14. pore-water chemistry of the sulu and celebes seas

Silver, E. A., Rangin, C., von Breymann, M. T., et al., 1991Proceedings of the Ocean Drilling Program, Scientific Results, Vol. 124

14. PORE-WATER CHEMISTRY OF THE SULU AND CELEBES SEAS: EXTENSIVE DIAGENETICREACTIONS AT SITES 767 AND 7681

Marta T. von Breymann,2 Peter K. Swart,3 Garrett W. Brass,3 and Ulrich Berner4

ABSTRACT

The combination of multiple sediment sources and varying rates of sediment accumulation in the Celebes andSulu seas have had significant impact on the processes of diagenesis, mineralization, and pore-fluid flow. Isotopicand mass-balance calculations help elucidate the various reactions taking place in these western Pacific basins,where ash alteration and basalt-seawater interactions are superimposed on the effects of sulfate oxidation of organiccarbon and biogenic methane and of dolomitization of biogenic carbonates. Based on the shape of the calcium andmagnesium depth profiles, two major reactive zones have been identified. The first is located near the zone of sulfatedepletion and is characterized by carbonate recrystallization, dolomitization and ash alteration reactions at bothOcean Drilling Program Sites 767 and 768. The second reactive zone corresponds to the bottom of the sedimentarysequence and is characterized by alteration reactions in the basement (Site 767) and in the pyroclastic depositsbeneath the sediment column (Site 768).

INTRODUCTION

Marginal basins experience major tectonic and paleocean-ographic changes during their evolution. The timing andextent of the volcanic, collisional, and paleoceanographicevents is recorded in their sedimentary sequences. During Leg124 five sites were drilled in two marginal basins in thewestern Pacific: the Celebes and the Sulu seas (Fig. 1). Inthese environments, the combination of multiple sedimentsources and varying rates of sediment accumulation have hada significant impact on the processes of diagenesis, mineral-ization, and pore-fluid flow. The objectives of this study wereto interpret the distributions of the dissolved constituents inthe pore waters, and their carbon, oxygen, and strontiumisotopes, in terms of diagenetic and alteration processes. Wehave focused our study on two sites: Site 767, located in thecentral Celebes Sea in a water depth of 4905 m, and Site 768,located in the southeastern part of the Sulu Sea in a waterdepth of 4385 m.

REGIONAL SETTINGThe Celebes Sea is a small ocean basin bounded in the

north by the Sulu Archipelago and in the south by the islandof Sulawesi (Fig. 1). During Leg 124, basaltic basement wasreached at 787 mbsf at Site 767. Drilling in this basinestablished that the volcanic basement is overlain by middleEocene red clays, indicating that the Celebes Sea originatedin an open ocean setting similar to that of the Philippine Sea.The Sulu Sea is a marginal basin located between northernBorneo and the central Philippine Archipelago. A distinctnortheast-trending bathymetric high, the Cagayan Ridge,subdivides the Sulu Sea into two sub-basins. Drilling at Site768 in the southeast basin recovered 1047 m of sediment and222 m of basement. Analysis of these sequences has shown

1 Silver, E. A., Rangin, C , von Breymann, M. T., et al., 1991. Proc. ODP,Sci. Results, 124: College Station, TX (Ocean Drilling Program).

2 Ocean Drilling Program, Texas A&M Research Park, 1000 DiscoveryDrive, College Station, TX 77845, U.S.A.

3 Rosenstiel School of Marine and Atmospheric Science, University ofMiami, 4600 Rickenbacker Causeway, Miami FL 33249, U.S.A.

4 Bundesanstalt für Geowissenschaften und Rohstoffe, Stillweg 2, D-3000,Hannover 51, Federal Republic of Germany.

that the Sulu Sea originated in a back-arc setting in thelate-early to early-middle Miocene, nearly concurrent withthe cessation of volcanism in the Cagayan Ridge. Earlyoutpourings of rhyolitic to dacitic pyroclastic flows markedthe early formation of the basin (Rangin, Silver, von Brey-mann, et al., 1990).

Although the basins have different tectonic origins andages, their proximity to volcanic arcs and continentalsources has led to similar patterns of sedimentation. Forexample, thick continentally derived turbidites in the upperMiocene at Site 768 coincide with those seen in the CelebesSea, suggesting a common source, very likely Borneo.These sequences correspond to a major compressive eventbetween the Philippine Mobile Belt, and the Sulu, Palawan,and Cagayan Ridges during the middle Miocene (Rangin andSilver, this volume).

Both sites (767 and 768) are located in deep water, resultingin most cases in low-carbonate sequences. The sediments atSite 767 record deposition within a deep basin below thecalcite compensation depth (CCD); carbonate turbidites arepresent as graded beds below 109 mbsf, and contain assem-blages of foraminifers that have been redeposited from shal-lower depths. Site 768 records a sedimentation regime belowthe CCD throughout its early depositional history, followed bya rapid increase in carbonate accumulation at the Pliocene/Pleistocene boundary (Linsley, this volume). The carbonatecontent of the sediments has been related to sediment source,mode of sedimentation, and changes in the CCD through time(Rangin, Silver, von Breymann, et al., 1990).

The organic carbon of the sediments in both basins ismostly of terrestrial origin, with the highest accumulationrates occurring in middle Miocene turbiditic sequences. Therelative proportion of marine organic material is higher insediments younger than upper to middle Miocene (Bertrand etal., this volume). The organic matter is characterized by lowlevels of maturity at Site 767, whereas Site 768 sedimentsreach the thermogenic hydrocarbon generation zone around450 mbsf (Rangin, Silver, von Breymann, et al., 1990; Ber-trand et al., this volume), a fact that is consistent with thehigher geothermal gradient of the Sulu Sea (Berner andBertrand, 1990).

Volcanic activity in the arcs that surround these basinsresults in extensive accumulation of air-fall ash and pyroclas-

203

M. T. VON BREYMANN, P. K. SWART, G. W. BRASS, U. BERNER

10°N —

10°N

10°S

125UE

110üE 120 F. 130°E

Figure 1. Location map for sites drilled in the Celebes and Sulu seas, from Rangin, Silver, vonBreymann et al., 1990.

tic flows in the Celebes and Sulu Seas (Pubellier et al., thisvolume). Thick, continentally derived turbidites occur at bothSites 767 and 768 in the upper middle Miocene, but are notpresent in the sections of Sites 769 and 770. To better evaluatethe effect of these sequences on the interstitial water chemis-try we have used, in some instances, comparisons with datafrom the sites located on bathymetric highs (Sites 769 and770), which are isolated from the deposition of turbidites ofterrigenous detritus.

METHODSInterstitial fluids were collected by pressure filtration.

Dissolved sulfate, alkalinity, ammonium, silica, calcium, andmagnesium were measured within 24 hr of sample collection.Methane contents were estimated by analysis of head spacesamples after incubation of the sediment at 60°C. The carbon-ate content of the sediments was measured coulometrically,and the total organic carbon (TOC) was obtained by Rock-Eval determinations. These analyses were done on board theJOIDES Resolution; the sample handling and the analyticaltechniques utilized are described in Rangin, Silver, von Brey-mann, et al. (1990).

Dissolved strontium and potassium were measured byatomic absorption spectrometry after appropriate sample di-

lution and matrix modification. The δ13C measurements ontotal dissolved CO2 were carried out in the BGR laboratories,Hannover, FRG, using a Finnigan MAT 250 mass spectrom-eter and are reported relative to the PDB (Pee Dee Belemnite)standard. 18O/16O ratios in the pore fluids were determinedusing a Finnigan-MAT 251 at the University of Miami, and arereported in per mill (%o>) deviation from SMOW. For thisdetermination, 1 cm3 of sample was equilibrated with CO2 at25°C for a period of 12 to 24 hr, after the method of Epsteinand Mayeda (1954). Results were normalized to standardsprepared and equilibrated at the same time.

RESULTSThe downhole distributions of alkalinity, sulfate, ammo-

nium, silica, calcium, magnesium, potassium, and strontiumat Sites 767 and 768 are shown in Figure 2, based on data listedin Table 1. Changes in gradient of the calcium and magnesiumprofiles suggest two main reaction zones: (1) the ash-carbon-ate-bearing sediments in the upper 200-300 m at both sites,and (2) the basalts at Site 767 and volcanogenic depositsoverlying basement at Site 768.

The concentration-depth profiles of calcium and magne-sium are nonlinearly correlated throughout the sediment col-umn. The downhole magnesium distributions suggest that

204

Site 767

PORE-WATER CHEMISTRY OF SULU AND CELEBES SEAS

Alk SO4 NH 4 Mg Ca SiO 2 Sr K

(mM) (mM) (µM) (mM) (mM) (µM) (µM) (mM).0 I0 0 30 0 1200 0 60 0 80 0 1000 0 400 0 15

- \

i

\

I0 0 30 0

Mg

(mM) (mM)2000 0

C 200(A

n

E 400 -

D.Φα

600 -

800

1 1 ' 1 V

>

\

\

I

Figure 2. Downcore distributions of dissolved alkalinity, sulfate, ammonium, magnesium, calcium, silica, strontium, and potassium at Sites 767and 768.

consumption of this element is largely taking place in theupper portion of the sedimentary sequences at both sites. Onthe other hand, the dissolved-calcium profiles indicate thatreactions in the underlying basalts are an important source forthis element, and that reactions in the sediments contribute toa lesser extent to the calcium increase relative to the magne-sium consumption. Within the sediment column, the eflFects ofalteration of volcanic material are superimposed on the dia-genetic alteration of biogenic carbonates.

It is difficult to quantitatively assign changes in the con-centration of dissolved constituents of the pore fluids todiagenetic reactions in the sediments and/or alteration reac-tions in the basement. These exercises have been somewhatsuccessful when applied to homogeneous sediments ofgeochemically simple areas. For example, carbonate recrys-tallization models have been used to explain changes in thecalcium and strontium concentrations in several deep-seacarbonate sediment sections (Baker et al., 1982). McDuff andGieskes (1976) and Gieskes and Lawrence (1981) have mod-eled changes in the dissolved calcium, magnesium, and oxy-gen isotope profiles of pore fluids in deep-sea sediments inareas where there is little effect from organic matter oxidationon the alkalinity, calcium, and magnesium distributions.

In the western Pacific basins where ash alteration andbasalt-seawater interactions are superimposed on the effectsof the sulfate reduction and dolomitization of biogenic carbon-ates, a fully quantitative evaluation of all the geochemicalprocesses is hardly possible. Nevertheless, in the followingsections we have used isotopic and mass-balance calculationsto help elucidate the various reactions taking place at Sites 767

and 768. These calculations represent a semiquantitativeapproach constrained by the complex and variable nature ofthe sediments in this environment.

Dissolved Strontium and Carbonate RecrystallizationEstimates

Both sites (767 and 768) show small increases in thedissolved strontium concentration downcore (Fig. 2). Suchincreases have been observed at other DSDP and ODP sitesand have been attributed to recrystallization of biogeniccarbonates to inorganic calcite, with the consequent release ofstrontium to the pore fluids (Baker et al., 1982). As thedissolved strontium and sulfate concentrations are belowcelestite saturation, the diffusive flux of strontium from thesediment and the degree of carbonate recrystallization at Sites767 and 768 may be estimated (Baker, 1986; Stout, 1985;Swart and Guzikowski, 1988). In the method outlined byBaker (1986), the strontium flux out of or into the sediments issummed over the time interval represented by two successivepore-water samples. The flux is given by

Flux = -Db dcldx (1)

where Db is the diffusion coefficient, here assumed to be 4 ×I0"6 cm2/s, de is the concentration gradient, and dx representsthe depth interval in question.

If we assume that the average Sr/Ca molar ratio of theoriginal biogenic sediment is 2 × I0"3, a typical value forpelagic calcite (Baker, 1982), then the total amount of stron-

205


Table 1. Concentration of dissolved constituents in the pore fluids of (A) Site 767 and (B) Site 768.

Sample Depth(mbsf)

Alk

(mM)so4

(mM)

NH4

(µM)Ca Mg

(mM) (mM)SiO2

(µM)Sr

(µM)

K(mM)

767B-1H-4,2H-4,3H-4,4H-4,5H-4-6H-4,7H-4,8H-3,9H-4,

10H-4,13X-1,17X-2,20X-1,23X-2,26X-4,29X-4,32X-4,35X-4,38X-2,41X-4,44X-3,47X-2,50X-4,53X-4,56X-3,59X-3,62X-4,65X-1,68X-4,71X-4,74X-4,75X-4,

145--150145--150145--150145-150

145--150145-150145--150145-150145-150145-150145-150145-150145-150145-150145-150145-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150

5.9514.9524.4533.9543.4552.9562.4070.4077.4586.95

111.05151.25178.75206.25238.55267.35296.30325.30351.40383.40410.90438.30470.00498.90525.00554.00585.00609.00643.00671.00700.00710.00

5.486.357.437.627.998.278.008.598.738.688.707.126.343.721.941.000.701.421.541.772.422.722.403.361.581.58

1.96

1.45

7.777.647.727.837.717.797.767.847.897.687.597.427.657.637.978.318.508.418.228.288.348.078.047.858.028.07

7.98

6.55

28.927.225.725.024.423.723.923.523.122.522.819.216.814.810.2

7.54.33.02.83.85.03.32.14.37.52.74.57.76.84.66.85.7

136221390392413590653766773820870

1086971838651692699797715713678523543429471473498422

402427302

10.1910.2410.4010.5910.5510.5810.5210.9511.0711.2511.6511.7211.9318.1625.0929.1130.3631.9432.2431.7530.9830.9730.9729.4830.7834.1738.0343.71

51.1657.2161.06

49.9549.0448.3348.4648.1047.3746.7246.6946.2445.1343.8642.2940.5232.6425.5619.7312.12

9.599.84

10.9614.3213.3113.0015.4413.1911.6012.22

10.948.274.586.27

536597654685689637674644642670742809910721390187178167159152109

74748293918491

9699

128241324328260

186196221225240240240230225221235

11.4

12.2

12.5

10.0

10.0

7.9

5.1

2.9

2.2

2.2

2.22.1

2.0

tium for each interval in the sediment can be calculated usingthe calcium carbonate data shown in Figure 3. We haveassumed an average calcium carbonate content of 2% for theupper 120 mbsf and 25% for the lower portion of the holes atSite 767. For the upper sections of Site 768 we have assumedan average value of 25%, and a 2% carbonate content for thesediments below 120 mbsf. By comparing the fractionalchange in dissolved strontium with that available in thesediment, we have estimated the amount of recrystallizationthat has occurred during the period represented by the sam-pling interval. By further assuming that there has been nosignificant change in the distribution of dissolved strontiumwith time, the percent carbonate recrystallized over a givensection may be estimated by summing all the intervals overwhich there are pore-water data available. The results of thisexercise are shown in Figure 4. Using different concentrationsof carbonate will result in significantly different estimates ofrecrystallization. For example, if the carbonate content in theupper 180 mbsf at Site 767 is reduced from 2% to 0.5%, theestimated rate of recrystallization increases so that the car-bonate component of the sediment is expected to be com-pletely recrystallized between 150 and 180 mbsf, which cor-responds to the maximum in the dissolved strontium profile(Fig. 2). Similarly, if at Site 768 the carbonate content used inthe previous calculations is reduced from 2% to 1%, thecarbonate component of the sediment will be 100% recrystal-lized by 500 mbsf (Fig. 4).

It is important to note that this method does not take intoaccount the amount of strontium incorporated into any diage-netic calcite, nor any contributions to the pore fluids by

diagenetic alteration of igneous minerals. The latter sourcemay be considerable, particularly in the deepest sedimentarysequences as evidenced by the strontium isotopic compositionof the pore fluids discussed in the following section. Theestimates given in Figure 4 are extremely dependent upon theassumptions listed above; nevertheless this calculation allowsus to place constraints on the system. Limits on the degree ofcarbonate recrystallization and the sedimentary sectionswhere this reaction is more likely to be occurring are useful forevaluating the oxygen and carbon isotope distributions.

Oxygen and Strontium IsotopesThe 87Sr/86Sr ratio in the pore fluids is shown relative to the

Burke et al. (1982) seawater curve in Figure 5. Normally inpelagic sediments dominated by carbonates, the 87Sr/86Srdistribution lies close to this curve. Sites 767 and 768 mighthave experienced a significant contribution of redepositedMiocene carbonates that can have a 87Sr/86Sr ratio as low as0.7086. Thus, although in the upper 200 mbsf of Site 767 thestrontium isotope values could represent recrystallization ofMiocene carbonates, the significant depletion observed in thedeepest analyzed sample at this site, and in most of Site 768samples, indicates that there has been isotopic exchangebetween the pore fluids and the volcanic sequences. Figure 6shows the depth distribution of the 87Sr/86Sr values superim-posed to the dissolved silica profiles at both Sites 767 and 768.The minima in these profiles coincide with a zone of ashalteration characterized by the replacement of volcanic glassby an authigenic smectite-phillipsite assemblage (Desprairieset al., this volume). This diagenetic zone is synchronous in

206


Table 1 (continued).

Sample Depth

(mbsf)

Alk

(mM)pH SO4

(mM)

NH4

(µM)

Ca

(mM)

Mg

(mM)SiO2

(µM)

Sr

(µM)

K

;mM)

768B-1H-2,2H-5,3H-5,4H-5,5H-4,6H-5,7H-5,8H-5,9H-5,

10H-5,13H-5,16H-5,19H-5,22H-4,26X-2,28X-4,31X-5,34X-2,37X-2,

768C-1R-3,4R-1,7R-3,

10R-3,13R-3,16R-1,19R-1,22R-2,26R-4,28R-4,31R-3,34R-5,34R-4,40R-3,43R-2,

145--150145-150145-150145-150145-150145-150145-150145-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150140-150

2.9511.4520.9530.4538.4549.4558.9568.4577.9587.45

115.95144.45172.90198.70222.20244.50274.90299.30328.40357.60383.50415.50444.30473.30499.30528.30558.70600.50619.80647.20679.20706.70734.33761.40

3.014.993.913.834.404.364.384.353.293.342.813.564.424.443.945.426.694.845.725.055.915.445.664.703.873.142.944.033.12

1.670.59

7.567.567.507.487.457.41

7.517.41

7.497.497.507.847.708.357.878.127.877.807.857.827.827.898.018.047.987.757.777.747.93

7.317.41

27.127.326.224.922.823.922.721.120.019.113.0

7.20.8

0.2

0.0

1.4

0.0

0.0

5.4

6.5

4.1

1.1

1.2

2.4

5.33.14.10.7

0.3

6.46.2

1.8

6.9

5.2

2 6

6 8

1161462842852802772774 1 9

4 5 3

4 8 8

561748814738751847506

11348259779 5 1

10931378

9111155

7148608464 6 8

5684054 1 2

10.86

11.0211.06

10.9511.2411.3911.4311.4311.6211.9412.0012.56

9.14

10.2411.5711.9511.1011.9712.5413.5614.9416.5217.4018.8620.7922.23

31.8035.6042.19

54.0463.28

51.4850.0748.8546.7646.2344.6742.9041.1739.4137.6833.8630.7826.75

23.0524.8625.1925.0927.4526.7223.3521.8020.1518.0919.0617.3714.83

10.609.079.00

9.093.30

411542625

6556746897 5 1

8 3 6

765793230206224

1581791 7 1

1 9 8

166

172211201156134146187148158121236207

1271 3 0

9 1

9 8

127148167167163167160158153

1531461391461 5 1

134127127136155158196136170153165167

10.1

9.3

8.9

8.6

8.8

6.7

4.6

2.5

2.0

2.5

1.8

2.7

1.9

1.82.1

both basins, and corresponds to tephra layers dated at 3.5 to4 Ma (Desprairies et al., this volume).

The oxygen isotopic composition of the pore fluids de-creases from approximately 0 %c SMOW to values between -2%o and -3 %o (Fig. 6). We have estimated the effect ofcarbonate diagenesis upon the δ 1 8 θ of the pore fluids at Site767, using the estimates of carbonate recrystallization pre-sented earlier, a geothermal gradient of 77°C/km and carbon-ate contents between 10% and 60%. Porosity is assumed todecrease from 80% to 40% by a depth of 800 mbsf (Rangin,Silver, von Breymann, et al., 1990). The results of thiscalculation indicate that carbonate recrystallization has anegligible influence on the oxygen isotopic composition of thepore waters. The higher geothermal gradient of Site 768(127°C/km, Vollbrecht and Kudrass, 1990) will result in anincrease the δ18 values in the lower portions of the section.Formation of opal-CT from biogenic silica would similarlyincrease the pore-water δ 1 8 θ, therefore diagenetic reactionsinvolving biogenic carbonates and silica are not likely to beresponsible for the observed trend in the oxygen isotopedistribution.

The most favorable diagenetic reactions for lowering theδ 1 8θ of the pore waters involve converting volcanic ash to clayminerals and the alteration of Layer 2 basalts (Lawrence etal., 1979). Such reactions must be responsible lowering of theδ 1 8θ of the fluids at Sites 767 and 768. The δ 1 8 θ of the porefluids has been used in conjunction with the isotopic compo-sition of authigenic minerals to determine that alteration of

CaCO3 (%) CaCO3 (%)

Q.03Q

00

2 0 0

4 0 0

arm

—

—

T Z :

«==

_--—

——̂̂

•

—

—

•

_ — —

i

50

—

i

f

_ — — ~

Sitei

75 11

—

- :

767i

00

2 0 0 •

4 0 0

6 0 0

8 0 0

1000

25 100

Figure 3. Carbonate content of sediments from Sites 767 and 768. Datafrom Rangin, Silver, von Breymann, et al., 1990./

207


Percent of Sediment Recrystallized20 40 60 80 100

768 (2)

200

(0

nE

αΦQ

400

600

800

767(1)

768(1)

Figure 4. Degree of recrystallization calculated for the carbonatecomponents at Sites 767 and 768 using parameters discussed in thetext. Data for Site 767 example (1) denote estimates using a carbonatecontent of 2% over the upper 120 mbsf, while Site 767 (2) representsa content of 0.5% over the same interval. For Site 768, example (1)was calculated using a content of 2% below 120 mbsf, whereas forexample (2) we used a value of 1% over the same interval.

ashes at these sites occurs under low-temperature conditions(Despraires et al., this volume).

Carbon Isotopes

The δ13C of the dissolved 2CO2 in the pore fluids arepresented in Table 2 and Figure 7. In this figure, downcoredistributions of dissolved sulfate, strontium, and methane areincluded for comparison. Of particular significance are thelight values of δ13C in the deeper samples analyzed, whichhave values for the SCO2, of -19 ‰ at Site 767 and of -42 ‰ atSite 768. The lightest value that could reasonably be explainedby oxidation of organic carbon is -22 %c, which corresponds tothe value measured for the sedimentary organic carbon in thisarea (Lallier-Verges et al., this volume). The δ13C distributionobserved in the pore fluids of various recent and ancientanoxic marine environments has been explained by a mixingprocess between buried oceanic dissolved carbonate (δ13C =+0.5 %o), metabolic CO2 generated from mineralization oforganic carbon (δ13C = -22 %c) and dissolution/recrystalliza-tion of fossil carbonates (δ13C = +2 ‰) (McCorkle, 1987;Emerson and Bender, 1981). Additionally, anaerobic methaneoxidation is now well established as an early diageneticprocess in reducing sediments (Devol, 1983; Whiticar andFaber, 1986), and has been shown to result in extremely lowδ13C values in the pore fluids of hemipelagic environments and

in the pore fluids expelled from the Oregon accretionarymargin (Suess and Whiticar, 1989).

The low δ13C values of the dissolved 2CO2 measured in theCelebes and Sulu Seas indicate that methane oxidation is notrestricted to depositional basins underlying areas of highprimary productivity or accretionary prism settings, but it alsooccurs in these deep basins where terrestrial turbiditic depos-its lead to the accumulation of high levels of biogenic methane(δ13C = -70 ‰, Berner and Bertrand, this volume). Subse-quent oxidation of this light carbon source will result 13C-depleted CO2 as measured here. The generally low organiccontent of these sediments precludes sulfate depletion, allow-ing this electron acceptor to be available to oxidize theupwardly diffusing CH4. This is further illustrated by compar-ing the total organic carbon, methane, and sulfate distribu-tions of the deep basin (Sites 767 and 768), with those of sitesdrilled in bathymetric highs (Sites 770 and 769) where noturbiditic deposits occur and no significant accumulation ofmethane was observed (Fig. 8).

The changes in the δ13C of the pore fluids (Fig. 7) can beexplained by a combination of mechanisms, namely: carbonaterecrystallization, organic matter degradation, and sulfate oxida-tion of methane. For a detailed look at these processes, we haveseparated the upper 340 mbsf of Site 767 into three intervals, asshown in Figure 7A. Interval la is characterized by sediments ofvery low carbonate content; the δ13C of the pore fluids in thisinterval corresponds to oxidation of particulate organic matter(δ13C -20 %o to -22 %o). Recrystallization of the carbonateturbidites below 140 mbsf was demonstrated by using the dis-solved strontium distribution as discussed earlier (Fig. 4). Thisprocess results in input of dissolved CO2 with δ13C of 0% ± 2%to the pore fluids (Interval Ib). The decrease in dissolved sulfatein Intervals Ib and II is largely driven by methane oxidation atabout 340 mbsf, which results in a sharp decrease in δ13C valuesof the pore fluids in Interval II.

At Site 768 (Fig. 7B) the δ13C of the dissolved 2CO2 inInterval la has values very similar to the buried seawater (δ13C= -2 ± 1 %o), and probably represents recrystallization reac-tions of these carbonate-rich sediments (Fig. 4). There isapparently very little of an isotope signature resulting from theoxidation of particulate organic carbon, which suggests thatthe sulfate profile is driven mostly by diffusion to the zone ofmethane oxidation. This is supported by the relationshipbetween sulfate and ammonium (Fig. 9). As methane oxida-tion generates no ammonium at all, there is a lot more sulfateconsumed relative to the amount of ammonium generated atthis site relative to Interval la of Site 767, where sulfate isconsumed by the oxidation of sedimentary organic carbon.The extremely negative δ13C values in Interval la at Site 768clearly indicate that the dissolved CO2 is being generated bysulfate oxidation of methane (δ13C = -74 ± 4 ‰, Berner andBertrand, this volume).

DISCUSSION

Two major "reactive-zones" have been identified at Sites767 and 768 based on the shape of the calcium and magnesiumdepth profiles that correspond to the base of Intervals I and III(Fig. 2). A sedimentary reactive-zone is located near the zoneof sulfate depletion and is characterized by carbonate recrys-tallization/dolomitization reactions and by ash alteration inInterval I in both sites. The second reactive-zone correspondsto alteration reactions in the basement (Site 767) and pyro-clastic flows beneath the sediment column (Site 768) anddetermines the geochemistry of the pore fluids in Interval III.Both reactive zones may involve precipitation of carbonatesand the formation of authigenic clays.

208

87S r / 8 6 S r00IDO

O

OJN-

oNO

O

ò

oCOoo

00

oó

00CO

oo

CO

oö


Sr (/xM)

100 200 300 400

Figure 5. (A) Strontium isotopic composition of the pore fluids vs. age at Sites 767 and768 relative to the Burke et al. (1982) seawater curve. In normal pelagic carbonate-dominated sediments the 87Sr/86Sr distribution lies close to this curve. (B) Dissolvedstrontium content of the pore fluids.

Interval I

The carbon isotopes of the carbonates and pore waters inSites 767 and 768 are indicative of reactions involving carbon-ate recrystallization partly driven by sulfate oxidation ofmethane. Carbonate recrystallization is also evident by thedissolved strontium distribution. Dolomitization reactions areinferred based on petrographic and XRD identification ofdolomites concurrent with the onset of methane accumula-tions and depletion of sulfate (von Breymann and Berner,1990). These reactions involve changes in the calcium andmagnesium distributions. This interval is coincident withaccumulations of ash at both Sites 767 and 768. Alteration ofthis material was evident by petrographic observations, and isprobably responsible for the depletion of dissolved silica(Figs. 2 and 6) and the poor preservation of diatoms in thesebasins (Rangin, Silver, von Breymann, et al., 1990).

At Site 768, the δ13C of the dissolved CO2 and the sulfate-ammonium relationship (Fig. 9) suggest a very small contri-bution to the system by oxidation of sedimentary organicmatter, so that the carbon isotope distribution is controlledmostly by carbonate recrystallization and by sulfate oxidationof methane, as given by

CH4 + SO4

2- - 2H2O + S2- (2)

Because there is no change in the calcium concentration inthe pore fluids in Interval, we postulate that dolomitization isoccurring by the following reaction:

CaCO3 + Mg2 +2HCO3- = CaMg(CO3)2

H2O (3)

Based on these relationships we are able to model theseprocesses very simply in terms of changes in dissolved sulfate(ΔSO4

2), so that

ΔSO4

2" = l/2ΔMg2+ - ΔXCO2 (4)

where Δ2CO2 and ΔMg2+ are the changes in total dissolvedCO2 (estimated from alkalinity and pH values) and magnesiumrelative to seawater values. Figure 10 shows the values forΔSO4

2• estimated using equation (4) vs. the ΔSO4

2" valuesmeasured in the pore fluids. The good correspondence shownin this figure suggests that the main mechanisms controllingthe dissolved alkalinity, sulfate, and magnesium distributionsmay be represented by equations (2) and (3).

At Site 767, however, the situation is more complex. Asdiscussed in relation to the δ13C distributions, the upper 340mbsf at this site can be subdivided into three intervals (Fig.7A). In Interval la, the main process is oxidation of sedimen-tary organic carbon as evidenced by the δ13C values of-10 ±2 %o. In this interval, the changes in dissolved calcium andmagnesium are small and the changes in alkalinity can beexplained in terms of microbial degradation of organic carbonby sulfate reduction:

2[CH2O(NH4)X] + SO4"2 = 2SCO2 + 2H2O (5)

where x represents the C/N regeneration ratio of thesedimentary organic carbon, and SCO2 is produced in a 1:2ratio relative to the amount of sulfate being consumed. Inter-vals Ib and II are characterized by recrystallization of carbon-ates, carbonate generation by methane oxidation, and calciumincreases relative to magnesium consumption in a 2:1 ratio(Fig. 2). These observations can be explained if dolomitizationwas occurring by the following reaction:

3CaCO3 + 2Mg2+ + 2HCO3" = 2CaMg(CO3)2 + Ca2+ +

CO2 + H2O. (6)

Using equations (2) and (6), the change in sulfate can beapproximated by

209


A

0.706 0.707 0.708 0.709 0.710U

6000 200 400 600 800 1000

SiO2 (µM)

100 -

200 -

300 -

400 -

500 -

600

K+ (mM)

B

Q.

Q

0.706 0.707 0.708 0.709 0.710

u

100

200

300

400

500

RΠΠ

r /)/

}

; y-

/ /

i •4

0 200 400 600 800 1000

SiO2 (µM)

100 -

200 -

300

400 -

500 -

600

K+(mM)

Figure 6. Downcore distributions of strontium and oxygen isotopes (solid lines) for(A) Site 767 and (B) Site 768. Dissolved silica and potassium distributions (dottedlines) are included for comparison.

210


Table 2. Isotope characterization of the pore fluids from (A) Site 767 and (B) Site 768.

Sample Depth δisc δiβo 87Sr/86Sr(mbsf) (o/oo) (o/oo)

767B-1H-4, 145--1502H-4, 145--1503H-4, 145--1504H-4, 145-1505H-4, 145-1507H-4, 145-1508H-3, 145-1509H-4, 145-150

10H-4, 145-15013X-1, 145-15017X-2, 145-15020X-1, 145-15023X-2, 145-15026X-4, 145-15029X-4, 145-15035X-4, 140-15038X-2, 140-15056X-3, 140-150

767C-1R, 140-150

Table 2 (continued).

Sample

768B-1H-2, 145-1502H-5, 145-1503H-5, 145-1504H-5, 145-1505H-4, 145-1506H-5, 145-1507H-5, 145-1508H-5, 145-1509H-5, 140-150

10H-5, 140-15013H-5, 140-15016H-5, 140-15019H-5, 140-15022H-4, 140-15026X-2, 140-150

768C-1R-3, 140-1504R-1, 140-150

16R-1, 140-15022R-2, 140-150

5.9514.9524.4533.9543.4562.4070.4077.4586.95

111.05151.25178.75206.25238.55267.35325.30351.40525.00

Depth

(mbsf)

2.9511.4520.9530.4538.4549.4558.9568.4577.9587.45

115.95144.45172.90198.70222.20357.60383.50499.30558.70

-8.32-10.54-1 1 .25-1 1 .74-1 1.90-1 1.76-12.10-11.79-1 1.45

-6.64-2.70-1.16-2.92-4.03-7.66

-18.98

δ13C

(°/oo)

0.0901.084

0.0480.0430.3240.235

-0.7560.108

-0.652

-2.136-1.319-1.189-1 .278-1.799

0.3520.3920.9850.8440.7250.7610.215

0.428-0.919-0.386-0.230-0.821-1.638

-2.232-2.455-2.583

6180

(o/oo)

-1.41-1 .84-1.90-2.20-2.44-2.58-2.98-2.92-2.85-3.05-7.79

-23.12-42.00

0.7091410.7090770.7090790.7090890.7089540.7088070.7087650.7087250.7086950.7087170.7087310.7086710.7084040.707856

0.7063090.7061280.7062900.706199

87Sr/86Sr

0.7092000.7091000.7090290.7089130.7089380.7088140.7087430.7085470.708437

0.707690

0.7072450.7073560.7072300.7071690.7070360.7068220.706618

ΔSO4

2" = 1/2 ΔCa2+ - ΔMg2+ - Δ2CO2, (7)

which, as shown in Figure 10, corresponds well with theactual observed change in the pore-water sulfate.

The above discussion indicates that the main mechanismdriving the calcium and magnesium profiles in Interval I ofSites 767 and 768 is related to carbonate geochemistry, mainlytriggered by oxidation of methane and precipitation of dolo-mite. However, this zone is coincident with high accumula-tions of ash at both sites. Alteration of this material wasevident in petrographic observations (Rangin, Silver, vonBreymann, et al., 1990) and constitutes a source of calciumand a sink for silica and magnesium.

Despraires et al. (this volume) conducted geochemical andmineralogical analysis of the volcanic ashes at both sites todetermine the nature of the secondary volcanic products inthese sequences. They showed that the largest diageneticeffects occur in tephras dated 3.5 to 4 Ma, which correspondto a zone between 233 and 329 mbsf at Site 767 and 129 to 165mbsf at Site 768. The alteration is characterized by a completereplacement of glass by a phillipsite-smectite assemblage. Thezeolites show a very low Ca content relative to Na and K.Authigenic smectites were shown to consist of a mixture ofspecies ranging from aluminous beidelite to nontronite. Thesesmectites display a positive correlation between potassiumand iron and are characterized by low magnesium content.

211


δ 1 3

-20

C I

-I0

S04 (mM) Sr {µM) CH 4 (ppm)

0 0 I0 20 30 0 200 400 0 5000 10000

CO

nE

αΦ

Q

u

100

200

3 0 0

4 0 0

500

p,nn

la I

Ib N.

/

r " s

-

-

BS 1 3 C (%o) S0 4 (mM) Sr tyxM) CH4 (ppm)

0 0 I0 20 30 0 200 400 0 5000 I0000

-GE

200-

Q.ΦO

300-

400

Figure 7. Downcore distribution of δ13C of the dissolved CO2 in the pore fluids of (A) Site 767 and (B) Site 768. Dissolvedsulfate and strontium profiles, as well as the methane content of the sediments, are included for comparison.

212


SO4 (mM)

10 20

CH4 (ppm)

2500

1 00 4-

2001-

500

5000

3001-

400f

500

100

200

300

400

500

SO4 (mM)

100h<nnE

Q.α>Q

200

300 J

CH4 (ppm)

0 5000

1 0 0 ^

10000

200f

300

1 00 -

200-

300

Figure 8. A. Sulfate, methane, and total organic carbon for sites in the Celebes Sea. Site 770 (solid line) is on a bathymetrichigh relative to Site 767 (dotted line). B. Sulfate, methane, and total organic carbon for sites in the Sulu Sea. Site 769 (solidline) is on a bathymetric high (Cagayan Ridge) relative to Site 768 (dotted line).

213


2000

10 20Sulfate (mM)

3 0

Figure 9. Dissolved sulfate vs. ammonium in the pore fluids. The linerepresents ammonium generated from microbiai sulfate reduction ofnormal sedimentary organic matter (equation 5); it is based on aregeneration ratio corresponding to a C/N = 1 5 . Samples from Site767, Interval la (open circles) lie within this line, indicating oxidationof organic matter. The samples from Sites 767, Intervals Ib and II(closed circles) and Site 768, Interval la (open squares) denotesamples with strong negative ammonium anomalies compared withthe observed sulfate reduction. This deficit reflects the lack ofammonium released from the anaerobic oxidation of methane (equa-tion 2). The small diamonds represent data from the Oregon marginwhere CH4-derived Σ C O 2 in the pore fluids has been thoroughlydocumented (Suess and Whitticar, 1989).

It is hard to estimate quantitatively the extent of thesereactions on the magnesium distribution because the silicauptake from the pore fluids is very small, and there is anunknown but small amount of silica available via dissolutionof biogenic components. A mass balance using average datafor the composition of the original glass (Pouclet et al., thisvolume), as well as for the authigenic smectite (Desprairies etal., this volume), allows us to estimate the change in silica tomagnesium molar ratio during alteration. This value corre-sponds to a %Si: ΔMg = 0.89: 0.26. Given that these sedi-ments are very poor in biosiliceous components (Rangin,Silver, von Breymann, et al., 1990), and that the secondaryclay minerals have a low magnesium content, ash alterationreactions are thought to have a small effect on the magnesiumdistribution. Dolomitization reactions are therefore likely toconstitute the main sink for this cation within the sediment, asevidenced by the results shown in Figure 10.

Interval III

Even though the formation of secondary phillipsite s con-stitutes a sink for potassium within the sediments, this ele-ment is also incorporated into celadonites, an alteration prod-uct observed in the pyroclastic sequences at Site 768 (Rangin,Silver, von Breymann, et al., 1990). There is a good correla-tion between the δ 1 8θ and the potassium distributions (Fig.11 A), which is in accordance with the fact that the alterationof igneous components results in oxygen isotope fractionationand potassium uptake.

A good relationship between a decrease in the δ 1 8θ valuesand a calcium increase has been observed in the pore fluids ofDSDP sites where the chemistry of the interstitial fluids isthought to be controlled by alteration of Layer 2 basalts(Gieskes and Lawrence, 1981). At Sites 767 and 768, IntervalsI and III show two distinct slopes for the Ca-δ18O relationship(Fig. HB). Interval 1, although it has a 2.5 %o change in theδ 1 8θ value, shows virtually no change in the dissolved calcium

-30 -20 -10

ΔSO4 measured

Figure 10. Estimated changes in the sulfate concentration (mM)relative to seawater, vs. measured values at Site 767, Interval Ib(circles) and Site 768, Interval la (squares).

-2

ß 60

03O

DSDP Site 336

Interval

-3 -2

(o/00)

Figure 11. A. Potassium vs. δ 1 8 θ for Sites 767 (circles) and 768(triangles). B. Calcium vs. δ 1 8θ for Sites 767 (circles) and 768(triangles). Data from DSDP Site 336 (squares) are included forcomparison. See text for discussion.

concentration, reflecting the fact that ash alteration reactionsdo not result in significant changes in the concentration of thiscation. The slope of the Ca-δ18O relationship for Interval III is

214


σ>

60

40 -

20 -

1 : 1

\ *

1 : 3

N. A

O 767

Δ 768

A 769

• 770

-

-

20 6 0 80

Ca (mM)

Figure 12. Relationship between calcium and magnesium for Sites 769and 770, as well as for Interval III of Sites 767 and 768.

compared in Figure 1 IB with that for DSDP Site 336 drilled onthe Iceland-Faeroe Ridge during Leg 38 (Lawrence andGieskes, 1981). The data indicate that in Interval III the mainmechanism controlling the calcium and oxygen isotope distri-butions is the alteration of volcanic sequences underlying thesediment column at Sites 767 and 768.

Carbonate-related diagenetic processes consume a signifi-cant amount of the dissolved magnesium from the pore fluids(Fig. 2) via dolomitization reactions. This has a potential effecton the nature of the alteration products in the basalt andpyroclastic sequences. At Site 769 on the Cagayan Ridge,where there are no CH4 accumulations and thus no likelycarbonate sink for the dissolved magnesium, this element ismainly consumed by the alteration of basalts (Rangin, Silver,von Breymann, et al., 1990). Correspondingly, at this site,Mg-rich smectites are the main minerals replacing the glass(R. Smith, this volume). A system depleted in dissolvedmagnesium may be conducive to the formation of celadonites,an alteration product observed at Site 768 but not at Site 769.Secondary smectites are also present at Site 768; these authi-genic clay minerals may have formed by alteration reactionsthat occurred before the rapid turbidite deposits of the middleMiocene or by isochemical transfer of magnesium by thereplacement of forsteritic olivine by hydrous ferric oxides(Bohlke et al., 1980).

These different alteration reactions are also suggested bythe Ca/Mg ratios in the pore waters. Sites 769 and 770 show a1:1 ratio between the calcium production and the magnesiumuptake, while the pore fluids of Interval III in the deep basinsites show a 3:1 correlation (Fig. 12). The lower magnesiumavailability at Sites 767 and 768 results in a different uptakeratio of this cation relative to calcium during alteration reac-tions relative to Sites 769 and 770.

ACKNOWLEDGMENTS

We thank M. Kurtz for use of the facilities for strontiumisotope analysis at Woods Hole Oceanographic Institution.The authors thank J. Leder and B. Haldane of the stableisotope laboratory at the University of Miami for assistance inpreparing samples and for helpful discussions. In addition, weacknowledge the scientific members and crew of ODP Leg 124for their assistance, in particular we thank the ODP chemistrytechnicians J. Powers and M. A. Cusimano for their skillfuland conscientious participation in the shipboard geochemicalwork. This research was supported by grants from the JOI/USSAC support program to M. T. von Breymann and G.Brass.

REFERENCES

Baker, P. A., 1986. Pore-water chemistry of carbonate-rich sedi-ments, Lord Howe Rise, Southwest Pacific Ocean. In Kennett, J.P., von der Borch, C. C , et al., Init. Repts. DSDP, 90: Washing-ton (U.S. Govt. Printing Office), 1249-1256.

Baker, P. A., Gieskes, J. M., and Elderfield, H., 1982. Diagenesis ofcarbonates in deep-sea sediments: evidence from Sr^/Ca2"1" ratiosand interstitial dissolved Sr2"1" data. J. Sediment. Petrol., 52:71-82.

Berner, U., Bertrand, P., and Scientific Party of Leg 124, in press.Evaluation of the paleogeothermal gradient at Site 768 (Sulu Sea).Geophys. Res. Lett. Böhkle, J. K., Honnorez, J., Honnorez-Guerstein, B. M., 1980. Alteration of basalts from site 396B,DSDP: petrographic and mineralogical studies. Contrib. Mineral.Petrol, 73:341-364.

Burke, W. H., Denison, R. E., Hetherington, E. A., Koepnick, R. B.,Nelson, H. F., and Otto, J. B., 1982. Variation of seawater87Sr/86Sr throughout Phanerozoic time. Geology, 10:516-519.

Devol, A. H., 1983. Methane oxidation rates in the anaerobic sedi-ments of Saanich Inlet. Limnol. Oceanogr., 28:738-742.

Emerson, S., and Bender, M., 1981. Carbon fluxes at the sedimentwater interface of the deep sea: calcium carbonate preservation. J.Mar. Res., 39:139-162.

Epstein, S., and Mayeda, T., 1953. Variation of 18O of waters fromnatural sources. Geochim. Cosmochim. Acta, 4:213-224.

Gieskes, J. M., and Lawrence, J. R., 1981. Alteration of volcanicmatter in deep-sea sediments: evidence from the chemical compo-sition of interstitial waters from deep-sea drilling cores. Geochim.Cosmochim. Acta, 45:1687-1703.

Lawrence, J. R., Drever, J. I., Anderson, T. F., and Brueckner, H.K., 1979. Importance of alteration of volcanic material in sedi-ments of Deep Sea Drilling Site 323: chemistry, 18O/16O and87Sr/86Sr. Geochim. Cosmochim. Acta, 43:573-588.

Lawrence, J. R., and Gieskes, J. M., 1981. Constraints on watertransport and alteration in the oceanic crust from the isotopiccomposition of pore water. J. Geophys. Res., 86:7924-7934.

McCorkle, D. C , 1987. Stable carbon isotopes in deep sea porewaters: modern geochemistry and paleoceanographic applica-tions. [Ph.D. dissert.]. Univ. Washington, Seattle.

McDuff, R. E., and Gieskes, J. M., 1976. Calcium and magnesiumprofiles in DSDP interstitial waters: diffusion or reaction? EarthPlanet. Sci. Lett., 33:1-10.

Rangin, C , Silver, E. A., von Breymann, M. T., et al., 1990. Proc.ODP, Init. Repts., 124: College Station, TX (Ocean DrillingProgram).

Stout, P. M., 1985. Interstitial water chemistry and diagenesis ofbiogenic sediments from the eastern equatorial Pacific. In Mayer,L., Theyer F., Thomas, E., et al. Init. Repts. DSDP, 85: Wash-ington (U.S. Govt. Printing Office), 805-820.

Suess, E., and Whiticar, M. J., 1989. Methane-derived CO2 in porefluids expelled from the Oregon subduction zone. Palaeogeogr.,Paleoclimatol., Palaeoecol., 71:119-136.

Swart, P. K., and Guzikowski, M., 1988. Interstitial-water chemistry anddiagenesis of periplatform sediments from the Bahamas, ODP Leg101. In Austin, J. A., Jr., Schlager, W., et al., Proc. ODP, Sci.Results, 101: College Station, TX (Ocean Drilling Program), 363-380.

Vollbrecht, R., and Kudrass, H. R., 1990. Geological results of a pre-sitesurvey for ODP drill sites in the SE Sulu Basin. In Rangin, C , Silver,E. A., von Breymann, M. T., et al., Proc. ODP, Init. Repts., 124:College Station, TX (Ocean Drilling Program), 105-111.

von Breymann, M. T., and Berner, U., 1990. Diagenetic dolomiteformation in the Sulu and Celebes seas from continentally derivedturbidite deposits rich in organic matter. Trans. Am. Geophys.Union, 71:1392 (Abstract).

Whiticar, M. J., and Faber, E., 1986. Methane oxidation in marineand limnic sediments and water columns. 12th Intl. Mtg. OrganicGeochem. 1985, Juelich. In Leythaeuser, D., and Rullkotter, J.(Eds.), Advances in Organic Geochemistry (Vol 10): New York,(Springer-Verlag), 759-768.

Date of inital receipt: 3 July 1990Date of acceptance: 11 February 1991Ms 124B-154

215

14. pore-water chemistry of the sulu and celebes seas

Documents