barite,bifsandbugs:evidencefortheevolutionof theearth...

Barite, BIFs and bugs: evidence for the evolution ofthe Earth’s early hydrosphere§

David L. Huston �, Graham A. LoganGeoscience Australia, GPO Box 378, Canberra, ACT 2601, Australia

Received 16 June 2003; received in revised form 19 November 2003; accepted 7 January 2004

Abstract

The presence of relatively abundant bedded sulfate deposits before 3.2 Ga and after 1.8 Ga, the peak in ironformation abundance between 3.2 and 1.8 Ga, and the aqueous geochemistry of sulfur and iron together suggest thatthe redox state and the abundances of sulfur and iron in the hydrosphere varied widely during the Archean andProterozoic. We propose a layered hydrosphere prior to 3.2 Ga in which sulfate produced by atmospheric photolyticreactions was enriched in an upper layer, whereas the underlying layer was reduced and sulfur-poor. Between 3.2 and2.4 Ga, sulfate reduction removed sulfate from the upper layer, producing broadly uniform, reduced, sulfur-poor andiron-rich oceans. As a result of increasing atmospheric oxygenation around 2.4 Ga, the flux of sulfate into thehydrosphere by oxidative weathering was greatly enhanced, producing layered oceans, with sulfate-enriched, iron-poor surface waters and reduced, sulfur-poor and iron-rich bottom waters. The rate at which this process proceededvaried between basins depending on the size and local environment of the basin. By 1.8 Ga, the hydrosphere wasrelatively sulfate-rich and iron-poor throughout. Variations in sulfur and iron abundances suggest that the redox stateof the oceans was buffered by iron before 2.4 Ga and by sulfur after 1.8 Ga.Crown Copyright ; 2004 Elsevier B.V. All rights reserved.

Keywords: hydrosphere evolution; atmosphere evolution; Archean; Proterozoic; biogeochemistry

1. Introduction

Variations in the amount of free atmosphericoxygen through geologic time have been the sub-

ject of considerable debate since Cloud [1] sug-gested that the Archean atmosphere containedmuch less oxygen than at present. He observedthat uraninite and pyrite, unstable in an oxygen-rich atmosphere, are present as detrital grains inArchean conglomerates. Subsequent work on Ar-chean and early Proterozoic soil pro¢les indicatesthat iron, immobile in an oxygen-rich atmosphere,was mobile [2,3]. Although disputed by some [4],most workers accept that these geologic observa-tions indicate that Archean atmospheric oxygenlevels were substantially lower than at present[1,2].

0012-821X / 04 / $ ^ see front matter Crown Copyright ; 2004 Elsevier B.V. All rights reserved.doi:10.1016/S0012-821X(04)00034-2

* Corresponding author. Tel. : +62-2-6249-9577;Fax: +62-2-6249-9971.E-mail addresses: [email protected] (D.L. Huston),

[email protected] (G.A. Logan).

§ Supplementary data associated with this article can befound at doi:10.1016/S0012-821X(04)00034-2.

EPSL 6994 5-3-04

Earth and Planetary Science Letters 220 (2004) 41^55

R

Available online at www.sciencedirect.com

www.elsevier.com/locate/epsl

mailto:[email protected]:[email protected]://www.elsevier.com/locate/epsl

It is tempting to infer that because Archeanatmospheric oxygen levels were low, the coexist-ing hydrosphere was reduced. However, the pres-ence of syngenetic sulfate in the early Archeansuggests at least local oxidizing redox conditionsin the hydrosphere1 [5]. These deposits, alongwith the abundance of banded iron formation(BIF), an indicator of reduced bottom waters [6]in the late Archean and early Proterozoic, suggestthe redox state of seawater may have varied sub-stantially.A substantial body of literature exists regarding

the evolution of the atmosphere and hydrosphereduring the early part of the Earth’s history, atopic that still receives a large amount of research,and perhaps more than its fair share of contro-versy. In the pursuit of research interests outsidethis general ¢eld, we have inferred a number ofgeological and geochemical anomalies that are notexplained by current models for the evolution ofthe Earth’s atmosphere and, particularly, hydro-sphere. As discussed below, these anomalies in-clude the abundance of bedded barite deposits,and the temporal distribution and geochemistryof BIFs.In this paper we use temporal distributions of

sulfate deposits and iron formations along withthe aqueous geochemistry of iron, sulfur, and ba-rium to present a unifying concept for the evolu-tion of the Earth’s hydrosphere before 1.0 Ga.This model builds on previous models, butpresents important, although in some cases subtle,di¡erences in an attempt to account for theanomalies mentioned above. The model also pla-ces quantitative constraints on variations in theabundance and redox state of aqueous compo-nents such as sulfur and iron, parameters thatare not well established in many current models.Although previous models invoking layered

oceans [7,8] partly explain our observations, thedata suggest a more complex evolution of the hy-drosphere through time. We intend to show thatchanges in the oxidation state of the atmosphereand the resulting impact on weathering had aprofound and perhaps unexpected impact on thehydrosphere. In particular, we hope to show thatchanges occur in both availability of sulfate andiron through time and that the dominant multi-valent ions that help bu¡er oxidation states of thehydrosphere varied in a surprising way.

2. The temporal distributions of sulfate and ironformation

The geologic record contains various sedimentsand deposits that re£ect the oxidation state of thehydrosphere and/or the atmosphere. We have re-viewed the temporal and spatial occurrences ofthe key sediment types (see Tables 1 and 22), thedetails of which are summarized below. Fig. 1,which is based on Tables 1 and 22, illustrates thedistributions of bedded sulfate deposits and BIFsprior to 1.0 Ga. These tables were compiled fromexisting literature. Table 22 is extensively updatedfrom compilations by James [9], Walker et al. [2],Isley [10] and Isley and Abbott [11]. It providesmore constrained ages for many BIFs and in-cludes data for BIFs not presented in the previouscompilations. In Table 22, BIFs are classi¢ed aseither Algoma- or Superior-type. Algoma-typeBIFs are generally small in size and associatedwith contemporaneous volcanic suites. In con-trast, Superior-type BIFs (Hamersley-type in Aus-tralia) are laterally extensive, generally in a shelfenvironment without a volcanic association [12].Before 3.2 Ga, bedded barite deposits are rela-

tively common (18 in Table 12), whereas otherbedded sulfates are not known and BIFs, al-though present, are not common and are exclu-sively of Algoma-type. Between 3.2 and 2.4 Ga,the geologic record is characterized by abundantBIF and a paucity of bedded sulfate deposits(only two known). The period between 2.4 and

1 Unless stated otherwise, comments relating to redox con-ditions refer only to the hydrosphere. No inference should bemade about atmospheric conditions, unless speci¢cally stated.By the term ‘oxidizing’ we imply that sulfate was the dominanthydrous sulfur species. By the term ‘reduced’ we imply thatH2S and/or HS3 were the dominant sulfur species. Under dis-equilibrium conditions with intermediate redox, sul¢te may beimportant. 2 See the online version of this article.

EPSL 6994 5-3-04

D.L. Huston, G.A. Logan / Earth and Planetary Science Letters 220 (2004) 41^5542

1.8 Ga is characterized by abundant BIF, andincreasingly abundant sulfate deposits, with gyp-sum and anhydrite becoming more common. Be-tween 3.0 and 1.8 Ga, both Superior- and Algo-ma-type BIFs are present, but the Superior-typedeposits account for the vast majority of BIF. Theperiod between 1.8 and 1.0 Ga lacks BIF butcontains abundant sulfate deposits. AlthoughBIFs of late Proterozoic and Phanerozoic agesare known (e.g. Rapitan), these deposits are theproducts of either unusual geologic conditions(e.g. global glaciation [13]) or hydrothermal vent-ing.As both the style and the abundance of BIF

vary with time, it is unlikely that these variationsare solely controlled by tectonic preservation. Inaddition, the true frequency of bedded sulfate de-posits is likely to be understated in the older rocksas these rocks are less likely to be preserved[14].

3. Sulfur isotope variations through time

Prior to 2.4 Ga, N34Ssulfate is relatively uniform(Fig. 1) with average values for all deposits inthe range of 3.8^5.4x (all N34S values are re-ported relative to the CDT standard), bar the BigStubby and Geco volcanic-hosted massive sul¢de(VHMS) deposits. Moreover, with the exceptionof the Dresser deposits [15], coexisting sul¢deminerals have a narrow range of 0P 4x [16].Sedimentary sul¢de elsewhere has N34S in therange of 0P 10x [13,17]. In contrast, depositsyounger than 2.4 Ga have higher and more vari-able N34Ssulfate, ranging from 4 to 39x (Fig. 1),similar to values in the late Proterozoic and Phan-erozoic [18]. Sedimentary sul¢des also have alarge range in N34S, from 330 to 60x [13,19,20].Farquhar et al. [21] showed that the fractiona-

tion of 33S relative to 32S also changes at about2.4 Ga. Syn- or diagenetic sulfate and sul¢deminerals deposited before 2.45 Ga show mass-independent fractionation of 33S, interpreted toresult from gas-phase photolysis of SO2, yieldinga large range of oxidized and reduced sulfur-bear-ing species [22]. Between 2.45 and 2.09 Ga 33Sfractionation shifted from mass independent tomass dependent. The lack of mass-independentfractionation after 2.09 Ga suggests that photo-chemical reactions ceased to be important, withbiogenic and redox reactions controlling sulfurspeciation. These observations have been indepen-dently supported by Mojzsis et al. [23].

4. The low-temperature geochemistry of sulfur,iron and barium

Fig. 2 illustrates the stability of Fe^S^O miner-als and barite, and the solubility of iron as afunction of sulfur content and redox at pH (7.8)and salinity of modern seawater (3% NaCl).These diagrams were calculated for 25 and 75‡Cto illustrate constraints on deposition of BIF andbarite.At 25 and 75‡C, magnetite is only stable at

sulfur levels below 1035 and 1033 that of modernseawater. Iron is most soluble in the magnetitestability ¢eld and least soluble in the pyrite ¢eld.

Fig. 1. Variations in the abundances of bedded sulfate depos-its and BIF and of N34S through time (based on Tables 1and 22).

EPSL 6994 5-3-04

D.L. Huston, G.A. Logan / Earth and Planetary Science Letters 220 (2004) 41^55 43

In the magnetite ¢eld, iron solubility locally ex-ceeds 100 ppm and decreases as gSO4/gH2S in-creases. Even if slow reaction rates inhibit pyriteformation, and the magnetite ¢eld expands tohigher sulfur levels (dashed lines in Fig. 2), sulfurconcentrations must still be much lower than thatof modern seawater. Under reduced conditions atcurrent seawater sulfur levels, the maximum ironsolubility in the pyrite ¢eld is only V1033 ppm,some ¢ve orders of magnitude lower than that inthe magnetite ¢eld. Although iron solubility ex-ceeds 1 ppm at the base of the hematite ¢eld, itdecreases with increasing gSO4/gH2S. Therefore,under sulfur-rich, highly oxidized conditions char-acteristic of modern seawater, iron is highly in-soluble. Magnetite is only stable under reducedconditions (gSO4/gH2S6 1032:5 at 25‡C), andhigh iron solubility is only possible under verylow ambient sulfur concentrations.In the presence of even low concentrations of

aqueous sulfate, barium is insoluble, forming bar-

ite. Therefore, barite deposition indicates rela-tively oxidized conditions (gSO4/gH2Ss 1032),even for £uids with low sulfur levels (at most1033 current seawater concentrations). In con-trast, the solubilities of gypsum and anhydriteare much higher: a high degree of evaporationis required to precipitate gypsum. Although bariteis the only sulfate mineral reported prior to 3.2Ga, the initial workers on the Dresser barite de-posit suggested that the barite replaced originalevaporative gypsum [25,26]. However, recentwork [27] suggests that most sulfate precipitatedoriginally as barite from low-temperature hydro-thermal emanations analogous to white smokers.Thus, the distributions of bedded sulfate min-

erals and iron formations plus constraints basedon the aqueous geochemistry of iron, sulfur andbarium set limits for gSO4/gH2S, total sulfur, andiron abundances in ancient seawater. These datarequire that before 1.8 Ga, ancient seawater hadmuch lower total sulfur and higher iron abundan-ces than modern seawater. Furthermore, total sul-fate levels between 3.2 and 2.4 Ga must have beenextremely low, although before 3.2 Ga sulfate lev-els may locally have approached modern levels.Variations in the abundances of the elementsmay be linked to the evolution of the atmosphereand hydrosphere, as suggested by Walker andBrimblecombe [28].

5. The origin of bedded sulfate deposits

Interpretation of the genetic signi¢cance of thedistribution of bedded sulfate deposits throughgeologic time requires an understanding of howthese deposits formed. Of particular interest arethe barite deposits : barite is the dominant or onlysulfate mineral in bedded sulfate deposits formedprior to 2.4 Ga.Because of the highly insoluble nature of barite,

formation of this mineral requires two discretesources, one for barium and another for sulfate.Jewell [29] suggests that barite can form in threeways: (1) diagenetic replacement, (2) hydrother-mal exhalation, or (3) biological precipitation.The ¢rst two mechanisms are most important,particularly in the Archean.

Fig. 2. gSO4/gH2S versus total sulfur diagrams calculated at25 and 75‡C at modern oceanic pH and salinity showing Fe^Ba^S^O mineral stabilities and Fe solubilities (calculated us-ing data generated from HCh [24]). Total sulfur concentra-tions are normalized to the level in modern open ocean sea-water.

EPSL 6994 5-3-04


Diagenetic replacement of original bedded gyp-sum has been proposed at Dresser in the PilbaraCraton and several occurrences in the KaapvaalCraton [25]. However, the most common originfor barite in the Archean appears to be hydro-thermal exhalation. Barite precipitates when a re-duced £uid carrying Ba2þ mixes with a £uid car-rying SO234 . In Phanerozoic VHMS deposits,barium-bearing hydrothermal £uid mixed withsulfate-bearing seawater at or just below the sea-£oor. As a consequence, barite concentrated to-ward the top of VHMS deposits [30].It is tempting to draw a direct analogy between

Phanerozoic VHMS barite and Archean barite,but given the controversy regarding the redoxstate of the Archean hydrosphere, other alterna-tives should be considered. The most likely alter-native is that an oxidized, sulfate-bearing hydro-thermal £uid mixed with barium-bearing, reducedseawater. Hydrothermal sulfate can be derived ei-ther from leaching of sulfate in evaporative, sedi-mentary rocks or by disproportionation of mag-matic SO2 [31]. As evaporative sulfate mineralsform by the evaporation of sulfate-bearing sea-water, the only alternative that is not ultimatelyderived from seawater is magmatically derivedsulfate.However, a number of characteristics of Arche-

an barite-bearing deposits favor seawater over amagmatic source for sulfate. Firstly, barite depos-its were more abundant before 3.2 Ga, whereasmagmatic activity was more abundant between3.2 and 2.4 Ga, a period of extensive formationof continental crust [32]. The formation of muchof the Yilgarn, Slave and Superior provinces be-tween 3.0 and 2.6 Ga involved extensive volca-nism. If magmatic derived sulfate was importantin forming Archean barite deposits, more depositsshould have formed between 3.2 and 2.4 Ga.Secondly, the isotope characteristics of the bar-

ite deposits are also inconsistent with a magmaticsulfate source. In Phanerozoic deposits where dis-proportionation of magmatic SO2 has been dem-onstrated, sulfate tends to be 15^30x enriched in34S relative to the parent SO2 [31]. Hence, unlessmagmatic N34S values during the Archean weresubstantially lower than the general range of2P 5x [33], the N34S of Archean barite is not

consistent with derivation from disproportionatedmagmatic SO2. Moreover, the mass-independent33S fractionation observed in Archean barite indi-cates that the sulfate formed in the atmosphere[21,23], which, again, is inconsistent with a mag-matic source.Thirdly, Huston et al. [34] interpreted that re-

gional alteration in the 3.24 Ga Panorama VHMSdistrict involving increases in Fe2O3/FeO of vol-canic rocks indicated extensive inorganic reduc-tion of seawater sulfate. Moreover, early ArcheanVHMS deposits have similar mineral zonation toPhanerozoic deposits, suggesting similar deposi-tional mechanisms, including interaction with sul-fate-rich seawater.In summary, the temporal distribution, sulfur

isotope characteristics and alteration/mineraliza-tion associated with bedded Archean barite-bear-ing deposits all point to seawater being the sourceof sulfate. Hence we take the presence of beddedbarite to indicate the presence of sulfate in con-temporaneous seawater.

6. The origin of BIF

Cloud [35] and Holland [36] ¢rst suggested thatiron and silica, the main components of BIFs,were derived from seawater. This concept is nowgenerally accepted, although some disagreementstill exists as to the ultimate origin of these com-ponents. Holland [36] suggested that Fe2þ couldbe provided either from terrestrial weathering inan anoxic environment, volcanic emanations intothe ocean, or bottom waters. Low-temperatureweathering and high-temperature alteration [37]of the ocean £oor could also be sources of iron.The concept that the iron and silica in BIFs

were derived from high-temperature alteration ofrocks below the sea£oor stems from comparingthe rare earth element patterns of BIFs with thosefrom modern hydrothermal £uids venting on thesea£oor. Both the BIFs and the venting £uids arecharacterized by positive europium anomalies,which are indicative of high-temperature altera-tion of volcanic rocks [37^39]. Although thismodel is generally accepted, systematic di¡erencesin the intensity of the europium anomaly exist

EPSL 6994 5-3-04


between Algoma- and Superior-type BIFs (Fig.3).Algoma-type BIFs are characterized by much

larger europium anomalies (s 1.8) than Superi-or-type BIFs (6 1.8). This di¡erence suggeststhat there is a much larger component of vol-canic-related hydrothermal emanations in Algo-ma-type BIFs, consistent with their close associa-tion with greenstone belts. The smaller europiumanomalies that characterize Superior-type BIFssuggest a lower input from volcanic-related hy-drothermal emissions and a higher contributionfrom other sources, including oceanic bottomwaters. Therefore the distribution and character-istics of Superior-type BIFs are a more accuraterecord of background processes that controlledthe overall chemistry of Archean and Proterozoicoceans.Isley [10] and Isley and Abbott [11] demon-

strated that major BIF depositional events corre-late strongly with global plume events. This anal-ysis was extended by Abbott and Isley [40] toconsider Algoma- and Superior-type BIFs sepa-rately. Inspection of their diagrams con¢rmsthat peaks in Algoma-type BIF deposition corre-spond to major global plume events, but that notall plume events correspond to Algoma-type BIFdeposition. The largest peak in Algoma-type BIFscorresponds to a major plume event at 2.75^2.70Ga, which can be related to extensive crustal

growth in many Archean greenstone belts. How-ever, although Abbot and Isley [40] also suggesteda correlation between mantle plumes for Superior-type BIFs, the updated age constraints providedin Table 2 indicate more complicated relation-ships. These constraints suggest that the V2.463Ga Superior-type BIF peak identi¢ed by Abbotand Isley [40] includes at least three discreteevents over a period of at least 163 million years,and these BIF events do not all correlate directlyto identi¢ed plume events. Moreover, theV1.890Ga Superior-type BIF event occurs toward theend of a V700 million year period of pulsedplume activity. Therefore, although some link be-tween mantle plume events and Superior-typeBIFs is probable, that link is not direct. Rather,hydrothermal activity associated with superplumeevents, along with other processes, contributediron to an oceanic reservoir, which depositediron during upwelling [2] or oxidation events notdirectly related to the plume events.

7. A model for the geochemical evolution of thehydrosphere prior to 1.0 Ga

Temporal variations in the abundances of ironformations and bedded sulfate, together with var-iations in N34Ssulfate, suggest that the evolution ofthe hydrosphere before 1.0 Ga can be divided intofour periods: (1) s 3.2 Ga, (2) 3.2^2.4 Ga, (3)2.4^1.8 Ga, and (4) 1.8^1.0 Ga. The geologicand isotopic characteristics, along with the low-temperature geochemistry of sulfur, iron and ba-rium, suggest that the evolution of the hydro-sphere was more complex than generally thought[7,13]. Figs. 4 and 5 present our model for thisevolution, which accounts for most of the con-straints outlined above.

7.1. s 3.2 Ga: sulfur-poor, strati¢ed oceans

The presence of barite and iron formation priorto 3.2 Ga indicate variable seawater sulfur con-tents and redox conditions. Magnetite, the mainmineral in iron formations, only forms at lowtemperatures under reduced redox conditionsand low sulfur levels. Barite forms at higher sulfur

Fig. 3. Variations with time of the average NASC-normal-ized europium anomaly for Algoma- (open circles) and Supe-rior-type (solid diamonds) BIFs. Table 2 presents the data,which were culled to minimize clastic input, used to makethis diagram.

EPSL 6994 5-3-04


levels, but only under oxidized conditions. Thewidespread occurrence of barite before 3.2 Gasuggests that sulfate-bearing seawater was a glob-al, not local, phenomenon.

The variable redox condition could be ex-plained by a strati¢ed ocean, with barite depositsforming in an oxidized upper layer, and Algoma-type BIF forming from Fe2þ-rich hydrothermalemanations in the reduced lower layer (Fig. 4).The presence of an upper oxidized layer coveringshelf areas may have prevented iron depositionand suppressed Superior-type BIF. This mayhave been further in£uenced by lack of shelf en-vironments early in Earth’s history. Di¡erences inthe operation of plate tectonics may have limitedthe extent of some of the environments displayedin Fig. 4 compared to later periods.As argued by a number of authors [1], the

Earth’s early atmosphere was oxygen-poor andreduced. Such conditions would stabilize gasessuch as H2S and SO2, both of which can undergophotolytic reactions, producing mass-independentsulfur isotope fractionation and oxidized sulfur-bearing gases such as SO3 [21]. Modelling by Pav-lov and Kasting [22] suggests that the products ofthese photolytic reactions will be incorporatedinto the hydrosphere through a combination of‘rainout’ and dissolution at the surface of theocean, with SO3 and SO2 being the two mostimportant oxidized sulfur-bearing gases. Thesegases would be dissolved into the hydrosphereby hydrolysis and disproportionation reactions,as follows:

SO3ðgÞ þH2OðlÞ ¼ SO234 þ 2Hþ; ð1Þ

Fig. 4. Model for hydrosphere evolution prior to 1.0 Ga.Photochemical sulfate oxidation e¡ectively shuts down be-tween 2.4 and 1.8 Ga and is not displayed in the diagramafter 2.4 Ga. The term oxidized refers to the presence of oxi-dized species such as sulfate and does not imply the presenceof free oxygen. Low levels of oxygen would have beenpresent in the photic zone and a signi¢cant oxycline mayhave developed by 1.8 Ga.

Fig. 5. Schematic diagram illustrating inferred changes in at-mospheric oxygen levels and oceanic gSO4/gH2S, total sulfurand iron abundances.

EPSL 6994 5-3-04


SO2ðgÞ þH2OðlÞ ¼ SO233 þ 2Hþ; ð2Þ

and

4SO2ðgÞ þ 4H2OðlÞ ¼ 3SO234 þH2SðaqÞ þ 6Hþ:

These processes would produce an upper oceaniclayer enriched in sulfate (cf. [41]), and possiblysul¢te [42]. The presence of sulfate in this layerwould then have allowed deposition of beddedsulfate minerals in hydrothermal and evaporativeenvironments, or possibly at the interface betweenthe oceanic redox layers. Many Paleoarchean bar-ite deposits are interpreted to have formed in shal-low water [25,26,43], and analogies with modernarc-related black smoker deposits suggest depthsof less than 2000 m for the VHMS deposits.Moreover, reduced atmospheric conditions wouldallow the 33S anomalies noted by Farquhar et al.[21].However, below this sulfate-enriched oxidized

upper layer, the ocean was reduced. Inorganic sul-fate reduction processes, where sulfate (and pos-sibly sul¢te) from this upper layer was reducedhydrothermally via interactions with Fe2þ inrock below [34], and biological sulfate (or sul¢te[42]) reduction, which may have evolved by V3.4Ga [15], combined to reduce sulfate produced byphotochemical reactions, eventually forming asteady state. These interactions bu¡ered seawaterredox, sulfur abundances and N34S. During thisperiod, N34Ssulfate and N34Ssulfide were both closeto zero, consistent with crustal values. These in-terpretations are consistent with those of Veizer etal. [44], who suggested that many aspects of Ar-chean oceanic chemistry were bu¡ered by equilib-rium with basaltic oceanic crust.

7.2. 3.2^2.4 Ga: sulfur-poor, reduced oceans

The period from 3.2 to 2.4 Ga di¡ers from theearlier period in lacking bedded sulfate depositsand having abundant BIFs, including abundantSuperior-type BIFs. This suggests that the oceansat this time lacked a signi¢cant sulfate-bearinglayer (Figs. 3 and 4). Moreover, the abundanceof BIF suggests that the ocean was generallymore reduced than the previous period, allowinghigh concentrations of Fe2þ in the bottom waters,

which may have been supplemented by hydrother-mal emanations as recorded in Algoma-type ironformations.As mass-independent 33S fractionation is re-

corded during this period [21,23], photochemicalreactions with atmospheric sulfur-bearing gaseswere still important. As the atmosphere was stillreduced during this period, the oxidative weath-ering of terrestrial pyrite to sulfate would nothave supplied signi¢cant levels of sulfur to theoceans. Importantly, the advent of sulfate-reduc-ing bacteria, in combination with hydrothermalinorganic sulfate reduction, would help to keepthe oceans virtually sulfate-free because in waterswith low sulfate concentrations biological sulfatereduction rapidly removes sulfate, precipitatingiron sul¢des from iron-rich ocean waters. Becauseof low atmospheric oxygen levels, the terrestrialweathering of sul¢des would not have suppliedsulfate at the rate of removal, and the oceanswould e¡ectively be scrubbed of sulfur throughiron sul¢de precipitation. By 3.2 Ga sulfate depos-its were virtually absent and photolytic reactionswere the main process of sulfate production[21,23]. We infer that biological sulfate reductionmay have been an important process by 3.2 Gaand that this led to the removal of sulfate as asigni¢cant component within ocean waters be-tween 3.2 and 2.4 Ga. This process may havebeen assisted or even initiated by a major meteor-ite bombardment at V3.2 Ga [46], which wouldhave perturbed the steady-state layered hydro-sphere inferred prior to 3.2 Ga. Like the periodbefore 3.2 Ga, the N34S composition of preservedsulfates and sul¢des from 3.2 to 2.4 Ga was lim-ited to that of crustal average [17,45].Biogeochemical evidence for oxygenic photosy-

thesis exists in sediments as old as 2.7 Ga [47].Eukaryotic steroids have a biosynthetic require-ment for free oxygen, these steroids being pre-served as steranes, which have been found in 2.7Ga sediments [48]. This helps set a lower limit foroxygen levels in the photic zone to at least 1%PAL at this time [49]. Upwelling of reduced, Fe-bearing waters into the photic zone would resultin extensive iron precipitation to form Superior-type BIF [2], as shown in Figs. 2 and 4.Further supporting evidence for sulfate-poor

EPSL 6994 5-3-04


seawater may also come from a global excursionin the N13C of kerogen between 2.8 and 2.5 Ga[50]. A model involving signi¢cant rates of meth-anogenesis coupled with methanotrophy has beenproposed by Hayes [54,55] to explain this excur-sion. For a methane cycle to be of global signi¢-cance it would require that other modes of organ-ic matter recycling, such as microbial aerobicrespiration, and bacterial sulfate reduction maynot have been as important within the water col-umn and upper sediments. Such a situation wouldcertainly be the case in anoxic, sulfate-poor andFe2þ-rich waters.

7.3. 2.4^1.8 Ga: transition to sulfur-rich, oxidizedoceans

The rise of atmospheric oxygen levels, whichstarted at 2.4^2.2 Ga [3], initiated one of themost sweeping changes in geologic processes af-fecting Earth’s history. This rise initiated condi-tions that in£uenced the evolution of the hydro-sphere over the next 600 million years (Figs. 3 and4). The rise in atmospheric oxygen levels has beenlinked to the increased sequestration of reducedcarbon by burial [50], to the loss of hydrogen tospace in a methane-rich atmosphere [51], tochanges in the redox of volcanic gases and/or hy-drothermal £uid that interacted with the crust[52,53], or to a major continental growth periodat the Archean^Proterozoic boundary [45]. Muchhigher levels of methanogenesis suggested for theprevious period of Earth’s history may have driv-en atmospheric methane levels and thus initiatedthe loss of hydrogen [51].The most signi¢cant consequence of the rise of

atmospheric oxygen levels was the initiation ofoxidative weathering [3]. An increasingly oxy-gen-rich atmosphere has important rami¢cationsfor the geochemistry of sulfur and iron [28]. Themobility of these elements reverses : sulfur be-comes increasingly mobile as oxidative weatheringof terrestrial sul¢des produces soluble sulfate, in-creasing sulfate levels in seawater. The earliestknown examples of bedded calcium sulfate depos-its are the 2.26 Ga anhydrite lenses in the GordonLake Formation in Ontario [56]. From this timethe abundance of bedded sulfate deposits in-

creased, so that by 1.8 Ga, these deposits wererelatively common (Table 1).The rise of atmospheric oxygen also would

have ended photochemical oxidation of reducedsulfur-bearing gases. Farquhar et al. [21] inter-preted the decrease in mass-independent fraction-ation of 33S between 2.45 and 2.09 Ga as havingresulted from ‘either a change in atmosphericcomposition or actinic £ux’. The ¢rst alternativeis consistent with our data, and suggests thatatmospheric oxygenation destabilized reducedsulfur gases, shutting down the photochemicalreactions responsible. During this period, the in-creasing £ux of sulfate derived from oxidativeweathering swamped the bu¡ering capacity of at-mospheric and volcanic sulfur reservoirs. Biolog-ical sulfate reduction, coupled with higher sea-water sulfate levels led to a larger range inv34Ssulfate�sulfide.The consequences of the oxygenation of the

Earth’s atmosphere did not proceed uniformly intime or space. The processes alluded to aboveprobably occurred at di¡erent rates, and the rateof change di¡ered in di¡erent geologic environ-ments. For instance, sulfate concentrations wouldhave risen more quickly in a small, closed basinrelative to an open oceanic basin (Fig. 4). As aconsequence, the e¡ects of the oxygenation of theatmosphere would have lead to both temporaland spatial variations in the distribution of sulfateand sul¢des.The cessation of signi¢cant BIF at about 1.8

Ga is one of the last consequences in the hydro-sphere of the oxygenation of the atmosphere. The¢nal increase in the oxidation state of bottomwaters would result in extensive deposition ofBIF, which may account for the spike in thequantity of iron formation between 1.9 and 1.8Ga (Fig. 1).During the period between 1.9 and 1.8 Ga, at

least 1.2U108 Mt of BIF was deposited (Table 2:[9]), which equates to 3.6U107 Mt of iron assum-ing an average BIF iron content of 30%. Giventhat the mass of seawater in the Proterozoic wassimilar to current day, the average iron concen-tration in Proterozoic seawater required to precip-itate this iron is 25 ppm. This value, or a value oftwo or three times higher, is consistent with con-

EPSL 6994 5-3-04


straints on iron levels for a reduced ocean indi-cated in Fig. 2. Even if the amount of BIF pro-duced was a factor of two or three more than thatpreserved, the concentration of iron required isconsistent with the values indicated in Fig. 2.The presence of sulfate deposits between 2.4

and 1.8 Ga indicates that the shallower parts ofthe oceans became progressively sulfate-rich, yetthe production of iron formations suggests thatthe bottom waters were reduced and sulfur-poor.Like the time prior to 3.2 Ga, this suggests aredox strati¢ed ocean. Several processes couldcause this inferred strati¢cation. Photosyntheticproduction of organic matter and its oxidativedestruction by microbial aerobic respiration, bac-terial sulfate reduction or even bacterial Fe3þ re-duction would also play important roles in de¢n-ing the redox gradients within ocean waters.Moreover, Fe2þ introduced via hydrothermal em-anations would also have maintained high ironlevels in bottom waters. Biological sulfate reduc-tion would have been an active process in oceanwaters and the sul¢de produced would have beentitrated by the Fe2þ-rich bottom waters, keepingthe lower ocean virtually sulfate-free.

7.4. 1.8^1.0 Ga: sulfur-rich, oxidized hydrosphere

The abundance of bedded sulfate deposits com-bined with the lack of iron formation between 1.8and 1.0 Ga suggests that the hydrosphere wassu⁄ciently oxidized and sulfur-rich to suppressiron solubility. Moreover, N34S variations re£ectbiogenic sulfate-reducing reactions [13]. Once the£ux of sulfate from terrestrial weathering of sul-¢des was high enough to swamp hydrothermalFe2þ £ux, biogenic sulfate reduction could thenbegin to titrate iron from ocean waters. Thegreatly increased £ux of sulfate from weatheringcompared to the supply of Fe2þ in the hydro-sphere ultimately led to the removal of Fe2þ

from the oceans [28]. This reverses the way thatiron titrated sulfur between 3.2 and 2.4 Ga. Thechemistry of the hydrosphere between 1.0 and 1.8Ga was more like modern day seawater, in com-parison to the period before 2.4 Ga. However,sul¢dic bottom water and redox gradients mayhave had a profound e¡ect on the distribution

of redox-sensitive bio-essential elements [7,8,57,58].

7.5. Redox bu¡ering of the hydrosphere

Data presented in this paper suggest not onlythat there were large variations in the redox stateand sulfur concentrations in the hydrosphere, butthat the element that bu¡ers the redox statechanged with time. Sulfur is presently the mostabundant multivalent element in seawater, its ox-idation state bu¡ering the redox state of seawater.As sulfate is virtually the only sulfur species inseawater in the open ocean, this seawater is geo-chemically oxidizing. To appreciably change theredox state of seawater requires the conversionof sulfate to sul¢de, as happens biologically inanoxic basins or thermochemically in sea£oor hy-drothermal systems. However, if the total sulfurcontent of seawater was low, and the iron contentwas high, the speciation of iron would govern theredox state of seawater. Moreover, the presence ofhigh Fe2þ would also prevent a rise in total sulfurconcentrations as sulfur could be titrated by bac-terial sulfate reduction. Therefore, the sulfur cycleprior to 2.4 Ga would be governed by high ironconcentrations along with the low rate of sulfateproduction coupled to rapid bacterial consump-tion. Prior to 3.2 Ga, bacterial sulfate reductionmay not have been a global phenomenon and sea-water did contain sulfate in the upper ocean.The rise in atmospheric O2 led to oxidative

weathering and an increased £ux of sulfate tothe ocean. Bacterial sulfate reduction wouldhave led to the titration of Fe2þ from the globaloceans. The O2 rise may have had its roots in low-sulfate, high-Fe2þ oceans, where methanogenesiswould have been a key component of the carboncycle. This may have provided the methane-richatmosphere where hydrogen loss gradually led toO2 increase [51].

8. Implications of hydrosphere evolution tomineralization

In addition to BIFs and barite deposits, a num-ber of other mineral deposit types have restricted

EPSL 6994 5-3-04


distributions in time [59,60]. For instance, paleo-placer uranium deposits, which require a reducedatmosphere and hydrosphere to form, are foundonly in rocks older than V2.0 Ga, whereas un-conformity-related uranium deposits formed afterthis time. In addition, sediment-hosted copperand shale-hosted zinc^lead (e.g. Mt Isa-type) de-posits entered the geologic record at 2.0 and 1.7Ga, respectively [60]. The unifying theme of thelatter three deposit types is the inference that theyformed from oxidized, sulfate-rich ore £uids. Aspointed out by Lambert et al. [60], Veizer [59] andothers, the oxygenation of the atmosphere, whichbegan at 2.4^2.2 Ga, allowed the generation ofsuch oxidized £uids. Another type of depositwhich appears to require oxidized £uids for eithermetal transport or deposition is iron oxide-hostedCuÂu deposits, including the giant OlympicDam deposit in South Australia. The oldest, andone of the better studied districts ^ Tennant Creekin the Northern Territory, Australia ^ has beendated at about 1.83 Ga [61]. Again the emergenceof this class of deposit appears to be linked to thedevelopment of an oxygenated atmosphere andhydrosphere.

9. Implications of hydrosphere evolution to leadisotope growth

The oxidation of the hydrosphere also in£u-enced temporal changes in the crust and mantle,including Th/U ratios and the growth of lead iso-topes. Modelling by Kramer and Tolstikhin [62]suggested that the ‘future’ paradox of lead iso-topes can be resolved by oxidation of the atmo-sphere (and hydrosphere) at about 2.0 Ga, a pro-cess which mobilized uranium into the oceans andoceanic crust. Collerson and Kamber [32] sug-gested that subduction of oceanic crust resultedin a decrease in depleted mantle Th/U ratios, be-ginning at between 2.2 and 1.8 Ga.Our analysis suggests that many of the conse-

quences of atmospheric oxidation proceeded un-evenly across the Earth’s surface. This may alsoapply to the introduction of uranium into thecrust. This hypothesis is illustrated by consideringlead from mineral prospects in the Pilbara Craton

(cf. [63]). In this region, crust was formed fromthe early Archean (V3.5 Ga) to the very earliestProterozoic (V2.4 Ga), with mineralizing eventsspanning virtually this entire history and continu-ing to even younger periods. Fig. 6 shows thechanges in the isotopic ratios of ore leads fromthis region relative to a theoretical lead isotopeevolution curve calculated using the WabigoonÛchi model of Thorpe et al. [64]. The ore leadsdeviate from the model evolution curve towardlower 208Pb/206Pb values at a model age of around2.4 Ga, which indicates U (232Th/238U integratedto present day) decreased and that uranium wasintroduced into the source rocks of the ores atthis time.As all of the deposits sampled are small in size,

it is likely that the source of Pb was local. There-fore, we hypothesize that uranium was introducedinto the Pilbara upper crust by circulation of oxi-dized sur¢cial £uids during incipient oxidation ofthe atmosphere and hydrosphere. This model isconsistent with the suggestion by Powell et al.[65] that BIFs in the Hamersley Province wereupgraded by oxidized £uids at 2.2 Ga. However,we are not implying that the uranium introduc-tion recorded by Pilbara lead isotopes is a globalevent. Rather, the inferred uranium metasoma-tism of the upper crust is probably a local phe-

Fig. 6. 208Pb/204Pb versus 206Pb/204Pb diagram showing theisotopic composition of ore leads from the Pilbara Craton.At V2.4 Ga the data progressively deviate from the Wabi-goonÛchi growth curve of Thorpe et al. [64] to lower U val-ues.

EPSL 6994 5-3-04


nomenon, with pandemic introduction of uraniuminto the crust and upper mantle at V2.0 Ga[32,62].

10. Conclusions

1. Bedded barite deposits are common in the Pa-leoarchean (s 3.2 Ga), but virtually lacking inthe Meso- and Neoarchean (3.2^2.4 Ga).

2. Variations in europium anomalies betweenSuperior- and Algoma-type BIFs indicate hy-drothermal emanations are not as importantan iron source for Superior-type deposits.

3. The evolution of the hydrosphere prior to 1.0Ga is complex, as follows:3.1. s 3.2 Ga: The oceans were mostly re-

duced and sulfur-poor, although a sul-fate-enriched upper layer was present.Sulfate in this upper layer was sourcedfrom atmospheric photolytic reactions.

3.2. 3.2^2.4 Ga: The oceans were relativelyuniform, reduced, sulfur-poor and iron-rich, although oxyclines may have ex-isted near the surface. These conditionsallowed the extensive development ofSuperior-type BIFs. Although atmo-spheric photolytic sulfate productionstill occurred, this sulfate was rapidlyreduced by biogenic and hydrothermalactivity, preventing the maintenance ofa sulfate-rich upper layer.

3.3. 2.4^1.8 Ga: Increasing atmosphericoxygen levels allowed oxidative weath-ering, which added sulfate to theoceans. This process proceeded at vary-ing rates across the globe, producinglayered oceans and an increasing num-ber of sulfate deposits in the geologicrecord, and culminating with the depo-sition of extensive BIF at 1.9^1.8 Ga asthe ¢nal oxidation of the hydrospherescrubbed the oceans of iron. Becausethe oxidation of the atmosphere re-moved reduced sulfur gases, photolyticsulfate reduction progressively shutdown.

3.4. 1.8^1.0 Ga: The oceans were uniformlysulfate-enriched and iron-poor. Beddedsulfate deposits became common andBIFs rare.

4. Prior to 2.4 Ga, Fe2þ was the main oceanicredox bu¡er. After 1.8 Ga, sulfate was themain redox bu¡er. From 2.4 to 1.8 Ga, theredox bu¡er shifted from Fe2þ to sulfate atvarious rates across the globe depending onbasin size and environmental factors.

5. The evolution of the hydrosphere had majorimpacts on the types of mineral depositsformed. Oxidized £uids became important be-tween 2.4 and 1.8 Ga, which allowed the for-mation of sediment-hosted copper, Mt Isa-typezinc^lead and iron oxide-hosted copper^golddeposits. Moreover, the presence of these oxi-dized £uids allowed upgrading of BIFs to formiron oxide deposits.

6. The introduction of uranium into the mantleby subduction of oxidized oceanic crust, there-by decreasing mantle Th/U ratios, was a globalprocess between 2.2 and 1.8 Ga. Lead isotopedata from the Pilbara Craton suggest that ura-nium addition to the upper crust may haveoccurred locally at V2.4 Ga in response tothe initial oxidation of the hydrosphere andatmosphere.

Acknowledgements

This contribution is the result of encourage-ment by Martin Van Kranendonk to expandupon ideas generated as part of the North PilbaraProject, a joint Geoscience Australia-GeologicalSurvey of Western Australia National GeoscienceMapping Accord Project. Alexander Larionovand Hannu Huhma are thanked for their help intracking down information on the age of someBIFs. Karin Orth provided unpublished geochem-ical analyses of BIF in the Kimberley Block ofWestern Australia. James Kasting, Ian Lambert,Karen MacKenzie, Peter Southgate, Shen-Su Sun,Jan Veizer and Malcolm Walter are thanked forreviewing early drafts of this contribution, andDean Hoatson and Janet Hope provided ¢nalpolishing. These reviews and ongoing discussion

EPSL 6994 5-3-04


with some of the reviewers, particularly MalcolmWalter and Shen-Su Sun, greatly improved the¢nal version of this contribution, which is pub-lished with permission of the CEO of GeoscienceAustralia. Terry Mernagh generated the data usedto construct Fig. 2.[BOYLE]

References

[1] P.E. Cloud Jr., A working model of the primitive Earth,Am. J. Sci. 272 (1972) 537^548.

[2] J.C.G. Walker, C. Klein, M. Sachidlowski, J.W. Schopf,D.J. Stevenson, M.R. Walter, Environmental evolution ofthe Archean^Early Proterozoic Earth, in: J.W. Schopf(Ed.), Earth’s Earliest Biosphere ^ Its Origin and Evolu-tion, Princeton University Press, Princeton, NJ, 1983, pp.260^290.

[3] R. Rye, H.D. Holland, Paleosols and the evolution ofatmospheric oxygen; a critical review, Am. J. Sci. 298(1998) 621^672.

[4] H. Ohmoto, Evidence of pre-2.2 Ga paleosols for theearly evolution of atmospheric oxygen and terrestrial bio-ta, Geology 24 (1996) 1135^1138.

[5] S. Vearncombe, M.E. Barley, D.I. Groves, N.J.McNaughton, E.J. Mikucki, J.R. Vearncombe, 3.26 Gablack smoker-type mineralization in the Strelley Belt, Pil-bara Craton, Western Australia, J. Geol. Soc. Lond. 152(1995) 587^590.

[6] N.J. Beukes, C. Klein, Models for iron-formation deposi-tion, in: J.W. Schopf, C. Klein (Eds.), The ProterozoicBiosphere ^ A Multidisciplinary Study, Cambridge Uni-versity Press, New York, 1992, pp. 147^151.

[7] I.B. Lambert, D.I. Groves, Early Earth evolution andmetallogeny, in: Handbook of Strata-Bound and Strati-form Ore Deposits vol. 8, 1981, pp. 339^447.

[8] A.D. Anbar, A.H. Knoll, Proterozoic ocean chemistryand evolution: A bioinorganic bridge?, Science 297(2002) 1137^1142.

[9] H.L. James, Distribution of banded iron-formation inspace and time, in: A.E. Trendall, R.C. Morris (Eds.),Iron-Formation: Facts and Problems, Elsevier, Amster-dam, 1983, pp. 471^490.

[10] A.E. Isley, Hydrothermal plumes and the delivery of ironto banded iron formation, J. Geol. 103 (1995) 169^185.

[11] A.E. Isley, D.H. Abbot, Plume-related ma¢c volcanismand the deposition of banded iron formation, J. Geophys.Res. 104 (1999) 15461^15477.

[12] G.A. Gross, Geology of iron deposits in Canada. I. Gen-eral geology and evaluation of iron deposits, Geol. Surv.Can. Econ. Geol. Rep. 22, 1965, 181 pp.

[13] D.E. Can¢eld, R. Raiswell, The evolution of the sulfurcycle, Am. J. Sci. 299 (1999) 697^723.

[14] J. Veizer, Recycling and preservation probabilities of sedi-ments, in: J.W. Schopf, C. Klein (Eds.), The Proterozoic

Biosphere ^ A Multidisciplinary Study, Cambridge Uni-versity Press, New York, 1992, pp. 63^65.

[15] Y.A. Shen, R. Buick, D.E. Can¢eld, Isotopic evidence formicrobial sulphate reduction in the early Archaean era,Nature 410 (2001) 77^81.

[16] D.L. Huston, Stable isotopes and their signi¢cance forunderstanding the genesis of volcanic-hosted massive sul-¢de deposits: a review, Rev. Econ. Geol. 8 (1999) 157^179.

[17] J.M. Hayes, I.B. Lambert, H. Strauss, The sulfur-isotoperecord, in: J.W. Schopf, C. Klein (Eds.), The ProterozoicBiosphere ^ A Multidisciplinary Study, Cambridge Uni-versity Press, New York, 1992, pp. 129^132.

[18] G.E. Claypool, W.T. Holser, I.R. Kaplan, H. Sakai, I.Zak, The age curves of sulfur and oxygen isotopes inmarine sulfate and their mutual interpretation, Chem.Geol. 28 (1980) 199^260.

[19] P. Gorjan, J.J. Veevers, M.R. Walter, Neoproterozoicsulfur-isotope variations in Australia and global implica-tions, Precambrian Res. 100 (2000) 151^179.

[20] H. Strauss, The isotopic composition of sedimentary sul-fur through time, Palaeogeogr. Palaeoclimatol. Palaeo-ecol. 132 (1997) 97^118.

[21] J. Farquhar, H.-M. Bao, M. Thiemens, Atmospheric in-£uence of the Earth’s earliest sulfur cycle, Science 289(2000) 756^758.

[22] A.A. Pavlov, J.F. Kasting, Mass-independent fractiona-tion of sulfur isotopes in Archean sediments: strong evi-dence for an anoxic Archean atmosphere, Astrobiology 2(2002) 27^41.

[23] S.J. Mojzsis, C.D. Coath, J.P. Greenwood, K.D. McKee-gan, T.M. Harrison, Mass-independent isotope e¡ects inArchean (2.5 to 3.8 Ga) sedimentary sul¢des determinedby ion microprobe analysis, Geochim. Cosmochim. Acta67 (2003) 1635^1658.

[24] Y. Shvarov, E. Bastrakov, HCh: a software package forgeochemical equilibrium modelling ^ user’s guide, Austr.Geol. Surv. Org. Rec. 1999/25, 1999, 61 pp.

[25] I.B. Lambert, T.H. Donnelly, J.S.R. Dunlop, D.I.Groves, Stable isotopic compositions of early Archaeansulphate deposits of probable evaporitic and volcanogenicorigins, Nature 276 (1978) 808^810.

[26] R. Buick, J.S.R. Dunlop, Evaporitic sediments of earlyArchaean age from the Warrawoona Group, NorthPole, Western Australia, Sedimentology 37 (1990) 247^277.

[27] W. Nijman, K.C.H. de Bruijne, M.E. Valkering, Growthfault control of early Archaean cherts, barite mounds andchert-barite veins, North Pole Dome, eastern Pilbara,Western Australia, Precambrian Res. 88 (1998) 25^52.

[28] J.C.G. Walker, P. Brimblecombe, Iron and sulfur in thepre-biologic ocean, Precambrian Res. 28 (1985) 205^222.

[29] P.W. Jewell, Bedded barite in the geological record,SEPM Spec. Publ. 66 (2000) 147^161.

[30] C.S. Eldridge, P.B. Barton, H. Ohmoto, Mineral texturesand their bearing on formation of the Kuroko orebodies,Econ. Geol. Monogr. 5 (1983) 241^281.

EPSL 6994 5-3-04


[31] A. Arribas Jr., Characteristics of high-sul¢dation epither-mal deposits, and their relation to magmatic £uid, Min.Assoc. Can. Short Course 23 (1995) 419^454.

[32] K.D. Collerson, B.S. Kamber, Evolution of the continentsand the atmosphere inferred from Th^U^Nb systematicsof the depleted mantle, Science 283 (2000) 1519^1522.

[33] H. Ohmoto, R.O. Rye, isotopes of sulfur and carbon, in:H.L. Barnes (Ed.), Geochemistry of Hydrothermal OreDeposits, 2nd edn., Wiley, New York, 1979, pp. 639^567.

[34] D.L. Huston, C.W. Brauhart, S.L. Drieberg, G.J. David-son, D.I. Groves, Metal leaching and inorganic reductionin volcanic-hosted massive sul¢de mineral systems: evi-dence from the paleo-Archean Panorama district, WesternAustralia, Geology 29 (2001) 687^690.

[35] P. Cloud, Paleoecological signi¢cance of the banded iron-formation, Econ. Geol. 68 (1973) 1135^1143.

[36] H.D. Holland, The oceans: a possible source of iron iniron-formation, Econ. Geol. 68 (1973) 1169^1172.

[37] L.A. Derry, S.B. Jacobsen, The chemical evolution ofPrecambrian seawater: evidence from REEs in bandediron formation, Geochim. Cosmochim. Acta 54 (1990)2965^2977.

[38] J.L. Graf Jr., Rare earth elements, iron formations andseawater, Geochim. Cosmochim. Acta 42 (1978) 1845^1863.

[39] C. Klein, N.J. Beukes, Time distribution, stratigraphy,and sedimentological setting, and geochemistry of Pre-cambrian iron-formations, in: J.W. Schopf, C. Klein(Eds.), The Proterozoic Biosphere: A MultidisciplinaryApproach, Cambridge University Press, New York,1992, pp. 139^147.

[40] D. Abbott, A. Isley, Oceanic upwelling and mantle-plumeactivity: paleomagnetic tests of ideas on the source of Fein early Precambrian iron formations, Geol. Soc. Am.Spec. Pap. 352 (2001) 323^339.

[41] C.E.J. De Ronde, D.M.D. Channer, K. Faure, C.J. Bray,E.T.C. Spooner, Fluid chemistry of Archean sea£oor hy-drothermal vents: implications for the composition of cir-ca 3.2 Ga seawater, Geochim. Cosmochim. Acta 61 (1997)4025^4042.

[42] G.W. Skyring, T.H. Donnely, Precambrian sulfur isotopesand a possible role for sul¢te in the evolution of biologicalsulfate reduction, Precambrian Res. 17 (1982) 41^61.

[43] K. Sugitani, K. Mimura, K. Suzuki, K. Nagamine, R.Sugisaki, Stratigraphy and sedimentary petrology of anArchean volcanic-sedimentary succession at Mt Golds-worthy in the Pilbara Block, Western Australia: implica-tions of evaporite (nahcolite) and barite deposition, Pre-cambrian Res. 120 (2003) 55^79.

[44] J. Veizer, W. Compston, J. Hoefs, H. Nielson, Mantlebu¡ering of the early oceans, Naturwissenschaften 69(1982) 173^180.

[45] Y. Godderis, J. Veizer, Tectonic control of chemical andisotopic composition of ancient oceans: the impact ofcontinental growth, Am. J. Sci. 300 (2000) 434^461.

[46] A.Y. Glikson, The astronomical connection of terrestrialevolution; crustal e¡ects of post-3.8 Ga mega-impact clus-

ters and evidence for major 3.2P 0.1 Ga bombardment ofthe Earth^Moon system, J. Geodyn. 32 (2001) 205^229.

[47] R.E. Summons, L.L. Jahnke, J.M. Hope, G.A. Logan,2-Methylhopanoids as biomarkers for cyanobacterial oxy-genic photosynthesis, Nature 400 (1999) 554^557.

[48] J.J. Brocks, G.A. Logan, R. Buick, R.E. Summons, Ar-chean molecular fossils and the early rise of eukaryotes,Science 285 (1999) 1033^1036.

[49] L.L. Jahnke, H.P. Klein, Oxygen requirement for forma-tion and activity of the squalene epoxidase in Saccharo-myces cerevisiae, J. Bacteriol. 155 (1983) 488^492.

[50] D.J. Des Marais, H. Strauss, R.E. Summons, J.M. Hayes,Carbon isotopic evidence for the stepwise oxidation of theProterozoic environment, Nature 359 (1992) 605^609.

[51] D.C. Catling, K.J. Zahnle, C.P. McKay, Biogenic meth-ane, hydrogen escape, and the irreversible oxidation ofearly Earth, Science 293 (2001) 839^843.

[52] L.R. Kump, J.F. Kasting, M.E. Barley, Rise of atmo-spheric oxygen and the ‘upside-down’ Archean mantle,Geochem. Geophys. Geosyst. 2 (2001) paper number2000GC000114.

[53] H.D. Holland, Volcanic gases, black smokers and thegreat oxidation event, Geochim. Cosmochim. Acta 66(2002) 3811^3826.

[54] J.M. Hayes, Geochemical evidence bearing on the originof aerobiosis, a speculative hypothesis, in: J.W. Schopf(Ed.), Earth’s Earliest Biosphere, its Origin and Evolu-tion, Princeton University Press, Princeton, NJ, 1983,pp. 291^301.

[55] J.M. Hayes, Global methanotrophy at the Archean^Pro-terozoic transition, in: S. Bengtson (Ed.), Early Life onEarth, Columbia University Press, New York, 1994, pp.220^236.

[56] E.M. Cameron, Evidence from early Proterozoic anhy-drite for sulphur isotopic partitioning in Precambrianoceans, Nature 304 (1983) 54^56.

[57] D.E. Can¢eld, A new model for Proterozoic ocean chem-istry, Nature 396 (1998) 450^453.

[58] Y. Shen, A.H. Knoll, M.R. Walter, Evidence for lowsulphate and anoxia in a mid-Proterozoic marine basin,Nature 423 (2003) 632^635.

[59] J. Veizer, Evolution of ores of sedimentary a⁄liationthrough geologic history; relations to the general tenden-cies in evolution of the crust, hydrosphere, atmosphereand biosphere, in: K. Wolf (Ed.), Handbook of Strata-Bound and Stratiform Ore Deposits vol. 3, Elsevier, Am-sterdam, 1976, pp. 1^41.

[60] I.B. Lambert, N.J. Beukes, C. Klein, J. Veizer, Protero-zoic mineral deposits through time, in: J.W. Schopf, C.Klein (Eds.), The Proterozoic Biosphere: A Multidisci-plinary Approach, Cambridge University Press, NewYork, 1992, pp. 59^62.

[61] D.M. Compston, I. McDougall, 40Ar^39Ar and K^Ar ageconstraints on the early Proterozoic Tennant Creek Block,northern Australia, and the age of its gold deposits, Aust.J. Earth Sci. 41 (1994) 609^616.

[62] J.D. Kramer, I.N. Tolstikhin, Two terrestrial lead isotope

EPSL 6994 5-3-04


paradoxes, forward transport modelling, core formationand the history of the continental crust, Chem. Geol. 139(1997) 75^110.

[63] D.L. Huston, S.-S. Sun, R. Blewett, A.H. Hickman, M.Van Kranendonk, D. Phillips, D. Baker, C. Brauhart, Thetiming of mineralization in the Archean North Pilbaraterrain, Western Australia, Econ. Geol. 97 (2002) 733^756.

[64] R.A. Thorpe, A.H. Hickman, D.W. Davis, J.K. Morten-son, A.F. Trendall, Constraints to models for Archean

lead evolution from precise Zircon U^Pb geochronologyfor the Marble Bar Region, Pilbara Craton, Western Aus-tralia, University of Western Australia, Geology Depart-ment and University Extension, Publication 22, 1992, pp.395^407.

[65] C.McA. Powell, N.H.S. Oliver, Z.-X. Li, D.McB. Martin,J. Ronaszeki, Synorogenic hydrothermal origin for giantHamersley iron oxide ore bodies, Geology 27 (1999) 175^178.

EPSL 6994 5-3-04


Barite, BIFs and bugs: evidence for the evolution of the Earth’s early hydrosphereIntroductionThe temporal distributions of sulfate and iron formationSulfur isotope variations through timeThe low-temperature geochemistry of sulfur, iron and bariumThe origin of bedded sulfate depositsThe origin of BIFA model for the geochemical evolution of the hydrosphere prior to 1.0 Ga3.2-2.4 Ga: sulfur-poor, reduced oceans2.4-1.8 Ga: transition to sulfur-rich, oxidized oceans1.8-1.0 Ga: sulfur-rich, oxidized hydrosphereRedox buffering of the hydrosphere

Implications of hydrosphere evolution to mineralizationImplications of hydrosphere evolution to lead isotope growthConclusionsAcknowledgementsReferences

barite,bifsandbugs:evidencefortheevolutionof theearth...

Documents