ARTICLE

Received 8 Dec 2014 | Accepted 9 Apr 2015 | Published 18 May 2015 DOI: 10.1038/ncomms8142 OPEN Rise to modern levels of ocean oxygenation coincided with the Cambrian radiation of animals

Xi Chen1, Hong-Fei Ling1, Derek Vance2, Graham A. Shields-Zhou3,4, Maoyan Zhu4, Simon W. Poulton5, Lawrence M. Och3, Shao-Yong Jiang1,6,DaLi1, Lorenzo Cremonese3 & Corey Archer2

The early diversification of animals (B630 Ma), and their development into both motile and macroscopic forms (B575–565 Ma), has been linked to stepwise increases in the oxygenation of Earth’s surface environment. However, establishing such a linkage between oxygen and evolution for the later Cambrian ‘explosion’ (540–520 Ma) of new, energy- sapping body plans and behaviours has proved more elusive. Here we present new molybdenum isotope data, which demonstrate that the areal extent of oxygenated bottom waters increased in step with the early Cambrian bioradiation of animals and eukaryotic phytoplankton. Modern-like oxygen levels characterized the ocean at B521 Ma for the first time in Earth history. This marks the first establishment of a key environmental factor in modern-like ecosystems, where animals benefit from, and also contribute to, the ‘homeostasis’ of marine redox conditions.

1 State Key Laboratory for Mineral Deposits Research, Department of Earth Sciences, School of Earth Sciences and Engineering, Nanjing University, 163 Xianlin Avenue, Nanjing 210023, China. 2 Department of Earth Sciences, Institute of Geochemistry and Petrology, ETH, Zu¨rich CH-8092, Switzerland. 3 Department of Earth Sciences, University College London, Gower Street, London WC1E 6BT, UK. 4 State Key Laboratory of Palaeobiology and Stratigraphy, Nanjing Institute of Geology and Palaeontology, Chinese Academy of Sciences, 39 East Beijing Road, Nanjing 210008, China. 5 School of Earth and Environment, University of Leeds, Leeds LS2 9JT, UK. 6 State Key Laboratory of Geological Processes and Mineral Resources, Department of Resource Science and Engineering, Faculty of Earth Resources, China University of Geosciences, Wuhan 430074, China. Correspondence and requests for materials should be addressed to H.-F.L. (email: hfl[email protected]) or to D.V. (email: [email protected]).

NATURE COMMUNICATIONS | 6:7142 | DOI: 10.1038/ncomms8142 | www.nature.com/naturecommunications 1 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8142

he delay between the origin of animals (B800 Ma, as isotope compositions (see Methods) of well-preserved organic-rich inferred from molecular clocks)1 and their early marine sediments of Cryogenian to early Cambrian age. We also Tdiversification in the Ediacaran Period (635–541 Ma)2 has assess local water column redox conditions using iron speciation been suggested to be due to sluggish oxygenation of the Earth (see Methods) in the same sediments. We find that the tempo of surface3,4. After the ‘Great Oxidation Event’ (GOE5) in the early ocean oxygenation, as reconstructed by the Mo isotope and B Proterozoic ( 2,400–2,100 Ma), atmospheric PO2 stayed within concentration profiles, was in step with the early Cambrian B0.01 and B10% of the present atmospheric level during the bioradiation. mid-Proterozoic (B2,100–800 Ma (refs 6,7)). It was only after the B termination of the Cryogenian glaciations ( 635 Ma) that Results oxygen levels in Earth’s surface environment began to increase 8,9 Geologic setting and samples. Late Neoproterozoic to Cambrian significantly again . Although the lower estimate of Ediacaran successions are well developed in South China. They are fossili- atmospheric PO2 may surpass the minimum oxygen requirement ferous, with several unique biotas in various environments of animals (o0.1% present atmospheric level)4, the deeper ocean 10–12 (Supplementary Note 1) and provide one of the best candidates likely remained predominantly anoxic . As a consequence, O2 13 14 for constraining the co-evolution of marine oxygenation and deficiency ,H2S toxicity and a scarcity of trace metal life9,14,40. We collected black shales and organic-rich cherts from micronutrients (such as Mo, Cu and Zn (ref. 15)) may have shallow-to-deep water successions on the south-eastern margin of continued to limit the ecological distribution of eukaryotes. the Yangtze platform (Supplementary Figs 1 and 2). The ages of Eukaryotes, especially animals and planktonic algae, only our samples span the mid-Cryogenian to the early Cambrian began to dominate the marine ecosystem during the ‘Cambrian (B660–520 Ma). Ediacaran and early Cambrian successions in B 12,16,17 explosion’ of biological diversity ( 520 Ma) . Many the Yangtze Gorges area represent shallow-water facies, while essential aspects of this biotic event, such as increased animal 18 19,20 deeper water facies occur in northwestern Hunan and southern body size , active locomotion, bioturbation , carbonate Anhui provinces (see Supplementary Note 1). biomineralization21, carnivory13,22 and cropping17,23,24, have been linked to a rise in atmospheric oxygen beyond the minimum requirement of animals, and/or widespread ocean Mo isotopes. Our Mo concentration and isotopic data, together oxygenation. However, redox conditions in the early Cambrian with published data from the late Archaean to early Cambrian, oceans, especially the deep ocean, are still controversial. Some are compiled in Supplementary Data 1 and shown in Fig. 1 and studies suggest widespread oxygenation13,25,26, while others Supplementary Fig. 2. After several moderate peaks ( þ 1.5% to propose a ferruginous (Fe2 þ -rich) or even euxinic (H S-rich) þ 1.7%) without a distinct increase in Mo abundance in the late 2 B 41–44 98/95 deep marine environment10,27. Moreover, it has been suggested Archaean ( 2,500–2,750 Ma) , presaging the GOE, d Mo that oceanic oxygen remained at levels much lower than the remained at low levels (o þ 1.3%) from the late B modern until the Devonian28. Palaeoproterozoic ( 2,300 Ma) to the mid-Neoproterozoic (B750 Ma)28,30,41,45–48. In the late Neoproterozoic, higher Here we use sedimentary molybdenum (Mo) isotope composi- 98/95 tions to trace the evolution of global ocean redox state over this d Mo values (up to þ 1.6%) are tied to increased Mo critical period for animal evolution. Global marine redox abundance in the aftermath of the three major glaciations conditions can be inferred from the sedimentary Mo record (Sturtian, Marinoan and Gaskiers), possibly linked to oxygenation events driven by high nutrient inputs from enhanced terrestrial because of its redox-sensitive deposition and isotope fractionation 6,8 98/95 mechanisms9,29–32. In the modern oxic openÀÁ oceans, Mo is present weathering . d Mo values approached 4 þ 1.5% around as the conservative oxyanion molybdate MoO2 À at relatively 550 Ma, and then reached þ 2% for the first time in Earth 4 B high concentrations (its salinity-normalized concentration is history during the earliest Cambrian ( 535 Ma, mid-Fortunian B107 nmol kg À 1)33. The modern open-ocean seawater (OSW) Stage). After that, coinciding with the first occurrence of trilobites Mo reservoir is enriched in heavy isotopes (modern and nearly all animal clades in the major phase of the ‘Cambrian 98/95 explosion’, both high d98/95Mo values (near þ 2.3%) and high d MoOSW ¼þ2.34%, relative to NIST-SRM-3134 (ref. 34)) 98/95 Mo concentrations (4100 p.p.m.) are found in samples from relative to the dominant input from rivers (d MoRivers ¼ þ 0.7% (ref. 35)). This arises because a major sink for Mo in the different locations, and cluster around the Cambrian Stage 2/3 modern oceans is the slow adsorption on particulate manganese boundary (521 Ma). There are four sulphidic samples out of eight 98/95 B oxides under widespread oxic conditions, and because this process samples having high d Mo values (4 þ 1.9%) between 525 and B520 Ma. Although the other half of these eight samples is accompanied by a À 3% isotopic fractionation DSediment–OSW, that is, d98/95Mo À d98/95Mo ¼À3% (refs 29,36)). were deposited under Fe-rich conditions (note that these samples Sediment OSW B still contain appreciable pyrite, with FePy/FeHR40.3 to 0.5), we When bottom-water dissolved oxygen is low or absent, H2Smay 98/95 be present in pore waters of organic-rich reducing sediments (on emphasise that the sediments analysed likely record d Mo some continental margins) or in the water column (for example, close to that of contemporaneous seawater (for the sulphidic Black Sea and regions of intense upwelling). Under these black shales) or a minimum value for seawater (for the conditions, Mo deposition can be accelerated by one to two ferruginous black shales). orders of magnitude37,38, leading to smaller isotopic fractionations DSediment–OSW ¼ 0to À 0.7% (ref. 39)) due to more quantitative Discussion sequestration from the dissolved phase in these settings. An Our data provide an estimate for the lower limit of coeval expansion of such sulphidic conditions will therefore cause seawater Mo isotopic composition and document the fact that 98/95 98/95 the d Mo value of seawater to decrease towards the riverine d MoOSW rose to a level higher than ever before during the input value. Seawater d98/95Mo will most likely be recorded in early Cambrian, peaking at modern levels (B þ 2.3%)at sediments deposited under euxinic conditions because of B520 Ma. The rarity of black shales deposited under fully quantitative removal of aqueous Mo. However, importantly, euxinic conditions makes a continuous record of the precise measured d98/95Mo values of sediments provide a minimum d98/95Mo of seawater difficult to obtain. However, where constraint on contemporaneous seawater isotopic composition sulphidic black shale data (or phosphatic data, which have been because all known sedimentary Mo sinks record suggested also to record seawater values25) are available, their d98/95Morseawater28. We determined Mo concentrations and d98/95Mo data are lower than those at 520 Ma—that is, þ 1.2%

2 NATURE COMMUNICATIONS | 6:7142 | DOI: 10.1038/ncomms8142 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8142 ARTICLE

a 80 Animal Classes diversity Phyla 0 6 Ediacara-type biota Max Bioturbation index Average 0 150 SSF genera Skeletal taxa Weakly biomineralized 0 Eukaryotic MP CJ phytoplankton b 2.34

2 δ 98/95

1 Mo (‰) 0.7 0

−1 Local redox condition 400 Sulfidic 300 Mo (p.p.m.) Ferruginous Anoxic 200 Unknown 100

0 S1S2 S3 S4 Cryogenian Ediacaran Early Cambrian

660 640 620 600 580 560 540 520 c 2.34

2 δ 98/95

1 Mo (‰) 0.7 0

−1 Local redox condition 400 Sulfidic 300 Mo (p.p.m.) Ferruginous Anoxic 200 Oxic Unknown 100

0 3,500 3,000 2,500 2,050 1,600 1,000 541 0 2,300 1,800 1,400 Million-years ago

Figure 1 | Compilation of Mo data together with biodiversity and degree of bioturbation during the Ediacaran–Cambrian transition. (a) Bioturbation indices and diversities of animals, skeletal taxa and eukaryotic phytoplankton1,20,59. MP: mesozooplankton appeared17; CJ: the Chengjiang Lagersta¨tte. (b) Mo data from the mid-Cryogenian to the early Cambrian. In the timescale, three major glaciations are marked as snowflakes, S1 to S4 denote the first four

Cambrian stages. The colour of the data points denotes local redox: sulphidic (FePy/FeHR40.7, red), ferruginous (FePy/FeHRo0.7, orange), anoxic (blue, when Fe speciation data are not available, Fe/Al40.5, trace metal enrichments and other sedimentary characteristics are used to discriminate anoxic conditions) and unknown (grey, no above mentioned data are available, also included are typical carbonates and phosphates, to which Fe-S-C systematics redox proxies cannot easily be applied). The dashed lines mark the average d98/95Mo value of modern seawater ( þ 2.34%) and the riverine input ( þ 0.7%). Data sources, filled circles: this study; open triangles: published data. Mo concentrations of samples from the early Cambrian Ni–Mo ore layer are not shown because of their exceptional enrichment in Mo (in the percent range). The green arrow marks the rising maximal d98/95Mo values. The graded green shading in a,b denotes postulated oxygenation of the ocean from the late Ediacaran to the early Cambrian. (c) Mo data from the

Palaeoarchaean to present (Supplementary Data 1). Symbols as in b, and green colour stands for oxic condition (FeHR/FeTo0.38 and/or Fe/Alo0.5). for 750 Ma (ref. 49), þ 0.2% for B551 Ma (ref. 50 and our data), geochemical records, such as iron speciation and uranium and þ 2% for phosphorites at B535 Ma (ref. 25). The early concentrations (Supplementary Fig. 3). Canfield et al.10 Cambrian marine oxygenation event delineated by maximal established an iron speciation database and found that anoxic d98/95Mo values in black shales is consistent with other ferruginous deep oceans were widespread and persistent during

NATURE COMMUNICATIONS | 6:7142 | DOI: 10.1038/ncomms8142 | www.nature.com/naturecommunications 3 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8142

a 1 b

0.07 98/95Mo 0.8 OSW 0.06 )

wOx 3D-contour of F 0.6 98/95Mo +2.34‰ OSW 0.04 Total + = +2.34‰ Max weakly oxic area A +2‰ / sOx

F +1.8‰ (<2%) wOx /( 0.4 +1.6‰ 0.02 A sOx

F +1.4‰ 0.2 Max euxinic area 0 1 0.8 −4 0.6 (<1%) −3 10 0.4 10−2 10 A 0.2 10−1 0 0.2 0.4 0.6 0.8 1 sOx /A 0 /A Total Total AEux FEux/FTotal

98/95 98/95 98/95 Figure 2 | Model of open ocean seawater d Mo (d MoOSW) in response to different proportions of Mo sinks. (a) Contours of d MoOSW as a function of the relative sizes of the euxinic (FEux), weakly oxic (FwOx) and strongly oxic (FsOx) sinks, assuming that the Mo cycle is in steady state. The black dot represents the modern budget. The upper shaded area would represent coupled expansion of euxinia and the strongly oxic condition relative to the weakly oxic condition, which is unrealistic28. The blue arrow denotes a possible oxygenation trend of the ocean, dictated by the Mo data and the model, 98/95 from the Neoproterozoic to the early Cambrian. (b) A 3D contour surface of d MoOSW ¼þ2.34% as a function of areal (A) fractions of the above redox settings, assuming that the Mo cycle is in steady state. The black dot represents modern redox area distributions; red and yellow dots represent maximum euxinic (1%) and weakly oxic (2%) areas, respectively.

14 later Neoproterozoic times. Li et al. analysed the iron speciation isotope fractionation DwOx–OSW ¼À0.7% (ref. 39)); and a sink systematics of samples obtained from a shore-to-basin transect under euxinic water (H2S is present in bottom waters, average Mo and proposed that sulphidic zones were sandwiched between isotope fractionation DEux–OSW ¼À0.5% (ref. 28)). We also note ferruginous waters on continental margins through the Ediacaran that in the modern oceans there are large areas of moderately oxic Period. Owing to geochemical similarities with Mo, U seafloor (mOx, B14% (ref. 38)), where the sulphidic zone is deep concentrations in sediments can also reflect the redox state of in the sediment column and the bottom-water O2 concentration the global ocean26, although the onset of U enrichment in is higher than or comparable to the weakly oxic condition. In sediments requires less reducing conditions than that of Mo51. these settings Mn oxide-related Mo is quantitatively remobilized Hence, U concentrations in sediments deposited under both from shallow sediments and released back to seawater37,38,54. sulphidic and anoxic non-sulphidic conditions can indicate the Such areas are important but not relevant to the mass balance size of the ocean uranium reservoir, which is proportional to the model since they are neither a sink nor a source for Mo. Our level of ocean oxygenation. Available data show that U modelling results (Fig. 2) suggest that Mo removal under concentrations in anoxic sediments are generally higher for the oxygenated bottom waters (including both strongly and weakly early Cambrian than for the Ediacaran, and also peaked at oxic conditions) must account for more than 94% of the total 98/95 B520 Ma (see Supplementary Fig. 3). Mo sink when d MoOSW reaches modern-like levels 98/95 98/95 Peak d Mo values indicate that oxygenation of the ocean ( þ 2.3%), but can be as low as 33% when d MoOSW is reached modern-like levels for the first time in Earth history at þ 1.6%, which is the highest observed value for the Proterozoic. B 98/95 98/95 520 Ma. The d MoOSW value of the ocean can attain such The modern-like d MoOSW value also requires that the high values under two alternative scenarios: either oxic waters strongly oxic sink accounts for more than 42% of the total 98/95 overwhelmingly dominated the global seafloor in a steady-state oxic sink, while a d MoOSW value of þ 1.6% would Mo cycle, or widespread mildly euxinic waters suddenly allow weakly oxic sink to account for up to 91% of the total consumed the ocean Mo reservoir in a catastrophic hydrogen oxic sink. sulphide-release event27. The latter scenario is inconsistent with The areal proportions of the three redox conditions are further the high d98/95Mo values found in this study for black shales both explored through incorporating Mo accumulation rates (see below and above the peak Mo concentration layer (that is, the Methods). There is a wide range in estimated Mo accumulation À 2 Ni–Mo-enriched layer, see Supplementary Note 1), while the peak rates for modern euxinic settings (F’Eux), from 12,000 mgm itself has only intermediate d98/95Mo values (o þ 1.4% yr À 1 (ref. 32) to 4,800 mgmÀ 2 yr À 1 (ref. 38), and we here (refs 52,53)). As increases in d98/95Mo and U concentrations choose the lower rate to avoid overestimating both the difference both exhibit a long-term trend through the early Cambrian, we between euxinic and oxic sinks, and the oxic area fraction. To apply an improved steady-state mass balance model (see Methods maintain not only mass but also isotope balance of the modern 98/95 À 2 À 1 and below), which demonstrates that d Mo values of ca. oceanic Mo cycle, a F’sOx value of 40 mgm yr , which is þ 2.3% indicate an unprecedentedly high level of marine higher than previous estimate of 27.5 mgmÀ 2 yr À 1 (ref. 38), is oxygenation. We divide the Mo sinks into three types with required. Our modelling results (Fig. 2) indicate that a 98/95 increasing Mo accumulation rates and decreasing Mo isotope d MoOSW value of þ 2.3% requires a limited extent of both fractionation37,38: a sink under strongly oxic water (denoted sOx, euxinic (o1%) and weakly oxic (o2%) areas. Therefore, strongly where O2 penetrates 41 cm below the sediment–water interface, and moderately oxic areas, with relatively high O2 concentrations average Mo isotope fractionation DsOx–OSW ¼À2.95% (ref. 29)); in bottom waters, must have covered 497% of the seafloor at B a sink under weakly oxic water (wOx, low O2 in bottom waters 521 Ma, at least episodically, in comparison with a coverage of 98/95 and H2S exists in organic-rich shallow sediments, average Mo 80% to satisfy a d MoOSW value of þ 1.6%. This oxygenation

4 NATURE COMMUNICATIONS | 6:7142 | DOI: 10.1038/ncomms8142 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8142 ARTICLE event was broadly contemporaneous with the Cambrian bior- Animal ecosystem engineers may also have contributed to adiation (Fig. 1). ocean oxygenation12,17. The early originating suspension-feeding 63 The early Cambrian bioradiation is characterized by the sponges had low O2 requirements , and may have helped emergence of nearly all major bilaterian body plans1. This consume the large dissolved organic carbon reservoir that acted developmental/morphological complexity is a long-term as a major redox buffer in the Proterozoic ocean12. Planktonic consequence of eukaryotic multicellularity, which had animals and algae64 diversified nearly simultaneously in the early originated by the late Mesoproterozoic (B1,200 Ma) and which Cambrian, and likely enhanced the efficiency of the biological diversified during the Ediacaran12,17,55. However, the Ediacara- pump12,17,65, lowering oxygen demand in the water column. type biota (B575–541 Ma) was dominated by soft-bodied sessile Once benthic motile animals became widespread, they deepened epibenthic osmotrophs, and there is no unambiguous evidence O2 penetration depth through bioturbation, which helped for motile bilaterians56,57, except towards the very end of the P-retention and in turn limited primary production and Ediacaran58. From the terminal Ediacaran to the earliest stabilized marine oxygenation12. Cambrian (B580–529 Ma), shallow bioturbation20 and In sum, a pervasively well-oxygenated ocean would have played phosphatic or aragonitic small shelly fossils characterize the a critical role in the ‘Cambrian explosion’, during which newly initial phase of the ‘Cambrian explosion’59. During Cambrian evolved animals and ecosystems affected both carbon and Stage 2 (B529–521 Ma), diverse deeper burrows and complex nutrient cycling, first facilitating and then stabilizing more trace fossils record active bioturbation of sediments and mark the widespread oxygenation in the world’s oceans12,17. Through a so-called ‘substrate/agronomic revolution’20, while small shelly combination of high-resolution palaeoenvironmental and fossils further diversified and calcitic taxa began to appear59,60. palaeobiological studies, we are reaching a more comprehensive Nearly all bilaterian body plans had appeared by early Cambrian understanding of the complex, dynamic interactions and Stage 3 (ref. 1), the major phase of the ‘Cambrian explosion’. feedbacks that helped to bring modern-like ecosystems into being. The apparently abrupt appearance of large, motile and diverse animal forms, following a prolonged and obscure history of Methods phylogenetic evolution, has frequently been explained by changes Samples. Fresh rock samples were powdered using an agate mill. Trace metal and in the physical environment, especially redox conditions1,13, Mo isotope analyses were carried out with an Element II ICP-MS and a Neptune although this conclusion is not without controversy4.Ourdata MC-ICP-MS, respectively, at the University of Bristol. Samples were digested by standard HF-HNO -HCl methods at 180 °C for 472 h. clearly indicate a spatial waning of anoxia (including euxinic and 3 non-sulphidic conditions) in the early Cambrian ocean, which may Mo isotope composition. Mass spectrometry and mass bias correction of Mo have been a pre-condition for the transition to a modern marine isotope data by the double spike technique were carried out as previously descri- ecosystem supporting diversified animals. If anoxia were bed35. Mo isotope data are reported using the d notation, relative to the Mo widespread, even though early animals may have inhabited standard NIST-SRM-3134 (with its d98/95Mo value of þ 0.25% (refs 34,66)) where ! patchy oxic environments, intrusion/upwelling of anoxic water 98 95 14,61 ð Mo= MoÞ would have stifled their success . Diminished euxinia may also d98=95Mo ¼ 1; 000Â Sample À 1 ð1Þ 15 ð98 =95 Þ Â : have relieved the Proterozoic trace metal micronutrient crisis . Mo Mo NIST À SRM À 3134 0 99975 Our data also indicate unprecedentedly widespread oxygena- The internal precision of Mo isotopic measurements ranges between 0.02% and tion in the early Cambrian ocean. The expansion of strongly and 0.05% (2SE of 30 integrations) for all our samples. The external reproducibility of a moderately oxic seafloor from o80% during the Precambrian to series of replicates (entire procedure from digestion of powder onwards) of a black 497% during the early Cambrian may have provided a shale sample is 0.02% (2SD, six replicates). In this study only samples with Mo-enrichment factor (MoEF, which is defined significant increase in the size of habitable space for animals. as the ratio of ([Mo]/[Al])Sample to ([Mo]/[Al])Crust) more than two are plotted, to Animals favour continental marginal areas that have abundant ensure that only the isotopic composition of authigenic Mo is considered. We did calculations to correct the d98/95Mo values for detrital Mo contribution39, assuming nutrients but are also prone to anoxia because of high primary 98/95 productivity and the consequent high O demand. When 497% that the total Mo is a mixture of the detrital (d Mo ¼ 0%) and authigenic Mo. 2 Samples having a difference 40.2% in d98/95Mo between the measured and the of the seafloor was well oxygenated, there must have been vast corrected values were removed. All Mo isotope data are presented without detrital space (460% of the continental margins) for animals to flourish, Mo correction. given that modern continental margins occupy B7% of the world’s seafloor. This expansion of oxygenated seafloor took place Iron speciation. We further filtered our samples to include only those deposited against the backdrop of the rising sea level in the early under anoxic conditions. We accessed the local redox conditions by iron specia- 62 Cambrian , creating numerous new oxygenated ocean margin tion. Iron in carbonate (FeCarb), oxides (FeOx) and magnetite (FeMag)was settings and liberating animals that possibly suffered from sequentially extracted at the Newcastle University using the method described in ref. 67. Iron in pyrite (FePy) is calculated from pyrite sulphur measured by the fluctuating redox conditions on the continental shelf during chromium reduction method, given a stoichiometry for pyrite of FeS2. Total Ediacaran time14,61. Moreover, importantly, the expansion of oxic organic carbon (TOC) contents were measured at the University College London seafloor to a modern extent may imply near-modern average using a Leco C/S analyser after acidification with 6 N HCl. marine oxygen concentrations. Although animals, especially those In general, under an anoxic (O2-free, with or without H2S) water column, highly reactive Fe (FeHR) usually constitutes more than 38% of total Fe (FeT) of sponge-grade and meiofauna, could survive at modest oxygen because of syngenetic formation of either iron sulphides in sulphidic waters or non- 4,63 levels , the large size and ecological dominance of the sulphidized minerals (Fe carbonate, FeIII oxides or magnetite) in ferruginous characteristic Cambrian fauna could not have developed waters. A FePy/FeHR ratio of 0.7 (ref. 11) is used to discriminate between without abundant oxygen. Metabolically expensive behaviours, ferruginous (o0.7) and sulphidic (40.7) anoxia in the water column. such as active locomotion, bioturbation and muscular carnivory, require a high oxygen consumption rate for efficient aerobic Mass balance model of Mo cycle. The parameters used in the model are listed in respiration13,20. Moreover, oxygenation would have favoured the Supplementary Table 1. We divide the Mo sinks into a strongly oxic sink (FsOx, bottom-water dissolved oxygen 410 mM, where O2 penetrates 41 cm below aerobic over anaerobic respiration in the deep ocean, resulting in the sediment–water interface), a weakly oxic sink (FwOx, bottom-water dissolved a decrease in alkalinity and thus suppressing carbonate oxygen o10 mM, with a shallow sulphidic zone in the reducing sediments) 21 37,38 deposition on/in the seafloor . This, consequently, would have and a euxinic sink (FEux,H2S is present in bottom water) . The moderately oxic seafloor, with a deep sulphidic zone in the sediment column, is not increased surface ocean carbonate supersaturation, which may relevant to the global Mo budget because Mn oxide-related Mo is quantitatively have reduced the physiological cost of carbonate skeleton released back into the water column during reduction. Because DEux–OSW 13 98/95 98/95 construction and facilitated the evolutionary arms race . ( ¼ d MoEux À d MoOSW) approaches 0 only if bottom-water [H2S]aq is

NATURE COMMUNICATIONS | 6:7142 | DOI: 10.1038/ncomms8142 | www.nature.com/naturecommunications 5 & 2015 Macmillan Publishers Limited. All rights reserved. ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8142

greater than 11 mM, an average fractionation of À 0.5% (DEux À OSW) is used for the 17. Butterfield, N. J. Macroevolution and macroecology through deep time. 28 98/95 euxinic sink . The Mo isotope composition of open ocean seawater (d MoOSW) Palaeontology 50, 41–55 (2007). is determined by the fractions of the fluxes to sediments in the three redox settings 18. Payne, J. L. et al. Two-phase increase in the maximum size of life over 3.5 (fEux, fwOx and fsOx) with the assumption of steady-state, and setting the input to the billion years reflects biological innovation and environmental opportunity. ocean at the modern value of the riverine flux: Proc. Natl Acad. Sci. USA 106, 24–27 (2009). 19. Bottjer, D. J., Hagadorn, J. W. & Dornbos, S. Q. The Cambrian substrate fEux þ fsOx þ fwOx ¼ fRivers ¼ 1 ð2Þ revolution. GSA Today 10, 1–7 (2000). Isotope mass balance is described by: 20. Ma´ngano, M. G. & Buatois, L. A. Decoupling of body-plan diversification and ecological structuring during the Ediacaran-Cambrian transition: evolutionary dRivers ¼ dEuxÂfEux þ dwOxÂfwOx þ dsOxÂfsOx ð3Þ and geobiological feedbacks. Proc. R Soc. B 281, 20140038 (2014). 98/95 where each dOutput equals (d MoOSW þ DOutput). 21. Higgins, J. A., Fischer, W. W. & Schrag, D. P. Oxygenation of the ocean and We define: sediments: consequences for the seafloor carbonate factory. Earth Planet Sci. f Lett. 284, 25–33 (2009). k ¼ sOx ð4Þ 22. Vermeij, G. J. The evolutionary interaction among species: selection, escalation, sOx f þ f sOx wOx and coevolution. Annu. Rev. Ecol. Syst. 25, 219–236 (1994). 98/95 and solve the above three equations to get the d MoOSW as a function of ksOx 23. Stanley, S. M. An ecological theory for the sudden origin of multicellular 98/95 and fEux, and then plot the contours of d MoOSW in Fig. 2. life in the late Precambrian. Proc. Natl Acad. Sci. USA 70, 1486–1489 À 2 À 1 The Mo output rates (F’Output,gm yr ) in various redox settings are (1973). assumed to be controlled by first-order kinetics with respect to the coeval Mo 24. Butterfield, N. J. Plankton ecology and the Proterozoic-Phanerozoic transition. reservoir in the open ocean (R), that is, each F’Output ¼ F’Output0 Â R/R0 (subscript 0 Paleobiology 23, 247–262 (1997). denotes the modern value). Replacing each fOutput ( ¼ FOutput/FRivers) in the isotope 25. Wen, H. et al. Molybdenum isotopic records across the Precambrian-Cambrian mass balance equation (3) with: boundary. Geology 39, 775–778 (2011).

0 A 26. Partin, C. A. et al. Large-scale fluctuations in Precambrian atmospheric and ðF ÂR=R0ÞÂðATotalÂf Þ Output0 Output ð5Þ oceanic oxygen levels from the record of U in shales. Earth Planet Sci. Lett. FRivers 369-370, 284–293 (2013). 98/95 27. Wille, M., Nagler, T. F., Lehmann, B., Schroder, S. & Kramers, J. D. Hydrogen together with the mass balance equation (2), we can get the d MoOSW as a A A A sulphide release to surface waters at the Precambrian/Cambrian boundary. function of areal fractions of the three redox settings fEux, fwOx and fsOx. In our sensitivity analyses (see Supplementary Fig. 4) we first investigated the Nature 453, 767–769 (2008). 28. Dahl, T. W. et al. Devonian rise in atmospheric oxygen correlated to the impact of setting both DEux À OSW and dRivers to zero. This results in even smaller areas of euxinia and greater predominance of oxic bottom waters for modern-like radiations of terrestrial plants and large predatory fish. Proc. Natl Acad. Sci. 98/95 USA 107, 17911–17915 (2010). d MoOSW, including a major contribution of strongly oxic conditions. In modern weakly oxic settings, although the uppermost few centimetres of 29. Barling, J., Arnold, G. L. & Anbar, A. D. Natural mass-dependent variations in sediments show heterogeneous DSediment–OSW ranging from À 1.2% to À 0.2%, the isotopic composition of molybdenum. Earth Planet Sci. Lett. 193, 447–457 the DSediment–OSW of deep sediments converge on a common value of À 0.7%, (2001). 39 which more likely represents the true average DwOx–OSW over long timescales . 30. Arnold, G. L., Anbar, A. D., Barling, J. & Lyons, T. W. Molybdenum isotope However, we further investigated the impact of uncertainty in DwOx–OSW. A smaller evidence for widespread anoxia in Mid-Proterozoic oceans. Science 304, 87–90 fractionation, such as DwOx–OSW of À 0.2%, results in a similar situation to the (2004). previous sensitivity analyses, that is, even greater contributions are required from 31. Ling, H.-F., Gao, J.-F., Zhao, K.-D., Jiang, S.-Y. & Ma, D.-S. Comment on 98/95 oxic sinks for modern-like d MoOSW. Although DwOx–OSW of À 1.2% would ‘Molybdenum isotope evidence for widespread anoxia in Mid-Proterozoic 98/95 allow smaller contributions from oxic sinks, modern-like d MoOSW still oceans’. Science 309, 1017c (2005). requires/reflects a predominance of oxic sinks. 32. Scott, C. et al. Tracing the stepwise oxygenation of the Proterozoic ocean. Nature 452, 456–459 (2008). References 33. Nakagawa, Y. et al. The molybdenum isotopic composition of the modern 1. Erwin, D. H. et al. The Cambrian conundrum: early divergence and later ocean. Geochem. J. 46, 131–141 (2012). ecological success in the early history of animals. Science 334, 1091–1097 34. Goldberg, T. et al. Resolution of inter-laboratory discrepancies in Mo isotope (2011). data: an intercalibration. J. Anal. At. Spectrom. 28, 724–735 (2013). 2. Xiao, S. & Laflamme, M. On the eve of animal radiation: phylogeny, ecology 35. Archer, C. & Vance, D. The isotopic signature of the global riverine and evolution of the Ediacara biota. Trends Ecol. Evol. 24, 31–40 (2009). molybdenum flux and anoxia in the ancient oceans. Nat. Geosci. 1, 597–600 3. Knoll, A. H. & Carroll, S. B. Early animal evolution: emerging views from (2008). comparative biology and geology. Science 284, 2129–2137 (1999). 36. Siebert, C., Na¨gler, T. F., von Blanckenburg, F. & Kramers, J. D. Molybdenum 4. Mills, D. B. & Canfield, D. E. Oxygen and animal evolution: did a rise of isotope records as a potential new proxy for paleoceanography. Earth Planet atmospheric oxygen ‘trigger’ the origin of animals? Bioessays 36, 1145–1155 Sci. Lett. 211, 159–171 (2003). (2014). 37. Scott, C. & Lyons, T. W. Contrasting molybdenum cycling and isotopic 5. Holland, H. D. The oxygenation of the atmosphere and oceans. Philos. Trans. R properties in euxinic versus non-euxinic sediments and sedimentary rocks: Soc. B 361, 903–915 (2006). refining the paleoproxies. Chem. Geol. 324-325, 19–27 (2012). 6. Lyons, T. W., Reinhard, C. T. & Planavsky, N. J. The rise of oxygen in Earth’s 38. Reinhard, C. T. et al. Proterozoic ocean redox and biogeochemical stasis. Proc. early ocean and atmosphere. Nature 506, 307–315 (2014). Natl Acad. Sci. USA 110, 5357–5362 (2013). 7. Planavsky, N. J. et al. Low Mid-Proterozoic atmospheric oxygen levels and the 39. Poulson Brucker, R. L., McManus, J., Severmann, S. & Berelson, W. M. delayed rise of animals. Science 346, 635–638 (2014). Molybdenum behavior during early diagenesis: Insights from Mo isotopes. 8. Canfield, D. E., Poulton, S. W. & Narbonne, G. M. Late-Neoproterozoic deep- Geochem. Geophys. Geosyst. 10, Q06010 (2009). ocean oxygenation and the rise of animal life. Science 315, 92–95 (2007). 40. McFadden, K. A. et al. Pulsed oxidation and biological evolution in the Ediacaran 9. Sahoo, S. K. et al. Ocean oxygenation in the wake of the Marinoan glaciation. Doushantuo Formation. Proc. Natl Acad. Sci. USA 105, 3197–3202 (2008). Nature 489, 546–549 (2012). 41. Wille, M. et al. Evidence for a gradual rise of oxygen between 2.6 and 2.5 Ga 10. Canfield, D. E. et al. Ferruginous conditions dominated later Neoproterozoic from Mo isotopes and Re-PGE signatures in shales. Geochim. Cosmochim. Acta deep-water chemistry. Science 321, 949–952 (2008). 71, 2417–2435 (2007). 11. Poulton, S. W. & Canfield, D. E. Ferruginous conditions: a dominant feature of 42. Voegelin, A. R., Na¨gler, T. F., Beukes, N. J. & Lacassie, J. P. Molybdenum the ocean through Earth’s history. Elements 7, 107–112 (2011). isotopes in late Archean carbonate rocks: implications for early Earth 12. Lenton, T. M., Boyle, R. A., Poulton, S. W., Shields-Zhou, G. A. & Butterfield, oxygenation. Precambrian Res. 182, 70–82 (2010). N. J. Co-evolution of eukaryotes and ocean oxygenation in the Neoproterozoic 43. Duan, Y. et al. Molybdenum isotope evidence for mild environmental era. Nat. Geosci. 7, 257–265 (2014). oxygenation before the Great Oxidation Event. Geochim. Cosmochim. Acta 74, 13. Sperling, E. A. et al. Oxygen, ecology, and the Cambrian radiation of animals. 6655–6668 (2010). Proc. Natl Acad. Sci. USA 110, 13446–13451 (2013). 44. Cabral, A. R. et al. Trace-element and multi-isotope geochemistry of Late- 14. Li, C. et al. A stratified redox model for the Ediacaran ocean. Science 328, 80–83 Archean black shales in the Caraja´s iron-ore district, Brazil. Chem. Geol. 362, (2010). 91–104 (2013). 15. Anbar, A. D. & Knoll, A. H. Proterozoic ocean chemistry and evolution: a 45. Siebert, C., Kramers, J. D., Meisel, T., Morel, P. & Na¨gler, T. PGE, Re-Os, and bioinorganic bridge? Science 297, 1137–1142 (2002). Mo isotope systematics in Archean and early Proterozoic sedimentary systems 16. Butterfield, N. J. Animals and the invention of the Phanerozoic Earth system. as proxies for redox conditions of the early Earth. Geochim. Cosmochim. Acta Trends Ecol. Evol. 26, 81–87 (2011). 69, 1787–1801 (2005).

6 NATURE COMMUNICATIONS | 6:7142 | DOI: 10.1038/ncomms8142 | www.nature.com/naturecommunications & 2015 Macmillan Publishers Limited. All rights reserved. NATURE COMMUNICATIONS | DOI: 10.1038/ncomms8142 ARTICLE

46. Kendall, B., Creaser, R. A., Gordon, G. W. & Anbar, A. D. Re-Os and Mo 65. Logan, G. A., Hayes, J. M., Hieshima, G. B. & Summons, R. E. isotope systematics of black shales from the Middle Proterozoic Velkerri and Terminal Proterozoic reorganization of biogeochemical cycles. Nature 376, Wollogorang Formations, McArthur Basin, northern Australia. Geochim. 53–56 (1995). Cosmochim. Acta 73, 2534–2558 (2009). 66. Na¨gler, T. F. et al. Proposal for an international molybdenum isotope 47. Kendall, B., Gordon, G. W., Poulton, S. W. & Anbar, A. D. Molybdenum measurement standard and data representation. Geostand. Geoanal. Res. 38, isotope constraints on the extent of late Paleoproterozoic ocean euxinia. Earth 149–151 (2014). Planet Sci. Lett. 307, 450–460 (2011). 67. Poulton, S. W. & Canfield, D. E. Development of a sequential extraction 48. Asael, D. et al. Coupled molybdenum, iron and uranium stable isotopes as procedure for iron: implications for iron partitioning in continentally derived oceanic paleoredox proxies during the Paleoproterozoic Shunga Event. Chem. particulates. Chem. Geol. 214, 209–221 (2005). Geol. 362, 193–210 (2013). 49. Dahl, T. W. et al. Molybdenum evidence for expansive sulfidic water masses in Acknowledgements B750 Ma oceans. Earth Planet Sci. Lett. 311, 264–274 (2011). 50. Kendall, B., Anbar, A. D., Gordon, G., Arnold, G. L. & Creaser, R. A. This work was partially supported by the National Key Basic Research Program of China Constraining the redox state of the Proterozoic deep oceans using the Mo Grant 2013CB835004 (to H.-F.L. and M.Z.), National Natural Science Foundation of isotope systematics of euxinic black shales. In Geological Society of America China Grand 41230102 (to S.-Y.J. and H.-F.L.) and 40872025 (to H.-F.L.), Sino-German Abstracts with Programs vol. 38, 56 (2006). Geoscientific Research Project FOR 736 (to G.A.S.-Z. and L.M.O.), Chinese Academy of 51. Algeo, T. J. & Tribovillard, N. Environmental analysis of paleoceanographic Sciences (visiting fellowship) and the 111 Project (to G.A.S.-Z.), and China Scholarship systems based on Molybdenum-Uranium covariation. Chem. Geol. 268, Council (to X.C.). We also thank Chris Coath (BIG) for his help with MC-ICP-MS; Tony 211–225 (2009). Osborn (UCL) for TOC analyses; Dan Wang and Tao Yang for their assistance in 52. Lehmann, B. et al. Highly metalliferous carbonaceous shale and Early laboratory work. X.C. also takes this opportunity to thank Qi Zhao, Ye Zhao, Jin Chen Cambrian seawater. Geology 35, 403–406 (2007). and Hu Xiao for their help during his stay in Bristol. 53. Xu, L., Lehmann, B. & Mao, J. Seawater contribution to polymetallic Ni-Mo- PGE-Au mineralization in Early Cambrian black shales of South China: Author contributions evidence from Mo isotope, PGE, trace element, and REE geochemistry. Ore H.-F.L., X.C., D.L., G.A.S.-Z., L.M.O., L.C., S.-Y.J. and M.Z. collected samples in the field. Geol. Rev. 52, 66–84 (2013). X.C., C.A. and D.V. carried out Mo isotopes and trace element analysis. X.C., L.M.O., 54. Morford, J. L. & Emerson, S. The geochemistry of redox sensitive trace metals D.L., S.W.P. and G.A.S.-Z. analysed Fe speciation and TOC. X.C., H.-F.L. and G.A.S.-Z. in sediments. Geochim. Cosmochim. Acta 63, 1735–1750 (1999). designed the study and prepared the original manuscript. X.C., H.-F.L., D.V., G.A.S.-Z., 55. Knoll, A. H., Javaux, E. J., Hewitt, D. & Cohen, P. Eukaryotic organisms in S.W.P. and M.Z. prepared the revised manuscript. All authors discussed the results and Proterozoic oceans. Philos. Trans. R Soc. B 361, 1023–1038 (2006). commented on the manuscript. 56. Gaucher, C., Poire´, D. G., Bossi, J., Bettucci, L. S. & Beri, A´ . Comment on ‘bilaterian burrows and grazing behavior at 4585 million years ago’. Science 339, 906 (2013). Additional information 57. Retallack, G. J. First evidence for locomotion in the Ediacara biota from the Supplementary Information accompanies this paper at http://www.nature.com/ 565 Ma Mistaken Point Formation, Newfoundland: COMMENT. Geology 38, naturecommunications e223 (2010). Competing financial interests: The authors declare no competing financial interests. 58. Fedonkin, M. A. & Waggoner, B. M. The Late Precambrian fossil Kimberella is a mollusc-like bilaterian organism. Nature 388, 868–871 (1997). Reprints and permission information is available online at http://npg.nature.com/ 59. Maloof, A. C. et al. The earliest Cambrian record of animals and ocean reprintsandpermissions/ geochemical change. GSA Bull. 122, 1731–1774 (2010). 60. Porter, S. M. Seawater chemistry and Early carbonate biomineralization. Science How to cite this article: Chen, X. et al. Rise to modern levels of ocean oxygenation 316, 1302 (2007). coincided with the Cambrian radiation of animals. Nat. Commun. 6:7142 61. Johnston, D. T. et al. Late Ediacaran redox stability and metazoan evolution. doi: 10.1038/ncomms8142 (2015). Earth Planet Sci. Lett. 335-336, 25–35 (2012). 62. Peters, S. E. & Gaines, R. R. Formation of the ‘Great Unconformity’ as a trigger This work is licensed under a Creative Commons Attribution 4.0 for the Cambrian explosion. Nature 484, 363–366 (2012). International License. The images or other third party material 63. Mills, D. B. et al. Oxygen requirements of the earliest animals. Proc. Natl Acad. in this article are included in the article’s Creative Commons license, unless indicated Sci. USA 111, 4168–4172 (2014). otherwise in the credit line; if the material is not included under the Creative 64. Vidal, G. & Moczydlowska-Vidal, M. Biodiversity, speciation, and extinction Commons license, users will need to obtain permission from the license holder to trends of Proterozoic and Cambrian phytoplankton. Paleobiology 23, 230–246 reproduce the material. To view a copy of this license, visit http://creativecommons.org/ (1997). licenses/by/4.0/

NATURE COMMUNICATIONS | 6:7142 | DOI: 10.1038/ncomms8142 | www.nature.com/naturecommunications 7 & 2015 Macmillan Publishers Limited. All rights reserved.