe oxides or carbonates) under anoxic ferruginous conditions1 (see Methods). We interpret ususual Ce enrichments across the Nama Group to indicate Mn-rich, low-oxygen conditions, supported by additional redox information from Fe-based proxies on the same samples. This enables us to distinguish fully anoxic, intermediate and well-oxygenated waters across a shelf-to-basin transect and compare these with the distribution of early skeletal animal life.

ResultsGeological setting. The succession was deposited B550541 Ma broadly coincident with the rst appearance of skeletal animals1618, as well as trace fossil evidence for motility15 and soft-bodied fossils belonging to the Ediacaran biota33. Our samples cover a range of palaeo depths from shallow inner ramp to deeper outer ramp waters, in the Kanies, Omkyk and Hoogland Members of the Kuibis Subgroup, and the Spitzkopf and Feldschuhorn Members of the upper Schwarzrand Subgroup5,16 (Fig. 2, Supplementary Table 1 and also see Supplementary Note 1 for full details of the geological setting). We focus on the rst known skeletal animals, Cloudina, a globally distributed eumetazoan of possible cnidarian afnity17,34,35; Namacalathus, interpreted as a stem group eumetaozan36 or triploblast lophophorate37 and reported from multiple localities; and Namapoikia, an encrusting possible cnidarian or poriferan known only from the Nama Group18.

Ce anomaly interpretation. In the Nama Group, the majority of REY distribution patterns are smooth and show a at or light REY-depleted shape on shale-normalized plots, positive La anomalies, low total rare earth elements (REE) concentrations and superchondritic Y/Ho ratios (467), all of which indicate preservation and extraction of original seawater signals (see

Supplementary Note 2 and supplementary Figs 110 for a description of all data). Four samples exhibit negative Ce anomalies (o0.9; Fig. 2), consistent with an oxic water column interpretation obtained for these samples from Fe speciation5. Ce anomalies are, as expected, absent from the persistently anoxic and ferruginous deepest water setting5 (Fig. 2). However, signicant positive Ce anomalies (1.302.15) are prevalent in inner ramp sections in both sub-basins (64 samples). In six cases, positive Ce anomalies are associated with anoxic ferruginous signals and in one case a positive Ce anomaly is associated with a sample that gives a robust oxic FeHR/FeT signal. However, for the majority of samples (B90%), FeT was o0.5 wt%, preventing a robust evaluation of water column redox conditions from Fe speciation alone. In these cases, samples have elevated Mn/Fe ratios (median 0.39), when compared with samples with no

positive Ce anomalies (median 0.14) and anoxic ferruginous

samples (median 0.10), which provides an independent

constraint on water column redox conditions, as discussed below (Fig. 3).

The regionally widespread positive Ce anomalies across the Zaris and Witputs Basins of the Nama Group imply a surplus of Ce sustained by a rain down of Mn (oxyhydr)oxides from shallow oxygenated surface waters and this is supported by the elevated Mn/Fe ratios of these samples (Fig. 3). Redox cycling of Mn (oxyhydr)oxides across the Mn(IV/II) redoxcline would leave ambient waters locally enriched in the Ce released during Mn(IV) reduction (Fig. 1). We therefore interpret positive Ce anomalies (41.3) to indicate intermediate manganous conditions (Table 1 and also see Supplementary Discussion for alternative Ce enrichment mechanisms). Where there is an absence of both positive Ce anomalies and any indication of enrichment in Fe (that is, FeHR/FeTo0.22 or FeTo0.5 wt%), we suggest that bottom waters were probably well oxygenated (which is consistent with FeHR/FeT signals in interbedded siliciclastics5), thus preventing the onset of both Fe and Mn reduction. Where data are equivocal (for example, FeHR/FeT between 0.220.38 and no Ce anomaly), we are unable to interpret redox conditions.

Positive Ce anomalies have not been widely reported from carbonate-rich sediments, but there are limited examples from iron formation38 and cherts39 in the earlier Paleozoic. Positive Ce anomalies, between 1.3 and 2.2, are also reported for late Ediacaran dolomites, from just a few samples in the possibly contemporaneous Krol Formation of northern India40. If these data were demonstrated to preserve seawater REY patterns, this strengthens the data from the Nama Group and hints that manganous conditions may have been a common feature of Ediacaran oceanic margins. By contrast, demonstrably contemporaneous terminal Ediacaran Ce anomaly data from the Yangtze platform, South China, show increasing Ce depletion towards the EdiacaranCambrian Boundary, indicating progressive oxygenation of the local marine environment41. Ce anomalies record local redox conditions and thus independent signals would be expected both within and between marine basins.

Redox conditions in the Nama group. The outer ramp was persistently anoxic and ferruginous (Brak section), and animals are absent from these settings5 (Fig. 4). The deep inner-ramp sections show periods of anoxic ferruginous, manganous and well-oxygenated conditions (Zebra River and Omkyk sections). In these settings, animals are notably absent from ferruginous and manganous waters, whereas well-oxygenated waters support abundant skeletal animals, up to 35 mm in diameter, and adjacent localities show trace fossil evidence for motility15. The shallowest inner ramp sections show high-frequency temporal uctuations

NATURE COMMUNICATIONS | 7:12818 | DOI: 10.1038/ncomms12818 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12818

541 Ma

a

Dreidoornvlagte

Zebra River

Zwartmodder

Grens

1.4

Kuibis subgroupSchwarzrand subgroup

Feldschuhorn

26 S

28 S

15 E

547 Ma

c

Ce/Ce*

2.4

OSIS RIDGE

550 Ma

South

0

Figure 2 | Summary of CeSN/Ce*SN and Fe-speciation data for nine localities. The location of nine sections within the Kuibis and Schwarzrand Subgroups of the Nama Group is shown on a simplied geological map of Namibia52,53,55,57 (a), as well as on a schematic cross-section, indicating average relative

water depth (b). The numbers along the basin prole relate to the relative position of each section as numbered in c. FeHR/FeT data for each location is shown for carbonate and siliciclastic rocks5, alongside CeSN=Ce SN data, screened for carbonate rocks showing seawater REY patterns (for example, molar

Y/Ho467) (c). Blue FeHR/FeT data indicate where Fe speciation predicts oxic conditions5 and positive Ce anomalies indicate where waters are interpreted to have been manganous. The presence of in situ biota is noted by grey lines5.

between anoxic ferruginous, manganous and well-oxygenated conditions (Zwartmodder, Arasab and Grens sections), as might be expected due to uctuations in the depth of the chemocline (Fig. 4). At Zwartmodder, skeletal animals are present in thin beds5, but there is only one skeletal horizon at Grens and no animal fossils at Arasab.

In contrast to these ecologies, the Driedoornvlagte pinnacle reef grew within a transgressive systems tract in a mid-ramp position, which was persistently well-oxygenated and hosts some very large skeletal animals5,18 (up to 1 m) and complex reef-building ecologies17. In the younger Schwarzrand Subgroup, which extends close to the EdiacaranCambrian Boundary (B547541 Ma), there is evidence for persistent well-oxygenated conditions5 and mid-ramp Pinnacle Reefs host mixed communities of large and small skeletal animals. At Swartpunt, abundant burrows and soft-bodied biota occur in siliciclastic horizons, where Fe speciation indicates oxic conditions5, whereas small in-situ skeletal animals are found in carbonate rocks throughout the succession5.

DiscussionOur geochemical and palaeontological data demonstrate a striking relationship between the precise redox condition of the water column and the presence and abundance of evidence for animal life. Constraints from the modern open ocean suggest that dissolved Mn(II), and therefore Ce(III), can start to build up in low concentrations in oxic waters with dissolved O2o100 mM22.

However, manganous conditions, whereby Mn becomes the dominant redox buffer, are achieved at lower oxygen concentrations. Reduced Mn can remain stable in the presence of up to 10 mM O2 (refs 21,42), although Mn oxidation has been reported locally at lower O2 concentrations where oxidation is catalysed by enzymatic processes43. Thus, active Mn cycling can occur in anoxic waters, but is commonly documented in partially oxic waters with at least 10 mM O2 (and up to 100 mM O2;

Fig. 1)21,42,44,45, which represents signicant oxygen depletion in comparison with modern fully oxygenated surface waters (B250 mM O2). The reduction potential for Ce is higher than that for Mn and so the 10 mM O2 constraint for manganous

4 NATURE COMMUNICATIONS | 7:12818 | DOI: 10.1038/ncomms12818 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12818 ARTICLE

waters may represent a lower limit on Ce cycling, as sufcient O2 to oxidize both Ce and Mn is required for the formation of Ce anomalies.

Our multi-proxy approach allows us to distinguish between fully anoxic and intermediate waters, which contained low but signicant amounts of oxygen. Where Fe speciation in Ce-enriched samples gives a robust anoxic signal (FeHR/FeT40.38),

Mn reduction may have persisted, but conditions must have been fully anoxic. However, the majority of samples interpreted to be manganous have insufcient FeT for Fe speciation (with 85% of these falling below 0.25 wt% FeT and 35% falling below 0.1 wt% FeT). Even very low oxygen concentrations (nM) are sufcient to prevent FeHR enrichments and thus the low FeT in shallower environments across the Nama Group may be indicative of oxic conditions46, and this is supported by persistent oxic FeHR/FeT ratios obtained from interbedded siliciclastics in some sections5.

We therefore suggest that the manganous zone occurred between well-oxygenated surface waters and deeper anoxic, ferruginous

waters, commonly overlapping with low but signicant concentrations of oxygen (at least B10 mM; Fig. 1).

Oxygen exerts an important control on ecosystem structure in modern environments, whereby low-oxygen environments are inhabited by smaller animals often lacking skeletons and forming low-diversity communities with simple food webs19. In general, skeletons are absent from modern oxygen minimum zones when O2 drops below 13 mM and large animals are often absent below 45 mM (refs 47,48). However, the importance of oxygen in supporting early animal ecosystems as they became increasingly complex in form, metabolic demand and behaviour through the Ediacaran Period is currently unresolved2,3,5,1113. In the Nama Group the majority of small skeletal animals (475%) and all evidence for large skeletal animals, motility, soft-bodied biota and complex or long-lived ecologies17 are found in sediments deposited from well-oxygenated waters (Fig. 4). The identication of low-oxygen, manganous water column conditions thus provides a compelling explanation for the general absence of biota in these settings and implies that poorly oxygenated conditions were insufcient to meet the relatively high oxygen requirements of these early skeletal animals5,15,17,33. If we take an upper O2 limit for Mn and Ce reduction of 10 mM O2, this suggests that Mn-enriched waters could theoretically support small, soft-bodied animals such as sponges14. In contrast, the absence of skeletal animals in Mn-enriched waters is consistent with the high energetic cost of skeletonization. Possible biomarkers for sponge animals appear in the fossil record at 4635 Ma (ref. 49), but it is possible that the availability of well-oxygenated habitats was necessary to support the later appearance of skeletonization, at B550 Ma. However, it is also unlikely to be that reaching an oxygenation threshold alone is sufcient to explain the appearance of skeletons50 and many have argued that the trigger for the rise of skeletonization may have been ecological, such as the rise of predation17,36,51.

Our approach highlights that intermediate redox conditions were probably widespread in the Ediacaran ocean, but have not previously been appreciated due to the inability of most commonly used proxies to identify such conditions. Our data suggest that low-oxygen water column conditions were insufcient to support early skeletal and reef-building animals, and thus the extent of suitable habitat space may have been less than previously identied. The widespread radiation of skeletal animals during the subsequent Cambrian explosion may have been facilitated by a global rise in the extent of habitable, oxygenated seaoor7, alongside other genetic and ecological factors. Our data therefore yield new insight into the debate on the role of oxygen in early animal evolution, suggesting that well-oxygenated waters were necessary to support the appearance of the skeletal animals and complex ecologies that are typical of the terminal Neoproterozoic.

5

13.17

12.01

Global carbonates

0.10

0.39

0.14

4

3

Mn/Fe

2

1

0.29

0

Ferruginous Manganous Oxic

Figure 3 | Average Mn/Fe ratios under different redox conditions. Mn/Fe ratios for samples identied as manganous (positive Ce/Ce* and low FeT or

oxic Fe-speciation signals), ferruginous (anoxic Fe-speciation signals) and oxic (oxic Fe-speciation signals, no positive Ce anomaly). Mn/Fe is enriched in manganous samples compared with global carbonate (0.29; ref. 70). Bars represent median values. Red outlines indicate dolomitized samples. Arrows indicate samples with exceptionally high Mn/Fe ratios that lie above the limit of the y axis.

Table 1 | Framework for co-interptetation of CeSN/CeSN* and Fe-speciation data on the same samples.

Ce anomaly

Fe-speciation Negative anomaly Equivocal (no anomaly) Positive anomaly Anoxic

FeHR/FeT40.38, Fepy/FeHRo0.7 NA Ferruginous Ferruginous

EquivocalFeHR/FeT 0.220.38 Oxic Unknown Manganous

OxicFeHR/FeTo0.22, FeTo0.5wt% (likely to be oxic) Oxic Oxic Manganous

Fe , highly reactive Fe; Fe , total Fe; NA, not applicable.

NATURE COMMUNICATIONS | 7:12818 | DOI: 10.1038/ncomms12818 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12818

25

541 Ma

a

Swartpunt

Pinnacle reefs

24

Well-oxygenated conditions

Manganous low-oxygen conditions

Ferruginous anoxic conditions

Redox

20

Small skeletons

Large skeletons

Soft-bodied biota

Local ecology Redox

Local ecology Redox

Kuibis subgroupSchwarzrand subgroup

Frequency

Burrows

15

10

6

6

Feldschuhorn

5

Animal fossils with in-situ geochemistry

2 2

0

Inner ramp Mid ramp

550 Ma

Fleeting oxia

Low oxygen

Well oxygenated

547 Ma

Zwartmodder

Omkyk

Local ecology Redox

Arasab

b

Zebra River

Redox

Driedoornvlagte

L Om U Omkyk Hoogland

Kanie

Local ecology Redox

Grens

Local ecology Redox Local ecology

Redox

Namapoikia reef building

1

Brak

Local ecology Redox

Inner ramp

Inner ramp

Wood et al., 2002 (ref. 18)

Penny et al., 2014 (ref. 17)

Deep inner ramp

Deep inner ramp

Inner ramp

OSIS RIDGE

Outer ramp

North South

Zaris Basin Witputs Basin

Figure 4 | Integrated redox interpretations compared with local ecology. A comprehensive redox interpretation is shown for each of the nine localities in the Nama Group, determined using combined Fe and Ce signals. In situ fossils (grey lines), and local ecologies5 and general ecology from the literature17,18 are shown alongside local water column redox conditions (b). The bar chart plots the frequency that different biota are found in each redox zone5 (a).

Large skeletal fossils and burrows are found exclusively in well-oxygenated settings and small skeletal fossils are largely restricted to well-oxygenated conditions, but may occur where conditions were only eetingly oxic.

Methods

Geological context and sample quality. The Nama Group is a well-preserved terminal Neoproterozoic carbonate and siliciclastic sequence, ranging from upper shore-line/tidal ats to below-wave-base lower shoreface, deposited in a ramp system B550541 Ma (refs 5,16,5255). Samples from nine shelf-to-basin sections within the Zaris and Witputs basins of the Nama Group encompass a range of palaeo-depths from outer- to inner-ramp settings (Supplementary Table 1). Stratigraphic correlations are well-established based on sequence boundaries and ash beds5,16. The age of the upper Nama Group is relatively well-constrained from U-Pb dating of three ash beds within the group, including one at 548.81 Ma in the Hoogland Member of the Kuibis Subgroup54, revised to 547.320.31 Ma (ref. 56). The base of the Nama Group is diachronous, but is between 553 and 548 Ma. The ProterozoicCambrian boundary is represented by a regionally extensive erosional unconformity near the top of the Schwarzrand Subgroup in the southern Basin53,54,57,58, which is overlain by incised-valley ll dated (U-Pb on an ashbed) at 5391 Ma (ref. 54). Therefore, the Nama Group section spans at least 7 Myr and extends to within 2 Myr of the EdiacaranCambrian Boundary52.

Unweathered samples were selected and powdered or drilled avoiding alteration, veins or weathered edges. For Zebra River section, powders were drilled from thin section counterparts, targeting ne-grained cements. Carbonate rocks in the Nama Group are very pure, but they have all undergone pervasive recrystallization. Less than 15% of the samples in this study are dolomitized and there is no petrographic evidence for deep burial dolomitization in the Nama Group5,55.

Samples were logged for fossil occurrences and sampled within established sequence stratigraphic frameworks, using detailed sedimentology5,52,53(see Supplementary Notes 1 and 2). The presence of different forms of skeletal biota, soft-bodied biota and trace fossils are reported for precise horizons where geochemical analyses have been performed5, indicated by grey lines in Figs 2 and 3.

General local ecology, supported by additional information from the literature, is also marked, without associated grey lines. Our sampling focused on carbonates and hence skeletal fossils are over-represented compared with soft-bodied biota and trace fossils. We dene large skeletal animals as 410 mm in any dimension, which includes Cloudina hartmannae, some Namacalathus and Namapoikia.

Rare earth elements in carbonate rocks. Rare earth elements and yttrium (REY) have a predictable distribution pattern in seawater and non-biological carbonate rocks should preserve local water column REY at the sedimentwater interface28. Ce anomalies develop progressively, but cutoff values are established to dene negative and positive anomalies. We dene a negative anomaly as CeSN=Ce SNo0:9, consistent with previous work59. A positive anomaly, using the same reference frame, would be dened as CeSN=Ce SN41:1. However, as positive anomalies are not previously described from carbonate sediments, we cautiously use a higher cutoff, CeSN=Ce SN41:3, to ensure any positive anomalies are environmentally signicant with respect to positive anomalies recorded from some modern manganous waters (1.212.43) (see Supplementary Note 3 and Supplementary Fig 11 for discussion of CeSN=Ce SN cutoffs). Although positive or negative Ce anomalies in carbonate rocks probably represent seawater redox conditions, the absence of any Ce anomaly(0.91.3) is somewhat equivocal and could result from anoxic water column conditions or overprinting of any Ce anomaly during diagenesis or leaching31. Alternately, Ce anomaly formation may be disrupted in surface waters because of wind-blown dust or photo-reduction of Mn oxides60. Fe (oxyhydr)oxides may also be REY carriers, but do not contain the clear Ce enrichments observed in Mn (oxyhydr)oxides (see Supplementary Note 4 for discussion of the role of Fe (oxyhydr)oxides in REY cycling).

Diagenetic phosphates, Fe and Mn (oxyhydr)oxides, organic matter and clays can potentially affect the REY signatures of authigenic sedimentary rocks if they are

6 NATURE COMMUNICATIONS | 7:12818 | DOI: 10.1038/ncomms12818 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12818 ARTICLE

partially dissolved during the leaching process6163. Care has been taken to partially leach samples, to isolate the carbonate phase without leaving excess acid, which may leach contaminant phases (see ref. 31 for detailed discussion of methodology). Powdered calcite samples were cleaned in Milli-Q water and pre-leached in 2% nitric acid, to remove adsorbed and easily exchangeable ions associated with clay minerals. The remaining sample was partially leached, also in 2% (w/v) nitric acid, to avoid contributions from contaminant phases such as oxides and clays31. The supernatant was removed from contact with the remaining residue, diluted with 2% nitric acid and analysed via inductively coupled plasma mass spectrometry in the Cross-Faculty Elemental Analysis Facility, University College London. This leaching method has been designed to extract the carbonate-bound REY pool without contributions from (oxyhydr)oxides or clays31. These same leachates were also analysed for major element concentrations (Mg, Fe, Mn, Al and Sr) via inductively coupled plasma optical emission spectrometry. Oxide interference was monitored using the formation rate of Ce oxide and the formation of 2 ions was monitored using Ba2. All REY concentrations were

normalized to post-Archean Australian Shale.

Standard solutions analysed after every ten samples were within 5% of known concentrations. Replicate analyses on the inductively coupled plasma mass spectrometry give a relative s.d. o5% for most trace elements, with a larger s.d. for the heavy REE that sometimes have non-normalized concentrations o0.5 p.p.b.

Carbonate standard material CRM 1c was prepared using the same leaching procedure as the samples and repeat analyses give a relative s.d. o5% for individual REY concentrations, and calculated Ce anomalies (average 0.80)

give a relative s.d. o3%.

Mn/Sr ratios are o1 for the majority (97%) of samples and d18Ocarb is 4 10%, indicating minimal open-system elemental and isotopic exchange

during diagenesis, and excluding deep burial dolomitization (SupplementaryFig. 12). Ce anomaly data are only presented for carbonates that preserve seawater REY features (smooth patterns with molar Y/Ho467)28,31, indicating they originate from the carbonate portion of the whole rock, without contributions from detrital or oxide phases. For samples with Y/Ho467, 85% also have PREE o2 p.p.m. and all have PREE o10 p.p.m. La anomalies, and small positive Eu and Gd enrichments are prevalent in samples with Y/Ho467 (Supplementary Fig. 11 and Supplementary Note 5 for discussion of Y anomaly thresholds). Positive Ce anomalies are associated with low Mn/Sr ratios (o1) and low Al, Zr, Ti, Fe and Mn contents in the leachate (o0.2 wt% for Fe and o500 p.p.m. for Mn), indicating minimal contamination due to diagenetic exchange, leaching of clays or FeMn (oxyhydr)oxide phases (Supplementary Fig. 12).

Rare earth elements in shales. Shales from throughout the Zebra River section, including inter-reef deposits and lateral subordinate shales between grainstone horizons, were fully digested using HNO3-HF-B(OH)3-HClO4 at the University of

Leeds. These full digestions include the dominant siliciclastic component, but would also encompass any subordinate (oxyhydr)oxide phases, organic matter or carbonate components. The full digestions were dried down, washed twice in 50% nitric acid and resuspended in 2% nitric acid for analysis on an inductively coupled plasma mass spectrometry in the Cross-Faculty Elemental Analysis Facility, University College London.

Relative to standardized shale composition, post-Archean Australian Shale64, the Zebra River shales show consistent patterns (Supplementary Fig. 13), with middle-REY enrichment (bell-shaped index 1.25) and negative Y anomalies

(shale-normalized Y/Ho 0.88), but no anomalous Ce behaviour. These patterns

resemble those reported for Fe (oxyhydr)oxides65,66 and may well derive in part from the high Feox contents of these shales (up to 1.2%). Shales carry a continental-type REY pattern and represent a baseline from which surface-solution fractionation of REY begins, and thus they are commonly used to normalize seawater REY patterns.

Fe speciation in carbonates and siliciclastics. The Fe speciation method quanties Fe that is (bio)geochemically available in surcial environments (termed FeHR) relative to FeT. Mobilization and subsequent precipitation of Fe in anoxic water column settings results in FeHR enrichments in the underlying sediment. The nature of anoxia (that is, sulde-rich or Fe-containing) is determined by the extent of suldation of the FeHR pool1. Fe speciation data for carbonate rock samples discussed here and accompanying interbedded siliciclastic rocks come from previously published data5. The Fe-speciation technique was performed using well-established sequential extraction schemes1,67. The method targets operationally dened Fe pools, including carbonate-associated-Fe (FeCarb),

ferric oxides (FeOx), magnetite (FeMag), pyrite Fe (FePy) and FeT. FeHR is dened as the sum of Fecarb (extracted with Na-acetate at pH 4.5 and 50 C for 48 h), Feox (extracted via Na-dithionite at pH 4.8 for 2 h), Femag (extracted with ammonium oxalate for 6 h) and Fepy (calculated from the mass of sulde extracted during CrCl2 distillation). FeT extractions were performed on ashed samples (8 h at550 C) using HNO3-HF-H3BO3-HClO4. All Fe concentrations were measured via atomic absorption spectrometry and replicate extractions gave a relative s.d. of o4% for all steps, leading to o8% for calculated FeHR. FePy was calculated from the wt% of sulde extracted as Ag2S using hot Cr(II)Cl2 distillation68. A boiling HCl distillation before the Cr(II)Cl2 distillation ruled out the potential presence of acid volatile suldes in our samples. Pyrite extractions give reproducibility for Fepy

of 0.005 wt%, conrming high precision for this method. Analysis of a certied reference material (PACS-2, FeT 4.090.07 wt%, n 4; certied

value 4.090.06 wt%) conrms that our method is accurate. Replicate analyses

(n 6) gave a precision of 0.06 wt% for FeT and a relative s.d. of o5% for the

FeHR/FeT ratio.

Calibration in modern and ancient marine environments suggests that FeHR/FeTo0.22 indicates deposition under oxic water column conditions, whereas

FeHR/FeT40.38 indicates anoxic conditions1. Ratios between 0.220.38 are considered equivocal and may represent either oxic or anoxic depositional conditions. For sediments identied as anoxic, Fepy/FeHR40.8 is diagnostic for euxinic conditions and Fepy/FeHRo0.7 denes ferruginous deposition1. Although originally calibrated for siliciclastics1,32, enrichments in FeHR/FeT can also be identied in carbonates deposited under anoxic water column conditions69. These FeHR enrichments can far exceed FeHR contents expected under normal oxic deposition, where trace amounts (B0.1 wt%) of Fe may be incorporated into carbonates, or precipitate as FeMn coatings69. However, although early dolomitization in shallow burial environments does not generally cause a signicant increase in FeHR, late-stage deep-burial dolomitization may signicantly increase FeHR69, but there is no petrographic evidence for deep-burial dolomitization in our samples5,16. Consistent with a recent calibration69, we have limited the application of Fe speciation to carbonate samples with 40.5 wt% FeT,

which buffers against the impact of non-depositional enrichments in FeHR69.

Where FeT is very low (o0.5wt%), this may indicate deposition under oxic conditions69. In addition, however, we stress that all of our redox interpretations based on Fe speciation in carbonates are entirely consistent with data from siliciclastic horizons interbedded with and/or associated with carbonate rocks contained within the same m- to dm-scale depositional cycle5.

Equivocal FeHR/FeT ratios could be a consequence of dilution of a high water column FeHR ux through rapid sedimentation32 (for example, in turbidite settings) or post-depositional transformation of unsuldized FeHR minerals to less reactive sheet silicate minerals8,67. Further, local FeHR enrichments can occur due to preferential trapping of FeHR in inner shore or shallow marine environments (for example, ood plains, salt marshes, deltas and lagoons). However, none of the presented FeHR data here are from rocks that show evidence of turbiditic deposition and are from dominantly open marine settings. Oxidative weathering may result in mineralogical transformation of Fe minerals. Oxidation of siderite would transfer Fecarb to the Feox pool and hence any interpretation of ferruginous or euxinic signals would remain robust. The weathering of pyrite to Feox would not affect interpretation of anoxic signals (FeHR/FeT40.38), but may reduce the Fepy/FeHR ratio, giving a false ferruginous signal in a euxinic sample. In the extreme and highly unlikely scenario that all Feox in our samples is a product of pyrite weathering, B10% of the anoxic samples would give a euxinic signal. However, signicant Fecarb (420% of the FeHR fraction) occurs in B57% of anoxic samples, indicating that the rocks have not been completely weathered and hence this extreme scenario is unlikely.

Data availability. All relevant data are available to download in the data repository associated with this manuscript and further details on the Fe-speciation data are available in ref. 5.

References

1. Poulton, S. W. & Caneld, D. E. Ferruginous conditions: a dominant feature of the ocean through earths history. Elements 7, 107112 (2011).

2. Johnston, D. T. et al. Late Ediacaran redox stability and metazoan evolution. Earth Planet. Sci. Lett. 335, 2535 (2012).

3. Planavsky, N. J. et al. Low Mid-Proterozoic atmospheric oxygen levels and the delayed rise of animals. Science 346, 635638 (2014).

4. Caneld, D. E., Poulton, S. W. & Narbonne, G. M. Late-Neoproterozoic deep-ocean oxygenation and the rise of animal life. Science 315, 9295 (2007).

5. Wood, R. A. et al. Dynamic redox conditions control late Ediacaran ecosystems in the Nama Group, Namibia. Precambrian Res. 261, 252271 (2015).6. Scott, C. et al. Tracing the stepwise oxygenation of the Proterozoic ocean. Nature 452, 456459 (2008).

7. Chen, X. et al. Rise to modern levels of ocean oxygenation coincided with the Cambrian radiation of animals. Nat. Commun. 6, 7142 (2015).

8. Poulton, S. W., Fralick, P. W. & Caneld, D. E. Spatial variability in oceanic redox structure 1.8 billion years ago. Nat. Geosci. 3, 486490 (2010).

9. Guilbaud, R., Poulton, S. W., Buttereld, N. J., Zhu, M. & Shields-Zhou, G. A. A global transition to ferruginous conditions in the early Neoproterozoic oceans. Nat. Geosci. 8, 466470 (2015).

10. Sperling, E. A. et al. Statistical analysis of iron geochemical data suggests limited late Proterozoic oxygenation. Nature 523, 451454 (2015).

11. Nursall, J. Oxygen as a prerequisite to the origin of the Metazoa. Nature 183, 11701172 (1959).

12. Lenton, T. M., Boyle, R. A., Poulton, S. W., Shields-Zhou, G. A.& Buttereld, N. J. Co-evolution of eukaryotes and ocean oxygenation in the Neoproterozoic era. Nat. Geosci. 7.4, 257265 (2014).

13. Buttereld, N. J. Oxygen, animals and oceanic ventilation: an alternative view. Geobiology 7, 17 (2009).

NATURE COMMUNICATIONS | 7:12818 | DOI: 10.1038/ncomms12818 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12818

14. Mills, D. B. et al. Oxygen requirements of the earliest animals. Proc. Natl Acad. Sci. USA 111, 41684172 (2014).

15. Macdonald, F. A., Pruss, S. B. & Strauss, J. V. Trace fossils with spreiten from the late Ediacaran Nama Group,Namibia: complex feeding patterns ve million years before the PrecambrianCambrian boundary. J. Paleontol. 88, 299308 (2014).

16. Grotzinger, J. P., Watters, W. A. & Knoll, A. H. Calcied metazoans in thrombolite-stromatolite reefs of the terminal Proterozoic Nama Group, Namibia. Paleobiology 26, 334359 (2000).

17. Penny, A. et al. Ediacaran metazoan reefs from the Nama Group, Namibia. Science 344, 15041506 (2014).

18. Wood, R. A., Grotzinger, J. P. & Dickson, J. A. D. Proterozoic modular biomineralized metazoan from the Nama Group, Namibia. Science 296, 23832386 (2002).

19. Gibson, R. & Atkinson, R. Oxygen minimum zone benthos: adaptation and community response to hypoxia. Oceanogr. Marine Biol. Annu. Rev. 41, 145 (2003).

20. Sperling, E. A. et al. Oxygen, facies, and secular controls on the appearance of Cryogenian and Ediacaran body and trace fossils in the Mackenzie Mountains of northwestern Canada. Geol. Soc. Ame. Bull. 128.3-4, 558575 (2015).

21. Johnson, K. S. et al. Manganese ux from continental margin sediments in a transect through the oxygen minimum. Science 257, 12421245 (1992).22. Klinkhammer, G. P. & Bender, M. L. The distribution of manganese in the Pacic Ocean. Earth Planet. Sci. Lett. 46, 361384 (1980).

23. Sholkovitz, E. R., Landing, W. M. & Lewis, B. L. Ocean particle chemistry: the fractionation of rare earth elements between suspended particles and seawater. Geochim. Cosmochim. Acta 58, 15671579 (1994).

24. Ohta, A. & Kawabe, I. REE (III) adsorption onto Mn dioxide (MnO2) and

Fe oxyhydroxide: Ce (III) oxidation by MnO2. Geochim. Cosmochim. Acta 65, 695703 (2001).25. De Carlo, E. H. & Green, W. J. Rare earth elements in the water column of Lake Vanda, McMurdo Dry Valleys, Antarctica. Geochim. Cosmochim. Acta 66, 13231333 (2002).

26. De Baar, H. J., German, C. R., Eldereld, H. & van Gaans, P. Rare earth element distributions in anoxic waters of the Cariaco Trench. Geochim. Cosmochim. Acta 52, 12031219 (1988).

27. Bau, M., Mller, P. & Dulski, P. Yttrium and lanthanides in eastern Mediterranean seawater and their fractionation during redox-cycling. Marine Chem. 56, 123131 (1997).

28. Webb, G. E. & Kamber, B. S. Rare earth elements in Holocene reefal microbialites: a new shallow seawater proxy. Geochim. Cosmochim. Acta 64, 15571565 (2000).

29. Banner, J. L., Hanson, G. N. & Meyers, W. J. Rare earth element andNd isotopic variations in regionally extensive dolomites from the Burlington-Keokuk Formation (Mississippian): implications for REE mobility during carbonate diagenesis. J. Sediment. Res. 58, 415432 (1988).

30. Himmler, T., Bach, W., Bohrmann, G. & Peckmann, J. Rare earth elements in authigenic methane-seep carbonates as tracers for uid composition during early diagenesis. Chem. Geol. 277, 126136 (2010).

31. Tostevin, R. et al. Effective use of cerium anomalies as a redox proxy in carbonate-dominated marine settings. Chem. Geol. 438, 146162 (2016).

32. Raiswell, R. & Caneld, D. E. Sources of iron for pyrite formation in marine sediments. Am. J. Sci. 298, 219245 (1998).

33. Hall, M. et al. Stratigraphy, palaeontology and geochemistry of the late Neoproterozoic Aar Member, southwest Namibia: Reecting environmental controls on Ediacara fossil preservation during the terminal Proterozoic in African Gondwana. Precambrian Res. 238, 214232 (2013).

34. Morris, S. C., Mattes, B. & Chen, M. The early skeletal organism Cloudina: new occurrences from Oman and possibly China. Am. J. Sci. 245260 (1990).

35. Warren, L. V. et al. Corumbella and in situ Cloudina in association with thrombolites in the Ediacaran Itapucumi Group, Paraguay. Terra Nova 23, 382389 (2011).

36. Wood, R. A. Paleoecology of the earliest skeletal metazoan communities: implications for early biomineralization. Earth Sci. Rev. 106, 184190 (2011).

37. Zhuravlev, A. Y., Wood, R. A. & Penny, A. M. Ediacaran skeletal metazoan interpreted as a lophophorate. Proc. R. Soc. B 282, 1818 (2015).

38. Planavsky, N. et al. Rare Earth Element and yttrium compositions of Archean and Paleoproterozoic Fe formations revisited: new perspectives on the signicance and mechanisms of deposition. Geochim. Cosmochim. Acta 74, 63876405 (2010).

39. Slack, J., Grenne, T., Bekker, A., Rouxel, O. & Lindberg, P. Suboxic deep seawater in the late Paleoproterozoic: evidence from hematitic chert and iron formation related to seaoor-hydrothermal sulde deposits, central Arizona, USA. Earth Planet. Sci. Lett. 255, 243256 (2007).

40. Mazumdar, A., Tanaka, K., Takahashi, T. & Kawabe, I. Characteristics of rare earth element abundances in shallow marine continental platform carbonates of Late Neoproterozoic successions from India. Geochem. J. 37, 277289 (2003).

41. Ling, H.-F. et al. Cerium anomaly variations in Ediacaranearliest Cambrian carbonates from the Yangtze Gorges area, South China: implicationsfor oxygenation of coeval shallow seawater. Precambrian Res. 225, 110127 (2013).

42. German, C. R. & Eldereld, H. Rare earth elements in the NW Indian Ocean. Geochim. Cosmochim. Acta 54, 19291940 (1990).

43. Clement, B. G., Luther, III G. W. & Tebo, B. M. Rapid, oxygen-dependent microbial Mn(II) oxidation kinetics at sub-micromolar oxygen concentrations in the Black Sea suboxic zone. Geochim. Cosmochim. Acta 73, 18781889 (2009).

44. Saager, P. M., De Baar, H. J. W. & Burkill, P. H. Manganese and iron in Indian Ocean waters. Geochim. Cosmochim. Acta 53, 22592267 (1989).45. Trefry, J. H., Presley, B. J., Keeney-Kennicutt, W. L. & Trocine, R. P. Distribution and chemistry of manganese, iron, and suspended particulates in Orca Basin. Geo Marine Lett. 4, 125130 (1984).

46. Raiswell, R. & Anderson, T. F. Reactive iron enrichment in sediments deposited beneath euxinic bottom waters: constraints on supply by shelf recycling. Geol. Soc. Lond. Special Publications 248, 179194 (2005).

47. Levin, L. A., Gage, J. D., Martin, C. & Lamont, P. A. Macrobenthic community structure within and beneath the oxygen minimum zone, NW Arabian Sea. Deep Sea Res. 47, 189226 (2000).

48. Savrda, C. E. & Bottjer, D. J. Oxygen-related biofacies in marine strata:an overview and update. Geol. Soc. Lond. Special Publications 58, 201219 (1991).

49. Love, G. D. et al. Fossil steroids record the appearance of Demospongiae during the Cryogenian period. Nature 457, 718721 (2009).

50. Zhang, S. et al. Sufcient oxygen for animal respiration 1,400 million years ago. Proc. Natl Acad. Sci. USA 113, 17311736 (2016).

51. Knoll, A. H. Biomineralization and Evolutionary History. Rev. Mineral. Geochem. 54, 329356 (2003).

52. Saylor, B. Z., Kaufman, A. J., Grotzinger, J. P. & Urban, F. A composite reference section for terminal proterozoic strata of Southern Namibia. SEPM J. Sediment. Res. 68, 12231235 (1998).

53. Saylor, B. Z., Grotzinger, J. P. & Germs, G. J. B. Sequence stratigraphy and sedimentology of the Neoproterozoic Kuibis and Schwarzrand Subgroups (Nama Group), southwestern Namibia. Precambrian Res. 73, 153171 (1995).

54. Grotzinger, J. P., Bowring, S. A., Saylor, B. Z. & Kaufman, A. J. Biostratigraphic and geochronologic constraints on early animal evolution. Science 270, 598604 (1995).

55. Grotzinger, J. & Miller, R. The Nama Group 2 Geological Society of Namibia, 2008).

56. Narbonne, G. et al. The Ediacaran Period. Geol. Time Scale 413435 (2012).57. Germs, G. J. B. Implications of a sedimentary facies and depositional environmental analysis of the Nama group in South West Africa/Namibia. Geol. Soc. South Africa 11, 89114 (1983).

58. Narbonne, G. M., Saylor, B. Z. & Grotzinger, J. P. The youngest Ediacaran fossils from southern Africa. J. Paleontol. 71.06, 953967 (1997).

59. Bau, M. & Dulski, P. Distribution of yttrium and rare-earth elements in the Penge and Kuruman iron-formations, Transvaal Supergroup, South Africa. Precambrian Res. 79, 3755 (1996).

60. Sunda, W. G. & Huntsman, S. A. Effect of sunlight on redox cycles of manganese in the southwestern Sargasso Sea. Deep Sea Res. A Oceanogr. Res. Papers 35, 12971317 (1988).

61. Bau, M., Koschinsky, A., Dulski, P. & Hein, J. R. Comparison of the partitioning behaviours of yttrium, rare earth elements, and titanium between hydrogenetic marine ferromanganese crusts and seawater. Geochim. Cosmochim. Acta 60, 17091725 (1996).

62. German, C. R. & Eldereld, H. Application of the Ce anomaly as a paleoredox indicator: The ground rules. Paleoceanography 5, 823833 (1990).

63. Shields, G. & Webb, G. Has the REE composition of seawater changed over geological time? Chem. Geol. 204.1, 103107 (2004).

64. Pourmand, A., Dauphas, N. & Ireland, T. J. A novel extraction chromatography and MC-ICP-MS technique for rapid analysis of REE, Sc and Y: revising CI-chondrite and Post-Archean Australian Shale (PAAS) abundances. Chem. Geol. 291, 3854 (2012).

65. Bau, M. et al. Discriminating between different genetic types of marine ferro-manganese crusts and nodules based on rare earth elements and yttrium. Chem. Geol. 381, 19 (2014).

66. Haley, B. A., Klinkhammer, G. P. & McManus, J. Rare earth elements in pore waters of marine sediments. Geochim. Cosmochim. Acta 68, 12651279 (2004).

67. Poulton, S. W. & Caneld, D. E. Development of a sequential extraction procedure for iron: implications for iron partitioning in continentally derived particulates. Chem. Geol. 214, 209221 (2005).

68. Caneld, D. E., Raiswell, R., Westrich, J. T., Reaves, C. M. & Berner, R. A. The use of chromium reduction in the analysis of reduced inorganic sulfur in sediments and shales. Chem. Geol. 54, 149155 (1986).

8 NATURE COMMUNICATIONS | 7:12818 | DOI: 10.1038/ncomms12818 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms12818 ARTICLE

69. Clarkson, M. O., Poulton, S. W., Guilbaud, R. & Wood, R. A. Assessing the utility of Fe/Al and Fe-speciation to record water column redox conditions in carbonate-rich sediments. Chem. Geol. 382, 111122 (2014).

70. Turekian, K. K. & Wedepohl, K. H. Distribution of the Elements in Some Major Units of the Earths Crust. Geol. Soc. Am. Bull. 72, 175192 (1961).

Acknowledgements

R.T., R.A.W., G.A.S.Z., S.W.P., R.G., F.B. and A.R.P. acknowledge nancial support from NERCs Co-evolution of Life and the Planet scheme (NE/1005978/1). Support was provided to M.O.C. and A.R.P. through the International Centre for Carbonate Reservoirs (ICCR). F.B. acknowledges support from the Laidlaw Hall fund. We are grateful for access to farms and thank A. Horn of Omkyk, U. Schulze Neuhoff of Ababis, L. and G. Fourie of Zebra River, C. Husselman of Driedornvlagte andL.G. Evereet of Arasab and Swartpunt. We thank Gary Tarbuck and Jim Davy for technical support, and Gerd Winterleitner and Tony Prave for help carrying out eld work.

Author contributions

R.A.W., K.H.H., R.T., A.M.P., F.B. and A.C. collected the samples. M.O.C., R.T., A.M.P. and F.B. prepared the samples. R.T. conceived the project and analysed the samples. R.T. interpreted the Ce anomaly data, after discussions with G.A.S., T.H., R.G. and

S.W.P. R.T. wrote the paper with S.W.P., G.A.S., R.A.W., and R.G. and input from all co-authors.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing nancial interests: The authors declare no competing nancial interest.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Tostevin, R. et al. Low-oxygen waters limited habitable space for early animals. Nat. Commun. 7:12818 doi: 10.1038/ncomms12818 (2016).

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

r The Author(s) 2016

NATURE COMMUNICATIONS | 7:12818 | DOI: 10.1038/ncomms12818 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9