Volcanic layers in Antarctic (Vostok) ice cores: Source identification and atmospheric implications Isabelle Basile-Doelsch, Jr Petit, S Touron, Fe Grousset, N Barkov
To cite this version:
Isabelle Basile-Doelsch, Jr Petit, S Touron, Fe Grousset, N Barkov. Volcanic layers in Antarc- tic (Vostok) ice cores: Source identification and atmospheric implications. Journal of Geophys- ical Research: Atmospheres, American Geophysical Union, 2001, 106 (D23), pp.31915-31931. 10.1029/2000JD000102. hal-00726370
HAL Id: hal-00726370 https://hal.archives-ouvertes.fr/hal-00726370 Submitted on 25 Jan 2021
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 106, NO. D23, PAGES 31,915-31,931,DECEMBER 16, 2001 Volcanic layers in Antarctic (Vostok) ice cores: Source identification and atmospheric implications
IsabelleBasile, 1'2 Jean Robert Petit, 2 StephanieTouron 3, Francis E. Grousset,4 and NartsissBarkov s
Abstract. Fifteen visible volcanicash layers (tephra) from Vostok ice coreshave been analyzedfor major elements,trace elements, and Sr andNd isotopecomposition. Comparisonof their geochemicalsignatures to lava compositionfrom the inventoryof Antarctic and subantarcticvolcanoes, which have been active over the last 0.5 million years,indicates that nine layersoriginate from activity of the SouthSandwich volcanic arc, three from southernSouth America, one from the AntarcticPeninsula (Bransfield Strait), and one from West Antarctica(Marie Byrd Land province).The large size of the tephra(up to 50 gm) requiresrapid atmospherictransfer from the volcaniccenters to East Antarctica. Rapid tropospherictransport from the southwesternAtlantic, penetratingEast Antarctica, therefore predominates during the period studied,whether in glacialor interglacialclimatic mode. In spiteof the low frequencyof occurrenceof visible tephralayers in Vostok core (one eventevery 20 kyr), the overall atmospheric pathwayof theseash events appears consistent with the almostcontinuous advection of continental dust from South America.
1. Introduction However, the dating of ice cores is not always a straightforwardtask. This is particularlytrue for sites The reconstruction of past atmospheric circulation with low snow accumulation such as those in East patterns as well as the investigation of long-term links Antarctica where the absence of seasonal variations of between volcanic aerosolsand climatic change are today proxies prevents the detection of annual layers. subjects of major interest and considerable debate. Moreover.deep ice corescover time periodsof up to half Representing potential archives of both climate and a naillionyears, and the tephrochronologybeyond the last volcanic events, ice cores can be an important sourceof 2000 years is still poorly documented.Unlike sulfuric data for suchstudies. Although ash layersin ice coresdo acid events, the tephra layers can potentiallyprovide not always provide information on the climatic impact of informationon sourcelocation through the "fingerprints" volcanism, they do provide a record of atmospheric left by their elementaland trace composition. This makes loadingand thc pathwaystaken by volcanicclouds. tla½study of tcphralayers of particularinterest for the Past volcanic activity is recorded as visible volcanic documentationof eruptivevolcanic events in the middle ash (tephra) layers and/or sulfate-rich layers. Sulfuric to late Pleistoceneperiod. The amount of available acid can be detectcdby an electrical method [Hammer, materialin ice is alwaysvery small, seldomexceeding a 1977] or by direct chemistry[Zielinski et al., 1994, 1996, few milligrams. This restrictsanalytic methodsand 1997]. The sulfuric acid record is an indicator of a large- prevents the application of dating inethods (e.g., scale atmosphericloading, and particular events, which 3•;Ar/4øAr).However, an inventory of thetephra in ice can are documentedeither by the observationsor geological provideevent markersthat are extremelyuseful when studies, are of interest because they could help in investigating the stratigraphic correlation between establishingthe ice core chronology. Conversely, the different ice cores[Fz.tji et al., 1999] or betweenice cores dating of ice cores by independentmethods (e.g., by and marine sedimentrecords [Zielinski et al., 1997]. In varve or layer countingas for the Greenlandice records) addition,it shouldbe possibleto link the tephrawith the can improve the documentation of the historical volcanic source if the amount of lava materials is observations[Zielinski et al., 1994]. sufficientfor absolutedating methods. For the southpolar regions,Smellie [1999] providesa reviewof upperCenozoic tephra records. Focusing on the •Ce•tre Europdende Rechercheet d'Enseignementen Antarctic ice records, volcanic ash layers from several G•osciencesde l'Environnement,Europ61e M•diterranden de l'Arbois, Aix-en-Provence, France. coreshave been alreadyexamined. In the Byrd ice core, :Laboratoire de Glaciologie et Gdophysiquede in West Antarctica (Figure 1), Go•v and Williamson l'Environnement, St. Martin d'H•res, France. [1971] identifiedahnost 2000 visible ashlayers, most of •Laboratoirede Gdologiedes ChainesAlpines, Institut them concentratedin ice correspondingto the last glacial Dolomieu, Grenoble, France. period. The vicinity of the Marie Byrd Land volcanic 4Ddpartementde Gdologieet Ocdanographie,Univ. of Bordeaux I, Talence, France. provinceand the possibletriggering of eruptiveactivity >Arcticand AntarcticResearch Institute, St. Petersburg, from bedrockisostatic adjustment consecutive to the ice Russia. sheet thickening have been suggestedto explain the abundanceof volcanic layers at this time [Kyle et al., Copyright2001 by the AmericanGeophysical Union. 1981; Pa/aix, 1985]. Among volcanoes,Mount Takahe, Papernumber 2000JD000102. located 450 km away t¾omthe drilling site, has been 0148-0227/01/2000JD000102509.00 proposedas a possiblesource [Kyle and •lezek, 1978;
31,9!5 31,916 BASILEET AL.: VOLCANICLAYERS OF VOSTOKCORE
SOUTHSANDWICHS ISLANDS 13 islands:(Pearce et al., 1995) (Bakeretal., 1989) OCEANIC PLATE ProtectorShoal, Zavodovski, Leskov, Visikoi,Candlemas, Vindication, (P.type MORB and OIB) Saunders,Montagu, Bristol, Freezland,• Verwoerd et al., 1989 SOUTH AMERICA Bellinghausen,Cook,Thule Bo: Bouvet;PE: Prince ii i . Edward;Ma: Madon;Cr: (Hawkesworthet al., 1979) Crozet;Ke: Kerguelen; (Futaand Stern, 1988) McD: Mc Donald; He: Heard;StP: St Paul; NSVZ = Northern South Am: Amsterdam; Volcanic Zone An: Anfipode;Pe: Peter SSVZ = Southern South Volcanic Zone AVZ = Antarctic Volcanic Zone
60'W 60'E
/ .Am
I NAZGA
PLATE 90'W } 90'E
BRANSFIELD STRAIT / (Smellie,1989) (Gonzals-Ferrand, 1981) / Br: Bridgeman;Pe: Penguin; Li: Livingston;De: Deception / /
120'E 120'W xx \ -,,
JAMES ROSS PROVINCE
.? (Smellie,1989) Pa' Paulet 150'E SN' Seal Nunatak
LEGEND MARIEBYRD LAND !80' VICTORIA LAND Ky•e,1989a, Rockoil, 1995 '•,• Mid-oceanridge ErebusProvince: (Kyle, 1989c) (LeMasuder etal., 1989) .. Er: Mt Erebus; Inferredboundary of (Mcintoshand Wilch, 1997) NEW ZEALAND the West Antarctic RSR: RoyalSociety Range (Rowley,1989) Rift System MelbourneProvince: Kyle, 1989b Hu: Hudson;To :Toney; (Ewartand Stipp, 1968) Me: Mt Melbourne;PI :The Pleiades Subduction Ta:Takahe; Wa: (Shaneetal., 1996) Waesche;Si: Siple; HP: HalletProvince (Mcintosh and Kyle, 1989) Extension of iCe shelves Mo: Moulton;Be: Berlin TVZ ßTaupo Volcanic Zone Ba: BallenyIslands (Wright and Kyle, 1989)
i WRhinplate volcano I Subductionvolcano 0 1000 2000 3000 4000 km Icecore site SP' Southpole; DC' DomeC; V' Vostok;B' Byrd. I • ,i • ,I BASILE ET AL.' VOLCANIC LAYERS OF VOSTOK CORE 31,917
Ky/e el al., 1981; Palais, 1985]. In the Dome C ice core sequenceever obtained. Climate is depicted by the (East Antarctica), only one tephra was found. It was at a isotope of the water and taken as a proxy of the depth of 726 m and has been identified as coming from temperature(Figure 2). Ice chronologywas obtainedby Mount Takahe [Kyle et a/.,1981]. In shallow ice cores developmentof a glaciologicalmodel, and the ageof the fi'om Vostok and South Pole, Kyle et al. [1984] and ice layers is estimatedto within -5% accuracy[Petit et Pa/ais el a/. [1987] studied tephra layers by SEM-EDX a/., 1999]. and electron microprobemethods and comparedtheir In this paperwe presentresults of the analysisof 15 results with the major element compositionof nearby levels to which we applied an extendedgeochemical sources. The authors suggested a common source tracer method (major elements,trace elements,and Sr (Candelmas Island in the South Sandwich Islands) for and Nd isotopiccompositions) for comparisonwith the one event at 3200 yearsB.P. (before present),providing a compositionof volcanicmaterial of the source.First of visible tephrahorizon for both sites.Therefore the tephra all, the elemental composition and the geochemical suggestan initial stratigraphiccorrelation between two signaturesof glassshards are usedto determinethe layers Antarctic ice cores -1200 km apart and-5000 km from exhibitinga similartectonic setting with respectto their the direct source. parental magma (subduction zone or within-plate In a deep Vostok ice core (called "3G") covering the volcanoes).Next, we establishan inventoryof volcanoes last 160 kyr, Palais et al. [1989] detected numerous surroundingAntarctica, located from 90øS up to 30øS cloudy and faint tephra layers after careful inspectionof latitude, which have been active at some time over the the core. The amount of available material was low, last 0.5 naillionyears. All thesevolcanoes are considered making quantitative analysis extremely difficult. As in as PSVs.This representsnumerous volcanoes (see Figure previousstudies, the sourcewas identifiedfor five layers 1) set within subductionor within-plate(continental or by the major element composition.However, from a oceanic)provinces. We then comparethe geochemical theoreticalpoint of view, this method has a number of signaturesof the tephrawith availablegeochemical data limitations for source identification: (1) the major for PSVs. We are aware that slight geochemical element compositionof tephra can be heterogeneous differencesmay resultfrom the useof differentanalytical within a given layer; (2), different volcanoes and/or techniques[e.g., Kyle et al, 1984] or from different different volcanic provincesmay produce tephra with eruption characteristics,i.e., lava flow or pyroclastic. similar major element composition;(3) conversely,for a Fortunately,they are not significantcompared to the •1 ¾•11 ¾kJl•l, llU• tilL., L.,L/IIIIJL//DILIL/11 111Cl,.y L'11CI,11•; 11•Jlll UII• geochemicaldifferences that exist between the two eruption to another or even within a given eruption if' tectonicsettings, which helpsus rapidlyanalyze the PSVs there is a zonedmagma chamber. andpredict the sourceof the tephra.We will thendiscuss To mitigate these problems, a more complete some of the implications in terms of atmospheric geochemicalidentity card can be establishedusing the pathwaysof air massesfor glacial and interglacial analysisof trace elementsand the strontium-neodymium periods. isotopic signature. This provides complementary constraintsto thosefi'om the major element composition 2. Samples and Analytical Methods and the comparison with sources. A matching source compositionwill, however, also be dependenton the data Ash layers were recovered during successive available fi'om the literatureand from the lava samples expeditionsat Vostok Stationfrom a shallow ice core which have been analyzed and published for each named "BHI," and from two deep ice cores named potentialsource volcano (hereinafter PSV). "4G2" and "5G" (Figure2). The 15 tephralayers studied Vostok cores contain a total of-20 levels of visible here(Table 1) were detectedvisually. The layersare from tephrawith significantamounts of material. These layers a few to 20 mm thick with a variable whitish, brownish, are located ahnost randomly in ice from the interglacial or grayishcolor. The total amountof materialis about1 period or glacial period (Figure 2). This stronglydiffers to 5 mg. The selected ice samples were first fi'om the Byrd core, howeverVostok is locatedin central decontaminatedby washing with deionizedwater in a East Antarctica, relatively far from volcanic centers.The dust-freeclean room. The inner part of the corewas then Vostok ice core provides a climatic record of the last melted, and two aliquots of the water were filtered 400,000 years [Petit et al., 1999], the longest climatic throughtwo Nucleopore©filters (13 mm diameterand
Figure 1. Locationin the SouthernHemisphere of volcanicprovinces mad PSVs in theirtectonic settings (activitydating back less than 500 kyr). The provincesand the mainreferences on tectonicsettings and datingare McMurdo volcanic group [Kyle, 1989a, 1989b and 1989c;?er•voed et al., 1989;Rocholl et al, 1995];Marie Byrd Landprovince (MBL) [LeMasurier,1989, Rowley et al., 1989];James Ross Islands volcmaicgroup and Bransfield strait volcanoes [Smellie, 1989; Tokarski,1991; Gonzales-Ferran, 1991 ]; SouthSm•dwich Islands (SSIs) [Bakeret al., 1989;Pearce et al., 1995];New Zealand[Ewart and Stipp, 1968;Sirekin and Siebert,1994; Shane et al., 1996]; SouthAmerica [Hawkesworth et al., 1979;Futa and Stern,1988]. Additional references are givenin the discussion(see text). Adapted from LeMasurier and Thomson[1989]. The locationsof Antarcticdrilling sitesare also shown.V, Vostok;PS, Admunsen- Scott SouthPole; B, Byrd; DC, Dome C. 31,918 BASILE ET AL.' VOLCANIC LAYERS OF VOSTOK CORE
-480 Deuterium-460 5D-440 O/oo -420 VostokCores
0- Age EGT4 yr 100 SSI 'P-- 104-BH1 3,500
200 ? -- 181-5G 6,870 I I 3OO
4OO i
5OO ssI •----- 547-4G235,170 6OO
7OO
8OO
9OO I
1000 BS 1 lOO I
1200 b-- 1280-4G2 90,600 1300 SA and SSI 1400 SA •-- 1431-4G2 101,840 1500
1600
1700 1800
1900
2000 1992-5G M•L • 1981-4G21996-5G 141,500 2•oo• SSI 2169-4G2 160,700 2200 • SSI i____ 2231-4G2 2300 SSI r-- 2260-5G1 170,770 SSI • _1 2326-4G2 179,330 2400 2500 SSI t.... 2502-4G2201,600 I 2600 SSVZ h -- 2587-5G1 213,700 2700 [
Figure2. Positionof the visibleash layers in the Vostokcore studied, recorded along with the isotopic composition(deuterium) of theice andthe age of the layer[Petit et al., 1999].The identifiedsources of ash layers are SSIs (South Sandwich Island), BS (Bransfield Strait), SA (South America), SSVZ (southernpart of southernvolcanic zone), MBL (Marie Byrd Land).
0.4 pm porosity)fitted for microscopeobservations. For microlitesof feldsparor ferroma•nesianminerals which eachlayer, one filter was embeddedwith epoxyresin and may or may not have been includedin the glassyphase polishedusing diamond pastes(grain size 6, 3, 1, and (Figure 3c). 0.25 pm) to obtain flat surfaces for quantitative The quantitativemajor elementcomposition of glass microprobeanalysis. shardswas obtainedusing a wavelength-dispersiveX-ray Tephra were observedwith an optical microscopein (WDX) electron microprobe.This method is more transmission]node (with naturaland polarizedlight) and accurate and precise and was preferred to scanning in reflection mode. Additional observations were electron microprobe (SEM) energy dispersive X-ray performedwith a scanningelectron microscope. The size spectrometry[Palais et al., 1987]. Analyses were of the tephra varies from a few micrometerup to 50 pm performedon particlesgreater than 5 pm becauseof the (see Table 1). Under the optical microscope,shards limitationof the requiredminimum for analyticalvolume. appearbrown, orange,or colorless.All the observations To preventthe lossof Na underthe beam [Nielsenand indicatedthat the tephra layers are a mixture of three $igurdsson,1981], we set the beam currentand the differenttypes of particles:(1) glassshards (70 to 90%) acceleratingvoltage to 6 nA and 15 kV respectively.The resultingfrom the quenchingof eruptedmagma (Figure numberof analyzedparticles varies from 7 to 33 for each 3b), (2) lithie fragments (5 to 30%) composedof layer.Results are presented in Table2. cryptocrystallineglass (i.e., glass partly recrystallized) Trace elementswere analyzed using an inductively probably from prior eruptions (Figure 3a), and (3) coupled plasma mass spectrometer(ICP-MS- VG BASILE ET AL.' VOLCANIC LAYERS OF VOSTOK CORE 31,919
Table 1. AsiaLayers Sampled From Different Vostok Ice The three samples illustrate different degrees of Cores:BH1,4G2 and5G a homogeneityinside a single tephra layer (Figure 4). The Maximal 2169 m layer is composed of basaltic andesite glass Number Thickness Depth Age EGT4 Number Size of shards and appearsvery homogeneous.By comparison of of Layer (m) (years) of Core Ash the 2260 m layer rangesfrom andesiteto dacite: SiO2 and Drilling (mm) MgO varying from 60 to 75% and from 3 to 0.3%, 103.14 3 450 104 BH1 30 50 respectively. This straight-line trend probably results 180.24 6 870 181 5G 15 20 fi'om fractional crystallization and the presence of a 546.24 32 170 547 4G2 20 40 zoned lnagma chamberprior to the eruption [Zielinski et 988.45 68 190 989 5G 19 40 a/., 1995]. The third example is the 1280 m layer which 1279.16 90 600 1 280 4G2 2 20-50 showstwo distinct coinpositionsfeaturing dacite (K20 1430.9 101 840 1 431 4G2 5 10 0.51%, noted 1280a) and rhyolite (K20 • 4.5%, noted 1980.37 140040 1 981 4G2 7 10 1280b), respectively.In addition,the size distributionof 1991.93 141 330 1 992 5G 2 50 the shards is bimodal, showing one mode at 20-40 lam 1995.23 141 690 1 996 5G 20 50 and anotherat 5-20 lam.This indicatesthat 1280 m tephra 2168.43 160 650 2 169 4G2 2 30 represents,in fact, two independenteruptive events, and 2230.29 168 180 2 231 4G2 20 40 the mixed volcaniccloud deposited tephra simultaneously 2259.02 171 390 2 260 4G2 2 40 at Vostok. This layer was divided into two subgroups: 1280a mad 1280b. 2z;25•4 179 250 2 326 4G2 5 40 2501.92 201 970 2 502 4G2 12 50 2586.15 213700 2587 5G1 8 20 3.2. Major Element Compositions and Classification of Tephra Layers "Age "GT4" is fi'omPetit et al. [1999].Ash layer label refersto the bottomdepth of the 1 m longice coreincrement The major element compositionsare given in Table 2 containingthe layer.Size of tephrawas measuredusing an and represent for each layer the average value from opticalmicroscope. multiple analyses. In Figure 5a we plot the mean composinon of alkalis (Na20+K20) versus SiO2 to PlasmaquadPQ 2 Turbo©) with a standardquantitative classifythe differenttephra layers accordingly. Four main method. The tephra sampleswere first dried (-5 mg of groupscould be identified accordingto the increaseof the total alkali content. sample obtained) then dissolved with a mixture of Tefion©-distilled acids [HF + HC104+ HNO3] [Falkner 1. Set I containstwo layers (1280a, and 2326 m) and is et al., 1995]. For five samples,the amount of available composedof dacite with a very low alkali content (K20 material was too small for accurate trace element • 0.5%). They belongto a low-K tholelite series. analysis.Results are presentedin Table 2. 2. Set 2 contains nine layers (104, 181, 547, 1981, Sr and Nd isotopiccompositions were measuredusing 1992, 2169, 2231,2260 and 2502 m) and is also tholeiitic a multicollector thermo-ionisation mass spectrometer n•agma but slightly more alkaline (Na20+K20 from 2 to (TIMS) (Finnigan MAT 261 ©). After dissolution,Sr and 4'% ) than set 1. Nd were chemically separated using ionic 3. Set 3 contains four layers (989, 2587, 1280b and chromatographiccolumns. We followed chemical and 1431 m) and representscalc-alkaline magma. However, massspectrometer procedures as previouslydescribed by their SiO: content is significantly different (•55% and Groussete! al. [1988]. Resultsare presentedin Table 2. •70%), and we thereforedivide them into two subgroups: Forconvenience, measured (14•Nd/•44Nd)Meas ratioswere set 3a with two layers (989 and 2587 m) and set 3b with normalized with respect to a standard of CHondritic two layers (1280b and 1431 m). The high TiO2 content Uniform Reservoir (CHUR) with a value of 0.512636 (TiO2 • 2.5%) in the 989 m layer (Table 2) could be a [.Jacobsene! 14/asserburg,1980] usingthe relationship signatureof tholeiitic or alkaline within-plate volcanism l•Nd(O)=((•4•Nd/•44Nd)Meas/( 14•'Nd/ 144 Nd)cHUR- 1)xl 04. [Albarade, 1992] (see below). 4. Set 4 containsone layer (1996 m) with a composition much different fi'om that of the other layers. It is a 3. Tephra Composition and Groupings trachytefrom an alkaline series(K20+Na20 •10%). 3.1. Variability of the Major Element Composition Tinis classification gives an initial indication of the tectonic setting of the volcanic sources'tholeiitic and Microprobe analyseswere performed only on glass calc-alkaline series typically come from subduction- shardsbecause they representthe magma at the time of related provinces, xvhereasthe alkaline series probably the eruption. To point out the variability in glass shard come from within-plate tectonicsources. compositionfrom a given layer, we represent,in Figure 4, the compositionof several single shards from three differentdepths (1280, 2169 and 2260 m) on a plot of the 3.3. Complementary Discrimination Using Trace concentrationof K20 and of MgO versusSiO2. The other Element Composition major elements are not representedbecause they show The formation of a magma follows three main steps: similar patternsand lead to the same conclusions.For the (1) the partial fusion of mantle rocks, (2) the collection discussion, we use the classification of volcanic rocks randmigration of the magma, and (3) the storageand based on alkaline elements versus SiO2 concentrations fractional crystallization.Different additional processes (Figure 5a) as suggestedby Cox e! al. [1979] and adapted may occur that modify the compositionduring magma by Le Maitre et al. [1989]. genesis. Magma may be mixed and contaminatedby 31,920 BASILE ET AL.: VOLCANIC LAYERS OF VOSTOK CORE
Figure 3. Microscopeobservations: (a) scanningelectron microscope picture of !ithic material composedof cryptocrysta!!ineglass; (b) scanningelectron microscope picture of glassyshard with a blockymorphology typical of phreatomagmaticashgrains [Heiken a,d Wohletz,1985]; (c) picture of the polishedsurface of a glassyshard observed with an optical microscope using transmitted light; microlites of p•'oxcnesappear in a spherolitestructure in theglassy matrix.
variouscrustal rocks with differentcompositions of availablematerial was unfortunatelytoo smallfor ICPMS continentalor oceanicorigin. Trace elementbehavior is or TIMS analysis. very sensitiveto all thesepetrogenetic processes. This Amongthe nine layersof set2, sevenlayers (104, 547, makesit possiblefor the origin of a magmato be 1992, 2169, 2231, 2260 and 2502 m) differ from the characterizedaccording to thetectonic setting (tholeiitic, 1981m layer with respectto the REE profile. Their calcaikalineor alkalineseries). Moreover, in a given profile is enriched by 6 to 17 times the chondritic tectonicsetting, each volcanicarea inheritsspecific abundance(Figure 5b) and a slight LREE depletion characteristicsof trace elements. Among trace elements, occurs(La/Gd = 0.8) (Figure5c). They alsohave a high rare earthelement (REE) enrichmentwith respectto a Zr/Nb ratio (62-76) (Figure 5c) and a Nd isotopic standard(normalized chondrite composition), negative composition(l•Nd(0)) from+6 to +8 aswell asa 87Sr/86Sr Europiumanomaly, light REE (LRRE) enrichmentor ratio from 0.70393 to 0.70478 (Figure5d), indicatinga depletiongiven by theratio of La to Gd, to Smor to Tb, ratherradiogenic signature. These data in additionto the Nb to Zr ratio,as well as Sr andNd isotopicsignatures tholeiitic composition indicate a typicalparental magma are the classical proxies used to characterize and from oceanic subduction zones. For convenience these discriminatebetween magmas and volcanic rocks sevenlayers will be called set 2* or "depletedLREE" [ Wilson, 1989]. We apply this approachto the layers hereinafter. compositionof theVostok tephra. The REEprofiles, the Set 3a includesthe 989 and 2587 m layers.The REE La/Gdversus Zr/Nb ratio and the isotopic composition of concentrationis 12 to 35 times higher than the chondritic set2, 3a and4 areshown in Figure5b, c, d. For set I and abundance(Figure 5b) with a slight LREE enrichment set 3b and $e 181 rn layer of set 2, the amountof (La/Gd-1.9) (Figure 5c). They have an intermediate BASILE ET AL.: VOLCANIC LAYERS OF VOSTOK CORE 31,921
Table 2. AnalyticalResults of Major andTrace Elements
Depth- 104-BH1 181-5G 547-4G2 989-5G 1280a-4G2 1280b-4G2 1431-4G2 1981-4G2 Drilling Number of 23 SD 14 SD 16 SD 31 SD 14 $D 33 $D 21 SD 13 SD Analyses giO2 60.23 1.52 53.16 1.05 60.90 1.20 55.50 1.93 64.62 1.15 66.62 0.96 66.30 0.88 56.12 0.95 TiO2 0.83 0.23 1.52 0.98 1.27 0.22 2.45 1.17 0.78 0.12 0.59 0.08 0.52 0.08 1.13 0.38 A120• 15.62 0.69 14.59 0.79 14.66 0.97 17.08 2.02 14.74 0.69 14.71 0.64 14.68 0.31 15.01 0.46 FeO 9.31 0.79 11.68 1.64 9.50 1.25 8.09 1.74 8.55 0.99 9.11 0.44 8.88 0.49 10.22 0.52 MnO 0.22 0.11 0.24 0.08 0.28 0.24 0.15 0.10 0.20 0.09 0.34 0.09 0.34 0.09 0.23 0.09 MgO 2.51 0.42 5.41 1.12 2.30 0.57 3.13 0.86 1.95 0.25 0.08 0.05 0.01 0.01 4.51 0.47 CaO 6.97 0.52 10.11 1.35 6.44 0.46 7.49 1.25 6.29 0.38 1.50 0.35 1.70 0.41 8.87 0.31 Na20 3.42 0.35 2.36 0.27 3.29 0.73 4.78 0.62 2.20 0.45 2.34 0.27 2.68 0.42 2.79 0.28 K20 0.49 0.10 0.47 0.19 0.86 0.11 0.67 0.26 0.47 0.07 4.46 0.24 4.76 0.19 0.69 0.08 P20• 0.15 0.14 0.21 0.20 0.14 0.12 0.42 0.22 0.09 0.06 0.06 0.05 0.06 0.04 0.25 0.15 SOs 0.05 0.07 0.11 0.14 0.15 0.11 0.11 0.10 0.10 0.23 CI 0.20 0.07 0.13 0.27 0.21 0.07 0.12 0.17 0.21 0.01 0.25 0.11 0.08 0.05 Total 100 100 100 100 100 100 100 100
Rb 11.3 24.3 113.4 Sr 142.7 356.8 244.5 Zr 49.5 204.3 47.4 Nb 0.7 5.9 43.0 Ba 91.9 168.7 102.1 472.0 La 2.4 3.8 11.2 97.1 Ce 7.6 11.2 27.1 198.0 Pr 1.3 1.9 3.8 23.1 Nd 7.2 8.6 22.5 92.1 Sm 2.6 2.4 5.3 18.0 Eu 1.0 0.8 1.6 3.5 Gd 3.3 2.9 5.5 16.3 Tb 0.6 0.5 1.0 2.8 I)y 4 3 3.5 5.1 16.4 Ho I 0 0.7 1.1 3.5 gr 2 9 2.3 3.2 10.2 Tm 0 4 0.4 0.4 1.4 Yb 2 8 2.9 2.9 10.0 I Ju 0 5 0.4 0.4 1.4 '•7Sr/'$•'Sr 0.704776 35 0.703932 22 0.703412 19 0.703304 20 143Nd/144Nd0.513046 12 0.513000 26 0.513003 13 0.512882 13 aNd(O) 8.00 7.10 7.16 4.80 '• Major elelnents(% weight)measured with an electronmicroprobe. Results are normalizedto 100%.Trace elements (in ppm) aremeasured by ICPMS.Isotopic compositions are measured by TIMS. Erroris 2cyx 10-'. Measured87 Sr/86 'St and 14 •Nd/ 86 88 14{• 144 ratioshave beencon'ected tbr massfractionation by normalizingto 'Sr/ Sr=0.1194 and 'Nd/ Nd=0.7219, respectively.Blanks 87 86 were negligible.Our measurementsfor the strontiumstandard (NIST SRM 987) yieldedan averagevalue tbr Sr/ 'Sr equal to 0.710205 {+/-0.00002) for 6 measurements,compared to the certifiedvalue of 0.710245. The neodymiumstandard (La Jolla) was analyzed5 timesand gave an averagevalue for 14•Nd/144Ndequal to 0.511846(+/-0.000015) compared to thecertified value of 0.5 ! 1865.The observeddifferences between our dataand the certifiedvalues are within the rangeof the measurementuncertainty.
Zr/Nb ratio (•35, Figure5c), but they differ by their Nd REE profiles show an enrichment30 to 300 times the and Sr isotopic compositions,which however are both chondritic abundance (Figure 5b) and a strong LREE includedin the array of the mantlecomposition (Figure enrichment(La/Gd = 5.3) (Figure 5c). They both display 5d). Thesecharacteristics suggest a parentalmagma from a negative Europium anomaly, a low (<7) Zr/Nb ratio, a subductionsetting, althoughthe high TiO2 content m•d similar [;Nd(0) signatures(4.76 and 4.97). Their (TiO: • 2.5%, Table 2) of the 989 m layeris unusualfor a 87Sr/86Srcompositions aredifferent, and the 1996 m layer subductionmagma. These compositions possibly suggest with a value of 0.70517 is more radiogenic than the a morecomplex tectonic environment[Albarbde, 1992]. 1981 m layer with 0.70330 (Figure 5d). We concludefor Althoughthe 1981 m layer from set 2 and the 1996 m the parentalmagma that the two layers originatefrom a layerfrom set4 belongto differentmajor elementtrends, magma with a tectonicsetting of continentalwithin-plate they bothdisplay a similarpattern for traceelements. The volcanism,but for the 1981 m layer, there is a setting 31,922 BASILE ET AL.' VOLCANIC LAYERS OF VOSTOK CORE
Table 2. (Continued)
Depth- 1992-5G 1996-5G 2169-4G2 2231-4G2 2260-4G2 2326-4G2 4G22502- 2587-5G1 Drilling Number of 26 SD 17 SD 20 SD 15 SD 20 SD 10 SD 16 SD 14 SD Analyses SiO2 54.12 0.86 63.68 1.15 55.52 0.47 56.67 0.95 66.26 5.76 65.59 0.46 62.59 0.71 56.53 0.81 TiO2 0.88 0.29 0.41 0.16 1.23 0.07 1.29 0.11 0.74 0.27 0.91 0.07 0.74 0.30 1.43 0.47 A1203 15.66 0.84 14.40 1.04 15.14 0.21 15.01 0.30 14.16 0.87 14.30 0.22 15.40 0.90 16.39 1.14 FeO 9.95 1.05 6.85 1.03 10.57 0.46 12.01 0.48 8.25 2.30 8.67 0.32 8.07 0.83 8.90 1.67 MnO 0.21 0.09 0.45 0.45 0.20 0.07 0.24 0.11 0.26 0.12 0.23 0.06 0.19 0.06 0.19 0.08 MgO 5.48 0.98 0.03 0.04 4.86 0.19 3.49 0.22 1.75 0.74 1.65 0.13 1.90 0.36 3.49 0.40 CaO 10.17 0.57 1.08 0.32 9.22 0.14 7.97 0.40 5.78 1.53 5.93 0.15 6.29 0.45 7.66 0.44 Na20 2.82 0.32 7.55 0.56 2.38 0.21 2.40 0.23 1.90 0.39 1.88 0.15 3.93 0.32 3.82 0.49 K20 0.42 0.11 5.08 0.34 0.61 0.07 0.62 0.06 0.51 0.15 0.47 0.04 0.48 0.08 0.98 0.22 P205 0.16 0.17 0.15 0.13 0.15 0.04 0.15 0.07 0.09 0.05 0.14 0.04 0.14 0.15 0.39 0.25 SO.• 0.04 0.05 0.04 0.07 0.06 0.07 0.13 0.11 CI 0.09 0.06 0.25 0.13 0.07 0.02 0.11 0.02 0.26 0.07 0.20 0.03 0.21 0.08 0.09 0.07 Total 100 100 100 100 100 100 100 100
Rb 10.0 191.2 16.2 15.2 19.1 11.2 26.1 Sr 123.7 75.9 103.9 206.1 138.2 140.7 374.2 Zr 59.6 560.6 94.9 91.4 89.5 68.7 136.0 Nb 09 79.2 1.2 1.5 1.1 1.1 3.7 Ba 84 5 107.5 97.4 150.4 97.5 98.8 270.0 La 28 107.6 3.1 4.8 3.0 2.6 12.7 Ce 77 218.6 8.9 13.6 8.5 7.7 29.4 Pr 12 25.4 1.4 2.1 1.3 1.2 4.1 Nd 72 100.1 10.3 12.2 9.7 8.5 21.2 Sm 21 19.8 2.8 3.8 2.8 2.5 5.4 Eu 0.7 2.3 0.8 1.3 0.9 0.9 1.4 Gd 2.5 15.9 3.6 4.4 3.6 3.1 5.1 Tb 0.5 2.7 0.7 0.9 0.8 0.6 0.9 Dy 2.8 14.6 3.5 5.4 3.6 3.6 4.7 Ho 0.6 3.0 0.8 1.2 0.8 0.8 1.0 Er 2.0 8.5 2.4 3.7 2.6 2.5 3.1 Tm 0.3 1.1 0.3 0.5 0.4 0.4 0.4 Yb 1.9 7.8 2.4 3.8 2.6 2.5 3.0 Lu 0.3 1.1 0.4 0.6 0.4 0.4 0.4 '•7Sr/86Sr 0.704680 33 0.705172 37 0.704416 14 0.703971 12 0.704266 16 0.704247 10 0.704137 7 143Nd/144Nd0.512941 41 0.512893 12 0.513014 30 0.513034 13 0.512985 38 0.512996 25 0.512827 12 aNd(O) 5.95 5.01 7.37 7.76 6.81 7.02 3.73
with a low alkali content comparableto tholeiitic The volcanic provincesin a within-plate setting are magmas. James Ross Province, Marie Byrd Land, and Victoria At this step, we may already deciphertwo main Land on the Antarctic continentalplate and the oceanic tectonicsettings for the studiedtephra: first the 1996m islands scattered over the Antarctic oceanic crust. The layer,which appears well apart,and probably the 1981m volcanic provinces in a subduction setting are the layer, both correspondto a settingof within-plate southern part of South America, the South Sandwich magmatism;second, the "depletedLREE" layers(set 2'), Islmqds(SSIs), the Bransfield Strait, and the New Zealand the 989 m layer, and the 2587m layer originatefrom volcanoes. different subduction zones. The geochemicalsignatures of PSVs are taken from the literature. Unfortunately, some PSVs are poorly 4. ComparisonWith PSV Signatures documentedand cannotbe compareddirectly with our data. We will thereforerestrict comparisons by usingonly 4.1. Inventory of PSVs the available chemical data which seem to be the most The location of the main volcanic centers with pertinent. documentedactivity over the last 500 kyr is shownin First of all, we can already rule out some PSVs from Figure 1 (with referencesherein). the inventoryregardless of the tectonicsettings by using BASILE ET AL.: VOLCANIC LAYERS OF VOSTOK CORE 31,923 the differentiation degree of the magmas. The least et al., 1989],but the 87Sr/86Srisotopic composition differentiated tephra at Vostok are basaltic andesites (0.70517) seems more radiogenic. However, this (Figure 5a) with a SiO2 content higher than 52%. Thus differencecould be nonsignificantbecause a radiogenic the PSVs, which emitted only basalt (i.e., no tendencyfor 87sr/assrwas already noted in the Marie differentiatedproducts), are unlike the sourcefor any of Byrd Land provincewith valuesup to 0.70422.This was the Vostok layers. So we can already discard the interpretedas contaminationof the magmaby continental following centers: Penguin, Greenwich, and Livingston crust[?anter e! al., 1994;Kyle et al., 1994].This, andthe Islands in the Bransfield Strait [Smellie, 1989; Gonzales- major elementscomposition which correspondsto an Ferrari, 1991]; Seal Nunatak and Paulet Islands in the alkaline magmaseries [LeMasurier et al., 1989], appear James Ross Province [Smellie, 1989]; and Antipodes therefore consistentwith the Marie Byrd Land lava Islands on the New Zealand continental shelf [Gamble composition.Such an alkaline compositioncould also and 7hornson, 1989]. We now focus discussion on suggestPeter Island as a PSV,but the radiogenic 87Sr/86Sr within-plateand subductionsettings. isotopicsignature is incompatiblewith its within-oceanic plate setting. For the 1981 m layer the REE enrichmentprofile 4.2. Within-Plate PSVs (Figure 6a) is similarto that of Marie Byrd Land, but Two tephra layers belong to this setting: 1981 and trace element depletion (Zr/Nb 2169 rn 2260 rn 1280 rn 6- 6 • i 4i MgO I 2 2• ' 2J • 1280a i J •, 12•80b 40 45 50 55 60 65 70 75 40 45 50 55 60 65 70 75 40 45 50 55 60 65 70 75 6.0• 6.0• 6.0- • 1280b K203.o al'" 3.o•i ß " 3.o!J • '1280a 0.0 • i ' i • • 0.0J• T j • ' • •'•"• ß • • 0.0;•, • • I • I , I ] r 40 45 50 55 60 65 70 75 40 45 50 55 60 65 70 75 40 45 50 55 60 65 70 75 SiO2 SiO2 SiO2 Figure 4. Electronmicroprobe analysis for layers at 2169 m (20 shards),2260 m (20 shards),and 1280 m (47 shards)plotted on Harker diagramsof K20 versusSiO2 and MgO versusSiO2. The 1280 m layer is dividedin two subgroups1280a and 1280b. 31,924 BASILE ET AL.' VOLCANIC LAYERS OF VOSTOK CORE ...... Phono#te Tephriphonohte ...... Rhyotite T•8•chy•f½ Phonotephrite- .:::..::. Trachyande•ite 6 ...... i BasaniCe "' "?...... '..... 5 ...... ,'"'::...... ::4...... '"•"•"-*•••'•;"250• ...... •_•..'"'•--• '•o• . 3 ...... • L..,.:/"*•'"""•<•:•.,• :;•';•' ½•:•...... 1280...... .... Picro-ba•att• ( Basaltic '• Andestie •...... Dac•te I ...... andesRe 40 45 50 55 60 65 70 75 SiO2 (wt %) (2) ....:; :;:i'...... (4) ::• 100 -..-.-..• ...... '"•• '...... • !:11 :...... ,• ======...... ,•,••,,.•,.•~...... ":.,•:..: 2231 (3a) .....:'...... :::::::-:::::::'...... .-I • ....; • OePr Nd SmEu 6d Tb• HoEr Tm Yb Lu • OePr Nd 8mEu 6d Tb Dy Ho Er Tm Yb Lu ß.•' 1981 ß0 4 m .....• (3a) ...... ß 2587 2 ...... ; 98; 1 iii•i . (2)* EMII 0 ...... •...... ;'•...... ;...... :•...... :'i...... •...... •...... •'...... • ...... ;...... •...... •...... '•...... •...... •...... ?...... • 0 !0 20 30 40 50 60 70 80 0.701 0.703 0.705 0.707 Zr/Nb s7Sr/•Sr Figure 5. Geochemicalcharacteristics of Vostok ash layers. The ash layer label refers to the bottom depthof the l m long ice coreincrement containing the layer. Seetext for definitionof sets1 to 4. Set 2* representsset 2 without 1981 m. (a) Alkali content versus SiO2. Error bars representthe standard deviation obtained from our measurements.(b) REE profiles normalized to chondrite composition [Taylor and McLennan, 1985]. (c) La/Tb versus Zr/Nb. (d) Isotopic compositionsin the theoretical model reservoirs[Hart et al., 1992]: DMM, depletedmantle; HIMU, U/Pb enrichedmantle; EMI, enrichedmantle with lower crustaffinity; EMII, enrichedmantle with uppercrust affinity; mantle array representsthe signatureof a large number of oceanicbasalts [Rollinson, 1993]. Errors bars (2c0 are smallerthan the symbol size. BASILE ET AL.' VOLCANIC LAYERS OF VOSTOK CORE 31,925 1000 1000 =. lOO• 100 J '•-- ...... • ...... • MarieByrd Lanq • 10-• '• Province• q V•ciodaLandProvinces I • La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu 1000-•Erebus andMelbourne • Provinces loo:: • lO•'•;• '•arieByrd Land 0 I 2 3 4 5 6 7 8 Zr/Nb 14 q .. WithinContinental Plate Provinces '• DMM lii-. 1412 • DM..•'tiI•/•xW•thin Oceanic P•ateProvinces 12 j '-?• ".. •-• Melbourne 10 •Mano•'• •%. 10 • *•. '%. / Province 8 tiEdward...... '•• 6 • HIMU..... et9• z0 • u-2-• Kergelen-• . H ' • -10 w -6 -12 • -8 'f ...... •...... %•%'• • ...... l ...... •T•m"• •? ...... T • w Gaussbergs.--.• 0.701 0.702 0.703 0.704 0.705 0.706 0.707 0.702 0.704 0.706 0.708 87Srl•eSr .7SrP6Sr Figure 6. Geochemicalcharacteristics of within-platePotential Source Volcanoes and relatedash layers (black circles).(a) REE profiles for 1981 and 1996 m layers;shaded areas represent the limits of most randleast enrichedprofiles of Victoria Land provinces[Rocholl e! al., 1995] and Marie Byrd Land province[Lemasurier et al., 1976]. (b) La/Tb versusZr/Nb signaturesof 1981 and 1996 m layers of Victoria Land provinces(open circle, Hallett province [Rocholl e! al., 1995]; open square,Erebus provinceand Melbourneprovince [Kyle, 1989b, 1989c]) and Marie Byrd Land province(open triangle) [Lemasurieret al., 1989]. (c) Isotopesignature of the 1981 and 1996 m layers (black circles);within continental(right) and oceanicplate (left) provincesare separatedfor legibility. See Figure 5d for reservoirdefinitions. Hallett province[Rocholl et al., 1995], Melbourneprovince [WOrner et al., 1989], Marie Byrd Landprovince [Kyle et al., 1994],Kerguelen islands [Dosso and Muthy, 1980], HeardIsland [Barling and Goldstein,1990], BouvetIsland [Verwoerdet al., 1989], Marion and Edward Islands[Hart, 1988]. We concludethat the sourceof the 1996 m layer is 4.3. Subduction Zone PSVs very likely located in Marie Byrd Land province. The For subductionPSVs the provincesare the SSIs, the source of the 1981 m layer is probably restrictedto a Bransfield Strait, the South American and the New within-continentalplate source(Victoria Land or Marie Zealandprovinces. Set 2* containsseven "depleted REE" Byrd Land province), but we cannotdiscriminate in any Vostok layersthat appearto correspondto the signature greaterdetail. of SSIs by their REE enrichmentprofiles (Figure 7a), 31,926 BASILE ET AL.: VOLCANIC LAYERS OF VOSTOK CORE lOO•i lOO 2231 ß 2169 2254 10! 25O2lgg2 La Ce Pr Nd Sm Eu Gd Tb Oy Ho Er Tm Yb Lu La Ce Pr Nd Sm Eu PodTo Dy Ho Er Tm Yb Lu 3' 2 •i'•...... '. AndeanSouth 8 lcanicZone •2 Strait 2 •/ 258'7 0 20 40 60 80 100 120 Zr/Nb 0.702 0.703 0.704 0.705 0.706 e7Sr•Sr 6 xx '•"'":• AndeanVolcanicSouthZone •-4 HIMU11258 '•o•-2 • /Ch•ie 0.701 0.702 0.703 0.704 0,705 0,706 0.707 0.702 0.704 0.7• 0.708 87Sr/•Sr S7Sr/SSSr Figure 7. Geochemicalcharacteristics of subductionpotential source volcanoes. (a-c) signaturesof the SSIssm•d the "depletedLREE" layers(black circles).(b, d, e, f) showthe BransfieldStrait and the Andem•signatures with the 989 and 2587 m layers.Shaded areas represent the limits of mostand least depletedREE profilesfor eachprovince. Crosses, Bransfield Strait [Weaver et al., 1979];open triangle, SSIs [Pearceet al., 1995];open circles, South America [Hawkesworth and al., 1979;Futa and Stern, 1988]. their trace elements (Figure 7b) and their isotope series, a tholeiitic series, and a calc-alkaline series of compositions(Figure 7c). Some Vostok layers displaya oceanic arcs. SSIs are also the only volcanoes that Sr compositionslightly more radiogenicthan the SSIs produce1ow-K tholeiitic magmasamong the PSVs. Thus lava composition given by Pearce et al., [1995]. usingthese arguments, we classifythe two eventsof set 1 However, such an isotopicsignature is typical of young as !ow-K tholeiitic (1280a and 2326 m) originatingfrom oceanic subductionfor which the SSIs (age < 3Ma) are the SSIs. the unique candidate among the PSVs. We have no The 989 and 2587m layers (set 3a) are basaltic explanationfor this difference, but such a shift between andesites(Figure 5) from calcalkalineseries lava which the in situ lavasand their relativepyroclastics has already do not correspondto SSIs composition(Figure 7a, 7b, been observed[e.g., Ewart and $tipp, 1968] and points and 7c). They neither correspondto a New Zealand out the limits of the comparisonbetween pyroclastics and signaturesince the Taupo volcaniczone producedonly lavas. rhyoliticemissions (SIO2>70%) over the last 0.5 million We can extend the SSIs comparison to layers for years[E•art and Stipp, 1968;Simkin and Siebert,1994]. which no trace elements are available. Pearce et al. Thusonly the BransfieldStrait or volcaniczones in South [1995] have shown that SSIs lava follows three distinct America could be the sources. Both are associated with a magmatic trends in major elements:a 1ow-K tholeiitic similarsubduction setting (oceanic crust subducted under BASILE ET AL.: VOLCANIC LAYERS OF VOSTOK CORE 31,927 continental plate). There is a slight geochemical by Pearce and coworkers.It was also found that lava from distinction between those two volcanic provinces, as Cook and Thule Islandswith tholeiiticsignatures could illustratedby the La/Sm versus *7Sr/a6Sr plot (Figure 7e), also be the source.The 547 m layer (K20=0.86% at where the Bransfield Strait displays a less radiogenic SiO2-60.90% and a La/Sm=l.58) correspondsmore strontium isotope signature and a lower LREE closely to a talcalkaline series in which Leskov and enrichmentthan the Andean volcanoes(see appendixfor Freezlandare represented[Pearce et al., 1995]. These tectonic justification). Thus regarding the isotopic two examplesshow onceagain the interestof the traces information and the high contentof TiO2 (Table 2), the andat the sametime the limitationof the interpretation 989 m layer seemsto correspondto the Bransfield Strait wlnenonly major element composition is usedand only signature [Smellie, 1989]. Unfortunately, for the scarce data on sources are available in the literature. BransfieldStrait, REE profilesand 143Nd/144Nd data are not available to confirm our suggestion.For the 2587 m 5.2. Atmospheric Transport layer, the REE profile (Figure 7d) suggests South America and probablythe southernsouth volcanic zone It is obvious that in addition to the frequency of the (SSVZ) as a possiblesource. eruptions,the height of injection, the circulation of We may also suggest a source for the 1280b and troposphericair masses,and the distancebetween the 1431m layers (set 3b) for which we do not have trace volcanic center and the location of the studied Antarctic element data. These layers belong to the calc-alkaline ice core site also greatly affectsthe occurrenceof tephra series, and taking into account their relative acidity in central Antarctica. For example, the Byrd ice core is (SIO2=66%), their low TiO2 content,and the rather small located,-500 km from the Marie Byrd volcanoes,and the size of the particles,we may suggestwith confidencean ice record, which covers the last • 100 kyr contains Andean origin. The 181 m layer belongsto set 2, which •2000 tephraevents. This gives an overall frequencyof 1 includes both within-plate and subductionvolcanoes. In per 50 years. For Vostok, a total of---20 tephra have been this case, the major elements are not sufficient to detected,mostly from SSIs and •-5000 km away. Over the discriminate,and no sourcecan be proposed. 400 kyr of the ice record,this gives a frequencyof 1 per 20,000 years. It is of interest to note the decreaseby 3 orders of magnitude in occurrenceof tephra when the 5. Discussion distanceto the sourcechanges by I order of magnitude. Interestingly, the occurrence of tephra in the Dome-F 5.1. Source Volcanoes core also in East Antarctica but closer to SSIs centers is We finally suggestthat the subductionvolcanoes of 50% higher (1.5 per 20000 years) than in the Vostok core the SSIs are sourcesof the "depletedLREE" layers (104, [l:•.t•i e• a/., 1999]. However, the transport is not in a 547, 1992, 2169, 2231, 2254, and 2502 m) as well as the direct line to Antarctica, and it is necessaryto take into 1280a and 2326 m layers. Sourcesfor the 1280b, 1431, accountnot only the distancebetween the volcanic center and 2587 m layers are probably located in the Andean and the Antarctic site but also the circulation of cordillera.The 2587 m layer is probablyrestricted to the troposphericair masses. southern south volcanic zone of South America. The transportof volcanic cloudsfrom the SSIs to East Bransfield Strait is the sourcefor the 989 m layer. The Antarctica seeins a priori consistent with the general 1996 m layer originatesfrom Marie Byrd Land province, atmospheric circulation with tropospheric air masses while the 1981 m layer is from West Antarctica. spiralingsouthward toward Antarctica.However, the size The SSIs appeartherefore as an important sourcefor of particles restricts the duration of transport because visible tephra layers discoveredin the Vostok ice core most of the ash particles are found in the first hundred (i.e., more than 65% of the events). The SSIs signature kilometers around volcanoes [Fisher, 1964]. For has alreadybeen suggestedfor two Vostok layers found example,20 [tin particlessettle rapidly in the atmosphere at depths of 100.8 and 550m from previous cores and accordingto Stoke's law fall by 500 to 1000 m per adjacentto our Vostok ice cores.Kyle e! al. [1984] and day. Taking into accountthe 4000 m elevation of the East Palais et al. [1987] studiedone layer at 100.8 m from a Antarctic plateau,the volcanicash would have to reach a shallow ice core that is equivalentto our 104 m layer on very great height after eruption. So assumingthat SSIs BH1, and ?alais et al. [1989] analyzedthe 550 m layer in volcanic eruptions can inject tephra up to a 6000 m the "3G" Vostok core that is equivalent to our 547 m altitude in the troposphere,corresponding to a volcanic layer from "4G2." From the major element composition explosivity index (VEI) of 2 to 3 [Newhall and Self, these authors proposedthe Candlemasvolcano (one of 1982], the transportduration to Vostok shouldnot exceed the 13 SSIs volcanoes)as the sourcefor the two layers. 2 to 4 days. This representsdirect advection with an Our data also suggesta SSIs source but from other apparentaverage horizontal speedof •100 km/h, a value volcanoes. From the study of Pearce e! al. [1995], that increaseswith particle size and with the lower level Candlemasvolcano belongs to the low-K tholeiitic trend reachedby the volcanic cloud. Thus the tephra event in (K20--•0.35% at SIO2---60%) and has a high LREE the ice core is probably associatedwith a particular depletion (La/Sm<0.8). By comparisonour results show synoptic situation leading to high eddies and strong for the 104 m layer a similar composition for major tropospheric perturbations which penetrate almost elements(K20=0.5% at SiO2-60.23%) but a low LREE directly to the center of Antarctica (Figure 8). This depletion (La/Sm=0.92). This compositioncorresponds schemeof transportcan be extendedto South America more closely to tholeiitic SSIs volcanoes such as (1280b, 1431 and 2587 m layers) and to Antarctic Bellingshausen,Saunders, or Visokoy Islandsas defined Peninsula (989 m layer) which representmore distant 31,928 BASILE ET AL.' VOLCANIC LAYERS OF VOSTOK CORE 30' 30'W SOUTH 8AN.DWtCH.• 60'W SLAN•S i BRANSF.!ELD STRAIT 90'W 90'E 180' 0 1000 2000 3000 4000km Figure 8. Sketchof possibleatmospheric paths suggested from the identificationof the sourcesof the 15 volcanicash layers. The width of each arrow is proportionalto the occurrenceof the pathways.The maximumparticle size is indicatedin the arrows.While the transportto Vostok is probablymore direct (a straightline), the shapesof the trajectoriesare inspiredfrom reversetrajectory models (C. Genthon, personalcommunication 2001) and Adriani et al. [1995]. See Figure I for volcanoand drilling site abbreviations. sources,but correspondingtephra layers in ice contain precipitation.The size of the dustparticles in the ice core smallerparticles (20 lamfor maximumsize)(Figure 8). is barely greater than -2 [tm and their corresponding In spite of the low number of tephra eventsin the gravitational settling velocity only 5 m/d. Moreover, as Vostok ice core, the occurrence of volcanic events continental dust deposition in Antarctic seems to be comingfrom the southwesternpart of the Atlantic does affectedby seasonalvariations, [Thompson et al., 1979], not significantlychange with climate(Figure 2). During we can deduce that the same is true for volcanic ash glacialperiods we foundsix layerscoming from SSIsand deposition.Our data do not, however, allow us to discuss one from the AntarcticPeninsula. During warm periods the seasonal effect of atmospheric circulation on the we tbund three layerscoming from SSIs and three from occurrence of volcanic dust at Vostok. South America. This rapid transportfrom the southern Only one layer in the Vostok core (1996 m) and part of the South Atlantic toward East Antarctica doesnot probably one event (726 m) in the Dome C core [Kyle et seem to be dependenton the climatic conditionsnor on al., 1981] originated from Marie Byrd Land volcanoes. theprobable northward position of thepolar front during This suggeststhat troposphericconditions carrying air glacialtimes [Petit et al., 1999]. This globalpathway masses from West to East Antarctica can occur. seemsconsistent with the almostcontinuous tropospheric However, considering the very high frequency of transportof continentaldust to Vostok [Groussetet aL, eruptions in Marie Byrd Land province [Gow and 1992;Basile et al., 1997; Basile, 1997] sincecontinental Williamson, 1971] m•d the relatively close distance dust originatesmostly from southernparts of South (---2300kin in direct line), such a pathway seemsvery Americafor glacial as well as for interglacialperiods unusual.Finally, note the absenceof visible New Zealand [Basile,1997]. However, unlike the volcanicpyroclastics tephra (distanceNew Zealand-Vostok:6000 km in direct injectedat high altitude by the eruption,continental line) in spite of the frequencyand violenceof eruptions particlesare mobilized at surfacelevel and injected in the from the active Taupo volcanic province [e.g., Walker, atmosphericcirculation. Their transportprobably follows 1981; Simkinand Siebert,1994]. the generalcirculation pattern, and transport to Antarctica Tephra eventsfrom this studydo not make it possible is longer allowing dust depositionand removal by to reconstructa step-by-steptrajectory, as we identify BASILE ET AL.' VOLCANIC LAYERS OF VOSTOK CORE 31,929 only the sourceand the size of particles. They probably appears to be consistent with the almost continuous correspondto a particularatmospheric circulation pattern. advection of continental dust from South America, but It would be of interestto know what synopticconditions the latter is probablygoverned by the spiralingthat takes led to such events and simulate them using a general place around Antarctica circulation. The atmospheric circulationmodel. Besidesthis, theseash layersrepresent pathway from West to East Antarctica for clouds stratigraphicmarkers which are independentof climate. containingvolcanic ash appearsonly once in the Vostok Their geochemicalcomposition opens possibilities for a core, randthere is no event from New Zealand. reliable ice core correlation with the existing or the Besidesbeing used as atmospherictracers, ash layers ongoing East Antarctic ice coring projects (e.g., Dome representstratigraphic markers that are also independent Fuji, EPICA, etc.). The abundanceof SSIs layers in the of climate. Thus precise and complete geochemical Vostok core, with different geochemical signaturesof analysesperformed on tephra opennew fields of research magmas, also opens possibilities of correlation with for core correlationand absolutedating of ice cores. South Atlantic marine sedimentcores documentingleads and lags between ocean and ice climatic records. An Appendix' Bransfield Straight and South additional use is the possible absolute dating of the American Geochemical Signatures Vostok core. In fact, Marie Byrd Land appearsto be a good candidate (1996m layer) since the alkaline The South American and Bransfield Straight volcanic compositionof the lava near the sourceis favorable for provinces are associatedwith the subductionof oceanic theapplication of the3øAr/4øAr method for thisvolcanic crust under continental crust. However, a slight area,a possibilitythat is understudy [ Wilch et al., 1999]. geochemicaldistinction exists between the two since the subduct•onstopped under the Antarctic Peninsulaaround 6. Conclusion 4 Myr ago [Tokarski, 1991]. The Bransfield Strait, still active, is related to the opening of an asymmetric We have attemptedto identify the volcanic sourceof marginal basin located between the paleovolcm•ic arc fifteen ash layers of the Vostok ice cores. We used a (South Shetland Islands) and the within-plate volcanic more systematicmethod than previous studies, first by area of the Jame Ross province [Hole et al., 1995]. Thus usingtrace element and isotopic signaturesin addition to the coinpositionof the magma integratesthe composition major element compositionand, second,by considering of the subductedfossil plate and a MORB (mean oceanic all the Antarctic and subantarcticvolcanic provinces ridge basalt) composition.This transitional signatureis active over the last 0.5 Myr. slightly different from that of the South American We first show the importanceof making precisemajor subduction volcanoes. These chemical differences are element m•alyses on a large set of glass shards. It is illustratedin Figure7 bythe La/Sm versus 87Sr/a6Sr plot. possible to detect heterogeneouslayers such as tephra The Bransfield Strait has a less radiogenic strontium from a zoned magma chamberor thosethat are the result Isotopic signatureand a lower LREE enrichmentthan the of the mixing of plumesfrom different eruptions. Andean volcanoes.This is probably related to a higher The geochemical identity card of each ash layer continental crust contribution in Andean lava. representsa fingerprint of the tectonic conditionsthat govern the magma genesis. We deduced that the 104, Acknowledgments. Vostok is a joint project amongRussia, France, and the United States. We acknowledge the Russian 547, 1280a, 1992, 2169, 2231, 2326, 2254, and 2502 m Antarctic Expedition,the IFRTP (Institut Francaisde Recherche layersas well as the 989, 1280b, 1431, and 2587 m layers en Territo,re Polaire), and the Office of Polar Program of NSF originate from subduction volcanoes. In contrast, the for logistic support.The project is supportedin Russia by the 1981 and 1996 m layers originate from within-plate Russian Ministry of Sciences. This work was funded by the volcanoes.The comparisonof the tephra signatureswith French INSU-CNRS Programme National d'Etude de la Dynalnique du Climat and VariEnte (Variabilit• de Antarctic and subantarcticvolcanic signaturesshow that l'Environnementde la Terre) programs.We are grateful to Jean the main sources of Vostok volcanic tephra are the Philippe Eissen for his help in the major element analyses volcanoes of the South Sandwich Islands since nine carried out in Brest (Institut de Recherche ct de layers (the "depletedLREE" layers) originate from this D6veloppement).We are also gratefulto C. Jeandelfor accessto volcanic arc. The other sources are the southern Andean the thermo-ionisation mass spectrometer at Paul Sabatier University (Toulouse). We thank F. Keller and M. Revel for volcanicprovince (1280b, 1431 and 2587 m layers),the their help in inductively coupled plasma mass spectrometryat Marie Byrd Land Province (i996 m layer), and the the lnstitut Dolomieu of JosephFourier University, Grenoble, BransfieldStrait (989 m layer). The sourceof the 1981 m and J.P. Balestrierifor technicalhelp. We thank F. Albar•de, H. layer is probably restrictedto a continentalwithin-plate Lapierre,M. Pateme,F. Guichard,and C. Genthonfor fruitful discussionsas well as the anonymousreviewers for their sourcebut has not yet been identified. We show that the suggestionsand improvements. sourcecan be identifiedwith a higher level of confidence when using trace elements and isotopic composition ratherthan major elementsalone. References The occurrenceof tephra eventssuggests preferential atmospheric trm•sport from the southern part of the Adl-iani, A., T. Deshler, G. Donfrancesco,and G.P. Gobbi, Polar western Atlantic toward the East Antarctic ice cap Stl-atosphericclouds and volcanicaerosol during spring 1992 associatedwith strong eddies penetrating Antarctica. over McMurdo Station, Antarctica: Lidar and particle counter comparisons,d. Geophys. Res. 100, 12, 25877- Occurrence of this transport does not appear to be 25897, 1995. dependenton climate, but our data are insufficient to Albar•de, F., How deep do commonbasaltic magmas form and discussthe seasonaleffect. This atmosphericscheme differentiate?, d. Geophys.Res., 97, 10,997-11,009, 1992. 31,930 BASILE ET AL.: VOLCANIC LAYERS OF VOSTOK CORE Baker, P.E., South Sandwich Island, in Volcanoes of the slab window formation along continentaldestructive plate Antarctic Plate and SouthernOceans, edited by Antarct. Res. margins, Volcanism Associated With Extension at Set., vol. 48, W. E. LeMasurier and J. W. Thomson,pp. 361- ConsumingPlate Margins, edited by J.L. Smellie,Geol. Soc. 395, AGU, D.C., 1989. Spec.Pub., 81,265-285, 1995. Barting,S. and S.T. Goldstein,Extreme isotopic variations in Jacobsen,S.B., and G.J. Wasserburg,Sm-Nd isotopicevolution Heard Island lavas and the nature of mantle reservoirs, of chondrites,Earth Planet. Sci. Lett., 50, 139-155, 1980. Nature, 348, 59-62, 1990. Kyle, P.R., McMurdo Volcanic Group, Western Ross Basile,I., Origine desa6rosols volcaniques et continentauxde la Embayment,Introduction, in Volcanoesof the Antarctic carottede Vostok (Antarctique),Ph.D. thesis,270 pp., Univ. Plate and Southern Oceans, Antarct. Res. Set., vol. 48, of JosephFourier, Grenoble, France, 1997. editedby W. E. LeMasurierand J. W. Thomson,pp. 19-25, Basile, I., F.E. Grousset,M. Revel, J.R. Petit, P.E. Biscaye and AGU, WashingtonD.C., 1989a. N.I. Barkov, Patagonianorigin of glacial dust depositedin Kyle, P.R., MelbourneVolcanic Province, in Volcanoesof the East Antarctica(Vostok, Dome C) duringglacial stages2, 4 Antarctic Plate and Southern Oceans,Antarct. Res. Set., vol. and 6, Earth Planet. Sci. Lett., 146, 573-589, 1997. 48, editedby W. E. LeMasurierand J. W. Thomson,pp. 48- Cox, K. G., J. D. Bell, and R.J. Pankhurst,The Interpretationof 80, AGU, WashingtonD.C., 1989b. IgneousRocks, 450 pp., Allen and Unwin, Concord,Mass., Kyle, P.R., ErebusVolcanic Province,in Volcanoesof the 1979. Antarctic Plate and Southern Oceans, Antarct. Res. Set., vol. Dosso, L., and V.R. Murthy, ANd isotopic study of the 48, editedby W. E. LeMasurierand J. W. Thomson,pp. 81- Kerguelen Islands: Inferenceson enrichedoceanic mantle 145, AGU, WashingtonD.C., 1989c. sources,Earth Planet. Sci. Lett.,48, 268-276, 1980. Kyle, P.R., andP.A. Jezek,Composition of threetephra layers Dupr6, B. and C.J. All•gre, Pb-Sr variation in Indian Ocean from the Byrd Station ice core, Antarctica,d. Volcanol. basaltand mixing phenomena,Nature, 303, 142-146, 1983. Geotherm. Res., 44, 225-232, 1978. Ewart, A., and J.J. Stipp, Petrogenesisof the volcanicrocks of Kyle, P.R.,P.A. Jezek,E. MosleyThompson, and L. Thompson, the Central North Island, New Zealand, Geochim. Tephralayers in the Byrd Stationice core and the Dome C Cosmochim.Acta, 32, 699-736, 1968. ice core, Antarctica and their climatic importance, J. Falkner, K.K., G.P. Klinkhammer, C.A Ungerer, and D.M. •/•91canol.Geotherm. Res.,] 1, 29-39, 1981. Christie, Inductivelycoupled plasma massspectrometry in Kyle, P.R., J. Palais,and E. Thomas,The Vostoktephra: An geochemistry,Annu. Rev. Earth Planet. Sci., 23, 409-449, importantenglacial stratigraphic marker?, Antarct. J. U.S., 1995. /9, 64-65, 1984. Fisher, R.V., Maximum size, median diameter, and sorting of Kyle,P.R., R.J. Pankhurst, S. Mukasa,K. Panter,J. Smellie,and tephra,Y. Geophys.Res., 69, 341-355, 1964. W. Mcintosh,Sr, Nd and Pb isotopicvariations in the Marie Fuji, Y., M. Kohno, H. Motoyama, S. Matoba, O. Watanabe,S. Byrd plume, west Antarctica,U.S. Geol. Surv. Circ. 1107, Fujita, N. Azuma, T. Kikuchi, T. Fukoaka,and T. Suzuki, 1COG8,p. 184, Berkley,Calif., 1994. Tephra layers in the Dome Fuji (Antarctica)deep ice core, Le Maitre, R.W., et al., A Classificationof IgneousRocks and Am•. Glaciol.,29, 126-130, 1999. Glossaryof Terms,Blackwell, Malden, Mass., 1989. Futa, K, and C.R. Stern, Sr and Nd isotopicand trace elelnent LeMasurier, W.E., Marie Byrd Land, in Volcanoesof the compositionof Quaternaryvolcanic centersof the southern Antarctic Plate and Southern Oceans, Antarct. Res. Ser., vol. Andes, Earth Planet. Sci. Lett., 88, 253-262, 1988. 48, editedby W. E. LeMasurierand J. W. Thomson,pp. 147- Gamble, J.A., and J.W. Thomson, Subantarcticvolcanoes of the 256, AGU, WashingtonD.C., 1989. Pacific Plate, in Volcanoes of the Antarctic Plate and LeMasurier,W.E., and D.C. Rex, Evolutionof linear volcanic SouthernOceans, edited by Antarct. Res.Set., vol. 48, W. E. rangesin Marie Byrd Land, West Antarctica,J. Geophys. LeMasurier and J. W. Thomson,pp. 361-395, AGU, D.C., Res. Lett., 94, 7223-7236, 1989. 1989. LeMasurier,W.E., D.C. Kyle P.R., andP.C. Rankin,Rare earth Gonzales-Ferran, M.O., The Bransfield rift and its active elementgeochemistry of volcanicrocks from the Executive volcanism,in GeologicalEvolution of Antarctica,edited by CommitteeRange Marie Byrd Land,Antarct. J. U.S., 11(4), M. R. A. Thomson,J. A. Crame and J. W. Thomson,pp 505- 263-267, 1976. 509, CambridgeUniv. Press.,New York, 1991. LeMasurier, W.E. and J.W. Thomson, Volcanoes of the Gow, A.J., and T. Williamson, Volcanic ash in the Antarctic ice Antarctic Plate and SouthernOceans, Antarct. Res. Ser., vol. sheet and its possibleclimatic implications,Earth. Planet. 48, editedby W. E. LeMasurierand J. W. Thomson,AGU, Sci. Lett., 13,210-218, 1971. WashingtonD.C., 1989. Grousset,F.E., P.E. Biscaye, A. Zindler, J. Prospero,and R. Newhall, C.G., and S. Self, Volcanic explosivityindex (VEI): Chester,Neodymium isotopes as tracersin marinesediments An estimateof explosivemagnitude of historicalvolcanism, and aerosols : North Atlantic, Earth Planet. Sci. Lett., 87, J. Geophys.Res., 87(2), 1231-1238, 1982. 367-378, 1988. Nielsen, C. H. and H. Sigurson, Quantitativemethod for Grousset,F.E., P.E. Biscaye, M Revel, J.R. Petit, K. Pye, S. electronmicroscope analysis of sodiumin syntheticglasses, Joussaumeand J. Jouzel,Antarctic (Dome C) ice core dustat AmericanMineralogist, 66, 547-552, 1981. 18 ky BP: Isotopicconstraints on origins,Earth Planet. Sci. Palais,J., Particlemorphology, composition and associatedice Lett., JJJ, 175-182, 1992. chemistryof tephralayers in the Byrd ice core:Evidence for Hammer, C.U., Pastvolcanism revealed by GreenlandIce Sheet hydrovolcaniceruptions, Ann. of Glaciol., 7, 42-48, 1985. impurities,Nature, 270, 483-486, 1977. Palais, J., P. Kyle, E. Mosley Thompson,and E. Thomas, Hart, R., Heterogeneousmantle domains: Signatures,genesis Correlationof a 3200 year old tephra in ice cores from and mixing chronologies,Earth. Planet. Sci. Lett., 90, 273- Vostok and South Pole stations, Antarctica, Geophys.Res. 296, 1988. Lett.,/4(8), 804-807, 1987. Hart, S.R., E.H. Hauri, L.A. Oscl•nam•, and J.A. Whitehead, Palais, J., J.R. Petit, C. Lorius, and Y.S. Korotkevitch, Tephra Mantle plumes and entrainmentisotopic evidence, Science, layersin the Vostok ice core: 160,000 years of southern 256, 517-520, 1992. hemispherevolcanism, Antarct. J., Rev.,24, 98-100, 1989. Hawkesworth,C.J., M.J. Norry, J.C. Roddick,P.E. Baker, P.W. 143 144 87 86 Panter,K.S., R. Pankhurst,P.R. Kyle, J. Smellie, S. Mukasa, Francis, and R.S. Thorpe, Nd/ Nd, Sr/ Sr, and andW. Mcintosh,Sr, Nd, O, andPb isotopesfrom the Mount incompatible element variations in calc-alkaline andesites Sidleyvolcano, Marie ByrdLand, Antarctica, (abstract), U.S. and plateau lava from South America, Earth Planet. Sci. Geol. Surv. Circ. 1107, ICOG8, p 241, Berkeley,California, Lett.,42, 45-57, 1979. 1994. Heiken G., and K. Wohletz, Volcanic ash, Los Alamos Nat. Pearce, J.A., P.E. Baker, P.K. Harvey, and I.W. Luff, Lab., Univ. of California Press, Berkeley, Calif., 246 pp., Geochemicalevidence for subductionfluxes, mantle melting 1985. and fractionalcrystallisation beneath the South Sandwich Hole, M.J., A.D. Saunders,G. Rogers, and M.A. Sykes, The Island Arc, J. Petrol., 36, 1073-1109, 1995. relationshipbetween magmatism, lithospheric extension and Petit, J.R., et al. Climate and atmospherichistory of the past BASILE ET AL.: VOLCANIC LAYERS OF VOSTOK CORE 31,931 420,000 yearsfrom the Vostokice core,Antarctica, Nature, Weaver, S.D., A.D. Saunders,R.J. Pankhurst,and J. Tarney,A 399. 429.-436. 1999. geochemicalstudy of magmatismassociated with the initial Rocholl, A., M. Stein, M. Molzahn, S.R. Hart, and G. W6rner, stagesof back-arcspreading, Contrib. Mineral. Petrol., 68, Geochemical evolution of rift magmas by progressive 151-169, 1979. trapping of a stratified mantle sourcebeneath the Ross Sea Wilch, T. I., W. C. Mcintosh, and N. W. Dunbar, Late Rift, Northern Victoria Land, Antarctica, Earth Planet. Sci. Quaternaryvolcanic activity in MarieByrd Land: Potential Lett, 131,207-224, 1995. 40Ar/39Ar-dated time horizonsin West Antarctic ice and Rollinson, H.R., Using geochemical data: Evaluation, marinecores, Geol. Soc.A•n. Bull., ]]], (10), 1563-1580, presentation,interpretation, in Geochemist•TSeries, 352 pp., 2000. Addison-Wesley-Longman,1993. Wilson, M., Igneous Petrogenesis.' ,4 Global Tectonic Rowley, P.D., J.W. Thomson, J.L. Smellie, T.S. Laudon, K.E. ,4pproach,466 pp., Chapman and Hall, New York, 1989. La Prade, and W.E. LeMasurier, Alexander Island, Palmer W6rnerG., H. Niephaus,J. Hertogen,and L. Viereck,The Mt Island and Ellsworth Land, in I/olcanoes of the Antarctic Melbournevolcanic field (Victoria Land, Antarctica),II; P/ate and Southern Oceans, Antarct. Res. Ser., vol. 48, Geochemistryand magmas genesis, Geol. dahrb., E38, 395- edited by W. E. LeMasurier and J. W. Thomson, pp. 257- 433, 1989. 300, AGU, WashingtonD.C., 1989. Zielinski,G.A., P.AoMayewski, L.D. Meeker,S.I. Writlow, Shane, P.A.R., T.M. Black, B.V. Alloway, and J.A. Westgate, M.S. Twickler,M. Morrison,D.A. Meese,A.J. Gow, and Early to middle Pleistocene tephrochronologyof North R.B. Alley,Record of volcanismsince 7,000 B.C. fi'omthe Island, New Zealand: hnplications for volcanism, tectonism GISP2 Greenlandice core and implicationsfor the volcano- and paleoenvironments,Geol. Soc. Am. Bull., 108(8), 915- climatesystem, Science, 264, 948-952,1994. 925, 1996. Zielinski,G.A., M.S. Germani,G. Larsen,M.G.L. Baillie, S.I. Sirekin T., and L. Siebert, I/olcanoes of the world, 349 Whitlow, M.$. Twickler, and K. Taylor, Evidenceof the pp.,Geosci.Press, Inc., Tucson,Arizona, 1994. Eldgja(Iceland) eruption in theGISP2 Greenland ice core: Sinclair, M., An objectivecyclone climatology for the southern Relationshipto eruption processes and climatic conditions in henrisphere,Am. Meteorol. Soc., 122, 2239-2256, 1994. thetenth century, Holocene, 5, 129-140,1995. Smellie, J.L., Graham Land and South Shetland Islands, in Zielinski,G.A., P.A. Mayewski,L.D. Meeker,S.I. Whitlow,and l/o/canoes of the Antarctic Plate and Southern Oceans, M.S. •Iwickler,110,000 yr recordof explosivevolcanism A•ttarct.Res. Set., vol. 48, editedby W. E. LeMasurierand J. fi'omthe GISP2 (Greenland) ice core, Quat. Res., 45(2), 109- W. Thomson,pp. 303-355, AGU, WashingtonD.C., 1989. 118, 1996. Smellie, J.L., The upper Cenozoic tephra record in the south Zielinski,G.A., P.A. Mayewski,L.D. Meeker,K. Gr6nvold, polar region: A review, Global Planet. Change, 21, 51-70, M.S. Germani,S.I. Whitlow, M.S. Twickler,and K. Taylor, 1999. Volcanic aerosol records and tephrochronologyof the Taylor, SR., and S.M. McLennan, The continental crust .' its Summit, Greenland,ice cores,•. Geophys.Res., 102. 625-26, compositionand evolution,312 pp., Blackwell Sci,. Malden, 640. 1997. Mass., 1985. Thompsol•, L.G., E. Mosley Thompson, and J.R. Petit, Glaciological interpretationof microparticle concentrations N. Barkov, Arctic and Antarctic ResearchInstitute, 38 fi'om the French 905 m Dome C, Antarctica core, in Sea BeringSt.. 199226St. Petersburg,Russia. level, ice and climatic change, 131 pp., IAHS Publ., I. Basile,Centre Europ•en de Rechercheet d'Enseignement Bentbrugge,Belgium, 1979. cn G6osciencesde l'Environnement,Europ61e M6diterran6cn de Tokarski, M.A.K., The late Cretaceous-Cenozoic structural l'Arbois, B.P. 80, 13545 Aix-en-Provence, France history of King George Island, South ShetlandIslands, and (basile(?bcerege.fi'). its plate tectonic setting, in Geological evolution of F. E. Grousset,D6partement de Gdologieet Ocdanographie, Antarctica, edited by M. R. A. Thomson,J. A. Crame, and J. Av. des Facultds,Univ. BordeauxI, 33405 Talence,France. W. Thomson, pp. 493-497, Cambridge Univ. Press., New J. R. Petit, Laboratoirede Glaciologieet Gdophysiquede York, 1991. !'Environnement, BP 96, 38402 St. Martin d'H&es, France. Verwoerd, W.J., L. Chevallier, and J.W. Thomson, Oceanic $. Tourol•, Laboratoirede G6ologiedes ChainesAlpines, islandson the Antarcticplate, Bouvetoya,in I/olcanoesof the Institut Dolomieu, 15 rue Maurice Gignoux,38301 Grenoble, Antarctic Plate and Southern Oceans, Antarct. Res. Set., vol. France. 48, editedby W. E. LeMasurierand J. W. Thomson,pp. 405- 410, AGU, WashingtonD.C., 1989. Walkel-, G.P.L., Characteristicsof two phreatomagmaticashes and fi•eir water-flushedorigins, 3'. k'olcanol.Geother;n. Rex., (Received November 2, 2000; revised March 9, 2001: 9, 395-407, 1981. Accepted March 19, 2001.)