Radiocarbon in Dissolved Organic Carbon, A ... - Wiley Online Library

WATERRESOURCES RESEARCH,VOL. 27,NO. 8, PAGES1975-1986, AUGUST1991

Radiocarbonin Dissolved OrganicCarbon, A PossibleGroundwater Dating Method' Case Studies From Western Canada LEONARD WASSENAAR1 AND RAMON ARAVENA Waterloo Centerfor Groundwater Research,Departmentof EarthSciences,University of Waterloo,Waterloo,Ontario,Canada JIM HEN DRY

National Hydrology ResearchInstitute, Saskatoon,Saskatchewan,Canada PETER FRITZ

Institutfiir Hydrologie,GSF Miinchen,Neuherberg,Germany

Thispaperexplores thefeasibility of using14Cin dissolved organic carbon(DOC)asanalternative isotopicgroundwaterdating method. Two hydrogeologicallycontrastinggroundwatersystemswere tested;the Cretaceousage Milk River aquifer, and low-permeability,organicrich, Wisconsinanage Prairie tills in southernAlberta, Canada. Comparisonsof radiocarbondata were made between DOC fractions, dissolvedinorganiccarbon(DIC), and severalDIC geochemicalage correctionmodelsalong

welldefinedflowpaths.Thedatapresented demonstrate that 14Cdatingof DOCfractions canindeed provide an alternative method for determiningisotopicgroundwaterages, under suitableconditions. However, detailed information may be required regarding(1) the geologicnature of the aquifer and its

flowsystem,(2) the isotopically conservative behaviorof DOC, (3) theinitial•4Cactivityof DOC in recharge, and(4)theeffectof bacterial redoxprocesses onthe •4Cactivityof DOC.In theMilk River aquifer,DOC •4Cagesweresuccessfully usedto estimate groundwater residence times,aswellasto furtherrefineinputparameter assumptions fortheDIC method. In thePrairietills,DOC 14Cageswere used to establish a maximum age for the pore waters in an environment where the DIC method is especially problematic.

INTRODUCTION

geochemicalreactionsthat occur along the flow path. Equations (1) and (2) allow for the correction of DIC radiocarbon Radiocarbondating of the dissolved inorganic carbonate agesfrom groundwater[e.g., Pearson and Hanshaw, 1970; species (DIC) in groundwater has been used by hydrologists Mook, 1972; Tamers, 1975; Wigley et al., 1978; Fontes and toestimateisotopic groundwater ages since the early work Garnier, 1979; Reardon and Fritz, 1978; Cheng and Long, ofMiinnich[1957].It was recognizedearly on, however, that 1984].

carbonate mineraldissolution[Ingersonand Pearson, 1964] A valid Q factor determinationfor most DIC correction andothergeochemical processesmay dilute the radiocarbon models, however, requires knowledge of several factors. contentof DIC with "dead" inorganic carbon, leading to anomalously old isotopicgroundwaterages. This led to the

These include major ion chemistry,p H, alkalinity, recharge

thatattemptto quantifya DIC dilutionfactor Q, whichis

aquifer carbonate minerals. Moreover, various other geochemical processes can occurin groundwaterthat affect

contentof theDIC and development of variousgeochemicaland isotopicmodels P(CO2),and•3Cand•4Cisotopic incorporated intotheequation for determining •4Cages, Age= 8266.6 In (At/Ao) + 8266.6 In Q

the DIC and its •4C activity, suchas sulfideoxidation,

(1)

where A0 is theinitial •4Cactivityof the DIC enteringthe

sulfate reduction, methanogenesis, lignite oxidation, and carbonate mineral precipitation. in contrastto the DIC, the carbonisotopegeochemistryof

groundwater flowsystem, At isthemeasured •4Cactivityof dissolvedorganiccarbon(DOC) in groundwatermay be less theDIC sample,and Q is the •4C dilutionfactor.The complex.In groundwatersystemsDOC may originatefrom dilution factorin itssimplest formcanbedefined as two possiblesources:(1) solubleorganicmatteroriginating from decomposingorganic carbon in the soil zone that is subsequentlytransportedto the groundwatervia recharge, where DICinitia • is the inorganiccarboncontentof the and (2) kerogenor buriedorganicdepositsindigenousto the

Q = (DlCinitial)/(DlCfinal)

(2)

groundwater at time= 0 or firstpointalongtheflowpathand aquifersediments[Thurman,1985a; Aravenaet al., 1989].

DIC•na• isthesumofDieinitial and"dead"DICadded from

Becausekerogenin ancientgeologicformationsis frequently highlyrefractory,insoluble,and presentin minuteamounts [Degens,1967],one may hypothesizethat DOC in the pore watersof manygeologicformations(with the exceptionof

1N'•-'•-•w atNational Hydrology Research Institute, Saskatoon, Saskatchewan, Canada. Copyright 1991 bytheAmerican Geophysical Union. Paper number 91WR00504.

0043-1397/91/91 WR-00504505.00

organicrich environments) wouldpredominantly originate from solubleorganicmatter rechargedfrom the soil zone. Dissolvedorganiccarbonmay be transportedwithin an 1975

1976

WASSENAARET AL.' RADIOCARBONIN DISSOLVED ORGANIC CARBON

CALGARY

51ON

50 ø

VAUXHALLo

•STUDY

AREA2

ETHBRIDGE STUDY

•_.•,•,/A REA1 49øN

114øW

113 ø

112 ø •".•._•111 ø

110øW

SWEETGRASS • HI LLS

0 I

40 I

km

I

Fig. 1. Locationof studyareas.Studyarea 1 are wells screenedalonga well-definedflow path in the Cretaceousage Milk River aquifer. Study area 2 (A and B) are piezometer nests in clayey Prairie till deposits.

aquifer system at groundwater velocity. Like radiocarbonin

The secondstudy area comprisestwo sites in organicrich,

the DIC pool, the •4C contentof the DOC will undergo low-permeability clayey glacial tills. Both study areasare exponential radioactive decay and could provide a suitable tracer for estimating groundwaterresidencetimes and ages, provided only one carbon source existsfor the DOC. Using radiocarbon in DOC fractions could provide a groundwater dating tool that may not require complexinorganicgeochemical correction models, as is required for DIC isotopic age dating. Radiocarbon studiesof DOC in groundwaterare relatively new [Spiker and Rubin, 1975; Thurman, 1985b; Murphy et al.,1989a,

b; Pettersson et al., 1989; Wassenaar et al.,

located in southern Alberta, Canada. In addition, complementary HMW DOC radiocarbon data are reportedfrom several shallow aquifers in Ontario, Canada, that shedlight on the potential effects of biogeochemicalprocessesonthe carbon isotopic evolution of DOC in groundwater.The

objectives weremetby analyzing•4Cin DOC fractions from wells alongwell-definedgroundwaterflow paths;evaluating the relevance of DOC radiocarbon ages in light of the potential organic carbon sources, the flow system,and geologicsetting;and comparingDOC radiocarbonageswith

1989, 1990a, b]. Spikerand Rubin [1975]first used •4C the established DIC •4Cmethodand severalDIC geochemanalysesof DOC from groundwaterto estimatethe degreeof ical agecorrectionmodels.The researchpresented hereis petrochemicalpollution from industrialpoint sources.Thur- partof an extensiveinvestigation of DOC andDIC cycling in man[1985b]analyzedthegeochemistry and14Ccontentof Canadiangroundwatersystems[Aravena et al., 1989;Was. one high molecular weight (HMW) groundwater humate sample from the Biscayne aquifer in Florida. Murphy et al. [ 1989a, b] used carbon isotopesof HMW and low molecular weight (LMW) DOC fractions to characterize and evaluate the origins of DOC fractions in the Stripa granite and the confined Milk River aquifer and to evaluate whether DOC could be used as a dating tool. Pettersson et al. [1989] reported and comparedradiocarbonactivitiesof three HMW DOC and DIC samples from granitic rocks in Sweden and

senaar, 1990; Wassenaar et al., 1989, 1990a, b]. STUDY

SITES

Milk River Aquifer

The Milk River sandstone aquiferunderlies much of southern Alberta,Canada,andis confinedabovebyshales ofthePakowkiFormation andbelowby shales ofColorado

suggested thatDOC i4c activities couldprovideanalterna- Group. Recharge totheaquiferoccurs wheretheMilkRiver tive isotopic groundwater dating method. In these studies the hydrologic flow systemswere not well defined, limiting

Formation outcrops in theSweetgrass Hills,Montana (Figure 1). The geochemistry of this artesianaquiferhasbeen

theevaluation of •4Cin DOC asa groundwater datingtool. extensively studied [Meyboom, 1960;Schwartz andMuhlen' The objective of this paper is to explore in greater detail bachs,1979;Phillipset al., 1986;Hendry andSchwart:, thefeasibility andpotential problems ofusing•4CinDOCas 1988,1990]andwasalsothe focusof an international an alternative isotopic groundwater dating method and to programme usingvarious methods for datingoldgroundwa' better establish grounds for further testing. For this research, two well-documented and hydrogeologicallycontrasting groundwater systems were selected for detailed investigations.The first study area is a well-definedhydrologic flow path in the confinedMilk River sandstoneaquifer.

ters[e.g.,HendryandSchwartz, 1988;Murphy etal., 1989b; Nolte et al., 1991].

Murphy et al. [1989a,b] presented a comprehensive

carbon isotopic andgeochemical characterization ofI)OC fractions intheMilkRiveraquifer using samples collected

WASSENAAR ET AL.' RADIOCARBON IN DISSOLVED ORGANIC CARBON

1977

throughout theaquifer.Althoughsomecomparisons were additionalinsightinto the problemsassociatedwith DIC

made between DOCandDIC•4Cages, meaningful interpre-radiocarbondatingin the tills. The organiccarboncycleof tations ofDOCradiocarbon datain termsof isotopic ground- the tills are discussed in a generalfashionhere;a thorough water agesin theaquiferwerehinderedby a lackof control evaluationof the organiccarboncycleandisotopiccompoonhydrologic flowpaths.HerenewDOCandDIC radiocar- sitionof dissolvedand solidorganiccarbonin the tills are bondataarepresented from flowingwellsscreened in the given by Wassenaaret al. [1990b]. eastern Milk River aquifer near the rechargezone and

sampled alonga we!l-defined hydrologic flowpathin the

ANALYTICAL

METHODS

eastern part of the aquifer (flow path 1 of Hendry and Schwartz [1988]).Well MR-85 is locatedapproximately 6 km Sampling Methods fromtherechargearea, and well MR-52 is locatedapproxiSamplingactivitieswere undertakenduringthe summers mately 12kmdowntheflowpathfromMR-85.Groundwater of 1987-1988.Groundwatersamplesfor DOC determinations velocities alongthis sectionof the flow path were estimated alone were filtered in the field through 0.45 filters into onthebasisof hydraulicparametersand the interpretationof precleaned25-mL glassbottles,and immediatelyacidifiedto 180distribution to rangebetweenabout0.5 and1.5m yr-• pH 2 with phosphoricacid. Sampleswere analyzedfor DOC [Hendry and Schwartz,1988;Drimmie et al., 1991].No content on a Dohrmann Carbon Analyzer within 48 hours significant geochemical redoxprocesses (i.e., NO•-, SOft after returning to the lab [Wassenaar et aI., 1990b]. reduction, methanogenesis)occur along this 20-km section Large groundwatersamples(20-120 L groundwater)were oftheaquiferflow path [Hendry and Schwartz, 1990]. collected from the study sites for the isolation of HMW and AlbertaPrairie Tills

Clayrich glacialtills cover most of the CanadianPrairies andoverlytheir Paleozoic provenance [Fenton, 1984]. The tillsrangebetween 35 and 60 m in thickness and were depositedduring the mid-late Wisconsinan glaciations [Stalker,1977; Rutter, 1984' Vreeken, 1989]. The low permeabilityof the clay tills have prompted interest in these formations as potential hazardous waste storage sites. Although the hydrologyand hydrochemistry of till pore waters in the Canadian Prairies has been thoroughly investigated [Hendry,1982, 1988;Hendry et aI., 1986; Keller et aI., 1986], conventional radiocarbon dating of the DIC of these till pore waterswas restricted by small sample sizes, and has only recently becomefeasible with the availability of accelerator massspectrometry(AMS). In southern Alberta, the tills can be divided into two

LMW DOC fractionsfrom groundwater.A HMW fraction of the DOC, in the form of aquatic fu!vic acids, was isolated

fromlargegroundwater samples usingtheXAD-8(TM)technique describedby Thurman and Malcolm [1981]. A LMW DOC fraction of the DOC was concentratedon Silicalite(T•d) molecular sieve for carbon isotopic analyses using techniques describedby Murphy et al. [1989a, b]. These two methods do not isolate 100% of the DOC; for the Milk River

aquifer they account for approximately 45% of the total DOC, whereas, in the tills they account for approximately 70-80% of the total DOC [Wassenaar, 1990]. Major ion chemistry of groundwater sampleswere determined using standard techniques. Dissolved inorganic car-

bon (DIC) samples for •3C and •4C determinations were quantitatively precipitated from 2-60 L groundwater samples by addition of BaC12 at pH > 10, and prepared for carbon isotopic analysesusing standard techniques.

distinct hydrogeologic zones:an upperoxidizedzone, anda Carbon Isotope Analyses lower unoxidized zone [Hendry, 1982].Active groundwater The HMW DOC samples (fulvic acids) eluted from the flowin the upper few meters of the oxidized zone is controlled by a network of interconnectedfractures. Pore waters in theupper5 m of fracturedtill are tritiatedand show

XAD-8 were desalted, freeze dried, and stored in sealed

glassvials [cf. Thurmanand Malcolm, 1981].Subsamplesof

isotopic evidenceof water loss by evaporation[Hend•y, the purifiedHMW humic substanceswere combustedin a 1988]. Groundwater movementin the deeperunoxidized pure oxygen atmosphere,and the resulting CO2 gas was andpurified cryogenically for •3Cand•4Canalyses. zone isdominated by matrix flow. Hydraulic conductivities trapped inthiszoneare estimatedto be of the order of 10-•ø m s-•

LMW DOC samples sorbed on Silicalite were thermally

orlower,resultingin negligiblegroundwatermovement. desorbed and combusted at 800øC under vacuum in the Pore waters in theunoxidized till zoneareverydepleted in presenceof cupric oxide. The CO2 produced from the

•80 and 2H,suggesting theyoriginated under cooler climaticcombustion of

sorbed organics was cryogenically isolated

One coalfragmentremoved conditions thanpresent[Hendry,1988].Stratigraphy of the for •3Cand •4Cdeterminations. tillswasdetermined fromcorestakenduringinstallation of froma coreat till siteB wasalsoanalyzedfor •4Cand 13C piezometers, and examinationof coresshowednumerous isotope content. DOC sampleswere analyzedfor their radiocarboncontent c0alfragments derivedfrom Cretaceous agebedrockare of •4Catomsby TandemAccelerator present in the upper 10-15 m of the oxidizedzone. Coal by directcounting

fragments werenotapparent in cores deeper in theunoxi- Mass Spectrometry(TAMS) at the University of Toronto dizedtill zone.

(IsotraceLaboratory). Milk River DIC sampleswere con-

Two tillsites used inaprevious study [Hendry, 1988] were vertedto CO2 gasby reactingwith 100% phosphoricacid for •4Cby the standard benzene technique at chosen for the evaluation of •4Cin DOC as a possibleandanalyzed groundwater datingmethod(studyArea2, A andB; Figure the Universityof Waterloo.SmallDIC samplesfrom the tills 1)in a leastdesirable hydrogeologic setting; negligiblewere analyzedfor •4C by TAMS at the Universityof groundwater movement and pore water association with Toronto. Radiocarbon activities of the DOC and D!C samnumerous Cretaceous age coal fragments.Also, it was plesare reportedin percentof moderncarbon(PMC)rela-

anticipated thatDOC radiocarbon datingcouldprovide tive to 95% the activityof oxalicacid standardin 1950.DOC

1978

WASSENAAR ET AL.' RADIOCARBON IN DISSOLVEDORGANICCARBON TABLE 1. RadiocarbonContent of HMW DOC in Three Shallow Aquifers in Ontario, Canada

SturgeonFalls

Depth

AllistonAquifer

Water table4, 1, 1 m

Aquifer

RodneyAquifer

96 PMC

101PMC

97-92PMC

aerobic high NO•-, SOft

02 reduction

0 2 reduction NO•- reduction SO• reduction

87 PMC

83 PMC

depth

8, 8, 2 m depth 16, 18 m depth

aerobic

SOl reduction

highNO•-, SO• 99 PMC

fermentation 75 PMC

fermentation 76 PMC

Allthree aquifers contain bomb tritium. The•4Cdatashow (1)theinitial14C content ofHMWDOC recharging the water tableand (2) the potentialeffectsof bacteriallymediatedoxidation-reduction

processes onthe •4Ccontent of HMW DOCalonggroundwater flowpaths.The Allistonaquiferhas no activeredoxprocesses occurringthatutilizeor produceDOC. In contrast,the SturgeonFalls and Rodneyaquifersarecharacterized by sequential anoxification, dentrification and/orsulfatereduction, andfermentation of organiccarbon[Wassenaar, 1990;Wassenaar et al., 1990a].The implications for

groundwater datingusing14Cin DOCarediscussed in thetext.

samples werenormalized to a (513C of-25%0.Radiocarbonbeds agesare calculatedusingthe equationfor radioactivedecay' 14

Age= -(t•/2/ln2) In a•4/ao

(3)

can be neglected. This information can be deduced indirectly by relative age comparisonswith other groundwa. ter dating methods. If the condition of conservancyis met,

thenthereduced14Cactivities of DOC alongtheflowpath wherea •4is themeasured •4Cactivityofthesample, ao TM is representradioactive decay alone and can be usedto detertheinitial•4Cactivityat t = 0 (100PMC)andti/2isthe•4C mine relative groundwater ages. However, if multipleDOG isotope half-life (5730 -+ 40). Analytical reproducibilityfor

sources exist along the flowpath (i.e., from buriedpeats,

PMC.

be developedfor DOC. As DOC is operationallydefined and extracted and represents a highly complex continuum 0f organic molecules, a dilution factor based on geochemical propertiesof DOC fractionswould be extremelydifficult.

a 14Cdilution factorQ (equation (1))would needt0 •4Cdeterminations of DOC by TAMS wasbetterthan-+0.8 kerogen)

The 8•3C values of the DOC fractions and DIC were determined on a VG Micromass 903 mass spectrometer at the Environmental Isotope Laboratory, University of Waterloo, and are reported relative to PDB, where -1

Second, the initial!4Ccontent(A0) of DOCrecharging the groundwatersystemmust be known. In contrasttothe

(4) DIC, whose•4Ccontentis initiallyderivedfromtheatm, sphere andsoilzoneCO2(100PMC),the 14Cactivity of andR is the ratioof 13C/12C in thesampleandtheinterna- DOC recharginga groundwater system reflectsa decom• tionalstandard. Analytical precision for •3Cdeterminationssitional mixture of modern and/or variable age organic (Ssample = [(Rsample )/(Rstd) ] X 1000

was betterthan -+0.05%o, •3C samplereproducibility was carbon sources in the soil and vadose zone. Therefore,the --+0.2%o. 14Cactivityof DOC at the initiationof theflowsystem may be notbe modern(