JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 105,NO. B3, PAGES 5997-6016,MARCH 10, 2000
The chemically zoned 1949 eruption on La Palma (Canary Islands): Petrologic evolution and magma supply dynamics of a rift zone eruption AndreasKlagell Max-Planck-Institutfar Chemic,Mainz, Germany
Kaj A. Hoernle and Hans-UlrichSchmincke GeomarForschungszentrum, Kiel, Germany
James D. L. White
GeologyDepartment, University of Otago,Dunedin, New Zealand
Abstract. The 1949 rift zoneeruption along the CumbreVieja ridgeon La Palmainvolved three eruptivecenters, 3 km spacedapart, and was chemicallyand mineralogically zoned. Duraznero cratererupted tephrite for 14 daysand shutdown upon the openingof Llano del Banco,a fissure that issuedfirst tephriteand, after 3 days,basanite. Hoyo Negro crateropened 4 dayslater and eruptedbasanite, tephrite, and phonotephrite, while Llano del Bancocontinued to issuebasanite. The eruptionended with Durazneroerupting basanite with abundantcrustal and mantle xenoliths. The tephritesand basanites from Durazneroand Llano del Bancoshow narrow compositional rangesand def'me a bhmodalsuite. Each batch ascended and evolved separately without significant intermixing,as did the Hoyo Negro basanite,which formedat lower degreesof melting.The magmasfractionated clinopyroxene + olivine+ kaersutite+ Ti-magnetiteat 600-800 MPa and possibly800-1100 MPa. Abundantreversely zoned phenocrysts reflect mixing with evolvedmelts at mantledepths. Probably as early as 1936,Hoyo Negrobasanite entered the deeprift systemat 200-350 MPa. Someshallower pockets of thisbasanite evolved to phonotephritethrough differen- tiationand assimilation of wall rock.A few monthsprior to eruption,a mixingevent in the mantle may havetriggered the final ascentof the magmas.Most of the eruptedtephrite and basanite ascendedfrom mantledepths within hours to dayswithout prolonged storage in crustalreservoirs. The CumbreVieja rift zonediffers from the rift zonesof Kilaueavolcano (Hawaii) in lackinga summitcaldera or a summitreservoir feeding the rift systemand in beingsmaller and lessactive with mostof the rift magmasolidifying between eruptions.
1. Introduction and typically,though not always,results in an eruption.Much Volcanic rift zones are common features of basaltic ocean magmais storedin the deeperpart of the rift system,forming crystalmush or isolatedpockets which differentiateand then islandand continental volcanoes, yet a full understandingof their eithersolidify or are erupted[Wright and Fiske, 1971; Ryan, role in the overallmagma supply system and in controlling 1988;Delaney et al., 1990].Mixing of distinctmagmas in the petrologicevolution remains elusive. An exceptionis Kilauea summitreservoir and in the deeprift is importantin controlling volcano(Hawaii), where interdisciplinary studies have provided the compositionof the eruptedlavas, although some magmas deeperinsight into the geometryand dynamicsof the volcano's bypassthe summit reservoir [Garcia et al., 1989,1996]. magmaplumbing system (see review by Tilling and Dvorak La Palmahas a well-developedactive volcanic rift zonealong [1993]),yet manyaspects, such as the geometryof thesummit a north-southtrending ridge [Middlemost,1972; Carracedo, reservoir,remain controversial [Pietruszka and Garcia,1999]. At 1994]. Only a few studies,however, have addressedmagma Kilaueaa nearlycontinuous supply of magmaaccumulates over transportand storage beneath and within this rift system[Klagel timein a summitreservoir at 2-4 krn depth,gradually inflating et al., 1997] and the evolutionof its magmas[Praegel, 1986; the summitregion. Episodic lateral injection of summitmagma Elliott, 1991].Here we discussthe petrology,geochemistry, and into adjacentrift zones(a rift event)releases accumulated stress magmasupply dynamics of the 1949 eruption,which serves as a casestudy for complexrift zoneeruptions on La Palma.We show •Now at Fachbereich Geowissenschaften,Universitltt Bremen, howthe eruptedmagmas can be relatedto storage,fractionation, Bremen,Germany. and mixing in reservoirswithin the mantle and the crust and comparethe situationon La Palma to that of Kilauea. The 1949 Copyright2000 by theAmerican Geophysical Union. eruptionis of particularinterest because a varietyof eruptive Papernumber 1999JB900334. processesoccurred during the 37 days of activityand because 0148-0227/00/1999JB900334509.00 eyewitnessaccounts indicate an interconnectionof widelysepa-
5997 5998 KLOGEL ET AL.' THE 1949 RIFT ZONE ERUPTION ON LA PALMA
N
o 5 km
Santa I Historiceruption Cruz [--•ß Recentsediments I----] CumbreVieja series '""• Older volcanicseries • Basalcomplex -,- -,- Cliff / steep ridge
--- Cumbre Vieja rift zone 1949
1585 1949 1712 / M25 Lanzarote•/S 1 29 La Palma/... •[ ._ / c;anary Islands 1646 28L •/'/ GO•m•er•• •Tenerife '• Fu ert e v•t enura/ / • / ""• Gran Canaria I / El Hierro I ! ATLANTICOCEAN
1971
Figure 1. Locationmap (inset)and simplifiedgeological map of La Palma [after Navarro and Coello, 1993; Ancocheaet al., 1994;Carracedo et al., 1999].Most historiceruptive centers are locatedalong the north-south trendingCumbre Vieja rift zone.The three vents of the 1949eruption are Duraznero (1880 m abovesea level) and HoyoNegro (1880 m) craterslocated on the Cumbre Vieja crestand Llano del Banco fissure (1300 m) at theridge flank to the west of the rift.
ratedvents. Furthermore, the late eruptedlavas carry a great La Palmahas been the mostactive of the CanaryIslands in variety of crustal and mantle xenoliths which were studied to historictimes with eruptionsrecorded in •-1440, 1585, 1646, obtainadditional information on magmaascent and storage. The 1677, 1712, 1949, and 1971 [Hernanddz-Pachecoand Valls, 1949eruption is alsoof interestbecause the erupted lavas show a 1982]. The island consistsof three main units: (1) the basal widerange in chemicalcomposition for a singleeruption (e.g., complex (3-4 Ma), which comprisesa Pliocene seamount basaniteand phonotephritewere simultaneouslyerupted from sequenceand a plutonic complex that have been uplifted and differentvents during one phase). tiltedto theirpresent position; (2) the oldervolcanic series (2 Ma to 0.7 Ma) which includes the Taburiente shield volcano, the 2. GeologicalSetting Bejenadoedifice, and the Cumbre Nueva series;and (3) the CumbreVieja series(0.7 Ma to present)which is confinedto the The CanaryIslands form a chainof sevenlarge volcanic southernhalf of the island [Hausen, 1969; Middlemost, 1972; islandslocated off the northwesternAfrican continental shelf Abdel-Monemet al., 1972;$chmincke, 1976; Staudigel and (Figure1). All islandsare underlain by oceaniccrust as indicated Schmincke,1984; Ancochea et al., 1994; Carracedoet al., 1999] by tholeiiticmid-ocean ridge basalt (MORB) gabbroxenoliths (Figure1). The north-southtrending Cumbre Vieja ridgeconsists occurringon Lanzarote,Gran Canaria, and La Palma[Hoernle, dominantly of mafic lava flows, cinder cones, and several 1998;Schmincke et al., 1998]. The ageof the crustis bracketed phonolitic plugs. Lava compositionsrange from basaniteto by paleomagneticanomalies S 1 (175 Ma), betweenthe eastern- phonolitewith clinopyroxene+ olivine+ kaersutiticamphibole as mostislands and the Africancoast, and M25 (155 Ma) between the majorphenocryst phases. Most historiceruptions were chem- La Palmaand Hierro,which are the westernmostand youngest ically and mineralogicallyzoned, commonly carrying mantle and islands[Roeset, 1982; Klitgord and Schouten,1986]. The crustalxenoliths during the final eruptive phase [Hernanddz- geologyof the CanaryIslands is summarizedby SchminckePacheco and Valls, 1982; Ibarrola, 1974; Praegel, 1986]. The [ 1976, 1982]. CumbreVieja is interpretedas a volcanicrift zone becausevents KLOGEL ET AL.: THE 1949 RIFT ZONE ERUPTION ON LA PALMA 5999
I Phreatomagmatic Firefountains xenoliths
Fissure eruption Llano del Banco •'
Stronglyphreatomagmatic ...... Ho,yo.N.e,gr.o.l.?, ! ? , . I...... June•6'28 30 July 4 6 8 10 12 14 '16 '18 '20 22 24 26 28 30 I Phase 1 I 2 ! 3 141
...... :•Tephrite group • Basunitegroup I I Phonotephritegroup
Figure2. Chronologyofthe 1949 eruption. Note the bimodality ofthe Duraznero and Llano del Banco eruptives, thesimultaneous eruption ofbasunite atLlano del Banco and phonotephrite atHoyo Negro, and the synchrony of events.Hoyo Negro produced a basanitic lapilli layer during its first days of activity and basanitic bombs during the finalphase, but further basunite occurrences inthe eruptive sequence areuncertain. The late basunite erupted from Duraznerocarried almost all xenolithsof the entireeruption.
and north-southtrending fissures,faults, and dikes are concen- whichrapidly descended toward the east coast. This final phase tratedalong its crest[Middlemost, 1972; Carracedo, 1994]. of theeruption terminated after 12 hours.
3. Chronology of the 1949 Eruption 4. Petrographyand Mineral Chemistry The shortsummary of the temporalevolution of the eruption On the basisof petrographyand wholerock chemistry,we given below is basedon works by Bonelli Rubio [1950], Romero dividethe volcanic rocks of the 1949eruption into three groups Ortiz [1951], and Martel San Gil [ 1960]; detailedaccounts of the (compareFigure 2): the tephritegroup comprises tephritic pyro- eruption chronology and volcanology are given elsewhere elasticsand lavas from Duraznero and Llano del Banco(June 24 [Kl•igelet al., 1999; Whiteand Schmincke,1999]. to July10); the basanite group contains basanitic lavas and pyro- Seismic precursors:The first weak seismic precursorswere elasticsfrom Llano del Banco, Hoyo Negro, and Duraznero (July 10-30); and the phonotephritegroup comprisestephritic to felt southof the Calderade Taburiente(Figure 1) from July 1936. Beginning in February 1949, strong earthquakesperiodically phonotephriticpyroelastics from Hoyo Negro (July 12-30).The HoyoNegro pyroelastics form <1% of thetotal erupted volume. shookthe whole island almostdaily. They were most intenseat the southerntip of the island,causing strong ground cracking. On 4.1. Tephrite Group June24, intenseseismicity and groundcracking in the proximity of MontafiaDuraznero marked the onsetof the eruption. Tephritescontain phenocrysts of clinopyroxene(3-5 vol %), Phase 1 (June 24 to July 8): Phreatomagmaticactivity from kaersutite(1-2 vol %), Ti-magnetite(1-2 vol %), andrare plagio- Duraznerocrater on the crestof CumbreVieja (1880 m abovesea clase,hauyne, and olivine. The groundmassconsists of glasswith level (asl)). Five vents openedsuccessively along a 400-m-long, plagioclaselaths (An53_56 [Kaiser, 1988])and clinopyroxene and north-southtrending fissure and eruptedtephritic pyroelastics. oxide microlites.Only a few, small xenolithsare present,the Phase2 (July 8-12): There was suddenopening of a fissureat mostcommon type being hauyne-bearing alkaline gabbros. Llano del Banco (1300 m asl) at the westernridge flank 3 km Clinopyroxenesare titaniferousaugites showing a wide com- north of Durazneroand cessationof activity at Duraznero(Figure positionalrange and variable, in manycases complex, zonation. 2). Llano del Banco issuedtephritic magma,producing a large Twoprincipal clinopyroxene types are recognized: type 1 (Mg# = lava flow which eventuallyreached the coast.Around July 10, 58-66) (Mg# = molarMg/(Mg+Fetot)x100) is greenishand Na- the lava becameless viscous,reflecting the arrival of basanitic rich and occursdominantly as euhedralto anhedralcores, some magmawhich makes up the bulk of the eruptedvolume. of whichare stronglyresorbed; type 2 (Mg# = 70-80) is pale Phase 3 (July 12-30): Accompanied by strong seismicity, brownishand occursas a mantlearound type I cores,as indi- Hoyo Negro (1880 m asl) opened500 m north of Duraznero.It vidualphenocrysts, or as groundmasscrystals (Table l a). The eruptedbasanite, tephrite, and phonotephritein violent phreato- green-coreclinopyroxenes are reverselyzoned with increasing magmatic explosions,while Llano del Banco simultaneously Mg numberand decreasingNa and Mn contentstoward the rim. issuedbasanitic lava at rates sometimesexceeding 14 m3/s. In mostcrystals, Ti stronglyincreases at the rim. Kaersutite(5-6 Eyewitness accounts indicate a close hydraulic connection wt % TiO2)is reverselyor irregularlyzoned with Mg# = 50-66in betweenthe two active vents,2.5 km apart in distanceand 580 m partlyrounded cores and Mg# = 69-71 at narrowlighter-colored apart in elevation.Eventually, the activitiesdeclined and ceased rims(Table lb). The crystalsare euhedral to mildlyrounded and entirelyon July 22 (Hoyo Negro) and July 26 (Llano del Banco) in some casesstrongly resorbed. Some kaersutitesshow fine- (Figure2). grained,polycrystalline rims or patches,indicating breakdown Phase4 (July 30): Explosive activity resumedat Duraznero reactions.Rare plagioclasecrystals distinct from microlites and Hoyo Negro (Figure 2). After 3 hours,Duraznero began to commonlyshow strong marginal and internal resorption erup/xenolith-richbasanite, producing lava fountains and a flow ("fingerprint"textures), are partly mantiedby newly grown 6000 KLf.)GELET AL.' THE 1949 RIFT ZONE ERUPTION ON LA PALMA
Tablela. RepresentativeClinopyroxene Analyses(Averaged) andCalculated Pressures andTemperatures of Tephritesand Basanites
TephriteSamples 1502-1 1502-2 1502-4 1502-51502-7 1502-8 1502-5 1502-6 1502-7 1502-4 1502-7 Type 2,rim 2,rim 2,rim 2,rim 2,rim 2,rim 2,core 2,core 2,core 1,core 1,core Points 4 4 3 3 3 3 6 4 1 8 4 SiO2 44.76 43.61 42.93 46.22 46.81 45.45 44.77 47.51 45.28 44.75 41.81 TiO2 3.90 4.55 4.57 3.49 3.12 3.79 3.21 2.54 3.01 2.74 4.02 A1203 7.67 8.65 8.95 6.79 6.19 7.19 8.08 5.25 7.80 8.13 10.!8 Cr203 0.01 0.04 0.03 0.02 0.01 0.03 0.07 0.03 0.10 0.02 0.02 FeO* 7.69 8.10 8.54 7.59 7.34 7.70 7.51 7.14 7.46 9.80 11.13 MnO 0.21 0.22 0.18 0.23 0.24 0.21 0.14 0.18 0.14 0.28 0.30 MgO 11.89 11.42 11.34 12.44 12.80 12.12 12.47 13.68 12.53 10.22 8.93 CaO 22.19 22.38 22.15 22.11 22.25 21.99 21.92 22.12 22.01 21.81 21.66 Na20 0.71 0.74 0.69 0.68 0.61 0.74 0.64 0.55 0.63 1.00 0.97 Sum 99.0 99.7 99.4 99.6 99.4 99.2 98.9 99.1 99.0 98.8 99.1 Mg# 73.4 71.5 70.3 74.5 75.7 73.7 74.7 77.3 75.0 65.0 58.9 T, øC 1122 1119 1119 1120 1113 1119 1170 1138 1169 1052 1054 P,MPa 749 690 677 721 644 709 932 531 926 798 794 BasaniteSamples 1507-1 1507-2 1507-3 1507-5 1507-6 !507-7 1507-11 1507-12 1507-14 1507-1 1507-7 Type 2,rim 2,rim 2,rim 2,rim 2,rim 2,rim 2,rim 2,core 2,core 1,core 1,core Points 4 4 4 4 4 4 4 14 5 2 10 SiO2 45.69 45.76 47.20 46.15 45.95 48.32 47.59 46.50 46.15 43.63 47.53 TiO2 3.41 3.15 2.63 2.67 2.86 2.31 2.35 2.36 2.69 3.20 1.52 A1203 7.07 7.50 6.18 7.17 7.50 4.85 5.79 7.18 7.01 9.38 5.32 Cr203 0.05 0.04 0.07 0.35 0.25 0.04 0.33 0.55 0.31 0.01 0.03 FeO* 7.85 7.88 7.54 7.37 7.50 7.64 7.23 6.78 8.05 9.19 11.29 MnO 0.12 0.13 0.15 0.12 0.11 0.16 0.13 0.12 0.13 0.23 0.49 MgO 12.28 12.42 13.34 12.93 12.73 13.90 13.58 13.38 13.28 10.19 9.96 CaO 22.38 21.96 22.07 22.17 22.30 21.93 22.03 22.06 21.70 21.68 21.60 Na20 0.53 0.58 0.48 0.50 0.53 0.44 0.46 0.52 0.49 0.94 1.27 Sum 99.4 99.4 99.7 99.4 99.7 99.6 99.5 99.5 99.9 98.7 99.1 Mg# 73.6 73.8 75.9 75.8 75.2 76.4 77.0 77.9 74.6 66.4 61.1 T, øC 1134 1141 1130 1132 1136 1116 1126 1223 1220 1052 1047 P,MPa 691 766 624 652 695 443 583 1090 1040 769 756 Mg# = molarMg/(Mg+Fetøt)x100. Seetext for details of the thermobarometric calculations.
plagioclase,and are interpreted asxenocrysts. Olivine occurs as thantype 1 andshows a gradualchange from FoB0 in thecore rarexenocrysts in schlieren and in a fewhybrid samples compo- towardFo84 maximum (60-300 gm) followed by a steepnormal sitionallybetween tephrites and basanites. zonationto Fo73_76at the rim (Figure3). Most olivinesare euhedral,and some contain spinel inclusions. 4.2. BasaniteGroup Thebasanites can be subdivided geochemically into the Llano delBanco, Duraznero, and Hoyo Negro subgroups, asshown in Tablelb. RepresentativeAmphibole Analyses of 1949 Tephrite section7.1. Basanites contain phenocrysts ofclinopyroxene (7-10 vol %), olivine(3-5 vol %), andminor Ti-magnetite. The Sample groundmassconsists ofglass with plagioclase laths (An60-67) and clinopyroxene,olivine, and oxide microlites. Partly resorbed 1502-1 1502-1 1502-3 1502-3 crystalsor glomerocrystsof plagioclase, amphibole, and large Type rim core rim core clinopyroxene(>2mm) are interpreted asdisaggregated wallrock SiO2 39.03 38.42 38.95 39.06 or xenoliths. TiO2 5.33 5.55 5.77 6.07 Clinopyroxenesare titaniferousaugites petrographically A1203 13.75 14.15 13.81 13.56 resemblingthose in tephrites.Similarly, greenish type 1 (Mg# = FeO* 10.54 12.93 10.25 11.26 60-66)and brownish type 2 (Mg# = 74-83)clinopyroxenes can MnO 0.19 0.16 0.07 0.19 MgO 13.66 11.86 13.46 12.24 bedistinguished, which differ from those in tephritesin having CaO 12.61 12.42 12.56 12.68 lowerTi andA1 and higher Cr contents,aswell as a slightly Na20 2.35 2.40 2.31 2.45 higherMg number (Table 1a). Many crystals, including reversely K20 1.04 1.16 1.16 1.17 zonedgreen-core clinopyroxenes, showa decreaseofMg number P205 0.15 0.10 0.13 0.15 andCr at theoutermost rim (<20gm); some clinopyroxenes F 0.19 0.14 0.18 0.15 CI 0.03 0.02 0.02 0.01 showcomplex zonation. Olivine occurs intwo varieties: type 1 is Sum 98.9 99.3 98.6 99.0 normallyzoned, having a core composition ofFo84 followed by a Mg # 69.8 62.1 70.1 66.0 steepdecrease to Fo73_76 atthe rim (20-90 gm); type 2 is larger KLOGEL ET AL.: THE 1949 RIFT ZONE ERUPTION ON LA PALMA 6001
4.3. Phonotephrite Group clinopyroxene-meltthermobarometer of Putirka et al. [1995] to clinopyroxenerim and melt compositionsyields 1110-1120øC This group comprisesphonotephrites to tephritesfrom Hoyo and 640-750 MPa (N=6). If appropriatetype 2 clinopyroxene Negro containingphenocrysts of amphibole(3-8 vol %), clino- cores(Mg # = 75-78) are related to the bulk rock composition, pyroxene (2-5 vol %), minor oxides, and rare apatite. The we obtain 1160-1170øC and 840-930 MPa (N=4) with two groundmassconsists of abundant plagioclase laths, clino- crystalsfalling outsidethis range (530 and 640 MPa) (Table la pyroxene, oxides, and glass. Most phenocrystsare strongly and Figure 4). zoned,with green-coreclinopyroxenes and kaersutiteswith light Greenish clinopyroxenes(type 1) are xenocrystic to the colored rims being particularly common. Many have corroded tephrite becauseof their low Mg # (58-66). If the relation of coreswith apatite,titanite, or hauyneinclusions resembling those Duke [1976] holdsfor evolvedcompositions, then this melt may in mafic and ultramafic xenoliths. Fine-grained, oxide-rich have had a Mg # of 22-29, which only matchesphonolite among reactionrims aroundkaersutites, in a few casesnewly overgrown, recentLa Palma rocks.Thus, to bracketthe crystallizationdepth are the rule in many phonotephritesbut are nearly absentin of greenishclinopyroxenes, we assigneda representativephono- others. Some kaersutitesshow sieve-like resorptiontextures. lite composition(Table 3) for the Putirka et al. [ 1995] thermo- Practicallyall rockscontain dark schlieren(up to 50 vol %) and barometer.This yielded 750-820 MPa (N=5) which adjoinsthe small (<5 mm) angular xenoliths and xenocrystsincluding rangefor brownishclinopyroxene rims. On the basisof their Mg olivine, titanite, apatite, and hauyne. Remarkably, the number, green-core clinopyroxenesmay have coexisted with phonotephritesonly contain fragmentsof mafic and ultramafic reverselyzoned kaersutitesfollowing the relation of Irving and cumulatesbut not of tholeiitic gabbros,in contrastto the xenolith Price [ 1981]. abundancesin the late Duraznerobasanite (see section4.4). 5.3. Basanite Thermobarometry 4.4. Xenoliths Following Duke [1976], clinopyroxenecoexisting with the Over 99% of the xenolithsof the 1949 eruptionoccur in the basanitemelt or the bulk rock (Tables 2 and 3) has Mg # = 73 terminalDuraznero basanite where they makeup almost1 vol % and Mg # = 81, respectively.Most brownishclinopyroxene rims [Klagel et al., 1999]. The xenolith suite consistsof (1) tholeiitic (type 2) haveMg # = 73-76 and appearto be in equilibriumwith gabbros(40% relative abundance)representing fragments of the the melt. The thermobarometerof Putirka et al. [1995] indicates Jurassicoceanic crust [Hoernle, 1998; Schminckeet al., 1998]; 1120-1140øCand 580-770 MPa (N=13); one phenocrystyields (2) alkaline hauyne- and titanite-bearinggabbros (35%) with 440 MPa. If the most magnesianclinopyroxene cores can be cumulustextures, interpreted as fragmentsof crustalintrusions; related to the bulk rock composition,then we obtain 1200- (3) olivine-bearingpyroxenites and kaersutitites(20%), inter- 1220øC and 910-1110 MPa (N-6) as first-order estimates. preted as cumulatesfrom mafic alkaline magmasrising beneath Greenishtype 1 clinopyroxenes,by analogywith the tephrite,are La Palma, which may have formed at both mantle and crustal xenocrysticand are interpretedto havegrown in a phonoliticmelt depths [Hansteenet al., 1998]; (4) basaniticfragments (5%), possibly at pressuresof 740-830 MPa (N=5) (Table l a and resemblingfragments of dikesor lava flows now buriedat depth; Figure 4). (5) at least five types of felsic, stronglyfused xenoliths(<1%) Using a melt Fe203/FeOratio of 0.30, the partitioncoefficient with syenitic,phonolitic and rhyolitic compositions,interpreted of Roederand Eroslie[ 1970] indicatesFo74 for olivinecoexisting as crustalfragments from variousdepths; and (6) spineldunites with the basanitemelt and Fo83for the bulk rock (Tables 2 and andharzburgites (<<1%) from the uppermantle [Kiwigel, 1998]. 3), which correspondswell with the phenocrystrim and core compositions,respectively (Figure 3). Olivine-meltthermometers 5. Thermobarometry with correctionsfor melt compositionand pressure[Putirka, 1997; Ford et al., 1983] indicate 1110-1130øC for the melt and 5.1. Redox State and Oxygen Fugacity 1190-1230øCfor the bulk rock at atmosphericpressure, and The bulk rock Fe203/FeO ratios of nine basanitesas highertemperatures for each 100-MPa pressureincrease (Figure determinedby FeO titration range from 0.34 to 0.69 (average 4). The temperaturesobtained for clinopyroxeneand olivine thus 0.47) overlapping with results from previous wet chemical correspondwell with each other. Accordingto Duke [1976] and analyses(0.30 to 0.55, average0.38) [Hernanddz-Pachecoand Irving and Price [1981], however, even the most magnesian Valls, 1982]. We calculatedthe oxygenfugacity (fo2) from the clinopyroxeneswere probably not in equilibrium with Fo84 redoxstate using the equationof Carmichaeland Ghiorso[ 1990] olivine (type 1 cores)but possiblywith FoB0(type 2), as hasalso for glassylava at atmosphericpressure. Fe203/FeO ratios from beennoted for the 1971 eruptionon La Palma [Praegel, 1986]. 0.30 to 0.69 yield fugacitiesof 0.8 to 2.6 log unit abovethe QFM Thereis a goodoverlap of the calculatedpressures for tephrite buffer for the bulk basanite.Since the magmashave undoubtedly and basanite,suggesting that both fractionatedat-600-800 MPa been oxidized during ascent and eruption, we consider and possibly at higher pressures.This is in accordancewith Fe203/FeO= 0.30 andfo2 = QFM+0.8(e.g., 10 -9 bars at 1100øC) clinopyroxenepreceding plagioclase as the fractionatingphase as the bestapproximation for the intratelluricmagma. The redox [Fisk et al., 1988]. The data are supportedby CO2-dominated stateof the 1949 tephrite is within the range for the basanite fluid inclusionsin phenocrystsof the basaniteshowing a bimodal [Hernanddz-Pachecoand Valls, 1982] yieldinga similaroxygen densitydistribution: Minimum pressuresof inclusionentrapment fugacity(QFM+ 1). are 600-680 MPa (upper mantle) for olivine and 200-340 MPa 5.2. Tephrite Thermobarometry Accordingto the.relation of Duke [1976], clinopyroxene•Supporting data tables are available on diskette or via AnonymousFTP coexistingwith the tephrite melt or the bulk rock (Tables 2 and3) from kosmos.agu.org,directory APEND (Usemame = anonymous, Password= guest).Diskette may be orderedfrom AmericanGeophysical hasMg # -- 73 andMg # = 76, respectively,which is withinthe Union, 200 FloridaAvenue, N.W., Washington,DC 20009 or by phone range for brownishtype 2 clinopyroxenes.Applying the at 800-966-2481;$15.00. Paymentmust accompany order. 6002 KLOGELET AL.:THE 1949RIFT ZONE ERUPTION ON LA PALMA
Table3. XRFWhole Rock Data of Representative 1949 Samples
Analysis 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Locality Dur Dur Dur Dur Dur LdB LdB LdB LdB HN HN HN HN HN Rocktype T T T/B B B T T B B B B B PT-A PT-A $iO2 45.43 45.22 44.73 43.41 43.76 45.29 45.31 44.06 43.95 44.12 42.42 43.54 48.75 48.67 TiO2 3.41 3.39 3.39 3.46 3.42 3.39 3.36 3.39 3.53 3.48 3.40 3.62 2.52 2.59 A1203 16.27 16.05 15.01 13.64 13.70 16.08 16.07 13.76 14.40 13.91 11.77 12.82 17.60 17.06 Fe•O3 11.89 11.65 12.42 13.62 13.70 11.65 11.50 13.61 13.26 13.80 13.57 13.44 9.43 9.59 MnO 0.21 0.21 0.21 0.20 0.20 0.21 0.21 0.20 0.20 0.20 0.19 0.21 0.21 0.21 MgO 4.93 4.91 6.40 7.97 8.18 4.96 4.90 8.23 7.36 8.03 11.76 9.84 3.37 4.04 CaO 9.17 9.09 9.76 10.63 10.69 9.12 9.02 10.53 10.14 10.85 10.64 10.97 7.14 7.14 Na•O 5.08 5.24 4.68 3.76 3.76 5.48 5.46 3.77 3.89 3.64 3.06 3.29 5.73 5.75 K•O 2.29 2.32 1.89 1.26 1.28 2.31 2.36 1.32 1.53 1.31 1.39 1.42 2.88 2.92 P•O5 0.90 0.87 0.81 0.72 0.70 0.85 0.89 0.69 0.75 0.69 0.69 0.84 0.76 0.74 $03 0.08 Majorelements inweight percent; trace elements inppm. Trace element dataof selected samples analyzed byICP-MS are indicated byasterisk. B, basanite;T,tephrite; PT,phonotephrite; Phon,phonolite; Dur,Duraznero; LdB,Llano del Banco; HN, Hoyo Negro; LLD, lower limit of detection; Mg # = molarMg/(Mg+Fetøt)x100. Samplesinanalyses areas follows: 1,KLA1502, scoriaceous layerbetween lithics atsouthern Durcrater rim; 2, KLA1519, chilled bomb 150 m SW ofDur crater; 3,KLA1521, welded spatter 100m west ofDur crater; 4,KLA1507, latescoriaceous lapJillfallout around Durcrater; 5,WHJ9224, lava fromlate Dur flow; 6, KLA1203, glassy crust ofearly LdB lava at 1100 m asl; 7, KLA1215, early LdB lava at 1270 m asl; 8, KLA1210, terminal aa lavafrom LdB; 9, KLA1213, scoriaceous lapillinear LdB vent; 10, KLA1508, scoriaceous lapillilayer within HN fallout 500 m NNW of crater; 11, KLA1513,chilled bomb 150 m NW of HN crater; 12, KLA1516, chilled bomb atHN crater rim; 13, KLA1512, cauliflower bomb 150 m NW of HN crater;14, KLA1514, late bomb 150 m NW of HN crater; 15, WHJ9254, breadcrust bombfrom ridge north of HN; 16, KLA1509, bomb 100 m west of HNcrater; 17,HUS251921, quenched bomb200 m NNW of HN crater; 18,KLA1517, chilled bomb atHN crater rim; 19, KLA1624, pahoehoe lava from1712 eruption; 20,KLA1613, prehistoric phonolite dome near Jedey village. KLOGEL ET AL.' THE 1949 RIFT ZONE ERUPTION ON LA PALMA 6003 Table 3. (continued) 6. Zoning and Reaction Rates of Phenocrysts 6.1. O!ivine PhenocrystZonations Analysis In order to assessthe magma ascentdynamics of the 1949 15 16 17 18 19 20 eruptionwe used diffusion kineticsto model the zonationsof Locality HN HN HN HN 1712 prehist. olivine phenocrysts.Because the effectsof simultaneouscrystal Rock type PT-A PT-B PT-B PT-B B Phon. growthcannot be quantifiedprecisely, we neglectthem and only SiO2 47.14 46.03 45.38 45.41 42.60 54.36 calculatediffusion times to constrainan upper limit for the actual TiO2 3.13 3.16 3.09 3.52 3.88 1.12 time to producethe zonations.We focuson Fe-Mg interdiffusion A1203 15.76 17.08 16.13 15.32 12.74 20.21 in olivine because it is well known and little affected by Fe203 11.18 10.32 10.77 12.08 13.84 4.75 simultaneousdiffusion of trace elements [Chakraborty, 1997] MnO 0.21 0.24 0.24 0.23 0.20 0.17 3.57 6.18 8.81 0.86 usinga diffusioncoefficient DFeMg of 1.3x10']6 m2 s-1 (see MgO 5.51 3.30 AppendixB). CaO 8.50 8.84 8.89 9.01 11.52 3.91 Na20 5.19 6.01 6.07 4.89 3.60 8.42 Normal Fe-Mg zonationsof olivine phenocryststoward the K20 2.49 2.70 2.69 2.38 1.62 4.18 rims are typically 20-90 [tm wide, with some of this variation P205 0.83 1.07 1.14 0.89 0.89 0.23 reflecting oblique cutting in thin section and anisotropy of SO3 0.06 < LLD < LLD 0.11 < LLD 0.35 diffusivity.Calculated diffusion times range from 2 to 45 days H20+CO2 0.36 0.27 0.31 0.42 0.39 0.23 (see Appendix B). Although Fe, Mg, Ni, and Ca have distinct 98.3 100.4 100.1 98.8 Total 100.4 99.0 diffusivities,their zonationwidths are similar (Figure 3), and we Mg# 49.4 38.8 39.6 50.3 55.8 26.4 concludethat crystalgrowth dominatedover diffusionand that the rims grew within a few daysat most.The similarityof zona- v 221 219 209 257 327 80 Cr 115 78 0.3 (lowercrust) for clinopyroxene(Figure 4), which is interpretedto 76 reflect two levels of magmastagnation during ascent[Hansteen 0.2 et al., 1998]. The rangefor olivinethus overlaps completely with 74 0.1 that obtained by clinopyroxene-meltbarometry, whereas the 72 rangefor clinopyroxeneoverlaps with thatfor fluid inclusionsin 70 t I I I I I I I I variousxenoliths in the 1949 basanite[Hansteen et al., 1998]. 0 50 100 150 200 250 300 350 JITI The data suggestthat the clinopyroxene-meltbarometer and Figure 3. Microprobe traversesthrough representative type 1 olivine fluid inclusionscould preservemantle pressures because and type 2 olivine phenocrystsin basanite.Note the similar magma stagnationat crustal levels was too short for re- compositionsof type 1 coresand type 2 maximaand their similar equilibration. rim zonationsand compositions. 6004 KLOGEL ET AL.: THE 1949 RIFT ZONE ERUPTION ON LA PALMA Table 2. ElectronMicroprobe Analyses of Basaniteand Tephrite defining a narrow compositional range. Dark schlieren and Glasses Used for Thermobarometric Calculations olivine crystalsin sometephrites reflect minor mixing duringthe i ,! transitionfrom tephrite to basaniteeruption. When plotted as a TephriteKLA1502 BasaniteKLA1507 time series,however, the tephritesand basanitesshow gradual Mean (5 Mean variationsof trace element concentrationsindicating a chemical zonationof the magmas(Figure 7). This is particularlystriking SiO2 50.11 0.28 44.91 0.20 for the basaniteK20/P2Os, Rb/Ba, andZr/La ratioswhich are not TiO2 2.57 0.14 4.02 0.13 A1203 18.85 0.16 15.24 0.12 fractionatedby the observedphenocryst phases. The dataindicate FeO* 7.43 0.18 11.76 0.13 subtle but systematic geochemical differences between the MnO 0.22 0.05 0.22 0.05 basanitesfrom the three eruptivecenters of 1949: from Llano del MgO 2.76 0.07 4.32 0.06 Banco to Duraznero to Hoyo Negro, the concentrationsof CaO 6.21 0.16 9.75 0.08 incompatibleand some compatibleelements increase and ratios Na20 6.68 0.08 4.94 0.05 K20 3.45 0.09 1.90 0.05 suchas K20/P20 5, Rb/Ba, and Zr/La decreaseslightly (Figure 6). P205 1.07 0.05 1.18 0.04 The basanitesfrom Hoyo Negro show larger compositional SO2 0.05 0.03 0.09 0.04 heterogeneitieswhich partly reflect phenocrystaccumulation F 0.24 0.04 0.22 0.05 (e.g., -25 vol % olivine and clinopyroxenecause high MgO, Ni, CI 0.14 0.02 0.06 0.02 andCr of sampleKLA 1513). Total 99.8 0.4 98.6 0.3 The phonotephritegroup of Hoyo Negro is compositionally mg# 39.8 0.7 39.7 0.5 heterogeneousand forms two distinctsubgroups: Phonotephrite A followsthe compositionalpath definedby otherrecent Cumbre In weightpercent. For tephrite N=55 and for basanite N=29, where N is numberof analyses. Vieja rocks,albeit at higher concentrationsof somecompatible and incompatibletrace elements(e.g., Cr, Ni, Zr, La), whereas phonotephriteB divergesfrom this pathextending to lowerMgO and Ni and to higher concentrationsof incompatibleelements at diffusionmodel (see AppendixB). On the basisof the above similarSiO 2 contents(Figures 5 and 6). There is alsoa tendency arguments,we concludethat the zonationsformed during a for kaersutite breakdown reactions to be more common and significantlyshorter period of a few weeksto months.The intense in phonotephriteA than in phonotephriteB samples. reversezonations of type 2 olivinesimply an increaseof the Somephonotephrites have trace element concentrations similar to Mg/Fe2+ ratio of thehost melt, which most likely reflects mixing far more evolvedCumbre Vieja rocks:For example,our unpub- with a moreprimitive melt. lisheddata from eight phonolitesrange from 562 to 872 ppm for 6.2. AmphiboleBreakdown Zr (660, maximafor Hoyo Negro), 110 to 204 ppm for Nb (203), and 101 to 149 ppm for La (179). It is mostremarkable that Hoyo Fine-grained,oxide-rich rims aroundkaersutite in some Negro erupted some of the most mafic as well as the most tephritesand most phonotephrites and the lack of suchrims along evolvedmagmas of the entireeruption, providing the highestand cleavagecracks indicate amphibole breakdown resulting from the lowestconcentrations of mostmajor andtrace elements. reaction with the melt which reduced its dissolved water content Overall, from basanitesto phonotephrites,Th/U decreases duringascent, rather than from surface oxidation. The width of systematicallyfrom 3.9 to 2.8, Rb/Ba increasesfrom 0.05 to 0.10, mostreaction rims is -10-40 •tm in phonotephritesand up to 10 and La/Sm increasesfrom 5.7 to 11.4, whereasSrn/Yb remains •tm in tephrites.Experiments with daciticmagma show that constantat 5.0ñ0.2. With few exceptions,samples of the reactionrims of similarthickness form within daysto weeksonce phonotephritegroup have Zr/Nb ratiosof 3.3 ñ 0.1 clearlybelow amphibolebecomes unstable due to decompressionand melt thosefor the tephriteand basanitegroups. All 1949 magmasare degassing[Rutherford and Hill, 1993].Other conditions being stronglyenriched in the light relativeto the heavy rare earth equal,the reaction rates increase with increasing temperature and elements(REE), showingmildly concaveupward shaped spectra decreasingmelt viscosity (M. Rutherford,personal communica- without notable anomalies (Figure 8a). Low Rb/Nb, Ba/Nb and tion, 1998). Althoughthese experiments cannot be directly Ba/La ratiosand troughs for K andTi in normalizedincompatible appliedto kaersutitebreakdown in tephritesand phonotephrites, element patterns (Figure 8b) are consistentwith derivationof the the resultsprovide a roughestimate for the timescaleinvolved. magmasfrom a HIMU-type source[Weaver, 1991] (HIMU - We thuspropose that the observedreaction rims formed within high238U/204pb). daysto weeksonce kaersutite became unstable due to magma Much of the majorand trace element variation of the 1949 and degassing. other CumbreVieja lavas is consistentwith removalof olivine, clinopyroxene,kaersutite, and minorTiomagnetite, which are the majorphenocryst phases and also the major'constituents(>98 vol 7. Major and Trace ElementChemistry %) of ultramafic cumulatexenoliths abundanton La Palma. The low MgO, Ni, and Cr concentrationsof eventhe mostprimitive 7.1. Compositionof EruptiveProducts rocks indicatethat the magmashave undergonesignificant The volcanicrocks from the 1949 eruptionshow a wide range fractionationof olivine and clinopyroxene.Clinopyroxene in chemicalcomposition (Table 3), butmost are within the range phenocrystswere fractionatedover the entire compositional of other recent Cumbre Vieja lavas. Llano del Banco and range,causing a decreasein CaO, CaO/A1203,and Cr from Durazneroeach produced lava flows and pyroclastics with similar basanitesto phonotephrites(Figure 6). Kaersutitebecomes a butessentially bimodal chemical compositions (Figures 5 and6). majorfractionating phase above -•45 wt % SiO2as indicatedby Botherupted tephrites (4.8-6.5 wt % MgO,44.8-46.3 wt % SiO2) minorkinks for Na20/K20 andTiO2 in Figure6, wherethe latter andbasanites (7.4-8.3 wt % MgO, 44.0-44.7wt % SiO2)each alsoreflects removal of Ti-magnetite.The fractionationof both KLOGEL ET AL.' THE 1949RIFT ZONE ERUPTION ON LA PALMA 6005 P [MPa] Depth [km] 0 0 Isochores: 0.50 Cpxin basanite 10 0.65 Moho m m m m mm m m m m m m m m m m m m m m mmmmmmmmm 5OO Isochores: OIin basanite 0.85 2o 0.88 Green Core Rim 3o Cpx? lOOO • I"l meph.Cpx 4- • I•l Bas.Cpx • •-",'-Bas. 01 4o :.:.:.• Fluid incl. I 900 1000 11 o0 1200 1300 T [øC] Figure 4. P-T-diagramshowing thermobarometric data for magmasand xenolithsof the 1949 eruption(Cpx, clinopyroxene-meltthermobarometer; O1, olivine-meltthermometer; Teph, tephrite;Bas, basanite).Mineral ther- mobarometerswere applied to phenocrystrim + melt (open symbols)and phenocrystcore + whole rock (solid symbols) compositions.Fluid inclusiondata for basanitephenocrysts (dotted) are from Hansteenet al. [1998]; numberson isochoresindicate inclusion density in gcm-3. The depthof the Moho is afterRanero et al. [ 1995]. Thick linesare solidi for peridotite+ CO2 + H20 [OlaJ•sonand Egglet, 1983] and dry peridotite[Takahashi and Kushiro, 1983]. I ' I ' I ' I ' I ..... 2, I ' I ' I ...... -•*.-•.••jd?:•;!•!•:•.,•..::.:,.:,:,:....•-,...... ß -...... -•-*------•---•------••::•••..••••i•-:- ,41 0 Tephntegroup ...... •a>•.•'"•' ""•fi"'"-••l••'•••••;•;'.•. • LdB. + Dur. basanites ...... -' a•l,,.:..•.•,•.;t:.•.:•:•:i•.•::.:,::.:•:.•.••:•%%--•.:•.:' :.:''•:•:-.-x:.•:'-•.•:•:;;•:•:•, ß•..•;•:-• ß H.N.Phonotephrite A ...... "• ••••'••""•:"•'"••• • H.N. Phonotephnteß B ...... "**•a;•,:,,•,F-•...... _,;,O•:,•-_-•a*-•"-'•,•,:;•N•,>.,..:,.**:.***-•-:•-•-•x•,:..----*-:•., , ...... •*'•::.n:':•-*l•:..•;e...:'...... ••••:••••:, , '-" 10 ...... Phono ...... -' / --•'•&':•j,•'- ...... •:•••.•••• ...... ::.•...:,,, ...... - -,, ,., - - . .•• .."•:•. E,--•.• •- , ..... ,, . -, - '* ...... "• ::•' '•'r' .. '"-, ß ...... ::: ...... ,:...... - '•--•'•"::•I m ...... I :' •Basalt_ iBasalticandesit.e 1 42 44 46 48 50 52 54 56 sio2 [wt%] Figure5. Totalalkalis versus silica (TAS) diagram with field boundaries from Le Maitreet al. [ 1989]showing the compositionalrange of the 1949eruptives as comparedto otherrecent lavas and intrusions of the CumbreVieja ridge(shaded field; data from Hernanddz-Pacheco and Valls[1982], Hernanddz-Pacheco and De la Nuez[1983], Elliott [1991],and our unpublished data, 1999). The tephritesfrom Llano del Banco(LdB) andDuraznero (Dur) showcomplete overlap and therefore are denoted with the samesymbol. Note the pronouncedbimodality of the tephriteand basanite groups (fields with solid lines, 77 samples),the small compositional range of bothgroups, and the widecompositional range of the phonotephritegroup (fields with dashedlines, 15 samples)from Hoyo Negro 6006 KLOGEL ET AL.: THE 1949 RIFT ZONE ERUPTION ON LA PALMA 12 Ni 10 Ti02 4 2.8 0.90.8f •:•'-":-%:--- CaO/AI203 .... o.5I- ' 0.4• -•.••: Sr , I , I t 1.2 '"& &',A 1.1 1 0.9 0.8 0.7 0.6 42 SiO2 Figure 6a. Variation diagramsof the 1949 eruptivesand other recentCumbre Vieja lavas (symbolsand data sourcesas in Figure 5). The tephriteand basanitegroups are enclosedwith solid lines, and Hoyo Negro phonotephritesA and B are enclosedwith dashedlines. Error barsshow 2c• analytical precision and are omitted wheresmaller than symbol size. Concentrations are in weightpercent for oxidesand parts per million for elements. The Harkerdiagrams show distinct fields for tephritesand basanites, with few samplesof intermediatecomposition, and well-definedphonotephrite subgroups (A and B) of Hoyo Negro.Arrows indicate high Cr and Ni contentsof sampleKLA1513 beingoutside the plot range. La 160 120 80 .--•;..::...•-g'7-[-. ß.'•.... ?...... ' •---':-'.:-:-::•.,,li•'--..* Syenitesfrom . :.!;..]--:-•.:•'n:,,.-_'•-"' basalcomplex -: 40 ' ]''' [''' t,,, [,,, 44 48 52 56 60 64 SiO2 Figure 6b. Plot of La versusSiO 2 indicatingthe high concentrationsof incompatibleelements in Hoyo Negro phonotephritesas comparedto highlyevolved La Palmarocks. KLOGEL ET AL.' THE 1949 RIFT ZONE ERUPTION ON LA PALMA 6007 ' I ' I ' I ' I ' I ' Rb/Ba least squaresmass balance for the major elements,and leaves q- large residualsfor Sr and Ba (Table 4). Plagioclasefractionation 2o is ruled out by the absenceof negativeEu and Sr anomaliesin all 0.09 samples(Figure 8), by increasein CaO/A1203 and Sr with Teph increasingSiO 2 (Figure 6), and by our barometry[cf. Fisk et al., 0.08 1988]. Kaersutitefractionation would causeincreasing Zr/Ba and decreasingK20/P205 from basaniteto tephrite,opposite to what 0.07 is observed(Figure 6c). Hypothesis2 is unlikely becauseno evolved La Palma rock, including felsic xenoliths (our 0.06 unpublisheddata, 1999), can producethe tephriteby mixing with 0.05 , I , I , I , I , I , I , I the basanite(compare Figures 5 and 6b). 1.2 1.6 2 2.4 2.8 3.2 3.6 4 We further tested whether another historic La Palma basanite K20/P205 could have producedthe 1949 tephriteby fractionalcrystalliza- tion. Surprisingly,mass balance calculationsfor these models ' ' ' I ' ' ' I ' ' ' I ' ' ' I ' ' ' I ' Rb/Ba yield generallybetter results than that for the 1949 basaniteas a ---+-- / A parent(sum of squaredresiduals (SSR) _<0.01, using data from 2o , ß 0.09 Elliott [1991] and Table 3; one model is shown in Table 4). ,•, A' Nevertheless,large residualsremain for Sr, Ba, and Zr (>100 0.08 0.07 [ B:'•'-•,••Teph Bas Tephrite Basanite Phase ! ' Phase 2 Phase3 ' Phase4 ..- .... '.':;.-.,.-.'...... 'b (14 days) 0.06 m Ce [ppm] ', 200 0.05 / , , , I , , , I , , , I , , , I , , , I , 0.48 0.52 0.56 0.6 0.64 0.68 180 Zr/Ba øøøo0"ø<>ooOo8,, 160 Nb/Zr I ' I ' I ' I ' I ' I ' I 140 0.3 2• ! _ A ...... :' ...... •' " 120 0.28 Zr/Ba m 0.26 0.68 217 . 2o Tepht [][]', 0.24 [] , 0.64 0.22 i + •___•Bas [] I , I [ I [ I [ I i I [ I 0.60 ! ! 2.6 2.8 3 3.2 3.4 3.6 3.8 ,, Sr/Zr 0o•oo o X 0.56 000 ,, O00• ' Figure6c. Ratioplots of incompatibleelements showing almost Rb/Ba m i no overlapof thefields for thetephrite (Teph) and basanite (Bas) ! groups.The Hoyo Negro phonotephrites A and B formirregular ! oo 08 ! fieldswith no apparentend-members and do not plot on a line O O O O withthe tephrite or basanitegroups. The lateDuraznero basanites 0.070 (shadedsquares) plot intermediatebetween the Llanodel Banco andHoyo Negro basanites. 0.065 : aaE!! [] [] 0.060 217 clinopyroxeneand kaersutite is consistentwith the depletionof 0.055 the middle relativeto the light and heavy REE and increasing 10 15 20 25 LaJSmratios from basaniteto phonotephrite. Sample sequence 7.2. RelationshipBetween Tephrites and Basanites Figure7. Systematiccompositional variations of tephritesand Theclearly distinct fields for the tephrite andbasanite groups basanites during the course ofthe eruption (error bars show 2o analyticalprecision). The late Duraznerobasanite from eruptive (Figures5 and 6) suggest thatthe two magmas arenot directly phase 4 (shaded squares) initially tends more tothe Hoyo Negro relatedtoeach other but represent distinct batches. Alternative (solid circles) and later to the Llano del Banco (open squares) hypothesesto produce the tephrite are (1) fractionalcrystal- basanite compositions. Thediagram is basedon 38 samples lizationand (2) mixingof basanitewith an evolvedmagma or unequivocallyrepresenting a time series as based on field studies wall rockassimilation. Hypothesis 1 requires---7 wt % of plagio- andthe eruptionchronology. Note thatthe sequenceof samples claseor kaersutitein the fractionatingassemblage, as shownby shownis notlinear in time(compare Figure 2). 6008 KLOGELET AL.' THE 1949 RIFT ZONE ERUPTION ON LA PALMA •,, ]OTephritegroup (#KLA1502) 7.3. Melting Depths 500 t•'x•,,_ I n LlanoBancobasanite (#KLA1210) %N•_ I mHoyo Negrobasunite (#KLA1513) The strong enrichment ofthe light and middle relative tothe • •x,, I A PhonotephriteA (#KLA1514) heavyREE of the1949 and other Cumbre Vieja lavas (Figure 8) •-x•%x,• I v PhonotephriteB(#HUS251921) indicates residual garnet inthe mantle source, which requires I-- St.Helena HIMU basalt depthsofmelt extraction inexcess ofabout 80 km if thesource lOO wasgarnet peridotite [e.g., Hirschmann andStolper, 1996]. A 50 contentscomparisonpredictedof the SiO2from and dryFeO* melting contentsexperimentsof erupted [Hiroselavas withand Kushiro,themagmas 1993]were suggestsderived meltingfrom a lherzolitedepthsofsource>100 (Figurekm (3GPa)9). The if possiblepresence of recycledoceanic crust in the form of pyroxenitein the source,however, as suggestedby the HIMU- lO typetrace element composition of the magmas,could significantlyaffect the SiO2 andFeO* contentsof theprimary • • • • • • • • • • • • • • • melts. Nevertheless,melting experimentson mixtures of La Ce Pr Nd Sm Eu Gd Tb Dy Ho Er Tm Yb Lu peridotiteand MORB [Kogisoet al., 1998], which could Figure8a. Rareearth element (REE) elementdiagram of repre- representrecycled oceanic crust, also yield compositionsthat are sentative1949 volcanics (ICP-MS data)normalized to average consistentwith meltingdepths of at least80-100 km for the 1949 chondrite[Boynton, 1984]. The datafor a primitiveHIMU basalt andother La Palmamagmas (Figure 9), placingthe source within from St. Helenaisland with 13.7 wt % MgO [Chaffeyet al., or below the baseof the lithosphere.The uniformSr-Nd-Pb 1989]are shown for comparison(HIMU = high238U/204pb). The isotopiccomposition of all historicbasunites [Elliott, 1991; K. concentrationsincrease from basunitesto phonotephriteswith strongenrichment of thelight relative to theheavy REE ranging Hoemle,unpublished data, 1999] suggestsderivation from a from220 (basunites)to 290 (tephrites)and 380 (phonotephrites). common source. The REE spectraare fairly smoothand have a mild concave upwardshape becoming more pronounced in the evolvedlavas. 7.4. Origin of the HoyoNegro Phonotephrites In mostvariation diagrams the HoyoNegro phonotephrites A andB appearto representmixing or crystalfractionation trends betweenthe most extremecompositions combined with some ppm in most cases).Although there are uncertaintiesin the contaminationof the magmasas petrographicallyobserved. In partitioncoefficients used, the large residualsdo not indicate ratioplots of incompatibleelements, however, these possibilities derivationof the tephritefrom a historicbasunite under closed- are ruledout becausethe formerend-member compositions in systemconditions. Radiogenic isotopes do not resolve this Figures5 and6a do not forma trendbut plot amidirregular questionbecause all historicbasunites from La Palma,as well as fields(Figure 6c). The relationshipsshown in Figure6 further the1949 tephrite, have identical Sr-Nd-Pb isotopic compositions indicate that neithergroup resembles a binarymixing line withinthe analytical precision [Elliott, 1991; K. Hoernle,unpub- betweenthe tephriteor basunitegroups and evolved rocks from lisheddata, 1999]. La Palma,such as phonolitesand syenites.In addition,neither We concludethat the tephriteand basunite groups represent twodistinct magma batches that evolved and ascended separately anderupted subsequently. This scenariois consistentwith very 500 -- _ limitedintermixing of the magmasdespite low melt viscosities (-•150and 18 Pa s,respectively, as calculated after Shaw [1972]). The distinctcompositions of the batchesat leastin partreflect sourcevariations such as degreeof melting[cf. Elliott, 1991], resultingin differentsilica activities and fractionation paths. The observedzonations within the tephrite and basunitebatches (Figure7) are consistentwith compositionalvariations of the primarymelts combined with incompletehomogenization after mixingevents in the mantle. The compositionaldifferences between the basunitesfrom the threeeruptive centers are mosteasily explained by different I I I I I I I I I I I I I I I I I I I I I I degreesof melting(lowest for HoyoNegro, highest for Llanodel Rb Th Nb K Ce Nd Zr Eu Gd Dy Yb Banco)of a commonsource, likely with phlogopitein the Ba U Ta La Sr Sm Hf Ti Tb Y Lu residuumfor all basunitesto fractionateRb/Ba and K20/P20 5. Figure 8b. Incompatibleelement diagram normalized to primi- The Duraznerobasunites have compositionsintermediate tive mantle[Hofmann, 1988] (ICP-MS dataexcept for Ti, K, Sr, betweenthe Llano del Banco and Hoyo Negro basunites (Figure andZr). The dataclosely resemble the patternfrom a St. Helena 6) andtherefore could have been produced by (1) intermediate HIMU basalt[Chaffey et al., 1989] characterizedby a markedK troughand low Rb/Nb,Ba/Nb, and Ba/La ratios [Weaver, 1991]. degreesof meltingor (2) shallowmixing of the two end- The pronouncedtroughs for Ba, Sm, andTi andthe relatively members.The latter is consistentwith massbalance calculations highU/Th ratioof themost evolved phonotephrite A sample are andcould explain the Duraznero basunites initially tending more consistentwith kaersutiteand apatite fractionation,whereas tothe Hoyo Negro and later to theLlano del Banco compositions assimilation of 1 wt % of apatitecan produce the relativelylow (Figure7). U/Th ratioof theextreme phonotephrite B sample. KLf2GELET AL.' THE 1949RIFT ZONE ERUPTION ON LA PALMA 6009 Table 4a. Resultsof FractionationCalculations Based on LeastSquares Mass Balance for Major Elements Model Bas-Teph1 Bas-Teph2 Bas-Teph3 Bas-PhonA Bas-PhonB Parent Dur basanite Dur basanite 1712 basanite HN basanite HN basanite KLA1507 KLA1507 KLA1624 KLA1516 KLA1516 Daughter Dur tephrite Dur tephrite Dur tephrite PhonotephriteA PhonotephriteB KLA 1502 KLA 1502 KLA 1502 WHJ9254 KLA 1517 Fraction wt % Olivine 6.3 4.2 3.9 4.6 4.6 Clinopyroxene 18.2 15.1 18.3 15.8 16.0 Kaersutite - 7.4 10.4 17.3 11.6 Plagioclase 6.6 Ti-magnetite 3.9 2.5 4.3 4.1 3.0 Apatite - - 0.7 0.4 - Residual melt 64.9 70.8 62.4 57.9 64.9 Bas-Tephi Bas-Teph2 Bas-Teph3 Bas-PhonA Bas-PhonB Calc Resid Calc Resid Calc Resid Calc Resid Calc Resid SiO2 44.20 -0.05 44.13 -0.12 42.96 -0.01 43.44 0.10 43.44 0.10 TiO2 3.35 -0.14 3.60 0.10 3.68 -0.19 3.64 -0.02 3.66 -0.04 A1203 13.77 0.01 13.49 -0.27 12.86 0.14 12.62 0.20 12.61 0.21 Fe203 13.76 0.02 13.58 -0.16 13.87 0.05 13.36 0.08 13.35 0.09 MnO 0.25 0.05 0.25 0.05 0.21 0.01 0.21 0.00 0.23 -0.02 MgO 8.04 0.00 7.98 -0.07 8.84 0.04 9.80 0.04 9.80 0.04 CaO 10.76 0.03 10.67 -0.05 11.57 0.06 10.96 0.01 10.96 0.01 Na20 3.78 -0.01 3.96 0.17 3.57 -0.04 3.56 -0.27 3.58 -0.29 K20 1.51 0.24 1.71 0.44 1.58 -0.04 1.70 -0.28 1.72 -0.30 P205 0.58 -0.14 0.65 -0.07 0.86 -0.03 0.69 0.15 0.59 0.25 SSR 0.10 0.36 0.07 0.23 0.30 Cr 36 0 63 27 36 0 51 -64 63 -64 Ni 17 -29 33 -13 18 -28 23 -54 39 -45 Rb 52 -4 47 -9 58 2 54 -16 49 -9 Sr 1202 -223 1241 -184 1625 179 1682 73 1636 18 Zr 405 -22 370 -57 469 42 553 4 502 7 Nb 119 0 108 -11 136 17 159 -5 144 -6 Ba 683 -103 616 -170 890 104 939 -1 850 1 La 100 -9 92 -17 145 36 152 -3 143 5 Ce 195 -20 177 -38 226 11 267 29 252 12 Nd 88 5 79 -4 90 7 108 24 103 16 Major elementsin wt %; traceelements in ppm. Bas,basanite; Teph, tephrite; Phon, phonotephrite; Dur, Duraznero;HN, HoyoNegro; SSR,sum of squaredresiduals; Calc, calculated; Resid, residual. See Table 3 for the compositionsof parentaland daughter magmas. the tephritenor the Llanodel Bancoor Duraznerobasanites can Table4b. PartitionCoefficients Used in Calculations have producedthe phonotephritesby crystal fractionation becausemass balance calculations using phenocrystand whole Olivine Clino- Kaersutite Plagio- Magnetite Apatite rock compositionsyield poor resultsfor major elements pyroxene clase (SSR> 0.6) andhigh residuals for manytrace elements (although increasedconcentrations of some incompatible trace elements Cr 0.9 8 0.34 0 15 0 suchas REE, Zr, andSr may partly reflect contamination as RbNi 100 0.0021.5 0.21.6 0.070 290 0 indicatedbytitanite xenocrysts insome Hoyo Negro rocks). Sr 0 0.13 0.4...0.8 1.8 0 5.1 Mass balancecalculations, however, show that the most Zr 0 0.12 0.18 0.05 0.1 0.1 primitive phonotephriteA and B compositionscan be derivedby Nb 0 0.01 0.15 0.01 0.1 0 crystal fractionation from relatively primitive Hoyo Negro Ba 0 0.0002 0.4...0.8 0.23 0 0.7 basanites(Table 4; therelatively large Cr andNi residualsmost La 0 0.06 0.17 0.19 0.003 6.8 likelyreflect crystal accumulation). These basanites arecharac- Ce 0 0.1 0.26 0.11 0.003 6.1 Nd 0 0.2 0.44 0.09 0.003 5.4 terized by higher incompatibleelement concentrations and lower SiO2 and Zr/Nb than the Llano del Banco and Duraznero basanites.Because of therestricted occurrence of phonotephrites Basedon compilations from Green [1994] and on additional data for at Hoyo Negro and becauseall of theserocks came from the same amphiboleand apatite [Irving and Price, 1981; Chazotet al., 1996], olivine[Beattie, 1994], and clinopyroxene [Skulski et al., 1994;Hart and vent,it is likelythat the primitive Hoyo Negro basanites are Dunn,1993]. Where arange ofcoefficients isgiven, the values yielding parentalto phonotephritesA and B. thebest solutions were used. 6010 KLOGELET AL.: THE 1949RIFT ZONEERUPTION ON LA PALMA 14 the rangeindicated by CO2-richfluid inclusionsfor otherCanary 13 • 1949 lavas Island lavas [Hansteenet al., 1998]. Magma transportin this • :•,•.Recentlavas depthrange is governedby dikesformed by hydraulicfracturing 12 2.5 •'/',, [e.g.,Lister and Kerr, 1991]. / 2.0/" Peridotire The evolution of the tephrite included mixing with a o"'0- 11 -.. (Mg# 85) phonoliticmagma as indicated by reverselyzoned clinopyroxenes 1.5 • I and kaersutiteshaving euhedral to roundedcores with low Mg •10 / Peridotite number.Magma mixing may have occurred in the inferredupper O 9 +MORB I \• mantle reservoirswhere the resultingtephrite subsequently 1.5 • resided. Likewise, the Llano del Banco and Duraznero basanites I Peridotite indicateat least two distinctmixing eventsbecause reversely (Mg # 89) zoned olivine and green-coreclinopyroxene have strongly differingMg number.We proposethat an ascendingmagma 6 batchcontaining Fo80 olivine and brownish clinopyroxene mixed 42 43 44 45 46 47 48 49 50 51 52 with a relativelysmall volume of phonoliticmagma at 600-800 SiC2 [wt%] MPa, producinggreen-core clinopyroxenes analogous to thosein the tephrite. The resulting magma was later mixed with a Figure9. A plot of SiO2versus FeO* for the 1949tephrite and relatively primitive melt producing Fo84 olivine as new basanitegroups and other recent Cumbre Vieja lavas with MgO > phenocrysts(type 1) and as a rim aroundFo80 olivine (type 2, 4 wt % comparedto partial melts of possiblesource material Figure 3). This mixing event occurreda few monthsprior to suggestsmelting pressures above 2.5-3.0 GPa. Fieldswith dashed eruption,as shownabove by diffusioncalculations. andsolid lines encircle the range of dry partial melts of two rela- Mixingof distinct magmas, including highly evolved melts, is tivelyfertile peridotites [Hirose and Kushiro, 1993] and of commonbeneath La Palma,as reflected in theabundance of mixturesof 50-67%peridotite with 33-50%MORB [Kogisoet green-core clinopyroxenesin historic and prehistoric lavas al., 1998]; numbersindicate experimental pressures in giga- studied by us and other workers [e.g., Praegel, 1986]. We pascals.Thick line with arrowshows calculated path for olivine fractionation. visualizemagma storage within the uppermantle occurring in a networkof partly interconnected"pockets" from which magma batchesare periodicallyexpelled, similar to scenariosproposed The compositionaldifferences between phonotephrites A and for other alkalic basalticprovinces [Duda and Schmincke,1985; B are interpreted to reflect different storage depths and Dobosiand Fodor, 1992]. This is consistentwith our geochemi- fractionationpaths. Calculatedmodels show that the principal cal data, indicatingthat the tephritesand basaniteswere not in path for phonotephrite A includes kaersutite and apatite prolongedcontact but evolved and ascendedseparately until fractionation,which is consistentwith decreasingP205 and surfaceactivity began. concaveupward REE patterns,high Rb/Ba ratios,and relativeTi 8.1.2. Ascent into the crust and filling of the rift system. depletionof the most evolvedrocks (Figures6 and 8), whereas The onsetof weak, sporadicseismicity in 1936 [Martel San Gil, the path for phonotephriteB includesolivine fractionationand 1960] is interpretedto reflectthe first batchof the 1949 magmas possibly •-1% of apatite assimilation(apatite is common in reachingthe crust. Magma stagnation,accumulation, or lateral cumulate xenoliths and as xenocryst inclusions). The more transportin the lower crustbeneath La Palmais manifestedby evolvedphonotephrite A and B samplesrequire contamination by fluid inclusionsin phenocrystsand xenolithsof the 1949 and wall rocksor phasessuch as titaniteto explaintheir incompatible othereruptions dominantly indicating pressures of 220-350 MPa, trace element contents.The higher kaersutite/olivineratio of closeto the Moho [Kl•igel et al., 1997; Hansteenet al., 1998] fractionatingphases for phonotephriteA than phonotephriteB (Figure 4). We interpretthis level as a major intrusionzone, may indicate deeper fractionationlevels and/or higher water possiblycomprising a crystal-richmush that may or may not contents,which is consistentwith more degassingduring ascent solidifybetween successive magma batches, and applythe term and more abundant reaction rims around kaersutites observed in "deeprift system". phonotephriteA samples. We propose that low-degree,relatively primitive basanite entered the rift system beneath Hoyo Negro in 1936 and stagnatedat three different levels, possibly within the crust 8. Discussion (Figure 10a). Over the next 13 years,this basanitedifferentiated to form (1) phonotephriteB at the shallowestlevel, (2) 8.1. Model of Ascent and Eruption of the 1949 Magmas phonotephriteA at an intermediatelevel, and (3) more evolved The simple model presentedbelow illustratesthe complexity Hoyo Negro basaniteat the deepestlevel; this reservoirwas also of magmaplumbing before and during the 1949 eruption.It is the largest. clearlynot the only possiblemodel and much of it is speculative Beginningin February1949, increasing magma pressure in the due to lacking information;however, we believe that it is the mantlereservoirs resulted in the propagationof dikesfeeding simplestmodel consistentwith the existing data and therefore tephritefrom the mantlethrough the deeprift into the shallowrift servesas an end-membertype model. system(Figure 10b). The magmaprobably stagnated at a level of 8.1.1. Storage and mixing in the mantle. After segregating neutralbuoyancy at a few kilometersdepth [Ryan, 1988; Lister from their sourceregion the ascendingmagmas stagnated and and Kerr, 1991], which is consistentwith fluid inclusion data differentiatedto varying extents in reservoirswithin the upper [Hansteenet al., 1998].The "rift event"produced strong, shallow mantle. Mineral thermobarometry(Figure 4) indicates main seismicityand ground cracking in southernmostLa Palmaduring fractionationlevels at circa 600-800 MPa (20-26 km depth) and the 3 monthsprior to eruption[Martel San Gil, 1960], a period possiblyalso at 800-1100 MPa (26-36 km) which overlapswith which coincidesremarkably well with the last mixing event KLOGEL ET AL.' THE 1949 RIFT ZONE ERUPTION ON LA PALMA 6011 ,, (A) July 1936 - February 1949 (B) February- June 23, 1949 km cumbreVieja 'rift system • .•---• --Sediments-- '•--Phon.. B •. .. .. lO . Deep Basalticcrust ..<=•::<=> ":<:;> = • •:.Hft:.system •=• • ..... =.. ,,UL••'==' ...... : • •:..:...... •.:,.,.:,.;:' ;,;,.•,,;..,;--.. '"• -.• .Mant!:e 20 o-.- •11 Tephdte •...... :'ll 30 •Ua:"'no.:-'de[:Banc.• ;".• Pho:notephrite •'. basanile •.basanite ;:".0 •:'.::O•der poc.kets...... ::I . ,, , ,...... , ...... , ...... (D)July 8- 11 km lO km LlanodelBanco • •Ho- o Negro 10 . . :.•! i•.i.•.--::•:•:•:!;i• .;:•:..•...... :.:.:.:•:• .? Figure 10. Schematicnorth-south cross section of La Palmaillustrating our modelof magmatransport and storage duringthe 1949 eruption(note that three-dimensionalrelations cannot be well illustratedby the drawing;see text for details). Lithologicalboundaries are from Ranero et al. [1995]. (a) Between 1936 and 1949, low-degree basaniteascends from the uppermantle to the deeprift systembeneath Hoyo Negro whereit stagnatesand differen- tiatesat distinctlevels, possibly within the crust.The basaniteand tephritegroup magmas reside and fractionatein the upper mantle. (b) At 3 monthsprior to eruption,following a mixing event in the mantle, basaniticmagma pushestephrite into the rift systemto causedike propagationand strongseismicity in southernLa Palma. (c) In phase1, Durazneroerupts tephrite and dikespropagate toward Llano del Banco. (d) In phase2, Durazneroshuts downand Llano del Bancofissure erupts tephrite. The basanitefollows the tephriteand eruptsa few dayslater. (e) In phase3, there is simultaneouseruption of basaniteat Llano del Bancoand basaniteand phonotephriteat Hoyo Negro, which are closelylinked throughthe rift system. (f) In phase4, Durazneroerupts basanite of intermediate composition.Partial collapse of magmareservoirs produces abundant xenoliths that are carriedto the surface. inferred for the basanites from diffusion calculations. We thus of the locusof strongestseismic precursors and over 1000 m proposethat the dramaticincrease in seismicityresulted from a higherin elevation.Intense ground cracking north of Duraznero sequenceof eventsinitiated by a mixingevent in the mantle.The waspresumably associated with shallowdike propagation along mixing producedthe Llano del Banco basanite,which subse- the rift [cf. Rubinand Pollard, 1987].On July8 a dike reached quentlyascended, encountered the tephrite,and pushedit ahead; Llano del Banco (3 km to the north and 580 m lower in however,the rising magmasdid not encounterthe pocketsof elevation),causing eruption of tephriteat increasedrates and Hoyo Negrobasanite and phonotephrites. shutdownof Duraznero(phase 2, Figure 10d). Over the next 2 8.1.3. Eruption of the magmas. On June24 a tephriticdike days,the gradual emptying of the tephritereservoirs in the mantle approachedthe surface at Duraznero and initiated phreato- was followed by the ascentof basanitealong the tephrite magmaticexplosions (phase 1, Figure 10c). Remarkably,the conduits,which produced minor intermixing. eruptionoccurred at the apexof the CumbreVieja, 10 km north Facilitatedby its lower viscosity,eruption of the basaniteat 6012 KLOGEL ET AL.: THE 1949 RIFT ZONE ERUPTION ON LA PALMA Llano del Banco induced further dike propagation and the hoursto days, during which rapid degassingand crystallization openingof Hoyo Negro on July 12 (phase3, Figure 10e). The causedthe outermostrim zonationof olivine phenocrysts.The apparentlyclose hydraulic connection between Llano del Banco magmasdid residebriefly in the lower crust,however, as shown and Hoyo Negro most likely occurredalong relativelyshallow by fluid inclusionsin phenocrysts(Figure 4), which most likely dikes [Kliigel et al., 1999], consistentwith the developmentof a reflectspassage through the deeprift system. fault systembetween the vents [Bonelli Rubio, 1950; Romero As the activityat Llano del Bancobegan to decline,tholeiitic Ortiz, 1951; Carracedoet al., 1999]. BetweenJuly 12 and 22, gabbroxenoliths began to appearin the lavas.After the activity some Llano del Banco basanitewas erupted at Hoyo Negro, ceasedentirely for 3 days a final magma pulse re-activated togetherwith Hoyo Negro basaniteand phonotephriteA and B, Durazneroand Hoyo Negro on July30 (phase4, Figure 1Of). The which presumablyresided within the rift systemas disconnected Duraznero basanite, intermediatein compositionamong the pocketsalongside the main 1949 magmapathway. Such magma basanites,could eitherhave beenformed by intermediatedegrees pockets can remain when parts of a storage system solidify of melting or could reflect mixing of the Llano del Banco and betweeninfrequent eruptions [Wright and Fiske, 1971; Garcia et Hoyo Negro end-members.We interpretthe occurrenceof crustal al., 1989; Wright and Helz, 1996]. Thus the Llano del Banco andmantle xenoliths almost exclusively in lava from the terminal basaniteeither actively flushedthe melts from these pocketsto phaseof the eruptionto reflect collapseof the plumbingsystem the surfaceat Hoyo Negro or triggeredtheir rise. Incomplete in the crust, and possibly the mantle, producing abundant mixing of the distinctbatches and abundantschlieren in the Hoyo fragmentswhich were entrainedinto the risingmagma [Kliigel et Negro rocks imply expulsion shortly after contact; in addition, al., 1999]. the low eruption rate at this vent allowed preservationof such 8.1.4. Transport of mantle xenoliths. Remarkably, the chemicalheterogeneities. Duraznerobasanite and the peridotiticmantle xenoliths it carries The bulk of the eruptedtephrite and basanitemust have passed have distinct ascent histories. The xenoliths indicate magma through the rift system quickly because clinopyroxene-melt contactfor tens of years and provide evidencefor prolonged barometry indicates mantle rather than crustal depths. Our storagein the deep rift system [Kliigel, 1998], whereas the diffusioncalculations for the basaniteindicate transfer times of basanitepassed through the crust within hours to days. In La Palma (A) Between eruptions (B) Eruption ...... D'•e-: p-,T'•i ½"•::•'•'•'' •'•"i..:;;l-:• ...... Kilauea (Hawaii) (C) Between eruptions (D) Eruption Summit caldera I¾øøanO ¸ '...... IOceanic "':" l ...... !...... '..... ":•"'•- 'cr•js.t. ,,,, , ,. ? ...... . ':'" ,,, ...... • .• .,• ,, . - ...... :..:.:•{•- .... :: .:.. •.:•.... Figure11. Simplified,schematic comparison of magma supply to Kilaueaversus La Palmarift zoneeruptions. Pocketswith older,evolved magmas are shown in white;numbers indicate depths in km. Thehypothetical cross sectionbeneath Kilauea is afterGarcia et al. [ 1996]. (a) BetweenLa Palmaeruptions, there is littleor no magma fed intothe volcanicedifice, which lacks a summitstorage zone. (b) DuringLa Palmaeruptions, spaced tens to hundredsof yearsapart, magma is transportedfrom mantle reservoirs through the deep and the shallow rift system towardthe surface. It mayencounter pockets of magmathat has been stored and differentiated at different stages. (c) BetweenKilauea eruptions, a continuous supply of sub-crustalmagma enters the summit reservoir inflating the volcanicedifice [Wright and Fiske,1971]. (d) DuringKilauea eruptions, spaced years to tensof yearsapart, magmais fedfrom below and from the summit reservoir into the rift zone,where it mingleswith magma that has beenstored and slightlydifferentiated. KLOGEL ET AL.: THE 1949 RIFT ZONE ERUPTION ON LA PALMA 6013 addition,the xenolithsare mantiedby cumulate-likeselvages serveas transient magma conduits and can propagate laterally for containingkaersutite, which is not a phenocrystphase of the up to tensof kilometers[Rubin and Pollard, 1987]. The shallow basanite.The xenolithswere thus transported from the mantleto rift systemof La Palmaprobably has a similarstructure, thelower crust by earliermagma batches, some of whichmay be expressedby thealignment of vents,eruptive fissures, and cracks associatedwith the weak seismicitybeginning in 1936. The along the CumbreVieja ridge; however,the lateral extent of earlierxenolith-bearing batches ponded in the deeprift system individualdikes appears to be significantlysmaller than on andproduced a crystal mush in whichthe xenoliths were stored Hawaii. until,late in the1949 eruption, conduit collapse spalied marginal It is not knownhow the presentrift systemof La Palma debris into the basanitewhich then carried them to the surface. formedor howit is relatedto theinferred lateral collapse of the Theminimum magma ascent rate of 0.2 m s-l necessaryto carry westernflank of the CumbreNueva at about0.7 Ma, at which fist-sizedperidotite xenoliths [Kliigel, 1998] is consistentwith time activity shiftedfrom the Taburienteshield in the northto the the inferredtransfer time throughthe crust. CumbreVieja rift zone in the south[Navarro and Coello, 1993; Ancocheaet al., 1994; Carracedoet al., 1999]. We also have no 8.2. ComparisonWith Kilauea Rift Zones constraints on whether the subcrustal conduits are located beneaththe apex of the CumbreVieja or beneaththe extinct In contrast to the two rift zones of Kilauea (Hawaii), no Taburiente shield volcano. The latter would imply significant collapsecaldera is associatedwith the activeCumbre Vieja rift of southwardtransport within the rift systemand is supportedby the La Palma, and there is no magma reservoirbeneath the summit southwardshift of seismicfoci of the 1949 precursors[Kl•gel et feeding the rift system(Figure 11). The absenceof a summit al., 1999]. reservoiron La Palma may reflect the longer periodsbetween eruptionsand the smaller volumesemitted per eruptionthan in 9. Conclusions Hawaii. Both of thesefeatures result from a much lower supply rate of basaniticmagma to La Palma (eruptive rate 0.15-0.37 The 1949 eruptionillustrates complex magma transport and km3 kyr-1 [Ancocheaet al., 1994] than of tholeiiticmagma to storagesystems beneath and within the CumbreVieja rift zoneof Kilauea(10-100 km3 kyr-• [Tillingand Dvorak,1993]). The net La Palma. Common features of the 1949 and other recent La effectof the lower supplyrates and absenceof a summitreservoir Palma eruptionsinclude eruption of two or more distinct lava is a storage system for La Palma in which only small and types,combined ridge and flank eruptions,abundant xenoliths in discontinuouspockets of residualmagma can survive from one the late lavas, and long contacttimes betweenmantle xenoliths eruptionto the next and in which pronouncedfractionation is and magma [Ibarrola, 1974; Hernanddz-Pachecoand Valls, commonplacebetween eruptions, even though the bulk of the 1982; Kl•gel et al., 1999]. Thus our proposedmodel of magma magmas(>90 vol %) is still basaltic. storageand transportmay be applicableto othereruptions along For La Palma, compositionallybimodal or multimodalerup- the CumbreVieja riff. tions,involving distinct magma batches produced and evolvedin The primary melts of La Palma magmasascend from at least the mantleand crust,are the norm. Kilauea rift zone eruptions,in 80-100 km depth and indicate residual garnet, and possibly contrast, commonly involve (1) magma accumulatedin the phlogopite,in the source.In additionto sourcevariations such as summitreservoir, (2) pocketsof magmastored along the rift, and degreeof melting, the highly variable major and trace element (3) magmanewly arrived from the mantle [Wright and Fiske, compositionof the eruptedlavas is controlledchiefly by frac- 1971; Wrightand Helz, 1996; Garcia et al., 1989] (Figure 11). tionationof olivine + clinopyroxene4- kaersutite4- Ti-magnetite Becausethe Kilauea systemis almost continuallysupplied with at 600-800 MPa and probably800-1100 MPa and by magma new magma,geochemical differences among these three compo- mixing.Mixing of mafic magmaswith batchesof highlyevolved nentsare more subtleyet can be considerablein somecases. Not melts within the mantle is common,as indicatedby reversely all eruptedlavas pass through the summitreservoir, but mixing of zonedphenocrysts such as green-coreclinopyroxenes, yet results distinct componentswithin Kilauea's rift system is common in relatively homogeneouswhole rock compositions.Magmas [Garcia et al., 1996]. This resultsin more gradualvariations in without direct geneticlinks to each other ascendseparately and the compositionof the erupted lavas, whereas the relatively are eruptedtogether to producestrongly zoned eruptions.The abruptchanges typical for La Palma imply minimal dampingof volcanic rocks of the 1949 eruption reflect derivationfrom at compositionalfluctuations and by inference smaller magma least three distinctmagma batches, all of which had a complex bodieswithin the rift system[Albarbde, 1993]. evolution involving mixing with distinct melts and shallow It is inferredthat significantmixing and lateralmagma trans- contamination. port beneathKilauea occur in the deeperpans of the rift system Upon enteringthe crust, ascendingmagmas are temporarily (3-10 km depth) consistingof more or less persistentmagma storedand laterallytransported in the deeprift systemat 200-350 bodies plus crystal mush [Ryan, 1988; Garcia et al., 1989; MPa, within the gabbroicoceanic crust, where cumulatesform Delaneyet al., 1990]. Xenolith and fluid inclusionstudies indi- and densexenoliths settle. Magma stagnationat this level is cate that La Palma has a deep rift systemas well, albeit at 8-13 consistentwith the commonpresence of tholeiiticMORB gabbro km depthwithin the oceaniccrust [Kliigel et al., 1997; Hansteen and ultramafic cumulatexenoliths, including plagioclase-and etal., 1998],which is consistentwith thoieiitic MORB gabbros apatite-bearing cumulates and "immature"cumulates with inter- and ultramaficcumulates being the mostabundant xenolith types. stitial glass and vesicles, all of which contain fluid inclusions We cannotdistinguish whether the deeprift beneathLa Palma is indicatinglower crustalpressures [Hansteen et al., 1998]. When a dike-like system, as proposedfor Kilauea [Delaney et al., the rising magmasenter the shallow rift systemat a few kilo- 1990], or a systemof sill-like bodiesas supportedby horizontal metersdepth, propagating dikes produce intense seismicity until lithologicalboundaries in the lower crust[Hill and Zucca, 1987; the eruptionbegins somewhere along the CumbreVieja ridgeand Lister and Kerr, 1991]. The deeperparts of Kilauea'srift zones its flank. are overlainby the shallowrift system,which, as deducedby During a typical La Palma eruption, most of the erupted Walker [1987], comprisesteeply dipping, subparalleldikes that magmaascends from mantledepths and passesthrough the deep 6014 KLf)GEL ET AL.: THE 1949 RIFT ZONE ERUPTION ON LA PALMA andshallow rift systemin a relativelyshort time, possibly hours are an initially constantolivine compositionScore and a constant to days.Some magma batches, especially those separated from interfacecomposition Gri m. FollowingCrank [1975], the concen- the majorconduits such as the HoyoNegro basanite and trationC at a giventime t and distancex is givenby: phonotephrites,maybecome strongly contaminated bywall rock andpockets of oldermagmas that have been stored and differen- C-Grim= erf( x ) (B1) tiatedwithin the rift system.Near the end of aneruption, partial Score-Grim 2x/-•' collapseof the magmaplumbing system produces fragments The diffusiontime for a givenprofile lengthL is calculatedby whichare erupted as xenoliths within the terminal lavas. In contrastto the rift zonesof Kilaueavolcano on Hawaii, La tdiff = • (B2) Palma lacks both a summitcaldera and a summitreservoir 16D feedingthe CumbreVieja rift system.Both volcanoesare inferred to havetheir shallowrift zonesunderlain by a deeperpart where In orderto place an upperlimit on the time necessaryto cause some lateral magma transportand mixing occurs.The magma the reversezonations of type 2 olivines the simplestdiffusion bodiesin the deeprift systembeneath La Palma are smallerand model consists of a one-dimensional,initially heterogeneous lesspersistent than thoseat Kilauea as reflectedby a largerrange olivinehaving a constamcore composition Score for x > 0 (FOB0) in compositionand more abruptcompositional changes during a and a constantmaximum composition Cma x for x < 0 (Fo84).It is singleeruption. This coincideswith a significantlylower rate of thus assumedthat a concentrationgradient exists a priori at t = 0, magmasupply. e.g. as a result of rapid growth of Fo84olivine. The solutionof this diffusionproblem is providedby (B2) with L being half of the zonationlength from coreto maximum[Crank, 1975]. Appendix A: Analytical Methods Acknowledgments. This work has benefitedgreatly from stimulating Whole rock analysesof major and trace elememswere carried discussions with T. Hansteen and P. Sachs. We thank K. Wolff and A. out on a Philips X'Unique X-ray spectrometerequipped with a Merkau for carryingout the XRF analyses,D. Garbe-Schonbergand T. Rh tube at Geomarin Kiel, usingfused beads and pressedpellets. Arpe for the ICP-MS analyses,and P. G16erfor technicalassistance with H20 and CO2 were analyzedon a RosemountCSA 5003 infrared the electronmicroprobe. Constructive reviews by M. Garcia,J. Stix, and photometer.Trace elements of selectedsamples were analyzedby D. Francishave significantlyimproved the manuscript.Our studieswere supportedby a stipendfrom the Studienstiftungdes deutschenVolkeG inductively coupled plasma mass spectrometry(ICP-MS) on a and a DFG fellowship(GRK 392/1-99) to A.K., by an Alexanderyon VG PlasmaQuad PQ1 at the Geologisch-Pal•iontologischesHumboldt fellowship to J.D.L.W., and by the DeutscheForschungs- Institut, Christian-Albrechts-Universit•itin Kiel (see Garbe- gemeinschaft(DFG grants Sa 594/2-1 and Schm 250/47-1) and the SchOnberg[1993] for details).Analytical precision(better than National ScienceFoundation (NSF grant EAR-9105113 to K.H.). This work presentsresults of the Ph.D. thesisof A.K. 5%) and accuracywere controlledby replicateanalyses, blanks, and standards.In all figures the major element data have been normalizedto 100 wt % on a H20- andCO2-free basis. References Mineral and glassanalyses were carriedout on a CamecaSX- Abdel-Monem, A., N. D. Watkins, and P. W. Gast, Potassium-argon 50 electronmicroprobe (EMP) at Geomar in Kiel. Analytical ages,volcanic stratigraphy,and geomagneticpolarity historyof the conditions included a beam current of 50 nA for olivine and 10 to CanaryIslands: Tensrife, La Palma and Hierro,Am. d. Sci., 272, 805- 20 nA for other phasesat an accelerationvoltage of 15 kV. A 825, 1972. focusedbeam was usedfor olivine, pyroxene,and oxides,and a Albar•de, F., Residence time analysis of geochemicalfluctuations in volcanicseries, Geochim. Cosmochim.Acta, 57, 615-621, 1993. rasteredbeam of 3x4 [tm was usedfor all otherphases. Counting Ancochea,A., F. Hemfin, A. Cendrero,J. M. Cantagrel,J. M. Faster, E. timeswere 20 s on peak and 10 s on background.Some pyroxene Ibarrola,and J. Coello, Constructiveand destructiveepisodes in the analyseswere carried out on a Jeol JXA 8900RL microprobeat building of a young oceanicisland, La Palma, Canary Islands,and the Institut air Geowissenschaften,Johannes Gutenberg- genesisof the Calderade Taburiente,d. l/olcanol.Geotherm. Res., 60, 243-262, 1994. Universitfitin Mainz, using a defocusedbeam and peak counting Beattie, P., Systematicsand energeticsof trace-elementpartitioning times of 15 s. betweenolivine and silicate melts: Implicationsfor the nature of mineral/meltpartitioning, Shem. Geol., 117, 57-72, 1994. Bonelli Rubio, J. M., Contribuci6nal estudiode la erupci6ndel volcan Appendix B: Diffusion Calculations del Nambroqueo San Juan (Isla de La Palma), 24 de Junio - 4 de Agostode 1949, Talleres,pp. 5-22, Inst. Geogr.y Catastral,Madrid, By extrapolatingthe data of Chakraborty [1997] we use a 1950. constantinterdiffusion coefficient DFeMg for FoB0olivine since Boynton, W. V., Geochemistryof the rare earth elements:meteorite the compositionaldependence is weak within the range of studies, in Rare Earth Element Geochemistry, edited by P. interest(F074 to F084). After correctingfor oxygenfugacity by Henderson,pp. 63-114, ElsevierSci., New York, 1984. Buening,D. K., and P. R. Buseck,Fe-Mg lattice diffusionin olivine, d. therelation DFeUg --- (fo2) 1/6 [Bushing and Buseck, 1973], we Geophys.Res., 78, 6852-6862, 1973. obtainDFemg = 1.3x10-16 m 2 s-1 alongthe [001]direction of Carmichael,I. S. E., and M. S. Ghiorso,The effect of oxygenfugacity on olivinefor a modeltemperature T = 1150øCand fo2 = 10-8 bar. the redox state of natural liquids and their crystallizingphases, in Owing to the anisotropyof olivine, diffusion along [010] is Modernmethods of igneouspetsology, edited by J. NiGhoilsand J. K. Russell,Mineral. Soc. of Am. Rev. Mineral., 24, 191-212, 1990. slowerthan along [001] by a factor of---4 [Bueningand Buseck, Carracedo,J. C., The CanaryIslands: An exampleof structuralcontrol on 1973]. the growthof large ocean-islandvolcanoes, d. l/olcanol. Geotherm. For olivine rims we use a simple one-dimensionaldiffusion Res., 60, 225-241, 1994. modelof an infinite halfspacewith a planarinterface toward the Carracedo,J. C., S. J. Day, H. Guillou, and P. Gravestock,Later stages of volcanicevolution of La Palma,Canary Islands: Rift evolution,giant hostmelt becausemost type 1 olivinesshow a plateauof constant landslides,and the genesisof the Calderade Taburiente,Geol. Soc. compositionat their coresand becausemost zoning profiles were Am. Bull., 111,755-768, 1999. measuredperpendicular to crystalfaces. The boundaryconditions Chaffey,D. J., R. A. Cliff, and B. M. Wilson,Characterization of the St KLOGEL ET AL.: THE 1949 RIFT ZONE ERUPTION ON LA PALMA 6015 Helenamagma source, in Magmatismin theOcean Basins, edited by Hofmann,A. W., Chemical differentiation ofthe Earth: The relationship A.D. Saundersand M. J. Norry,Geol. Soc. Spec. Publ., 42, 257-276, betweenmantle, continental crust, and oceanic crust, Earth Planet. 1989. Sci.Lett., 90, 297-314, 1988. Chakraborty,S., Ratesand mechanismsof Fe-Mg interdiffusionin Ibarrola,E., Temporalmodification of thebasaltic materials from 1971 olivineat 980ø-1300øC,,/. Geophys.Res., 102, 12,317-12,331,1997. eruptionof the Teneguiavolcano (La Palma,Canary Islands), Chazot,G., M. A. Menzies,and B. Harte,Determination of partition EstudiosGeol., Vol. Teneguia, 49-58, 1974. coefficientsbetween apatite, clinopyroxene, amphibole, and melt in Irving,A. J., andR. C. Price,Geochemistry and evolution of lherzolite- naturalspinel lherzolites from Yemen; Implications for wet melting of bearingphonolitic lavas from Nigeria, Australia, East Germany and the lithosphericmantle, Geochim. Cosmochim. Acta, 60, 423-437, NewZealand, Geochim. Cosmochim. Acta, 48, 1309-1320,1981. 1996. Kaiser,C., The 1949 eruption of VolcanSan Juan, La Palma,Canary Crank,J., TheMathematics of Diffusion,414 pp., OxfordUniv. Press, Islands:Petrology and lava flow dynamics,diploma thesis, Univ. New York, 1975. Bochum,Bochum, Germany, 1988. Delaney,P. T., R. S. Fiske,A. Miklius,A. T. Okamura,and M. K. Sako, Klitgord,K. D., andH. Schouten,Plate kinematics of thecentral Atlantic, Deepmagma body beneath the summitand rift zonesof Kilauea in The Geology of North America, vol. M, The WesternNorth volcano,Hawaii, Science,247, 1311-1316, 1990. AtlanticRegion, edited by P. R. Vogt andB. E. Tucholke,pp. 351- Dobosi,G., and R. V. Fodor,Magma fractionation,replenishment, and 378, Geol. Soc.of Am., Boulder,Colo., 1986. mixing as inferred from green-coreclinopyroxenes in Pliocene Kltigel,A., Reactionsbetween mantle xenoliths and host magma beneath basanite,southern Slovakia, Lithos, 28, 133-150, 1992. La Palma (CanaryIslands): Constraints on magmaascent rates and Duda,A., andH.-U. Schmincke,Polybaric differentiation of alkalibasalt crustalreservoirs, Contrib. Mineral. Petrol., 131,237-257, 1998. magmas:Evidence from green-coreclinopyroxenes (Eifel, FRG), Kltigel,A., T. H. Hansteen,and H.-U. Schmincke,Rates of magma Contrib. Mineral. Petrol., 91,340-353, 1985. ascentand depthsof magmareservoirs beneath La Palma (Canary Duke, J. M., Distributionof the period four transitionelements among Islands),Terra Nova, 9, 117-121, 1997. olivine,calcic clinopyroxene and mafic silicateliquid: experimental Kltigel, A., H.-U. Schmincke,J. D. L. White, and K. A. Hoernle, results,./.Petrol., 17, 499-521, 1976. Chronologyand volcanology of the 1949multi-vent rift-zone eruption Elliott, T. R., Elementfractionation in the petrogenesisof oceanisland on La Palma(Canary Islands), ./. Volcanol.Geotherm. Res., in press, basalts,Ph.D. thesis, The OpenUniv., Milton Keynes,England, 1991. 1999. Fisk,M. R., B. G. J. Upton,C. E. Ford,and W. M. White,Geochemical Kogiso,T., K. Hirose, and E. Takahashi,Melting experimentson andexperimental study of the genesisof magmasof RaunionIsland, homogeneousmixtures of peridotiteand basalt:Application to the IndianOcean, ./. Geophys.Res., 93, 4933-4950, 1988. genesisof oceanisland basalts, Earth Planet. Sci. Lett., 162, 45-61, Ford,C. E., D. G. Russell,J. A. Craven,and M. R. Fisk, Olivine-liquid 1998. equilibria:Temperature, pressure and composition dependence of the Le Maitre, R. W., P. Bateman,A. Dudek,J. Keller,M. J. LameyreLe crystal/liquidcation partition coefficients for Mg, Fe2+, Ca andMn, d. Bas, P. A. Sabine,R. Schmid, H. Sorensen,A. Streckeisen,A. R. Petrol., 24,256-265, 1983. Wooley, and B. Zanettin,A Classificationof IgneousRocks and Garbe-Sch0nberg,C. D., Simultaneousdetermination of thirty-seven Glossaryof Terms,Blackwell, Malden, Mass., 1989. trace elementsin twenty-eightinternational rock standardsby ICP- Lister, J. R., and R. C. Kerr, Fluid-mechanical models of crack MS, Geost.Newsl., 17(1), 81-97, I993. propagationand their applicationto magmatransport in dykes, Garcia,M. O., R. A. Ho, J. M. Rhodes,and E. W. Wolfe, Petrologic Geophys.Res., 96, 10,049-10,077, 1991. constraintson rift-zoneprocesses; results from episode1 of the Puu Martel San Gil, M., El Volcande San Juan, 239 pp., Talleresde Artes Oo eruptionof Kilaueavolcano, Hawaii, Bull. Volcanol.,52, 81-96, Gr/tficas,Madrid, 1960. 1989. Middlemost,E. A. K., Evolutionof La Palma,Canary archipelago, Garcia, M. O., J. M. Rhodes,F. A. Trusdell,and A. J. Pietruszka, Contrib. Mineral. Petrol., 36, 33-48, 1972. Petrologyof lavasfrom the PuuOo eruptionof Kilaueavolcano, III, Navarro, J. M., and J. Coello, Sucesi6nde episodiosen la evoluci6n The Kupaianahaepisode (1986-1992), Bull. Volcanol.,58, 359-379, geo16gicade La Palma,mapa geo16gico de La Palma,Inst. Nac. para 1996. la Conserv.de la Nat., Madrid, 1993. Green, T. H., Experimentalstudies of trace-elementpartitioning Olafsson,M., and D. H. Eggler, Phase relationsof amphibole, applicableto igneouspetrogenesis - Sedona 16 yearslater, Chem. amphibole-carbonate,and phlogopite-carbonate peridotite: Petrologic Geol., 117, 1-36, 1994. constraintson the asthenosphere,Earth Planet. Sci. Lett., 64, 305- Hansteen,T. H., A. Kltigel, and H.-U. Schmincke,Multi-stage magma 315, 1983. ascentbeneath the Canary Islands:Evidence from fluid inclusions, Pietruszka,A. J., and M. O. Garcia,The size and shapeof Kilauea Contrib. Mineral. Petrol., 132, 48-64, 1998. volcano'ssummit magma storage reservoir: A geochemicalprobe, Hart, S. R., andT. Dunn,Experimental cpx/melt partitioning of 24 trace Earth Planet. Sci. Lett., 167, 311-320, 1999. elements,Contrib. Mineral. Petrol., 113, 1-8, 1993. Praegel,N. O., The petrologyand geochemistry of Volcb.n Teneguia,La Hausen,H., Some contributionsto the geologyof La Palma, Commun. Palma, Canary Islands, Ph.D. thesis, Univ. of Copenhagen, Phys.Math. Finn. Acad.Sci., 35, 1-140, 1969. Copenhagen,1986. Hernand6z-Pacheco,A., and J. De la Nuez, Las extrusionessalicas del Putirka,K., Magmatransport at Hawaii:Inferences based on igneous Surde la Isla de La Palma(Canarias), Estud. Geol., 39, 3-30, 1983. thermobarometry,Geology, 25 (1), 69-72, 1997. Hernand6z-Pacheco,A., and M. C. Vails, The historiceruptions of La Putirka, K., M. Johnson,R. Kinzler, J. Longhi, and D. Walker, PalmaIsland (Canarias), Arquipelago, Rev. Univ. Azores,Ser. C., Thermobarometryof mafic igneousrocks based on clinopyroxene- Nat., 3, 83-94, 1982. liquid equilibria,0-30 kbar, Contrib.Mineral. Petrol., 123, 92-108, Hill, D. P., and J. J. Zucca,Geophysical constraints on the structureof 1995. Kilauea and Mauna Loa volcanoes and some implications for Ranero,C. R., M. Torne, and E. Banda, Gravity and multichannel seismomagmaticprocesses, in Volcanismin Hawaii, editedby R.W. seismicreflection constraints on the lithosphericstructure of the Decker,T.W. Wright, and P.H. Stauffer,U.S. Geol. Surv. Prof CanarySwell, Mar. Geophys.Res., 17, 519-534, 1995. Paper, 1350, 903-917, 1987. Roeder,P. L., and R. F. Emslie,Olivine-liquid equilibrium, Contrib. Hirose,K., andI. Kushiro,Partial melting of dryperidotites at high Mineral.Petrol., 29, 275-289, 1970. pressures:Determination of compositions of melts segregated from Roeser,H. A., Magneticanomalies in the magnetic quiet zone off peridotiteusing aggregates ofdiamond, Earth Planet. Sci. Lett., 114, Morocco,in Geology ofthe Northwest African Continental Margin, 477-489,1993. editedby U. Rad,K. Hinz,M. Sarnthein,and E. Seibold,pp. 61-68, Hirschmann,M. M., andE. M. Stolper,A possiblerole for garnet Springer-Verlag,NewYork, 1982. pyroxeniteinthe origin of the "garnet signature" inMORB, Contrib. Romero Ortiz, J., La erupci6n delNambroque enla isla de La Palma, Bol. Mineral.Petrol., 124, 185-208, 1996. Inst.Geol. Min. Esp., 63, 3-163, 1951. Hoernle,K., Traceelement and Sr-Nd-Pb isotopic geochemistry of Rubin,A.M., andD. Pollard,Origins of blade-like dikes in volcanic rift Jurassicocean crust beneath Gran Canaria(Canary Islands): zones,in VolcanisminHawaii, edited by R. W. Decker,T. W. Wright Implicationsfor the generation of OIB reservoirsand for crustal andP. H. Stauffer,U.S. Geol. Surv. Prof Pap.,1350, 1449-1470, contaminationof ascending OIB magmas, J. Petrol., 39, 859-880, 1987. 1998. Rutherford,M. J., andP.M. Hill, Magmaascent rates from amphibole 6016 KLf,)GEL ET AL.: THE 1949 RIFT ZONE ERUPTION ON LA PALMA breakdown:An experimentalstudy applied to the 1980-1986Mount Walker,G. P. L., The dikecomplex of Koolauvolcano, Oahu: Internal St. Helenseruptions, or. Geophys. Res., 98, 19,667-19,685,1993. structureof a Hawaiianrit• zone, in Volcanismin Hawaii, edited by Ryan,M.P., The mechanicsand three-dimensional internal structure of R.W.Decker, T.W. Wright, and P.H. Stauffer, U.S. Geol. Surv. Prof. activemagmatic systems: Kilauea volcano, Hawaii, o r. Geophys.Res., Pap., 1350, 961-993, 1987. 93, 4213-4248, 1988. Weaver, B. L., The origin of ocean island basalt end-member Schmincke,H.-U., Geologyof theCanary Islands, in Biogeographyand compositions:Trace element and isotopic constraints, Earth Planet. Ecologyof theCanary Islands, edited by G. Kunkel,pp. 67-184, Junk Sci.Lett., 104, 381-397, 1991. BV, The Hague,Netherlands, 1976. White,J. D. L., andH.-U. Schmincke,Phreatomagmatic eruptive and Schmincke,H.-U., Volcanic and chemicalevolution of the Canary depositionalprocesses during the 1949 eruption on La Palma (Canary Islands,in Geologyof theNorthwestern African Continental Margin, Islands),J. Volcanol.Geotherm. Res., in press,1999. editedby U. vonRad, K. Hinz,M. Sarnthein,and E. Seibold,pp. 3- Wright,T. L., andR. S. Fiske,Origin of thedifferentiated and hybrid 20, Springer-Verlag,New York, 1982. lavasof Kilaueavolcano, Hawaii, or. Petrol., 12, 1-65,1971. Schmincke,H.-U., A. Kltlgel,T. H. Hansteen,K. Hoernle,and P. v. d. Wright,T. L., andR. T. Helz,Differentiation and magma mixing on Bogaard, Samplesfrom the Jurassicocean crust beneathGran Kilauea'sEast Rit• Zone: A furtherlook at the eruptions of 1955and Canaria,La Palmaand Lanzarote (Canary Islands), Earth Planet.Sci. 1960;2, The1960 lavas, Bull. Volcanol., 57, 602-630,1996. Lett., 163, 343-360, 1998. Shaw, H. R., Viscositiesof magmaticsilicate liquids: An empirical methodof prediction,Am. or. Sci., 272, 870-893, 1972. Skulski,T., W. Minarik, and E. B. Watson,High-pressure experimental K. A. Hoernle and H.-U. Schmincke,Geomar Forschungszentrum, trace-elementpartitioning between clinopyroxene and basalticmelts, Wischhofstr.1-3, D-24148 Kiel, Germany. (khoemle•geomar.de; Chem. Geol., 117, !27-148, 1994. hschmincke•geomar.de) Sparks,R. S. J., and H. Pinkerton,Effect of degassingon theologyof A. Kltigel, Universitltt Bremen, FachbereichGeowissenschaf•en, basalticlava, Nature, 276, 385-386, 1978. Postfach 330440, D-28334 Bremen, Gemany. (akluegel•uni- Staudigel,H., andH.-U. Schmincke,The Plioceneseamount series of La bremen.de) Palma/CanaryIslands, or. Geophys. Res., 89, 11,195-11,215,1984. J. D. L. White,University of Otago,Geology Department, PO Box 56, Takahashi,E., and I. Kushiro, Melting of a dry peridotite at high Dunedin,New Zealand.(james.white•stonebow. otago.ac.nz) pressuresand basaltmagma genesis, Am. Mineral., 68, 859-879, 1983. Tilling, R. I., and J. J. Dvorak,Anatomy of a basalticvolcano, Nature, (ReceivedSeptember 30,1998; revised August 9, 1999; 363, 125-133, 1993. acceptedSeptember 20, 1999.)