Sulfate Reduction and Diffusion in Sediments of Little Rock Lake ...
Sulfate Reduction and Diffusion in Sediments of Little Rock Lake, Wisconsin Author(s): N. R. Urban, P. L. Brezonik, L. A. Baker, L. A. Sherman Source: Limnology and Oceanography, Vol. 39, No. 4 (Jun., 1994), pp. 797-815 Published by: American Society of Limnology and Oceanography Stable URL: http://www.jstor.org/stable/2838417 . Accessed: 09/05/2011 19:08 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . http://www.jstor.org/action/showPublisher?publisherCode=limnoc. . Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
American Society of Limnology and Oceanography is collaborating with JSTOR to digitize, preserve and extend access to Limnology and Oceanography.
http://www.jstor.org
Limnol. Oceanogr., 39(4), 1994, 797-815 ? 1994, by the American Society of Limnologyand Oceanography,Inc.
in sediments Sulfatereduction and diffusion of LittleRock Lake, Wisconsin N. R. Urban,IP. L. Brezonik,L. A. Baker,2and L. A. Sherman3
Environmental Engineering Program, ofMinnesota,122CME Bldg,500 Pillsbury University Dr.,Minneapolis55455 Abstract
Ratesofsulfate diffusion andreduction weremeasured insediments ofLittleRockLake,an oligotrophic, lakein northern soft-water Wisconsin.Laboratory ofkineticsofsulfatereduction measurements found half-saturation constants(20-30 ,umolliter-') and Q,0 values (2.6) similarto valuesreportedin the literature. Sulfate reduction underinsituconditions insediment coreswaslimitedbysulfate andfollowed similaruptakekinetics as in laboratory experiments. Somevariationin kineticparameters was evident as a function oflocationin thelake.No seasonalvariationwas observedin sulfate reduction ratesin the lakesediments, andlittoral andpelagicsitesexhibited similarrates.Ratesofsulfate reduction weremuch higher thanfluxesofsulfatecalculatedfrompore-water profiles. Pore-water profiles also indicatedlittle indiffusive difference fluxes amongpelagicandlittoral sitesandamongseasons.Thediscrepancy between diffusive fluxes andsulfate reduction ratesisascribedtohighratesofoxidation ofreducedsulfur. Nonlinear ratesofsulfatereduction and calculatedturnover timesofsediment sulfide poolssupport thehypothesis thatsulfideoxidationoccursnearlyas rapidlyas sulfatereduction.
There is a general perception that sulfate reductionis relativelyunimportantforcarbon oxidation in freshwatersediments. Although sulfatereductionaccounts for50-100% of anaerobic carbon oxidation in marine systems (e.g. J0rgensen1989; Howarth 1984; Capone and Kiene 1988), it is thoughtto account for only 10-30% in hypolimneticsediments of lakes (e.g. Kelly and Rudd 1984; Kuivila et al. 1989; Ingvorsenand Brock 1982). Even though freshwatersulfate-reducing bacteria have low half-saturationconstantsforboth sulfateand acetate (e.g. Ingvorsenet al. 1981; Schoenheit et al. 1982; Lovley and Klug 1983) thatenable them to outcompete methanogensin surface sediments,the relativelylow diffusiveinputs of sulfateto lake sediments limit the depth overwhichsulfatereductioncan occur(Lovley and Klug 1983). In oligotrophiclakes,diffusive fluxesof sulfatemore nearlyequal ratesof or-
' Presentaddress:LakeResearchLaboratory, EAWAG, Kastanienbaum, CH-6047Switzerland. 2 Present address:Dept.CivilEngineering, ArizonaState University, Tempe85287. 3 Present address:SoilScienceDept.,University ofWisconsin,Madison53706. Acknowledgments This workwas fundedby theU.S. EPA laboratory at Corvallis,Oregon. Weacknowledge thehelpinthefieldreceivedfromCarl Mach,JaniceTacconi,and Ed Weir,and thelaboratory assistanceofWendyCrescenzoand BethannBarankovic.
ganicmatteroxidation,and sulfatereduction is thoughtto accountfora largerfraction of anaerobiccarbonoxidationthanin eutrophic lakes (Lovleyand Klug 1983). The low concentrations of FeS in lake sedimentsrelative to marinesediments(e.g.Bemer1984; Davison 1988; Urban 1994) are interpreted as anotherindicationthatlittlesulfideis produced fromsulfatereduction in lakes. Because ratesof sulfatereductionin lake sedimentsare thought to be limitedby diffusiveinputsofsulfate(Lovleyand Klug 1983), ratesof diffusive influxoftenare equatedto grossratesof sulfatereduction.Diffusiveinputsof sulfateare low as a resultoflow conof sulfate;concentrations centrations in lakes (10-500 ,umolliter-') are twoto threeorders of magnitudelowerthanin marinesystems. Becausenetdiffusive fluxesofsulfateintolake sediments areproportional tolakesulfateconcentrations (e.g.Kellyet al. 1987;Bakeret al. 1986),sulfatereductionhas been modeledas a first-order processwithrespecttolakesulfate concentrations. For diffusive fluxesofsulfateintolake sedimentsto equal grossor totalratesof sulfate reduction, othersourcesofandsinksforsulfate withinsedimentsmustbe negligible. Landers and Mitchell(1988) demonstrated thatformationofsulfateesterscouldbe an important sinkforsulfatein sediments. Possiblesources forsulfatein sedimentsincludehydrolysis of such sulfateestersand oxidationof reduced
797
798
Urban et al.
sulfur.King and Klug (1982) have shown that pelagicsedimentsduringdifferent seasonsto hydrolysisof sulfateesters provides a negli- examineeffects of sulfateconcentrations, orgiblecontributionofsulfateto sulfate-reducing ganic carbon availability,and temperature. and sulfateconcenbacteria in WintergreenLake. To date, how- Responsesto temperature ever, therehas been no quantificationof the trationalso weremeasuredin sedimentslurDiffusivefluxesof sulrateof regenerationof sulfatein sedimentsvia riesin thelaboratory. fateintothe sedimentswerecalculatedfrom sulfideoxidation. profilesmeasuredin multipleloSeveral recentreportssuggestthat regener- pore-water all seasons.Resultsindicatethat ation of sulfatein sedimentsmay be impor- cationsduring tant.First,ratesof sulfideproductiontypically sulfatereductionis muchfasterthansulfate into the sediments, thatsulfatereare much higherthan rates of reduced-sulfur diffusion at accumulationin sediments(e.g. Chanton et al. ductionis limitedby sulfateconcentration 1987; Bernerand Westrich1985). Second, di- anygivendepthwithinthesediment, and that inratesofsulfate rectlymeasuredgrossratesofsulfatereduction seasonalvariations reduction areminor.Fromtheseresultswe (e.g. Dunnette 1989; Kuivila et al. 1989; King anddiffusion thatreoxidation ofsulfide is an important and Klug 1982) are much higherthandiffusive infer fluxesof sulfateinto lake sediments (Urban mechanism formaintaining thesupplyofsul1994). Such highratesof sulfatereductioncan fatein thesediments. be maintainedonlyifanothersourceof sulfate exists besides diffusionfrom the lake water. Methods - LittleRockLakeis a small Biotic and abiotic reductionof iron and manSitedescription Wisconsin ganese oxides by sulfidecan produce polysul- (17 ha) seepagelake in northern fides,elemental sulfur,thiosulfate,polythio- (45059'N,89042'W).The lake has no stream nates, sulfite,and sulfate (e.g. dos Santos inlets or outlets and receives -99% of water and the reAfonso and Stumm 1992; Pyzik and Sommer inputsfromdirectprecipitation 1981; Allerand Rude 1988; Burdigeand Neal- mainderfromgroundwater inflow.Consethelakewateris verylowindissolved son 1986). Intermediateoxidation states of quently, = 0.002 molliter-1)and sulfurdo not accumulatein sediments;ifthey solids(ionicstrength are formedtheymust be furtheroxidized to alkalinity (25 ,ueqliter-1).The lake has two sulfateor reducedagain to sulfide.Rapid rates basinsofapproximately equal surface area;the of sulfide oxidation have been measured in northbasin has a maximumdepthof 10 m anaerobic marine and lake sediments (Els- (meandepth,3.8 m) and thesouthbasinhas gaard and J0rgensen1992; J0rgensen1990b; a maximumdepthofonly6.3 m (meandepth, J0rgensenand Bak 1990). In the studies of 3.1 m). The lake is dimictic and both basins J0rgensen(1 990b) and J0rgensenand Bak stratify buttheshallowness weaklyin winter, (1990), elemental sulfurand thiosulfatewere of thesouthbasinpreventsformation of any identifiedas initial productsof sulfideoxida- significant in summer.The small hypolimnion tion; subsequent disproportionationand oxi- hypolimnion (8% ofbasinvolume)inthenorth dation of both of theseproductsproduced sul- basin does experienceoxygendepletionand fate. In the study of Elsgaard and J0rgensen becomesanoxicin some years.The lake has (1992), thiosulfateagain was shown to be an been thesiteof an experimental acidification importantintermediate,but theyinferredthat project(Brezoniket al. 1986) since1983.The a pathway for oxidation of sulfideto sulfate twobasinsweredividedfromeach otherat a withoutformationof thiosulfatewas also im- narrowpointwitha polyvinyl curtainin 1984. portant.The quantitativesignificanceof sul- From1985to 1991,thenorthbasinwas acidfide oxidation as a source of sulfate in lake ifiedto successively lowerpH values(5.6, 5.1, and 4.7 each for2-yrperiods)byaddingconsedimentshas yet to be determined. sulfuricacid. Experimental acidifiThe objectives of this studywere to deter- centrated concentraminetheratesofsulfatereductionand thecon- cationcausedan increasein sulfate trolson thisprocessin sedimentsofLittleRock tionin thenorthbasinfrom28 to 70,uM.The Lake, Wisconsin. Rates of sulfatereduction, southbasinreceivedno acid and remainedat measured in intact sediment cores under in its initialpH (6.1) and sulfateconcentration on thelimsitu conditions,were compared in littoraland (28 ,uM).Additionalinformation
Sulfate reductionand diffusion nology,sediment,and water chemistryof the lake can be found elsewhere(e.g. Brezonik et al. 1986, 1993; Baker et al. 1989, 1992). Pore-waterprofiles-Profiles of sulfate in pore waterwere measured frequentlyover the 5-yrperiod from1983 to 1988 (Shermanet al. 1994; Weir 1989; Perry 1987). Pore-water equilibrators(Hesslein 1976) were installedat fivelocationsin thelake on multipleoccasions. The equilibrators,fittedwitha plastic grateto hold them at a fixeddepth in the softpelagic sediments,were leftin the lake for 3 weeks. Sites included littoral (1-m water depths in both basins, 3.5 m in the south basin) and pelagic (9 and 7 m in the northbasin, 5 m in both basins) sediments.Littoralsedimentsare 90% sand; thepelagicsedimentscontain> 40% organicmatterand have a porosity>90%. Followingretrievalofequilibratorsfromthe lake, pore-watersamples were withdrawnand preservedwithin45-60 min. Samples forredox-sensitivespecies (Fe, Mn, H2S) were removed first,preservedwith 0.3 M HCI (metals) or 0.04 M ZnAc (sulfide),and measured within2 d. Sulfatewas measuredby ion chromatography,sulfideand dissolved ironby colorimetry(Cline 1969; Am. Public Health Assoc. 1984), and Mn by atomic absorption withgraphitefurnace. spectrophotometry Fluxes of sulfateto the sedimentswere calculated from37 profilesbased on Fick's first law and the assumption that the profilesrepresentsteadystate conditions.Diffusioncoefficients(fromLi and Gregory1974) were correctedfortemperatureand porosity(measured by weightloss of sedimentsupon drying).A discussion of the concentrationsand fluxesof all measured ions in these same pore-water profilesis given by Sherman et al. (1994).
Measurement ofsulfatereduction in intact sedimentcores-To measure gross or instantaneousratesofsulfatereduction,we tookshort cores fromthe lake; they were injected with trace amounts of 35S042-, incubated at lake temperaturesfor0-17 h, and analyzed forreduced 35S species (J0rgensen1978). Cores were collectedon threeoccasions (August,October, and March) to study the influenceof season and temperature.The lake was weakly stratifiedand ice-coveredin March, stronglystratified in August, and isothermal in October. Sedimenttemperatureson these dates were 4, 23, and 1 C, respectively.Cores were taken
799
from 5-m depths in both basins on each occasion. In addition,coresweretakenfrom0.5 and 8-m depths in the northbasin and 1r-m depth in the south basin in March and from 0.5-m depth in both basins and 7-m in the northbasin in October. To ensurerecoveryof the intact sediment-waterinterface,we used small box cores (30 x 30 x 30 cm; Wildco Co.). Immediatelyupon retrievalof each box core, we inserted7-21 60-ml plastic syringes (with tip-end cut off)by hand into each box core,simultaneouslydrawingup on theplunger. Afterall "minicores" (- 10 cm long) had been inserted,the bottoms were capped and thecoresretrievedand storedin thedarkwhile enrouteto the lab. In the laboratory,the cores were injected with35S042, incubatedat in situtemperatures in the dark,frozen,and lateranalyzed forvarious S fractions.The plasticsyringeshad 5-mm holes at 2-cm intervalsfilledwithsiliconesealer to allow injectionof 35S, MoO42-, acetone, ethanol,or SO42-. Cores receivingonly35SO42were injected at each port with 0.1 ml of carrier-free 35S042- (Amersham) containing7.413.5 nCi35S.The cores were thenincubatedin the dark at the appropriate temperaturefor periods of 0-17 h. Following incubation,the cores were plunged into a bath of dry ice in acetone for 5 min and then stored at - 20?C untilanalyzed (1-10 d later). A minimumof fiveminicoreswas collected fromeach site.One or tworeceivedno 35S042and were analyzed only forpore-wateranions (SO42-, Cl-, F-). One spiked core was frozen immediately (t = 0) to test for recoveryof 35S042-. The remaining3-5 cores were incubated for0.7-17 h beforefreezing.In addition, on each samplingdate, 2-5 cores wereinjected with 0.1 ml 0.2 M Na2S to measure the analyticalrecoveryand to determinewhetheroxidation occurred duringincubation, freezing, ofMoO42-, acetate,ethor storage.The effects anol, and S042- also were examined on two ofthesamplingdates. These coresweretreated as described above, with the sole exception thatthe injected solutioncontained either0.9 M acetate, 17 M ethanol,3-17 mM S042, or 0.9 M MoO42 in addition to the 35S042. These additions resultedin concentrationsin the pore water of 20 mM acetate, 0.4 M ethanol, 70-400,M S042, and 20 mM MoO42. The frozencores were sectioned into 2-cm
800
Urban et al.
increments,thawed underN2 in a solution of 35SO42- and 10 mM Na2SO4. Aftertheywere 2.5 M zinc acetate, then distilledwith 20 ml shaken, the vials were incubated in the dark of 1.5 M HCI for 1.5 h at 90?C. The H2S re- for 0-40 h at temperaturesof 4, 10, 15, 23, leased into the nitrogenstreamwas trappedin and 30?C. Four replicateswere used at each 20 ml of0.2 M NaOH. RecoveryofNa2S stan- temperature,togetherwith a controlcontaindards added directlyto the distillationflasks ing molybdate(10 mM). Althoughvials were was 100%. Recoveryof sulfideadded to cores not shakenduringthe incubations,linearityof beforeincubationrangedfrom55 to 86%, av- sulfatereductionwithtime suggestedthatdiferaging67%; therewas no decreasein recovery fusion limitationwas not importantover the with increasingstoragetime (2 h-2 months). incubation times used. Incubations were The ratesreportedbelow are not correctedfor stopped by immersingvials into a bath of dry this recoveryand hence may underestimate ice in acetone. Samples were storedfrozenfor the actual rate of sulfatereduction. up to 1 monthbeforeanalysis.Beforeanalysis, The solution fromeach trap was diluted to 3 ml ofZn acetatewas added to thevials which 25 ml, and 2 ml were added to 4 ml of scin- werethenthawedundernitrogen.Sampleswere tillationcocktailin plastic minivialsthatwere then distilledin hot acid (12 ml of 6 M HCI) subsequentlycounted on a Beckman LS1800 for90 min under a streamof N2. Sulfidewas scintillationcounter.Quench correctionswere trapped in 0.2 M NaOH which subsequently determinedfromthe H numbermeasured for was made up to a volume of 50 ml. Two mileach sample and a quench curvemeasuredsep- lilitersof this solution were added to 4 ml of arately.Sulfidewas measured colorimetrically scintillationcocktailformeasurementofH235S. (Cline 1969) on an additional subsample from The acid solutionwas centrifuged, rinsedtwice, the NaOH trap. and the solution plus rinseswere made up to After distillation of acid-volatile sulfides a total volume of 100 ml beforemeasurement (AVS), the samples were centrifuged, decant- of 35SO42- by liquid scintillationcounting. Sedimentslurriesalso wereused to examine ed, and rinsedtwicewith 3 M MgSO4. At this point,the sedimentswere dried at 100?C. The the dependence of reductionrates on sulfate supernatantand rinseswerecombined and di- concentration.Surfacesedimentsfromthe 5-m lutedto 100 ml. From thissolution,2 ml were site in the northbasin were collected with an added to 4 ml of scintillationcocktail,and f Ekman dredgeand storedat 4?C for 1 week to activitywas measured as above. Any incor- allow sulfatein the sedimentsto become deporation of 35S042- into sulfate esters (cf. pleted. About 10 ml of sedimentwas placed Landers and Mitchell 1988) that are not hy- into 20-ml vials (screwcap vials with septa) drolyzedduringtheacid distillation(cf.Urban which were then capped and purgedwith N2 and Brezonik 1993) would cause an overesti- for30 min. Concentrationsof sulfatewere admate of the rate of sulfatereduction. Cores justed with a 50 mM solution of Na2SO4 to incubated without 35S042- were frozen as rangefrom5 to 1,000 ,uM;duplicatemeasureabove, sectioned,and thawed underAr. Sam- ments were made at each sulfateconcentraples were then centrifugedunder Ar, and the tion. Afterreceiving10 ,ulof solutioncontaindecanted material was analyzed by ion chro- ing 6.5 nCi 35S042-, the vials were incubated matography (Dionexmodel10)forS042-, Cl-, for2 h on a shakertable in the dark at 2 1C. Followingincubation,thereactionwas stopped and F-. Laboratory measurements ofkinetics ofsul- by immersingthe vials in a bath of dryice in fate reduction-The temperaturedependence acetone. Samples were thawedunderN2, cenof microbial sulfatereductionwas measured trifuged, and both 35S042- and S042- were in sedimentslurriesin the laboratory.Surface measured in the decanted material. sedimentscollectedwithan Ekmandredgewere Two methods were used to calculate rates thoroughlymixed, distributedinto vials (20- of sulfatereduction,R (,umolcm-3h-1), from ml screwcapvials withsepta),and purgedwith the radiotracerexperiments.Typically, rates nitrogenfor 20 min. Vials were then equili- are calculated accordingto brated at the appropriatetemperaturefor 24 R = (35 Sreduced/35SO42 injected) h. At that point, each vial was injected with 1jS042-]-(a/t). (1) 0.3 ml of a solution containing0.2 ,uCiml-'
and diffusion Sulfatereduction 35Sreduced (nCi) is thesumofall reducedforms
801
mained constant during the course of incu-
theamount bations, Eq. 2 was furthersimplifiedto firstcon- orderkinetics: ofradiotracer injected,[SO42-]thesulfate centration (,umolcm-3) in the sample,a the -kf [SO42-] SAox. (3) d([35S042-])=dt isotopefractionation factor,and t the incurate constantequal to V//(Km 1978;Howarthand kfis a first-order bationtime(e.g.J0rgensen 1989).The + [SO42-]).Integrationofthisexpressionyields Teal 1979; Fossingand J0rgensen isotopefractionation factor(1.03-1.06; J0rln([35S042-]/[35S042 injected]) = -k t. (4) in thisstudy,as in gensen1978)wasneglected in all Because of the large fractionof 35S reduced, manyothers,becausetheuncertainties othermeasurements farexceedthiscorrection rate constantscould be obtained as the slopes vs. time. factor.Use of thisequationis based on the ofplotsofln([35SO42-]/[35SO42-injected]) assumptionsthatsulfateuptakefollowsMo- Because oftheformationofend productsother re- than AVS, decreases in 35SO42- were used nod kinetics,thatsulfateconcentrations thatthefrac- ratherthanformationof [35S]AVSto calculate mainconstant during incubation, tionof35S reducedis small,and thatthereis rateconstants.Sulfatereductionrateswerecalno backoxidationofreduced35S(e.g.Hobbie culated as the rate constanttimes the sulfate 1973;Howarthand Merkel1984).In thepres- concentration(Eq. 3). entstudy,thisequationwas used onlywhen to exist(viz. in Results theseconditions werethought frompore-waterequilibrathe laboratory kineticassayswithhighconSulfateprofiles centrations tors-All profiles of sulfate in pore water ofadded SO42-). As discussedbelow,theabove assumptions showed depletion of sulfate below the sediwerenotmetintheassayswithintactsediment ment surface(see figure3 of Sherman et al. cores.In briefincubations(15-60 min),large 1994). Over a depthintervalof 1-10 cm, confractions (5-40%) ofthe35S werereducedeven centrationsof sulfatedecreased fromambient appearedto re- lake values (13-45 ,umolliter-1)to concentrathoughsulfateconcentrations production ofre- tions between 2 and 14 ,umolliter-1. Among mainconstant. Furthermore, duced 35Swas not linearwithtime.Nonlin- the 33 profiles,seven showed peaks in sulfate earitywas attributedto reoxidationof the at or below the sedimentsurface. reduced35S(discussedbelow).The usualmodDiffusivefluxescalculated by application of el had to be modified to accountforthisreox- Fick's law rangedfrom0 to 4.9 ,umolm-2 h-l idationas follows: (Table 1). The average of all fluxes was 1.5?0.94 jmol M-2 h-I (mean ? SD, n= 33). + [S04])} [S042 -] d([35SO4 2- ])/dt = { - Vf/(Km of 35Srecovered, 35S042-injected (nCi)
An average deviation of 22% was observed in fluxes from three pairs of replicate profiles (2) [Sred] SAred. measured on one sampling date. There were and back ratecon- no statisticallysignificantdifferencesin the Vfand Vb are theforward KmandKmb arethehalf-saturation sites (Tacon- magnitudeof fluxesamong different stants, stantsforthe forwardand reversereactions, ble 1). At a given site, fluxeswere higherin [Sred]is the concentrationof reduced sulfur summer than in winterand spring.Also, the andSAoxand depthat whichsulfatedepletionbegan in porereoxidation, speciesundergoing activitiesin theoxidized water profileschanged seasonally; depletion SAredarethespecific ofthisiso- began higherabove the sediment surface in and reducedpoolsofS. Integration topedilutionequationforthecase ofMonod summerthan in winter(Shermanet al. 1994). kineticsdoes notyieldexpressions readilyap- The factoraccountingfor the most variance protocolwe fol- (20%) in the magnitudeof the fluxeswas the plicableto theexperimental we sulfate concentrationin the overlyingwater lowed(cf.Blackburn1979).Consequently, used a simplerapproach-examininginitial column (Fig. 1A). Linear regression(flux = reactionrates.At thestartofall experiments,0.38 + 0.04-[S042-]) yieldsa rate constantof thespecific ofthereduced-sulfur pools 0.37?0.16 (SE, n = 33) m yr-1. Temperature activity is zero,and thesecondtermin Eq. 2 can be accounted foran additional 10% of the varire- ance in fluxes(Fig. 1B). The activationenergy dropped.Because sulfateconcentrations *SAOX + { Vbl(Kmb
+ [Sred]) }
Urban et al.
802
Table 1. Summaryof sulfatediffusionratesmeasured with pore-waterequilibratorsin LittleRock Lake.
Site (waterdepth)
No. of meas- Flux (Mmol m-2h-') urements Mean Range
Littoral 3 South basin (1 m) South basin (3.5 m) 3 3 Northbasin (1 m) Pelagic South basin (5 m) North basin (5 m) North basin (7 m) North basin (9 m)
1.3-3.7 1.3-1.8 1.0-1.5
E [SO42-]* (jumolliter-') Range
2.2 1.5 1.3
A
0.7-2.2 0.5-4.9 0.5-2.4 0-1.9
1.4 1.6 1.5 1.2
4
x
22-28 25-28 30-33 12-27 24-40 12-35 15-50
* Sulfateconcentration1-5 cm above the sedimentsurface.
calculated froman Arrheniusplot of rate convs. inversetemperstants(flux/concentration) ature (significantat P = 0.05) was only 20.9 kJmol-I, correspondingto a QIo of < 1.5. The residual fluxesnot explained by sulfateconcentrationsand temperaturewereinverselyrelated to site depth, but stepwise multiple regressionsindicated this factoraccounted for only 5% of the variance in fluxes.
and reductionin intact Sulfurdistribution
l
E 0
2
U
7 7 4 6
l
0
0
?
_
10
0
so42-
0.6
0.4
B
0 0
0
n
:30.2 D
30 40 50 20 Concentration (A,mol liter-')
0.0
0
_
0
0 00
0
-0.4
sedimentcores-Sulfate profilesin pore waters 0.0034 0.0036 0.0032 of the shortsyringecores (Fig. 2) provide im1/Temperature (?K-') do notcause thatartifacts portantconfirmation errorsin the measured rates of sulfatereducFig. 1. A. Rates of sulfate diffusionwere positively tion(cf.Kellyand Rudd 1984). Profilesshowed correlatedwithconcentrationsof sulfatein the lake water that sulfatewas depleted in the sedimentsat overlyingthe sediments.Stepwise multiplelinear regresanalysisindicatedthatthis factoraccounted for20% all sites, with minor seasonal differencesbut sion of the variance in diffusionrates. B. Diffusionrates also consistent differencesbetween sites. Differ- appeared to be a functionof temperature.Stepwise mulences among sites in the northbasin were not tipleregressionsindicatedthattemperatureaccounted for readily apparent, but consistent differences 10% of the variance in diffusionrates.An Arrheniusplot were observed in the south basin. Concentra- (fluxeswere normalized to sulfateconcentrationsto give rate constants)shows much scatterbut indicatesthat the tions of sulfatewere always lower in all core activationenergyfortherate-controlling processwas only incrementsat the south basin 5-m site than at 5 kcalmol-h. othersites. In contrast,sulfateconcentrations and inventorieswerehigherat the southbasin littoralsite in both October and March. In of some cores may reflectoxidationofreduced general,profileslooked similarto profilesmea- sulfuras a resultof exposureof this sectionto sured with pore-waterequilibrators;concen- air duringcapping of the syringes.In general, trationsdecreased from25-40 ,umolliter-' in sulfateconcentrationsbetween 5- and 10-cm the lake water to 5-20 ,umolliter-1within 5 depthin theminicores(3-20 AM) weresimilar cm of the sedimentsurface.Similarityof the to those measuredin pore-waterequilibrators. profilesin Fig. 2 to profilesfrompore-water Oxidation ofreducedS or hydrolysisoforganic equilibrators indicates that insertion of the S may have caused these concentrationsto be minicoresdid not cause movement of S042- overestimated;the similaritybetween minifromthe overlyingwater into the sediments. cores and pore-water equilibrators despite handlingstrategiessuggeststhat Increasedconcentrationsin thebottomsection vastlydifferent
Sulfate reductionand diffusion
so 0
30 60 90
0
2-
30
concentration 60
I
0
30
I
60
803
(pbmol liter-') 0
30
60
q
0
30
60
I
-~~~~~~~~~~~~~~~~~~~~~C
0
..-......-........ .......... ........... 0 ........... ..............t. --------------------t............-. .-...... 10
V
09"
0
Vi
2
6
~f
8
0-Aug 0Oct -Ma r
South B. 0.5m
North B3.0.5m
'lSouth basin 5m North B. 5m
North B. 8m
Fig.2. Pore-water profiles ofsulfate intheintactcoresusedformeasurement ofsulfate reduction ratesweresimilar to profiles obtainedwithpore-water equilibrators. At the 5-m sitein thesouthbasin,concentrations in thecores werelowerthanconcentrations in theequilibrators. consistently sites. Seasonaltrendswereevidentonlyat thelittoral Theconcentrations shownfortheoverlying waterrepresent withpore-water themeanandrangeofall valuesrecorded at thesesites. equilibrators
such effectsmay not be important.Overesti- riesin March werecomparableto or even highat thesedepths er than inventoriesin October (Table 2). mationofsulfateconcentrations could cause an overestimateof the rates of Reduction of 35S042- to [35S]AVS occurred sulfatereduction. in all cores, but calculation of rates of reducInventories of AVS measured in all short cores used formeasurementof sulfatereduction did not show the expected seasonal and Table 2. Inventories(mmol S m-2) of AVS in sedispatialtrends(Table 2). ConcentrationsofAVS mentsof Little Rock Lake. (Not measured-nm.) generallywereconstantwithdepthin thecores Mar Oct Aug or showed a subsurfacemaximum between 2 (mean?SD) Site (waterdepth) and 4 cm. Inventorieswerelowestin the south basin littoralsitebut weresimilarin all pelagic Littoral 2.4?1.4 1.8?1.2 nm South basin sites independentof water depth. The north nm North basin 47?5 29?11 basin littoralsite was in a shelteredbay, and Pelagic the sedimentscontained considerableleaf lit22?1.8 19?5.4 4.6?0.7 South basin (5 m) ter as well as higher AVS content than the 26?4.0 25?6.4 North basin (5 m) 8.0?0.9 south basin littoralsite. Seasonally,lowest in37?7.6 nm 19?2.4 Northbasin (8 m) ventorieswere observed in August; invento-
804
Urban et al. 1.00
VI)
0
0~~~~
0.20
0) 0.5
0
a)
ci
-U
OT
0.5
0.15
o)U
v
0.91
U~~~~~~~~~~
~s0.25
(1 0) o
0.015
0.0 T30.10
0
\
V
10 5 15 20 Incubation time (hours) Fig. 3. Recoveryof 35S fromintactcoresdecreased o2 0.00 0.25 0.50 0.75 1.00 withincreasing timeof incubation.Shownhereare the o averages(0) and standarddeviations(errorbars)forall of 35S Recovery samples(all dates,all sites,all depthintervals). Recovery w (fracfion of amount injected) in controlswithmolybdate (notshownhere)was 100% Fig.5. Recovery(as a fraction ofthatinjected)of35S evenafter4 h ofincubation, butrecovery in time-0samas AVS decreasedwithdecreasing totalrecovery of 35S pleswas only87%. (i.e. [35S]AVS+ 3"SO42-).Decreasingrecovery of 35S iS to resultfromformation thought of otherend products tionis problematic. Calculationofratesmust besidesAVS. The rateof formation of theseotherend wasproportional tothefonnation ofAVS.Shown takeintoaccounttheformation of end prod- products here are recoveries (per core) for all minicores in March. uctsotherthanAVS and the relatively large conversionof 35S. Recoveryof 35S generally
was < 100% and decreased withincreasingincubationtime(Fig. 3). Failure to recover100% 0.000 0.025 0.050 0.075 0.100 of the 35S within the AVS and SO42- pools 0 probablyresultedfromformationofend products other than [35S]AVS. Incorporation of 35S042- into organic sulfurcompounds (both -2 sulfateestersand carbon-bondedsulfur),elemental sulfur,and pyritehas been observed previouslyin sediments of Little Rock Lake E (Urban and Brezonik 1993; Baker et al. 1989) as well as otherlakes (Fossing and Jorgensen 04 1989; Landers and Mitchell 1988; Rudd et al. 1986a). The distributionsof [35S]AVS in the -6 cores (Fig. 4) generallyexhibitedthe same pattern;activitieswere lowest in the topmostincrementand increasedwithdepth.Formation of otherend productsappeared to be propor-8 tional to the formationof [35S]AVS (Fig. 5) and also increased with depth in the cores. Sulfatereductionratesbased only on appear-1 0 _ L _ I ance of [35S]AVS may, therefore,underestiFig. 4. Depthprofiles of [35S]AVSgenerally showed mate the total rate of sulfatetransformation. increasing activities withdepth.Shownhereare theavIf eitherfirst-order or Monod kineticsapply erageand SE (n = 27) forall coresincubatedlongerthan 25 minin March.Decreasedactivity towardthesurface to the uptake of unlabeled S042, the disapresultsfromincreasedconcentrations ofSO42-. pearance of 35S042- should be firstorderwith [35S]AVS as fraction of 35S injected
Il
and diffusion Sulfatereduction
805
0.0 respectto concentrations of35SO42- provided ofunlabeledSO42-does thattheconcentration notchangeduringtheexperiment. Analysisof v -0.5 pore-water profilesof SO42- beforeand after 0 incubations showedno decreaseinsulfate conalthoughan increaseof 20% was ZD -1.0 centrations, observedin somecases.Plotsofln(35S042-rec/ 35SO 2-.) and ln([35S]AVS/35SO42-i,j) vs. time generally appearedlinearforthefirst1-3 h of X-1 -1.5 incubation andleveledoffat longerincubation times(Fig.6; 3504 2-inis theamountof35SO42-2.07,.1 injectedin each increment, and 35S042rec is 6 4 0 2 theamountof 35SO42- recovered).Rate conFime (h) stants(k,% h-) werecalculatedbylinearreappearedconstant Fig. 6. Rates of sulfatereduction gressionas the slope of suchplotsforall in- for onlythefirst1-3 h. A line (based on nonlinearrecubationtimes
American Society of Limnology and Oceanography is collaborating with JSTOR to digitize, preserve and extend access to Limnology and Oceanography.
http://www.jstor.org
Limnol. Oceanogr., 39(4), 1994, 797-815 ? 1994, by the American Society of Limnologyand Oceanography,Inc.
in sediments Sulfatereduction and diffusion of LittleRock Lake, Wisconsin N. R. Urban,IP. L. Brezonik,L. A. Baker,2and L. A. Sherman3
Environmental Engineering Program, ofMinnesota,122CME Bldg,500 Pillsbury University Dr.,Minneapolis55455 Abstract
Ratesofsulfate diffusion andreduction weremeasured insediments ofLittleRockLake,an oligotrophic, lakein northern soft-water Wisconsin.Laboratory ofkineticsofsulfatereduction measurements found half-saturation constants(20-30 ,umolliter-') and Q,0 values (2.6) similarto valuesreportedin the literature. Sulfate reduction underinsituconditions insediment coreswaslimitedbysulfate andfollowed similaruptakekinetics as in laboratory experiments. Somevariationin kineticparameters was evident as a function oflocationin thelake.No seasonalvariationwas observedin sulfate reduction ratesin the lakesediments, andlittoral andpelagicsitesexhibited similarrates.Ratesofsulfate reduction weremuch higher thanfluxesofsulfatecalculatedfrompore-water profiles. Pore-water profiles also indicatedlittle indiffusive difference fluxes amongpelagicandlittoral sitesandamongseasons.Thediscrepancy between diffusive fluxes andsulfate reduction ratesisascribedtohighratesofoxidation ofreducedsulfur. Nonlinear ratesofsulfatereduction and calculatedturnover timesofsediment sulfide poolssupport thehypothesis thatsulfideoxidationoccursnearlyas rapidlyas sulfatereduction.
There is a general perception that sulfate reductionis relativelyunimportantforcarbon oxidation in freshwatersediments. Although sulfatereductionaccounts for50-100% of anaerobic carbon oxidation in marine systems (e.g. J0rgensen1989; Howarth 1984; Capone and Kiene 1988), it is thoughtto account for only 10-30% in hypolimneticsediments of lakes (e.g. Kelly and Rudd 1984; Kuivila et al. 1989; Ingvorsenand Brock 1982). Even though freshwatersulfate-reducing bacteria have low half-saturationconstantsforboth sulfateand acetate (e.g. Ingvorsenet al. 1981; Schoenheit et al. 1982; Lovley and Klug 1983) thatenable them to outcompete methanogensin surface sediments,the relativelylow diffusiveinputs of sulfateto lake sediments limit the depth overwhichsulfatereductioncan occur(Lovley and Klug 1983). In oligotrophiclakes,diffusive fluxesof sulfatemore nearlyequal ratesof or-
' Presentaddress:LakeResearchLaboratory, EAWAG, Kastanienbaum, CH-6047Switzerland. 2 Present address:Dept.CivilEngineering, ArizonaState University, Tempe85287. 3 Present address:SoilScienceDept.,University ofWisconsin,Madison53706. Acknowledgments This workwas fundedby theU.S. EPA laboratory at Corvallis,Oregon. Weacknowledge thehelpinthefieldreceivedfromCarl Mach,JaniceTacconi,and Ed Weir,and thelaboratory assistanceofWendyCrescenzoand BethannBarankovic.
ganicmatteroxidation,and sulfatereduction is thoughtto accountfora largerfraction of anaerobiccarbonoxidationthanin eutrophic lakes (Lovleyand Klug 1983). The low concentrations of FeS in lake sedimentsrelative to marinesediments(e.g.Bemer1984; Davison 1988; Urban 1994) are interpreted as anotherindicationthatlittlesulfideis produced fromsulfatereduction in lakes. Because ratesof sulfatereductionin lake sedimentsare thought to be limitedby diffusiveinputsofsulfate(Lovleyand Klug 1983), ratesof diffusive influxoftenare equatedto grossratesof sulfatereduction.Diffusiveinputsof sulfateare low as a resultoflow conof sulfate;concentrations centrations in lakes (10-500 ,umolliter-') are twoto threeorders of magnitudelowerthanin marinesystems. Becausenetdiffusive fluxesofsulfateintolake sediments areproportional tolakesulfateconcentrations (e.g.Kellyet al. 1987;Bakeret al. 1986),sulfatereductionhas been modeledas a first-order processwithrespecttolakesulfate concentrations. For diffusive fluxesofsulfateintolake sedimentsto equal grossor totalratesof sulfate reduction, othersourcesofandsinksforsulfate withinsedimentsmustbe negligible. Landers and Mitchell(1988) demonstrated thatformationofsulfateesterscouldbe an important sinkforsulfatein sediments. Possiblesources forsulfatein sedimentsincludehydrolysis of such sulfateestersand oxidationof reduced
797
798
Urban et al.
sulfur.King and Klug (1982) have shown that pelagicsedimentsduringdifferent seasonsto hydrolysisof sulfateesters provides a negli- examineeffects of sulfateconcentrations, orgiblecontributionofsulfateto sulfate-reducing ganic carbon availability,and temperature. and sulfateconcenbacteria in WintergreenLake. To date, how- Responsesto temperature ever, therehas been no quantificationof the trationalso weremeasuredin sedimentslurDiffusivefluxesof sulrateof regenerationof sulfatein sedimentsvia riesin thelaboratory. fateintothe sedimentswerecalculatedfrom sulfideoxidation. profilesmeasuredin multipleloSeveral recentreportssuggestthat regener- pore-water all seasons.Resultsindicatethat ation of sulfatein sedimentsmay be impor- cationsduring tant.First,ratesof sulfideproductiontypically sulfatereductionis muchfasterthansulfate into the sediments, thatsulfatereare much higherthan rates of reduced-sulfur diffusion at accumulationin sediments(e.g. Chanton et al. ductionis limitedby sulfateconcentration 1987; Bernerand Westrich1985). Second, di- anygivendepthwithinthesediment, and that inratesofsulfate rectlymeasuredgrossratesofsulfatereduction seasonalvariations reduction areminor.Fromtheseresultswe (e.g. Dunnette 1989; Kuivila et al. 1989; King anddiffusion thatreoxidation ofsulfide is an important and Klug 1982) are much higherthandiffusive infer fluxesof sulfateinto lake sediments (Urban mechanism formaintaining thesupplyofsul1994). Such highratesof sulfatereductioncan fatein thesediments. be maintainedonlyifanothersourceof sulfate exists besides diffusionfrom the lake water. Methods - LittleRockLakeis a small Biotic and abiotic reductionof iron and manSitedescription Wisconsin ganese oxides by sulfidecan produce polysul- (17 ha) seepagelake in northern fides,elemental sulfur,thiosulfate,polythio- (45059'N,89042'W).The lake has no stream nates, sulfite,and sulfate (e.g. dos Santos inlets or outlets and receives -99% of water and the reAfonso and Stumm 1992; Pyzik and Sommer inputsfromdirectprecipitation 1981; Allerand Rude 1988; Burdigeand Neal- mainderfromgroundwater inflow.Consethelakewateris verylowindissolved son 1986). Intermediateoxidation states of quently, = 0.002 molliter-1)and sulfurdo not accumulatein sediments;ifthey solids(ionicstrength are formedtheymust be furtheroxidized to alkalinity (25 ,ueqliter-1).The lake has two sulfateor reducedagain to sulfide.Rapid rates basinsofapproximately equal surface area;the of sulfide oxidation have been measured in northbasin has a maximumdepthof 10 m anaerobic marine and lake sediments (Els- (meandepth,3.8 m) and thesouthbasinhas gaard and J0rgensen1992; J0rgensen1990b; a maximumdepthofonly6.3 m (meandepth, J0rgensenand Bak 1990). In the studies of 3.1 m). The lake is dimictic and both basins J0rgensen(1 990b) and J0rgensenand Bak stratify buttheshallowness weaklyin winter, (1990), elemental sulfurand thiosulfatewere of thesouthbasinpreventsformation of any identifiedas initial productsof sulfideoxida- significant in summer.The small hypolimnion tion; subsequent disproportionationand oxi- hypolimnion (8% ofbasinvolume)inthenorth dation of both of theseproductsproduced sul- basin does experienceoxygendepletionand fate. In the study of Elsgaard and J0rgensen becomesanoxicin some years.The lake has (1992), thiosulfateagain was shown to be an been thesiteof an experimental acidification importantintermediate,but theyinferredthat project(Brezoniket al. 1986) since1983.The a pathway for oxidation of sulfideto sulfate twobasinsweredividedfromeach otherat a withoutformationof thiosulfatewas also im- narrowpointwitha polyvinyl curtainin 1984. portant.The quantitativesignificanceof sul- From1985to 1991,thenorthbasinwas acidfide oxidation as a source of sulfate in lake ifiedto successively lowerpH values(5.6, 5.1, and 4.7 each for2-yrperiods)byaddingconsedimentshas yet to be determined. sulfuricacid. Experimental acidifiThe objectives of this studywere to deter- centrated concentraminetheratesofsulfatereductionand thecon- cationcausedan increasein sulfate trolson thisprocessin sedimentsofLittleRock tionin thenorthbasinfrom28 to 70,uM.The Lake, Wisconsin. Rates of sulfatereduction, southbasinreceivedno acid and remainedat measured in intact sediment cores under in its initialpH (6.1) and sulfateconcentration on thelimsitu conditions,were compared in littoraland (28 ,uM).Additionalinformation
Sulfate reductionand diffusion nology,sediment,and water chemistryof the lake can be found elsewhere(e.g. Brezonik et al. 1986, 1993; Baker et al. 1989, 1992). Pore-waterprofiles-Profiles of sulfate in pore waterwere measured frequentlyover the 5-yrperiod from1983 to 1988 (Shermanet al. 1994; Weir 1989; Perry 1987). Pore-water equilibrators(Hesslein 1976) were installedat fivelocationsin thelake on multipleoccasions. The equilibrators,fittedwitha plastic grateto hold them at a fixeddepth in the softpelagic sediments,were leftin the lake for 3 weeks. Sites included littoral (1-m water depths in both basins, 3.5 m in the south basin) and pelagic (9 and 7 m in the northbasin, 5 m in both basins) sediments.Littoralsedimentsare 90% sand; thepelagicsedimentscontain> 40% organicmatterand have a porosity>90%. Followingretrievalofequilibratorsfromthe lake, pore-watersamples were withdrawnand preservedwithin45-60 min. Samples forredox-sensitivespecies (Fe, Mn, H2S) were removed first,preservedwith 0.3 M HCI (metals) or 0.04 M ZnAc (sulfide),and measured within2 d. Sulfatewas measuredby ion chromatography,sulfideand dissolved ironby colorimetry(Cline 1969; Am. Public Health Assoc. 1984), and Mn by atomic absorption withgraphitefurnace. spectrophotometry Fluxes of sulfateto the sedimentswere calculated from37 profilesbased on Fick's first law and the assumption that the profilesrepresentsteadystate conditions.Diffusioncoefficients(fromLi and Gregory1974) were correctedfortemperatureand porosity(measured by weightloss of sedimentsupon drying).A discussion of the concentrationsand fluxesof all measured ions in these same pore-water profilesis given by Sherman et al. (1994).
Measurement ofsulfatereduction in intact sedimentcores-To measure gross or instantaneousratesofsulfatereduction,we tookshort cores fromthe lake; they were injected with trace amounts of 35S042-, incubated at lake temperaturesfor0-17 h, and analyzed forreduced 35S species (J0rgensen1978). Cores were collectedon threeoccasions (August,October, and March) to study the influenceof season and temperature.The lake was weakly stratifiedand ice-coveredin March, stronglystratified in August, and isothermal in October. Sedimenttemperatureson these dates were 4, 23, and 1 C, respectively.Cores were taken
799
from 5-m depths in both basins on each occasion. In addition,coresweretakenfrom0.5 and 8-m depths in the northbasin and 1r-m depth in the south basin in March and from 0.5-m depth in both basins and 7-m in the northbasin in October. To ensurerecoveryof the intact sediment-waterinterface,we used small box cores (30 x 30 x 30 cm; Wildco Co.). Immediatelyupon retrievalof each box core, we inserted7-21 60-ml plastic syringes (with tip-end cut off)by hand into each box core,simultaneouslydrawingup on theplunger. Afterall "minicores" (- 10 cm long) had been inserted,the bottoms were capped and thecoresretrievedand storedin thedarkwhile enrouteto the lab. In the laboratory,the cores were injected with35S042, incubatedat in situtemperatures in the dark,frozen,and lateranalyzed forvarious S fractions.The plasticsyringeshad 5-mm holes at 2-cm intervalsfilledwithsiliconesealer to allow injectionof 35S, MoO42-, acetone, ethanol,or SO42-. Cores receivingonly35SO42were injected at each port with 0.1 ml of carrier-free 35S042- (Amersham) containing7.413.5 nCi35S.The cores were thenincubatedin the dark at the appropriate temperaturefor periods of 0-17 h. Following incubation,the cores were plunged into a bath of dry ice in acetone for 5 min and then stored at - 20?C untilanalyzed (1-10 d later). A minimumof fiveminicoreswas collected fromeach site.One or tworeceivedno 35S042and were analyzed only forpore-wateranions (SO42-, Cl-, F-). One spiked core was frozen immediately (t = 0) to test for recoveryof 35S042-. The remaining3-5 cores were incubated for0.7-17 h beforefreezing.In addition, on each samplingdate, 2-5 cores wereinjected with 0.1 ml 0.2 M Na2S to measure the analyticalrecoveryand to determinewhetheroxidation occurred duringincubation, freezing, ofMoO42-, acetate,ethor storage.The effects anol, and S042- also were examined on two ofthesamplingdates. These coresweretreated as described above, with the sole exception thatthe injected solutioncontained either0.9 M acetate, 17 M ethanol,3-17 mM S042, or 0.9 M MoO42 in addition to the 35S042. These additions resultedin concentrationsin the pore water of 20 mM acetate, 0.4 M ethanol, 70-400,M S042, and 20 mM MoO42. The frozencores were sectioned into 2-cm
800
Urban et al.
increments,thawed underN2 in a solution of 35SO42- and 10 mM Na2SO4. Aftertheywere 2.5 M zinc acetate, then distilledwith 20 ml shaken, the vials were incubated in the dark of 1.5 M HCI for 1.5 h at 90?C. The H2S re- for 0-40 h at temperaturesof 4, 10, 15, 23, leased into the nitrogenstreamwas trappedin and 30?C. Four replicateswere used at each 20 ml of0.2 M NaOH. RecoveryofNa2S stan- temperature,togetherwith a controlcontaindards added directlyto the distillationflasks ing molybdate(10 mM). Althoughvials were was 100%. Recoveryof sulfideadded to cores not shakenduringthe incubations,linearityof beforeincubationrangedfrom55 to 86%, av- sulfatereductionwithtime suggestedthatdiferaging67%; therewas no decreasein recovery fusion limitationwas not importantover the with increasingstoragetime (2 h-2 months). incubation times used. Incubations were The ratesreportedbelow are not correctedfor stopped by immersingvials into a bath of dry this recoveryand hence may underestimate ice in acetone. Samples were storedfrozenfor the actual rate of sulfatereduction. up to 1 monthbeforeanalysis.Beforeanalysis, The solution fromeach trap was diluted to 3 ml ofZn acetatewas added to thevials which 25 ml, and 2 ml were added to 4 ml of scin- werethenthawedundernitrogen.Sampleswere tillationcocktailin plastic minivialsthatwere then distilledin hot acid (12 ml of 6 M HCI) subsequentlycounted on a Beckman LS1800 for90 min under a streamof N2. Sulfidewas scintillationcounter.Quench correctionswere trapped in 0.2 M NaOH which subsequently determinedfromthe H numbermeasured for was made up to a volume of 50 ml. Two mileach sample and a quench curvemeasuredsep- lilitersof this solution were added to 4 ml of arately.Sulfidewas measured colorimetrically scintillationcocktailformeasurementofH235S. (Cline 1969) on an additional subsample from The acid solutionwas centrifuged, rinsedtwice, the NaOH trap. and the solution plus rinseswere made up to After distillation of acid-volatile sulfides a total volume of 100 ml beforemeasurement (AVS), the samples were centrifuged, decant- of 35SO42- by liquid scintillationcounting. Sedimentslurriesalso wereused to examine ed, and rinsedtwicewith 3 M MgSO4. At this point,the sedimentswere dried at 100?C. The the dependence of reductionrates on sulfate supernatantand rinseswerecombined and di- concentration.Surfacesedimentsfromthe 5-m lutedto 100 ml. From thissolution,2 ml were site in the northbasin were collected with an added to 4 ml of scintillationcocktail,and f Ekman dredgeand storedat 4?C for 1 week to activitywas measured as above. Any incor- allow sulfatein the sedimentsto become deporation of 35S042- into sulfate esters (cf. pleted. About 10 ml of sedimentwas placed Landers and Mitchell 1988) that are not hy- into 20-ml vials (screwcap vials with septa) drolyzedduringtheacid distillation(cf.Urban which were then capped and purgedwith N2 and Brezonik 1993) would cause an overesti- for30 min. Concentrationsof sulfatewere admate of the rate of sulfatereduction. Cores justed with a 50 mM solution of Na2SO4 to incubated without 35S042- were frozen as rangefrom5 to 1,000 ,uM;duplicatemeasureabove, sectioned,and thawed underAr. Sam- ments were made at each sulfateconcentraples were then centrifugedunder Ar, and the tion. Afterreceiving10 ,ulof solutioncontaindecanted material was analyzed by ion chro- ing 6.5 nCi 35S042-, the vials were incubated matography (Dionexmodel10)forS042-, Cl-, for2 h on a shakertable in the dark at 2 1C. Followingincubation,thereactionwas stopped and F-. Laboratory measurements ofkinetics ofsul- by immersingthe vials in a bath of dryice in fate reduction-The temperaturedependence acetone. Samples were thawedunderN2, cenof microbial sulfatereductionwas measured trifuged, and both 35S042- and S042- were in sedimentslurriesin the laboratory.Surface measured in the decanted material. sedimentscollectedwithan Ekmandredgewere Two methods were used to calculate rates thoroughlymixed, distributedinto vials (20- of sulfatereduction,R (,umolcm-3h-1), from ml screwcapvials withsepta),and purgedwith the radiotracerexperiments.Typically, rates nitrogenfor 20 min. Vials were then equili- are calculated accordingto brated at the appropriatetemperaturefor 24 R = (35 Sreduced/35SO42 injected) h. At that point, each vial was injected with 1jS042-]-(a/t). (1) 0.3 ml of a solution containing0.2 ,uCiml-'
and diffusion Sulfatereduction 35Sreduced (nCi) is thesumofall reducedforms
801
mained constant during the course of incu-
theamount bations, Eq. 2 was furthersimplifiedto firstcon- orderkinetics: ofradiotracer injected,[SO42-]thesulfate centration (,umolcm-3) in the sample,a the -kf [SO42-] SAox. (3) d([35S042-])=dt isotopefractionation factor,and t the incurate constantequal to V//(Km 1978;Howarthand kfis a first-order bationtime(e.g.J0rgensen 1989).The + [SO42-]).Integrationofthisexpressionyields Teal 1979; Fossingand J0rgensen isotopefractionation factor(1.03-1.06; J0rln([35S042-]/[35S042 injected]) = -k t. (4) in thisstudy,as in gensen1978)wasneglected in all Because of the large fractionof 35S reduced, manyothers,becausetheuncertainties othermeasurements farexceedthiscorrection rate constantscould be obtained as the slopes vs. time. factor.Use of thisequationis based on the ofplotsofln([35SO42-]/[35SO42-injected]) assumptionsthatsulfateuptakefollowsMo- Because oftheformationofend productsother re- than AVS, decreases in 35SO42- were used nod kinetics,thatsulfateconcentrations thatthefrac- ratherthanformationof [35S]AVSto calculate mainconstant during incubation, tionof35S reducedis small,and thatthereis rateconstants.Sulfatereductionrateswerecalno backoxidationofreduced35S(e.g.Hobbie culated as the rate constanttimes the sulfate 1973;Howarthand Merkel1984).In thepres- concentration(Eq. 3). entstudy,thisequationwas used onlywhen to exist(viz. in Results theseconditions werethought frompore-waterequilibrathe laboratory kineticassayswithhighconSulfateprofiles centrations tors-All profiles of sulfate in pore water ofadded SO42-). As discussedbelow,theabove assumptions showed depletion of sulfate below the sediwerenotmetintheassayswithintactsediment ment surface(see figure3 of Sherman et al. cores.In briefincubations(15-60 min),large 1994). Over a depthintervalof 1-10 cm, confractions (5-40%) ofthe35S werereducedeven centrationsof sulfatedecreased fromambient appearedto re- lake values (13-45 ,umolliter-1)to concentrathoughsulfateconcentrations production ofre- tions between 2 and 14 ,umolliter-1. Among mainconstant. Furthermore, duced 35Swas not linearwithtime.Nonlin- the 33 profiles,seven showed peaks in sulfate earitywas attributedto reoxidationof the at or below the sedimentsurface. reduced35S(discussedbelow).The usualmodDiffusivefluxescalculated by application of el had to be modified to accountforthisreox- Fick's law rangedfrom0 to 4.9 ,umolm-2 h-l idationas follows: (Table 1). The average of all fluxes was 1.5?0.94 jmol M-2 h-I (mean ? SD, n= 33). + [S04])} [S042 -] d([35SO4 2- ])/dt = { - Vf/(Km of 35Srecovered, 35S042-injected (nCi)
An average deviation of 22% was observed in fluxes from three pairs of replicate profiles (2) [Sred] SAred. measured on one sampling date. There were and back ratecon- no statisticallysignificantdifferencesin the Vfand Vb are theforward KmandKmb arethehalf-saturation sites (Tacon- magnitudeof fluxesamong different stants, stantsforthe forwardand reversereactions, ble 1). At a given site, fluxeswere higherin [Sred]is the concentrationof reduced sulfur summer than in winterand spring.Also, the andSAoxand depthat whichsulfatedepletionbegan in porereoxidation, speciesundergoing activitiesin theoxidized water profileschanged seasonally; depletion SAredarethespecific ofthisiso- began higherabove the sediment surface in and reducedpoolsofS. Integration topedilutionequationforthecase ofMonod summerthan in winter(Shermanet al. 1994). kineticsdoes notyieldexpressions readilyap- The factoraccountingfor the most variance protocolwe fol- (20%) in the magnitudeof the fluxeswas the plicableto theexperimental we sulfate concentrationin the overlyingwater lowed(cf.Blackburn1979).Consequently, used a simplerapproach-examininginitial column (Fig. 1A). Linear regression(flux = reactionrates.At thestartofall experiments,0.38 + 0.04-[S042-]) yieldsa rate constantof thespecific ofthereduced-sulfur pools 0.37?0.16 (SE, n = 33) m yr-1. Temperature activity is zero,and thesecondtermin Eq. 2 can be accounted foran additional 10% of the varire- ance in fluxes(Fig. 1B). The activationenergy dropped.Because sulfateconcentrations *SAOX + { Vbl(Kmb
+ [Sred]) }
Urban et al.
802
Table 1. Summaryof sulfatediffusionratesmeasured with pore-waterequilibratorsin LittleRock Lake.
Site (waterdepth)
No. of meas- Flux (Mmol m-2h-') urements Mean Range
Littoral 3 South basin (1 m) South basin (3.5 m) 3 3 Northbasin (1 m) Pelagic South basin (5 m) North basin (5 m) North basin (7 m) North basin (9 m)
1.3-3.7 1.3-1.8 1.0-1.5
E [SO42-]* (jumolliter-') Range
2.2 1.5 1.3
A
0.7-2.2 0.5-4.9 0.5-2.4 0-1.9
1.4 1.6 1.5 1.2
4
x
22-28 25-28 30-33 12-27 24-40 12-35 15-50
* Sulfateconcentration1-5 cm above the sedimentsurface.
calculated froman Arrheniusplot of rate convs. inversetemperstants(flux/concentration) ature (significantat P = 0.05) was only 20.9 kJmol-I, correspondingto a QIo of < 1.5. The residual fluxesnot explained by sulfateconcentrationsand temperaturewereinverselyrelated to site depth, but stepwise multiple regressionsindicated this factoraccounted for only 5% of the variance in fluxes.
and reductionin intact Sulfurdistribution
l
E 0
2
U
7 7 4 6
l
0
0
?
_
10
0
so42-
0.6
0.4
B
0 0
0
n
:30.2 D
30 40 50 20 Concentration (A,mol liter-')
0.0
0
_
0
0 00
0
-0.4
sedimentcores-Sulfate profilesin pore waters 0.0034 0.0036 0.0032 of the shortsyringecores (Fig. 2) provide im1/Temperature (?K-') do notcause thatartifacts portantconfirmation errorsin the measured rates of sulfatereducFig. 1. A. Rates of sulfate diffusionwere positively tion(cf.Kellyand Rudd 1984). Profilesshowed correlatedwithconcentrationsof sulfatein the lake water that sulfatewas depleted in the sedimentsat overlyingthe sediments.Stepwise multiplelinear regresanalysisindicatedthatthis factoraccounted for20% all sites, with minor seasonal differencesbut sion of the variance in diffusionrates. B. Diffusionrates also consistent differencesbetween sites. Differ- appeared to be a functionof temperature.Stepwise mulences among sites in the northbasin were not tipleregressionsindicatedthattemperatureaccounted for readily apparent, but consistent differences 10% of the variance in diffusionrates.An Arrheniusplot were observed in the south basin. Concentra- (fluxeswere normalized to sulfateconcentrationsto give rate constants)shows much scatterbut indicatesthat the tions of sulfatewere always lower in all core activationenergyfortherate-controlling processwas only incrementsat the south basin 5-m site than at 5 kcalmol-h. othersites. In contrast,sulfateconcentrations and inventorieswerehigherat the southbasin littoralsite in both October and March. In of some cores may reflectoxidationofreduced general,profileslooked similarto profilesmea- sulfuras a resultof exposureof this sectionto sured with pore-waterequilibrators;concen- air duringcapping of the syringes.In general, trationsdecreased from25-40 ,umolliter-' in sulfateconcentrationsbetween 5- and 10-cm the lake water to 5-20 ,umolliter-1within 5 depthin theminicores(3-20 AM) weresimilar cm of the sedimentsurface.Similarityof the to those measuredin pore-waterequilibrators. profilesin Fig. 2 to profilesfrompore-water Oxidation ofreducedS or hydrolysisoforganic equilibrators indicates that insertion of the S may have caused these concentrationsto be minicoresdid not cause movement of S042- overestimated;the similaritybetween minifromthe overlyingwater into the sediments. cores and pore-water equilibrators despite handlingstrategiessuggeststhat Increasedconcentrationsin thebottomsection vastlydifferent
Sulfate reductionand diffusion
so 0
30 60 90
0
2-
30
concentration 60
I
0
30
I
60
803
(pbmol liter-') 0
30
60
q
0
30
60
I
-~~~~~~~~~~~~~~~~~~~~~C
0
..-......-........ .......... ........... 0 ........... ..............t. --------------------t............-. .-...... 10
V
09"
0
Vi
2
6
~f
8
0-Aug 0Oct -Ma r
South B. 0.5m
North B3.0.5m
'lSouth basin 5m North B. 5m
North B. 8m
Fig.2. Pore-water profiles ofsulfate intheintactcoresusedformeasurement ofsulfate reduction ratesweresimilar to profiles obtainedwithpore-water equilibrators. At the 5-m sitein thesouthbasin,concentrations in thecores werelowerthanconcentrations in theequilibrators. consistently sites. Seasonaltrendswereevidentonlyat thelittoral Theconcentrations shownfortheoverlying waterrepresent withpore-water themeanandrangeofall valuesrecorded at thesesites. equilibrators
such effectsmay not be important.Overesti- riesin March werecomparableto or even highat thesedepths er than inventoriesin October (Table 2). mationofsulfateconcentrations could cause an overestimateof the rates of Reduction of 35S042- to [35S]AVS occurred sulfatereduction. in all cores, but calculation of rates of reducInventories of AVS measured in all short cores used formeasurementof sulfatereduction did not show the expected seasonal and Table 2. Inventories(mmol S m-2) of AVS in sedispatialtrends(Table 2). ConcentrationsofAVS mentsof Little Rock Lake. (Not measured-nm.) generallywereconstantwithdepthin thecores Mar Oct Aug or showed a subsurfacemaximum between 2 (mean?SD) Site (waterdepth) and 4 cm. Inventorieswerelowestin the south basin littoralsitebut weresimilarin all pelagic Littoral 2.4?1.4 1.8?1.2 nm South basin sites independentof water depth. The north nm North basin 47?5 29?11 basin littoralsite was in a shelteredbay, and Pelagic the sedimentscontained considerableleaf lit22?1.8 19?5.4 4.6?0.7 South basin (5 m) ter as well as higher AVS content than the 26?4.0 25?6.4 North basin (5 m) 8.0?0.9 south basin littoralsite. Seasonally,lowest in37?7.6 nm 19?2.4 Northbasin (8 m) ventorieswere observed in August; invento-
804
Urban et al. 1.00
VI)
0
0~~~~
0.20
0) 0.5
0
a)
ci
-U
OT
0.5
0.15
o)U
v
0.91
U~~~~~~~~~~
~s0.25
(1 0) o
0.015
0.0 T30.10
0
\
V
10 5 15 20 Incubation time (hours) Fig. 3. Recoveryof 35S fromintactcoresdecreased o2 0.00 0.25 0.50 0.75 1.00 withincreasing timeof incubation.Shownhereare the o averages(0) and standarddeviations(errorbars)forall of 35S Recovery samples(all dates,all sites,all depthintervals). Recovery w (fracfion of amount injected) in controlswithmolybdate (notshownhere)was 100% Fig.5. Recovery(as a fraction ofthatinjected)of35S evenafter4 h ofincubation, butrecovery in time-0samas AVS decreasedwithdecreasing totalrecovery of 35S pleswas only87%. (i.e. [35S]AVS+ 3"SO42-).Decreasingrecovery of 35S iS to resultfromformation thought of otherend products tionis problematic. Calculationofratesmust besidesAVS. The rateof formation of theseotherend wasproportional tothefonnation ofAVS.Shown takeintoaccounttheformation of end prod- products here are recoveries (per core) for all minicores in March. uctsotherthanAVS and the relatively large conversionof 35S. Recoveryof 35S generally
was < 100% and decreased withincreasingincubationtime(Fig. 3). Failure to recover100% 0.000 0.025 0.050 0.075 0.100 of the 35S within the AVS and SO42- pools 0 probablyresultedfromformationofend products other than [35S]AVS. Incorporation of 35S042- into organic sulfurcompounds (both -2 sulfateestersand carbon-bondedsulfur),elemental sulfur,and pyritehas been observed previouslyin sediments of Little Rock Lake E (Urban and Brezonik 1993; Baker et al. 1989) as well as otherlakes (Fossing and Jorgensen 04 1989; Landers and Mitchell 1988; Rudd et al. 1986a). The distributionsof [35S]AVS in the -6 cores (Fig. 4) generallyexhibitedthe same pattern;activitieswere lowest in the topmostincrementand increasedwithdepth.Formation of otherend productsappeared to be propor-8 tional to the formationof [35S]AVS (Fig. 5) and also increased with depth in the cores. Sulfatereductionratesbased only on appear-1 0 _ L _ I ance of [35S]AVS may, therefore,underestiFig. 4. Depthprofiles of [35S]AVSgenerally showed mate the total rate of sulfatetransformation. increasing activities withdepth.Shownhereare theavIf eitherfirst-order or Monod kineticsapply erageand SE (n = 27) forall coresincubatedlongerthan 25 minin March.Decreasedactivity towardthesurface to the uptake of unlabeled S042, the disapresultsfromincreasedconcentrations ofSO42-. pearance of 35S042- should be firstorderwith [35S]AVS as fraction of 35S injected
Il
and diffusion Sulfatereduction
805
0.0 respectto concentrations of35SO42- provided ofunlabeledSO42-does thattheconcentration notchangeduringtheexperiment. Analysisof v -0.5 pore-water profilesof SO42- beforeand after 0 incubations showedno decreaseinsulfate conalthoughan increaseof 20% was ZD -1.0 centrations, observedin somecases.Plotsofln(35S042-rec/ 35SO 2-.) and ln([35S]AVS/35SO42-i,j) vs. time generally appearedlinearforthefirst1-3 h of X-1 -1.5 incubation andleveledoffat longerincubation times(Fig.6; 3504 2-inis theamountof35SO42-2.07,.1 injectedin each increment, and 35S042rec is 6 4 0 2 theamountof 35SO42- recovered).Rate conFime (h) stants(k,% h-) werecalculatedbylinearreappearedconstant Fig. 6. Rates of sulfatereduction gressionas the slope of suchplotsforall in- for onlythefirst1-3 h. A line (based on nonlinearrecubationtimes
