Sedimentology (1990) 37, 997-1009

Anaerobic diagenesis within Recent, Pleistocene, and Eocene marine carbonateframeworks

FRANCIS 1. SANSONE*t, GORDON W. TRIBBLE1*t, CHRISTINE C. ANDREWStandJEFFREY P. CHANTON§

*Department o/OceaIWgraphy, tHawaii Institute o/Geophysics, and tHawaii Institute 0/Marine Biology, University0/Hawaii at MaIWa, 1000 Pope Road, HOIWlulu, Hawaii 96822, USA, §Department o/OceaIWgraphy, Florida State

University, Tallahassee, Florida 32306-3048, USA

ABSTRACT

Porewaters from a variety of Recent, Pleistocene, and Eocene lithified marine carbonate frameworksdisplayed similar chemical characteristics: highly depleted concentrations of dissolved oxygen ( < 20 jlM),

elevated levels of dissolved methane (25-5000 nM), and near-seawater sulphate levels. These porewatersalso had low pH values (7,5-7,9), and contained elevated concentrations of sulphide (4-10 jlM), dissolvedinorganic carbon (2,05-2,46 roM), and inorganic nutrients. Hydrocarbon composition data indicate thatthe methane is biogenic, whereas the methane ()I3c values ( - 47·4 ±2·7%0) suggest that it has been subjectto oxidation. The porewater dissolved inorganic carbon (j13C values varied from - 0·6 to - 3'9%0,suggesting input of carbon dioxide from organic matter oxidation. We conclude that anaerobic diagenesisinvolving bacterial degradation of organic matter is a common process inlithified marine carbonates andhypothesize that it may be an important factor controlling their carbonate geochemistry.

INTRODUCTION

Although much research has been conducted on earlydiagenesis in marine carbonate sediments, relativelylittle attention has been paid to such processes withinmarine carbonate frameworks, which are fundament­ally different from carbonate sediments in that theyare dominated by the advection of seawater throughtheir structure. In contrast, sediments are dominatedby diffusive processes at depths below the zone offaunal bioturbation/irrigation (e.g. Berner, 1980). Inaddition, it has been commonly assumed that theinteriors ofcarbonate frameworks are aerobic becauseof the inward advection of oxygenated seawater, butthis assumption has not generally been tested byanalyses of porewater composition.

Marine geochemists commonly view the interiorsof marine carbonate frameworks as sites for thedeposition of fine-grained material and the precipita­tion of carbonate cements (e.g. Land & Goreau, 1970;

I Present address: US Geological Survey, 677 Ala MoanaBlvd, Suite 415, Honolulu, Hawaii 96813, USA.

Scoffin, 1972; Friedman, Amiel & Schneidermann,1974; MacIntyre, 1977; Ginsburg, 1983), usuallyinvoking inorganic processes as the controlling mech­anisms. Other studies, however, indicate that organicmatter diagenesis is widespread in Recent carbonateframeworks, as evidenced by depleted oxygen and pHlevels, and elevated concentrations of dissolved inor­ganic carbon (DIe), sulphide, inorganic nutrients,and methane (e.g. DiSalvo, 1971; Oberdorfer &Buddemeier, 1983; Risk & Miiller, 1983; Sansone,1985; Sansone et al., 1988a,b, 1990; Tribble, Sansone& Smith, 1990).

Unfortunately, little information is available on thebiogeochemistry of the interstitial waters of oldercarbonate frameworks. Because these waters can beused to provide information on the nature and extentof organic and inorganic diagenesis, vJe collected andanalysed porewaters from a variety of Recent, Pleis­tocene, and Eocene lithified marine carbonates. Ourdata show evidence for active suboxic and anaerobic

997

998 F. J. Sansone et aI.

(b)

Porewater from the cased Davies Reef wells Wascollected with a peristaltic pump (Oberdorfer &Buddemeier, 1986). The SH, NMFS, and MML wells'on Oahu, which are continuously pumped, weresampled by drawing aliquots from the outlet streamsof the wells' centrifugal pumps. The CI well wassampled by lowering a length ofTygon tubing into theuncased well, and then extracting water with aperistaltic pump. A 5-1 Niskin bottle was used tosample the well at Marathon Key. Water was sampledfrom MHSS with diver-operated Niskin bottles.

The pH value (±O·OI standard deviation) and O2

(±5~.,) were measured immediately after samplingusing a battery-powered digital pH meter and a YSIModel 58 dissolved oxygen meter, respectively. Meth­ane and sulphide samples were collected directly inglass containers for later laboratory analysis. Dis­solved sulphide was trapped as ZnS; methane sampleswere preserved with HgCI2 • Samples for nutrients,titration alkalinity (TA), sulphate, and major dissolved

SAMPLE LOCATIONS

Fig. 1. Location of sample sites on (a) Davies Reef, GreatBarrier Reef; (b) Oahu, Hawaii; and (c) the Florida Platform.

METHODS

organic matter oxidation in all of the systems studied.These results suggest that such processes are verycommon in marine carbonate frameworks and mayexert a control on the carbonate geochemistry of thesesystems through their influence on interstitial waterchemistry.

Samples were collected from existing salt-water wellsdrilled into (a) the windward reef flat of Davies Reef,Great Barrier Reef, Australia (Oberdorfer & Budde­meier, 1986), (b) buried carbonate frameworks atseveral coastal locations on the island of Oahu,Hawaii, and (c) the Florida Platform at MarathonKey (Huggins & Block, 1985) (Fig. I). Data are alsoreported for Mud Hole Submarine Springs (MHSS), awarm (36°C) saline submarine spring located at 7 mdepth on the West Florida continental shelf (Kohout,Henry & Banks, 1977; Fanning et al., 1981) that isapparently the result of geothermally driven internalconvection within the Florida Platform (Kohout,1965). The locations and geological settings of thesample sites are described in Table I and illustratedin Figs 1-3.

(c)

(0)

km ~TRADEWINDS0....· --'----!;

60

DAVIES REEF

"II'~.'"KEY WEST<G. .... MK

km

FLORIDA

kmI ! I

o 20

o

N

+

Anaerobic diagenesis in carbonate frameworks

Table 1. Locations and geological settings of the saltwater wells sampled. SH, NMFSand MML subsurface geology adapted from Ferrall (1976).

Davies Reef, Great Barrier Reef. Modern windward reefflat: coral in anunconsolidated coral sand and gravel matrix, underlying a lithified reef plate(l8°48'S, 147°39'E). Wells were located 50 m behind the windward reef margin(FS and FD) and at the mid-point of the reefflat (MS and MD). Wells wereconstructed in 1983 by M. Sandstrom and co-workers, Australian Institute ofMarine Science, Townsville, Qld (Oberdorfer & Buddemeier, 1986).

Oahu, HawaiiSH (Snug Harbor Algal Raceway, University of Hawaii Marine Center). Lithified

interlayered Pleistocene coral and alluvium underlying c. 8 m of Pleistocenebarrier reef and 10m of lithified carbonate lagoonal deposits and corallinedebris; located on the southern shore ofOahu. This well is not usually pumped.

NMFS and MML (US National Marine Fisheries Service, and University ofHawaii Marine Mammal Laboratory). Pleistocene barrier reef flat underlyingc. 10m oflithified carbonate lagoonal deposits and coralline debris; locatedadjacent to Kewalo Basin on the southern shore of Oahu. The NMFS and MMLwells are continuously pumped at 3000 and 1500 I min -I, respectively.

CI (Coconut Island, Hawaii Institute of Marine Biology). Modern lagoonal patchreef overlain with c. 0·3 m of dredged reef rock; located in Kaneohe Bay on theNE shore ofOahu. This well is not usually pumped.

Florida PlatformMK (Marathon Key). Eocene marine carbonate platform, sampled from a higWy

porous limestone sequence (the so-called 'Boulder Zone'; Kohout, 1965) within25 km of the platform margin (24°42'N, 81 °5'W).

MHSS (Mud Hole Submarine Springs). Submarine geothermal spring, 19 kmoffshore on the West Florida continental shelf (26°15'N, 82°00, W).

999

ions were filtered through Whatman GF/F glassmicrofibre filters (0'7 J1.m poresize) into plastic bottles;nutrient bottles were washed with 10% Hel prior touse. Nutrient samples were kept frozen until analysis;TA, sulphate, and major dissolved ion samples were

kept refrigerated until analysis. All samples wereanalysed within 2 weeks of collection.

Hydrocarbon measurements used cryogenic cold­trapping followed by gas chromatography with flameionization detection (±2%; Brooks, Reid & Bernard,

a

E 10-:r:f- 20a...W0 3 0

40

MML SNUG

_ ..~-:-: .... , _ _ FILL AND ALLUVIUM~,_::::==::::-::-::-: __=_N_M_FSt-_-+H....A=R=B=O:;;R;::;::l

·t~~TSWI~E-D~BRiS- //-

~~~~~', './ :/tt--------_, ALLUVIUM AND CINDER , 'f\ , ' ALLUVIUM ;' r-'\..." \ ,,,,,,'....-----, \ I

........ ALLUVIUM wiTH I \ ,\ CORALLI NE I ' I

',GRAVEL I " ./\ I

.......... ~.",/

CORAL AND CORALLINE DEBRIS50 -I.----------r-----------,-----------,r--..J

1000 500 a

DISTANCE INLAND FROM SHORE (m)Fig. 2. N-S subsurface section of the southern coast of Oahu (modified from Ferrall, 1976) showing the approximate locationof wells sampled.

1000 F. J. Sansone et al.

MARATHON WELL...~-----,(a)

ANHYDRITE

50 i faakm

o

____2~DRITE

? .............. TERTIARY

---~ -- ...............COLD ?

WATER . '" UPPER .....SEEPS "', CRETACEOUS, , , ,

....LOWER 'CRETACEOUS

0

1000

E 2000

::r:f-a.. 3000w0

4000

5000

6000

MARATHONWELL

0

200

E

::r: 400f-a..w0 600

800

1000

STRAITS OF FLORIDA(b)

Fig. 3. E-W subsurface section of the Florida Platform (modified from Kohout et al., 1977; Paull et al., 1984; Huggins & Block,1985) showing (a) the location of the Marathon well and the Florida Escarpment cold-water seeps, and (b) the subsurfacestructure adjacent to the Marathon well.

1981; Kipphut & Martens, 1982). Sulphide wasmeasured colorimetrically (Cline, 1969). Inorganicnutrients were determined using a Technicon Auto­Analyzer II (Smith et al., 1981). Dissolved organiccarbon (DOC) was measured using an automatedinfra-red analyser. TAwas determined using thesingle-point method (±0·0 I meq 1- 1) of Smith &Kinsey (1978), except for MK samples, which weredetermined by the multi-point method of Edmond(1970) (±0·004 meq 1- 1). Dissolved inorganic carbon(DIC) was calculated from pH, temperature and TA[after accounting for the contribution ofB(OH)4, S2 -,HS - , NH3 , SiO(OH) -3, and HP04

2- to TA; Dickson

(1981)] using the dissociation constants of Mehrbachet al. (1973); the precision of calculated DIC valueswas to·02 mM (Tribble et al., 1990). Sulphate wasmeasured (to'85 mM) with a Dionex 2010i ionchromatograph with an AS-4 analytical column,except for Davies Reef and MHSS samples, whichwere measured gravimetrically (t 1·3 mM) as BaS04.Salinity was measured (to·003%o) with an AGEMinisal 2100 laboratory induction salinometer. Dis­solved calcium in Oahu samples was measured (±0·02 mM) by EGTA titration (Kremling, 1983). Dis­solved calcium and magnesium in MK samples weredetermined (t 2%) with a Leeman Labs PlasmaSpec-I

Anaerobic diagenesis in carbonate frameworks 1001

inductively coupled plasma emission spectrometer;dissolved sodium was determined (± 3%) by flameatomic absorption spectrometry (Perkin-Elmer 603).226Ra and 222Rn were measured (± 10%) on I-Isamples using the methods of Broecker (1965) asmodified by Mathieu (1977).

Samples for methane stable isotope analyses weretransferred directly into evacuated 2·5-1 glass bottleswith butyl rubber stoppers and preserved with 20 mlof saturated HgCI2 solution. In the laboratory,methane was removed from the samples by spargingwith He and oxidized to CO2over CuO at 800°C; theCO2was purified and then sealed in break-seals. Ten­milJilitre DIC stable isotope samples from MarathonKey were transferred by syringe into evacuated 50-mlvials with rubber septa; the samples were treatedwithin 4 h of collection with 5 ml of He-sparged 30%phosphoric acid saturated with CuS04 . In the labora­tory, CO2was removed from the samples by spargingwith He, purified, and then sealed in break-seals.Analyses of both types of samples were performed bythe Center for Applied Isotope Studies, University ofGeorgia, using a Finnigan-MAT 251 mass spectro­meter; instrumental standard deviations were 0,02­0'14%0' The (j13C values of DIC in samples from theother sites were determined by Global Geochemicalafter extraction of total CO2 in a vacuum line at theUniversity ofHawaii (Kroopnick, 1974); instrumentalstandard deviations were 0·03%0'

Calcite and aragonite saturation states were calcu­lated as the product ofdissolved calcium and carbonateconcentrations divided by the apparent solubilityconstant. The apparent solubility constant for calcitewas calculated as a function of salinity and tempera­ture according to Plath, Johnson & Pykowicz (1980).Aragonite calculations were made using apparentsolubility constants 1·53 times those for calcite, theratio determined by Morse, Mucci & Millero (1980)for 35%0 salinity and 25°C. The range of salinities andtemperatures of our samples were 34,2-35,6%0 and19-28°C, respectively. In both calculations we used acalcium-to-salinity ratio of 11·9 mg 1- 1%0 -1, which isthat of tropical surface seawater (e.g. Tribble et al.,1990). In comparison, calcium-to-salinity ratios meas­ured in Checker Reef, Oahu framework porewaterwere 11·7-12'1 (op. cit.); however, these latter ratioswould have changed the calculated saturation statevalues by less than 2%.

Laboratory experiments were performed to monitordiagenetic changes in seawater incubated with freshcoral rubbl.e and carbonate sand in gas-tight vessels.Twenty-litre plastic containers were tightly wrapped

with aluminium foil, loaded to the top with substrate,filled with Kaneohe Bay.seawater, and then sealedwithout a gas headspace. Sequential samples of theseawater were collected over time from an externalspigot and analysed for DIC, TA, and dissolvedoxygen (Ingvorsen & J0rgensen, 1979); argon-sparged(i.e. degassed) seawater was added to the vessels aftereach sampling to ensure that no gas headspacedeveloped. Control experiments with filtered, argon­sparged seawater without added solid substrate main­tained oxygen concentrations of 3% air-saturation (±1%) over a period of 370 h, thereby indicating theintegrity of the apparatus with respect to atmosphericexchange.

RESULTS

The chemical composition and carbonate saturationstate of the porewaters from the sample sites aresummarized in Table 2, and the measured relationshipbetween TA and DIC is plotted in Fig. 4. The DaviesReef data in Table 2 are for only the four wellssampled by both Oberdorfer & Buddemeier (1983)and Sansone (1985), whereas the Davies Reef data inFig. 4 are for all of the wells sampled by the former.Table 3 compares the major-ion/salinity and radionu­elide/salinity ratios ofthe Florida Platform porewaterswith those ofadjacent seawater. Table 4 lists the (j13C_

2.6,,.------------------,

2.5

24

2.3

2.2

2.1 +----r'-----,,....---r--......--.-----r--~1.9 2.0 2.1 2.2 2.3 24 2.5 2.6

DIC(mM)

Fig. 4. TA (meq 1-1) vs. DIC in (0) surface seawateroverlying Davies Reef, and in the following wells: (x)Davies Reef, (e) NMFS Oahu, <,.) MML Oahu, (.) SHOahu, and (.) CI Oahu. Data in this figure have beennormalized to a standard salinity of 35·0%0 in order tofacilitate comparisons between the different sampling sites.The linear regression for Davies Reef well data (line A) isTA=0·91 DIC+0·29 (r2 =0'95), and for Oahu well data(line A') is TA = 1·02 DIC - 0·07 (r 2 = 0,99). Line B is thehypothesized oxic respiration pathway (see text). DaviesReef data from Buddemeier & Oberdorfer (1986).

ootv

Table 2. Composition and carbonate saturation state of framework porewaters.

Site Depth O2 CH4 NH4 N02+ P04 Si DOC H 2S pH TA S04 Sal. Temp. DIC Saturation state(m) (PM) (nM) (pM) N03 (pM) (PM) (pM) (pM) (meq I-I) (mM) (Yoo) (DC) (mM)

Calcite Aragonite(pM)

Davies ReefFS 2·5 1·9 150 63·0 0·01 1·10 39·0 200 9·8 7·76 2·27 28·3 35·3 28 2·16 2·12 1·39FD 5·3 4·5 373 32·0 0·37 1·20 28·0 125 9·3 7·74 2·32 29·2 35·3 28 2·21 2·08 1·36MS 0·5 4·5 <25 7-4 0·32 0·74 7·7 100 2·2 7·76 2-20 27-4 35·3 28 2·05 2·06 1·35MD 1·9 0·6 62 56·0 0·07 1·10 30·0 24·0 7·75 2·41 30·1 35-4 28 2-30 2-20 1·44

Oahu ~SH 45·0 20·0 137 11·0 1·80 0·76 161 28 4·0 7-48 2·32 34·9 27 2·31 1·16 0·76 ~NMFS 20·0 19·0 109 12·0 0·13 1·30 170 28 4·0 7·58 2·16 34·6 24 2·13 1·22 0·80

~MML 19·0 16·0 - 8·8 0·08 1·10 172 27 4·0 7·56 2·18 34·2 25 2·15 1·21 0·79 ;:sCI 5·5 0·0 84·9 21·9 1·10 3-40 191 4·3 7·60 2·48 34·2 26 2-43 1·55 1·01 isCI 7·5 0·0 92·6 24·4 0·20 3·50 186 4·3 7·54 2·49 34·2 25 2-46 1·32 0·87

;:s~

(1)

Florida Platform ....MK 305 12 1050 0·25 0·07 0·45 39·0 0·5 7-61 2·30 29·6 35·4 - 2·05 2·16 1·41 ~

MK 365 11 1040 0·52 0·19 0·52 - 0·8 7·63 2·26 30·2 35·3 21 2-42 2·10 1·38MK 380 0 5000 0·49 0·04 0·10 7·4 91 - 7·87 2·32 35·6 19 2-21 2·09 1·37MHSS 0 14 101 0·90 0·15 0·34 94·0 4·0 7·30 2-19 28·6 34·9 36 2·23 0·92 0·60

Surface seawater (Oahu)232 7 0·45 0·03 0·03 5·5 55 <I 8·19 2·22 28·2 34·9 26 1·96 4·49 2·94

Depths are distance below water surface in wells. Site abbreviations are as indicated in Table I. Hyphens indicate data not available. CI samples from 5·5 and 7·5 mwere collected at high- and low-tide, respectively. Davies Reef data (except DOC) are from Oberdorfer & Buddemeier (1983) and Sansone (1985). MK samples from 305and 365 m were collected in January 1988. 02' pH, DIC, TA, and temperature data for MK at 380 m are from Huggins & Block (1985); other samples from this depthwere collected in June 1986. MHSS CH4 data are from C. Winn and B. Tilbrook (pers. comm., 1986); other MHSS data are from Fanning et al. (1981). Surface seawatersamples were collected in Kaneohe Bay near the CI site.

Anaerobic diagenesis in carbonate frameworks 1003

that the decomposition of siliceous organisms may besignificant in these systems; alternatively, inorganicsilica dissolution may be important, particularly forthe Oahu sites, which may be influenced by basaltweathering products known to enter nearshore waters(Mackenzie et al., 1981).

(b)

8

100

(c)

(0)

X

/1

DO

8040 60

Time (h)

20

8

6

DO

TA x cP4

/#

aa 2 6

DIG (mM)

8

00

6

00

DIG EJ

(mM) X

fJ<x

/1X !1f1<.

/1/1

/1 /1 /1/1 /1

50

100

150Oxygen

(uM)

O+-"................--f!;.I.L>_....D..-r*__~~X~_--{}-__X::...j

a 10 15 20 25Time (h)

Fig. S. Incubation experiments using fresh carbonate materialand Kaneohe Bay seawater: (a) TA (meq 1-1) vs. DIC, (b)DIC vs. elapsed time, (c) oxygen concentration vs. elapsedtime. The following solid substrates were used: (D) unsortedsurface reef rubble with attached epifaunal community, (x)intertidal carbonate dredge rubble, and (6) subtidal carbon­ate sand.

*Data from Smith & Kroopnick (1981).tData from Whiticar et al. (1986).

DISCUSSION

DIC and b13C-CH4 values obtained for Oahu andFlorida Platform framework porewaters. The resultsof the laboratory incubation experiments are shownin Fig. 5.

Site Depth Ca2+ Mg2+ Na+ 222Rn 226Ra(m)

(mgl-1%0 -1) (d.p.m.I-1%0 -1)

MK 365 11·6 27·5 326 0·840 0·727MHSS 0 15·9 35·1 307 26·6 2-96Deep 500- 11·8 37·0 308

tropical 1000seawater

Site Depth ol3C-DIC 013C-CH4

(m) (%0) (%0)

SH 45 -3·92NMFS 20 -3-63MML 19 -3,53CI 5·5 -2·32CI 7·5 -2'32MK 305 -1,47 -45·9

-45·9MK 365 -0·62 -50,5MK 378Tropical surface

seawater +2·0*Biogenic -60to

methane -1I0t

Table 4. Isotopic composition of framework porewaters.

Table 3. Major-ion/salinity and radionuclide/salinity ratiosof Florida Platform porewaters and adjacent seawater.MHSS and seawater data from Fanning et al. (1981).

Porewater composition

Several consistent patterns can be observed in theporewater data. despite the very different geologicalsettings and d~pths. Dissolved oxygen is highlydepleted compared to surface tropical seawater con­centrations, an:d pH is significantly lower thanseawater values. Ammonia, phosphate and methane,which are produ~ed during the suboxic and anaerobicoxidation of organic matter (e.g. Froelich et al., 1979;Jergensen, 198~), are enriched from typical surfaceseawater concentrations. In contrast, N02 +N03

levels are low, as would be expected for these oxidizedspecies under anaerobic conditions (e.g. Downes,1988). The consistently elevated silica levels suggest

1004 F. J. Sansone et al.

Porewater salinities and dissolved sulphate levelsare not significantly different from seawater values.These results imply that organic matter is beingtransported into the frameworks by seawater advec­tion, and that the decomposition of this materialresults in compositional changes in the porewaters ascompared to the adjacent seawaters. The porewaterconcentrations, however, are obviously the result ofboth production and consumption reactions duringthe transport of the porewaters from the adjacentseawaters to the points of sampling within theframeworks.

Similar results have recently been obtained fromdetailed 'well point' sampling of the porewaters ofChecker Reef, a modern patch reef in Kaneohe Bay,Oahu (Sansone et al., 1988a,b; Tribble et al., 1990).A transect across this reef showed suboxic andanoxic conditions throughout the reef framework withstrong concentration gradients from the windwardand leeward margins of the reef towards the centre.The highest levels of diagenetic end products wereobserved in the reef interior, and were similar to thehighest concentrations reported here (Table 2), exceptthat the maximum dissolved silica was lower (100 jlM),

and the maximum DIC and TA were higher (3,2 mMand 3-4 meq 1- 1, respectively). The importance oforganic matter diagenesis was emphasized by theobserved linear relationships between porewater DICand dissolved ammonia, phosphate and silica (Tribbleet al., 1990).

Major ion ratios for MHSS were found to be similarto those for seawater, thereby leading Kohout et al.(1977) to conclude that the MHSS waters were not theresult of physiochemical processes such as the dis­solution of evaporite deposits within the platform.Similarly, MK porewaters sampled in the presentstudy had major ion ratios similar to that of seawater(Table 3). In contrast, sediment porewaters sampledby the submersible ALVIN in the vicinity of the muchdeeper (3300 m) Florida Escarpment cold water seepsshowed significant differences in the major ion ratiosas compared to seawater (Paull & Neumann, 1987).These latter differences may be due to the dissolutionof evaporite deposits known to occur at depths below1000 m in the Florida Platform (Fig. 3), which maybe too deep to affect the interstitial fluids sampled atMHSS or the Marathon Key well.

Calcium carbonate saturation states were depressedin all framework porewaters sampled relative tosurface seawater (Table 2), a result largely due to theinput of carbon dioxide, which lowered the porewaterpl;:l values. Similar results were noted with the

framework porewaters of Checker Reef, Oahu, inwhich aragonite saturation state values were close toI (Tribbleetal., 1990).

The 222Rn/salinity and 226Ra/salinity ratios meas­ured in MHSS water by Fanning et al. (1981) weremuch higher than those measured in MK porewater(Table 3), perhaps reflecting the geothermal processesacting on MHSS water (Kohout, 1965). However,222Rn in MK porewater was found to be nearequilibrium with its parent, 226Ra, whereas in theMHSS water the 222Rn ratio was nearly 10 timeslarger than the 226Ra ratio. This observation is mostprobably due to a greater degree of system closure inthe MK well than in the MHSS vent.

Finally, the porewaters from Marathon Key areparticularly interesting as they are qualitativelydifferent from the porewaters drawn from the othercarbonate frameworks in at least one respect: bothammonia and phosphate are distinctly lower than atthe other sites, although methane is dramaticallyhigher. While the high methane values may indicatea nearby reservoir of fossil methane, it is also possiblethat organic diagenesis in Marathon Key occurs inextremely reducing microenvironments which favourmethanogenesis more than at the other areas. Overall,it is unclear if the low nutrient and high methaneconcentrations are a consequence of the much greaterage of the platform, the much greater depth of thewell, the increased distance from the ocean, or someother factor.

Evidence for microbially mediated diagenesis

The data presented here indicate that bacterialdegradation of organic matter is occurring in thecarbonate frameworks investigated and is probablythe major factor causing geochemical deviations ofthe interstitial water from surface seawater conditions.The depletion of oxygen and the build up of sulphideand methane is consistent with microbial respiratoryactivity, as is the accumulation of DIC and inorganicnutrients (e.g. J0rgensen, 1983; Tribble et al., 1990,and references therein). Fanning et al. (1981) alsoconcluded that microorganisms were probablyresponsible for the anoxic conditions found in MHSSwaters. Furthennore, the release of isotopically lightcarbon dioxide from the oxidation of organic matter(e.g. Presley & Kaplan, 1968) is consistent with theobserved light DIC Ol3C values for frameworkporewaters as compared to overlying seawater (Table4). However, isotope mass-balance calcu,lations usingDIC data from Table 2 and ol3C-DIe data from

Anaerobic diagenesis in carbonate frameworks 1005

Table 4 yield an unrealistic range of (PC values of-12 to -69%0 for the organic matter being oxidizedto form DIe. This discrepancy may reflect additionalprocesses such as isotopic exchange between thecarbonate frameworks and the associated porefluids.

The hypothesized biogenic origin of the porewatermethane is supported by determinations of the C 1-C3

hydrocarboncontentoffoursamplesofMKporewaterfrom 378 m depth. The methane/(ethane + propane)ratio was 2003 (±282); in comparison, Bernard,Brooks & Sackett (1976) concluded that biogenichydrocarbons from Gulf of Mexico sediments haveratios> 1000.

The presence of elevated methane levels indicatesthe existence of the highly reducing conditions neededfor methane production (Rudd & Taylor, 1980),although such reducing environments may largelyexist in microzones within the frameworks. Theexistence of microenvironments with restricted circu­lation (e.g. Jl1Jrgensen, 1977), in sizes ranging fromsubmillimetre to tens of centimetres, is consistent withthe heterogeneous structure of carbonate frameworks(e.g. Garrett et al., 1971; Scoffin, 1972). The impor­tance of microenvironments is also suggested by thecoexistence ofoxygen, sulphide, sulphate and methanein several of the samples, demonstrating that thesewaters are not in thermodynamic equilibrium withrespect to these species. We hypothesize that sulphideand methane are produced in anoxic subenvironmentsthat allow them to escape into the oxygen-containingbulk interstitial water where their oxidation is kineti­cally limited (e.g. the half times for sulphide oxidationshould be > 100 min, based on the measured pore­water Oz/HzS concentration ratios; Millero, 1986).Note that the low HZS/S04 and CH4 /DIC ratios inthe porewater suggest that the oxidation of sulphideand methane is unlikely to make significant differencesin the sulphate and DIC concentrations.

Further evidence of micro-zonation is the presenceof high sulphate concentrations in the bulk porewater,conditions which normally allow sulphate reducers toout-compete methanogens for organic substrates inmarine sediments, thereby preventing methanogen­esis (e.g. Rudd & Taylor, 1980; Sansone, 1985).Methane released from sulphate-depleted microen­vironments into the bulk porewater would, however,be subject to oxidation, as is indicated by the heavymethane stable carbon isotope ratios observed inthe Marathon Key porewater (-47·4±2·7%0,Table 4), and in the porewater of Checker Reef, Oahu(-46·5±0·8%0, data not shown); the oxidation ofmethane results in increases in b13C-CH4 values due

to the preferential consumption of the light isotope(Barker & Fritz, 1981; Coleman, Risatte & Schoell,1981; Alperin, Reeburgh & Whiticar, 1988).

Diagenetic pathways

The data in Table 2 show a I: I ratio between DICand TA (illustrated in Fig. 4 as pathways A and A'),suggesting that the in-situ geochemistry is dominatedby the oxidation of organic matter during sulphatereduction (after the removal of dissolved oxygen byaerobic respiration) according to the equation:

2 CH ZO+S04Z

- = 2 HC03 - +HzS (I)

where CH zO is used schematically to representorganicmatter. Although sulphate reduction is known to be amajor diagenetic pathway in nearshore anaerobicsediments (e.g. Berner, Scott & Thomlinson, 1970;Thorstenson & Mackenzie, 1974), and has been shownto occur in many carbonate sediments associated withcoral reefs (e.g. Skyring & Chambers, 1976; Macken­zie et al., 1981; Hines & Lyons, 1982; Pigott & Land,1986), it appears that the extent and significance ofthis process within lithified carbonate frameworks hasnot been appreciated. We have recently demonstratedthat up to 75% of the organic matter remineralizationoccurring within Checker Reef, Oahu is due tosulphate reduction (Tribble et al., 1990).

We recognize that a similar I : I stoichiometry canalso result from oxic respiration combined withcarbonate dissolution, but we feel that sulphatereduction is the major reaction pathway for tworeasons. First, levels of dissolved oxygen along thereaction pathway are far too low to account for theobserved increases in DIe. All of the wells haddissolved oxygen concentrations of less than 20 JlM, sothe potential contribution of oxic respiration to thereaction pathway is small. Secondly, the increases inDIC for the Davies Reef data are paralleled by risingsulphide concentrations.

It is also possible that the I: I ratio for DICITAmay be due to sulphate reduction coupled withcarbonate dissolution and FeS formation:

2 CHzO+SO/- +Fez+ +2 CaC03

=4HC03 -+2Ca2++FeS. (2)

Although we cannot determine the extent to whichthis reaction occurs, it will certainly be limited by theavailability of dissolved iron in the porewater. In anycase, the DICand TA changes in the water chemistrywould still result from the anaerobic/suboxic oxidationof organic matter.

1006 F. J. Sansone et al.

The consistent offset of the Davies Reef and MKdata from the Oahu data (Fig. 4) is probably due tothe methodological differences in measurements ofpH and TA, because the slopes ofthe linearregressionsofthe individual data sets are not significantly different(0,98 vs. 1·03 meq I -lmM- 1

, respectively). The offsetof the MHSS datapoint is likely to be too large to beexplained by methodological differences, and hencemay reflect geochemical processes distinctive to thishigher-temperature system.

The characteristics of the wells sampled in thisstudyprovide evidence for the existence ofa diagenetictrend resulting from sulphate reduction along thepathways A and A' (Fig. 4).

(I) Oahu Pleistocene wells. The wells MML andNMFS are continuously pumped at high flowrates. Consequently, these wells are diluted withoxygenated fresh seawater, as reflected by thelocation of their data points at the bottom ofpathway A'. The SH well is not pumped continu­ously, presumably resulting in greater isolationand consequently more extensive diagenetic trans­formation in the porewater, which is reflected inthe location of the SH data point at the midpointof pathway A'.

(2) CI well. The CI well data are located in the upperextreme of pathway A'. This well is not pumped,and is subject to tidal flushing only; it is thereforethe most isolated from fresh seawater of the Oahuand Davies Reef wells, and is the site for the mostdiagenetically altered porewater in these wells.

(3) Davies Reef wells. The MS site is distinctlyshallower than the other Davies Reef wells, anddata from this station fall at the bottom ofpathwayA, thus reflecting the comparatively limited extentof diagenesis. This well also contains the lowestlevels of dissolved methane. The sulphide contentofthe Davies Reef wells increases with increasingDIC and TA (Table 2), which is indicative ofincreasing diagenesis along pathway A.

Further evidence that anaerobic organic matteroxidation may influence internal framework geochem­istry is provided by the laboratory experiments whichmonitored diagenetic changes in seawater incubatedwith fresh coral rubble and carbonate sand in gas­tight vessels. In these experiments the DIC and TAconsistently changed linearly along pathway A, in onecase reaching a DIC value of 6 mM within 35 h (Fig.5a,b). The experiments demonstrate that for a varietyof carbonate substrates there are similar patterns inDIC vs. TA during organic matter respiration,

although there are different kinetics. These data,combined with the results of our porewater sampling(Fig. 4), lead us to hypothesize that linear increases inDIC and TA due to organic matter oxidation may bea regular feature of carbonate framework porewaters.

In order to link the chemistry of frameworkporewater with that of the overlying seawater, wepostulate an oxic respiration pathway having anincrease in DIC with little change in TA (Fig. 4:pathway B):

2 CH 20 +2 O2 = 2 H 2C03 • (3)

At seawater pH most of the added carbonic acid willdissociate, causing a resultant drop in pH. Note alsothat the increase in DIC of the overlying seawater asit is converted along pathway B to the porewater atthe bottom of the anoxic pathway A is roughlyequivalent to the DIC increase which would resultfrom the consumption of the oxygen dissolved inwarm surface seawater (approximately 230 JiM) ac­cording to reaction (2) above. Only small amounts ofoxygen are observed in any of the frameworkporewaters sampled (Table 2), a further indication ofactive oxygen consumption within frameworks.

It is likely that the consumption of oxygen occursrapidly in the systems reported here because no datapoints fall between pathway B and the point repre­senting surface seawater (Fig. 4). This hypothesis issupported by the laboratory experiments describedabove, which also showed that aerobic respiration canremove 90% of the dissolved oxygen in seawaterincubated with fresh coral rubble and other carbonatesediments within a few hours (Fig. 5c). More detailedsampling, particularly in suboxic (as opposed toanoxic; Froelich et al., 1979) carbonate environments,will be needed to determine whether pathway B (Fig.4) is actually present in carbonate frameworks.

CONCLUSIONS

The data presented here suggest that anaerobicdiagenesis is a common process in marine carbonateframeworks, and may be more important thanpreviously recognized. Anaerobic, organically linkedprocesses may exert important controls on the carbon­ate geochemistry of such environments because of theresultant changes in local pH and/or carbonatesaturation state (e.g. Berner, 1980; Ehrlich, 1981;Bathurst, 1983; MacIntyre, 1984). Microbial process­ing will consequently need to be considered in modelsof dissolution and cementation in marine lithified

./

Anaerobic diagenesis in carbonateframew~rks 1007

carbonates, as it has been for carbonates in anaerobicterrigenous sediments (e.g. Hudson, 1977; Nelson &Lawrence, 1984; Ritger, Carson & Suess, 1987) andcarbonate sediments (Mackenzie et al., 1981; Pigott& Land, 1986). In addition, anaerobic organic matterdecomposition in carbonate frameworks may besources of methane and/or ammonia fluxes intoadjacent seawater (Sansone, 1985; Tribble et al.,1988), with subsequent utilization by aerobic chemo­lithotrophs. As an example, it is possible that thisprocess may be supporting the abundant macrofaunalcommunities recently found around submarine cold­water methane seeps (Fig. 3) at the base of the FloridaPlatform (Paull et al., 1984, 1985).

ACKNOWLEDGMENTS

We thank C. Winn and B. Tilbrook for the Mud HoleSubmarine Springs data, and R. W. Buddemeier andD. W. Kinzie for the Davies Reef alkalinity vs. DICdata. We also thank S. V. Smith for extensivediscussions, and F. T. Mackenzie, K. Chave and K.Crook fot critical comments. Reviews were providedby L. M. Walter and J. H. Macquaker. B. Tilbrookcontributed assistance for the Oahu methane analyses,and T. Walsh performed Oahu nutrient analyses. Wealso thank the Florida Keys Aquaduct Authority, theUS National Marine Fisheries Service, the Universityof Hawaii Marine Mammal Laboratory, and theUniversity of Hawaii Snug Harbor Algal Raceway foraccess to their wells. This study was supported byNational Science Foundation grants OCE84-00820and OCE86-{)0803 (Chemical Oceanography Pro­gramme), and 121308-303 (US/Australia CooperativeScience Programme), and by the Australian Instituteof Marine Science. A portion of this research wasconducted as part of the US/Australia MicrobialEcology of Coral Reefs (MECOR) project. HawaiiInstitute of Geophysics Contribution No. 2329;Hawaii Institute of Marine Biology Contribution No.817.

REFERENCES

ALPERIN, M.J., REEBURGH, W.S. & WHItTICAR, M.J. (1988)Carbon and hydrogen isotope fractionation resulting fromanaerobic methane oxidation. Global biogeochem. Cycles,2,279-288.

BARKER, J.F. & FRITZ, P. (1981) Carbon isotope fractionationduring microbial methane oxidation. Nature, 293, 289­291.

BATHURST, R.G.C. (1983) Early diagenesis of carbonatesediments. In: Sediment Diagenesis (Ed. by A. Parker &B. W. Sellwood), pp. 349-377. Reidel, Dordrecht, 427 pp.

BERNARD, B.B., BROOKS, J.M. & SACKEtT, W.M. (1976)Natural gas seepage in the Gulf of Mexico. Earth planet.Sci. Lett., 31, 48-54.

BERNER, R.A. (1980) Early Diagenesis. Princeton UniversityPress, Princeton, 241 pp.

. BERNER, R.A., SCOtT, M.R. & THOMLINSON, C. (1970)Carbonate alkalinity in the pore waters of anoxic marinesediments. Limnol. Oceanogr., IS, 544-560.

BROECKER, W.S. (1965) The application of natural radon toproblems in ocean circulation. In: Symposium on Diffusionin Oceans and Fresh Waters (Ed. by T. Ichiye), pp. 116­145. Lamont Geological Observatory.

BROOKS, J.M., REID, D.F. & BERNARD, B.B. (1981) Methanein the upper water column of the northwestern Gulf ofMexico. J. geophys. Res., 7,11029-11040.

BUDDEMEtER, R.W. & OBERDORFER, J.A. (1986) Internalhydrology and geochemistry of coral reefs and atollislands: key to diagenetic variations. In: Reef Diagenesis(Ed. by J.H. Schroeder & B.H. Purser), pp. 91-111.Springer-Verlag, Heidelberg, ISS pp.

CLINE, J.D. (1969) Spectrophotometric determination ofhydrogen sulfide in natural waters. Limnol. Oceanogr., 14,454-458.

COLEMAN, D.O., RISAlTE, J.B. & SCHOELL, M. (1981)Fractionation of carbon and hydrogen isotopes by meth­ane-oxidizing bacteria. Geochim. Cosmochim. Acta, 45,1033-1037.

DICKSON, A.G. (1981) An exact definition of total alkalinityand a procedure for the estimation of alkalinity and totalinorganic carbon from titration. Deep-Sea Res., 28, 609­623.

DISALVO, L.H. (1971) Regenerative functions and microbialecology of coral reefs. II. Oxygen metabolism in theregenerative system. Can. J. Microbiol., 17, 1091-1100.

DoWNES, M.T. (1988) Aquatic nitrogen transformations atlow oxygen concentrations. Appl. environ. Microbiol., 54,172-175.

EDMOND, J .M. (1970) High precision determination oftitration alkalinity and total carbon dioxide content of seawater by potentiometric titration. Deep-Sea Res., 17, 737­750.

EHRLICH, H.L. (1981) Geomicrobiology. Marcel Dekker, NewYork, 393 pp.

FANNING, K.A., BYRNE, R.H., BRELAND, J.A., BETZER,P.R., MOORE, W.S., ESLINGER, R.J. & PYLE, T.E. (1981)Geothermal springs of West Florida continental shelf:evidence for dolomitization and radionuc1ide enrichment.Earthplanet. Sci. Lett. 52, 345-354.

FERRALL, C.c., JR (1976) Subsurface geology of Waikiki,Moiliili and Kakaako with engineering application. MSthesis, University of Hawaii, Honolulu, 175 pp.

FRIEDMAN, G.M.,AMIEL, A.J. &SCHNEIDERMANN, N. (1974)Submarine cementation in reefs: example from the RedSea. J. sedim. Petrol., 44,816-825.

FROELICH, P.N., KLINKHAMMER, G.P., BENDER, M.L.,LUEDTKE, N.A., HEATH, G.R., CULLEN, D., DAUPHIN, P.,HAMMOND, D., HARTMANN, B. & MAYNARD, V. (1979)Early oxidation of organic matter in pelagic sediments ofthe eastern Atlantic: suboxic diagenesis. Geochim. Cos­mochim. Acta, 43, 1075-1090.

1008 F. J. Sansone et al.

GARRETT, P., SMITH, D. L., WILSON, A. O. & PATRIQUlN, D.(1971) Physiography, ecology, and sediments of twoBermuda patch reefs. J. Geol., 79, 647---{)68.

GINSBURG, R.N. (1983) Geological and biological roles ofcavities in coral reefs. In: Perspectives on Coral Reefs (Ed.by D. J. Barnes), pp. 148-153. Australian Institute ofMarine Science, Townsville, Qld, Australia, 277 pp.

HINES, M.E. & LYONS, W.B. (1982) Biogeochemistry ofnearshore Bermuda sediments. I. Sulfate reduction ratesand nutrient generation. Mar. Ecol. Prog. Ser., 8, 87-94.

HUDSON, J.D. (1977) Stable isotopes and limestone lithifica­tion. J. geol. Soc. London, 133, 637-660.

HUGGINS, J.C. & BLOCK, D.L. (1985) Deep Wells ofOTECApplication. Florida Solar Energy Center, Cape Canaveral,Fla., 76 pp.

INGVORSEN, K. & J0RGENSEN, B.B. (1979) Combinedmeasurement of oxygen and sulfide in water samples.Limnol. Oceanogr., 24, 390-393.

J0RGENSEN, B.B. (1977) Bacterial sulfate reduction withinreduced microniches of oxidized marine sediments. Mar.Bioi., 41, 7-17.

J0RGENSEN, B.B. (1983) Processes at the sediment-waterinterface. In: The Major Biochemical Cycles and TheirInteractions (Ed. by B. Bolin & R.B. Cook), pp. 477-509.Wiley, Chichester.

KIPPHUT, G.W. & MARTENS, C.S. (1982) Biogeochemicalcycling in an organic-rich coastal marine sediment. 3.Dissolved gas transport in methane-saturated sediments.Geochim. Cosmochim. Acta, 46, 2049-2060.

KOHOUT, F.A. (1965) A hypothesis concerning cyclic flow ofsalt water related to geothermal heating in the FloridianAquifer. Trans. New York Acad. Sci., Ser. 2, 28, 249-271.

KOHOUT, F.A., HENRY, H.R. & BANKS, J.E. (1977) Hydro­geology related to geothermal conditions of the FloridianPlateau. In: The Geothermal Nature ofthe Floridian Plateau(Ed. by D.L. Smith & G.M. Griffin), Bureau of Geology,Florida Department of Natural Resources Spec. Publ., 21,1-41.

KREMLING, K. (1983) Determination of the major constitu­ents. In: Methods of Seawater Analysis (Ed. by K.Grasshoff, K. Ehrhardt & K. Kremling), pp. 247-268.Verlag Chemie, Weinheim, 419 pp.

KRooPNICK, P. (1974) The dissolved Oz-COZ_13C system of

the eastern equatorial Pacific. Deep-Sea Res., 21, 211-227.LAND, L.S. & GOREAU, T.F. (1970) Submarine lithification

of Jamaican reefs. J. sedim. Petrol., 30,457-462.MACINTYRE, I.G. (1977) Distribution of submarine cements

in a modern Caribbean fringing reef, Galeta Point,Panama. J. sedim. Petrol., 47,503-516.

MACINTYRE, I. G. (1984) Extensive submarine lithificationin a cave in the Belize barrier reef platform. J. sedim.Petrol., 54,221-235.

MACKENZIE, F.T., RISVET, B.L., THORSTENSEN, D.C., LER­MAN, A. & LEEPER, R.H. (1981) Reverse weathering andchemical mass balance in a coastal environment. In: RiverInputs to Ocean Systems (Ed. by J.D. Burton & D. Hisma),pp. 152-187. NATO and UNESCO, Switzerland.

MATIUEU, G. (1977) Appendix I. Radon-222/radium-226technique ofanalysis. In: Annualreport to ERDA, Transportand Transfer Rates in the Waters of the Continental Shelf,pp. 1-30. Contract EY 76-S-02-2185, Energy ResearchDevelopment Agency, Washington, DC.

MEHRBACH,C.,CULBERSON,C.H., HAWLEY,J.E. &PYTKOW-

ICZ, R.M. (1973) Measurementofthe apparent dissociationconstants of carbonic acid in seawater at atmosphericpressure. Limnol. Oceanogr., 18, 897-907.

MILLERO, F.J. (1986) The thermodynamics and kinetics ofthe hydrogen sulfide system in natural waters. Mar. Chem.,18,121-147.

MORSE, J.W., MUCCI, A. & MILLERO, F.J. (1980) Thesolubility of calcite and aragonite in seawater of 35%0salinity at 25°C and atmospheric pressure. Geochim.Cosmochim. Acta, 44, 85-94.

NELSON, C.S. & LAWRENCE, M.F. (1984) Methane-derivedhigh-Mg calcite cement in Holocene nodules from theFraser River Delta, British Columbia, Canada. Sedimen­tology, 31, 645---{)54.

OBERDORFER, J.A. & BUDDEMEIER, R.W. (1983) Hydrogeol­ogy of a Great Barrier Reef: implications for groundwaterin reef and atoll islands. Eos, 64, 735. .

OBERDORFER, J.A. & BUDDEMEIER, R.W. (1986) Coral-teefhydrology: field studies of water movement within abarrier reef. Coral Reefs,S, 7-12.

PAULL, C.K., HECKER, B., COMMEAU, R., FREEMAN-LYNDE,R.P., NEUMANN, C., CORSO, W.P., GOLUBIC, S., HOOK,J.E., SIKES, E. &CURRAY, J. (I 984) Biological communitiesat the Florida Escarpment resemble hydrothermal venttaxa. Science, 226, 965-967.

PAULL, C.K., JULL, A.J., TooLlN, L.J. & LINICK, T. (1985)Stable isotope evidence for an abyssal seep community.Nature, 317,709-711.

PAULL, C.K. & NEUMANN, A.C. (1987) Continental marginbrine seeps: their geological consequences. Geology, 15,545-548.

PIGOTT, J.D. & LAND, L.S. (1986) Interstitial water chemistryof Jamaican reef sediment: sulfate reduction and sub­marine cementation. Mar. Chem., 19, 355-378.

PLATH, D.C., JOHNSON, K.S. & PYKTOWICZ, R.M. (1980)The solubility of calcite-probably containing magnes­ium-in seawater. Mar. Chem., 10, 9-29.

PRESLEY, B.J. & KAPLAN, I.R. (1968) Changes in dissolvedsulfate, calcium and carbonate from interstitial water ofnear-shore sediments. Geochim. Cosmochim. Acta, 32,1037-1048.

RISK, M.H. & MOLLER, H.R. (1983) Porewater in coralheads: evidence for nutrient regeneration. Limnol. Ocean­ogr.,28,1004-1008.

RITGER, S., CARSON, B. & SUESS, E. (1987) Methane-derivedauthigenic carbonates formed by subduction-inducedpore-water expulsion along the Oregon/Washington mar­gin. Bull. geol. Soc. Am., 98, 147-156.

RUDD, J.W. & TAYLOR, C.D. (1980) Methane cycling inaquatic environments. Adv. aquatic Microbial., 2,77-149.

SANSONE, F.J. (1985) Methane in the reef flat porewaters ofDavies Reef, Great Barrier Reef, Australia. Proc. Sth Int.Coral ReefCong., 3, 415-419.

SANSONE, F.J., ANDREWS, C.C., BUDDEMEIER, R.W. &TRIBBLE, G.W. (l988a) Well point sampling of reefinterstitial waters. Coral Reefs, 7,19-22.

SANSONE, F.J., TRIBBLE, G.W., BUDDEMEIER, R.W. &ANDREWS, C.C. (1988b) Time and space scales ofanaerobic diagenesis within a coral reef framework. Proc.6th Int. Coral ReefSymp., 3, 367-372.

SCOFFIN, T.P. (1972) Fossilization of Bermuda patch reefs.Scknce,178,1280-1282.

SKYRING, G.W. & CHAMBERS, L.A. (1976) Biological sulfate

Anaerobic diagenesis in carbonate frameworks 1009

reduction in carbonate sediments of a coral reef. Aust. J.mar. freshwater Res., 27,595---602.

SMITH, S.Y., KIMMERER, W.J., LAWS, E.A., BROCK, R.E. &WALSH, T.W. (1981) Kaneohe Bay sewage diversionexperiment: perspectives on ecosystem responses tonutritional perturbation. Paci! Sci., 35, 279-395.

SMITH, S.Y. & KINSEY, D. W. (1978) Calcification and organiccarbon metabolism as indicated by carbon dioxide. In:Coral Reefs: Research Methods (Ed. by D.R. Stoddart &R.E. Johannes), UNESCO Monographs on OceanographicMethodology, 5, 469--484. UNESCO, Paris.

SMITH, S.Y. & KROOPNICK, P. (1981) Carbon-13 isotopefractionation as a measure of aquatic metabolism. Nature,294,252-253.

THORSTENSON, D.C. & MACKENZIE, F.T. (1974) Timevariability of pore water chemistry in recent carbonatesediments, Devil's Hole, Harrington Sound, Bermuda.Geochim. Cosmochim. Acta, 38,1-19.

TRIBBLE, G.W., SANSONE, F.J., LI, Y.-H., SMITH, S.Y. &BUDDEMEIER, R.W. (1988) Material fluxes from a reefframework. Proc. 6th Int. Coral ReefSymp. 2, 577-582.

TRIBBLE, G.W., SANSONE, F.J. & SMITH, S.Y. (1990)Stoichiometric modeling of carbon diagenesis within acoral reef framework. Geochim. Cosmochim. Acta, 54,2439-2449.

WHlTICAR, M.J., FABER, E. & SCHOELL, M. (1986) Biogenicmethane formation in marine and freshwater environ­ments: CO2 reduction vs. acetate fermentation-Isotopicevidence. Geochim. Cosmochim. Acta, SO, 693-709.

(Manuscript received 15 June 1989; revision received 4 May 1990)