Carbon isotope ratios of Phanerozoic marine cements: Re-evaluating the global carbon and sulfur systems

Geochimicaet CosmochimicaActa, Vol. 61, No. 22, pp. 4831-4846, 1997 Copyright© 1997Elsevier ScienceLtd Printed in the USA. All fights reserved 0016-7037/97 $17.00 + .00

Pergamon

PH S0016-7037(97) 00361 -X

Carbon isotope ratios of Phanerozoic marine cements: Re-evaluating the global carbon and sulfur systems ScoTr J. CARPENTER1 and KYGER C LOHMANN2 ~Department of Geosciences, The University of Texas at Dallas, Richardson, Texas, USA 2Department of Geological Sciences, The University of Michigan, Ann Arbor, Michigan, USA (Received March 20, 1997; accepted in revised form July 21, 1997) Abstract--Original 613C values of abiotically precipitated marine cements from a variety of stratigraphic intervals have been used to document secular variations in the 613C values of Phanerozoic oceans. These, together with the 634S values of coeval marine sulfates, are used to examine the global cycling of carbon and sulfur. It is generally accepted that secular variation in 613C and 634S values of marine carbonates and sulfates is controlled by balanced oxidation-reduction reactions and that their long-term, steadystate variation can be predicted from the present-day isotopic fractionation ratio ( A c / A s ) the ratio of the riverine flux of sulfur and carbon (Fs/Fc). The predicted slope of the linear relation between 6 ~3Cc~ and 634Ssulfat,values is approximately -0.10 to -0.14. However, temporal variation observed in marine cement ~3C values and the ~34S values of coeval marine sulfates produces a highly significant linear relation (r 2 = 0.80; a > 95%) with a slope of -0.24; approximately twice the predicted value. This discordance suggests that either the Phanerozoic average riverine FslFc was 1.6-3.3 times greater than today's estimates or that an additional source of 34S-depleted sulfur or 13C-enriched carbon, other than continental reservoirs, was active during the Phanerozoic. This new relation between marine 6 ~3C and 634S values suggests that the flux of reduced sulfur, iron, and manganese from seafloor hydrothermal systems affects oceanic 02 levels which, in turn, control the oxidation or burial of organic matter, and thus the 613C value of marine DIC. Therefore, the sulfur system (driven by seafloor hydrothermal systems) controls the carbon system rather than organic carbon burial controlling the response of 6a4s values (via formation of sedimentary pyrite). Secular variation of marine S7Sr/a6Sr ratios and 6180 values argues for a coupling of 613C and ~34S values to variation in the relative contribution of seafloor hydrothermal and continental weathering fluxes. These trends indicate that the early Paleozoic was dominated by low temperature silicate weathering, whereas the Late Paleozoic to Modern was dominated by high temperature seawater-basalt interactions. Variation in Proterozoic 613C¢~b and 63*S~lf~ values produces a slope that is greater than that of the Phanerozoic ( -0.50 vs. -0.24). This steeper slope is consistent with other geochemical data that indicate relatively high seafloor hydrothermal fluxes during the late Precambrian. We speculate that the dramatic evolutionary changes of the Neoproterozoic-Paleozoic transition occur during a waning of seafloor hydrothermal fluxes and a concomitant decrease in 02 consumption that permitted the oxygenation of seawater thought to be critical in metazoan evolution. Copyright © 1997 Elsevier Science Ltd 1. INTRODUCTION

sulfur cycle is negligible (e.g., Berner and Raiswell, 1983; Lasaga et al., 1985; Kump and Garrels, 1986; Francois and Gerard, 1986; Kump, 1989; Berner, 1987, 1990). In contrast, Walker (1986) suggested that the flux of sulfur from seafloor hydrothermal systems plays a significant role in the global sulfur budget which, in turn, affects the oxygen and carbon systems. The assumption that the flux of reduced sulfur species from seafloor hydrothermal systems is negligible simplifies the isotopic and mass balance considerations as the sulfur fluxes for riverine input and sedimentary pyrite removal are approximately equal. However, there is evidence that the sulfur flux from seafloor hydrothermal systems is globally significant (e.g., Von Damm et al., 1985a,b, 1995; Humphris et al., 1995). The introduction of an additional sulfur flux into seawater clearly requires a re-examination of this and related systems (see Walker, 1986 for discussion). The relation between the 513C values of marine carbonate and ~534S values of marine sulfate provide a means of conducting this re-examination (e.g., Garrels and Lerman, 1981, 1984; Bemer and Raiswell, 1983; Kump and Garrels, 1986; Walker, 1986). In order to relate long-term patterns of chemical variation

During the past three decades, numerous studies have demonstrated that the history of Phanerozoic oceans is one of dynamic chemical exchange between the hydrosphere and lithosphere (e.g., Compston, 1960; Weber, 1967; Schidlowski et al., 1976; Veizer and Hoefs, 1976; Holland, 1978, 1984; Holser, 1977; Claypool et al., 1980; Scholle and Arthur, 1980; Veizer et al., 1980). Secular variation in ocean chemistry is recorded as changes in the carbon, strontium, sulfur, and oxygen isotopic composition of ancient sedimentary minerals. Detailed chronologies have been documented for S7Sr/S6Sr ratios of marine carbonates (e.g., Burke et al., 1982; Denison et al., 1997) and 634S values of marine evaporites (e.g., Holser, 1977; Claypool et al., 1980). Variations in the 613C and 61sO values of marine carbonates have also been documented (e.g., Weber, 1967; Veizer and Hoefs, 1976; Veizer et al., 1980; Lindh, 1983; P o p p e t al., 1986; Walker and Lohmann, 1989; Carpenter et al., 1991 ). Most global geochemical models have assumed that the contribution of seafloor hydrothermal fluxes (via high temperature seawater-basalt interaction) to the carbon-oxygen4831

S. J. Carpenter and K. C Lohmann

4832

Table 1. Steady-state dl~13C/d~345 calculated from flux and fractionation parameters employed in models of the global carbon-sulfur exogenic cycle. F~ Study Veizer et al. (1980) Walker (1986) Lasaga et al. (1985) Kump and Garrels (1986) Lindh (1983) (Data Only) This Study (w/DMS) This Study (w/o DMS)

(in l0 t2 mol/yr)

1.9 3.0 1.5 1.5 . 4.0 3.8

31.1 25.0 16.3 14.0 . . 25.0 25.0

F~/F~ A~

A¢

At/As

dt~13C/d6345

0.06 -40 -25 0.12 -40 -25 0.09 -35 -26 0.11 -35 -25 . . . 0.16 2.9* 2.3* 0.15 2.7* 2.3*

0.62 0.62 0.74 0.71 . 0.81 0.85

-0.13 -0.14 -0.10 -0.11 0.26 -0.24 -0.24

* Values calculated for all fluxes.

to changes in geological process, it is necessary to understand the effects of different geologic processes on the variation of each isotopic tracer. These include a variety of oxidation-reduction, silicate-equilibration, and basalt-seawater reactions (e.g., Garrels and Perry, 1974; Claypool et al., 1980; Veizer et al., 1980; Berner and Raiswell, 1983; Berner et al., 1983; Lasaga et al., 1985; Kump, 1989). On the basis of theoretical arguments and empirical relations in contemporary oceans, Garrels and Perry (1974) postulated a redox balance between the sulfur and carbon reservoirs: 4FeS2 + CaCO3 + 7CaMg(CO3)2 + 7SIO2 + 15H20 = 15CH20 + 8Ca504 -t- 2Fe203 + 7MgSiO3

(1)

This reaction provides for the direct coupling of the carbon and sulfur exogenic cycles by assuming constant atmospheric 02 and marine SO4 concentrations and serves as the focus for modeling changes in fluxes among the oxidized and reduced reservoirs and in isotopic composition of oceanic sulfate and bicarbonate (Garrels and Lerman, 1981, 1984; Berner and Raiswell, 1983; Kump and Garrels, 1986; Walker, 1986). Using an isotopic mass balance approach, a steady-state value for the slope of variation of carbon and sulfur isotopic change (drI3C/d6345) can be related to fluxes and fractionation factors by the relation:

dt~13C/dt~345 : - - ( 1 5 / 8 ) X ( A c / A s ) ) <

(fs/Fc) (2)

where - ( 1 5 / 8 ) is the carbon-sulfur mole ratio from Eqn. 1, Ac/As is the carbon-sulfur isotopic fractionation ratio, and Fs/Fc is the flux ratio of sulfur and carbon measured from present-day river fluxes (Veizer et al., 1980; Walker, 1986; Kump and Garrels, 1986). This relation assumes that a steady-state between the carbon and sulfur cycles has been achieved and that the production of oxygen from excess carbon burial has been matched by a compensatory oxidation of reduced sulfur. As noted by Kump and Garrels (1986), the rate of change of the isotopic composition of carbon should be more rapid than the complimentary shift in sulfur, due to the comparatively larger reservoir size of sulfur in ocean water. This difference may induce short-term deviations in carbon composition about a steady-state, 6J3C-634S correlation line. However, when examined on a scale of Phanerozoic time, the measured d613C/d634S slope should approximate the steady-state value.

Applying this relation, a steady-state df~3C/d634S slope can be predicted from the Fs/Fc ratio of present-day rivers (Table l). This study provides an additional prediction, based on the flux parameters of Lasaga et al. (1985), which evaluates the change in effective Ac/As fractionation factors resulting from a dilution effect of weathering evaporites and carbonates contributing to the total river fluxes. However, when such effects are considered, the changes in Ac/As and d613C/d634S are minor. Predicted variation in dfiaC/d634S ranges from - 0.10 to - 0.14 (Table 1 ). The magnitude of this d613C-dt334Svalue is of importance in evaluating the parameters employed in models of the exogenic carbon-sulfur cycle. Clearly, an overall reciprocal relation between changes in 6 ~3C and 634S values has been confirmed by numerous studies (Veizer et al., 1980; Lindh, 1983; Holser, 1984). The magnitude of this slope, however, may be controlled by either Fs/Fc (Berner and Raiswell, 1983 ) or Ac/As, when examined on a short timescale where steady-state conditions have not been attained (Kump and Garrels, 1986). Thus, when examined on a Phanerozoic timescale, if empirically determined slopes lie within the range of - 0 . 1 0 to -0.14, these would support current estimates of Fs/Fc and A c / A s ratios and, by implication, would limit the source of sulfur and carbon to continental inputs. Slopes differing from this range require that additional fluxes of sulfur and/or carbon with different isotopic compositions must be considered (e.g., Walker, 1986). Addition of such reservoirs would modify the effective fractionation factors ( A c / A s ) or result in variable sulfur and carbon fluxes into or out of the ocean (Berner and Raiswell, 1983). The purpose of this study is to examine the record of secular variation in Phanerozoic marine 613C values derived from abiotic marine cements and to evaluate the parameters presently employed in models of the global carbon-sulfur cycle (Fig. 1). 2. THE MARINE CARBONATE RECORD 2.1. Equilibrium Precipitation of Marine Cements Marine carbonates provide the most accurate and complete record of secular variation of 6~3C values in paleoceans. However, as summarized by Veizer et al. (1980), numerous complications must be considered to determine original marine isotope records. The composition of mineral phases, for

C isotope ratios of Phanerozoic marine cements Pbamrozo~ Mu~

,

......................

C . ~ m ~ and su/fa~s

.................

........... +................

i

......................

+ ....

'

0

2Q..... ¸, 10

12

14

16

18

20

°++ 22

24

26

28

30

8~S Marine Sulfate Fig. 1. Plot of 6 t3C vs. 634Svalues of Phanerozoic marine carbonates and sulfates. 613C values are from marine cements (compiled for this study), and 634S values are from Claypool et al. (1980). See Table 2 for actual values and descriptions. The inset compares the slope derived from the empirical 6 1 3 C - 6 3 4 S data in this study with similar data from Veizer et al. (1980) and Lindh (1983).

example, must reflect real variation in ocean chemistry rather than disequilibrium fractionation during precipitation; this is particularly important when evaluating biogenic mineral phases that may exhibit "vital effects." Samples must also represent carbonate precipitation in well-mixed surface waters or chemical variation may simply reflect local facies effects. In addition, because diagenesis can significantly alter the initial composition, such effects must be identified if the magnitude of secular shifts is to be resolved. Finally, the effects of primary mineralogy must be considered because of significant carbon isotope fractionation between coevally precipitated calcite and aragonite (e.g., Rubinson and Clayton, 1969; Gonzalez and Lohmann, 1985; Carpenter et al., 1991; Romanek et al., 1992). This is particularly important in light of studies demonstrating cyclic variation in the mineralogy of abiotic marine carbonates (e.g., Sandberg, 1983, 1985; Wilkinson et al., 1985). Using secular variations in marine carbonate 613C values (data from published whole rock analyses of marine micrites), Veizer et al. (1980) demonstrated a negative correlation between 6t3C and 634S values for the Phanerozoic. The general agreement between the slope calculated from these data and that from the isotope mass-balance calculation in Eqn. 2, has been cited as independent confirmation of flux and fractionation parameters used in model calculations. However, comparison of values in Veizer et al. (1980) and numerous detailed studies (e.g., Tan and Hudson, 1974; Davies, 1977; Scholle and Arthur, 1980; Brand, 1982; Czerniakowski et al., 1984; Given and Lohrnann, 1985; Meyers and Lohrnann, 1985; Moldovanyi and Lohmann, 1985; Hurley and Lohmann, 1989; Carpenter and Lohrnann, 1989) suggests that the 6 t3C values of Veizer et al. (1980) significantly underestimate the actual magnitude of Phanerozoic secular variation. In this study, a research strategy has been adopted to

4833

minimize those factors that are likely to induce variation other than those reflecting secular changes in ocean chemistry. To avoid complications arising from local facies effects, this study focuses on low-latitude, shallow shelf, and reefal carbonates. Reefal buildups are typically developed in well-mixed, shallow waters at the edges of continental shelves and are restricted geographically to low latitude settings (e.g., Opdyke and Wilkinson, 1990). Choice of a suitable carbonate component for analysis is more difficult. Numerous studies have examined the isotopic composition of Phanerozoic biogenic carbonates (e.g., Compston, 1960; Lowenstam, 1961; AI-Aasm and Veizer, 1982; P o p p e t al., 1986). However, it has been demonstrated that the majority of biogenic carbonates are precipitated in varying degrees of carbon isotope disequilibrium (e.g., Veizer, 1983a; Wefer, 1985; McConnaughey, 1989; Carpenter and Lohmann, 1995). More critically, the inherent variation present in biogenic carbonate, even within individual taxa, may limit resolution of oceanic 6 t3C value variation in ancient materials to no better than _ 1%o (e.g., Carpenter and Lohmann, 1995). Carbon and oxygen isotope data from modern marine cements indicate that these abiotic precipitates are formed in isotopic equilibrium with ambient fluids and that they have a relatively small overall range of values (a total range of 0.6%0 for both 6 1 3 C and 6tso values; Fig. 2). The isotope data plotted in Fig. 2 represent analyses of co-occurring aragonite and Mg-calcite marine cements from the windward reef flat of Enewetak Atoll (data from Carpenter et al., 1991). Marine cements have 6t3C and 61sO values that are inherently less variable and higher than co-occurring biogenic or micritic carbonates (e.g.,

Modern Marine Cements Enewetak A t o l l - Windward Reef Flat t

. . . .

•

i

. . . .

I

,

,

+

,

t

,

++

,

t

I

,

,

)

t

. . . .

0

t

,

I

I

+ +

i

I

[

+ +

o

......................~.......................i.........................+........................i....................... +........................i........................

. . . .

-7

! ~ . . . . . . . . -6 -5

i I .... -4

~ i .... -3

i , . . . . -2

i i . . . . -1

0

I

0

81 SO (PDB) Fig. 2. 613C and 6too values of Holocene marine cements from the windward reef fiat of Enewetak Atoll, Marshall Islands. Rectangles are estimates of the equilibrium low Mg-calcite (LMC), high Mg-calcite (HMC), and aragonite (ARAG) compositions for this location. The difference between aragonite and calcite 6 m3Cvalues is 1.2%o. From Carpenter et al. (1991).

S. J. Carpenter and K. C Lohmann

4834

Table 2. Sulfur, carbon, and strontium isotope data for selected time periods. Age

6~4S*

~I3C

87Sr/86Sr Mineralogy Reference

Holocene Pliocene Cretaceous (Maastrichtian) Cretaceous (Aptian) Jurassic (Kimmeridgian) Triassic (Norian) Permian (Kazanian) Pennsylvanian (Moscovian) Lower Mississippian (Visean) Lower Mississippian (Tournaisian) Upper Devonian (Famennian) Upper Devonian (Frasnian) Middle Devonian (Givetian) Upper Silurian (Pridolian) Middle Silurian (Ludlovian) Upper Ordovician (Hirnantian) Upper Ordovician (Ashgillian) Middle Ordovician (Llanvimian) Lower Ordovician (Tremadocian) Lower Cambrian (Waucoban)

21.4 21.8 18.8 16.2 17.1 15.9 11.7 14.7 16.0 16.9 23.0 23.3 21.0 26.5 27.5 28.0 28.2 29.0 29.3 29.5

3.0 3.2 2.8 3.0 3.0 2.8 4.3 3.5 3.5 4.0 2.5 2.5 1.5 1.5 1.5 6.5 0.0 0.5 - 1.0 -0.5

0.70917 0.70900 0.70765 0.70740 0.70690 0.70780 0.70680 0.70830 0.70800 0.70790 0.70810 0.70800 0.70780 0.70870 0.70870 0.70790 0.70785 0.70830 0.70900 0.70870

HMC HMC IMC C C C HMC HMC C IMC IMC IMC C C C C C C C C

1, 2 3 4 5 6 7 8 9 10, 11 12 13, 10 13, 14 15, 10 16 17 18 19 20, 21 19 22, 23

* Claypool et al. (1980); 1. Gonzalez & Lohmann (1985); 2. Carpenter et al. (1991); 3. Frank & Lohmann (1996); 4. Wilson and Opdyke (1996); 5. Moldovanyi& Lohmann (1985); 6. Unpublished Data--Smackover Fm.; 7. Unpublished Data--Alps & Sicily; 8. Given & Lohmann (1985); 9. Graber (1989); 10. Meyers & Lohmann (1985); 11. Dunn (1988); 12. Carpenter (1991); 13. Hurley & Lohmann (1989); 14. Carpenter & Lohmann (1989); 15. UnpublishedData--Wirbelau Quarry; 16. Breining (1985); 17. Cercone & Lohmann (1986); 18. Marshall and Middleton, 1990; 19. Stepanek (1984); 20. Ross et al. (1975); 21. Unpublished Data--Efna Fm.; 22. James and Klappa (1983); 23. Unpublished Data--Shady Ls.

Keith and Weber, 1965; Weber and Woodhead, 1970; Erez, 1978; Graham et al., 1981; Swart, 1983; Wefer, 1985; Gonzalez and Lohmann, 1985; McConnaughey, 1989; Carpenter et al., 1991 ). Because inorganic precipitation results in a consistent equilibrium fractionation between carbonate and ambient dissolved inorganic carbon (DIC), abiotic marine cements may be the preferred material for evaluating variation in paleocean chemistry (e.g., Gonzalez and Lohmann, 1985; Carpenter et al., 1991). Moreover, because such inorganic carbonate precipitation requires extensive wave and tidal pumping (e.g., Land and Goreau, 1970; Ginsburg and James, 1976), marine cements associated with shelf-margin buildups should provide an accurate record of the isotopic composition of low latitude, surface waters. This supposition is supported by isotope data from modern marine carbonates. In a similar manner to those described above, primary marine 6J3C values have been determined for marine cements from various stratigraphic intervals during the Phanerozoic (Table 2). These represent time intervals on the order of 1 - 2 million years. For the majority, primary isotopic values were also determined for several coeval units, providing independent corroboration of the global character of marine 613C estimates. Because, in some cases, these values represent extrapolated endmember compositions, measured standard deviations cannot be calculated. Based on the maximum variation observed in modem marine cements (___0.3%o), and allowing for error in the extrapolation to primary values, these compositions are conservatively precise to within -+0.5%~.

2.2. Diagenetic Alteration and Preservation of Marine ~13C Values Because modem marine cements have metastable mineralogies (aragonite (ARAG) and high Mg-calcite (HMC)), diagenetic modification of original marine isotope ratios must be considered when discussing secular variations in Phanerozoic marine carbonates. Several studies have shown that preservation of original marine 613C values is possible over a wide range of water/rock ( W / R ) ratios and degrees of diagenetic alteration (e.g., Carpenter and Lohmann, 1989; Banner and Hanson, 1990; Carpenter et al., 1991 ). Diagenetic alteration or water-rock interaction commonly occurs in shallow marine carbonates when drops in sea level subaerially expose shallow marine carbonates (composed of biogenic and abiotic precipitates with a range of stabilities). Meteoric water, charged with CO2 from oxidized organic matter in soils, reacts with metastable phases (to varying degrees and at various scales) to produce more stable low Mg-calcite (LMC) (e.g., Lohmann, 1987). Water-rock interaction modeling of these cement compositions indicates that alteration of marine cements often occurs at low W / R ratios ( W / R < 1000; e.g., Meyers and Lohmann, 1985; Lohmann, 1987; Banner and Hanson, 1990; Quinn et al., 1991 ). Retention of marine 613C values is not surprising given the mass balance calculations associated with the reaction of marine carbonates in meteoric waters (e.g., Lohmann, 1987; Banner and Hanson, 1990; Quinn et al., 1991 ). Compared with the carbon derived from dissolving metastable carbonates, rainwater gains relatively little

C isotope ratios of Phanerozoic marine cements carbon from the oxidation of organic matter in soils (via solution of CO2 in water where PCO2 is approximately 10-35 for the atmosphere and 10 -2 for soils; e.g., Lohmann, 1987; Banner and Hanson, 1990). Evolved meteoric fluids (those that have experienced subsequent dissolution of ARAG and HMC) are dominated by carbon from the dissolving marine rock. This yields preservation of marine 613C values even during neomorphic alteration of ARAG and HMC. The preservation of carbon isotope ratios in bulk carbonate samples has been exploited by various workers examining events in earth history (e.g., Magaritz, 1989; Derry et al., 1992; Knoll et al., 1996). Here we report data from microsamples of marine cement milled from petrographically wellcharacterized materials (typically reef margin grainstones). This sampling technique avoids mixing of various quantities of carbonate components such as fossils and other allochems, cements, and micrite (each with different isotopic compositions). Given the isotopic heterogeneity of ancient marine carbonates, due either to original inherent variability or diagenesis, detailed studies of ancient marine cements may provide the most accurate and precise estimate of the 613C value of shallow marine DIC. We will describe two types of carbon isotope preservation that have been observed in ancient marine cements. Comparison of the isotopic compositions of ancient marine cements with those of modem analogues will also aid our understanding of the preservation of carbon isotope ratios (e.g., Gonzalez and Lohmann, 1985; Carpenter et al., 1991; Fig. 2). The first example of preservation is the mineralogic transformation (ARAG to LMC or HMC to LMC) over a range of W / R ratios which produces a physical mixture of two carbonate phases each with different 61sO and 613C values (e.g., Given and Lohmann, 1985; Frank and Lohmann, 1996). This heterogeneity permits, with the help of microsampling techniques, the determination of original marine 6 ~3C values. The second type of preservation is the result of water-rock interaction at low W / R ratios (e.g., Meyers and Lohmann, 1985; Lohmann, 1987) and, in many cases, the outright preservation of original marine carbonate (e.g., Carpenter and Lohrnann, 1989; Carpenter et al., 1991). Here we briefly describe selected examples of these types of preservation in ancient marine cements. 2.3. Permian Marine Cements

Given and Lohmann (1985) have shown that original marine 6180 and 613C values can be measured and estimated from diagenetically altered aragonite and Mg-calcite cements (Fig. 3). In this case, late-stage diagenetic calcite spars are physically intergrown with calcitized marine cements that have been stabilized to LMC at low W / R ratios. Microsampling of these mixtures produces 6~80 and 613C values that form linear, covariant trends which diverge from primary marine cement values toward diagenetic endmember values. Importantly, the trends for both the former aragonite and Mg-caicite cements converge at different marine 6 ~sO and 6 ~3C values, with relative positions comparable to those of modem aragonite and Mg-calcite marine cements (e.g., Gonzalez and Lohmann, 1985; Carpenter et al., 1991; Figs. 2, 3). The measurement of both aragonite and Mg-calcite

4835

equilibrium 6~gO and 6~3C values is compelling evidence for preservation of original marine 6~3C values. 2.4. Late Devonian and Early Mississippian Marine Cements

Carpenter and Lohmann (1989), Carpenter et al. ( 1991 ), and Carpenter ( 1991 ) have shown that Late Devonian (Frasnian) metastable marine cements (Intermediate Mg-calcite (IMC): 2.5-7 mol% MgCO3) are preserved in the rock record without diagenentic alteration and thus record primary isotopic compositions. We conclude that portions of these cements have escaped diagenetic alteration on the basis of compelling petrographic evidence (alteration of margins of crystals near micro- and macro-pores; see Carpenter and Lohmann, 1989) and intra- and inter-basin correlation of various chemical tracers (613C and 61so values, S7Sr/S6Sr, and Sr/Mg ratios; see Carpenter et al., 1991). In addition, the strontium, carbon, and oxygen isotope data from these cements have overall variability comparable to modem marine cements (Carpenter et al., 1991; Fig. 2). As there is a great body of evidence that indicates that water-rock interactions in marine cements produces isotopic heterogeneity rather than homogeneity (e.g., Given and Lohmann, 1985; Meyers and Lohamnn, 1985; Frank and Lohmann, 1996; Figs. 3, 4, 5); the clustering of data in Figs. 2, 4, and 5 argues for a lack of diagenetic alteration in portions of these marine cements. It is also unlikely that several locations would undergo the same postdepositional reactions thereby producing the same isotopic and chemical compositions (e.g., 6~3C and 6180 values, STSr/86Sr, and Sr/Mg ratios; Carpenter et al., 1991 ). Petrographic and geochemical evidence indicates that alteration fabrics result from interaction with meteoric and burial fluids (Carpenter and Lohmann, 1989; Carpenter et al., 1991; Carpenter, 1991 ). Regardless of possible recrystallization phenomena, even diagenetically altered marine cements (inclusion-rich cements) retain their original marine carbonate 613C values (Fig. 4). This alteration is the result of interaction with meteoric water at low W / R ratios (Carpenter and Lohmann, 1989; Carpenter et al., 1991). Although there is some temporal variation in marine 6 ~3C values within the Leduc Reefs of Alberta (~0.5%o, Carpenter and Lohmann, 1989), the +2.5%0 value is representative of the majority of these Middle Frasnian marine cements. Similar preservation is observed in marine cements from the Lower Mississippian (Toumaisian) Pekisko Formation of Alberta, Canada (Fig. 5). Carpenter ( 1991 ) has described the cements with the same petrographic characteristics as those from the Late Devonian of Alberta (i.e., inclusionpoor and inclusion-rich). Again, the inclusion-poor cements are interpreted to be unaltered, marine IMC, whereas inclusion-rich cements have undergone diagenetic alteration at low W / R ratios and still retain original marine 613C values (Fig. 5). 2.5. Anomalous Late Ordovician 813C values

Brenchley et al. (1994) described a globally significant increase in both 6 tsO and 613C values during a short-lived

4836

S.J. Carpenter and K. C Lohmann Lower Mississippian (Tournaisian) Marine Cements Pek~ko Fro.: 9-17-81-15W5

Permian Reef Complex - Guadalupe Mtns.

i

i

I

6

,~,i

....

r . . . . . . . .

i . . . . . . . .

i ....

i ....

i

g

5 4

i................ i~..... /'[o

~

/

....

..... . . . . . .

4

4

8 o%i 3

2 .....................................

7 ~ ...............

er Reef-Ara~onite [

i

....... -~!

I

........ -9 .s

,,~, -r

1 ~', -6

,~ . . . . -s

, .... -4

i .... -~

~,. . . . -2

i .... -~

1

°

0

1-1 0

..........................................................

. . . . . . . . . . . . . . . . . . . . . .

I . . . . . . . .

5

4 .................................................................... ................... 4

i

go

2 .................. ! ........... -

i

.................

i

•

QO

'

~ •

o

i...................................

•

,

i

i

i

2

7o 0 .................. .....~.

-

0

~

i

-2

.... .i1

•

ll~lusicm-free Marine Cement] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

0

Inclusion-rlch

i .... -~o

i .... -9

Marine Cement

i .... -s

81SO

- 1

]

i .... -7

p........ -6

Inclusioo-rieh Marine Cement

•

Inclusion-free Marine Cement

..................

~ "

-8

-7

-6

-5

-4

-3

-2

0

1

-1

81 S O ( P D B )

Late Devonian (Frasnian) Marine Cements Golden Spike Reef." 3-27.51-27W4 i,

o

1

. . . . i . . . . ; . . . . i . . . . i . . . . i . . . . i . . . . ~. . . . f . . . . .

-1

(PDB)

glaciation during latest Ordovician time (Hirnantian). Kump et al. (1995) have confirmed these carbon isotope results by analyzing Late Ordovician carbonates from Nevada. Carbon isotope ratios from Late Ordovician marine cements and brachiopods from central Sweden have ~ 13C values that are significantly higher than other early Paleozoic marine cement values: +2 to +3%~; Kullsberg Limestone (Tobin and Walker, 1996) and +6.5%o; Boda Limestone (Marshall and Middleton, 1990). Brenchley et al. (1994) suggest that this carbon isotope excursion was caused by a modification of

l , l l

~

2

1

Fig. 3.6 ~3C and ~5~80 values from calcitized aragonite and Mgcalcite marine cements from the Permian Reef complex of SE New Mexico. Data from Given and Lohmann (1985).

5

% o

-9

8~SO

~iooa

......... !.................... ~............... i° ............... ~ ............. ÷................... ~ 3

:2

.........

7o l

..................... i

-s

i ......... -4

.3

2 -2

(PDB)

Fig. 4.6L3C and ~80 values of Upper Devonian (Frasnian) calcite marine cements from the Golden Spike reef, Alberta, Canada. The inclusion-poor cements are unaltered marine carbonate, and the inclusion-rich cements are diagenetically altered at relatively low W/ R ratios (see Carpenter and Lohmann, 1989 and Carpenter et al., 1991 for petrographic and chemical descriptions). Original marine 613C values are preserved in both cases. From Carpenter and Lohmann (1989).

Fig. 5. 6~3C and 6JsO values of Lower Carboniferous (Middle Tournaisian) calcite marine cements from the Pekisko Formation (Waulsortian Mounds), Alberta, Canada. As with Devonian marine cements, inclusion poor cements are unaltered, and inclusion rich cements have undergone diagenetic alteration at low W/R ratios. Data from Carpenter ( 1991 ).

ocean circulation. Kump et al. (1996) suggest that a modification of the global carbon cycle (triggered by the Taconic orogeny) allowed a significant drawdown of atmospheric CO2 and global cooling. This glaciation is relatively shortlived ( ~ 1 myr.) and occurs during a long-term, greenhouse time period (Gibbs et al., 1997). A comparison of Late Ordovician marine cement 613C values indicates that the Hirnantian (+6.5%~) is anomalous in comparison with other Early Paleozoic marine carbonates. Given the global nature of this short-term excursion, we must assume that such values represent nonsteady-state conditions that occur relatively infrequently during the Phanerozoic. The lack of a concomitant change in 634S values is probably due to a short-term decoupling of the carbon and sulfur systems. This is in contrast to the system outlined in our study. A similar positive 613C value excursion occurs during the Late Permian ( + 6 to +8%0, Gruscyzinski et al., 1989; Mii et al., 1997) that corresponds with a relative decrease in marine ~34S values (e.g., Kramm and Wedepohl, 1991; Carpenter and Mitterer, 1996). Clearly further examination of these anomalous time periods is warranted. Because of the short-lived nature of these 613C value excursions, we have not included these values in our analysis of Phanerozoic marine 6 ~3C and 634S values. Lindh ( 1983 ) has also excluded similar data on this basis. However, we would be remiss if we excluded these data without discussion. 3. I M P L I C A T I O N S

OF

PHANEROZOIC

8t'C AND 8~*S VALUES Phanerozoic marine cement ~ 3 C values range from -1.0%o to +4.3%o (Table 2). Our range of 6~3C values is comparable to that of Lindh (1983) ( - 2 . 0 6 to +3.97%~), who used mean values for 10 million year intervals to calcu-

C isotope ratios of Phanerozoic marine cements late the relation between carbonate 6 ~3C values and sulfate 634S values (Fig. 1). The data compilation methods used by Lindh (1983) are also comparable to those of our study. Lindh (1983) states that he has used a majority of the citations compiled by Veizer et al. (1980) but has added an additional thirty studies of marine 613C values and twentyfive studies of marine 634S values. Lindh (1983) compiled only those data that the respective authors were convinced were truly marine and in which isotope ratios are not suspected of being altered by diagenesis (Lindh, 1983). Lindh (1983) also excluded data that showed sudden shifts in 613C values to avoid biasing the linear regression. Our work also excludes data that represent rapid high amplitude changes in 6~3C values (e.g., Late Ordovician). In cases where we present our isotope analyses, extensive petrographic and geochemcial analyses have been conducted to ascertain primary marine cement 6~3C values (e.g., Given and Lohmann, 1985; Carpenter and Lohmann, 1989; Carpenter et al., 1991). One difference in these datasets is that the Lindh (1983) compilation uses a variety of marine carbonate matedais (predominantly micrite and various skeletal carbonates) whereas we have exclusively used abiotic marine cements. Both compilations yield comparable t5~3C-6~S value slopes (-0.26 (Lindh, 1983) vs. -0.24 (this study); Fig. 1). The measured amplitude of variation in marine cement 6~3C values is nearly double that of the secular trend measured by Veizer et al. (1980). Two factors likely control this increase in amplitude over the Veizer et al. (1980) dataset: (1) removal of diagenetic alteration effects which tend to attenuate variation and (2) exclusive analysis of abiotic marine precipitates. Our slightly higher 6~3C values (in comparison with those of Lindh, 1983), may also be due to analysis of abiotic precipitates (e.g., Gonzalez and Lohmann, 1985; McConnaughey, 1989; Carpenter et al., 1991). Modeling attempts subsequent to the Veizer et al. (1980) and Lindh (1983) data compilations are faced with a dilemma. How can the findings of Veizer et al. (1980), which have been used to explain carbon, sulfur, and oxygen mass balance calculations (and do not include fluxes associated with seafloor hydrothermal activity; e.g., Lasaga et al., 1985; Berner, 1987, 1989; Berner and Canfield, 1989; Kump, 1989), be reconciled with the results of Lindh (1983), the 6~3C values of which have been extensively used in these models? The simultaneous use of these results seems contradictory (see Walker, 1986). Therefore, reinterpretation of the relation between marine carbonate 6 ~3C values and sulfate 6~S values in the context of seafloor hydrothermal flux estimates is needed. Our correlation between marine cement 613C values and marine sulfate 634S values (Claypool et al., 1980) is highly significant (r 2 = 0.80; a > 95%), with a slope of the principal axis equal to -0.24. 613C = -0.24 6~*S + 7.4

(3)

This slope is nearly two times as large as that calculated by Veizer et al. (1980) and strongly suggests that parameters used in the model calculations (Eqn. 2) are in error. As the carbon-sulfur mole fraction ( - 1 5 / 8 ) is fixed by the stoichiornetry of the overall equation, the larger slope derived in

4837

this study requires an increase in either the FslFc or Ac/As parameter (see Eqn. 2). It is unlikely that variation in Ac/As could account for the increased slope (d613CId634S). Canfield and Teske (1996) report a relatively constant sulfur fractionation factor for bacterial sulfate reduction during the Phanerozoic. Poppet al. (1989) indicate that the 613C value of geoporphyrins decreases as atmospheric pCO2 increases during the Cenozoic-Meosozoic. They also note that the 613C value of terrestrial organic matter has remained relatively constant during the same time interval. However, Farquhar et al. (1982) have shown that the 6~3C value of C3 plants decreases with increases in pCO2. Given the generally accepted concept that the Mesozoic and Early Paleozoic were times of high pCO2, 613Cco~gshould be low and 613Cc~bo~t~should be high for these time periods. Marine cement data presented here do not support such a pCO2-related change. However, the large increase in marine cement 6~3C values may bear some relation with the evolution of vascular land plants and marine phytoplankton. Our recalculation, which accounts for transfers between oxidized and reduced reservoirs of carbon and sulfur, leads to a similar Ac/As (0.85) relative to values of previous studies (0.51-0.71; Table 1). As noted by Bemer and Ralswell (1983), changes in FslFc could resuk from a shift in the site of organic carbon burial from oceanic to continental environments. On the basis of data presented here, it appears that input values derived from the present-day fiver fluxes (FslFc) and used in previous models, grossly underestimate the average FslFc for the Phanerozoic. Solving for Fs/Fc, using the steeper slope derived in this study and the Ac/ As value of other studies, indicates that Phanerozoic FslFc values would have to be 1.6-3.3 times that of today (Table 3). Another possibility is that the oxidation-reduedon cycle is not limited to continental inputs of carbon and sulfur. If so, both the flux ratio (Fs/Fc) and fractionation ratio (Ac/ As) must be recalculated using the following equations: Fx(total) = Z (Fxi + Fxj + • . - )

(4)

Ax (total) = Z ([Fxl Axd + [Fx/A~j] + . • .) (F~ + Fxj + • • • )

(5)

where x is either carbon or sulfur, and i and j are possible reservoirs such as rivers and mantle. An important component of the global carbon-sulfur cycle may be the addition of mantle-derived sulfur to the oceans via high-temperature seawater-basalt interaction (e.g., Edmond et al., 1979; Veizer, 1983; Von Datum et al., 1995). When oxidized to sulfate, this sulfur source affects the global sulfur mass balance as well as the isotopic composition of seawater (Walker, 1986). Using an average 634S value for seafloor hydrothermal sulfide sulfur of +3.5%o (compiled from various sources--East Pacific Rise 21°N) and the equations above, the flux of mantle-derived sulfur (Fs mantle) relative to the flux of river sulfur (Fs aw) required to generate a 6~3C-6~S value slope of -0.24 has been determined (Tables 3, 4). The most significant aspect of these calculations is the inclusion of sulfur fluxes (as H2S or sulfide minerals) from seafloor hydrothermal systems (0.5-0.8 × 1012 moll

4838

S.J. Carpenter and K. C Lohmann

Table 3. Increase in sulfur flux required to approximate a steadystate dtS13C/d6345 of -0.24. FS mantle

Study

Fs ,ve,/Fs~o,t

Fs ,,~jflFs....

(in 10~2mol/yr)

Veizer et al. (1980) Walker (1986) Lasagaet al. (1985)

3.30 1.60 2.50

1.3 0.4 2.5

2.4 1.1 3.8

Note: Increasesin sulfurflux partitionedbetweenriver (Fs~,ve,)and mantle (Fsm~t~o)sourcesrelativeto publishedcontinentalsulfur fluxes(Fs~,t).

yr). These values are comparable to previous H2S flux estimates (1 x 1012 mol/yr; Edmond et al., 1979). If there is a net positive flux of sulfur to seawater from seafloor hydrothermal systems, the modern marine sulfur budget is not balanced (both with respect to mass and isotopic composition; e.g., Walker, 1986). These imbalances require the use of short-term fluxes for either hydrothermal anhydrite precipitation or dimethyl-sulfide ((CHa)2S or DMS) production to achieve a mass balance in today's oceans. Although the modern flux is negligible, another potential sulfur sink is the deposition of evaporites on continental shelves. This flux has not been modeled but is significant at various time periods during the Phanerozoic (e.g., Late Permian). Fluxes for removal of sulfur from seawater via anhydrite precipitation during high temperature seawaterbasalt interactions and/or DMS production by phytoplankton have been included in our model (Table 4). These fluxes do not greatly affect the 634S value of seawater sulfate as there is little or no isotopic fractionation during formation. (Tables 1, 3, 4). We have used the input parameters listed in Table 4 and Eqns. 4 and 5 to model the isotope ratios and fluxes needed to produce the empirically derived d613C/ d~34S value of -0.24. High temperature precipitation of anhydrite during seawater-basalt interaction is a potentially large sink for sulfur (3.75 x 1012 mol/yr; Edmond et al., 1979; McDuff and Edmond, 1982; Von Damm et al., 1985a) and is comparable to the river sulfate flux (3.7 × 10 ~2 mol/yr; Edmond et al., 1979 and 3.0 x 10 ~2 mol/yr; Meybeck, 1979). Although hydrothermal anhydrite precipitation could potentially provide a modern mass balance for the sulfur system, it is thought to be a geologically short-term sink as anhydrite dissolves at temperatures below approximately 150°(2 (Edmond et al., 1979; McDuff and Edmond, 1982; Von Datum et al., 1995; Humphris et al., 1995b). The time needed to dissolve anhydrite in seafloor hydrothermal systems and the degree to which it is dissolved is important in understanding the modern mass balance of sulfur. The absence of large quantities of anhydrite in massive sulfide deposits and presence of extensive dissolution breccias in these systems (Humphris et al., 1995b) clearly indicate that retrograde dissolution of anhydrite occurs. Assuming a spreading rate of 5 mm/yr, and that high temperature areas adjacent to active seafloor hydrothermal systems occur within 0.5 to 5 km of black smokers, we estimate that the time needed to produce off-axis temperatures of less than 150°C is approximately 105-106 years. Therefore, it appears that precipitation of anhydrite in hydrothermal systems is

not a significant sink for sulfur at timescales of greater than ]06 years. If we assume that approximately 20% of this anhydrite survives retrograde conditions, then the previously stated anhydrite flux of 3.75 x 10 ~2 mol/yr is reduced to 7.5 x 10 u mol/yr. This significantly modifies mass balance calculations and necessitates a sink for the sulfur being added to seawater by seafloor hydrothermal systems. Another potential sulfur flux is the removal of dimethylsulfide to the atmosphere by marine phytoplankton (e.g., Restelli and Angeletti, 1993). This flux is estimated to be 1.1 × 10 ~2 mol S / y r by Brimblecombe et al. (1989) and between 0.6 and 1.25 x 10 ~2 mol S / y r by Andreae and Crutzen (1997). Analysis of marine aerosol particles indicates that the 6345 value of DMS is approximately + 15.6%o (_+3.1%o) (Calhoun et al., 1993). Given these preliminary results, DMS may play a role in isotopic mass balance calculations. However, photooxidation of DMS produces SO2 and subsequent incorporation by atmospheric particulates suggests that the residence time of DMS in the atmopshere is not geologically significant (Restelli and Angeletti, 1993; Andreae and Crutzen, 1997). Incorporation of DMS-derived sulfur in soils may enhance the effectiveness of this flux, but the effects of these processes are currently not well understood. The evolution of marine phytoplankton and related DMS production may also effect paleoclimate and the ~345 value of ancient oceans. Coccolithophorids were evolved and well established by the Middle Mesozoic (Bramlette, 1958; Tappan and Loeblich, 1973; Perch-Nielsen, 1985 ) and they have been documented in rock as old as the Pennsylvanian-Permian (Gartner and Gentile, 1972; Minoura and Chitoku, 1979; Pirini-Radrizzani, 1970). Thus, it is likely that evolution of phytoplankton sometime in the late Paleozoic produced a modification of the sulfur system with a potentially significant impact on global climate (Andreae and Crutzen, 1997). Given that the ~348 value of DMS appears to be several permil lower than modern seawater, further examination is warranted.

3.1. 02 Consumption and Carbon-Sulfur Cycling Walker (1986) has discussed the sulfur fluxes associated with seafloor hydrothermal systems and the impact of these fluxes on the negative correlation between 6~3C and 634S values, burial of organic matter, and the 02 content of seawater. Like the Walker (1986) discussion of Cretaceous isotope data, we draw a similar conclusion that seafloor hydrothermal systems are responsible for the steep slope of this negative correlation. Implicit in our model is the concept that the flux of sulfur (as well as Fe, Mn, CI-h, and H2) from seafloor hydrothermal systems controls the burial of organic carbon through the consumption of 02 during inorganic and organically-mediated oxidation reactions (e.g., Jannasch, 1995; Lilley et al., 1995). By combining this concept with the accepted model linking the carbon and sulfur systems (i.e., burial of organic carbon and subsequent bacterial sulfate reduction and sedimentary pyrite formation), we have a more complete estimate of these related systems. This has important implications for understanding the evolution of oxygen in the ocean and atmosphere (e.g., Berner and Can-

C isotope ratios of Phanerozoic marine cements field, 1989; Lasaga, 1989; Canfield and Teske, 1996), and carbon and phosphorus cycling (e.g., Van Cappellen and Ingall, 1994, 1996; Filipelli and Delaney, 1994; Wheat et al., 1996). Van CappeUen and Ingall ( 1994, 1996) estimate that modem O2 removal fluxes are 1.65 X 10 ~2 mol/yr for FeS2 weathering and 3.75 × 1012 mol/yr for oxidation of organic matter. Their model assumes that the flux of reduced sulfur from seafloor hydrothermal systems (and thus related consumption of 02 via sulfide weathering) is negligible. If we consider the oxidation reactions described by Jannasch (1995), there are numerous pathways by which 02 is consumed in seafloor hydrothermal systems. If O2 is consumed on a mole for mole basis, then the consumption of 02 by seafloor hydrothermal sulfur alone is on the order of 1 × 1012 mol/yr. This is significant when compared with other modem O2 flux estimates. If 02 is also consumed by various reactions with Fe, Mn, and CI-L (e.g., Jannasch, 1995; Lilley et al., 1995 ), O2 fluxes associated with seafloor hydrothermal systems could be comparable with those associated with oxidation of organic matter. Therefore, determining the scale at which reduced species of S, C, Fe, and Mn are produced in sea:floor hydrothermal systems is important in understanding the relation between the C-O-S-P systems. The impact of a variable seafloor hydrothermal flux on the C-S system is dependent on the amount of O2 in the atmosphere and ocean. There is presently more 02 in the atmosphere than is dissolved in the ocean (Atmosphere: 3.8 × 1019 moles 02; Ocean: ~1016-1017 moles of 02; Kump and Garrels, 1986). During the pre-Cambrian, it is likely that these conditions were significantly different, with lower all-around 02 contents (e.g., Walker et al., 1983; Veizer, 1983). Thus, variation in the seafloor hydrothermal flux of S, Fe, Mn, CH4, etc. could potentially have a profound impact on atmospheric and oceanic O2 contents (e.g., Veizer, 1983). We suggest that under such conditions, elevated seafloor hydrothermal activity alone or in concert with modified ocean circulation (e.g., sluggish circulation and stratification) could produce widespread ocean anoxia and subsequent burial of organic matter. Conversely, diminished seafloor hydrothermal activity could increase 02 contents. Betts and Holland ( 1991 ) have shown that the burial efficiency of organic matter (the ratio of the quantity of organic carbon that is buried to that which reaches the sediment water interface) in modem sediments is more strongly corre-

Table 4. Input parameters for revised carbon-sulfur isotopic massbalance calculations. Process/Material

634S (in %°)

6~3C (in %0

Flux (in 1012mol/yr)

River Sulfate HydrothermalSystemSulfides HydrothermalSystemAnhydrite SedimentaryPyrite Dimethyl-sulfide River Bicarbonate Organic Matter Marine Carbonates

10 3.5 20 -20 16 ----

------4 -25 1

3.2 0.58 to 0.76* -2.7 to 0 - 1.3 -1.2 to 0 25 -6.6 - 18.4

* Range of values dependenton values assignedto other fluxes.

4839

lated with sedimentation rate than the 02 content of bottom waters. This is in contrast to the Emerson (1985) conclusion that only sedimentation rates above 10 cm/kyr significantly increase carbon preservation. Although sedimentation rate plays an important role in the burial/oxidation of marine organic matter, a decrease in bottom water 02 concentration will decrease the depth to which 02 will penetrate sediments. If the role of other oxidizing agents (NO3, MnO2, FeO (OH), and SO2 ) in organic carbon oxidation is negligible (e.g., Emerson, 1985), bottom water O2 concentrations must ultimately be related to organic carbon fluxes. For the purposes of this discussion we directly link the 02 content of ambient seawater the burial/oxidation of organic matter. Thus, low 02 contents enhance burial of organic matter and high 02 levels enhance oxidation of organic matter. As we cannot call upon unusually high or low sedimentation rates for individual time periods (e.g., Cretaceous vs. Permian vs. Devonian) we assume that sedimentation rates are constant.

3.2. Explanations and Implications of the 8~3C and 8 ~ Value Trend The relation between the C-O-S systems, orogenic activity, and sea level is enigmatic (e.g., Veizer, 1985). However, interpretation of marine 613C and 634S values is possible if taken in the context of coeval strontium and oxygen data, modem seafloor hydrothermal fluxes, and large-scale tectonic activity. We attempt to explain the observed 613C-634S value relation in this manner (Figs. 6, 7). As suggested by Walker (1986), the carbon and sulfur systems appear to be more complicated than the balance between riverine inputs and burial of organic carbon and sedimentary pyrite formation. Seafloor hydrothermal fluxes figure prominently in our reinterpretation. Invoking variation in seafloor hydrothermal activity as the process that controls these various isotopic tracers over Phanerozoic type timescales requires a mechanism for producing changes in mantle heating and convection. Supercontinent assembly and breakup is a likely mechanism (e.g., Anderson, 1982, 1994; Gurnis, 1988; Zhong and Gumis, 1993). We suggest that ocean chemistry responds to episodic or periodic changes (both long- and short-term) in the generation of oceanic crust. The relation between ocean chemistry changes and seafloor hydrothermal activity, mantle convection, eruption of large igneous provinces (LIPs; Coffin and Eldholm, 1993, 1994), and super-continent assembly and break-up (e.g., Veevers, 1990) is worthy of further examination (e.g., Storey et al., 1995; Yale et al., 1996). Furthermore, time periods associated with oceanic anoxic and mass extinction events, and evaporite and phosphorite deposition should be re-examined for potential contributions from seafloor hydrothermal systems (e.g., Carpenter and Mitterer, 1996). On the basis of Phanerozoic 6~3C-634S value variation alone, it is not be possible to unambiguously determine which fluxes have changed to produce the observed variations. Additional isotopic and/or elemental tracers characteristic of either continental lithosphere or mantle contributions are required to adequately test the relative contributions of each in defining the evolutionary pathway of paleocean chemistry. For example, secular variation in STSr/a6Sr ratios

4840

A

S.J. Carpenter and K. C Lohmann 0.70950

@, 0.70900

I. r~

. . . . . . . . . . . . . . .

.

.

.

!

.

0.70850 ..................

0.70800

..............

0,70750

0.70700 ...... ....... 0.70650 - lO

' , ,s,, , , s-. = . o . ?, ,/,,+

15

,

, = o , .

20

25

r

30

~ S (CDT) B

0.70~0

.... "H"

0.-,0900

:

~c)

_

I

US, .

Ms

;

r~

g 0.70700

. . . . . . . . .

~

~mJ

~Sr/16$r= 0.70853- 1.9x 104lil3C;r1= 0.19 0.70650

~

, , , ,

-2

-1

0

-

1

~ilaC

2

3

4

5

(PDB)

Fig. 6. (a) Plot of marine carbonate 87Sr/86Sr ratios and marine evaporite 6 34S values (from Table 2; Denison et al., 1997; Claypool et al., 1980, respectively). (b) Plot of marine carbonate 87Sr/86Sr ratios and marine cement 5 ~3C values (from Table 2; Denison et al.. 1997). Lines and inset equations represent linear regressions of the plotted data. In both plots, data from the lower Paleozoic and the late Paleozoic-Mesozoic plot separately; Devonian, Pliocene and Holocene data occupy intermediate positions. Compare with Fig. 1. Although there is only a weak correlation between these values, likely due to the different cycling rates for each element, there is a general trend relating low 87Sr/a6Sr ratios and 634S values (mantlederived Sr and S) and high 6~3C values (enhanced organic carbon burial ).

has been documented in the marine carbonate record (Burke et al., 1982; Denison et al., 1997; Fig. 6). Although the reasons for short-term, second order variations are, as yet, unresolved (e.g., Brass, 1976; Holland, 1984; Richter et al., 1992; Richter and Turekian, 1993; Yale et al., 1996), there is general agreement that relatively high 875r/86Sr ratios indicate a dominance of continental lithosphere contribution, whereas low 87Sr/S6Sr ratios reflect an increased contribution of mantle-derived strontium to seawater. Interestingly, the general covariation of secular trends for 634S values and 875r/S6Sr ratios suggests that, like strontium isotope ratios, sulfur isotope ratios of paleoceans may, in part, reflect a long-term balance between continental weath-

ering and high temperature seawater-basalt exchange reactions (Fig. 6a). Likewise, the relation between 613C values and S7Sr/86Sr ratios suggests that there is a link between mantle fluxes and burial of organic matter (Fig. 6b). Division of Late Paleozoic-Mesozoic and Early Paleozoic data is readily seen (Fig. 6). Although the correlation between strontium and carbon and strontium and sulfur isotope ratios is not exceptional (due in part to the different cycling times for each element and the different processes acting to change each isotope ratio) these correlations suggest a link between mantle fluxes and the C-O-S system. Veizer et al. (1980), Garrels and Lerman (1981), and Walker (1986) have described the linear relation between 6 ]~C and 634S values as a function of the relation between reduced and oxidized forms of carbon and sulfur and the associated fractionation factors. Conventional wisdom suggests that the Phanerozoic relation between these two tracers indicates the balance of organic carbon burial and the formation of sedimentary pyrite. We choose to depart from this line of reasoning because there is compelling evidence that seafloor hydrothermal fluxes of sulfur should be incorporated into isotopic mass balance calculations (e.g., Edmond et al., 1979; McDuff and Edmond, 1982; Walker, 1986; Von Damm et al., 1995; Humphris et al., 1995b), By interpreting the relative position of time periods along the linear array of t5~3C and 634S values and in the context of other secular trends (875r/86Sr ratios, 6~80 values), the Phanerozoic can be divided into pre- and post-Devonian time periods, with the Devonian, Late Cretaceous, Tertiary, and Holocene having intermediate values (Figs. 1, 7). Our interpretation of the variation observed in Phanerozoic marine /5~3C and (~348 values relies on the contribution of mantle-

Phanerozoic Marine Cements and Sulfates

4 ~

~

~ZY

/ ~

I' o

.,°!' "" "2

~ ~--

1O

I

12

i.,

-T.-~--

~

14

16

i

, [ ....

18

20

~534SM a r i n e

i

22

I

4 r

I

24

I

I ....

26

I ....

28

"2

30

Sulfate

Fig. 7. Schematic representation of processes responsible for the observed 6~3C and 634S values of marine carbonates and sulfates. The Late Permian is a time of high seafloor hydrothermal activity and the early Paleozoic (Cambrian-Silurian) is a period of low seafloor hydrothermal activity. The intermediate values for the Devonian, Late Cretaceous, Pliocene, and Holocene suggest an intermediate amount of hydrothermal activity and may indicate time periods of transition into and out of the late Paleozoic-Mesozoic high in seafloor hydrothermal activity.

C isotope ratios of Phanerozoic marine cements derived sulfur relative to other fluxes. As a result, the relative position of time periods in Fig. 1 are indicative of the relative importance of seafloor hydrothermal fluxes and the concomitant response of the carbon system to changes in O2 (Fig. 7). In this scenario, seafloor hydrothermal sulfur fluxes are high during the late Paleozoic to Mesozoic and low during the early Paleozoic (Cambrian-Silurian). This hypothesis is supported by high STSr/S6Srratios (Denison et al., 1997; Fig. 6) and low/5 tso values of early Paleozoic marine carbonates (e.g., Walker and Lohmann, 1989; Carpenter et al., 1991). Marked by Phanerozoic highs in/5 t3C values and lows in 87Sr/SrSr ratios and/534S values, the late Permian is truly an unusual time period (e.g., Claypool et al., 1980; Given and Lohman, 1985; Kramm and Wedepohl, 1991; Martin and Macdougall, 1995; Denison et al., 1997; Figs. 1, 6, 7). These anomalous isotope ratios correlate with significant volumes of gypsum (Castille Fm. of W. Texas/New Mexico and Zechstein of Germany) and organic-rich phosphorite deposition (Phosphoria Fm. of the Western United States). These events also coincide with the eruption of the Siberian Traps (the largest known LIP/flood basalt; e.g., Renne et al., 1995). Carpenter and Mitterer (1996) and Yale et al. (1996) have described the relation between eruption of large igneous provinces (LIPs), the surface expression of mantle plumes, and seafloor hydrothermal activity during this time period. These events have been discussed by various authors and suggest elevated mantle fluxes associated with the assembly and break-up of Pangea (e.g., Veevers et al., 1994; Veevers and Tewari, 1995; Yale et al., 1996; Carpenter and Mitterer, 1996). Late Permian ocean chemistry may have responded to an influx of mantle-derived sulfur from seafloor hydrotherreal systems which contributed to burial of organic matter and widespread gypsum and phosphorite deposition (Carpenter and Mitterer, 1996). The Devonian and Late Cretaceous have /513C and/534S values that are intermediate between those of the late Paleozoic-Mesozoic (elevated seafloor hydrothermal activity) and the early Paleozoic (diminished seafloor hydrothermal activity). These compositions may reflect the transitions associated with the assembly and break-up of Pangea, respectively (Figs. 1, 7). The Pliocene and Holocene/513C and/534S values may be the result of increased weathering fluxes from the Himalaya mountains and/or decreases in seafloor hydrothermal fluxes relative to the Mesozoic. Interestingly, Holocene (and Pliocene) 6~3C and/534S values are near the median values for the Phanerozoic. On the basis of/5~3C and 634S values presented here, the Late Paleozoic and Mesozoic is a time of elevated seafloor hydrothermal fluxes probably associated with mantle convection and the assembly and break-up of Pangea (e.g., Anderson, 1982, 1994; Gurnis, 1988; Zhong and Gurnis, 1993; Fig. 7). Heller et al. (1996) have suggested that there is no geochemical evidence that indicates that the Cretaceous was a time of significantly elevated seafloor hydrothermal activity and that there is no evidence for a superplume event between ~80 and 120 Ma as suggested by Larson (1991a,b). Although not arguing for a superplume, the/513C and/534S values from Cretaceous and other Mesozoic samples argue that seafloor hydrothermal fluxes were higher than modern values during this interval (Fig. 7). Walker (1986) sug-

4841

gested that the relation between /513C and /534S values of Cretaceous carbonates and sulfates indicated a greater contribution of mantle-derived sulfur (via seafloor hydrothermal systems), than previously suggested by Veizer etal. (1980). Our data are consistent with the conclusions of Walker (1986). Likewise, similar data from Proterozoic marine carbonates and sulfates also support these conclusions (see next section; Strauss, 1993). Modeling of marine STSr/S6Sr ratios indicates that the time interval during the Cretaceous normal superchron requires a moderate increase in the seafloor hydrothermal flux of strontium ( ~ 15-20% higher than modern flux estimates; Yale et al., 1996). Therefore, we conclude that seafloor hydrothermal fluxes were elevated in the MidCreteaceous (relative to the Holocene), regardless of the relation between seafloor hydrothermal activity, oceanic crust production, and sea level. The relation between sea level and mid-ocean ridge volume described by Gaffin (1987) is difficult to reconcile with our findings. Using the Hays and Pitman (1973) and the Gaffin (1987) conceptual models, the low sea levels of the Permian-Triassic should indicate smaller ridge volumes and lower seafloor hydrothermal fluxes (thus high /534S values and S7Sr/S6Sr ratios). However, our findings suggest that seafloor hydrothermal fluxes were at a Phanerozoic high. Veizer (1985) also found a similar relation between sea level and carbon, sulfur, and strontium isotope ratios. Carpenter and Mitterer (1996) suggested that the assembly of Pangea produced mantle heating and a thermally buoyant supercontinent. This, together with a significant glacial ice volume during the Permian (Frakes et al., 1992), may have produced a low relative sea level during a time with a high seafloor hydrothemal flux. These conclusions suggest that the use of first order sea level variation as a proxy of oceanic crust production (and by extrapolation, seafloor hydrothermal activity) should be re-examined. 3.3. Proterozoic Carbon-Sulfur Isotope Relations Strauss (1993) has also described a linear relation between 6 '3C and 634S values in Proterozoic marine carbonates and sulfates: 613C = -0.50/534S + 13.4 (r 2 = 0.68). This relation has been determined by replotting the sulfur isotope data from Strauss (1993) and the carbon isotope data of Derry et al. (1992). The steeper d/s13C/d/534S slope of the Strauss (1993) trend ( - 0 . 5 0 vs. -0.24 for the Phanerozoic data presented here) is interpreted as the result of larger seafloor hydrothermal fluxes during the Proterozoic. Assuming that only the seafloor hydrothermal flux of sulfur has increased and that other sulfur and carbon fluxes and fractionations are similar to modern values, this steeper slope requires that the Proterozoic seafloor hydrothermal flux of sulfur be approximately 50% higher than that of the Phanerozoic. Asmerom et al. (1991) and Derry et al. (1992) have also suggested that seafloor hydrothermal fluxes were higher during the Neoproterozoic. Asmerom et al. ( 1991 ) estimate that for much of the Neoproterozoic the river/hydrothermal Sr flux ratio is 75% (_+15) lower than modern values. Considering the errors associated with these independent estimates, this correspondence is encouraging. Derry etal. (1992) have suggested that 02 levels in Neo-

4842

S.J. Carpenter and K. C Lohmann

proterozoic seawater were controlled in part by the oxidation of H2S, Fe, and Mn produced by seafloor hydrothermal systems. On the basis of the very different d~13C/dt~345 slopes for the Proterozoic and Phanerozoic, we agree with this assessment. In addition, we conclude that the dramatic evolutionary changes of the late Neoproterozoic-early Phanerozoic were brought about by the waning of seafloor hydrothermal fluxes and a subsequent increase in the O2 content of seawater (e.g., Knoll, 1991 ). We suggest that the decreased demand for O2 by hydrothermal systems, and the resulting availability of 02 in seawater contributed significantly to the rapid evolution of metazoans observed in the late Neoproterozoic and early Paleozoic (e.g., Raft and Raff, 1970; Runnegar, 1991). The modifications of biogeochemical cycles at the Proterozoic/Cambrian transition described by Logan et al. (1995) are not mutually exclusive with this proposed mechanism for increasing the O: content of Phanerozoic seawater. These evolutionary changes and reorganization of biogeochemical cycles occur at the same time that marine STSr/S6Sr ratios increase dramatically, indicating a relative decrease in the seafloor hydrothermal to continental weathering fluxes (e.g., Asmerom et al., 1991; Denison et al., 1997). These 87Sr/SrSr ratios remain relatively high throughout the early Paleozoic (Cambrian-Devonian) and roughly coincide with the time interval following the breakup of Gondwanaland prior to the assembly of Pangea (e.g., Veevers, 1990; Stem, 1994). Support for diminished seafloor hydrothermal fluxes during the early Paleozoic is provided by t~13C-6345 value relations (this study; Lindh, 1983; Strauss, 1993), 87Sr/86Sr ratios (Asmerom et al., 1991; Kaufman et al., 1993), 6~3C values (Walker and Lohmann, 1989; Derry et al., 1992; Kaufman et al., 1993) and 6~80 values (Walker and Lohmann, 1989; Carpenter et al., 1991 ). Together, these tracers indicate that the early Paleozoic was a period of relative mantle quiescence that was conducive to rapid metazoan evolution.

justified given data presented here and by Lindh (1983) ( - 0 . 2 4 and -0.26, respectively). We suggest that oxidation of sulfide minerals, HzS, Fe, Mn, and CH4, etc. associated with seafloor hydrothermal activity consumes dissolved 02 in seawater. The extent to which 02 is consumed controls the oxidation or burial of marine organic matter. Given this role, seafloor hydrothermal activity probably plays a significant role in producing oceanic anoxic events. Therefore, global sulfur fluxes (particularly those from seafloor hydrothermal activity) control global carbon fluxes. This is unlike currently accepted models of carbon and sulfur cycling which have organic carbon burial fluxes controlling bacterial sulfate reduction and the flux of sulfur into formation of sedimentary pyrite. Two sulfur sinks, anhydrite precipitation during high-temperature seawater-basalt interaction and dimethyl-sulfide formation by marine phytoplankton, are required to balance the modem sulfur budget (assuming negligible evaporite deposition). These fluxes appear to be of short duration (<106 yr) and may, therefore, have little impact on longterm flux estimates. However, as flux estimates for these processes are not well constrained, inclusion of these processes in isotopic mass balance calculations may be warranted. The dt~13C/d~348 relation for Proterozoic marine carbonates and sulfates is significantly steeper than the analogous trend for the Phanerozoic ( - 0 . 5 0 vs. - 0 . 2 4 ) . To achieve this slope, we estimate that Proterozoic seafloor hydrothermal fluxes must be approximately 50% higher than modem values. This estimate is consistent with previously reported flux estimates for this time period. The dramatic evolutionary changes of the NeoproterozoicPaleozoic transition coincide with an apparently significant decrease in seafloor hydrothermal fluxes. We speculate that this waning of seafloor hydrothermal activity increased the amount of dissolved 02 in seawater and facilitated the rapid evolution of metazoans observed in the Vendian and early Cambrian.

4. CONCLUSIONS Using an isotopic mass balance model modified for seafloor hydrothermal system fluxes, d~13C/d6345 can be predicted from Fs/Fc values estimated for today's oceans. A d613C/d634S value of approximately - 0 . 2 4 accounts for seafloor hydrothermal venting of reduced sulfur (H2S and sulfide minerals), anhydrite precipitation during seawaterbasalt interaction, and the production of dimethyl-sulfide by phytoplankton (Table 3 ). These calculations are comparable to actual Phanerozoic 613C-634S data and suggest that seafloor hydrothermal systems and related fractionations were operative to varying degrees throughout the Phanerozoic. On the basis of our re-examination of marine carbon and sulfur systems, we make the following conclusions: The flux of mantle-derived sulfur into seawater from seafloor hydrothermal systems is significant and plays an important role in determining marine 634S values. Without this flux, the observed dt~13C/d~34S relation is significantly underestimated ( - 0 . 1 0 vs. - 0 . 2 4 ) . The generally accepted fluxes associated with d~13C/d¢~34Svalues of - 0 . 1 0 are not

Acknowledgments--We thank the many colleagues whose efforts have provided the extensive data on which this synthesis is based. We are grateful to J. C. G. Walker, R. K. Given, B. H. Wilkinson, R. M. Mitterer, M. Arthur, R. Berner, and L. Kump for their valuable contribution of ideas, encouragement, and critical reviews of earlier versions of the manuscript. Three anonymous reviews also significantly enhanced the quality of this work. This research was supported in part by funds from Marathon Oil Company, National Science Foundation Grant EAR-8115840 to KCL and by Texas Advanced Research Program Grant #009741-064 to SJC. REFERENCES

A1-Aasm I. S. and Veizer J. (1982) Chemical stabilization of lowMg calcite: An example of Brachiopoda. J. Sediment. Petrol. 52, 1101-1109.

Anderson D. L. (1982) Hotspots, polar wander, Mesozoic convection, and the geoid. Nature 297, 391-393. Anderson D. L. (1994) Superplumes or supercontinents? Geology 22, 39-42. Andreae M. O. and Crutzen P. J. (1997) Atmospheric aerosols: biogeochemical sources and role in atmospheric chemistry. Science 276, 1052-1058. Asmerom Y., Jacobsen S.B., Knoll A.H., Bunerfield N.J., and Swett K. (1991) Strontium isotopic variations of Neoproterozoic

C isotope ratios of Phanerozoic marine cements seawater: Implications for crustal evolution. Geochim. Cosmochim. Acta 55, 2883-2894. Banner J. L. and Hanson G. N. (1990) Calculation of simultaneous isotopic and trace element variations during water-rock interaction with applications to carbonate diagenesis. Geochim. Cosmochim. Acta 54, 3123-3137. Berner R. A. (1987) Models for carbon and sulfur cycles and atmospheric oxygen: application to Paleozoic geologic history. Araer. J. Sci. 287, 177-196. Berner R. A. (1989) Biogeochemical cycles of carbon and sulfur and their effect on atmospheric oxygen over Phanerozoic time. Palaeogeogr. Palaeoclimatol. Palaeoecol. 75, 97-122. Berne, R. A. (1990) Atmospheric carbon dioxide levels over Phanerozoic time. Science 249, 1382-1386. Beruer R. A. and Canfield D. (1989) A model for atmospheric oxygen over Phanerozoic time. Amer. J. Sci. 289, 333-361. Berner R. A. and Raiswell R. (1983) Burial of organic carbon and pyrite sulfur in sediments of Phanerozoic time: A new theory. Geochim. Cosmochim. Acta 47, 855-862. Berner R. A., Lasaga A. C., and Garrels R. M. (1983) The carbonate-silicate geochemical sycle and its effect on atmospheric carbon dioxide over the past 100 million years. Amer. J. Sci. 283, 641683. Betts J. N. and Holland H. D. ( 1991 ) The oxygen content of ocean bottom waters, burial efficiency of organic carbon, and the regulation of atmospheric oxygen. Palaeogeogr. Palaeoclimatol. Palaeoecol. 97, 5-18. Brandette M. N. (1958) Significance of cocoliths in calcum-carbonate deposition. GSA Bull, 69, 121-126. Brand U. (1982) The oxygen and carbon isotope composition of Carboniferous fossil components: Seawater effects. Sedimentology 29, 139-147. Brass G. W. (1976) The variation of the marine 87Sr/86Srratio during Phanerozoic time: Interpretation using a flux model. Geochim. Cosmochim. Acta 40, 721-730. Brass G. W. and Turekian K. K. (1977) The strontium isotopic composition of seawater and sea-water oceanic-crust interaction. Earth Planet. Sci. Len. 34, 165-166. Breining K. A. (1985) Stable isotopic cathodoluminescence study of the diagenetic evolution of a Silurian shelf reef: Evaluation of the timing and mechanism of dolomitization. M.S. Thesis: Univ. Michigan. Brenchley P. J. et al. (1994) Bathymetric and isotopic evidence for a short-lived Late Ordovician glaciation in a greenhouse period. Geology 22, 295-298. Brimblecombe P., Hammer C., Rodbe H., Ryaboshapko A., and Boutron C. F. (1989) Human influences on the sulphur cycle. In Evolution of the Global Biogeochemcial Sulphur Cycle ( ed. P. Brimblecombe and A. Lein), pp. 77-121. Wiley. Burke W. H., Denison R.E., Hetherington E. A., Koepnick R. B., Nelson N. F., and Otto J. B. (1982) Variation of seawater 87Sr/ 86Sr throughout Phanerozoic time. Geology 10, 516-519. Calhoun J., Delmas R., Nielsen H., Gravenhorst G., and Jouzel J. (1993) Stable sulfur isotope ratios: source indicators. In Dimethylsulphide: Oceans, Atmosphere, and Climate (ed. G. Restelli and G. Angeletti), p. 183. Kluwer Academic Publ. Canfield D. E. and Teske A. (1996) Late Proterozoic rise in atmospheric oxygen concentration inferred from phylogenetic and sulphur-isotope studies. Nature 382, 127-132. Carpenter S.J. (1991) Isotopic and minor element chemistry of Devonian-Carboniferous abiotic marine calcite. Ph.D. dissertation, Univ. Michigan. Carpenter S. J. and Lohmann K~ C. (1989) 6180 and 6 m3Cvariations in Late Devonian marine cements from the Golden Spike and Nevis reefs, Alberta, Canada. J. Sediment. Petrol. 59, 792-814. Carpenter S. J. and Lob.mann K. C. (1992) Sr/Mg ratios of modern marine calcite: Empirical indicators of ocean chemistry and precipitation rate. Geochim. Cosmochim. Acta 56, 1837-1849. Carpenter S. J. and Lohmarm K. C. (1995) 61sO and 613C values of modern brachiopod shells. Geochim. Cosmochim. Acta 59, 37493764. Carpenter S. J. and Mitterer R. M. (1996) Global change, phosphorite deposition and superplume-related seafloor hydrothermal ac-

4843

tivity during the Late Permian. Geol. Soc. Amer. Abstr. Prog., 225, Vol. 8. Carpenter S.J., Lohmann K. C., Holden P., Walter L. M., Huston T., and Halliday A. N. (1991) 6180 values, 87Sr/~Sr and Sr/Mg ratios of Late Devonian abiotic marine calcite: Implications for composition of ancient seawater. Geochim. Cosmochim. Acta 55, 1991-2010. Cercone K. R. and Lohmann K. C. (1986) Diagenetic history of a middle Silurian pinnacle reef from the Michigan Basin. In Reef Diagenesis (ed. J. H. Schroeder and B. H. Purser), pp. 381-398. Elsevier. Claypool G. E., Holser W.T., Kaplan I. R., Sakai H., and Zak I. (1980) The age curves of sulfur and oxygen isotopes in marine sulfate and their mutual interpretation. Chem. Geol. 28, 199-260. Coffin M. F. and Eldholm O. (1993) Scratching the surface: Estimating dimensions of large igneous provinces. Geology 21, 515-518. Coffin M.F. and Eldholm O. (1994) Large Igneous Provinces: Crustal Structure, Dimensions, and External Consequences. Rev. Geophys. 32, 1-36. Compston W. (1960) The carbon isotopic compositions of certain marine invertebrates and coals from the Australian Permain. Geochim. Cosmochim. Acta 18, 1-22. Czemiakowski L.A., Lohmann K.C., and Wilson J.L. (1984) Closed-system marine burial diagenesis: Isotopic data from the Austin Chalk and its components. Sedimentoiogy 31, 863-877. Davies G. R. (1977) Former magnesian calcite and aragouite submarine cements in upper Paleozoic reefs of the Canadian Arctic: A summary. Geology 5, 11-15. Denison R. E., Koepnick R. B., Burke W. H., Hetherington E. A., and Fletcher A. (1997) Construction of the Silurian and Devonian seawater 87Sr/86Sr curve. Chem. Geol. (lsot. Geosci. Sect.) 146, 109-121. DePaolo D. J. and Ingram B. L. (1985) High-resolution stratigraphy with strontium isotopes. Science 227, 938-941. Derry L. A., Kanfman A. J., and Jacobsen S. B. (1992) Sedimentary cycling and environmental change in the Late Proterozoic: Evidence from stable and radiogenic isotopes. Geochim. Cosmochim. Acta 56, 1317-1329. Dunn P. A. (1988) Carbon and oxygen isotopic secular variation in Devono-Carboniferous marine carbonates of western Europe. M.S. Thesis, Univ. Michigan. Edmond J.M. et al. (1979) Ridge crest hydrothermal activity and the balances of major and minor elements in the ocean: The Galapagos data. Earth Planet. Sci. Lett. 46, 1-18. Emerson S. (1995) Organic carbon preservation in marine sediments. In The Carbon Cycle and Atmospheric C02: Natural Variations Archean to Present (ed. E.T. Sunquist and W.S. Broecker); Amer. Geophys. Union Geophys. Manogr. 32, 78-87. Erez J. (1978) Vital effects on stable-isotope composition seen in foraminifera and coral skeletons. Nature 273, 199-202. Farquhar G.D., O'Leary M.H., and Berry J.A. (1982) On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves. Australian J. Plant Physiol. 9, 121-137. Filipelli G. M. and Delaney M. L. (1994) The oceanic phosphorus cycle and continental weathering during the Neogene. Paleoceanography 9, 643-652. Frakes L. A., Francis J. E., and Syktus J. I. (1992) Climate modes of the Phanerozoic. Cambridge Univ. Press. Francois L. M. and Gerard J-C. (1986) A numerical model of the evolution of ocean sulfate and sedimentary sulfur during the last 800 million years. Geochim. Cosmochim. Acta 50, 2289-2302. Frank T. D. and Lohmann K. C. (1996) Diagenesis of fibrous magnesian calcite marine cement: Implications for the interpretation of 6180 and 613C values of ancient equivalents. Geochim. Cosmochim. Acta 60, 2427-2436. Gaffin S. (1987) Ridge volume dependence on seafloor generation rate and inversion using long term sealevel change. Amer. J. Sci. 287, 596-611. Garrels R. M. and Lerman A. (1981) Phaneruzoic cycles of Sedimentary carbon and sulfur. Natl. Acad. Sci. Proc. 78/8, 46524656. Garrels R. M. and Lerman A. (1984) Coupling of the sedimentary

4844

S.J. Carpenter and K. C Lohmann

sulfur and carbon cycles--an improved model. Amer. J. Sci. 284, 989-1007. Garrels R.M. and Perry E.A. (1974) Cycling of carbon, sulfur, and oxygen through geologic time. In The Sea: Marine Chemistry (ed. E. D. Goldberg), pp. 303-336. Wiley. Gartner S. and Gentile R. (1972) Problematic Pennsylvanian coccoliths from Missouri, Micropaleontology 18, 401-404. Gibbs M. T., Barron E. J., and Kump L. R. (1997) An atmopsheric pCO2 threshold for glaciation in the Late Ordovician. Geology 25, 447-450. Ginsburg R. N. and James N. P. (1976) Submarine botryoidal aragonite in Holocene reef limestones, Belize.Geology 4, 431-436. Given R. K. and Lohmann K. C. (1985) Derivation of the original isotopic composition of Permian marine cements. J. Sediment. Petrol. 55, 430-439. Gonzalez L.A. and Lohrnann K.C. (1985) Carbon and oxygen isotopic composition of Holocene reefal carbonates. Geology 13, 811-814. Graber E. R. (1989) Sedimentology, diagenesis, and secular 613C variations in the Upper Horquilla Limestone (Pennsylvania/Permian) of New Mexico. Ph.D. dissertation, Univ. Mich. Graham D. W., Corliss B. H., Bender M. L., and Keigwin k D., Jr. (1981) Carbon and oxygen isotopic disequilibria of recent deepsea benthic foraminifera. Mar. Micropaleontol. 6, 483-497. Gruszczynski M., Hoffman A., Malkowski K., Zawidzka K., Halas S., and Zeng Y. (1989) A brachiopod calcite record of the oceanic carbon and oxygen isotope shifts at the Permian/Triassic transition. Nature 337, 64-68. Gurnis M. (1988) Large-scale mantle convection and the aggregation and dispersal of supercontinents. Nature 332, 695-699. Hays J. D. and Pitman W. C., IH (1973) Lithospheric plate motion, sea level changes, and climatic and ecological consequences. Nature 246, 18-22. Heller P.L., Anderson D.L., and Angevine C.L. (1996) Is the middle Cretaceous pulse of rapid sea-floor spreading real or necessary? Geology 24, 491-494. Holland H. D. (1978) The Chemistry of the Atmosphere and Oceans. Wiley. Holland H.D. (1984) The Chemical Evolution of the Atmosphere and Oceans. Princeton Univ. Press. Holser W. T. (1977) Catastropic chemical events in the history of the ocean. Nature 267, 403-408. Holser W. T. (1984) Gradual and abrupt shifts in ocean chemistry during Phanerozoic time. In Patterns of Change in Earth Evolution (ed. H. D. Holland and A. F. Trendall), pp. 123-143. SpringerVerlag. Humphris S. E., Zierenberg R. A., Mullineaux L. S., and Thomson R. E. (eds.) (1995a) Seafloor Hydrothermal Systems. Amer. Geophys. Union Monogr. 91. Humphris S.E. et al. (1995b) The internal structure of an active sea-floor massive sulfide deposit. Nature 377, 713-716. Hurley N.F. (1986) Geology of the Oscar Range Devonian reef complex, Canning Basin, Western Australia. Ph.D. Dissertation, Univ. Michigan. Hurley N. F. and Lohmann K.C. (1989) Diagenesis of Devonian reefal carbonates in the Oscar Range, Canning Basin, Western Australia. J. Sediment. Petrol. 59, 127-146. James N. P. and Klappa C. F. (1983) Petrogenesis of Early Cambrian reef limestones, Labrador, Canada. J. Sediment. Petrol. 53, 1051-1096. Jannasch H.W. (1995) Microbial interactions with hydrothermal fluids. AGU Monogr. 91, 273-296. Kaufman A. J., Jacobsen S. B., and Knoll A. H. (1993) The Vendian record of strontium and carbon isotopic variations in seawater: Implications for tectonics and paleoclimate. Earth Planet. Sci. Letts. 120, 409-430. Keith M. L. and Weber J. N. (1965) Systematic relationships between carbon and oxygen isotopes in carbonates deposited by modern corals and algae. Science 150, 498-501. Knoll A. H. ( 1991 ) End of the Proterozoic Eon. Scientific American, (October), 64-73. Knoll A. H., Bambach R. K., Canfield D. E., and Grotzinger J. P.

(1996) Comparative Earth history and late Permian mass extinction. Science 273, 452-457. Kramm U. and Wedepohl K. H. ( 1991 ) The isotopic composition of strontium and sulfur in seawater of Late Permian (Zechstein) age. Chem. Geol. 90, 253-262. Kump L. R. (1989) Alternative modeling approaches to the geochemical cylces of carbon, sulfur, and strontium isotopes. Amer. J. Sci. 289, 390-410. Kump L. R. and Garrels R. M. (1986) Modeling atmospheric 02 in the global sedimentary redox cycle. Amer. J. Sci. 286, 337-360. Kump L. R., Gibbs M.T., Arthur M.A., Patzkowsky M.E., and Sheehan P. M. (1995) Hirnantian glaciation and the carbon cycle. In Ordovician Odyssey (ed J. D. Cooper et al.); 7th Intl. Symp. Ordovicain System SEPM, 299-302. Kump L. R., Arthur M. A., Gibbs M., Patzkowsky M., and Pinkus D. (1996) Isotopic and modeling constraints on late Ordovician carbon dioxide levels: Climate paradox solved? Geol. Soc. Amer. Annu. Mtg. Abstr. Prog., 226, Vol. 28. Land L. S. and Goreau T. F. (1970) Submarine lithification of Jamaican reefs. J. Sediment. Petrol. 40, 457-462. Larson R. L. (1991a) Geological consequences of superplumes. Geology 19, 963-966. Larson R.L. (1991b) Latest pulse of Earth: Evidence for a midCretaceous superplume. Geology 19, 547-550. Lasaga A. C. (1989) A new approach to the isotopic modeling of the variation of atmospheric oxygen through the Phanerozoic. Amer. J. Sci. 289, 411-435. Lasaga A. C., Bemer R. A., and Garrels R. M. (1985) A geochemical model of atmospheric CO2 fluctuations over the last 100 million years. In The Carbon Cycle and Atmospheric C02: Natural Variations Archean to Present (ed. E.T. Sunquist and W.S. Broecker); Amer. Geophys. Union Geophys. Monogr. 32, 397411. Lilley M. D., Feely R. A., and Trefry J.H. (1995) Chemical and biological transformations in hydrothermal plumes. AGUMonogr. 91, 369-391. Lindh T. B. (1983) Temporal variations in t3C and 34S and global sedimentation during the Phanerozoic. M.S. thesis, Univ. Miami. Logan G.A., Hayes J.M., Hieshima G.B., and Summons R.E. (1995) Terminal Proterozoic reorganization of biogeochemical cycles. Nature 376, 53-56. Lohmann K. C. (1987) Geochemical patterns of meteoric diagenetic systems and their application to studies of paleokarst. In Paleokarst (ed. P. W. Choquette and N. P. James), pp. 58-80. SpringerVeflag. Lowenstam H. A. (1961) Mineralogy J80/160 ratios, and strontium and magnesium contents of recent and fossil brachiopods and their beating on the history of the oceans. J. Geol. 69, 241-260. Magaritz M. (1989) 13C minima follow extinction events: A clue to faunal radiation. Geology 17, 337-340. Marshall J. D. and Middleton P. D. (1990) Changes in marine isotopic composition and the late Ordovician glaciation. J. Geol. Soc. London 147, 1-4. Martin E. E. and Macdougall J. D. (1995) Strontium and neodymium isotopes at the Permian/Triassic boundary: A record of climate change. Chem. Geol. (Isor Geosci. Sect.) 125, 73-99. McConnaughey T. (1989) 13C and 180 isotopic disequilibrium in biological carbonates: I. patterns. Geochim. Cosmochim. Acta 53, 151-162. McDuff R. E. and Edmond J. M. (1982) On the fate of sulfate during hydrothermal circulation at mid-ocean ridges. Earth Planet. Sci. Letr 57, 117-132. Meyers W. J. and Lohmann K. C. (1985) Isotope geochemistry of regionally extensive calcite cement zones and marine components in Mississippian limestones, New Mexico. In Carbonate Cements (ed. N. Schneidermann and P. M. Harris); SEPMSpec. Publ. 36, 223-239. Meybeck M. (1979) Concentrations des eaux fluviales en elements majeaurs et apports en solution aux oceans. Rev. Geol. Dyn. Geogr. Phys. 21, 215-246. Mii H-s., Grossman E. L., and Yancey T. E. (1997) Stable carbon and oxygen isotope shifts in Permian seas of West Spitsbergen-Global change or diagenetic artifact? Geology 25, 227-230.

C isotope ratios of Phanerozoic marine cements Minoura N. and Chitoku T. (1979) Calcareous nannoplankton and problematic microorganisms found in late Paleozoic limestones. J. Fac. Sci. Hokkaido Univ. (Ser. IV), Vol. 19, p. 199-212. Moldovanyi E. P. and Lob.mann K. C. (1985) Isotopic and petrographic record of phreatic diagenesis: Lower Cretaceous Sligo and Cupido formations. J. Sediment. Petrol. 53, 972-985. Opdyke B. N. and Wilkinson B. H. (1990) Paleolatitude distribution of Phanerozoic marine ooids and cements. Palaeogeogr. Palaeoclimatol. Palaeoecol. 78, 135-148. Perch-Nielsen K. (1985) Mesozoic calcareous narmoplankton. In Plankton Stratigraphy, pp. 329-427. Cambridge Univ. Press. Pirini-Radrizzani C. (1971) Coccoliths from Permian deposits of eastern Turkey. In Proc. ll-Plankton Conference (Rome, 1970), (ed. A. Farinacci), pp. 993-1001. Roma, Edizioni Tecnoscienza. Popp B. N., Anderson T. F., and Sandberg P. A. (1986) Textural, elemental and isotopic variation among constituents in Middle Devonian limestones, North America. J. Sediment. Petrol 56, 715-727. Popp B. N., Takigiku R., Hayes J. M., Louda J. W., and Baker E. (1989) The post-Paleozoic chronology and mechanism of ~3C depletion in primary marine organic matter. Amer. J. Sci. 289, 436-454. Quinn T. M., Lohmann K. C., and HaUiday A. N. (1991) Strontium isotopic variation in shallow water carbonate sequences: Stratigraphic, chronostratigraphic, and eustatic implications for the record at Enewetak Atoll. Paleoceanography 6, 371-385. Raft R. A. and Raft O. C. (1970) Respiratory mechanisms and the metazoan fossil record. Nature 228, 1003-1004. Renne P.R., Zichao A., Richards M.A., Black M.T., and Basu A. R. (1995) Synchrony and causal relations between PermianTriassic boundary crises and Siberian flood volcanism. Science 269, 1413-1416. Restelli G. and Angeletti G. (1993) Dimethylsulphide: Oceans, Atmosphere, and Climate. Kluwer Academic Publ. Richter F.M. and Turekian K.K. (1993) Simple models for the geochemical response of the ocean to climatic and tectonic forcing. Earth Planet. Sci. Lett. 119, 121-131. Richter F. M., Rowley D. B., and DePaolo D. J. (1992) Strontium isotope evolution of seawater: The role of tectonics. Earth Planet. Sci. Lett. 109, 11-23. Romanek C. S., Grossman E. L, and Morse J.W. (1992) Carbon isotopic fractionation in synthetic aragonite and calcite: Effects of temperature and precipitation rate. Geochim. Cosmochim. Acta 56, 419-430. Ross R. J., Jaanusson V., and Friedman I. (1975) Lithology and origin of Middle Ordovician calcareous mudmound at Meiklejohn Peak, southern Nevada. U.S. Geological Survey Prof. Paper #871. Rnhinson M. and Clayton R.N. (1969) Carbon-13 fractionation between aragonite and calcite. Geochim. Cosmochim. Acta 33, 997-1007. Runnegar B. ( 1991 ) Precambrian oxygen levels estimated from the biochemistry and physiology of early eukaryotes. Palaeogeogr. Palaeoclimatol. Palaeoecol. 97, 97-111. Sandberg P. A. (1983) An oscillating trend in Phanerozoic nonskeletal carbonate mineralogy. Nature 305, 19-22. Sandberg P. A. (1985) Nonskeletal aragonite and CO2 in the Phanerozoic and Proterozoic. In The carbon Cycle and Atmospheric C02: Natural Variations Archean to Present (ed. E.T. Sunquist and W. S. Broecker); Amer. Geophys. Union Geophys. Monogr. 32, 585-594. Schidlowski M., Junge C. E., and Peitrek H. (1976) Sulfur isotope variations in marine sulfate evaporites and the Phanerozoic oxygen budget. J. Geophys. Res. 82, 2557-2565. Scholle P. A. and Arthur M. A. (1980) Carbon isotope fluctuatins in Cretaceous pelagic limestones: Potential stratigraphic and petroleum exploration tool. Bull. Amer. Assoc. Petrol. Geol. 64, 6787. Stepanek B. E. (1984) Diagenesis of paleokarst collapse breccias in the Ordovician El Paso and Montoya groups, west Texas. M.S. thesis, Univ. Michigan. Stern R.J. (1994) Arc assembly and continental collision in the Neoproterozoic east African orogen: Implications for the consoli-

4845

dation of Gondwanaland. Annu. Rev. Earh Planet. Sci. 22, 319351. Strauss H. (1993) The sulfur isotopic record of Precambrian sulfates: New data and a critical evaluation of the existing record. Precambrian Res. 63, 225-246. Storey M., Mahoney J. L, Saunders A.D., Duncan R.A., Kelley S. P., and Coffin M. F. (1995) Timing of hot spot-related volcanism and the breakup of Madagascar and India. Science 267, 852855. Swart P. K. (1983) Carbon and oxygen isotope fractionation in scleractinian corals: A review. Earth ScL Rev. 19, 51-80. Tan F. C. and Hudson J. D. (1974) Isotopic studies of the paleoecology and diagenesis of the Great Estuadne Series (Jurassic) of Scotland. Scottish J. Geol. 10, 91-128. Tappan H. and Loeblich A. R. Jr. (1973) Evolution of the oceanic plankton. Earth Sci. Rev. 9, 207-240. Tobin K. J. and Walker K. R. (1996) Ordovician low- to intermediate-Mg calcite marine cements from Sweden: Marine alteration and implications for oxygen isotopes in Ordovician seawater. Sedimentology 43, 719-735. Van Cappellan P. and Ingall E. D. (1994) Benthic phosphorus regeneration, net primary production, and ocean anoxia: A model of the coupled marine biogeochemical cycles of carbon and phosphorus. Paleooceanography 9, 677-692. Van Cappellan P. and Ingall E. D. (1996) Redox stabilization of the atmosphere and oceans by phosphorus-limited marine productivity. Science 271, 493-496, Veevers J. J. (1990) Tectonic-climatic super-cycles in the billionyear plate-tectonic con: Permian Pangean icehouse alternates with Cretaceous dispersed-continents greenhouse. Sediment. Geol. 68, 1-16. Veevers J. J. and Tewari R. C. (1995) Permian-Carboniferous and Permian-Triassic magmatism in the rift zone bordering the Tethyan margin of southern Pangea. Geology 23, 467-470. Veevers J. J., Cunaghan P. J., and Shaw S. E. (1994) Turning point in Pangean environmental history at the Permian/Triassic (P/Tr) boundary. Geol. Soc. Amer. Spec. Paper 288, 187-196. Veizer J. (1983a) Trace elements and isotopes in sedimentary carbonates. In Carbonates: Mineralogy and Chemistry (ed. R. Reeder); Rev. Mineral. 11, 265-300. MSA. Veizer J. (1983b) Geologic evolution of the Archean - Early Proterozoic Earth. In Earth's Earliest Biosphere, Its Origin and Evolution (ed. J. W. Schopf), pp. 240-259. Princeton Univ. Veizer J. (1985) Carbonates and ancient oceans: Isotopic and chemical record on time scales of 107-109 years. In The carbon Cycle and Atmospheric COz: Natural Variations Archean to Present (ed. E. T. Sunquist and W. S. Broecker); Amer. Geophys. Union Geophys. Monogr. 32, 595-601. Veizer J. and Hoefs J. (1976) The nature of 1sO/16C and 13C/~2C secular trends in sedimentary carbonate rocks. Geochim. Cosmochim. Acta 40, 1387-1395. Veizer J., Holser W. T., and Wilgus C. K. (1980) Correlation of 13C/~2C and 34S/32S secular variation. Geochim. Cosmochim. Acta 44, 579-587. Von Datum K. L., Edmond J. M., Grant B., Measures C. I., Walden B., and Weiss R. F. (1985a) Chemistry of submarine hydrothermal solutions at 21°N, East Pacific Rise. Geochim. Cosmochi~ Acta 49, 2197-2220. Von Datum K.L., Edmond J.M., Measures C.I., and Grant B. (1985b) Chemistry of submarine hydrothermal solutions at Guaymas Basin, Gulf of California. Geochim. Cosmochim. Acta 49, 2221-2237. Vun Datum K. L. et al. (1995) Evolution of the East Pacific Rise hydrothermal vent fluids following a volcanic eruption. Nature 375, 47-50. Walker J. C. G. (1986) Global geochemical cycles of carbon, sulfur, and oxygen. Marine Geology 70, 159-174. Walker J. C. G. and Lohrnann K. C. (1989) Why the oxygen isotopic composition of seawater changes through time. Geophys. Res. Lett. 16, 323-326. Walker J. C. G., Klein C, Schidlowski M., Schopf J. W., Stevenson D.J., and Walter M. R. (1983) Environmental evoution of the Arcbean-Early Proterozoic Earth. In Earth's Earliest Biosphere,

4846

S.J. Carpenter and K. C Lohmann

Its Origin and Evolution (ed. J.W. Schopf), pp. 260-290. Princeton Univ. Weber J. N. (1967) Possible changes in the isotopic composition of carbon and oxygen and atmospheric carbon reservoir over geologic time. Geochim. Cosmochim. Acta 31, 2343-2351. Weber J. M. and Woodhead P. M. J. (1970) Carbon and oxygen isotope fractionation in the skeletal carbonate of reef-building corals. Chem. Geol. 6, 93-117. Wefer G. (1985) Die Verteilung Sabiler Isotope in Kalkschalen Mariner Organismen. Geol. Jahrbuch. Wheat C. G., Feely R. A., and Mottl M. J. (1996) Phosphate removal by oceanic hydrothermal processes: An update of the phosphorus budget in the oceans. Geochim. Cosmochim. Acta 60, 3593-3608.

Wilkinson B. H., Owen R. M., and Carroll A. R. (1985) Submarine hydrothermal weathering, global eustasy, and carbonate polymorphism in Phanerozoic marine oolites. J. Sediment. Petrol. 55, 171-183. Wilson P. A. and Opdyke B. N. (1996) Equatorial sea-surface temperatures for the Maastrichtian revealed through remarkable preservation of metastable carbonate. Geology 24, 555-558. Yale L. B, Carpenter S. J., and Dalton J. (1996) Modeling the effects of Large Igneous Provinces on the 87Sr/S6Sr ratio of Seawater. Geol. Soc. Amer. Abstr. Prog., 428, Vol. 28. Zhong S. and Gurnis M. ( 1993 ) Dynamic feedback between a continentlike raft and thermal convection. J. Geophys. Res. 98, 12,21912,232.