Timescales of hydrothermal scavenging in the South Pacific Ocean from 234 Th, 230 Th, and 228 Th

HAL Id: hal-02511018

https://hal.archives-ouvertes.fr/hal-02511018

Submitted on 18 Mar 2020

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

Timescales of hydrothermal scavenging in the South Pacific Ocean from 234 Th, 230 Th, and 228 Th

Frank Pavia, Robert Anderson, Erin Black, Lauren Kipp, Sebastian Vivancos, Martin Fleisher, Matthew Charette, Virginie Sanial, Willard Moore, Mikael

Hult, et al.

To cite this version:

Frank Pavia, Robert Anderson, Erin Black, Lauren Kipp, Sebastian Vivancos, et al.. Timescales

of hydrothermal scavenging in the South Pacific Ocean from 234 Th, 230 Th, and 228 Th. Earth

and Planetary Science Letters, Elsevier, 2019, 506, pp.146 - 156. �10.1016/j.epsl.2018.10.038�. �hal-

02511018�

Contents lists available atScienceDirect

Earth and Planetary Science Letters

www.elsevier.com/locate/epsl

Timescales of hydrothermal scavenging in the South Pacific Ocean from

Th,

Th, and

Th

FrankJ. Pavia^a,b,∗, RobertF. Anderson^a,b, ErinE. Black^c,d, LaurenE. Kipp^c,d,

Sebastian M. Vivancos^a,b,Martin Q. Fleisher^a,Matthew A. Charette^d,Virginie Sanial^e, Willard S. Moore^f,Mikael Hult^g,Yanbin Lu^h,Hai Cheng^h,i, Pu Zhang^h,

R. Lawrence Edwards^h

aLamont-DohertyEarthObservatoryofColumbiaUniversity,Palisades,NY,USA

bDepartmentofEarthandEnvironmentalSciences,ColumbiaUniversity,NewYork,NY,USA

cMassachusettsInstituteofTechnology/WoodsHoleOceanographicInstitutionJointPrograminOceanography/AppliedOceanScienceandEngineering,USA dDepartmentofMarineChemistryandGeochemistry,WoodsHoleOceanographicInstitution,WoodsHole,MA02543,USA

eDivisionofMarineScience,UniversityofSouthernMississippi,StennisSpaceCenter,MS39529,USA fDepartmentofEarthandOceanSciences,UniversityofSouthCarolina,Columbia,SC29208,USA gEuropeanCommission,JointResearchCentre,DirectorateforNuclearSafetyandSecurity,Geel,Belgium hDepartmentofEarthSciences,UniversityofMinnesota,Minneapolis,MN,USA

iInstituteofGlobalEnvironmentalChange,Xi’anJiaotongUniversity,Xi’an,China

a r t i c l e i n f o a b s t ra c t

Articlehistory:

Received3May2018

Receivedinrevisedform23October2018 Accepted26October2018

Availableonlinexxxx Editor:D.Vance Keywords:

thorium

hydrothermalactivity scavenging GEOTRACES EastPacificRise

Hydrothermalactivityinthedeepoceangeneratesplumesofmetal-richparticlescapableofremoving certaintraceelementsfromseawaterbyadsorptionandsedimentation.Thisremovalprocess,knownas scavenging,canbeprobedusingtheinsolubleradiogenicisotopesofthorium(Th),whichareproduced at a known rate in the water column via the decay of soluble uranium (²³⁴Th, ²³⁰Th) and radium (²²⁸Th) isotopes. We present dissolved andparticulate measurementsofthese threethoriumisotopes in a hydrothermal plume observed in the southeast Pacific Ocean onthe GEOTRACES GP16 section.

Since their half-lives varyfrom days (²³⁴Th) to years (²²⁸Th) to tens of thousands of years (²³⁰Th), the combinationoftheirsignalscanbeused tounderstandscavengingprocessesoccurringonawide rangeoftimescales.Scavengingisamulti-stepprocessinvolvingadsorptionanddesorptionontoparticles, followedbyparticleaggregation,sinking,andeventualsedimentation.Weusethoriumisotopestostudy howhydrothermalactivityaffectsthesesteps.Therateconstantsfornetadsorptionof²³⁴Thdetermined herearecomparabletopreviousestimatesfromhydrothermalplumesintheAtlanticandNorthPacific Oceans. Thepartitioning of²³⁴Thand²³⁰Thbetweenlarge and smallparticlesismore similar inthe hydrothermal plumethan aboveit, indicatingfasteraggregation of particleswithinthe hydrothermal plume atstations nearbythe EastPacificRise thaninwaters outsidetheplume. Inaddition torapid scavenging and aggregation near the ridge axis, we also infer continuous off-axis scavenging from observationsand modelingof²²⁸Th/²²⁸Ra activityratios. Thedegreeofdepletionofthethreethorium isotopesincreasesinorderofhalf-life,withtotal²³⁴Thactivityclosetothatofitsparent²³⁸U,but²³⁰Th showingnearly70%depletioncomparedtoexpectedvaluesfromreversiblescavenging.Bymodelingthe variationsindepletionforthedifferentisotopes,weshowthatmuchofthe²³⁰Thremovalisinherited fromscavengingeventshappeninglongbeforethemostrecenthydrothermalinputs.

©2018ElsevierB.V.Allrightsreserved.

1. Introduction

Submarine hydrothermal vents emit hot, reducing fluids that are highly enriched in trace metals such as iron (Fe) and man-

* Correspondingauthorat:Lamont-DohertyEarthObservatoryofColumbiaUni- versity,Palisades,NY,USA.

E-mailaddress:fpavia@ldeo.columbia.edu(F.J. Pavia).

ganese (Mn) relative to the deep ocean (German and Seyfried, 2014).Upon reactingwithseawater,dissolvedFe andMn precipi- tate toformparticles,withFe initiallyforming sulfidesclosest to theridgeaxis(e.g.Feelyetal.,1987),andbothFeandMnforming oxides overlonger distances(Feelyetal., 1996). Thesemetallifer- ousparticleshavehighlyreactivesurfacesthatcanreadilyremove trace metals,phosphorus, andcarbon fromsolution (Feely etal., https://doi.org/10.1016/j.epsl.2018.10.038

0012-821X/©^{2018ElsevierB.V.Allrightsreserved.}

Fleisher, M. Q., Charette, M. A., Sanial, V., Moore, W. S., Hult, M., Lu, Y., Cheng, H., Zhang, P. and Edwards, R. L.: Timescales of hydrothermal

scavenging in the South Pacific Ocean from 234Th, 230Th, and 228Th, Earth

and Planetary Science Letters, 506, 146–156, doi:10.1016/j.epsl.2018.10.038,

2019.

1990;Germanetal.,2015; Kadkoetal.,1994),makinghydrother- malsystemsimportantsinksfora widerangeofelements inthe ocean.

The south Pacific was identified as a major locality of hy- drothermal activity by the discovery of volcanic ³He emanating from the East Pacific Rise (EPR) at 2500 m between 15^◦–20^◦S (Lupton and Craig, 1981). In 2013, the GEOTRACES GP16 cruise followedthissouthPacific heliumplume west oftheEPR.In ad- ditionto³He,the plumewas foundtobe highly enrichedindis- solvedmetalslike Fe(dFe) andMn (dMn)over 4000 kmwestof theEPR (Resingetal., 2015).Enrichments inparticulateFe (pFe) andMn (pMn) were alsofoundinthe GP16hydrothermalplume (Fitzsimmonsetal., 2017). WhilepMnanddMnwereboth found tohavepeaks alongthe sameisopycnalas³He,the pFe anddFe peaksdescendedbelow the³He isopycnal, indicatingremoval by sinkingparticles,despitetheapparentconservativebehaviorofdFe previouslyinferredby Resingetal. (2015). Modelingstudieshave shownthat hydrothermalFe,stabilizedinsolutionandcarriedby abyssaloceancirculation,could upwellandsupport newprimary production,particularlyintheSouthernOcean(Resingetal.,2015;

Tagliabue et al., 2010). A better understanding of removal pro- cesses and their associated timescales in hydrothermal plumes couldpotentiallyimprovemodelsofFestabilizationandremoval.

Chemicalscavengingencapsulatesthesumofprocessesrespon- siblefortheremovalofionsfromtheocean byoceanicparticles.

Theseindividual processes include adsorption anddesorption re- actions,aggregationanddisaggregationofparticles,andeventually particlesettlingandsedimentationattheseafloor(BaconandAn- derson,1982).Netscavengingratesresponsiblefortheremovalof elementsfromthe ocean integrateacross therates oftheseindi- vidualprocesses.

The radiogenic isotopes of thorium (²³⁴Th, ²³⁰Th, ²²⁸Th) are powerfultoolsforunderstandingthe kineticsofoceanicscaveng- ingprocesses.Thesethoriumisotopesareproducedatwell-known ratesinseawaterby thedecay ofsolubleuranium(²³⁸U→^234Th,

234U→^{230Th)andradium(228Ra}→^{228Th).Thoriumishighlyinsol-} ubleinseawater,withascavengingresidencetimeontheorderof 10–40 yrinthedeepocean(HendersonandAnderson,2003).Tho- riumadsorbs onto particles that subsequently settle through the watercolumn, generating radioactivedisequilibrium withrespect totheirmoresolubleparents.TherapidremovalofThisapparent fromtheopen oceanactivity ratioof ²³⁰Th to its parent ²³⁴Uof

∼^{0.00002(MooreandSackett,1964).Withrespectivehalf-livesfor}

234Th, ²³⁰Th, and²²⁸Thof 24.1days (Knight andMacklin,1948), 75,587 yr(Chengetal.,2013),and1.91 yr(Kirbyetal.,1956),tho- riumisotopescanconstrainscavengingbehaviorintheoceanona rangeoftimescalesspanningmonthstomillennia.

Pavia et al. (2018) studied hydrothermal scavenging of ²³⁰Th and²³¹Pa inthe GP16 hydrothermal plume, finding large deple- tions in totaland dissolved ²³⁰Th and²³¹Pa coincidentwith en- richmentsin the particulate phase. The authors determined that this intense scavenging was largely the result of iron and man- ganesecoatingsonparticles,andthat scavengingwas continuous overthe4000 kmextentoftheplume.Inthisstudy,weusecom- binedmeasurementsofdissolvedandparticulate²³⁴Th,²³⁰Th,and 228Thandtheirparent activitiestostudythekineticsofthe indi- vidualprocessesinvolvedinscavenging, includingadsorptionand desorption, particle aggregation, andthe net scavengingremoval ofthoriumfromthewatercolumn. Weusetheseobservationsto assesstheimportanceoflocalanddistalhydrothermalactivityon 230Th scavengingpreviously observed in the GP16 hydrothermal plume,andtostudythe timescalesoverwhichdifferentscaveng- ingprocessesactinhydrothermalsettings.

Fig. 1.Sitemapofthestudyarea.NetvelocitiesofRAFOSfloatsdeployedalongthe EPR(HautalaandRiser,1993; LuptonandJenkins,2017) areshownassolidwhite arrows.Samplingsitesdiscussedinthispaperaretealdots.Locationsofactivehy- drothermalventsites(http://vents-data.interridge.org)areshownasblackXmarks.

DashedwhitearrowshowsproposedflowpathofwatersenteringtheGP16off-axis hydrothermalplumediscussedinSection4.3.

2. Materialsandmethods 2.1. Cruisesetting

SamplesweretakenonboardtheR.V.ThomasG.Thompsondur- ing the GEOTRACES GP16 cruise (TGT303) between Ecuador and Tahiti from 25 October to 20 December, 2013. We focus on the three sampling locations at or nearest downstream of the EPR:

station 18 atthe EPR and stations 20and 21, lessthan 250 km to the west along 15^◦S latitude (Fig. 1). Since the hydrothermal plume was interrupted by a discontinuity dueto mixingofnon- plumewatersatthenextstationtothewest(Jenkinsetal.,2018;

Lupton and Jenkins, 2017), we restrict our analysis to the three stationsnearesttotheEPR.

2.2. Samplecollectionandanalysis

Dissolved²³⁰Thandtotal²³⁴ThsamplesweretakenfromNiskin bottles, with²³⁰Thcollected frombottleson aconventional steel rosette and deep ²³⁴Th samples taken from bottles hung above in-situpumps. Dissolved²²⁸Thand ²²⁸Ra sampleswere collected bypumpingwaterfilteredat0.8 µmover in-lineQMAandSupor filters using McLane in-situ pumps, and then over MnO2 coated acrylic cartridges (Henderson et al., 2013; Maiti et al., 2015).

Particulatesamplesforall Thisotopeswere collected viabattery- operated McLane in-situ pumps in two size classes: 0.8–51 µm, the smallsize fraction(SSF) and>51 µm, thelarge size fraction (LSF). All data from the GEOTRACES GP16 section presented in thispaperareavailableinthe2017GEOTRACESIntermediateData Product (https://www.bodc.ac.uk/geotraces/data/idp2017/). Much of the data is also archived at the Biological and Chemical OceanographyData Management Office(BCO-DMO),including to- tal ²³⁴Th and ²³⁸U (https://www.bco-dmo.org/dataset/643213), particulate ²³⁴Th (https://www.bco-dmo.org/dataset/643316), dis- solved ²³⁰Th (https://www.bco-dmo.org/dataset/643639), particu- late ²³⁰Th (https://www.bco-dmo.org/dataset/676231), and ²²⁸Ra (https://www.bco-dmo.org/dataset/650340).

2.2.1. ²³⁴Th

The methodsfor analyzingGP16samples fortotal andpartic- ulate²³⁴Thhavebeensummarizedpreviously(Blacketal., 2018).

Briefly, 4 L samples foranalysis of total ²³⁴Th were spiked with 230Thasayieldmonitor,pre-concentratedbyco-precipitationwith MnO2, and collected on Whatman quartz microfiber (QMA) fil- ters. Particulate ²³⁴Th samples were taken in two size fractions

usingin-situpumps.The LSFparticles werecollected onto apre- filter,andrinsedontosilverfilters.SSFparticleswerefilteredonto QMA filters. The activity of ²³⁴Th for both total and particulate sampleswasdeterminedusinganti-coincidencebetacountersand correctedforbackgroundradioactivity.Tocomputetheradioactive disequilibrium of ²³⁴Th, ²³⁸U was predicted for each sample by theU-salinityrelationshipofOwens etal. (2011).Dissolved²³⁴Th is calculated as the difference betweenthe total and particulate pools.

2.2.2. ²³⁰Th

The ²³⁰Th data and methods in this paper have been previ- ouslypublished(Paviaetal.,2018).Seawatersamples(∼^{5 L)were} filtered over 0.45 µm Acropak capsule filters at sea and acidi- fiedto pH= ^{2 using}redistilled6M hydrochloricacidforstorage andon-shoreanalysis. Size-fractionatedparticulate²³⁰Th samples were taken using in-situ pumps, with LSF particles collected on a Sefar polyester mesh prefilter, and SSF particles collected on paired 0.8 µmSupor filters.Dissolved sampleswere spiked with 229Th,co-precipitated usingiron oxyhydroxide, then digested us- ing HF, HNO3, and HClO4. Particulate samples were spiked with 229Thanddissolved usingHNO3 andHClO4,followedbyiron co- precipitationandsubsequentredissolution.Forbothdissolvedand particulate samples,thorium isotopes were then separatedusing anionexchange chromatography(BioRad AG1-X8).Concentrations of²³⁰ThweredeterminedonaThermoELEMENTXRsinglecollec- torinductively-coupledplasmamassspectrometerinpeakjumping mode.Dissolved²³⁰Thdatapresentedherehavebeencorrectedfor detrital²³⁰Thpresentinthedissolvedpoolfromthedissolutionof continentalmaterial (e.g.Roy-Barman etal., 2009), andingrowth from²³⁴Udecayduring samplestorage.The ²³⁴Uactivityineach sample was computed by multiplying the ²³⁸U-salinity relation- ship(Owensetal.,2011) bytheoceanic²³⁴U/²³⁸Uactivityratioof 1.1468(Andersenetal.,2010).

2.2.3. ²²⁸Thand²²⁸Ra

Filteredseawaterwas pumpedover MnO2-coated cellulosefil- tersbyMcLanein-situpumpstocollectdissolved²²⁸Thand²²⁸Ra samples.Typically,1500–1700Lofseawaterwasfilteredatanav- erageflowrateof6.5Lmin⁻¹ withanaveragecartridgecollection efficiency of66±17%. Particulate samples were collected in the SSF on QMA filters via the same in-situ pumps. Full procedures fortheanalysisof²²⁸RaontheGP16sectionhavebeenpublished (Kipp et al., 2018b). The MnO2 cartridges and QMA filters were countedfor²²⁸Thvia²²⁰RnemanationwiththeRaDeCCalphade- layedcoincidencesystem(Charetteetal.,2015; Maitietal., 2015;

MooreandArnold,1996).Thismethoddetectsadsorbed²²⁸Thca- pable ofreleasing ²²⁰Rn; the resultingparticulate activityshould thereforebeconsidered alower limit,astheremaybeadditional 228Thin phasesthat trap its ²²⁰Rn frombeing released.Fordis- solvedsamples,afterRaDeCCanalysis,theMnO2-coatedcartridges wereashed andgammacountedfor²³⁴Thand²²⁶Raonhighpu- rity,well-typegermaniumdetectors.Thecartridge-based²³⁴Thand 226Rameasurementswere comparedtosmall-volume(∼4 L) beta countingmeasurements of²³⁴Thand²²⁶RacollectedfromNiskin bottleshung atthedepth ofeach sample,andthe ratioof ²³⁴Th and ²²⁶Rameasured on the small-volume samples to that mea- sured on the MnO2-coated cartridges was used to calculate the collectionefficiencyof²²⁸Thand²²⁸Raonthecartridges(Maitiet al.,2015).

228Raactivitiesweremeasuredongammadetectorslocatedun- dergroundattheLaboratoire SouterraindeModaneinFranceand the HADES laboratory in Belgium. The underground location of these laboratories serves to minimize the amount of cosmic ra- diationreachingthedetectors,reducingthedetectionlimits.

3. Results

At stations 18 and20there is a clearsignature of hydrother- mal scavengingobserved intheprofiles of all threethorium iso- topes below 2200 m (Fig. 2). The most distinct signal for the shorter lived ²³⁴Th and ²²⁸Th is enrichment in the particulate phase. In background,non-plume influenced deepwaters, partic- ulate²³⁴Thmakes up lessthan5% ofthe total²³⁴Th,withactiv- ities of1.4–1.8 mBq/kg.In thehydrothermalplume atstations18 and20,particulate²³⁴Threaches apeak of32% ofthetotal pool and consistently has activities of 9–12 mBq/kg. Particulate ²³⁰Th increases from background values near 2 µBq/kg to more than 4 µBq/kg,reaching57%ofthetotalpool.Particulate²²⁸Thincreases from0.1–0.5 µBq/kginnon-plumewatersto3.5–4.5 µBq/kgwithin the plume,peakingat36%ofthetotal ²²⁸Th.Atstation 21,there isa slightenrichmentinparticulate²³⁴Th to3.97 mBq/kgandin particulate²²⁸Thto1.76 µBq/kg.

There is little sign of significant excess in or deficit of total 234Threlativetoitsparent²³⁸U.Asteadystatemassbalancemodel of²³⁴Thcanbeusedtodeterminescavengingandremovalratesof 234Thinhydrothermalplumes(Kadko,1996).Themassbudgetfor dissolved²³⁴Thcanbewritten:

∂²³⁴Th_diss

∂t =!₂₃₈

U−^234Thdiss"

λ− ^JTh (1)

where²³⁸Uisthedissolved²³⁸Uinventory(Bq/m²)intheplume, defined here as the depth interval between 2200 and 3000 m, 234Thdiss isdissolved ²³⁴Th inventory (Bq/m²) in the plume, λ is the decay constant of ²³⁴Th in yr⁻¹ and J_Th is the net rate at which dissolved ²³⁴Th isadsorbed ontoparticles in Bq m⁻²yr⁻¹. Assumingsteadystate,wecansolvefor J_Th,thenthenetadsorp- tion rate constant of ²³⁴Th in the plume (k1, in units yr⁻¹) by dividingthescavengingratebythedissolved²³⁴Thinventory:

k₁=₂₃₄^J_Th^Th

diss (2)

At station18, we derive J_Th=106±16 Bq m⁻² day⁻¹ andk1= 2.99±0.47 yr⁻¹. At station20, we find J_Th=147±31 Bq m⁻² day⁻¹andk₁=².32±⁰.51 yr⁻¹.Atstation21, J_Thandk₁decrease to35.5±^{17 Bq m−2 day−1 and1.21}±^{0.58 yr−1} respectively.Our estimatesof J_Tharecomparabletopreviousvaluesof152 Bq m⁻² day⁻¹ foundbyOwens etal. (2015) attheMid-AtlanticRidgeand 81.6–202 Bq m⁻² day⁻¹ foundby Kadko etal. (1994) atthe Juan deFucaRidge.

Unlike ²³⁴Th, there is a large deficitof total ²³⁰Thin the hy- drothermal plume atstations18–21 relativeto theconcentration profiles expectedforremovalby reversible scavenging.Reversible scavengingshoulddrivedissolved, particulate,andtotal²³⁰Thac- tivitiesto linearlyincrease withdepth throughoutthe watercol- umn (BaconandAnderson,1982).Abovethehydrothermalplume this is the case along GP16 (Pavia et al., 2018); however, be- low 2200 mat stations 18–21, total ²³⁰Th sharplydeclines from 15–17 µBq/kg to lessthan 5 µBq/kg (Fig. 2). A similar signal is seen in the dissolved phase, andthe depletion ofboth total and dissolved ²³⁰Thinthehydrothermalplumeisroughly similar be- tween the three stations. Whereas the imprint of hydrothermal scavengingsignalson²³⁴Thessentiallydisappeared bystation 21, roughly 250 kmfromtheEPR,thehydrothermaldepletionofdis- solvedandtotal²³⁰Thandenrichmentofparticulate²³⁰Thextends over4000 kmfromtheridgeaxis(Paviaetal.,2018).

Fig. 2.Profilesofatstation18(toprow),station20(middlerow),andstation21(bottomrow)ofthoriumisotopesandtheirparentactivitiesfromtheGP16section.Dissolved 238U,andtotal,particulate,anddissolved²³⁴ThareshowninpanelsA,D,andG.Dissolved,particulate,andtotal²³⁰ThareshowninpanelsB,E,andH.Dissolved²²⁸Ra,and dissolved,particulate,andtotal²²⁸ThareshowninpanelsC,F,and I.Errorbarsrepresent1-sigmauncertainty,andaresmallerthanthesymbolsizewherenotvisible.

4. Discussion

4.1.Sizepartitioningof²³⁴Thand²³⁰Th

Theisotopesofthoriumareexpectedtohaveidenticalchemical scavengingbehavior.Differencesinobservedscavengingintensities ofthedifferentisotopescan thereforebe attributedto thediffer- enttimescalesoverwhichtheyintegrate,whicharerelatedtotheir half-lives.One ofthe keyvariablesindeterminingthe scavenging ratesand sinking fluxesof particles istheir size (Burdand Jack- son, 2009). Smaller particles typically have a larger surface area tovolume ratio,allowing forgreateradsorption per unit massof particles(HoneymanandSantschi,1989),whilelargerparticlesare exportedfasterby gravitationalsettling (BurdandJackson,2009).

Measuringthoriumisotopes withvaryinghalflivesinparticlesof differentsize classescan beused to constrainthe aggregationof

small particles, withlarge surface area to mass ratios,into large particlesthatsettlemorerapidly.

On the GP16 section, particulate ²³⁴Th measurements were madeonboththeLSFandSSFthroughoutthewatercolumn. Par- ticulate ²³⁰Th was mostly measured in theSSF, butthere are 44 LSFmeasurements,including17inthehydrothermalplumeatsta- tions 18–21that permit comparison with ²³⁴Th. ²²⁸Th was only measured inthe SSF,so we canonly comparethe sizepartition- ingof²³⁴Thand²³⁰Th.Todothis,weusethefractionofthorium activityfoundinthelargesizefractionrelativetothetotalpartic- ulateactivity(Leeetal.,2018):

fLSF= [^pTh]LSF

([^pTh]LSF+ [^pTh]SSF) ⁽³⁾

where [pTh] is the particulate activity of thorium. For the data availableoutsidethehydrothermalplume,aslightlyhigherfraction