Trinity University Digital Commons @ Trinity

Geosciences Faculty Research Geosciences Department

11-10-1990

Compositional Diversity of Late in a Transect Across the Southern Cascades: Implications for Zone Magmatism

W. P. Leeman

Diane R. Smith Trinity University, [email protected]

W. Hildreth

Z. Palacz

N. Rogers

Follow this and additional works at: https://digitalcommons.trinity.edu/geo_faculty

Part of the Earth Sciences Commons

Repository Citation Leeman, W. P., Smith, D. R., Hildreth, W., Palacz, Z., & Rogers, N. (1990). Compositional diversity of late Cenozoic basalts in a transect across the southern Washington Cascades: Implications for subduction zone magmatism. Journal of Geophysical Research: Solid Earth, 95(B12), 19561-19582. http://doi.org/ 10.1029/JB095iB12p19561

This Article is brought to you for free and open access by the Geosciences Department at Digital Commons @ Trinity. It has been accepted for inclusion in Geosciences Faculty Research by an authorized administrator of Digital Commons @ Trinity. For more information, please contact [email protected]. JOURNAL OF GEOPHYSICAL RESEARCH,VOL. 95, NO. BI2, PAGES19,561-19,582, NOVEMBER 10, 1990

Compositional Diversity of Late Cenozoic Basalts in a Transect Across the SouthernWashington Cascades' Implicationsfor SubductionZone Magmatism

WILLIAM P. LEEMAN,• DIANE R. SMITH,2 WESHILDRETH, 3 ZEN PALACZ,4.5 AND NICK ROGERS4

Major volcanoesof the SouthernWashington Cascades (SWC) includethe large stratovolcanoesof Mount St. Helens(MSH) andMount Adams(MA} and the IndianHeaven (IH) and SimcoeMountain (SiM) volcanicfields. There are significant differences among these volcanic centers in termsof theircomposition and evolutionary history. The stratovolcanoesconsist largely of andesitic to daciticlavas and pyroelastics with minorbasalt flows. IH consistsdominantly of basalticwith minor andesitclavas, all eruptedfrom monogeneticrift andcinder cone vents. SIM hasa poorlyexposed andesitcto rhyolitecore but mainly consists of basalticlavas erupted from numerous widely dispersed vents; it has the morphologyof a shield . Distribution of mafic acrossthe SWC is related to north-northwesttrending faults and fissure zones that indicate a significantcomponent of east-west extensionwithin the area.There is overlapin eruptivehistory for theareas studied, but it appearsthat peakactivity was progressivelyolder (MSH (<40 Ka), IH (mostly<0.5 Ma), MA (<0.5 Ma), SIM (I.-4 Ma)) and morealkalic towardthe east. A varietyof compositionallydistinct mafic magma types has been identifiedin the SWC, includinglow large ion lithophileelement (LILE) tholeiitic basalts, moderateLILE calcalkalicbasalts, basalts transitional between these two, LILE-enriched mildly alkalic basalts,and basalticandesites. Compositional diversity among basaltic lavas, both within individual centersas well as acrossthe arc, is an importantcharacteristic of the SWC traverse. The fact that the basaltic magmaseither show no correlation between isotopicand trace element componentsor show trends quite distinctfrom those of the associatedevolved lavas, suggeststhat their compositionalvariability is attributableto subcrustalprocesses. Both the primitivenature of the erupted basaltsand the fact that they are relatively commonin the SWC sectoralso imply that such magmashad little residencetime in the crust. A majority of the SWC basalticsamples studies are indistinguishablefrom oceanicisland basalts(OIB) in terms of trace elementand isotopiccomposi- tions, and more importantly, most do not display the typical high field strengthelement (HFSE) depletion seenin subduction-relatedmagmas in volcanicarcs elsewhere.LILE enrichmentand HRSE depletioncharacteristics of most arc magmasare generallyattributed to the role of fluids releasedby dehydrationof subductedoceanic lithosphere and to the effectsof sedimentsubduction. Because most SWC basalts lack these compositionalfeatures, we concludethat subductedfluids and sedimentsdo not play an essentialrole in producingthese magmas.Rather, we infer that they formed by variable degree melting of a mixed mantle source consistingmainly of heterogeneouslydistributed OIB and mid-ocean ridge source domains. Relatively minor occurrencesof HFSE-depleted arclike basalts may reflect the presence of a small proportion of slab-metasomatizedsubarc mantle. The juxtaposition of such different mantle domains within the lithospheric mantle is viewed as a consequenceof (1) tectonic mixing associatedwith accretionof oceanicand islandarc terranesalong the Pacific margin of North America prior to Neogenetime, and possibly(2) a seawardjump in the locus of subduction at about 40 Ma. The Cascadesarc is unusual in that the subductingoceanic plate is very youngand hot. We suggestthat slabdehydration outboard of the volcanicfront resultedin a diminishedrole of aqueousfluids in generatingor subsequentlymodifying SWC magmascompared to the situation at most convergent margins. Furthermore, with low fluid flux conditions, basalt generationis presumablytriggered by other processesthat increasethe temperatureof the mantle wedge (e.g., convective mantle flow, shear heating, etc.).

INTRODUCTION Range has been consideredto be a more or less typical subduction-relatedarc [e.g., McBirney, 1978;McBirney and The CascadeRange is a continentalvolcanic arc situated White, 1982], although petrogenetic studies have largely alongthe PacificNorthwest margin of North America.It has focusedon the more topographicallyimposing young strato- beena site of vigorousmagmatic activity in responseto volcanoesof the High Cascades. We have undertaken pet- obliquesubduction of the Farallon and Juande Fuca plates rologicstudies of major central volcanoesand nearbybasal- duringapproximately the last 38 m.y. [cf. Lux, 1982;Vet- tic fields within an east-west transect across the planckand Duncan, 1987].In manyrespects, the Cascade southern Washington Cascades to document the spatial distributionof magmatypes in this portion of the arc. This lKeith-Wiess Geological Laboratories, Rice University, Hous- regionis an area of transition[Weave• and Malone, 19871 ton, Texas. "GeologyDepartment, Trinity University, San Antonio, Texas. which separatesthe isolatedcentral volcanoesto the north '*U.S.Geological Survey, Menlo Park, California. (Mount Rainier, Mount Baker, and Glacier Peak) from the 4Departmentof Earth Sciences, Open University, Milton Key- broadvolcanic of the OregonCascades to the south nes, England. which has a widespreadand nearly continuousvolcanic SNowat VGIsotopes Limited, Winsford, Cheshire, England. cover from both large stratovolcanoes(e.g., Mount Hood, Copyright1990 by theAmerican Geophysical Union. Mount Jefferson,Mount Washington,and the Three Sisters) Pa•r Number89JB03490. and numerousmonogenetic vents [cf. Luedke and Smith, 0148-0227/90189JB-03490505.00 !982]. Although magmaticactivity in the vicinity of the

19,561 19,562 LEEMAN ET AL.: CHEMICAL DIVERSITY OF CASCADES ARC BASALTS

SouthernWashington Cascades (SWC) transectdates back volcanicand sedimentary rocks. Each of thesevolcanoes to at least 36 Ma [cf. Evarts et al., 1987; Korosec, 1987a, b], comprisesa variety of magmatypes ranging from basalt to this paper concentrateson magmaticactivity within this part rhyodacite[Hoblitt etal., 1980; Hildreth and Fierstein, 1985• of the Cascadesduring the past few million years (mainly Ortet al., 1983].Distributed among them are the Indian late Plioceneto Recent) and principallyon petrogenesisof Heaven(iH) andnumerous smaller lava fields, small shiekt the basaltic lavas and implications for subcrustalmagmatic volcanoes,and monogenetic vents comprising mainly ba•. processes related to Cascades subduction. ticlavas [e.g., Hammond and Korosec, 1983]. Many of the lattervents are distributed in roughlyNNE to NNWtrend. TECTONIC SETTING OF THE SWC ingalignments orfissure systems that presumably areper. pendiculartothe direction of leastprinciple horizontal stress The Cascadessubduction zone is somewhatatypical com- [e.g.,Nakamura, 1977]. Loci of the largestratovolcanoes pared to other convergent margins [Duncan and Kultn, straddlesuch structures, and most notably Mount St. Helens 1989]. There are no deep (> 100 km) earthquakesassociated [C.S. Weaveret al., 1987]and Mount Hood [Williams etal., with the present Juan de Fuca subduction [Weaver and 1982]appear to be situatedabove zones of localcrustal Baker, 1988], althougha subductingslab has been imagedto extensionalong active NW trendingstrike-slip zones. greater depths by seismic tomography [Rasmussen and MountSt. Helens has been intermittently active during the Humphreys, 1988]. Because the Juan de Fuca spreading last 36,000years [Mullineaux, 1986], during which time it ridge is exceptionally close to the subductionzone, the producedpredominantly dacitic magmas[Smith and Lee. subductingplate is relatively thin and warm; this factor may man, 1987].in the last •-2000years, its eruptiveproducts explain the paucity of deep earthquakesand may influence havebecome more diverse, ranging from basalt to rhyo- the dip angle of the subductingplate as well as the rate of dacite.However, true basaltic lavas were produced only subduction.There is no topographicallyexpressed trench duringthe CastleCreek eruptive period (2200-1700 yea.rs along the convergent margin, possibly because the trench B.P. [Hoblitt et al., 1980]).Somewhat older (about 690,• has been filled with terrigenoussediments owing to high to 8,000years B.P. [Hammondand Korosec,1983]) basaltic rates of sedimentationalong the margin. to andesiticlavas erupted from smaller vents in thevicinity Convergenceof the Juande Fuca and North Americanplates (e.g., Marble Mountain, , Soda Peak, andTrout presentlyoccurs at a slow rate (approximately4 cm/yr) and CreekHill; seeFigure 1); someof thesevents lie alongthe at an oblique angle (about N50øE) to the continentalmargin NW trendingMount St. Helens seismiczone [C. S. Weaver [Riddihough,1984; Rogers, 1985];both of theseparameters et aI., 1987]. have varied significantlywith time [Rea and Duncan, 1986; The Indian Heaven and King Mountain-Mount Adams Verplanckand Duncan, 1987]. Teleseismicimaging of the volcanicfields are associatedwith NNE to NNW trending subductingJuan de Fuca plate suggeststhat it is segmented structures (fissures or normal fault zones) that are delineated and dippingrelatively steeply (>45 ø) beneaththe Oregon- by alignmentsof vents. Most volcanicactivity in the Indian Washington Cascades [Weaver and Michaelson, 1985; Heaven field postdatesthe last magneticreversal at 670,000 Michaelson and Weaver, 1986]. Because of the oblique years. B.P. [Hammond and Korosec, 1983; M. A. Korosec, subduction,the regionalstress field is not orientedsymmet- personal communication1989]; these rocks are predo•- rically with respect to the subductionzone. Rather, the nantly primitive basalts, althougha few basalticandesires greatestprincipal stresscorresponds to roughly north-south and andesitesare present.Most eruptive productsof Mount horizontal compressionand the least principal stress to Adams (a compositestratovolcano) are youngerthan about approximatelyeast-west horizontal extension [Zoback and 450,000years B.P. [Hildreth and Fierstein, 1985];they range Zoback,, 1980;Smith, 1982;Kim et al., 1986]. Evidence for in composition from basalt to rhyodacite, althoughthe east-west extension elsewhere in the Cascades arc includes basaltstypically erupted from monogeneticor shieldvents (1) earthquakefault-plane solutionsin Northern California distributed on the flanks of the stratovolcano. Due south of [e.g., Klein, 1979], and (2) geologicalfeatures such as Mount Adams, and east of Mount Hood, the NNE trending abundance of basaltic lavas, north-south orientation of Hood River fault zone in northern Oregonis associatedwith chainsof cindercones, and major normalfaults throughout a small field of Plio- basaltic lavas. East of Mount the range [McBirney, 1978;Hammond, 1979;McKee et al., Adams, the Plio-Pleistocene Simcoe (Indian Rock) volcano 1983;Priest, 1983;Bacon, 1985;Smith et al., 1987;Guffanti is a large shieldlikestructure with exposuresof dacitesand and Weaver,1988]. The OregonHigh Cascadesare actually rhyolites in its dissected core [Sheppard, 1967]. However, situatedwithin a N-S trendinggraben [Taylor, this issue].A its surroundingflanks are blanketedby basalticlavas which belt of elevated heat flow coincident with this structure erupted from numerousfissures, cinder cones, and s .mall presumablyreflects shallow magma intrusion and hydrother- shield vents that define either N-S or NNW trending mal circulation concentrated within an axial extensional ments. Available K-Ar ages indicate that nearly all Simcoe zone [Blackwell et al., 1982]. lavasare older than 1 Ma; mostrange between 2 and4 Ma [Farooquiet al., 1981;Luedke and Smith, 1982;Phillips et GEOLOGIC SETTING OF THE $WC aI., 1986; Anderson, 1987a, b]. AND SAMPLE DISTRIBUTION Althoughonly approximateestimates of magmavo!urr•s are availablefor this sectorof the Cascadesarc [cf. SherrM Generalgeologic features of the SWC are shownin Figure and Smith, this issue], it is clear that basaltic lavas are 1. The large central volcanoesof Mount St. Helens (MSH) commonexcept within the large stratovolcanoes,and even andMount Adams (MA) andthe shieldlikeSimcoe (SIM) (or therebasaltic lavas occur in significantvolumes [cL ttam- IndianRock) volcano define an easterlytrend some 100 km mondand Korosec,1983]. For comparison,basa}tic roc•ks in width.The exposed basement consists mainly of Cenozoic are also abundantin the southernCascades [Luedke a,nd LEEMANET AL ' CHEMICAl.DIVERSITY OF'CASCADES ARC []ASALTS 19,563

12•00

. St. Helen$ Mt. A

\ \

VENTS <1 m.y.old 1-5 m.y.old From= 4•I Basalt I Luedkeand Smith, 1982 Campbelland Bentley, (• Andesitc [1 1981 Farooqui et al., 198! O BasaltDacite to I'-1 Hammond,1980 Fig.1. Geologicsetting of the SWC traverse showing extent of late Pliocene toRecent volcanic rocks (stippled areas)and locations of major vents [from Leudke and Smith, 1982]. Miscellaneous sampling localities indicated by lettersinclude Marble Mountain (MM), Soda Peak (SP), West Crater (WC), (TCH),and Hood River valley(HR). Conspicuousvent alignments (outlined by dashedlines) are believed to be fault-controlled.Major structures(indicated by conventionalsymbols) offsetting to P!ioceneage volcanic deposits include ENE trendingfolds and thrust faults, NW trending strike-slip faults, and north trending normal faults [after Harnrnond, 1980; Campbelland Bentley, 1981; Farooqui et al., 1981].These structural elements are compatible with N-S regional compression(trI N-S) and are distinct from Neogene structural patterns ofthe Basin and Range province (tri vertical).

Smith,1982; McKee et al., 1983;Hughes and Taylor,1986; activationspectrometry. All analytical methodsused are Walkerand Naslund,1986], and it is estimatedthat they describedby Smith and Leemah [1987], and error estimates compriseas much as 85% by volumeof the Tertiaryto for eachelement are giventherein. Briefly, on the basisof modernOregon Cascades [McBirney and White, 1982].In replicateanalyses, accuracy and precision of majorelement theSWC traverse, basaltic volcanism was not common prior contentsare estimatedto range between 1 and 2%, and of to Pliocenetime; thereafter, basaltic magmatism is charac- most trace elements between 2 and 10% of the amount teristicof the entiretraverse, although such lavas are more reported. For elements determined by more than one voluminoustoward the east. It is unclear whether this method,either averagesor resultsby the most precise relationreflects fundamental differences in the supplyof methodare reported.Major elementanalyses are reported basalticmagma across the transect,or the possibilitythat on a "dry" basisrecalculated to 100% with all iron as FeO*. eruptionefficiency of mariemagmas is somehowpropor- tionalto the lifetimeof the eruptivecenters. The main Isotopic compositionsof Sr, N d, and Pb were measured objectiveof thispaper is to providea conciseoverview of usingVG Isotopes54E and Finnigan-MAT25! massspec- spatialvariations in magmacomposition based on repre- trometersat the Open University, England. Ratios are sentativesamples collected from all of the localitiesdis- normalized to correct for mass fractionation as follows: cussed. 86Sr/88Sr = 0.1194, ]46Nd/•Nd = 0.7219.Pb isotope ratios were correctedfor ---0.I% fracti.onationper amu basedon replicatesof the NBS-981common lead standard. Accuracy ANALYTICAL METHODS of the reported ratios is about -+0,(RX)03(0.004%) or better Majorelement compositions were determinedfor more for 87Sr/86Sr,ñ0.00002 (0.003%) or betterfor t43Nd/mNd, than150 whole rock samples using a combinationof X ray and --.0.02(0.1%) or betterfor all leadisotope ratios; internal fluorescence(XRF) spectrometry and inductively coupled precision of the analysesis usually better than theseesti- plasmaoptical emission spectroscopy. Many of these rocks mates.The 87Sff86Sr ratios are, reposed relative to valuesof werealso analyzedfor trace elementcontents using the 0.70800 and 0.71028 for the Eimer & Amend and NBS-987 .aforementionedmethods along with instrumental neutron st.andards,respectively; analyses of BCR-I during the 19,564 LEEMANET AL.: CHEMICAt,DIVERSITY OF CASCADESARC BASALTS

TABLE 1. Distributionof Magma Types in Major Eruptive mingtonite),Fe-Ti oxides,and olivine (rare). Biotite an• Centers of the SWC K-feldsparoccur in somerhyodacites and rhyolites (e.g., SlM). Compositional St. Indian Types Helens Heaven Adams Simcoe

Basalts GEOCHEMISTRYOF LAVAS FROM THE SWC Tholeiitic -- xx x -- Transitional x x x w Perhapsthe most important observation of this study is Calcalkalic x xx xx w the relativelywide range in compositionsof basalticlavas Alkalic • m x xx eruptedwithin a relativelysmall region within the Cascade Basaltic x x x • arc. Selectedanalyses are presentedin Table2 to illustrate the magmaticvariants so far recognized;complete data will Andesites xx x xx x? be presentedin separatepapers discussing each voltaic Dacites, xx m x x rhyodacites, center.Analyses from MSH includethe "transitional"Cave rhyolites Basalt(DS-27, 82-56)from the southflank, calcalkalicba- salts(DS-6, DS-9, and H-SH17 [from Halliday et at., 198:31). Dash, not recognized; x?, present but volume unknown; x, and basalticandesite (DS-72) from the (now obliterated) presentto common; xx, common to abundant. northernflank, andesitic(SH-25) and dacitic (SH-24)te- phras; all of these samples represent the Castle Creel• courseof this studyaveraged 0.70500. The 143Nd/144Nderuptive episode (1700-2200 years B.P.). The most evolved MSH magmas are representedby a rhyodacitic lava from ratiosare reportedrelative to valuesof 0.51182and 0.51264 East dome (DS-62); further analysesof more evolvedMSH for La Jolla Nd oxide and BCR-1, respectively. Oxygen rocks are given by Hoblitt et al. [1980], Ha!liday et al. isotopicanalyses were measuredon whole rock and plagio- claseseparates at Menlo Park (MA), University of Wiscon- [1983],and Smith and Leeman[1987]. The dominantly basalticIH field hasbeen described by Hammond[1980] • sin (IH), the GeologicalSurvey of Japan (MSH), and South- ern MethodistUniversity (SIM). The /5180ratios are Hammond and Korosec [1983], who mapped and named many of the flow units. Analyses are given for low-K reported relative to SMOW with values of 9.6%0for the tholeiitic basalts (DS80A, DS15A-80), ca!c-alkalic ba•ts NBS-28 quartz standard;duplicate analyses indicate preci- (DS9A-81, DS3B-80), basalticandesites with relic amphibote sion of about -+0.1%o.Results of the isotopic analyses are (DS69A-80, DS42-81, and DS23B-81; all from Berry Moun- shownin Figures8-12, and representativeanalyses are given tain), and an olivine (DS-63) which is the most in Table 2. For comparison, only a few Pb, Nd, and Sr evolved sample known from the IH field. Analyses for isotopic analyses have been published previously for SWC Mount Adams include low-K tholeiites (83-49, MA434), volcanoes [Church and Tilton, 1973; Church, 1976; Halliday calc-alkalic basalts (82-31, 82-26, MA226, MA421, MA20), a et al., 1983]; a survey of ore Pb data [Church et al., 1986] basaltic andesire (MA148), and the Aiken andesitc includes samplesfrom the SWC. flow (82-24). Additional major element analyses for lavas from the Mount Adams area are given by Korosec [1987a] MINERALOGY OF L^v^s FROM THE SWC and Anderson [1987a]. Samples from the SIM field include Detailed studies of petrology, mineralogy, and of the --!-4 Ma basalts from peripheral vents from the west (L83- magmaticevolution of the individual eruptive centers have 65, L83-97), southeast (L83-92, L83-94), and south (L83-67, been [e.g., Smith and Leeman, 1987] or will be presented Maryhill flow) and a dacite (L84-65) and rhyolite (L84-67) elsewhere. Table 1 briefly summarizes the distribution of from the eroded core of the volcano near indian Rocks. magmatypes recognizedat the major SWC volcanic centers; Major element analysesof additionalSIM lavas are given:by criteria for subdividing the basaltic variants are discussed Ort et al. [1983] and Anderson [1987a]. Not only are there below. Those from the Simcoe field are notably distinct in differencesamong the variouseruptive centerssampled, bu.t that they commonly carry conspicuousphenocrysts of pla- there are also significantvariations within singleeruptive gioclase, olivine, and occasionallyclinopyroxene; in addi- centers. The wide compositionalspectrum of roughlycon- tion, small xenoliths of spinal lherzolite and diverse crustal temporaneousbasaltic magmasreveals that petrogenetic were found in some of these lavas. Basalts from processesare complex and that not all of these magmaswere the other localities also tend to be porphyritic (plagioclase derived from common source materials. Second, mostof the and olivine +_ rare clinopyroxene), but their phenocrysts young Cascadesbasalts studied differ from thosein other usually are relatively small (usually no more than a few volcanicarcs; in manyrespects they moreclosely resemble millimeters in length); however, basaltswith large (up to a oceanic(or within-plate)basalts despite their obviousasso- few centimeters)plagioclase megacrysts are present at In- ciation with an active subduction complex. dian Heaven. Hydrous phenocrysts are typically absent Major Element Compositionsand Definitions from the mafic lavas. The only known exception is the of Basalt Types notable occurrence of rare, partially resorbed amphibole phenocrystsin one basalticunit from the Indian Heaven field Approximately150 major element analyses are compared (Berry Mountain unit of Hammond and Korosec [1983]). in Figure2, whichshows the discriminantdiagrams (ff Andesitic and dacitic lavas and pyroclasticdeposits domi- Miyashiro[1974] and Gilt [!981]. The AFM (alkalies-FeO*- nate the main stratovolcanoes.These are typically porphy- MgO)diagram (not shown) of Irvineand Baragar [1971] was ritic with ubiquitousplagioclase phenocrysts and assorted usedto subdividethe mafic lavas into tholeiiticand catcfl- combinationsof maficphenocrysts including clinopyroxene kalic magmaseries. Although most samplesstudied are and orthopyroxene,amphibole (hornblendeand/or cum- broadlycalc-alkalic, tholeiitic basalts occur at !H andM.A, LEEMANET AL.: CHEMICAL DIVERSITY OFCASCADES ARCBASALTS 19,565 andtransitional basalts occur at MSH. Figure 3 showsthat distinctionsare further emphasized in Figure5 whereinthe bas•tsfrom MSH, IH, and MA have relativelyhigh Mg # majorityof SWCbasalts (and several from the Oregon High !!00 x molarMgO/[MgO + FeO*]; notethat theseare Cascades[Hughes and Taylor, 1986])have Ta/Yb-Th/Yb mi•mal valuesas somefraction of the iron actuallyoccurs ratiosthat plot along a so-called"mantle array" definedby as Fe;O3);in particular,some from IH resemblehigh-Mg mid-oceanridge (MORB) and oceanic island (OIB) basalts. Aleutianbasalts (Mg # ---70[cf. Gust and Perfit, 1987]).The Onlya few maficlavas and assorted evolved lavas actually 'basalts(and andesites) have relatively highalumina contents plotin thefield of volcanicarc magmas.Those SWC basalts {typicallybetween 16 and 18%). It can be noted in this conformingto the mantlearray display other similarities to diagramthat the basalticlavas with Mg # greaterthan 55 OIB-likemagmas: e.g., low boroncontents (1-4 ppm[Lee- conformrather closely to an idealizedolivine + clinopyrox- man, 1987])and Ba/La (<18 for most [Perfit et al., 1980; ene fractionation trend. In fact, many of these basalts do Morris and Hart, 1983])compared to basaltsfrom virtually containphenocrystic olivine and clinopyroxenealong with all other modernarcs. None of the SWC basalttypes plagioclase,implying evolution at relativelyhigh pressure (> analyzed so far have detectable concentrations of robe -8 kbar [Gust and Perfit, 1987]). However, alkali contents [Morris et al., 1989].In theseregards, the modernCascade vary considerablyin SWC basalts(0.2-2.4% K20' Figure arc appearsto be unusual.The questionarises as to whether 23.!..The SIM basaltsstand out in terms of their compara- thesefeatures reflect unusual conditions of magmagenesis tivelyhigh alkali contents;they comprisealkalic and subal- and, if so, hasthis beena persistentsituation or a relatively kalic groups(cf. Figure 2a) the most potassicof which is recent turn of events. characterized by normarive nepheline. Most other SWC Another remarkable feature of the SWC basalts is their basaltsare hypersthene normarive. compositionaldiversity, which is furtherillustrated in Figure Evolved lavas from MSH, MA, and SIM volcanoes con- 6, where Ba and Nb are plotted versusSiO2. First, the sistof medium-to high-K andesites,dacites, and rhyolites. ranges in Ba (---20x), Nb (---!2x), and BaJNb (---9x) are Volcanicrocks from each center define coherentcomposi- substantial.The SIM alkalic basaltsresemble many OIBs in tionaltrends on silica variation diagrams(cf. Figure 2), but a having high Nb (20-60) and low BaJNb (10-20). Low-K fewMSH "dacites" (oxidizedpumices) have conspicuously tholeiitic basalts from IH and MA have the lowest Ba and Nb h'[ghalumina (Figure 3) and relatively low silica contents.In but have Ba/Nb ratioswhich overlap with thosein the a!ka!ic terms of most other major and trace element constituents basalts.Because both Ba and Nb are stronglypartitioned andtheir mineralogy,these rocks are indistinguishablefrom into basaltic magmas during melting and crystallization otherMSH dacites.Smith and Leemah [ 1987]suggested that processes,it is unlikely that the BaYNbratio can be easily thesefeatures may reflect alteration processes(as opposed fractionated[cf. HoJknann,1988]. Thus there may be some to accumulationof plagioc!asephenocrysts, for example); form of petrogeneticlinkage among all of the SWC basalts Ha!liday et al. [1983] also report one such analysis. In (e.g., they could represent variable degrees of fusion of general,alkalinity of SWC magma suitesincreases eastward commonsource materials). Second, the HFSE-dep!eted IH with distance from the subduction zone, as does the maxi- lavas have elevated Ba and relatively high Ba/Nb ratios mum silica enrichment attained. However, there is consid- (30-90); these lavas are unlikely to be related to the other erablevariability among !H and MA maficlavas which range basalttypes via closedsystem evolution or by partial melting from1ow-K tholeiitic to high-K calcalkalicbasalts. Inspec- of similarsources. Some mechanismis requiredto explain tion of basaltanalyses in Table 2 havingcomparable MgO both their Nb depletionand their Ba enrichment.Nb deple- contents(e.g., 6%) showsthat the major and trace element tion is presumably"source-related" as it is very difficultto compositionsdo not covary in any systematicfashion. conceiveof magmacontamination models that can produce this effect while retaininga basalticbulk composition[cf. Hickey et al., 1986].BaJNb possibly can changewith degree TraceElement Compositions of melting,so longas the sourcecontains a minorphase that Representativetrace element patterns for SWC mafic preferentiallyincorporates Nb [cf. Reagan and Gill, !989]. lavasare shownin Figure4 normalizedto the "primordial It is also useful to consider how these elements are mantle"values of Wood[1979]. These projections facilitate distributedin more evolved lavas to appreciatethe efficacy comparisonsof elementsthat display similar degreesof of various petrogeneticprocesses. Figure 6 clearly shows "incompatibility"relative to dominantmantle (or pheno- that MA and MSH have distinctiveevolutionary patterns. At cryst)mineral assemblages. Most basalticlavas from MSH, MSH, Ba/Nb increasessixfold from basaltsto rhyodacites MA, andSIM, andlow-K IH basaltshave relatively smooth suggestingthat Nb is more compatible than Ba. This is convex-upwardtrace element profiles characteristicof oce- apparently the case at MSH, as Smith and Leeman [1987, anicisland basalts (OIBs). Only a few IH (e.g., DS23B-81, Table 7] calculatedaverage bulk DBa (0.54)

TABLE2. RepresentativeLavas Fr•, ,the

Sample

DS-27 82-56 DS-72 DS-6 DS-9 H-SH17 SH-25 SH-24

Volcano MSH MSH MSH MSH MSH MSH MSH MSH

SiO: 48.64 51.23 52.96 46.96 48.99 49.96 58.0! 64.15 TiO2 1.47 1.55 1.43 2.04 2.10 1.95 1.49 0.67 AI203 17.50 17.45 18.33 ! 6.85 ! 6.70 ! 6.98 16.65 17.37 FeO* 11.69 9.59 8.22 I0.14 11.14 9.37 9.46 5.08 MnO 0.16 0.15 0.13 0.15 0.15 0.15 0.12 0.07 MgO 6.58 6.17 5.46 6.62 6.68 7.00 2.20 !.73 CaO 9.60 9.38 7.96 8.63 8.72 9.01 5.81 4.52 Na20 3.66 3.62 4.27 3.89 3.79 3.89 4.62 4.68 K20 0.56 0.65 0.99 1.31 1.36 1.28 1.32 1.54 P•.O5 0.15 0.20 0.25 0.41 0.37 0.40 0.31 0.17

Cs 0.36 0.38 0.31 0.25 0.61 0.36 1.15 2.06 Ba 118 225 316 312 309 345 35! Rb 11 15 23 25 24 20 31 39 $r 321 363 528 620 590 577 383 427 Y 23.0 27.0 21.3 24.9 26.2 24.0 Zr 132 139 154 203 200 174 189 130 Nb 8.7 11 12.5 32.7 28.5 22 V 210 184 212 214 230 153 69 Ni 85 64 66 97 84 82 21 22 Co 41.7 38.1 28.3 36.3 38.4 18.8 10.4 Cr 175 140 113 196 155 159 74 56 Sc 33.3 31.0 19.1 25.9 28.5 28.2 19.6 8.3 Zn 83 77 74 74 79 72 75 54 La 8.1 10.2 11.3 21.2 19.8 22.6 17.2 11.9 Ce 21.1 25.0 26.8 45.3 45.7 42.0 40.2 24.8 Nd Sm 3.52 4.02 3.42 5.54 5.25 5.05 5.10 2.62 Eu 1.29 1.32 1.21 1.79 1.80 1.84 1.58 0.85 Tb 0.69 0.77 0.49 0.74 0.78 0.78 0.82 0.• Yb 2.35 2.41 1.61 2.11 2.20 2.40 2.61 1.33 Lu 0.33 0.38 0.25 0.33 0.30 0.36 0.40 0.19 Hf 3.03 3.20 3.10 4.30 4.50 4.70 4.90 3.30 Ta 0.57 0.56 0.67 2.35 2.13 contaminated 1.00 0.43 Th 0.94 1.40 1.36 2.47 2.26 3.10 2.59 2.76 B 3.4 5.7 4 5.1 U 0.75 0.82 0.70 Be 1.0 1.5 1.7 1.6 1.7 1.5 Pb 5 6 8 5 8 6

Sr87/86 0.70313 0.70318 0.70304 0.70301 0.70295 Nd143/144 0.512997 0.513003 0.513012 0.513007 0.513035 Pb206/204 18.869 18.812 18.961 18.862 18.756 Pb207/204 15.573 15.577 15.590 ! 5.564 15.546 Pb208/204 38.513 38.495 38.569 38.406 38.361 8180 5.71 6.19 5.79 5.68 5.89

Sample

83-49 MA434 82-26 82-31 MA226 MA421 MA20 MAI48

Volcano MA MA MA MA MA MA MA MA SiO2 48.95 48.38 52.73 5 !.42 50.21 49.34 48.46 54.59 TiO2 1.52 1.57 1.27 1.66 1.62 !.93 1.89 !.29 A1203 16.82 17.35 17.48 17.18 16.11 15.52 16.46 16.13 FeO* 10.69 11.03 8.79 10.06 10.41 10.45 10.23 8.71 MnO 0.17 0.16 0.15 0.17 0.14 0.15 0.15 0.12 MgO 7.89 7. ! 1 6.46 5.78 7.40 8.41 7.94 5.57 CaO 10.40 10.51 8.79 8.94 9.56 9.10 9.16 7.74 Na20 3.04 3.31 3.38 3.64 3.07 3.53 3.73 3.85 K20 0.33 0.37 0.74 0.87 1.05 1.13 1.44 1.56, P2Os 0.17 0.23 0.22 0.27 0.43 0.45 0.55 0.42 Cs 0.21 0.38 0.38 0.35 0.9 Ba 88 90 306 233 404 311 371 527 Rb 4.5 6.3 !0 16 16 19 24 37 Sr 298 357 501 354 862 680 862 783 Y 25.0 27.0 23.0 29.5 27.0 21.5 27.0 25.0 Zr 111 !33 144 208 172 !7! 205 236 LEEMANET AL.: CHEMICALDIVERSIT• Or C^SC^DESARC B^S^LX'S 19,567

SouthernWashington Cascades

Sample

DS-62 DS80A DS 15A-80 DS9A-81 DS3B-80 DS69A-80 DS42-81 DS23B-81 DS63

MSH IH IH IH IH IH IH IH IH 6,9,08 48.30 49.29 49.66 50.53 52.37 52.66 52.21 59.02 0,.35 1.16 1.35 1.10 1.40 1.40 1.13 1.32 0.96 t5.91 17.51 17.91 17.37 17.51 16.78 16.04 16.88 17.86 3.67 10.98 10.59 9.82 9.99 8.08 7.36 7.65 6.50 0.03 0.18 0.18 0.17 0.15 0.!4 0.12 0.13 0.11 0..• 8.37 6.50 8.59 5.85 7.92 8.41 6.02 3.49 322 9.89 10.44 9.51 9.98 8.38 8.49 8.85 6.45 4,86 3.23 3.36 3.10 3.49 3.62 3.62 4.18 4.31 1,76 0.22 0.22 0.48 0.86 0.94 1.74 2.19 1.05 0,!i 0.16 0.15 0.19 0.23 0.37 0.43 0.57 0.26

2.14 0.11 0.21 0.27 46,2 63 57 130 291 268 563 947 314 .q) 2.5 5 6.2 15 13 21.5 22 16 369 259 248 382 512 569 1058 1650 685 14.4 18.8 20.8 21.9 21.4 19.1 14.7 20.5 12.6 167 75 96 107 151 157 184 218 170 8.5 3.8 5.5 6.5 9.7 18.1 7.9 12.3 12 .• 187 211 187 224 184 169 18! !24 5 110 58 133 41 186 183 64 56 5.7 48.4 41.4 42.9 39.1 40.9 34.7 28.3 20.9 23 208 !82 254 14l 373 334 134 81 5,3 29.1 37.9 33.6 37.7 23.4 20.1 20.1 !3.8 56 64 74 67 83 76 85 78 15.1 4.0 3.7 7.9 17.5 17.2 39.7 65.9 13.1 31.2 11.6 10.6 18.7 40.0 41.3 88.9 141.0 30.0 9.9 11.9 25.7 46.7 70.1 3,00 2.26 2.39 2.90 5.77 5.37 7.89 10.90 3.10 0.87 1.01 1.05 1.25 1.87 1.72 2.36 3.22 1.33 O.39 0,52 0.57 0.71 0.83 0.80 0.79 I. 10 0.39 1,24 1.81 2.14 2.51 2.43 1.98 !.50 2.02 1.15 0.19 0.29 0.34 0.40 0.39 0.31 0.22 0.31 O. 16 4.10 1,85 2.30 2.32 4.10 3.95 4.84 5.57 4.00 0.57 0,26 0.33 0.38 0.64 1.33 0.54 0.80 0.64 3.40 0.48 0.76 3.29 2.05 6.85 9.51 1.33 22 0.9 1.! 3.6 3.3 2.9 5.6 1.00 0.60 0.41 0.71 1.90 2.00 1.8 4 6 20 8

0.70391 0.70304 0.70354 O.70333 0,70369 O.70358 O.70350 0.512885 0.512997 0.512930 0,512980 0.512960 0.512'961 0,512912 18.875 18.960 18.820 18.886 18,934 15.577 15,634 15.554 15,550 15,576 38.538 38.790 38.398 38.502 38.609 7.61 5.79 5.53 5.64 5.60 5.88 6.06

Sample

82-24 L83-92 L83-94 L83-65 L83-67 L83-97 L83-62 L84-65 L84-67

: MA SiM SIM SIM SIM SIM SIM SIM SIM

59.41 49.33 49.14 47.04 50.48 49.98 49.07 63.78 74.26 1.24 2.24 2.57 3.58 2.25 2.26 2.54 0.87 0,13 16.74 16.78 15.50 17.04 16,43 17,12 16.77 16.63 13.59 6.76 11.06 11.99 12.53 10.55 !0.97 !2,89 5.10 2.13 0.12 0.15 0.16 0.16 0.16 0,!6 0.19 0.!1 0.03 3.21 6.56 6.88 6.48 6.06 5,80 4.00 1,37 0.08 5.90 9.75 9.39 8.15 7,20 6.50 6,65 3.83 0,22 4.16 3.14 3.22 3.10 4.13 4.28 4.44 4.92 4.71 2.15 0.66 0.76 1.37 2.00 2.17 2.24 3.10 4.83 0.31 0.35 0.39 0.56 0.74 0.76 1.20 0.29 0.02

1.8 ÷ 2.38 0.75 421 181 194 339 554 640 747 425 18 58 8.7 13 23 34 38 38 7,6 159 383 554 497 772 86 ! 838 646 514 4 •,0 23.0 27,0 24,0 27,0 25,0 36.0 34,0 !8,0 3! 1 156 164 205 2,57 262 3,07 355 431 19,568 LEEMAN ET AL.: CHEMICAL DIVERSITY OF CASCADESARC BASALTS

TABLE 2, Sample

83-49 MA434 82-26 82-31 MA226 MA421 MA20 MA148 ,,

Volcano MA MA MA MA MA MA MA

Nb 8.2 10 11.0 15 16 26 33 17 V 178 182 157 179 198 166 178 164 Ni 103 87 80 69 100 166 134 100 Co 50.0 42.0 35.5 40.2 42.0 40.0 30.9 Cr 252 199 160 141 228 278 278 149 Sc 32.0 32.8 27 29.0 26.3 23.3 25.0 •.8 Zn 65 73 74 78 86 84 78 98 La 6.89 10.l 13.9 14.0 30.4 24.6 30.0 39.0 Ce 17.5 22.2 31.5 37.0 60.7 48.0 61.0 77.0 Nd 12.2 17.0 16.4 20.0 34.0 28.0 31.0 37.0 Sm 3.40 4.26 3.85 6.89 6.07 6.45 7.10 Eu 1.24 1.21 1.34 1.87 1.76 1.94 1.79 Tb 0.62 0.77 0.61 0.79 0.71 0.86 0.84 Yb 2.50 2.56 1.98 2.02 1.81 2.20 1.90 Lu 0.33 0.28 0.32 0.30 Hf 2.53 2.94 3.40 3.99 4.05 4.45 5.• Ta 0.49 0.69 0.67 1.01 1.91 2.23 I. 13 Th 0.63 0.82 2.00 3.13 2.44 3.35 7.30 B 1.7 3.4 U 0.22 0.80 0.94 0.80 0.95 1.80 Be Pb 1.0 4 3 4 7 6 7 ! 1 Sr87/86 0.70291 0.70281 0.70345 0.70321 0.70320 0.703-69 Nd143/144 0.512987 0.512860 Pb206/204 18.838 18.688 18.928 Pb207/204 15.565 15.500 15.560 Pb208/204 38.422 38.148 38.475 /5280

, decreasein Ba/Nb with increasingSiO2, all of which seem some form of mixing between compositionally distinctend- more consistentwith crystal fractionationprocesses. This is members. Either melting of previously mixed sourcesor not to say that there is an obvious linkage between evolved contamination of ascending magmas would be reasonable lavas and any of the associated basalts, which are them- causesfor the observednear-linear arrays. Simplevariatiom selves somewhat variable in composition. In passing, the in degree of melting or crystallization are plausiblecauses strongBa depletion seen in the SIM rhyolite is suggestiveof for the different trends only if the two arrays result from extreme fractionation involving alkali feldspar; the high Nb differential involvement of a Nb-buffering residualphase, and very low B•Nb ratio in this rock do not support The SWC basalts which conform to the "mantle array" significantinvolvement of crust(which typically haslow Nb definedby OIB and MORB analysescould representmelting and high Ba/Nb). of a mixed mantle material comprising OIB and MORB Variation in Ba and Nb contentsis displayed directly source components, or possibly magma-wall rock interac- (Figure7a), and alsoas normalizedto Zr, which is a slightly tions involving such mantle components.The absenceof more compatible trace element (Figure 7b). For basaltic significantLILE/HFSE fractionationsin most young$WC magmas,this elementratio plot minimizesscatter related to basalts is unusual for magmas.On the other variationsin degree of crystallizationor partial melting. In hand, the rarer HFSE-depleted basalts could result from both plots two distincttrends emergefor the SWC basaltic combinedfractional crystallization and assimilationof a high lavas. First, a "mantle array" is againevident, here embel- BaJNb contaminant;however, rocks similar to estimated lished by data from representativesamples of OIB (e.g., averagecrustal composition [Taylor and McLennan,1985] Tristan da Cunha lB. L. Weaver et al., 1987) and normal would be inadequate contaminants.Ba enrichmentin arc MORB (NMORB) [Le Roexet al., 1983].These comparative magmascommonly has been attributedto subductionof sampleswere chosen simply becausethey approach ex- Ba-richpelagic sediments [cf. Kay, 1984].As an example,in tremes in Ba and Nb contents of oceanic basalts, but the Figure7b a mixingvector is shownthat headstoward a• light stippledfield in Figure 7b representsa compilationof estimatedaverage composition for Pacificpelagic sediments several hundred Pacific island basalts (W. P. Leeman, un- (e.g., "PAWMS" compositionof Hole et al. [1984]).The publisheddata and literature survey, 1990). Ba/Nb for this data for SWC andesites and more evolved samples• group of SWC basalts is ---11.5. Second, a small number of possiblyconsistent with assimilationof Ba-richsediments HFSE-depleted IH basalts(and a few miscellaneousbasalts becauseSiO2, Ba, and Ba/Zr are all positively correlated i•a from West Crater and the Hood River valley) define a theserocks. Some other explanation seems required for the distinct linear array which trends toward a high Ba/Nb IH basalticlavas. As pointedout by manyauthors [e.g., component.Strong large ion lithophi!eelement (LILE) ver- Pearce,1983; Hole et al., 1984;Tatsumi et al., 1986],hig:h susHFSE correlationsin both of thesearrays could reflect LiLE/HFSE ratios couldalso reflect sourcemetasomatism LEEMANET AL.: CHEMICAL DIVERSITY OFCASCADES ARC'BASALtS 19,569 lcontinued)

Sample

82-24 L83-92 L83-94 L83-65 L83-67 L83-97 L83-62 L84-65 L84-67

MA SIM SIM SIM SIM SIM SIM SIM SIM 25 21 22 37 50 53 57 34 76 114 173 186 195 129 106 80 49 3 • 80 87 33 137 85 18 7 4 20.0 42.3 45.5 46.0 38.1 36.9 32.1 6.7 0.3 32 160 159 32 144 !25 29 13 6 17,0 24.1 24.8 18.4 15.8 15,0 14.4 6.6 0.4 73 102 106 101 96 98 124 78 114 28.7 16.1 17.1 26.2 37.8 40.5 49.1 38.6 21.1 59.9 35.6 37.5 56.1 76.0 78.4 99.4 84.2 40.3 2'6.3 22.0 23.0 32.0 38.0 39,0 54.0 39.0 !6.4 5,69 5.51 6.13 6.68 7.91 7.96 11.00 7.73 3.61 1.50 2.01 2.18 2.36 2.67 2.61 3.49 1.85 0.16 0.77 0.86 1.06 1.00 1.12 1.01 1.46 1.15 0.70 2.43 1.82 2.16 1.98 2.15 2.07 2.90 3.00 2.50 0.37 0.28 0.32 0.28 0.34 0.31 0.44 0.48 0.38 6.40 3.90 4.27 4.80 5.71 5.91 7.02 7.60 12.80 1.60 1.39 !.58 2.70 3.52 3.82 4.25 2.07 4.45 6.80 1.70 1.67 2.64 4.18 4.22 4.61 9.80 18.70 12 2 2.2 1.2 2 1.5 2.20 0.40 0.70 0.70 1.50 1.50 1.40 3.00 4.30

7 8 5 5 7 8 5 7 12

0.70289 0.70287 0.70316 0.70303 0.70302 0.512958 0.512997 0.512923 0.512912 0.512880 19.409 18.680 15.604 15.537 38.773 38.349 6.08 5.84 5.41 5.78 6.36

by LILE-rich "fluids" released during dehydration of sub- mal analytical precision, but much of the observed range ductedoceanic crust. However, it is difficult to see how such (5.4-6.4%0)corresponds to whole rock analysesfor the older large-scaleprocesses could selectively enrich the sourcesof SIM lavas. Further analysesof SIM feldsparseparates will only a few magma batches without similarly affecting very be carded out in the future to see if the whole rock data are largevolumes of the mantle wedge. reliable. Data for IH and MSH whole rock and plagioclase samplesare more restricted, averagingabout 5.8%c,whereas Isotopic Compositions /5t80progressively increases for moresilicic MSH (whole rock or pumice c!ast) samples. Because plagioclase sepa- Sr, Nd, Pb, and 0 isotopic compositionshave been ratesfrom MSH dacitepumices have 8•80 between 6.5 and measuredfor selected samples to further constrain the 6.7%0 (W. Hildreth, unpublished data, 1989), either the sourcecharacteristics of SWC basalts. These data are briefly whole rock trends may result from secondary alteration summarizedhere and will be evaluatedin detail in papersin preparation.The observedranges in/5•80 (5.4-7.7%0), $TSr/ (e.g., hydration) or else the p!agioclasesmay be out of 86Sr(0.7028-0.7040), •43Nd/144Nd (0.51284--0.51304), 2ø6pb/ isotopic equilibrium with their host magma.We tentatively •Pb (1:8.68-19.40),2ø7pb/2ø4pb (15.53-15.66), and 2ø8pb/ favor the latterpossibility because /5180 in theserocks is '•Pb (38.4-39.2)are significant, and these ratios display strongly correlatedwith radiogenicisotope ratios and major importantcorrelations with other geochemical data. Isotopic and trace element contents. We believe that such correla- rangesfor other Cascadevolcanoes are comparable[cf. tionswould be unlikelyif 8180values were seriously mod- Petermanet al., 1970; Church and Tilton, 1973; Church, ified by alterationprocesses' this is indicatedby one of the 1976;Leemah, 1982; Grove et al., 1982,1988; Bacon, 1989]. older MSH dacitesanalyzed by Halliday et al. [1983], which Forcomparison, analyses of Eoceneterrigenous sediments is deletedfrom the figures in thispaper because its 8!80 is fromthe OregonCoast Range (representative of likely extremely low (---4%o).Furthermore, plagioclasesfrom the subductedmaterials or possiblysedimentary sequences in May !8, !980, eruptiondisplay abundanttextural evidence thesubarc crust) have present-day isotopic ratios as follows for multipleepisodes of resorptionand reaction with the host [Petermanet al., 1967, 1981;Heller et al., 1985;W. P. magma [Pearce et al., 1987], and there are documented Leeman,unpublished data, 1989]:/5180 (--• 10%o), 87Sr/S6Sr examples of phenocryst-matrixisotopic disequilibria else- (0.7064}.71!),143Nd/144Nd (0.5!22-0.5124); the range in Pb where in the Cascades [Mazzone and Grant, 19881. As isotopiccomposition of thesesediments is shownin later indicatedin Figure8b, the rate of increasein/5•SO in MSH figures[cf. Church,1976]. rocks exceedsthat predictedfor fractionationof observed TheSWC data are portrayed as a functionof silicacontent phenocrysts. .inFigure 8 to illustratehow they vary with bulk composi- Radiogenicisotopic ratios do not vary consistentlybe- tion.Oxygen compositions of the basalts vary beyond nor- tween the basalts and the more evolved lavas. The basalts 19,570 LEEMANET AL.' CHEMICALDIVERSITY OF Ca.SCADES ARC BASALTS

ß i ß i ß i ß •

. (afterMcDonald & Katsura,19 24 sub-alkalic alkalic 22 I, ! S o 8 o 2O

0 0 • Aleutian 18

16

14

12

10 ' •.'•)I+--Cpx•'•IORB% •

(after Gill, 1981) 8 ß2:01 +P!ag 0 20 40 60 8O 0 in

Fig. 3. A!203-Mg # diagram showing Cascadeslavas relatixe•t fields for mid-ocean ridge basalts IMORB), island arc volcanicrock basalts ,,o • dacites (IAV), and Aleutian high-magnesianbasalts [after Gust and Per 1987].Vectors are showncorresponding to approximatecomposi- tional variations resulting from removal of olivine and either cli- nopyroxene (1, high pressure)or plagioclase(2. low pressure).All MSH sampleswith Mg # near or below 40 and highA!20 3 {>1 are dacitic whole pumice clasts from early eruptive periods.

b

I I [ I I I I each volcanic center are not identical, or perhaps exen

o linked. (afterMiyashiro, 1974) Figure 9 shows isotopic compositions plotted xersu Ba/Nb to emphasize the distinctions in trace element varia- tions noted earlier. For the basaltic rocks, there is little © TH correlationbetween Ba/Nb, /•80, or Pb isotopicratios, although878['/86Sr correlates positively with BaJNb Ba/Zr, not shown). However, this trend belies considerable complexity when it is recalled that Ba/Nb exhibits opposite • o correlations with silica in MSH (+) and MA ½-• evolved lavas.The high Ba/Nb IH basaltshave 87Sr/S"Srratio• similar to those in the more evolved SWC lavas but retain

c relativelyhigh •43Nd/•44Nd. However, it appearsthat LILE HFSE fractionation (i.e., Ba, Sr enrichment) and elexated 45 50 55 60 65 70 75 87Sr/a6Srin these basalticmagmas result from different Si02 processesthan those involved in productionof SWC inter- mediateand silicic magmas.For example, the basaltsdift•r Fig. 2. Variation diagrams showing {a) total alkalies, (b) K20, and (c) FeO*/MgO versusSiO2. Classificationcriteria are discussed in that they display positive correlation betx•eenSr content by Gill [1981]. This plot compares samples from the four major and 87Sr/86Sr,and Sr contentscorrelate positivel,•x•ith volcanic complexes (MSH, IH, MA, and SIM) and from miscella- abundancesof the incompatibletrace elements/e.g., neous minor vents shown by letters in Figure !. A few silicic lavas La, Ba,Th). Thesefeatures suggest that elevation of S7Sr with very high FeO*/MgO plot off the field of view. 86Sroccurred at pressuresabove the stabilityof plagiocla•e such that Sr behaved as an incompatible element (e.g.. subcrustaldepths [cf. Leemahand Hawkesworth,1986]). Isotope-isotope correlations provide additional con- cover nearly the entire range in Sr, Nd, and Pb isotopic straints.Sr-Nd variationsin SWC volcanic rocks(Figure compositions but show no significant correlations between mimicthe fieldsof oceanicbasalts, with more evolvedrock• silica and these ratios. Isotopic variations in the evolved apparentlydisplaced to theright of the "mantlearray" and lavas are more systematic, but differ between volcanoes. In with 143Nd/144Ndshifted slightly to lowervalues than in the MSH lavas,SiO2 is positivelycorrelated with 87Sr/86Sr and basalts.The highBa/Nb lavas are distinctfrom other IH negativelycorrelated with 143Nd/144Nd'a weak negative basaltsin thatthey display increasing 87Sr/$6Sr with onl3 correlationexists between SiO2 andthe Pb isotopicratios. In slightdecrease in •43Nd/144Nd.SIM basaltsdefine a distinct MA lavas SiO2 is poorly (negatively?)correlated with all trendslightly below the "mantlearray," where,t.',MSH these ratios. It is evident that open system processes must evolvedlavas trend slightly to the rightof thearra). Pb have operated to produce evolved SWC magmas and that isotopedata define distinctive positive correlations bcmeen petrogeneses of the basaltic and more evolved magmas at 2ø6pb/2ø4pband both 2ø7pb/2ø4pband 2øspb/2t•pb{Figure !1) LEEMANE¾ AL.: CHEMICAL DIVERSITY OF CASCADES ARC BASALTS 19,571

lOOO 100• Mount St. Helens 'Irid[a'ri n' " DS69A'-80 basalts DS-6 basalts • DS,23•1 T DS-27DS-9 DS3B-80 '"--0--- DS-72 DS42-81 DS9A-81 -t- I-I-SH17 DS80A lOO ,,, ,

!o

I i I i i i i i ! I i i i ! i I i RbTh Ba K NbTa LaCeSr P NdSmZ• Hf '13 Y Yb RbThBa K NbTa LaCeSr P NdSmZY Hf 'n Y Yb

ß I I I i i ii I ,11iii I" I [ i i[ I" [ i I [ I[ i, [ 1 ! Mt.Adams area '";' Simcoebasalts --o- un- --0--- L83-r•, :. basalts ß t MA42! 1 , -- MA434 / • L8•.87 !00 r • MA226 -J 10•

• 10

d i i i I i ! i i i i i I i i I I ! RbThBa K NbTa LaCeSr P NdSmZ• Hf Ti Y Yb RbTh Ba K NbTa LaCeSr P NdSrnZr Hf Ti Y Yb

Fig. 4. "Incompatible" trace element abundancesin selectedSWC basalts(all <53% SiO.,) as normalizedto Wood's [1979] estimatedprimitive mantle composition(model MORB source). Representativesamples showing the rangein basalt compositionsare shownfor each major volcanoas follows: (a) MSH, (b) IH, (c) MA, and (d) SIM. Line profilesare omitted for MA samplesbecause analytical data are lesscomplete and La, Ce, and Nd were measuredby relatively impreciseXRF methods.Note that t•w SWC basaltsdisplay relative depletionsof HFSE (e.g., Nb, Ta) which are characteristicof volcanic arc magmas.Only a few samplesfrom IH {alsofrom someof the miscellaneousvents, not shown) display these features; these samplesalso exhibit relative depletionsof P205. Profiles for the other three volcanoesare much more internally similar.

and lie between fields for oceanic basalts (representedby basaltic magmas may be more susceptibleto effects of PacificMORB [White et al., 1987] and the northernhemi- crustal contamination.Analytical uncertaintiesin the avail- sphere regression line (NHRL [Hart, 1984]) for oceanic able XRF Pb analysespreclude detailed discussion of the basalts)and that for NE Pacificsediments, many of'which systematics.Production of thehigh •r•Pb/2øaPb basalts via havea large componentof terrigenousdetritus [Church, contaminationis possibleif averagecrust resembles the ore 1976].Compositions of Cascadeore leads [Churchet al., lead data but is unlikelyif Pb compositionsof MSH silicic 1986]overlap with the mostradiogenic lavas. Average Juan magmas(Figure 11 [alsocf. Hal!iday et al., 1983])are more deFuca Ridge basalt is indicatedfor reference[Hegner and representativeof possiblecrustal contaminants (cf. Figure Tatsumoto,!987], alongwith the modelcrustal Pb isotope 8a). 'growthcurve of Staceyand Kramers[1975]. The trendof the Alternatively,radiogenic Pb signaturesin SWC basalts Cascadearray suggests that the arc magmas might contain a could be attributed to inheritance from subduction-modified c'•:omponentof Pb similar to that in the sediments or the crust, mantlesource regions [cf. Morris and Hart, 1.983;Hk'key et as was suggestedby Church and Tilton [1973] and Church a!., !986]. For severalreasons this possibilityis considered [1976].However, the new results show that basaltic magmas implausible.Figure 12 shows little or nocorre!at•n between containmore :radiogenicPb than in the associatedsilicic Sr and O isotopiccompositions of SWC basalts,whereas magmas.One possibleexplanation for this trend is that there is a weak negativecorrelation between Pb and O owingto theirlow Pbcontents and higher temperatures, the compositions.Although oxygen isotopic data are not as 19,572 LEE•^N ET AL.' CttEMICAL DIVERSITY OF CASC&DES ARC BASALTS

2.0 lOOO

evolved SWC lavas 800 ß MSH 1.5 A IH O SIM 600 ß OR HC W CAS

1.0 4O0

volcanic arcs 200 0.5

0 high Ba suite 80 0.0 • 0 2 4 6 8

Th/Yb 60 oo o Fig. 5. Th/Yb-Ta/Yb diagramshowing SWC lavas;a few basalts o from the Oligocene-Pliocene Western Cascades (W CAS (W. P. o Leeman, unpublished data, 1990)) and Pleistocene central High o 40 Cascades{OR HC [Hughes and Taylor, 1986])of Oregonare shown o o• for comparisonalong with the field for typical volcanic arc magmas [Pearce, 1983]. Distinct stippled fields are shown for HFSE- depleted{low Ta/Yb) basaltsfrom the IH field and a "mantle array" 20 defined by most other SWC basalts,and by typical oceanicisland basalts(OIB). Samplesplotting between these two fieldsare evolved lavas (>53% SiO,).

i , i , i . i ,

100 - sensitive as Pb isotopes to this form of source contamina- ,.. tion, the observed Pb-O isotopic trend (unlikely to be an 8O artifact of analytical uncertainty) is inconsistent with in- volvement of significant quantities of marine sediments ..,: :: .. whichtypically have high /5•80 [e.g., Heller et al., 1985]. 60 Furthermore,in the SWC lavasas a whole, thereare weakly negativecorrelations between 2ø7pb/2ø4pb and a varietyof 40 other compositionalfeatures that typify sediments(e.g., high BaJZr,Rb/Zr, K/Zr, BaJNb,BaJLa, SiO2, and 87Sff86Sr); ADoo thesecorrelations (not shown)seemingly preclude sediment 20 contaminationof either the mafic magmasor their mantle o ß.:.;:..- • sourceas explanationsfor the high2ø7pb/2ø4pb ratios. Fi- 0 ' i , ! , i . i . i . nally,the high2ø7pb/2ø4pb ratios could be inherited from old 50 55 60 65 70 75 subcontinentallithospheric mantle, but there is no geologic Si02 evidence to support the existence of such a reservoir near the SWC transect. Pb, Sr, and O compositionsof MSH Fig. 6. Variations in {a) Ba, (b) Nb, and {c) Ba/Nb versusSiO., Basaltic lavas define virtually the entire range in Ba and intermediateand silicic rocks are correlatedand suggest that concentrations. Most of these rocks have low BaYNb(<201, tspical thesemagmas contained a significantcrustal component in of OIB' those with higher Ba/Nb are HFSE-depleted IH bas',fit• concert with their proposedcrustal anatecticorigin [Smith (dark stippled field). Evolved lavas from MA {light stipplinglshoa and Leemah, 1987].The SIM basaltsare notable because, Ba and Nb enrichmentsand slight Ba/Nb decreasewith increasin• ratherthan follow the trend of other Cascadesdata in Figure SiO2, consistentwith dominantly fractional crystallizationdifferen- tiation.Those from MSH (mediumstippling) display compositional 11, analyses of these rocks approximate the NHRL line. variationsinconsistent with simplecrystal fractionation(see text}. Thus there is little reasonto suspectthat these magmas containmuch of a subductedslab component. Another line of reasoningagainst significant contribution I and 20). Figure 13 compares B/Nb ratios •ith other of subducted sediment to SWC magmas involves the parameters.All SWC basaltscontain less than 5 ppmB, and geochemistryof boron.Strong correlation of B with løBe average---2 ppm, essentiallyoverlapping with OIB value•. concentrationsin arc magmas[Morris et al., 1990]provides Thosewhich conform to the mantlearrays discussed earlier compelling evidence that the observed B enrichments are have high TaJSm(0.25-0.5) and low B/Nb ratios1<0.2} related to subruction processes.Leemah [1987] demon- similarto thosein O1B. The distinctiveHFSE-depleted IH stratedthat magmasfrom nearly all volcanic arcs are sys- basaltsdisplay higher B/Nb (up to 0.5) andradiogenic Sr, as tematicallyenriched in B relative to within-platemagmas, well as enrichmentsin St, Ba, and other incompatible andthat OIB lavashave low B/Nb ratios(<0.1), whereasarc elements. Even andesiticSWC lavas have B/Nb belox•(}.5. magmashave muchhigher ratios (>0.5, commonlybetween In contrast,MSH daciteshave high B (up to 24 ppm)and LEENIANET AL ' CHEMICALDi•,œRSir• or C XSCADI:.SARC BAS•XLTb 19,.'"3

SWC basaltic lavas mafic magma.,,.A po.,,.,,iblcimplication of theseresults is that lOOO the marie magmasu'ere producedgith little input from the subducted oceanic slab compared to the situation in other Mantle volcanic arcs. array 8oo 0 DISCUSSION e MSH A I-I The foregoingdescriptive information placesa numberof rn Adams constraint?,on the nature of magma generation as,,ocmtcd o SIM 4oo -!- MIsc with Cascades subduction processes. but perhap.-,more importantly illustratesboy, little we reall3 understandthese I TdC ß N-MORB processes. The modern 0ate P!iocene and younger) Ca.,,- 2oo cades magmatism. at least in southern Washington. stands out as beingunusu tl comparedto most volcanic arcs (except as noted below). Some of the most notable feature,are o o 20 40 60 80 lOO summarized here. Compared to many other continental margin arcs [cf. Nb

, Lee,tan. 1983]. basaltic volcanism is relativel3 common in to PAWMS @ Ba!Zr = 62 the SWC. This is apparentlytypical of much of the Cascades arc [e.g., Hughes and Taylor, 1986: Httgh•s, this issue;

4 Gltffanti et al., this issue; Taylor, this issue], although true evolved basalts are not everywhere present [cf. Luedke and Smith, ß MSH & MA 1982; Bacon et al. this issue] and more evolved !a•,xs

3 dominate the major stratovolcanoes.Although Basin and Range style deformation appears to have recently propa- gated northwestward into the southern Cascades and to the

2 backarc region of central Oregon [cf. Wells and Heller, 1988], it seemsunlikely that this developmentis responsible for the abundance of basaltic lavas in the Cascades. First,

1 basaltic magmatism was volumetrically significant in central •(• • soucemtxtng e g Oregon throughout the Neogene (since ---40 Ma) when convergence rates were much higher than at present [Res 0 and Duncan, 1986; t ½rplancl,and Dttncan, 1987; Priest, this 0.0 0.1 0.2 0.3 issue]; initiation of Basin and Range style deformation in the Nb/Zr Great Basin postdates and could not have influenced this early Cascade volcanism. Also, as emphasized here orien- Fig. 7. (a) Ba versus Nb and (b) Ba/Zr versus Nb/Zr in SWC tations of !ate Neogene structures in the SWC are inconsis- basaltic!a,,as (<53% SiO2). For clarity, the compositionalfield for tent v, ith Basin and Range style extension [cf. Zoback and e,ohed MA and MSH rocks is shown only by light stippling. Zoback, 1980]. The abundance of basalt in the central HFSE-depletedIH basalts(dark stippling) and samplesfrom some Cascades may be enhanced by additional thetots such as the o the miscellaneoussmaller vents display Ba and BaIZr enrich- ments tog ard approximate composition of pelagic sediments (as presenceof thinner or more mafic crust relative to adjacent represented by PAWMS average of Hole et al. [1984]). Average segmentsin northern Calilbrnia or north central Washington upper crust (point UC [Tayh)r and McLennan, 1985]) is plotted to [McBirnev, 1978], proximity to segmentation boundaries shog that melting of, or contamination by, such material cannot [Hltg]!c,• t t a!., 1980; Weaver and Michaelson, 1985' explain the high Ba.Zr and Ba/Nb values observed. Remaining xamplesdispiss coupled enrichment in Ba and Nb and define a Gu•htlti and Weaver, 1988], or development of intra-arc "mantle,trra5" trend (medium stippling)similar to that for oceanic extension related to along-strike variations in obliquity of b•'alts (representedby Tristan de Cunha [B. L. Weaver eta!., convergence [Rogers, 1985] or decrease in convergence 1987], BHVO-I Hawaiian tholelite •point H lcf. Govindaraju, rates with time [Verplanck attd Dirtroan, 1987]. At this stage 1984]),NMORB [Lc Roex et al., 1983];the stippledfield coversthe we feel it is premature to criticall) evaluate these possibili- rangelk)r several hundred OIB analyses((W. P. Leeman, unpub- lisheddata and literature compilation, 1990)). The strong!inearity of ties, much lessadopt any singleexplanation for high basaltic thisarras suggests that thisgroup of SWC basaltscould be derived magma productionin the Cascades. b) meltinga mixedreservoir consisting of OIB and MORB source Another striking anomaly concerns the prevalence in the mantlecomponents. SWC of basaltic magmas that have strong geochemical similarities to oceanic island basalts •OIB•; many of those at SIM are true alkali oilvine basalts (some containing sparse B Nb t 1-3). Thu.,,the rangein B/Nb in thesevarious MSH spinal!herzolite xenoliths). Alkali basaltsalso occurin back maema•(along with other isotopicand trace element data) is arc regionsof severalvolcanic arcs (e.g. Japan, Argentina inconsistent,,,•ith cogeneticrelations because magmatic pro- [cf. Gill, 1981; Sat•tiYartlaand Nesbitt, 1986; Stern ½t al., •.cs•csare unlikely to fractionate such highly incompatible 1986; l•llttto: and Stern, 1989]), and it is tempting to drag elements.Near-linear covariation of B/Nb andother param- analogiesbetween the Cascadesand back arc ,,etting,, cter' in the evolved lavas (Figure 13) more likely results especiallyconcerning the SIM basalts.More rare!5, such frommixing between partial melt.,, of the underlyingcrust magmasoccur within intra-arc extensional ones tc.g., New IhighB. possiblyderived from forearc sediments) and low-B Hebrides,Fiji [Gill, 1981,l • 84; (;ill and Whthtn, 1989a,hi). 19,574 LEE MAN ET ,kL.: CttFMICAL DIVERSITY OF CASCADESARC BASALTS

15.65

15.60

O

1535 O O

O a b 1530 L a ...... • ' ' 5 ' • ' ' ' • ' • ' • ' 45 50 55 60 65 70 75 45 50 55 60 65 70 75

03131 ß . ' - '. - ' . ' 0.7040 14'3N•/1•41•d ...... 0.7038 ..??:::.:.•i•::•ii!?•:ii:i?•:•i!!}?.... ß ..-::iis:?:i?:!•:: ":?:iii::ziii!i•.Sl•:•:- 03130• d• 0.7036 .-:•::::iii•?':•?."• ;::•....

03129 0.7032

r• 0.7030

C 03128 ...... 0.7028 45 50 55 60 65 70 75 45 50 55 60 65 70 75 Si02 Si02

Fig. 8. (a) Pb, (b) O, (c) Nd, and (d) Sr isotopiccompositions versus SiO, for SWC volcanic rocks. Basaltic rocks definevirtually the gamut of radiogenic isotopic compositions, butthey are relatively uniform in/5180. HFSE-depleted IH lavas(dark stippling) display significant correlations between SiO2 and '•07~ Pb/204 Pb, •5! 8 O, and 87 Sr/86 Sr. Evolved lavas from MSH (medium stippling) and MA (light stippling) are distinctive (although oxygen isotopic data are not availablefor theMA samples);this is especiallytrue for 87Sr/86Sr trends. Each volcano has experienced open-system evolution, but specific processesmust differ in each case. A vector indicating fractional crystallization (FC) effects is shownfor 8180to illustratethat MSH magmascannot be relatedsimply by thisprocess; they morelikely represent mixtures between silicic crustal melts (i.e., the dacites) and basaltic magmas [cf. Smith and Leemah, 1987]. The 1o• 143Nd/144Nd(also 87Sr/•6Sr = 0.7041)mafic sample from MSH is a gabbroicxenolith from oneof the dacitedome eruptions (Goat Rocks); its relation to other MSH basalts is unclear.

Both OIB-like and typical arc basalts occur in volcanoes monly attributed to the influence of aqueous fluids on phase from Central America [cf. Reagan and Gill, 1989], the equilibria and element transport in the subductionzone western Mexican volcanic belt [Verma and Nelson, 1989; environment(possibly augmented by crustalcontamination Luhr et al., 1989], southwestern Japan and nearby Korea of arc magmas).Specifically, most recent arc petrogenetic and China [Nakamura eta[., 1989], and other arcs with modelspropose that dominantlyaqueous fluids are released young subducting plates [Reagan and Gill, 1989]. Shosho- from the downgoingslab of oceanic lithosphere,and that nitic magmas also occur along the volcanic front in the thesefluids may promotemelting of the slab or, following northern seamount province of the Mariana-Volcano arc upwardmigration, of the mantle wedge. Under highpressure [Stern eta!., 1988]. Unlike these examples, OIB-like basalts conditionsthe fluid phasecould preferentiallyextract from are distributed across the entire SWC traverse and are the the slab and transportsoluble components {e.g., LIL trace rule rather than the exception. elements) into the overlying wedge; melting of suchmetaso- Relative to OIB magmas, for example, most volcanic arc matizedwedge mantle could produce arc magmas[cf. Ellam basaltic lavas are characterized by anomalous enrichments and Hawkesworth, 1988; Tatsumi, 1989]. It has alsobeen of alkali and alkaline earth metals, by depletions of HFSE, proposedthat variablewater activity {for example)could and by high LILE/HFSE (e.g., B',x/Nb)and high LILE light stabilizephases with relatively high partition coefficients for rare earth element (LREE) (e.g., Ba/La) ratios [Green, 1980; Nb, Ta, andTi withinarc magma sources [Saunders et Perfit et al., 1980; ArctdltS and Joltrison, 1981; Morris and 1980;Arctdtts and Powell, 1986];depletion of suchpha•e• Hart, 1983; lto and Stern, 1985; Hickey eta!., 1986; with increasingextent of meltingcould possibly produce Sakuyama and Nesbitt, 1986; Tatsumi et al., 1986; Hildreth rangeof magmatypes with differing degrees of Nbdepletion. and Moorbath, 1988]. As discussed by these authors, the Thecoexistence of OIB-!ike(no Nb depletion)and tb. pical unique chemical characteristics of arc m,tgmas are corn- are (Nb-depleted)basalt types at Turrialbavolcano LEEM^N ET ^L.' CHEMICALD!X, ERSIT'• OF C^SC^DI-•ARC Lt-xb-xl l'b ! , 5

15.65 8 '

15.60

o

15.55

a b 15.50 = .... " ' ' 5 ' ' , , , , , t ß 0 20 40 60 80 100 0 20 40 60 80 100

0.5131 .... 0.7040

0.7038 ß

0.5130 :•,:...c...'.:"'"::" 41 0.7036 :'•..... On ,. ß . 12•a ': .....:..,,::.*.•'• .:::'"*' 0.7034 n" ' Oep"n 0.5129 o 0.7032 0.7030' c d 0.5128 ...... 0.70• ..... • 0 20 40 60 80 100 0 20 4O 60 80 100 Ba/Nb Ba/Nb Fig. 9. Ba./Nb ratio versus (a) Pb, (b) O, (c) Nd, and (d) Sr isotopic compositions. Excepting the high-Ba IH variants (dark stippling), SWC basaltshave BaJNb<20 and display poor correlationswith all isotopic parametersexcept 87Sr/SbSr;they generally resemble typical OIB lavas[cf. Morrisand Hart, 1983].B'a/Nb in HFSE-depletedIH basalts (dark,,tippling) is correlatednegatively with 2ø7pb/'ø4pb and positively v,ith 87SrffhSr,but the, rocksare distinctive from more evolved lavas in the transect (e.g., MA and MSH andesitesand dacites; light stipplingt. Ba.Nb enrichment seems to have no significant correlation with O isotopic compositions.

.Rica)has been attributed to such a process [Reagan and Implications of Basalt Diversity in the SB C Gill, 1989]. There are perhapstwo extreme VlC• points concerningthe Hob ever, there are few SWC basalts with arclike HFSE- origin of basalt divcrsit• in the C'tscadc•. For example, the depletedcompositions regardless of distance from the sub- spectrumof ba.,,altcomposition• may reflect v:tricd degree.,, ductionzone. Those from the 1H field are closely associated of modification (e.g., by wall rock interaction) of relativel} withessentially contemporaneous OIB-like variants.Basalts primitive magmaproduced from a uniform mantle reservoir ha•ing similar compositionaldiversity to that in the SWC [cf. Leemah, 1983; Hildreth and Moorbath, 1988]. At the occur elsebhere in the Cascade arc [cf. Hughes attd Taylor, 1986;Hughes, this issue; C!ynne, this issue; Bullen and other extreme, the compositional variations ms} be source- Clynne,this issue]. Although there seemsto be a higher related and reflect partial melting of heterogeneousmantle proportionof basalts exhibiting HFSE-depletion toward the beneath the arc, in which ca.•,cthe hcterogcneity may result northern[Green, 1981] and southern ends of the arc [Bragg- at !cast partly from subductionprocease.,, [e.g. El!am and man et al., 1987; T. Bullen and M. Clynne, personalcom- ttawl, esworth, 1988;/'lank and Lungmuir, 1988]. To evalu- munication,1.989], the SWC basaltsthat we have studieddo ate these possibilities {or an intermediate position}, •e not appear to bc anomalous with respect to other contem- consider briefly the compo.,,itionalct n•traint.,, provided by poraneous Cascades basalts. Whether OIB-likc bas dis were lavas of the SWC as bell a..,their geologic setting. commonthroughout the history of the arc, or are character- The strongsimilarity to OIB and the absenceof significant isticonly of the modernCascades is an importantquestion correlations between i,,otopic and trace element parameters thatc, nnotbe resolvedwithout further study.Nevertheles.•, (Figures8, 9, and 1•) in mo.,,tSWC basalt• ,tppearto argue in recentgeologic history there appearsto havebeen a against ,,ignificantcrustal contaminationof these mugmss •ariety of basalticsources simultaneously available b ithm {with possibleexception of the Pb isotopicdatum. l'hc rarer x•hat must be a heterogeneousmantle beneath the Ca.,,cade.•. HFSE-depleted basalt,, ma• •n part reflect.some form ot 19,576 LFEMANET AL.: CHEMICALDIVERSITY OF CASCADESARC BASALTS

0.5131 39.2

IMORB t 39.0

0.5130 38.8 0 dacites andesites 38.6 0.5129 38.4

0.5128 38.2

Mantle array 38.0 MSHI 0.5127 37.8 JDF avg o[]A IH/ S,M1 Ja 37.6 0.5126 I ' ' ' '' 18.6 18.8 19.0 19.2 19.4 0.7030 0.7035 0.7040 0.7045 15.70 * * 6Sr Geochron Fig. 10. Sr-Nd isotopiccompositions of SWC lavas relative to typical MORB, estimated "bulk earth," and the "mantle array" 15.65 defined by typical oceanic island basalts [cf. Hofinann, 1988]. Evolved rocks from MA and MSH (stippledfield) are shiftedtoward SK'' ' ' ' I'' ' ' ,-- bulk earth from the basaltic lavas, which closely resemble many 15.60 OIBs. NHRL

ß 15.55 magma contamination process because isotopic composi- Pacific MORB tions (O, Sr, Pb) display weak correlations with bulk com- 15.50 positions {Figures 8, 9, and 12); trends for these basaltsare usually distinct from those of more evolved lavas, indicating • JDFavg b the need for different contaminants in each case. However, 15.45 because the various SWC basalt types have overlapping Mg 18.6 18.8 19.0 19.2 19.4 # and transition metal abundances (cf. Table 2), it is unlikely that their differences in incompatible trace element contents can be explained solely by contamination of a common Fig. 11. (a) 2øspb/2øapband (b) 2ø7pb/2øapbversus -'ø6pb'2ø41• parental magma type. If Nb depletionwith respectto Ba and diagram for SWC lavas relative to the fields of NE Pacificsediments (outlined field [Chttrch, 1976]) and Pacific MORB {dark stipplin La in these lavas is caused by enrichment of LIL and REE [White eta!., 1987]). The regression line for northern hemisphere elements over Nb, then a significantly greater proportion of oceanic basalts (NHRL [Hart, 1984]), and an average of Juandr subducted component is required than can be reconciled Fuca Ridge basalts [Hegner and Tatsttmoto, 1987] also pm•id½ with the relatively minor shifts in isotopic composition estimates of oceanic crust composition. The "geochron" {shown only in Figure I lb) is the locusof modelsingle-stage Pb composi- (especially Pb) and B concentrations associated with the tions, and the Stacey and Kramers, [1975] growth curve (Six• required shifts in BafNb. Thus we conclude that basaltic representsthe time-dependentevolution of conformablePb ore• magmasin this region tapped a heterogeneousmantle source which approximate average crust. Actual Cascadesore lead• that is in part characterized by high LILE/HFSE enrich- [Churcheta!., 1986]plot in the cross-hatchedfield at the radiogemc ments (similar to those in normal arc magma sources) and in endof thevolcanic rock array. The highest 2ø7pb/2ø4pb samples are predominantlybasalts (also see Figure 8a), whereasdacites from part consists of mantle similar to MORB and OIB source MSH (light stippling)plot nearerthe NHRL line. SIM basaltsare material. If the relative volumes of such components are distinctivebecause they fall nearNHRL (in FigureI lb) andexhibit reflected by extrusive proportions of the respective basalt no anomalousenrichment in 2ø7pb/2ø4pb. variants, then the source must be dominated by the MORB- and OIB-like end-members. Furthermore, Nb/Zr ratios cor- relate with magma type, being lowest in the tholeiitic and America at least by latest Cretaceous time. The basement highest in the alkalic basalts (cf. Figure 7). Though specula- was further modifiedby accretionof Paleogeneoceanic tive, this relationship may signify greater contribution of the fragmentsand forearc sediments. As pointedout by AveL•- MORB-like end-member to magmas produced by higher 11emantand 01dow[1988], obliqueconvergence along the degrees of melting, and vice versa. A heterogeneousdistri- Pacificmargin resulted in complexjostling and transport of bution of a lower melting temperature OIB source compo- the pre-Cenozoicterranes. Thus underlying lithospheric nent within a more refractory matrix of MORB source mantle(to the extent that such rocks are preserved belo• the material could explain such melting behavior. Such a model accretedterranes) likely includes scraps and fragments of source is certainly consistentwith the geologic and tectonic subarcand suboceanic mantle domains, perhaps comingled history of the Pacific Northwest margin. by tectonicmixing. We wouldexpect there to be heteroge- The Cascade arc is a relatively young edifice built largely neouslydistributed lithospheric mantle domains having both upon a basement of amalgamated oceanic and island arc OIB and MORB sourceaffinities. Prior to initiationof terranes that were laterally translated and acefeted to North Cascades volcanism, calc-alkalic magmatism was concen- LEEMAN ET AL.: CHEMICAL D!XFR$1T• OF CASCXDt•. ',, ARC Bg.•AI 'rs 19..•"" tratedin broad belts farther to the east [cf. Old(,,' ½tel., 75 1 $ ]. TheChallis volcanic belt tNE Washingtonand central Idaho •as active bet•een 55 and 45 Me, and a younger 70 C!arno-JohnDay-Salmon Creek belt teast centraland SE Oregonto SW Idaho)was active between about 45 and20 65 .Me[•\ortnan et el., 1986;Norman and Leemah,1989: Hart I I.. 19'•9]:the•e volcanicbelts are inferredto be relatedto • 60 ,m ancestralsubduction zone located somewherenear the presentCascade axis. It hasbeen proposed that accretion of 55 the Coast Range block tspecifica!ly the Paleocene-Eocene oceanicisland chain of Duncan [1982]) between 48 and 38 50 \ta led to gradual demise of this older arc and initiation of the modern Cascade subduction zone and volcanic arc [3bllp.•ot1attd Cox, 1977;SnareIv et el., 1980:Duncan and 0.7040 Ixtdrn,1989]. Timing of this event is constrainedby the 43

Ma ageof the youngestCoast Rangevolcanic rocks •,Siletz 0.7038 Ri•er Volcanics [Dttncan, 1982]), and by the 38-40 Ma ages of earliest Cascades volcanic rocks flux, 1982: Phillips et 0.7036 el., 1986: Verplanck and Dttncan, 1987]. A likely conse- quenceof a seawardjump in positionof the subductionzone 0.7034 is that the mantle wedge under the present Cascade arc must 0.7032

15.64 0.7030 b

15.62 0.7028 ,• I i I I ß MSHI OIB 15.60 0.4

15.58 0.3 MSH dacltes & andesires 15.56

15.54

0.1

i I i i , ß.. :'.); HFSE-depletad basalts 15.52 c

i ß 0.0 0.7040 o I 2 .• B/Nb 0.7038 • - &', Fig. 13. (a) SiO2,(b) S"Sr/t't'Sr,and tc) Ta•Sm vcrsu.s B Nbfor SWC volcanic rocks. B Nb ratios in most •,olc,mic arc magma.,, exceed 0.5, whereas OIBs are characterized b5 x.alues belog 0.1 0.7036 .& •l• MSH dacites. ß & andesites fLeeman, 1987,also manuscriptin preparation,1990]; ,also in arc megmssB correlate.,,positively with robe,which is a robust 0.7034 indicatorof incorporationof subductedoceanic sediments and crust [Morris et el., 1990].The absenceof B enrichmentin SVsC basalt• suggeststhat the• containlittle subductedmaterial. Only a few MA 0.7032 lavas are plottedin this figureo•ing to limited B anal• ses.

0.7030 consistesscntiqll• of suboceanic{,i.e., ,L,,theno.,,pheric)man- o o tle, and possiblya remnant of the mantle •edee (perhaps 0.7028 • I ' ' ' ' 5 7 8 metasomatizedby slab-derivedfluids) formcr!• a,,,,,ociated with the ance.•,tral arcs. Thu,; the mantle underlying the Cascadesarc ma3 consistlargely of mixedMORB and OIB Fig. 12. The gitgOversus (a) 2ø7pb/2ø4pband (b) S7Sr/S6Sr.source material tastheno,,phere),depleted mantle (oceanic MSH dec,res(light stippling)and andesires(Three pointswith lithosphere)and remnant,,, of metasomati/edsubarc mantle. SrS•Sr betg. cen 0.7034 and 0.7036) display significant correlations This model providesa context in v,hich the ,,ourcesof betv•eenall threeisotopic parameters. The basaltic lavas (excluding modern Cascade arc m:tgm,t.•,may be evaluated. ttFSE-dcpletedIH laxas; dark stipplingldisplay a weaknegative •.orrelationbetv, een O and Pb compositions,but neitherof these The paucityof typical arc basaltsargues against the correlates,gmficantly with Sr composition. applicabilityof conventional•ubduction one model.,, invol•- 19,578 LEEMANET AL.: CHEMICALDIVERSITY OF CASCADESARC BASALTS ing pervasivemetasomatism of the mantle wedgeby slab- Plio-Pleistocenetime voluminousbasaltic magmas have derivedfluids. If subductioncomponents (fluids, sediments) beenproduced, including tholeiitic, calc-alkalic, and al.kgic were added to the subarc mantle in the past (e.g., during types.The role of crustalcontamination in modifying Cas. more rapid plate convergence),it is difficult to cadebasaltic magmas is consideredto be small,whereas understandwhy high LILE/HFSE sourceshave not been associatedandesitic and daciticmagmas clearly have tappedmore frequentlyby youngSWC magmas.We there- evolvedby opensystem processes and containcrustfl fore conclude that the underlying mantle wedge was not contributions.Some dacites (e.g. MSH) are interpreted• metasomatizedor, if so, this processonly producedlocal crustalanatectic melts formed in responseto stagnation½• heterogeneities(as in relicts of fossilsubarc mantle associ- hot basalticmagmas at or just above the crust-mantle ated with some of the accreted terranes). The apparently boundary(e.g., the MASH zonesof Hildrethand Moorbath minor contribution of slab-derived fluids may relate to the [1988]).The primitive nature of theerupted basalts and their youthfulness(< 10-20 Ma at the trench) and warmth of the relativelycommon occurrence inthe SWC sector both imply downgoingslab [Duncan and Kulm, 1989].Using conductive that suchmagmas had little residencetime in thecrust; t,his thermal models, Abbott and Lyle [1984] concluded that view is supportedby the occurrenceof mantlespinel lher. dehydrationof sucha young slab would likely resultin loss zolite xenolithsin some SIM basalts. The fact that the of volatiles(H20 and CO2) well beforereaching a position basalticmagmas either show no correlation between iso.to• beneaththe volcanicfront (roughly 150 km depth [Rasmus- and trace elementcomponents or showtrends quite disti.nct senand Humphreys,1988]). From petrologicconsiderations, fromthose of theevolved lavas suggests that their com•. Wyilie [1984, Figure 9] and Tatsumi [1989] also concluded sitionalvariability can be attributedto subcrustalprocesses.. that dehydration of hot slab would occur at relatively shal- The majorityof basalticsamples studied are indistinguish- low depths, outboard of the volcanic front. Also, despite able from oceanic island basalts(OIB) in trace element..aM high sedimentationrates in the CascadiaBasin, the size of isotopiccompositions, with exceptionof high2ø?pb/-x•4pb in the accretionaryprism along the NW Pacific margin indi- cates that a large portion of that sediment has not been some. Althoughsimilar Pb isotopicdata are commonly attributed to some form of sediment involvement in other subducted[Snavely et al., 1980; Duncan and Kulm, !989]. volcanic arcs, in the Cascades this interpretationis m Thus, compared to arcs like the Andes, there may be a supported by other observations. The absence of LILE relatively small flux of subducted sediment and negligible input of dehydration fluids to metasomatize the Cascade enrichment and HFSE depletion distinguishesmost SWC basalts from typical volcanic arc magmas. Becausethese magma sources. However, the above view of the Cascade arc raises features are generally attributed to the role of fluidsreleas• important questions about what triggers melting there and by dehydration of subductedoceanic lithosphereand to the the extent to which aqueous fluids are necessary in produc- effects of sediment subduction, we conclude that these ing arc magmasin general. In other words, why does the arc factors are not essential in producing the SWC basalts.We exist? We suggest that the Cascade arc may represent a infer that these basaltsformed by variable degreemelting g "dry" end-member in the continuum of volcanic arcs, in a mixed mantle sourceconsisting mainly of heterogeneously which melting results from adiabatic upwelling associated distributed OIB and MORB source domains. Minor occur- with convective mantle flow (e.g., "corner flow" of Hsui et rences of HFSE-depleted (arclike) basalts in the SWC may al. [1983]). Dehydration of the subducted hot slab and reflect the presence of a small proportion of s!ab- storageof fluids in shallowlithosphere that is not involved in metasomatized subarc mantle (possibly remnants of older magmagenesis would require high temperatures(approach- accretedisland arc terranesor of the Paleogeneforearc). The ing the dry mantle solidus)to produceCascades magmas. In juxtaposition of such different mantle domains within tl•e a sense, more vigorousconvention in the mantle wedge may lithospheric mantle is viewed as a consequenceof (!} tec- be a consequenceof subduction of hot oceanic lithosphere; tonic mixing associatedwith accretionof oceanicand island in this case, mantle isotherms would not be as severely arc terranesalong the Pacificmargin of North Americaprior depressedas in most subductionzones, and higher temper- to Neogene time, and possibly (2) a seawardjump in the atures would be expected below the volcanic front. Our locus of subduction during the mid to late Eocene. suggestionthat the source of most SWC basalts is a mixture Finally, the Cascadearc is unusualis that the subducting of MORB and OIB sourcedomains is not really new, as such oceanicplate is veryyoung and hot. Becausedehydration of components previously have been proposed on other thisslab probably occurs outboard of the volcanicfront, the grounds[e.g., Kay, 19'78;Morris and Hart, 1983; Gill, 1981, role of aqueousfluids in generatingarc magmasis strongly 1984]. However, because the slab contributions so charac- diminishedcompared to the situation at most convergent teristic of most other arcs have been repressed,we are able margins, as judged by compositions of the magmaspro- to discern more clearly the effects of processesoperating in duced.Under low fluidflux conditions,basalt generation ,is the mantle wedge. We speculate that elsewhere such pro- presumablytriggered by other processesthat raisethe cessesmay be overprintedand even masked by slab contri- temperatureof the mantlewedge to its solidus(e.g., convec- butions. tive mantle flow, shear heating,etc.). Althoughfluid- enhancedmelting is an attractivefeature f•)r modelsof subductionzone magmaproduction, it may not be an essen- CONCLUSIONS tial one. The QuaternaryCascades provide a "dry" perspec- The present study documentsthat the geochemistryand tiveon what is surelya widespectrum of arcmodels; inthis tectonic features of the Cascade arc are unusual if not case,reduced slab contributions offer the advantagethat '• unique.Basaltic magmatism has been significantthroughout cansee in moredetail the effects of processesthat occur ia the historyof the arc. Particularly in the SWC sector, since the mantle wedge. I.,EEMANET AL.: CHEMICAl.DIVERSI'tY OF CASCADESARC BASALTS 19,579

Ackm•wledgments.W.P.L. thanksM. A. Menziesfor his aid in Church, S. E., A. P. LeHuray, A. R. Grant, M. H. Detevaux, and arranginga sabbatical visit to theOpen University where much of J. E. Gray, Lead-isotopicdata from sulfide minerals from the thinswork was accomplished. C. J. Hawkesworthkindly provided CascadeRange, Oregonand Washington,Geochim. Cosmochim. •accessto the isotope facilities there, and the Open University staff Acta, 50, 317-328, 1986. providedsupport inmany ways. We are indebted toJ. G. Fitton and Clynne, M. A., Stratigraphic, lithologic, and major element D. James,who provided some critical XRF analysesat the Univer- geochemicalconstraints on magmatic evolution at Lassen vol- sityof Edinburgh, and to TomBullen, who provided Nd isotopic canic center, California, J. Geophys. Res., this issue. a.--•ysesofthe high Ba/Nb IH lavasat theU.S. GeologicalSurvey Duncan, R. A., A captured island chain in the Coast Range of {MealoPark). M. Dehn,E. Pestana,and several students at Rice Oregon and Washington, J. Geophys. Res., 87, 10,827-10,837, Umversityassisted with samplepreparation and ICP analyses. 1982. Oxygenisotopic data were provided by M. Matsuhisa(Geological Duncan, R. A., and L. D. Kulm, Plate tectonic evolution of the Sarveyof Japan)and R. Harmon(Southern Methodist University), Cascadearc-subduction complex, in The Geologyof North Amer- armJ. W. Valley(University of Wisconsin!provided D. R. S. access ica, vol. N, The Eastern Pacific Ocean and Hawaii, edited by E. to hislaboratory for furtheranalyses. W. H. thanksJim O'Neil for L. Winterer, D. M. Hussong, and R. W. Decker, pp. 413-43.8, accessto his stableisotope facilities and Judy Fiersteinfor her GeologicalSociety of America, Boulder, Colo., 1989. o•oingsupport and collaboration in the study of MountAdams. Ellam, R. M., and C. J. Hawkesworth, Elemental and isotopi.c Fin•ally,W. P. L. and D. R. S. acknowledgePaul Hammondfor variations in subduction related basalts' Evidence for a three providinganorientation to southernWashington geology and shar- componentmodel, Contrib. Mineral. Petrol., 98, 72-80, 1988. inghis preliminary geologic maps with us. A largeportion of this Evarts, R. C., R. P. Ashley, and J. G. Smith, Geologyof the Mount researchwas funded by NationalScience Foundation support to W. St. Helens area: Record of discontinuousvolcanic and p!utonic P. L. {grantsEAR82-14876 and EAR85-12172).Z. A. P. was activity in the Cascadearc of southernWashington, J. Geophys. supportedby NERC postdoctoralfellowship. The manuscripthas Res., 92, !0,155-10,169 1987. beenimproved significantly by reviewsof Fred Frey, Mike Ko- Farooqui, S. M., J. D. Beaulieu, R. C. Bunker, D. E. Stensland,and rosec,George Priest, and Mark Reagan.We alsothank Pat Muffler for his role in inspiringus to prepare this paper. R. E. Thoms, Dalles Group: Neogene formations overlying the Basalt Group in north-central Oregon, Ore. Geol., 43, 131-140, 1981. REFERENCES Gill, J., and P. Whelan, Early rifting of an oceanicisland arc (F•i) Abbott,D., and M. Lyle, Age of oceanic plates at subductionand producedshoshonitic to tholeiitic basalts,J. Geophys.Res., 94, volatilerecycling, Geophys. Res. Lett., I1,951-954, 1984. 4561--4578, 1989a. Anderson,J. L., Geologic map of the Klickitat 15' Quadrangle, Gill, J., and P. Whelan, Postsubductionocean island alkali basaltsin Washington,Open File Rep. 87-14, 13 pp., Wash. State Div. of Fiji, J. Geophys.Res., 94, 4579-4588, 1989b. Geol. and Earth Resour., Olympia, 1987a. Gill, J. B., Orogenic Andesites and , 390 pp., Anderson,J. L., Geologic map of the Goldendale 15' Quadrangle, Springer-Verlag,New York, !981. Washington,Open File Rep. 87-15, 9 pp., Wash. State Div. of Gill, J. B., Sr-Pb-Nd isotopicevidence that both MORB and OIB Geol. and Earth Resour., Olympia, 1987b. sourcescontribute to oceanic island arc magmasin Fiji, Earth Arculus,R. J., and R. W. Johnson, Island-arc magma sources:A Planet. Sci. Lett., 68, 443-458, 1984. geochemicalassessment of the roles of slab-derivedcomponents Govindaraju,K., Compilationof working values and samplede- and crustal contamination, Geochem. J., 15, 109-133, 1981. scriptionfor 170international reference samples of mainlysilicate Arculus, R. J., and R. Powell, Source component mixing in the rocks and minerals, Geostand. Newsl., special issue, 8, 3-16, regionsof arc magmageneration, J. Geophys.Res., 91, 5913- !984. 5926, 1986. Green, N. L., Geologyand petrologyof Quaternaryvolcanic rocks, AveLallemant,H. G., and J. S. Oldow, Early Mesozoic southward Garibaldi Lake area, southwesternBritish Columbia: Summary, migrationof Cordilleran transpressionalterranes, Tectonics,7, Geol. Soc. Am. Bull., 92, 697-702, 1981. 1057-1075, 1988. Green, T. H., island arc and continent-buildingmagmatism---A B',•con,C. R., Implicationsof silicic vent patternsfor the presence review of petrogeneticmodels based on experimentalpetrology of largecrustal magma chambers, J. Geophys.Res., 90, 11,243- and geochemistry,Tectonophysics, 63, 367-385,1980. 11,252, 1985. Grove, T. L., D.C. Gedach,and T. W. Sando,Origin of ca!c-alkalic Bacon,C. R., M. A. Lanphere, and J. R. O'Neil, Strontiumand serieslavas at MedicineLake volcanoby fractionation,assimila- oxygenisotopes in volcanicrocks near Crater Lake, Oregon,and tion and mixing, Contrib.Mineral. Petrol., 80, !60-182, 1982. their bearingon arc magmatism,Geological, Geophysical, and Grove, T. L., R. J. Kinzler, M. B. Baker, J. M. Donnelly-Nolan, TectonicSetting of the , edited by L. J.. P. and C. E. Lesher, Assimilationof granite by basaltic magmaat Muffler, C. S. Weaver, and D. D. Blackwell, U.S. Geol. Surv. Burnt lava flow, Medicine Lake volcano, northern California: OpenFile Rep., 89-178, 521-555, 1989. Decouplingof heatand mass transfer, Contrib. Mineral. Petrol., Blackwell,D. D., R. G. Bowen, D. A. Hull, J. F. Riccio, and J. L. 99, 320-343, 1988. Steele, Heat flow, arc volcanism, and subduction in northern Guffanti, M., and C. S. Weaver, Distributionof late Cenozoic Oregon,J. Geophys.Res., 87, 8735-8754, 1982. volcanicvents in the CascadeRange: Volcanic arc segmentation Brique.u,L., H. Bougault,and J. L. Joron,Quantification of Nb, Ta, andregional tectonic considerations, J. Geophys. Res., 93, 65!3- Ti, andV anomaliesin magmasassociated with subductionzones: 6529, 1988. petrogenicimplications, Earth Planet Sci. Lett., 68, 297-308, Guffanti,M., M. A. Clynne,J. G. Smith,L. J.P. Muffler,and T. D. !984. Bullen,Late Cenozoicvolcanism, subduction, and extensionin Bruggman,P. E., C. R. Bacon,P. J. Aruscavage,R. W. Lerner, the Lassenregion of California,southern Cascade Range, J. L. J. Schwartz, and K. C. Stewart, Chemical analyses of rocks Geophys.Res., this issue. andglass separates from Crater Lake National Park and vicinity, Gust,D. A., and M. R. Perfit,Phase relations of a high-Mgbasalt Oregon,U.S. Geol. Surv.Open File Rep., 87-56,36 pp., 1987. from the AleutianIsland Arc: implicationst•r primaryisland arc Bullen,T. D., and M. A. Clynne, Trace elementand isotopic basaltsand high-A! basalts, Contrib. Mineral. Petrol., 97, 7-!8, constraintson magmaticevolution at Lassenvolcanic center, J. 1987. Geophys.Res., this issue. Ha!liday,A. N., A. E. Fallick,A. P. Dickin,A. B. Mackenzie, Cam•11, N. P., andR. D. Bentley,Late Quaternary deformation W. E. Stephens,and W. Hildreth,The isotopicand chemical of the ToppenishRidge uplift in southcentral Washington, evolutionof Mount St. Helens, Earth Planet. Sci. Lett., 63, Geology,9, 519-524, 1981. 241-256, t983. C.hurch,S. E., The CascadeMountains revisited: A re-evaluationin Hammond,P. E., A tectonicmodel for theevolution of theCascade lightof new lead isotopicdata, Earth Planet Sci. Lett., 29, Range,in Pact.'ficCoast Pa!eogeography Symposium, vol. 3, 175-188, 1976. CenozoicPaleogeography of the Western:United States, edited Church,S. E., and G. W. Tilton, Lead and strontiumisotopic by J. M. Armentrout,M. R. Cole,and H. J. R. Terbest,pp. st•.tMiesin the CascadeMountains: Beating on andesRegenesis, 2!9-237, PacificSection, Society of EconomicPaleontot•sts Gecd.Soc. Am, Bull., 84, 431-454, 1973. and Mineralogists,Los Angeles,Calif., 19'79. 19,580 LEEMANETAL..' CHEMICAL DIVERSITY OFCASCADES ARCBASALTS

Hammond,P. E., Reconnaissancegeologic map and cross sections Washington,Open File Rep.87-5, 39 pp., Wash.State Div. Geol. of southernWashington Cascade Range, lat. 45ø30'--47ø15'N, and Earth Resour., Olympia, 1987a. long.120ø45'-122ø22.5'W, Dep. of EarthSci., PortlandState Korosec,M. A., Geologicmap of the HoodRiver Quadmngle• Univ., Portland, Oreg., 1980. Washingtonand Oregon, Open File Rep.87-6, 42 pp., Wash. e Hammond,P. E., andM. A. Korosec,Geochemical analyses, age Div. Geol. and Earth Resour., Olympia, 1987b. dates,and flow-volume estimates for Quaternary volcanic rocks, Leeman,W. P., Tectonicand magmaticsignificance of stro•t•m southernCascades Mountains, Washington, Open File Rep. 83- isotopicvariations in Cenozoicvolcanic rocks from the westera 13,36 pp., Wash. State Div. Geol. and Earth Resour., Olympia, United States, Geol. Soc. Am. Bull., 93,487-503, 1982. 1983. Leeman,W. P., The influenceof crustalstructure on composi Hart,S. R., A large-scaleisotopic anomaly in thesouthern hemi- of subduction-relatedmagmas, J. Volcanol.Geotherm. Res., 18, sphericmantle, Nature, 309, 753-757,1984. 561-588, 1983. Hart, W. K., M.D. Norman,R. W. Carlson,and W. P. Leemah, Leeman,W. P., Boron geochemistryof volcanicarc m Andesitcpetrogenesis in the northwesternBasin and Range: Evidence for recycling of subductedoceanic lithosphere,Eas Implicationsfor the geochemicalevolution of subcontinentalTrans. AGU, 68, 462, 1987. mantle,Geol. Soc. Am. Abstr.Programs, 21, 91, 1989. Leeman, W. P., and C. J. Hawkesworth, Open magmaticsystems-- Hegner,E., andM. Tatsumoto,Pb, Sr, and Nd isotopesin basalts someisotopic and trace elementconstraints, J. Geophys.Res., andsulfides from the Juande FucaRidge, J. Geophys.Res., 92, 91,5901-5912, 1986. 11,380--11,386, 1987. Le Roex,A. P., H. J. B. Dick, A. J. Erlank,A.M. Reid,F. A. Frey, Heller,P. L., Z. E. Peterman,J. R. O'Neil, andM. Shafiqullah, and S. R. Hart, Geochemistry,mineralogy and petrogenesis Isotopicprovenante of sandstonesfrom the Eocene Tyee Forma- lavas erupted along the southwestIndian Ridge betweenthe tion,Oregon Coast Range, Geol. Soc. Am. Bull.,96, 770-780, Bouvet triple junction and 11 degreesEast, J. Petrol., 24, 267- 1985. 318, 1983. Hickey,R. L., F. A. Frey,D.C. Gerlach,and L. Lopez-Escobar,Luedke, R. G., and R. L. Smith, Map showingdistribution, corn. Multiplesources for basalticarc rocks from the SouthernVol- position,and age of late Cenozoicvolcanic centers in Oregon canicZone of the Andes(34ø--41øS): Trace element and isotopic Washington,U.S. Geol. Surv. Misc. Invest. Map, I-I09!D, 1982. evidencefor contributionsfrom subductedoceanic crust, mantle Luhr, J. F., J. F. Allan, I. S. E. Carmichael,S. A. Nelson,and T• andcontinental crust, J. Geophys.Res., 91, 5963-5983,1986. Hasenaka, Primitive calc-alkalineand alkaline rock typesf Hildreth,W., andJ. Fierstein,Mount Adams: Eruptive history of an the western Mexican volcanic belt, J. Geophys. Res., 94, 451• andesite-dacitestratovolcano at the focus of a fundamentally 4530, 1989. basalticvolcanic field, U.S. Geol. Surv.Open File Rep., 85-521, Lux,D. R., K-Arand 40Ar.•Ar ages of mid-Tertiary volcanic rocka 44-50, 1985. from the western Cascade Range, Oregon, Isochron West,33, Hildreth,W., and S. Moorbath,Crustal contributions to arc mag- 27-32, 1982. matismin the Andesof centralChile, Contrib. Mineral. Petrol., MacDonald, G. A., and T. Katsura, Chemical composition 98, 455-489, 1988. Hawaiian lavas, J. Petrol., 5, 82-133, 1964. Hoblitt,R. P., D. R. Crandall,and D. R. Mullineaux,Mount St. Mazzone,P., and N. K. Grant, Mineralogicaland isotopic evid Helenseruptive behavior during the last 1500 years, Geology, 8, for phenocryst-matrixdisequilibrium in the Gamer Mo.mtaia 555-559, 1980. andesitc,Contrib. Mineral. Petrol., 99, 267-272, 1988. Holmann,A. W., Chemicaldifferentiation of the Earth:The rela- McBirney,A. R., Volcanicevolution of the CascadeRange, Ann•t. tionshipbetween mantle, continental crust, and oceanic crust, Rev. Earth Planet. Sci., 6, 437-456, 1978. Earth Planet. Sci. Lett., 90, 297-314, 1988. McBirney,A. R., and C. M. White, The CascadeProvince, 'at Hole, M. J., A.D. Saunders,G. F. Marriner, and J. Tarney, Andesites:Orogenic Andesites and Related Rocks,edited b• Subductionof pelagicsediments: Implications for the originof R. S. Thorpe,pp. 115-135,John Wiley, New York, 1982. Ce-anomalousbasalts from the Mariana Islands, J. Geol. Soc. McKee,E. M., W. A. Duffield,and R. J. Stem,Late Miocene earlyPliocene basaltic rocks and their implications for cru.stal London, 141,453-472, 1984. structure,northeastern California and south-centralOregoa, Hsui, A. T., B. D. Marsh,and M. N. Toksoz,On meltingof the Geol. Soc. Am. Bull., 94, 292-304, 1983. subductedoceanic crust: Effects of subductioninduced mantle Michaelson,C. A., andC. S. Weaver,Upper mantle structure from flow, Tectonophysics,99, 207-220, 1983. teleseismicP wave arrivalsin Washingtonand northernOreg•, Hughes,J. M., R. E. Stoiber,and M. J. Cart,Segmentation of the J. Geophys.Res., 91, 2077-2094,1986. Cascadevolcanic chain, Geology,8, 15-17, 1980. Miyashiro,A., Volcanicrock seriesin islandarcs and Hughes,S.S., Mariemagmatism and associated tectonism of the continentalmargins, Am. J. Sci., 274, 321-355,1974. centralHigh CascadeRange, Oregon, J. Geophys.Res., this Morris,J. D., andS. R. Hart, Isotopicand incompatible elemea! issue. constraintson the genesis of islandarc volcanics from Cold Bay Hughes,S.S., andE. M. Taylor,Geochemistry, petrogenesis, and andAmak Island,Aleutians, and implications for mantle tectonicimplications of centralHigh Cascademarie platform ture,Geochim. Cosmochim. Acta, 47, 2015-2030,1983. lavas, Geol. Soc. Am. Bull., 97, 1024--1036,1986. Morris,J. D., W. P. Leeman,and F. Tera,The subduction compo- Irvine, T. N., and W. R. A. Baragar,A guideto the chemical nentin islandarc lavas: Constraints from Be isotopesand B-lle classificationof the commonvolcanic rocks, Can. J. Earth Sci., 8, systematics,Nature, 331, 31-36, 1990. 523-548, 1971. Mullineaux,D. R., Summaryof pre-1980tephra-fall deposits Ito, E., andR. J. Stern,Oxygen and strontiumisotopic investiga- eruptedfrom Mount St. Helens, Washington state, USA, tions of subductionzone volcanism: the case of the Volcano Arc Volcanol., 48, 17-26, 1986. and the Marianas Island Are, Earth Planet. Sci. Lett., 76, Munoz,J. andC. R. Stern,Alkaline magmatism within the segmeat 312-320, 1985. 38ø-39øSof the Plio-Quaternaryvolcanic belt of the Kay, R. W., Aleutianmagnesian andesites: Melts from subdueted SouthAmerican continental margin, J. Geophys.Res., 94, 4545- PacificOcean crust, J. Volcanol.Geotherm. Res., 4, 117-132, 4560, 1989. 1978. Nakamura,E., I. H. Campbell,and M. T. McCulloch,Che.m• Kay, R. W., Elementalabundances relevant to identificationof geodynamicsin a backarc region around the Sea of magmasources, Philos. Trans. R. Soc.London Ser. A, A310, Implicationsforthe genesis ofalkaline basalts inJapan, Korea, 535-547, 1984. andChina, J. Geophys.Res., 94, 4634-4654,1989. Kim, K., S. A. Dischler,J. R. Aggson,and M.P. Hardy,The state Nakamura,K., Volcanoesaspossible indicators of tectonic st•ss of in-situ stressesdetermined by hydraulicfracturing at the orientation,J. Volcanol.Geotherm. Res., 2, 1-16,1977. HanfordSite: Richland, Washington, Rep. RHO-BW-ST-73P, 172 Norman,M.D., andW. P. Leeman, Temporal evolution of pp., RockwellHanford Oper., Richmond, Wash., 1986. zoic-Cretaceousmagmatism andits relation to tectonicsett'mg, Klein, F. W., Earthquakesin LassenVolcanic National Park, southwesternIdaho, U.S.A., Earth Planet. Sci. Lett., 94, 78-96, California,Bull. Seismol.Soc. Am., 69, 867--875,1979. 1989. Korosec,M. A., Geologicmap of the MountAdams Quadrangle, Norman,M.D., K. R.McElwee, R. A. Duncan,and W. P. •, LEEMANET AL.: CHEMICAL DIVERSITY OFCASCADES ARC'BASALlIS !9,581

K-Arages of OligoceneSalmon Creek Volcanics,Owyhee Moun- Smith,D. R., andW. P. Leeman,Petrogenesis of Mount St. Helens t•ns, SW Idaho• lsochron West, 46, 15-19, 1986. daciticmagmas, J. Geophys.Res., 92, !0,313-10,334,1987. Oldow,J. S., A. W. Bally, H. G. AveLallemant,and W. P. Leeman, Smith, G. A., Late Cenozoicstructures on the ColumbiaPlateau: PMnerozoicevolution of the North AmericanCordillera {United Implicationsfor tectonicsin the PacificNorthwest, Eos Trans. States and Canada!, in The Geology of North America: An AGU, 63, 1116, !982. Overview,edited by A. W. Bally and A. R. Palmer,pp. 139-232, Smith,G. A., L. W. Snee,and E. M. Taylor, Stratigraphic, GeologicalSociety of America, Boulder,Colo., 1989. sedimentologic,and petrologic record of lateMiocene subsidence (•M,K. M., R. W. Tabor,and V. A. Frizell, Jr., Chemicalanalyses of thecentral Oregon High Cascades, Geology, 15,389-392, 1987. of selectedTertiary and Quaternary volcanic rocks, Cascade Snavely,P. D., Jr.,H. C. Wagner,and D. L. Lander,Interpretation Range,Washington, U.S. Geol. Surv. OpenFile Rep., 83-1, 1-14, of the Cenozoicgeologic history, central Oregon continental 1983• margin:Cross-section summary, Geol. Soc.Am. Bull., 91, !43- Pearce,J. A., Roleof subcontinentallithosphere in magmagenesis 146, 1980. at active continental margins, in Continenta! Basalts and Mantle Stacey,J. S., andJ. D. Kramers,Approximation of terrestriallead Xenoliths,edited by C. J. Hawkesworthand M. J. Norry, pp. isotopeevolution by a two-stagemodel, Earth Planet. Sci. Lett., 23{)--249,Shiva, Cambridge, Mass., 1983. 26,207-221, 1975. Pe•arce, T. H., J. K. Russell, and I. Wolfson, Laser-interferenceand Stern, C. R., K. Futa, S. Saul, and M. A. Skewes,Nature and Normarskiiinterference imaging of zoningprofiles in plagioclase evolutionof thesubcontinental mantle lithosphere below southern p•Mnocrystsfrom the May 18, 1980, eruption of Mount St. SouthAmerica and implications for Andeanmagma genesis, Rev. Helens, Washington, Am. Mineral., 72, 1131-1143, 1987. Geol. Chile, 27, 41-53, 1986. Pe•t, M. R., D. A. Gust, A. E. Bence, R. J. Arculus, and S. R. Stem, R. J., S. H. Bloomer, P.-N. Lin, E. Ito, and J. Morris, Taylor,Chemical characteristics of island-arcbasalts: Implica- Shoshoniticmagmas in nascentarcs: New evidencefrom subma- fi•onsfor mantle sources, Chem. Geol., 30, 227-256, 1980. rine volcanoesin the northernMarianas, Geology, 16, 426-430, Peterman,Z. E., C. E. Hedge,R. G. Coleman,and P. D. SnareIcy, 1988. Jr., 87Sr/•Srratios in someeugeosynclinal sedimentary rocks Tatsumi,Y., Migrationof' fluid phases and genesis of basaltmagmas andtheir bearing on the originof graniticmagma in orogenicbelts, in subductionzones, J. Geophys.Res., 94, 4697-4707,1989. Earth Planet. Sci. Lett., 2, 433-439, 1967. Tatsumi,Y., D. L. Hamilton,and R. W. Nesbitt,Chemical charac- Peterman,Z. E., I. S. E. Carmichael,and A. L. Smith,Sr87/Sr 8•' teristicsof fluidphase released from a subductedlithosphere and ratiosof Quaternarylavas of the CascadeRange, northernCali- originof arc magmas:Evidence t¾om high-pressure experiments fornia, Geol. Soc. Ant. Bull., 81,311-318, 1970. and natural rocks, J. Volcano!. Geotherrn. Res., 29, -"03-309, Peterman,Z. E., R. G. Coleman, and C. E. Bunker, Provenanceof 1986. Eocenegraywackes of the Floumoy Formationnear Agness, Taylor, E. M., Volcanichistory and tectonicdevelopment of the Oregon•a geochemicalapproach, Geology, 9, 81-86, 1981. central High CascadeRange, Oregon,j. Geophys.Res., this •illips, W. M., M. A. Korosec, H. W. Schasse,J. L. Anderson, issue. and R. A. Hagen, K-Ar ages of volcanic rocks in southwest Taylor, S. R., and S. M. McLennan, The Continental Crust: Its Washington,lsochron West, 47, 18-24, 1986. Compositionand Evolution, 312 pp., Blackwell Scientific, Bos- Plant,T., and C. H. Langmuir, An evaluationof the global varia- ton, Mass., !985. tionsin the major element chemistry of arc basalts,Earth Planet. Verma, S. P., and S. A. Nelson, Isotopic and trace element Sci. Lett., 90, 349-370, 1988. constraints on the origin and evolution of alkaline and calc- Priest,G. R., Overview of the geologyof the central Oregon alkaline magmasin the northwesternMexican volcanic belt, J. CascadeRange, in Geology and geothermalresources of the Geophys. Res., 94, 4531-4544, 1989. centralOregon Cascade Range, edited by G. R. Priest, and B. F. V erplanck,E. P., and R. A. Duncan,Temporal variations in plate Vogt, Spec.Pap. Oreg. Dep. Geol. Miner. Ind.., 15, 3-28, 1983. convergenceand eruptionrates in the WesternCascades, Oregon, Priest, G. R., Volcanic and tectonic evolution of the Cascade Tectonics, 6, 197-209, 1987. volcanicarc, 44000' to 44ø52.5'N,J. Geophys.Res., this issue. Walker, J. R., and H. R. Naslund,Tectonic significance of mildly •smussen,J., and E. Humphreys,Tomographic image of the Juan alkaline Pliocenelavas in Klamath River gorge,Cascade Range, de Fuca plate beneathWashington and westernOregon using Oregon, Geol. Soc. Am. Bull., 97, 206-212, 1986. teleseismicP wave traveltimes, Geophys. Res. Lett., 15, 1417- Weaver, B. L., D. A. Wood, J. Tamey, and J. L. Joron, Geochem- 1420, !988. istry of ocean island basaltsfrom the South Atlantic: Ascension, Rea,D. K., andR. A. Duncan,North Pacificplate convergence: a Bouvet, St. Helena, Gough, and Tristan da Cunha, in Alkaline quantitativerecord of the past 140 m.y., Geology,I4, 373-376, IgneousRocks, edited by J. G. Fitton and B. J. Upton, SpecPubl. 1986. Geol. Soc. London, 30, 253-267, 1987. Reagan,M. K., andJ. B. Gill, Coexistingcalcalkaline and high- Weaver, C. S., and G. E. Baker, Geometryof the Juande Fucaplate niobiumbasalts from Turrialba volcano, Costa Rica: Implications beneath Washington---Evidencefrom seismicity and the !949 South Puget Sound earthquake,Bull. Seismol. Soc. Am., 78, forresidual titanates in arcmagma sources, J. Geophys.Res., 94, 264-275, 1988. 4619-4633, 1989. Weaver, C. S., and S. D. Malone, Overview of the tectonic setting Ri,ddihough,R. P., Recent movementsof the and recentstudies of eruptionsof Mount St. Helens, Washington, system,J. Geophys. Res., 89, 6980-6994, 1984. J. Geophys. Res., 92, 10,149-10,154, 1987. Rogers•G. C., Variationin Cascadevolcanism with margin orien- Weaver, C. S., and C. A. Michaelson, Seismicityand volcanismin tation,Geology, 13, 49.5-498,1985. the PacificNorthwest: Evidencefor the segmentationof the Juan Sakuyama,M., andR. W. Nesbitt,Geochemistry of the Quaternary de Fuca Plate, Geophys, Res. Lett., 12, 215-218, 1985. volcanicrocks of the northeastJapan arc, J. Volcanol.Geotherm. Weaver, C. S., W. C. Grant, and J. E. Shemeta, Local crustal Res.,29, 413--450,1986. extensionat Mount St. Helens, Washington,J. Geophys.Res., 92 Saunders,A.D., J. Tarney,and S. D. Weaver,Transverse 10,170--10,178, 1987. geochemicalvariations across the AntarcticPeninsula: lmplica- Wells, R. E., and P. L. Heller, The relative contribution of accre- tionstbr the genesis of calc-alkalinemagmas, Earth Planet. Sci. tion, shear, and extension to Cenozoic tectonic rotation in the Lett., 46, 344--360,1980. Pacific Northwest, Geol. Soc. Am. Bull., !00, 325-338, 1988. Sheppard,R. A., Geologyof theSimcoe Mountains volcanic area, White, W. M., A. W. Hofrnann, and H. Puchelt, Isotope geochem- Washington,Geol. Map GM-3,Wash. State Div. of Minesand istry of Pacific mid-ocean ridge basalt, J. Geophys. Res., 92, Geol.,Olympia, 1967. 4881-4893, 1987. Shenod,D. R., andJ. G. Smith,Quaternary extrusion rates from Williams, D. L., D. A. Hull, H. D. Ackerman, and M. H. Beeson, the CascadeRange, northwestern United States and British The Mount Hood region: Volcanic history, structure, and geo- Columbia,J. Geophys. Res., this issue. thermal energy potential,J. Geophys, Res., 87, 2767-2781, !982. Simpson,R. W., andA. V. Cox,Pa!eomagnetic evidence for Wood, D. A., A variably veined sub-oceanicupper mantle,--geaetic tectonicrotation of theOregon Coast Range, Geology, 5,585-589, significancefor mid-oceanridge basaltsfrom geochemicalevi- !977. dence, Geology, 7, 499-503, 1979, 19,582 LEEMAN ET AL.: CHEMICAL DIVERSITY OF CASCADESARC BASALTS

Wyllie, P. J., Constraintsimposed by experimentalpetrology on Z. Palacz,VG Isotopes Limited, Ion Path, Road 3 Winfiord, possible and impossible magma sources and products, Philos. Cheshire,CW7 3BX, England. ' Trans. R. Soc. London Ser. A, A310, 439-456, 1984. N. Rogers,Department of Earth Sciences, Open Univer•ty, Zoback, M. L., and M.D. Zoback, States of stress in the conter- Milton Keynes,MK7 GAA, England. minous United States,J. Geophys.Res., 85, 6113-6156, 1980. D. R. Smith,Geology Department, Trinity University, San tonio, TX 78284.

W. Hildreth, U.S. Geological Survey, 345 Middlefield Road, Menlo Park, CA 90425. (Received May 11, 1989- W. P. Leemah, Keith-Wiess Geological Laboratories, Rice Uni- revisedNovember 3, 1989; versity, Houston, TX 77251. acceptedNovember 6, 1989.)