Volatiles in glasses from Mauna Loa Volcano,Hawai'i: implications for magma degassing and contamination,and growth of Hawaiian volcanoes

Glasses from Mauna Loa pillow basalts, recent subaerial vents, and inclusions in olivine were analyzed for S, Cl, F, and major elements by electron microprobe. Select submarine glasses were also analyzed for H₂O and CO₂ by infrared spectroscopy. The compositional variation of these tholeiitic glasses is dominantly controlled by crystal fractionation and they indicate quenching temperatures of 1,115-1,196 °C. Submarine rift zone glasses have higher volatile abundances (except F) than nearly all other submarine and subaerial glasses with the maximum concentrations increasing with water depth. The overwhelming dominance of degassed glasses on the submarine flanks of Mauna Loa implies that much of volcano's recent submarine growth involved subaerially erupted lava that reached great water depths (up to 3.1 km) via lava tubes. Anomalously high F and Cl in some submarine glasses and glass inclusions indicate contamination possibly by fumarolic deposits in ephemeral rift zone magma chambers. The relatively high CO₂ but variable H₂O/K₂O and S/K₂O in some submarine rift zone glasses indicates pre-eruptive mixing between degassed and undegassed magma within Mauna Loa's rift system. Volatile compositions for Mauna Loa magmas are similar to other active Hawaiian volcanoes in S and F, but are less Cl-rich than Ll'ihi glasses. However, Cl/K2O ratios are similar. Mauna Loa and Ll'ihi magmas have comparable, but lower H₂O than those from Kilauea. Thus, Kilauea's source may be more H₂O-rich. The dissimilar volatile distribution in glasses from active Hawaiian volcanoes is inconsistent with predictions for a simple, concentrically zoned plume model.     

Trace element abundances of high-MgO glasses from Kilauea, Mauna Loa and Haleakala volcanoes, Hawaii

We performed an ion-microprobe study of eleven high-MgO (6.7–14.8 wt%) tholeiite glasses from the Hawaiian volcanoes Kilauea, Mauna Loa and Haleakala. We determined the rare earth (RE), high field strength, and other selected trace element abundances of these glasses, and used the data to establish their relationship to typical Hawaiian shield tholeiite and to infer characteristics of their source. The glasses have trace element abundance characteristics generally similar to those of typical shield tholeiites, e.g. L(light)REE/H(heavy)REE_C1 < 1. The Kilauea and Mauna Loa glasses, however, display trace and major element characteristics that cross geochemical discriminants observed between Kilauea and Mauna Loa shield lavas. The glasses contain a blend of these discriminating chemical characteristics, and are not exactly like the typical shield lavas from either volcano. The production of these hybrid magmas likely requires a complexly zoned source, rather than two unique sources. When corrected for olivine fractionation, the glass data show correlations between CaO concentration and incompatible trace element abundances, indicating that CaO may behave incompatibly during melting of the tholeiite source. Furthermore, the tholeiite source must contain residual garnet and clinopyroxene to account for the variation in trace element abundances of the Kilauea glasses. Inversion modeling indicates that the Kilauea source is flat relative to C1 chondrites, and has a higher bulk distribution coefficient for the HREE than the LREE. Received: 4 February 1997 / Accepted: 27 August 1997     

Ages,rare-earth element enrichment,and petrogenesis of tholeiitic and alkalic basalts from Kahoolawe Island,Hawaii

Kahoolawe Island, Hawaii (18×11 km), is a basaltic shield volcano with caldera-filling lavas, seven identified postshield vents, and at least two occurrences of apparent rejuvenated-stage eruptive. We examined 42 samples that represent all stages of Kahoolawe volcano stratigraphy for their petrography, whole-rock major-and trace-element contents, mineral compositions, and K–Ar ages. The two oldest shield samples have an average age of 1.34±0.08 Ma, and four postshield samples (3 are alkalic) average 1.15±0.03 Ma; ages of 1.08 and 0.99 Ma for two additional tholeiitic samples probably are minimum ages. Whole-rock major- and trace-element and mineral compositions of Kahoolawe shield and caldera-fill laves are generally similar to the lavas forming Kilauea and Mauna Loa tholeiitic shields, but in detail, Kahoolawe shield lavas have distinctive compositions. An unusual aspect of many postshield Ka-hoolawe lavas is anomalously high REE and Y abundances (up to 200 ppm La and 175 ppm Y) and negative Ce anomalies. These enrichments reflect surficial processes, where weathering and soil development promoted REE-Y transport at the weathering front. Major element abundances (MgO, 10–6 wt.%) for shield and caldera-fill basalts are consistent with fractionation of ol+px+pl in frequently replenished magma reservoirs. In general, tholeiitic basalts erupted from late vents are higher in SiO₂ than the shield lavas, and temporal differences in parental magma compositions are the likely explanation. Alkalic basalts that erupted from vents are comparable in composition to those at other Hawaiian volcanoes. Trace-element abundance ratios indicate that alkalic basalts represent either relatively lower degrees of melting of the shield source or a distinct source. Apparent rejuvenated-stage basalts (i.e., emplaced after substantial Kahoolawe erosion) are tholeiitic, unlike the rejuvenated-stages at other Hawaiian volcanoes (alkalic). Kahoolawe, like several other Hawaiian volcanoes, has intercalated tholeiitic and alkalic basalts in the postshield stage, but it is the only volcano that appears to have produced tholeiitic rejuvenated-stage lavas.     

Petrogenesis of Tholeiitic Lavas from the Submarine Hana Ridge, Haleakala Volcano, Hawaii

Hana Ridge, the longest submarine rift zone in the Hawaiianisland chain, extending from Maui 140 km to the ESE, has a complexmorphology compared with other Hawaiian rift zones. A totalof 108 rock specimens have been collected from the submarineHana Ridge by six submersible dives. All of the rocks (76 bulkrocks analyzed) are tholeiitic basalts or picrites. Their majorelement compositions, together with distinctively low Zr/Nb,Sr/Nb, and Ba/Nb, overlap those of Kilauea lavas. In contrast,the lavas forming the subaerial Honomanu shield are intermediatein composition between those of Kilauea and Mauna Loa. The compositionalcharacteristics of the lavas imply that clinopyroxene and garnetwere important residual phases during partial melting. The compositionsof olivine and glass (formerly melt) inclusions imply that regardlessof textural type (euhedral, subhedral–undeformed, deformed)olivine crystallized from host magmas. Using the most forsteriticolivine (Fo_90·6) and partition coefficients     

Boron isotopic constraints on the source of Hawaiian shield lavas

Boron isotopic compositions of lavas from three representative Hawaiian shield volcanoes (Kilauea, Mauna Loa, and Koolau) were analyzed by thermal ionization mass spectrometry. The boron isotopic composition of each sample was analyzed twice, once with and once without acid leaching to evaluate the effect of posteruptive boron contamination. Our acid-leaching procedure dissolved glass, olivine, secondary zeolite, and adsorbed boron; this dissolved boron was completely removed from the residue, which was comprised of plagioclase, pyroxenes, and newly formed amorphous silica. We confirmed that an appropriate acid-leaching process can eliminate adsorbed and incorporated boron contamination from all submarine samples without modifying the original ¹¹B/¹⁰B ratio. On the other hand, when the sample was weathered, i.e., the olivine had an iddingsite rim, ¹¹B/¹⁰B of the acid-resistant minerals are also modified, thus it is impossible to get the preeruptive ¹¹B/¹⁰B value from the weathered samples. Through this elimination and evaluation procedure of posteruptive contamination, preeruptive δ¹¹B values for the shield lavas are −4.5 to −5.4‰ for Koolau (N = 8), −3.6 to −4.6‰ for Kilauea (N = 11), and −3.0 to −3.8‰ for Mauna Loa (N = 6).Historical Kilauea lavas show a systematic temporal trend for B content and Nb/B coupled with other radiogenic isotopic ratios and trace element ratios, at constant δ¹¹B, indicating little or no assimilation of crustal materials in these lavas. Uncorrelated B content and δ¹¹B in Koolau and Mauna Loa lavas may also indicate little or no effect of crustal assimilation in these lavas. The source of KEA-component (identical to the so-called Kea end member in Hawaiian lavas) of the Hawaiian source mantle, represented by Kilauea, should be derived from lower part of subducted oceanic crust or refractory peridotite in the recycled subducted slab. The systematic trend from Kilauea to Koolau—decreasing δ¹¹B coupled with decreasing εNd as well as increasing ⁸⁷Sr/⁸⁶Sr and ²⁰⁶Pb/²⁰⁴Pb—is consistent with involvement of subducted sediment components in the EMK(enriched Makapuu)-component, represented by Makapuu-stage of Koolau lavas.     

Isotope Compositions of Submarine Hana Ridge Lavas, Haleakala Volcano, Hawaii: Implications for Source Compositions, Melting Process and the Structure of the Hawaiian Plume

We report Sr, Nd, and Pb isotope compositions for 17 bulk-rocksamples from the submarine Hana Ridge, Haleakala volcano, Hawaii,collected by three dives by ROV Kaiko during a joint Japan–USHawaiian cruise in 2001. The Sr, Nd, and Pb isotope ratios forthe submarine Hana Ridge lavas are similar to those of Kilauealavas. This contrasts with the isotope ratios from the subaerialHonomanu lavas of the Haleakala shield, which are similar toMauna Loa lavas or intermediate between the Kilauea and MaunaLoa fields. The observation that both the Kea and Loa componentscoexist in individual shields is inconsistent with the interpretationthat the location of volcanoes within the Hawaiian chain controlsthe geographical distribution of the Loa and Kea trend geochemicalcharacteristics. Isotopic and trace element ratios in Haleakalashield lavas suggest that a recycled oceanic crustal gabbroiccomponent is present in the mantle source. The geochemical characteristicsof the lavas combined with petrological modeling calculationsusing trace element inversion and pMELTS suggest that the meltingdepth progressively decreases in the mantle source during shieldgrowth, and that the proportion of the recycled oceanic gabbroiccomponent sampled by the melt is higher in the later stagesof Hawaiian shields as the volcanoes migrate away from the centralaxis of the plume. KEY WORDS: submarine Hana Ridge; isotope composition; melting depth; Hawaiian mantle plume     

Chemical compositions of Kilauea east-rift lava, 1968-1971

The major element chemical compositions of lava from four eruptionson the east rift zone of Kilauea between August 1968 and October1971 reflect three petrologic processes:

1. Production of chemically distinct batches of magma in the mantle.
2. Separation of olivine, augite, and plagioclase from liquidduringflow in the rift-zone conduits.
3. Mixing of differentmagmas during ascent to the surface.

Chemically none of the four Kilauea east-rift eruptions matchesthe preceding summit eruption in Halemaumau that ended in July1968. The Mauna Ulu eruption, May 1969 to October 1971 (thelast of flie east-rift eruptions), can be divided into fiveolivine-controlled and chemically distinct variants. Three ofthese characterize the first seven months of the eruption andare closest in composition to the 1967–8 Halemaumau eruption.Variants 4 and 5 were erupted later and have compositions thatare distinctly different from that of the 1967–8 eruption.Major differences are higher Al₂O₃ (0?15–0?23 per cent),and lower K₂O (0?07–0?10 per cent) and TiO₂ (0?12–0?23per cent) in variants 4 and 5 at the same MgO content. Somelavas from eruptions in August and October 1968 and February1969, have olivine-controlled magma compositions that are identicalto mixtures of Mauna Ulu variants 1–3 and the 1967–8composition. This observation fits an hypothesis advanced earlierby T. L. Wright and R. S. Fiske that magmas in the central magmachamber become mixed with magmas in the rift zone and can beidentified as mixing components of rift eruption magmas beforethey appear as distinctive magmas in summit eruptions. Lavas representing mixing of olivine-controlled magma with differentiatedmagma were erupted in October 1968, February 1969, and in Mayand December 1969. The changes in amount of K₂O and TiO₂ during the latter partof the 1969–71 Mauna Ulu eruption are the reverse of theoverall secular change in composition of Kilauea summit lavasfrom pre-1750 through 1967–8. The K₂O and TiO₂ contentsof the latest overflows during the 1969–71 Mauna Ulu eruption(April 1971) are comparable to that of lava erupted at Kilaueasummit prior to 1750. The changing chemistry of Kilauea magma is found to be of useas a ‘tracer’ in the complex Kilauea conduit system.Application of these data to older lava sequences is difficulbecause of the complexity of the processes controlling lavacomposition and the absence of detailed information about thetime-space chemical variation during individual eruptions.     

A Quantitative Insight into the Dependence Dynamics of the Kilauea and Mauna Loa Volcanoes,Hawaii

Hawaiian volcanoes such as Kilauea and Mauna Loa have drawn the attention of researchers for quite some time and numerous theories abound hinting at a possible inverse relationship between the two. Most of these analyses are intrinsically qualitative and are bereft of data-driven statistical justification. The present work attempts to address this issue adopting a more mathematical approach and endeavours to examine the existence of such a relationship through the novel use of a smoothing statistic termed as the empirical recurrence rates ratio. Additionally, it is shown that useful knowledge about the possible interplay between these two volcanoes is coded into this single statistic and based on it; construction of new dependence measures such as the two introduced, becomes simpler and much more intuitive. The recent decade is witnessing an increased activity of Kilauea and the methods proposed here can be successfully implemented to safeguard human lives and property against the unpredictable advances of all-engulfing molten lava flow.     

Origin of the Differentiated and Hybrid Lavas of Kilauea Volcano, Hawaii

Kilauea Volcano has erupted lava from its summit caldera andfrom two rift zones that extend from the summit towards theeast and south-west. Lavas erupted from the summit of the volcanodiffer from each other principally in their content of olivineand define lines of ‘olivine control’ on magnesiavariation diagrams. Lavas erupted on the rift zones may be similarin composition to the summit lavas or may be differentiatedby processes that involve minerals other than olivine. All ofthe differentiated lavas have less than 6·8 per centMgO and plot off the extension of olivine control lines forthe summit lavas. Prehistoric vents (before A.D. 1750) fromwhich differentiated lavas have been erupted are found on theeast rift zone and in the western Koae fault zone adjacent tothe south-west rift zone; historic vents for differentiatedlavas are confined to the east rift zone. Twenty-one new analysesare presented for several of the east rift differentiates andfor the newly discovered differentiates adjacent to the south-westrift zone. The differentiates have MgO as low as 3·9per cent and SiO₂ as high as 56 per cent; both extremes arefound in the prehistoric lavas adjacent to the south-west rift. Detailed petrochemical studies suggest the following conclusions:

1. Thechemical composition of magma erupted at Kilauea summitvarieswith the date of eruption. Lavas erupted before 1750,duringthe eighteenth and nineteenth centuries, and in the twentiethcentury form groups that can be distinguished chemically. Ona lesser scale, each Kilauea summit eruption in the twentiethcentury has a chemistry that is distinctive with respect tothe chemistry of every other summit eruption.
2. During lateprehistoric time pockets of differentiated magmawere formedwithin the rift zones by separation of the liquidremainingafter partial crystallization of bodies of summitmagma. Thisprocess presumably is still going on within theeast rift zone,but the more recently separated liquids havenot yet been eruptedto the surface. The relative time at whichthese differentiatedmagmas were produced can be estimated fromcalculations basedon their chemical compositions, which showthat the differentiatescould lie on the liquid line of descentfor Kilauea summit magmaof prehistoric composition but noton any liquid line of descentfor younger summit magmas.
3. Lava from some eruptions, notablythe early part of the 1955eruption on the lower east rift,has the composition of theliquid fraction as it is generatedwithin the rift. Lava compositionsof other eruptions, includingthose of the later lavas of 1955,are best explained by mixingof magma supplied from a centralreservoir beneath Kilauea summitwith the differentiated liquidin the rift. Lava from each summiteruption is unique chemically,so it is possible to recognizeits presence or absence as componentsof mixing in such mixedlavas. It appears that summit magmaof composition characteristicof the 1952 and 1961 Halemaumaueruptions contributed to thecomposition of the mixed lavasproduced in the latter part ofthe 1955 eruption. Summit magmaof 1961 composition is alonesufficient to explain the compositionof mixed lavas eruptedin 1960 and 1961. In rift lavas eruptedfrom 1962 to 1965, thecomposition of lava erupted in Halemaumauin 1967, in additionto the 1961 composition, is a componentof mixing, and it isthe dominant summit component in the compositionof the two1965 eruptions. The proportion of summit magma todifferentiatedmagma needed to explain the composition of lavaserupted onthe upper east rift increases from 1961 to 1965;this increaseindicates that the differentiated magma was beingdiluted andused up by repeated flooding of this part of therift zone bymagma supplied from the central reservoir.
4. The fact that componentsof ‘summit composition’appear in rift eruptionsbefore they appear undiluted in Halemaumausuggests that thecentral reservoir is vertically zoned. Rifteruptions are fedfrom lower levels where younger magma is available,and summiteruptions are fed from the relatively older magmaabove. Thechemical distinction between lava of successive summiteruptionsimplies that significant convective mixing of magmadoes nottake place throughout the central reservoir.
5. The unique anduniform composition of lava of each successivesummit eruptionalso suggests that summit eruptions end whenall of the magmaof one composition has been erupted. The magmaerupted fromthe upper levels of the reservoir during one cycleis continuallyreplaced from below by younger magma of differentcomposition.In order for eruption to be renewed in Halemaumau,new magmafrom the mantle must be held in storage at intermediatelevelsbefore it attains an ‘eruptive state’.
6. The hypothesispresented in 2–4 above permits qualitativepredictionsconcerning future lava compositions. The compositionof thenext lava to be erupted in Halemaumau is expected tobe distinctfrom that of the 1967 eruption, and this compositionwill presumablybe identified in rift eruptions occurring between1967 and thetime of its appearance in Halemaumau.
7. Differentiates of prehistoricage also were apparently formedin the same way as those ofhistoric age, but the mixing cannotbe described quantitativelybecause of poor control on the stratigraphyand the compositionsof erupted lavas. One lava in the Koaegroup, that from YellowCone, appears to be a mixture of a picriticmagma (12 per centMgO) with a differentiated liquid with lessthan 2·5per cent MgO and nearly 60 per cent SiO₂.

     

Volatiles in Basaltic Glasses from Loihi Seamount, Hawaii: Evidence for a Relatively Dry Plume Component

New H₂O, CO₂ and S concentration data for basaltic glasses fromLoihi seamount, Hawaii, allow us to model degassing, assimilation,and the distribution of major volatiles within and around theHawaiian plume. Degassing and assimilation have affected CO₂and Cl but not H₂O concentrations in most Loihi glasses. Waterconcentrations relative to similarly incompatible elements inHawaiian submarine magmas are depleted (Loihi), equivalent (Kilauea,North Arch, Kauai–Oahu), or enriched (South Arch). H₂O/Ceratios are uncorrelated with major element composition or extentor depth of melting, but are related to position relative tothe Hawaiian plume and mantle source region composition, consistentwith a zoned plume model. In front of the plume core, overlyingmantle is metasomatized by hydrous partial melts derived fromthe Hawaiian plume. Downstream from the plume core, lavas tapa depleted source region with H₂O/Ce similar to enriched Pacificmid-ocean ridge basalt. Within the plume core, mantle components,thought to represent subducted oceanic lithosphere, have waterenrichments equivalent to (KEA) or less than (KOO) that of Ce.Lower H₂O/Ce in the KOO component may reflect efficient dehydrationof the subducting oceanic crust and sediments during recyclinginto the deep mantle. KEY WORDS: basalt; Hawaii; mantle; plumes; volatiles     

Volatiles in pillow rim glasses from Loihi and Kilauea volcanoes,Hawaii

Volatiles and major elements in submarine glasses from Loihi seamount and Kilauea volcano. Hawaii were analyzed by high temperature mass spectrometry and the electron microprobe. Loihi glasses are subdivided into three groups: tholeiitic, transitional and alkali basalts. The glasses are evolved: Mg numbers range from 48–58. The alkalic lavas are the most evolved.Total volatiles range from 0.73 to 1.40 wt.%. H₂O shows a positive linear correlation with K₂O content [H₂O = 0.83 (± .09) K₂O + 0.08 (± .06)]. Concentrations of H₂O are higher in the alkalic lavas, but Cl and F abundances are highly variable. Variations in ratios of incompatible elements (K₂O, P₂O₅, H₂O) indicate that each group was derived from a distinct source. CO₂ contents range from 0.05 to 0.19 wt.% but show no systematic correlation with rock type or Mg #. A well-defined decrease in glass CO₂ content with increasing vesicularity is shown by the alkalic lavas. CO₂ may have been outgassed from the tholeiitic and transitional magmas prior to eruption during storage in a shallow magma chamber. Reduced carbon species (CO and CH₄) were found in small amounts in most of the alkalic samples. Although the redox histories of Hawaiian lavas are poorly known, these new data indicate the presence of a reduced source for Loihi magmas.The Kilauea tholeiitic glasses are evolved (Mg # 48.3 to 55) and have higher H₂O contents (av. 0.54 wt.%) than Loihi tholeiites (av. 0.42 wt.%) at the same Mg # (~55). Cl is distinctly lower in Kilauea glasses (0.01 wt.%) compared to Loihi glasses (0.09 wt.%). The data indicate significant source differences for the two volcanoes, consistent with results of other geochemical studies.Loihi tholeiites have distinctly higher ³He/⁴He ratios than Kilauea tholeiites and are the highest measured in submarine basalts (KURZ et al., 1983). These high ratios have been used to invoke a primitive source for Loihi basalts. The high Cl content of these basalts, the highest we have ever measured in submarine basalts, may be a fingerprint of this primitive source, as previously noted for Icelandic basalts (Schillinget al. 1980).     

Empirical estimation of magnetite/liquid distribution coefficients for some transition elements

Magnetite/liquid distribution coefficients have been calculated that are consistent with observed variations in contents of V, Sc, Cr, and Ti in lavas from Craters of the Moon lava field, Idaho. In particular, our average distribution coefficient for V (27±8) is in close agreement with that determined experimentally at the temperature and oxygen fugacity appropriate for crystallization of these lavas. Although this value is not very precise, it shows that V contents of at least some calcalkaline series are consistent with generation of andesitic and related magmas from basaltic parental magmas by crystal fractionation, involving removal of appreciable amounts of magnetite.Current address: Geology Department, Rice University. Houston, Texas 77001, USA     

Perspectives on basaltic magma crystallization and differentiation: Lava-lake blocks erupted at Mauna Loa volcano summit, Hawaii

Explosive eruptions at Mauna Loa summit ejected coarse-grained blocks (free of lava coatings) from Moku'aweoweo caldera. Most are gabbronorites and gabbros that have 0–26 vol.% olivine and 1–29 vol.% oikocrystic orthopyroxene. Some blocks are ferrogabbros and diorites with micrographic matrices, and diorite veins (≤ 2 cm) cross-cut some gabbronorites and gabbros. One block is an open-textured dunite.

The MgO of the gabbronorites and gabbros ranges  7–21 wt.%. Those with MgO > 10 wt.% have some incompatible-element abundances (Zr, Y, REE; positive Eu anomalies) lower than those in Mauna Loa lavas of comparable MgO; gabbros (MgO < 10 wt.%) generally overlap lava compositions. Olivines range Fo_83–58, clinopyroxenes have Mg#s  83–62, and orthopyroxene Mg#s are 84–63 — all evolved beyond the mineral-Mg#s of Mauna Loa lavas. Plagioclase is An_75–50. Ferrogabbro and diorite blocks have  3–5 wt.% MgO (TiO₂ 3.2–5.4%; K₂O 0.8–1.3%; La 16–27 ppm), and a diorite vein is the most evolved (SiO₂ 59%, K₂O 1.5%, La 38 ppm). They have clinopyroxene Mg#s 67–46, and plagioclase An_57–40. The open-textured dunite has olivine  Fo_83.5. Seven isotope ratios are ⁸⁷Sr/⁸⁶Sr 0.70394–0.70374 and ¹⁴³Nd/¹⁴⁴Nd 0.51293–0.51286, and identify the suite as belonging to the Mauna Loa system.

Gabbronorites and gabbros originated in solidification zones of Moku'aweoweo lava lakes where they acquired orthocumulate textures and incompatible-element depletions. These features suggest deeper and slower cooling lakes than the lava lake paradigm, Kilauea Iki, which is basalt and picrite. Clinopyroxene geobarometry suggests crystallization at < 1 kbar P. Highly evolved mineral Mg#s, < 75, are largely explained by cumulus phases exposed to evolving intercumulus liquids causing compositional ‘shifts.’ Ferrogabbro and diorite represent segregation veins from differentiated intercumulus liquids filter pressed into rigid zones of cooling lakes. Clinopyroxene geobarometry suggests < 300 bar P. Open-textured dunite represents olivine-melt mush, precursor to vertical olivine-rich bodies (as in Kilauea Iki). Its Fo_83.5 identifies the most primitive lake magma as  8.3 wt.% MgO. Mass balancing and MELTS show that such a magma could have yielded both ferrogabbro and diorite by ≥ 50% fractional crystallization, but under different fO₂: < FMQ (250 bar) led to diorite, and FMQ (250 bar) yielded ferrogabbro. These segregation veins, documented as similar to those of Kilauea, testify to appreciable volumes of ‘rhyolitic’ liquid forming in oceanic environments. Namely, SiO₂-rich veins are intrinsic to all shields that reached caldera stage to accommodate various-sized cooling, differentiating lava lakes.     

Primary alkaline magmas associated with the Quaternary Alligator Lake volcanic complex,Yukon Territory,Canada

The Alligator Lake complex is a Quaternary alkaline volcanic center located in the southern Yukon Territory of Canada. It comprises two cinder cones which cap a shield consisting of five distinct lava units of basaltic composition. Units 2 and 3 of this shield are primitive olivine-phyric lavas (13.5–19.5 cation % Mg) which host abundant spinel lherzolite xenoliths, megacrysts, and granitoid fragments. Although the two lava types have erupted coevally from adjacent vents and are petrographically similar, they are chemically distinct. Unit 2 lavas have considerably higher abundances of LREE, LILE, and Fe, but lower HREE, Y, Ca, Si, and Al relative to unit 3 lavas. The ⁸⁷Sr/⁸⁶Sr and ¹⁴³Nd/¹⁴⁴Nd isotopic ratios of these two units are, however, indistinguishable. The differences between these two lava types cannot be explained in terms of low pressure olivine fractionation, and the low concentrations of Sr, Nb, P, and Ti in the granitoid xenoliths relative to the primitive lavas discounts differential crustal contamination. The abundance of spinel lherzolite xenoliths and the high Mg contents in the lavas of both units indicates that their compositional differences originated in the upper mantle. The Al and Si systematics of these lavas suggests that, compared to unit 3 magmas, the unit 2 magmas may have segregated at greater depths from a garnet lherzolite mantle. The identical isotopic composition and similar ratios of highly incompatible elements in these two lava units argues against their differences being a consequence of random metasomatism or mantle heterogeneity. The lower Y and HREE contents but higher concentrations of incompatible elements in the unit 2 lavas relative to unit 3 can be most simply explained by differential partial melting of similar garnet-bearing sources. The unit 2 magmas thus appear to have been generated by smaller degrees of melting at a greater depth than the unit 3 magmas. The contemporaneous eruption of two distinct but volumetrically restricted primary magmas from adjacent vents at the Alligator Lake volcanic complex suggests that volcanism in this region of the Canadian Cordillera is controlled by localized, small batch processes.     

Recycled oceanic crust in the Hawaiian Plume: evidence from temporal geochemical variations within the Koolau Shield

The subaerial surface of Koolau volcano is composed of lavas that define the distinctive endmember composition for Hawaiian shield lavas, known as the Koolau component, now designated as the Makapuu-stage. The geochemical characteristics of lavas recovered by the Koolau Scientific Drilling Project (KSDP) show that this distinctive composition forms a <300-m thick veneer. Below this veneer, from ~300m to 470 m below sea level, Koolau shield lavas transition to a composition similar to Mauna Loa lavas, now designated as the Kalihi-stage. This transition was gradual, occurring over >80 ka; therefore it was not caused by an abrupt event, such as a landslide. Among all Koolau shield lavas, there are correlations between radiogenic isotopic ratios of Sr, Nd and Pb and compositional characteristics, such as SiO₂ content (adjusted to be in equilibrium with Fo₉₀ olivine), Sr/Nb, La/Nb and Th/La. These long-term compositional and isotopic trends show that as the shield aged, there was an increasing role for an ancient recycled marine sediment component (<3% of the source) accompanied by up to 20% SiO₂-rich dacitic melt. This melt was generated by partial melting of garnet pyroxenite, probably kilometers in size, that formed from recycled basaltic oceanic crust. In detail, time series analyses of depth profiles of Al₂O₃/CaO, Sr/Nb, La/Nb and Th/La in the KSDP drill core show correlations among these ratios indicating that recycled oceanic crust contributed episodically, ~29 ka period, to the magma source during the prolonged transition from Kalihi- to Makapuu-stage lava compositions. The long-term geochemical trends show that recycled oceanic crust was increasingly important as the Koolau shield moved away from the plume and encountered lower temperature.     

The redox states of basic and silicic magmas: a reflection of their source regions?

At present the best estimates of the oxygen fugacity of spinel-lherzolites that could be the source material of basic magmas is about five log units below the Ni–NiO buffer to one above it. However partially glassy basic lavas, ranging from MORBs to minettes, all with olivine on their liquidus, cover a wider range, and may have oxygen fugacities that extend to four log units above NNO. Surprisingly the range of oxygen fugacities observed in silicic lavas and ashflows with quartz phenocrysts is smaller, despite a crustal dominated evolution. The peralkaline silicic lava type pantellerite is the most reduced, equivalent to MORBs, whereas the large volume ashflows with phenocrysts of hornblende and/or sphene are the most oxidised. As the concentration of water in the basic lavas is correlated with increase in their redox state, it would seem that water could be the agent of this increase. That this is unlikely is seen in the behavior of silicic ashflows and lavas, particularly those of Yellowstone. Here the silicic magmas of the last 2Ma contain about 2 wt% FeO_(total), and typically phenocrysts of fayalite, quartz and Fe–Ti oxides. Despite extensive exchange of the ¹⁸O of the magma with meteoric water after caldera collapse (Hildreth et al. 1984), there is no displacement of the redox equilibria. Thus the thermal dissociation of molecular H₂O to H₂, and its subsequent diffusive loss to cause oxidation must have been minimal. This could only be so if the activity of water was small, as it would be if H₂O reacted with the silicate liquid to form OH groups (Stolper 1982). The conclusion is that silicic magmas with small amounts of iron and large amounts of water do not have their redox states reset, which in turn presumably reflect their generation. By analogy basic magmas with large amounts of iron and far less water are even less likely to have their redox equilibria disturbed, so that their oxygen fugacities will also reflect their source regions. The effect of pressure on the ferric-ferrous equilibrium in basic magmas can be calculated from experimental measurements of the partial molar volumes of FeO and Fe₂O₃, and their pressure derivatives V/P, in silicate liquids. The effect of pressure is found to be about the same on the liquid as it is for various solid oxygen buffers. Accordingly there should be mantle source regions covering the same range of oxygen fugacity as that found in basic lavas, but so far samples of spinel-lherzolite of equivalent oxygen fugacity to minettes or other potassic lavas have not been found. Whether or not the redox state of phlogopite-pyroxenites is equivalent to these potassic lavas cannot be established without experiment.     

The Aurora volcanic field,California-Nevada: oxygen fugacity constraints on the development of andesitic magma

 The Aurora volcanic field, located along the northeastern margin of Mono Lake in the Western Great Basin, has erupted a diverse suite of high-K and shoshonitic lava types, with 48 to 76 wt% SiO₂, over the last 3.6 million years. There is no correlation between the age and composition of the lavas. Three-quarters of the volcanic field consists of evolved (<4 wt% MgO) basaltic andesite and andesite lava cones and flows, the majority of which contain sparse, euhedral phenocrysts that are normally zoned; there is no evidence of mixed, hybrid magmas. The average eruption rate over this time period was ∼200 m³/km²/year, which is typical of continental arcs and an order of magnitude lower than that for the slow-spreading mid-Atlantic ridge. All of the Aurora lavas display a trace-element signature common to subduction-related magmas, as exemplified by Ba/Nb ratios between 52 and 151. Pre-eruptive water contents ranged from 1.5 wt% in plagioclase-rich two-pyroxene andesites to ∼6 wt% in a single hornblende lamprophyre and several biotite-hornblende andesites. Calculated oxygen fugacities fall within –0.4 and +2.4 log units of the Ni-NiO buffer. The Aurora potassic suite follows a classic, calc-alkaline trend in a plot of FeO^T/MgO vs SiO₂ and displays linear decreasing trends in FeO^T and TiO₂ with SiO₂ content, suggesting a prominent role for Fe-Ti oxides during differentiation. However, development of the calc-alkaline trend through fractional crystallization of titanomagnetite would have caused the residual liquid to become so depleted in ferric iron that its oxygen fugacity would have fallen several log units below that of the Ni-NiO buffer. Nor can fractionation of hornblende be invoked, since it has the same effect as titanomagnetite in depleting the residual liquid in ferric iron, together with a thermal stability limit that is lower than the eruption temperatures of several andesites (∼1040–1080°C; derived from two-pyroxene thermometry). Unless some progressive oxidation process occurs, fractionation of titanomagnetite or hornblende cannot explain a calc-alkaline trend in which all erupted lavas have oxygen fugacites ≥ the Ni-NiO buffer. In contrast to fractional crystallization, closed-system equilibrium crystallization will produce residual liquids with an oxygen fugacity that is similar to that of the initial melt. However, the eruption of nearly aphryic lavas argues against tapping from a magma chamber during equilibrium crystallization, a process that requires crystals to remain in contact with the liquid. A preferred model involves the accumulation of basaltic magmas at the mantle-crust interface, which solidify and are later remelted during repeated intrusion of basalt. As an end-member case, closed-system equilibrium crystallization of a basalt, followed by equilibrium partial melting of the gabbro will produce a calc-alkaline evolved liquid (namely, high SiO₂ and low FeO^T/MgO) with a relative f _{O 2} (corrected for the effect of changing temperature) that is similar to that of the initial basalt. Differentiation of the Aurora magmas by repeated partial melting of previous underplates in the lower crust rather than by crystal fractionation in large, stable magma chambers is consistent with the low eruption rate at the Aurora volcanic field. Received: 7 July 1995 / Accepted: 19 April 1996     

Diabasic intrusion and lavas,segregation veins,and magma differentiation at Kahoolawe volcano,Hawaii

A mafic sill-like intrusion, ~5?×?30 m, exposed along the eastern shoreline of Kahoolawe Island, Hawaii, represents tholeiitic magma emplaced as diabase among caldera-filling lavas. It differentiated from ~7.8 wt.% MgO to yield low-MgO (2.9 wt.%) vesicular segregation veins. We examined the intrusion for whole-rock and mineral compositions for comparison to Kahoolawe caldera-fill lavas (some also diabasic), to the Uwekahuna laccolith (Kilauea), and to gabbros, diabases, and segregations and oozes of other tholeiitic shield volcanoes (e.g., Mauna Loa and Kilauea lava lakes). We also evaluate this extreme differentiation in terms of MELTS modeling, using parameters appropriate for Hawaiian crystallization environments. Kahoolawe intrusion diabase samples have major and trace element abundances and plagioclase, pyroxene, and olivine compositions in agreement with those in gabbros and diabases of other volcanoes. However, the intrusion samples are at the low-MgO end of the large MgO range formed by the collective comparative samples, as many of those have between 8 and 20 wt.% MgO. The intrusion’s segregation vein has SiO₂ 53.4 wt.%, TiO₂ 3.2 wt.%, FeO 13.5 wt.%, Zr 350 ppm, and La 16 ppm. It plots in compositional fields formed by other Hawaiian segregations and oozes that have MgO <5 wt.%—fields that show large variances, such as factor of ~2 differences for incompatible element abundances accompanying SiO₂ from ~49 to 59 wt.%. Our MELTS modeling assesses the Kahoolawe intrusion as differentiating from ~8 wt.% MgO parent magma beginning along oxygen buffers equivalent to FMQ and FMQ-2, having magmatic H₂O of 0.15 and 0.7 wt.% (plus traces of CO₂ and S), and under 100 and 500 bars pressure. Within these parameters, MELTS calculates that <3 wt.% MgO occurs at ~1,086 to 1,060 °C after ~48 to 63 % crystallization, whereby the lesser crystallization percentages and lower temperatures equate to higher magmatic H₂O, leading to high SiO₂, ~56–58 wt.%. To contrast, greater crystallization is calculated for lower H₂O, for which it achieves less SiO₂, <55 wt.%. While MELTS reliably predicts SiO₂ approaching 58 wt.% for differentiation beyond <4 wt.% MgO, and shows that Kahoolawe intrusion’s segregations and those of Kilauea and Mauna Loa are all reasonably accommodated by the modeled parameters and SiO₂ differentiation curves, MELTS fails where it predicts that Fe enrichment is more robust under FMQ than FMQ-2 buffers. That failure not withstanding, MELTS differentiation from liquidus temperatures ~1,205–1,185 °C (depending on the various parameters) gradually increases fO₂ (up to ~0.4 log units, as normalized to FMQ) until magnetite crystallizes at ~1,090–1,085 °C, which reduces absolute fO₂ ~1 to 1.5 log units. The modeled Kahoolawe intrusion, then, exemplifies how tholeiitic magma differentiation can produce extreme SiO₂ and incompatible element compositions, and how Hawaiian segregations from shallow intrusions and lava lakes can be generally modeled under compositional and physical parameters appropriate for Hawaiian tholeiitic magmatism.     

Physicochemical parameters of deep-seated mantle plumes

Thermodynamic analysis of experimental data has demonstrated that FeO activity in silicate melts identical in composition to natural magmas can be described by the regular-solution model, which takes into account interactions of all cations with Si and interaction of Ca with Al. Using this model, we propose an oxygen barometer for spinel + magma phase association. In contrast to the earlier proposed methods for estimation of oxygen chemical potential, this barometer can work in the PJ-domain close to the liquidus of magmatic process. The new oxygen barometer has been applied to magmas related to mantle plume activity, including Siberian meimechites, Hawaiian picrites, and picrites from the Emeishan large igneous province (LIP) and Greenland. We have shown that most magmas related to the activity of deep-seated mantle plumes are characterized by a higher relative chemical potential of oxygen than magmas of mid-ocean ridges. Thermodynamically calculated stability fields of rocks with different carbon-containing phases show that under PJ-conditions of the lower mantle, the ascending mantle plumes are characterized by relatively high oxygen fugacity. Formation of diamond in the lower mantle requires more oxidizing conditions as compared with the major part of this geosphere, where the presence of Fe-Ni alloy is predicted. We have put forward a hypothesis that the main reason for the oxygen fugacity increase in particular domains of the lower mantle is a shift of redox equilibria toward a decrease in the amount of Fe-Ni alloy, up to its disappearance, with temperature growth.     

Geochemical Constraints on the Role of Oceanic Lithosphere in Intra-Volcano Heterogeneity at West Maui, Hawaii

Stratigraphically well-constrained sequences of late shield-buildingstage lavas from West Maui volcano, Hawaii, show age-dependentcompositional variability distinct from that seen in shield-stagelavas from any other Hawaiian volcano. These distinctions aredefined by ²⁰⁶Pb/²⁰⁴Pb–²⁰⁸Pb/²⁰⁴Pb variation as well as⁸⁷Sr/⁸⁶Sr correlation with ²⁰⁶Pb/²⁰⁴Pb and trace element compositions.The West Maui lavas from stratigraphically higher in the sequencehave major and trace element and Sr–Pb–Hf–Ndisotopic compositions similar to Kea-type lavas sampled at theyounger Mauna Kea and Kilauea volcanoes, indicating that theKea compositional end-member of Hawaiian lavas has remainedhomogeneous over