V31C-1441 0800h
Contrasting Isotope Systematics in Canarian Rift Zones: La Palma and Tenerife
The Canary Islands are dominated by basaltic volcanism, like Hawaii, but in contrast show a greater range of differentiation and isotopic heterogeneity and differences in magma chamber and rift zone structures. Our study focuses on the NW rift zone which extends from the Pico-Teide volcanic complex on the island of Tenerife. Basaltic to phonolitic volcanism along the rift spans approximately 20,000 years. Major element data show a large range of differentiation with MgO concentrations ranging from 0.85 to 8.26%. Samples are strongly enriched in highly incompatible elements, La averages 320x chondrites, (La/Yb)$_{N}$ averages 21.1, and no Eu anomaly was observed. High-precision Pb isotope data from nearly all samples collected along the $\sim$15km length of the NW rift show a very restricted range. Most samples range in $^{206}$Pb/$^{204}$Pb from $\sim$19.732 to 19.775 with rare samples as low as 19.675, and $^{207}$Pb/$^{204}$Pb from $\sim$15.594 to 15.621, a range only slightly greater than analytical error ($\pm$180 ppm 2$\sigma$). This small range contrasts with that observed for other Canarian volcanoes over a similar time span. For example, over a period of 20,000 years Bejenado volcano on La Palma has $^{206}$Pb/$^{204}$Pb values that range from $\sim$19.415 to 19.965 and $^{207}$Pb/$^{204}$Pb values that range from $\sim$15.425 to 15.645. Our NW rift zone data are consistent with Simonsen et al. (2000) who report a somewhat wider range of $^{206}$Pb/$^{204}$Pb values (19.653 to 19.785, with two outliers at $\sim$19.32) for "younger basalts" of Tenerife. The range for the entire island is 19.323 to 20.085. The limited isotopic range observed on the NW rift and associated with the Pico-Teide volcanic complex contrasts strongly with isotopically heterogeneous volcanoes on La Palma (Bejenado and Taburiente). We have interpreted this heterogeneity to reflect multiple small, isolated magma chambers or conduits that preserve original isotope heterogeneities in the mantle. The homogeneity observed on Tenerife may indicate a subsurface magmatic configuration that is distinct from the western Canarian volcanoes. The relatively large Teide magma system may have a homogenizing effect, even on magmas erupting along the rift zone. Alternatively, a single homogeneous mantle source may have provided all the erupted magmas on Tenerife over the past 20,000 years compared to multiple heterogeneous sources in the western islands. Simonsen, S.L., Neumann, E.R., Seim, K. (2000) {\it J. Vol. and Geochem. Res.}, {\bf 103}, 299-312.
V31C-1442 0800h
Geochemistry of the Baie Charrier Basaltic Section, Courbet Peninsula, Kerguelen Archipelago
The Kerguelen Archipelago, located in the Southern Indian Ocean, is the third largest oceanic island and represents the emergent part of the Northern Kerguelen Plateau. We present new petrographic and geochemical data from the Baie Charrier basaltic section on the northern Courbet Peninsula. This dataset complements the 1000 m Mont Crozier basaltic section located 16 km to the southwest in the central part of the peninsula. Isotopic compositions of the Crozier basalts are interpreted to be representative of the enriched component of the Kerguelen mantle plume. The Baie Charrier basalts are mildly alkalic and olivine-phyric, with a range of MgO contents (3.1-16.7 wt.%) significantly larger than that observed in the Crozier section. Mafic phenocrysts are normally zoned, with olivine core compositions of Fo$_{86-70}$ and clinopyroxene core Mg\#'s of 0.88 to 0.79. Mineral-melt equilibria constraints indicate a maximum MgO content for the Baie Charrier parental magmas of 8-10 wt.%. Major and trace element concentrations, as well as Sr, Nd, Pb, and Hf isotopic compositions, identify four distinct volcanic units within the Baie Charrier section, reflecting temporal changes in volcanism derived from a heterogeneous source region. A comparison of the Baie Charrier and Crozier sections shows similar Zr/Nb ratios of 6-7.5, enrichments in incompatible and light rare earth elements, and highly radiogenic Pb compositions ($^{206}$Pb/$^{204}$Pb = 18.35-18.64). This suggests that both sections are derived from the same source region and provides further evidence that the Courbet Peninsula is a single volcanic unit. However, the abundance of olivine-phyric basalts at Baie Charrier, and their absence at Crozier, suggests that the eruptive center of this volcano may not coincide with the present geographic center of the peninsula. The Baie Charrier basalts contain a strong signature of the enriched plume component and provide additional constraints on the composition of the Kerguelen mantle plume.
V31C-1443 0800h
Crustal Zircons from the Podiform Chromitites in Luobusa Ophiolite, Tibet
For the past decade, diamonds and unusual mineral asemblages were reported in podiform chromitites of the Luobusa ophiolite, southern Tibet, China (Bai 1993, Bai 2000, Yan 2001) by heavy mineral separation. These include (1) native elements, (2) alloys, (3) carbide, (4) platinium group elements (PGE) and arsenides, (5) silicates (6) oxide, (7) carbonates, (8) minerals with unusual compositons. Despite many questions as to these minerals above still remain open, these mineral inclusions would provide us the important infomation on the formation of the podiform chromitites. In this study, over 100 zircons were discovered by heavy mineral separation of podiform chromitite in Luobusa ophiolite. The discovery of accessory zircons in chromitites allowed us to date the formation of the chromitite and history of tectonic evolutions. Here we report the U-Pb age and mineral inclusions of zircons and discuss with unusually old age zircons. 20 zircon grains in chromitites from No. 1 site were analyzed. Zircons from the chromitites in Luobusa ophiolite are usually euhedral-subhedral and some are rounded. Cathodoluminescence images of these zircons indicate that some zircons have clear oscillatory zoning, whereas other zircons show apparent homogeneous overgrowth. U-Pb dating of these zircons by LA-ICP-MS yielded two different ages. One group has relatively younger age, 107-534Ma, which plots nearly on a concordia line. Another group has older age 1460-1822Ma, which plots off the concordia line. There is insignificant difference of apparent ages within a single zircon grain. For example, a zircon has 1650 Ma in the core, whereas does 1654 Ma in the rim. We identified several mineral inclusions, quartz, feldspar, mica, apatite, within both yonger and older zircons using laser-Raman spectrometry and EPMA. No high-pressure minerals or mantle minerals were identified. This means that these unusually old zircons were formed in low-pressure crustal emvironment. Where did the zircons in chromitites come from? It has been recognized that this ophiolite was formed at 110-120 Ma based on radiolaria in cherts overlying the pillow lavas (ALLEGRE et al., 1984; ZIABREV et al., 2003). In this study, the minimum age of 107 Ma, which we obtained from zircon in chromitites, is consistent to the age of the ophiolite. But, all other ages of zircons are much older than that of ophiolite. Yang et al. (2001) also reported U-Pb zircon ages of 450-910 Ma and Re-Os iridosmine age of 400 Ma from chromitites in Luobusa ophiolite. Mineral inclusions within zircons are crustal materials, which means that these zircons were crystallized in the low pressure crustal condition. Thus these zircons within chromitites are interpreted as xenocrysts from old crustal materials. Recently, old-age zircons (330 to 1600 Ma) were also reported from the Mid Atlantic Ridge MORBs (PILOT et al., 1998). They suggest one possibility that these old-age zircons may have derived from old continental crustal material, which have assimilated with the MORB magma during ascent. Moreover, Archean zircons were reported from pyroxenite dikes in Jormua ophiolite (PETRI et al., 2003). YU et al. (2001) reported that zircons from chromitites in Luobusa ophiolite have shorter inter-atomic distances of Zr-O and Si-O bonds. As a result, they concluded that Tibetan-zircons were derived from the high-pressure mantle environment. Judging from the line of evidence mentioned above, it is highly possible that these zircons captured by chromitites were originated from recycled crustal materials, convecting through upper mantle.
V31C-1444 0800h
Immiscible Transition from Carbonate-rich to Silicate-rich Melts in Eclogite+CO$_{2}$ and Genesis of Ocean Island Melilitite
Derivation of highly silica-undersaturated lavas such as olivine melilitites and melilite nephelinites from the mantle has been attributed to the effects of CO$_{2}$. However, experimental studies have so far failed to demonstrate equilibrium of melilititic melts with a four-phase peridotite assemblage. Instead, the liquidus mineralogy of these silica-undersaturated magmas at high-pressures appears to be dominated by cpx$^{1}$. Although, experimental partial melts from natural peridotite+CO$_{2}$ span a continuum from carbonatite to alkali-basalts$^{2}$, ocean-island melilitites have distinctly higher TiO$_{2}$, FeO*, and CaO/(CaO+MgO)$^{3,4}$ than compositions derived thus far from a carbonated lherzolite source. Partial melting experiments of a nominally anhydrous, natural eclogite with a small amount of added carbonate (SLEC1; 5 wt.% bulk CO$_{2}$) were performed to investigate the transition between carbonate and silicate melts with increasing temperature. Experiments were conducted in a piston cylinder at 3 GPa from 1050 to 1400 $\deg$C. Garnet and cpx appear in all the experiments and ilmenite is observed from 1075 to $\sim$1200 $\deg$C. An Fe-bearing calcio-dolomitic melt is present from the solidus (1050-1075 $\deg$C) up to 1375 $\deg$C. Beginning at 1275 $\deg$C, it coexists with a silica-poor silicate melt. Textural criteria indicate only a single CO$_{2}$-rich silicate melt phase at 1400 $\deg$C, coexisting with garnet and minor cpx. The liquidus temperature is estimated to be $\sim$1415 $\deg$C from the melt fraction-temperature trend. With increasing temperature, the carbonate melt becomes richer in SiO$_{2}$ ($\sim$2 to 5 wt.%) and Al$_{2}$O$_{3}$ ($\sim$0.75 to 2.25 wt.%) and poorer in CaO ($\sim$30 to 25 wt.% from $\sim$1200 to 1375 $\deg$C). Compositions of silicate partial melts change systematically with increasing temperature, increasing in SiO$_{2}$ ($\sim$36 to 41 wt.%), Al$_{2}$O$_{3}$ ($\sim$4.5 to 9.5 wt.%), MgO ($\sim$9.5 to 13 wt.%), CaO ($\sim$8 to 14 wt.%) and decreasing in TiO$_{2}$ ($\sim$14 to 2.5 wt.%), FeO ($\sim$20 to 13 wt.%), Na$_{2}$O ($\sim$3.3 to 1.7 wt.%). A wide temperature interval of coexisting carbonate and silicate partial melts of carbonated eclogite is distinct from the continuous transition from carbonate to silicate melts observed in carbonated peridotite systems$^{2,5}$. At high-temperature, the silicate melts generated from SLEC1 are comparable to strongly silica-undersaturated, alkalic OIB lavas and closely resembles ocean island melilitite and nepheline melilitite$^{3,4}$ in its SiO$_{2}$, FeO*, MgO, CaO, TiO$_{2}$, and Na$_{2}$O content. They are also similar to melilite bearing lavas of continental affinity, though the match is not as close. Although the SLEC1 derived immiscible silicate melts are lower in Al$_{2}$O$_{3}$ than primitive alkalic OIB lavas, liquids richer in Al$_{2}$O$_{3}$ may be produced at slightly lower pressures. Geochemical and geodynamical investigations of carbonated eclogite sources for melilitic volcanic series thus merit consideration. 1. Brey, G and Green, D. H. 1977, CMP 61, 141-162. 2. Hirose, K. 1997, GRL 24, 2837-2840. 3. Clague, D. A. and Frey, F. A. 1982, JP 23, 447-504. 4. Hoernle, K. and Schmincke, H.-U. 1993, JP 34, 573-597. 5. Moore, K. R. and Wood, B. J. 1998, JP 39, 1943-1951.
V31C-1445 0800h
Partial Melting and Contact Metamorphism of Basic Dykes in the Aureole of an Ocean Island Pyroxenitic Intrusion; Fuerteventura, Canaries
The hypabyssal root zone of an ocean island volcano can be directly observed in the uplifted Basal Complex of Fuerteventura (Canary Islands). This complex records a long-lasting magmatic activity characterised by the intrusion of numerous magma batches as small plutons, dykes, dyke swarms and ring-dyke complexes of alkali-gabbros, pyroxenites, syenites and carbonatites. The high heat flow generated by the successive magma pulses feeding the overlying subaerial volcanic centres induced remarkable metasomatic and metamorphic processes, culminating with the low-pressure partial melting of basic dykes within the heterogeneous Tierra Mala gabbro-pyroxenite-syenite formation, in a contact metamorphic aureole around a layered pyroxenite-gabbro intrusion (PX1). The migmatites are characterised by a dense, zebrated network of closely spaced millimetre-wide leucocratic veins with perfectly preserved igneous textures. Only limited partial melting over a small distance (15-20m) appears to have been generated around an apophysis of PX1 (250m wide), due to a presumably limited heat input. It is thus possible to follow the evolution of this phenomenon within particular dykes with increasing distance from the intrusion. Results show that lithologies are mineralogically and chemically highly variable and behave differently in the contact aureole. Dykes of mafic, foid-microgabbros with igneous intersertal textures show increasing alteration of diopside to kaersutite with increasing proximity to the PX1 apophysis, up to 16m from the contact. Subsequently diopside crystals begin to appear smaller and recrystallised, and within the first 4m the dykes almost completely lose their small leucocratic component, leaving a framework texture dominated by sub- to euhedral diopside, Fe-Ti oxides, poikiloblastic Ti-rich biotites and minor kaersutite. This is reflected by the whole-rock SiO2, alkali, and LILE concentrations, which decrease as you approach the intrusion. Trace element contents are comparable to those of standard OIBs and there seems to be no significant change in REE composition with distance from PX1. However, dykes of diopside phyric microfoidolite that appear to have partially melted over a larger distance (up to 16m), contain trace element concentrations higher than that of common OIBs, and show a significant depletion in Si, Al, K and most incompatible trace elements, including LREE, in the anatectic samples close to the contact. As with the previous lithology, much leucocratic material was lost within the first metre of the contact, but relic diopside phenocrysts reside in abundance, and within a distance of 2m from the intrusion there is a rapid increase in kaersutite and feldspathic material (andesine and k-feldspar). Segregation of leucosomes into discrete veinlets or small pods is evident in this rock type and dykes of amphibole phyric foid-microgabbros. Previous theories of an enrichment in K, Sr, Y and REE for all migmatites relative to their protolith thus appear to have been disproved. Also, the idea that all rocks close to the PX1 intrusion are enriched in REE relative to OIBs has also been disproved by the foid- microgabbro series, which, on the contrary, are depleted in some trace-elements. Hence, we can conclude that the enrichment of some of these dykes must be due to a process unrelated, and prior to, the intrusion of PX1.
V31C-1446 0800h
Melt Inclusions in Primitive Olivines from Padloping Island, Baffin Bay
Tertiary picrites from Padloping Island, Baffin Bay, are thought to be the result of early activity of the Iceland Plume. The olivines found in these lavas are primitive to very primitive with Fo-content ranging from 88 to 93. However, they are not in equilibrium with the liquid, suggesting that they have been derived from earlier cumulates. Therefore, by obtaining the composition of the melt inclusions in these olivines we might be able to gain some insight in the composition of, possibly primary, mantle derived melts. We have optically evaluated a large set of olivines and it is clear that they all contain primary inclusions ranging in size from $\sim$10 to $\sim$100 micron in diameter. The inclusions are randomly distributed and not related to any fractures. They contain glass, included spinel minerals, daughter minerals and volatiles, suggesting relatively slow cooling after entrapment. Electron microprobe analysis of daughter minerals has suggested a primitive composition for the initial trapped liquid, which has evolved into a residual Si-rich liquid after the growth of Al-rich clinopyroxene. In order to analyse the original trapped melt homogenisation experiments have been carried out under ambient pressure using a heating stage under controlled fO$_{2}$ conditions (purified He atmosphere). These experiments suggest that our liquidus temperature could be as high as $1350\deg$C. However, the fact that we are unable to complete our homogenisation, i.e. let the gas phase dissappear, suggests that the melt was either oversatured in CO$_{2}$-rich volatiles or that the entrapment has taken place at higher pressures than can be obtained by our experiments. Major element compositions for the homogenised inclusions have been obtained by electron microprobe. Our first results suggest that the early melts, in equilibrium with Fo 93, are low in total alkalis and range from 51 to 56 wt% SiO$_{2}$, i.e. a composition ranging from basalt to basaltic andesite. There are two mantle rocks that could generate these type of melts of which some are quartz normative. Very shallow level melting of a peridotite and melting at higher pressures of an eclogite. Further work should allow us to decide which mantle rock has contributed and to what extent.
V31C-1447 0800h
Evaluating the role of crustal interaction with mantle-derived magmas: Sr-Nd-O isotope geochemistry of a Quaternary, caldera-forming, phonolitic eruptive sequence; the Diego Hernandez Formation, Tenerife, Canary Islands (Spain)
The Sr-Nd-Pb isotope composition of Miocene-Pliocene basalts from the Canary Islands has been well documented (e.g., Simonsen et al., 2000), and the data define a restricted range of compositions resulting from a mixture of two or more mantle sources. Trace-element and O-isotope variations among younger basalts and more highly differentiated eruptive products from these islands require the involvement of either hydrothermally altered MORB crust or the volcanic edifice during their ascent and petrogenesis (e.g., Wolff et al., 2000). Here we present new Sr-Nd-O data from all of the major pyroclastic deposits of the Diego Hernandez Formation (DHF), a Quaternary, caldera-forming sequence that includes the entire spectrum of magma types found on the island. Sr ($^{87}$Sr/$^{86}$Sr$_{i}$ = 0.70302 - 0.70311) and Nd ($^{143}$Nd/$^{144}$Nd = 0.51288 - 0.51290) ratios of DHF pumices fall within the range of Canary Islands basalts, over-lapping the low $^{87}$Sr/$^{86}$Sr end of the array. These data are also consistent with the Tenerife basalt isotope data of Simonson et al. (2000). Sr and Nd isotope variations among the phonolitic pyroclastic deposits have a more restricted range than those of the less differentiated and older basalts on the island. Initial isotopic data from syenite xenoliths, representative of at least two plutons within the volcanic edifice, suggests that they have Nd isotope compositions overlapping the highest $^{143}$Nd/$^{144}$Nd ratios of DHF phonolites but have a more radiogenic Sr isotopic composition. Mean \delta$^{18}O$ values produced by laser fluorination of feldspar phenocrysts from DHF pumices range from 5.63 to 6.33 \permil VSMOW. The lower \delta$^{18}O$_{FSP}$ values are within the range of MORB oxygen isotope values, but the higher \delta$^{18}O$_{FSP}$ values are lower than expected for some of the more differentiated units. \delta$^{18}O$_{FSP}$ values also display a negative correlation with $^{87}$Sr/$^{86}$Sr ratios. [1]Simonsen et al., 2000; JVGR, vol. 103, p. 299;[2]Wolff et al., 2000; JVGR, vol. 103, p. 343;[3]Hickes et al., 2003, Eos Trans. AGU, 84(46), V12A-0565
V31C-1448 0800h
MORB Mantle Hosts the Missing Eu in the Continental Crust
The continental crust (CC) is an important geochemical reservoir. The origin and evolution of CC remains debated [1-13]. Estimation of bulk CC composition is model dependent [1-8,12]. A reliable upper CC composition has been obtained from shales and other fine-grained sedimentary rocks [1,2,15,16] due to the mixing/homogenization effects of sedimentary processes [14-16]. The upper CC has a negative Eu anomaly (NEuA or Eu/Eu* $<$ 1), 0.65, which points to a missing Eu storage. The upper CC is $<$10km thick and cannot represent the bulk CC [1-3], which has a mean thickness of $\sim$36km. To determine the composition of deep CC is required, but the task is hard [1-3]. The approach of combining geochemical data on deep crustal rocks (xenoliths and locally exposed amphibolites and granulites) with seismic and heat flow data to estimate the average deep CC composition is creative [1-4,12]. As some granulites show Eu/Eu* $>$ 1, the missing Eu in the upper CC may be in the deep CC. Indeed, the NEuA in the upper CC is largely reduced in model bulk CC compositions: 0.829 [4], 0.962 [1], 0.974 [3]. By assuming mantle derived igneous rocks display no Eu anomaly [1,12], the upper CC NEuA would come from intracrustal differentiation. For example, the granitic upper CC resulted from anatexis of mafic rocks in the deep CC with Plag as a residual phase (holding Eu) in the granulite residues [1-4,12]. This interpretation is sensible, but the protoliths of most of these granulites are found to be underplated mantle melts [1,17-19]. The NEuA ($\sim$0.8) and the more felsic nature of bulk CC composition in China encouraged the authors [12] to suggest that the more mafic lower CC rocks were tectonically removed. We show that high quality data on 306 fresh MORB glass samples (2-10 wt% MgO) from the Pacific and Atlantic [20-25] exhibit varying Eu/Eu* (0.68-1.18) and significant correlations: R[MgO-Eu/Eu*] = 0.876, R[MgO-Sr/Sr*] = 0.809 and R[Eu/Eu*-Sr/Sr*] = 0.875. 148 samples show Eu/Eu* $>$ 1. For Eu/Eu* $>$ 1, MgO $>$ 7.6 wt% or T[liquidus] $>$1185$\pm$10$\deg$C. For MORB, Plag begins to crystallize at $\sim$1200$\pm$10$\deg$C. These observations demonstrate that primitive MORB melts all have POSITIVE Eu and Sr anomalies. We interpret these anomalies as inherited from MORB source. Eu and Sr behave similarly because of the same charge [2+] and ionic radius (~1.31$\AA$ for CN=6). The larger radius may make Eu[2+] more incompatible than the smaller Sm and Gd, but the divalent Eu[2+] goes mostly with Sr into M2 site of Cpx, making it more compatible than the trivalent Sm and Gd. If the depleted MORB mantle indeed resulted from bulk CC extraction in the early Earth [9], then that event would have preferentially extracted the more incompatible trivalent REEs into CC, leaving the divalent Eu[2+] in the MORB mantle. Hence, we suggest MORB mantle host the missing Eu in CC. References: [1] Taylor, McLennan, RG, 1995; [2] Rudnick, Fountain, RG, 1995; [3] Rudnick, Nature, 1995; [4] Wedepohl, GCA, 1995; [5] Weaver, Tarney, Nature, 1984; [6] Bowring, Science, 1995; [7] McCulloch, Bennett, GCA, 1994; [8] Abbott et al., EPSL, 1997; [9] Hofmann, EPSL, 1988; [10] Condie, 282pp., 1997; [11] Albarede, Tectonophys, 1998; [12] Gao et al., EPSL, 1998; [13] Niu et al., JPet, 2003; [14] Goldschmidt, 148pp, 1938; [15] Taylor, McLennan, 312 pp, 1985; [16] Goldstein, Jacobsen, EPSL, 1988; [17] Rudnick, Taylor, JGR, 1987; [18] Rudnick, GCA, 1992; [19] Rudnick, Nature, 1990; [20] Niu, Hekinian, EPSL, 1997; [21] Niu, Batiza, EPSL, 1997; [22] Castillo et al., EPSL, 1998; [23] Niu et al., JGR, 1999; [24] Regelous et al., EPSL, 1999; [25] Niu et al., EPSL, 2001.
V31C-1449 0800h
Mineral Compositions from the Hawaii Scientific Drilling Project (HSDP): Preliminary Results Part III - Olivine
Olivines are the most prevalent phenocrysts in Hawaiian lavas and thus offer a more complete temporal coverage of magmatic processes compared to phenocrysts of more restricted occurrence. As in parts I and II (see accompanying abstracts) we examine olivine compositions from HSDP core samples, which provide a temporal view of the Mauna Kea magma plumbing system. In general, mineral compositions are valuable because they provide an archive of liquid compositions; the compositional record of liquids is always murky due to the facility with which liquids are blended. But minerals, such as olivine, are resistant to re-equilibration and hence provide a much more complete record of the range of liquid compositions that enter a magma plumbing system. Olivine compositions are of particular interest since the Fo content (Mg/(Mg + Fe) ratio) is highly sensitive to T. Olivines that are sampled through the HSDP core thus provide a window into temperature variations across the Hawaiian hot spot. Most olivines are not in Fe-Mg exchange equilibrium with their host whole rock samples. This results is unsurprising and has been noted in numerous studies of olivine phenocrysts at Hawaii. This lack of equilibration limits the precision with which crystallization temperatures can be extracted. But comparisons of Mg/(Mg + Fe) (Fo content) with core depth reveal interesting qualitative results. Most notable are two aspects: (a) Fo contents exhibit nearly constant maximum values throughout the 3000 m of core, near Fo = 90. (b) In contrast, minimum Fo values appear to decrease with decreasing depth. The first observation is counter to what was expected. As Mauna Kea moved over the hot spot, one might expect maximum temperatures to decrease in response to the tapping of cooler mantle material. But near constant maximum values suggest that the hottest parts of the plume are delivered to Hawaiian conduits throughout their eruptive history. In contrast, the decrease in minimum Fo values with time might indeed reflect some T variation across the plume, and is consistent with observations of plagioclase (see part II). We hypothesize that the progressive decrease of Fo contents reflects a gradual decline in eruptive recurrence intervals: as magma supply rates wane, and eruption recurrence intervals increase, the magma conduit may progressively become choked off at various depths, with the result that individual magma batches are stored at a greater range of depths for longer periods of time. But even as supply rates decline, the hottest magmas that are supplied to a conduit during the shield-building stage seem somehow to still be available to the conduit during waning periods of volcanic activity.
V31C-1450 0800h
Mineral Compositions from the Hawaii Scientific Drilling Project (HSDP): Preliminary Results Part I - Clinopyroxene
Drill core samples recovered from the HSDP have sampled Mauna Kea volcano, Hawaii, to depths of over 3000 meters. These samples provide a temporal view of the evolution of the Mauna Kea magma plumbing system not easily discerned from surface sampling. We utilize mineral compositions from the HSDP as a monitor of the depths at which magmas stagnated and partially crystallized. Mineral compositions provide a record of magma stagnation, and due to their resistance to re-equilibration and mixing, provide an archive of the range of liquid compositions that entered the magma plumbing system. Most interesting is the question as to whether the Mauna Kea magma plumbing system changed during the construction of the volcanic edifice, or during the waxing and waning of magma supply as Mauna Kea passed over the Hawaiian mantle hot spot. Here we use clinopyroxenes to examine magma transport, and in two accompanying abstracts we compare these results to those obtained from the analysis of plagioclase and olivine compositions. We utilized the models of Putirka (1999) to test clinopyroxene-melt equilibria and the models of Putirka et al (2003) to calculate the P-T conditions of crystallization of clinopyroxene, where appropriate. Equilibrium tests suggest that clinopyroxene phenocrysts are out of equilibrium with whole rock compositions. As an attempt to recover an equilibrium liquid, we adjusted whole rock compositions by adding or subtracting olivine so as to achieve Fe-Mg exchange equilibrium between (calculated) residual liquid and average olivine phenocryst compositions. This strategy required the subtraction of large amounts of olivine, 20% on average. Approximately 68% of cpx phenocrysts were calculated to approach equilibrium with these corrected liquid compositions; only these phenocrysts are used for the calculation of P and T. The distribution of cpx phenocryst depth estimates is not unlike those found for Mauna Kea in earlier studies (Putirka, 1997; Yang et al., 1999). There is no evidence for ponding at any particular depth, such as the Moho; instead, partial crystallization appears to occur throughout a conduit that extends to depths well into the mantle (see Ryan, 1988). However, these estimates lose precision as our corrections for olivine accumulation cause our `liquid compositions' to migrate from an observable quantity; each individual depth estimate is imprecisely located. While the P-T record of HSDP clinopyroxenes is imprecise, clinopyroxenes record no striking temporal trends when composition is compared to core depth. Taken at face value, this lack of temporal variation may indicate that clinopyroxenes crystallized shortly after magma mixing (after olivine crystallization) and prior to plagioclase saturation.
V31C-1451 0800h
Mineral Compositions from the Hawaii Scientific Drilling Project (HSDP): Preliminary Results Part II - Plagioclase
As noted in part I, the drill core samples recovered from the HSDP provide a temporal view of Mauna Kea volcano, Hawaii. Presently, drill core samples extend to depths of 3000 meters, which represents an eruptive age of approximately 550 ka. This depth interval probably represents about 1/2 the life span of Mauna Kea, since the base of the crust is thought to lie some 6 km beneath the drill hole site. Our samples thus portray the latter half of Mauna Kea evolution. In part I we examined clinopyroxene grains; here we present an analysis of plagioclase compositions from the HSDP, and utilize new models that allow the calculation of crystallization depths and temperatures for plagioclase + liquid pairs (Putirka, 2004, Am Min, in press). As with clinopyroxenes (see part I), we apply a saturation surface model as a filter to test whether plagioclase grains are consistent with equilibrium with their whole rock compositions. Approximately 57 % of plagioclase grains pass this filter. Unlike clinopyroxene, plagioclase compositions change with time. Minimum depth estimates at any particular core depth (i.e., age) all indicate shallow level storage. But maximum crystallization depths increase with decreasing age of the edifice (decreasing depth in the HSDP core). Plots of An and Ab contents vs. core depth reveal similar systematics in mineral compositions with core depth, hence these temporal variations do not reflect liquid composition variations alone. The base of the core, at 3 km, should sample Mauna Kea near the height of its shield-building phase; in these flows all depth estimates are less than 10 km. Interestingly, maximum crystallization depths increase with time, and the greatest depth estimates and the largest depth range are observed for post-shield samples. We propose that plagioclase crystallization reflects a change in magma supply rate. It might be expected that the supply rate of magma at Mauna Kea should be near a maximum for samples recovered at 3000 m, and that supply rates would diminish, and eruption recurrence intervals increase, as one moves upward in the core. Given this expected change, plagioclase compositions may record a late-stage clogging of the magma conduit as recurrence intervals decline. During periods of high magma supply rates the conduit is open; temperatures are sufficiently high so that the plumbing system is rarely clogged, and plagioclase crystallization, which may occur only following sustained periods of storage, occurs only at the shallowest depths. As magma supply rates decline, recurrence intervals become too long to keep the conduit fully open. Magma is thus trapped at progressively greater depths, and stored long enough to achieve plagioclase saturation at depth. The progressive change in plagioclase depths thus reflects a dwindling magma supply rate as Mauna Kea moves off of the Hawaiian hot spot.
V31C-1452 0800h
Ocean Floor Metamorphism Recorded in the Taitao Ophiolite
Ocean floor metamorphism is important process of interaction between sea-water and oceanic crust, namely surface environment and solid earth. We reconstructed a complete section of oceanic ridge emplaced near the Chile Triple Junction based on the geological investigation of the Miocene-Pliocene Taitao ophiolite. The Taitao ophilolite consists of harzburgite, massive and layered gabbro, sheeted dike complex and sequence of pillow lavas and sedimentary rocks in ascending order from the south. These sequences generally strike east-west and dip to the north. Our results revealed that a major shear zone is present between harzburgite and 4 Ma arc-derived granitoid in the south. The thickness of harzburgite is about 1 km. The harzburgite and overlying gabbros are folded into complex pattern together, and the boundary between them is mainly tectonic. The gabbro is about 3 km thick. Sheeted-dike complex has two types of intrusions with different directions; NNE-SSW trending in the northern block, and NW-SE trending in the southern blocks. Piles of pillowed basalts and terrigenous sediments with westward younging were inferred to be associated with sheeted-dike complex of the northern block. Observation of thin sections and analysis by electronprobe microanalyser revealed that metamorphic grade ranges from zeolite facies, throuth the greenschist facies, to amphibolite facies, and increases progressively downwards. The zone boundaris are subparallel to the lithostratigraphy. Metamorphic facies series of the Taitao ophiolite corresponds to the low-pressure type. According to our results, ocean floor metamorphism is well-preserved in the Taitao ophiolite.
V31C-1453 0800h
Carbon flux and C/Nb ratios in the mantle in ridge context
The hypothesis that undegassed MORB would display constant CO$_{2}$/Nb-ratios of \sim 239 (ppm/ppm) obtained by Saal et al (2002) is not supported by available data. Incorrectely quoted by these authors, the MORB popping rock shows a CO$_{2}$/Nb-ratio higher by a factor of \sim 3 (Javoy and Pineau, 1991) than the ratio measured among their carbon under-saturated glass inclusions from the Siquieros transform fault. In order to better constrain the variability of CO$_{2}$/Nb-ratios among MORB, we have analysed and compiled volatile, trace and major element data from 14 \degN and 34 \degN basaltic glasses on the Mid Atlantic Ridge. The sample set includes N-, E-MORB and alkali basalts. Carbon (and water for alkali basalts) have been affected by degassing and initial carbon and water concentrations were reconstructed using methods described previously (e.g. Pineau et al. 2004). The variations in \delta D-values (-88 to -55 per mil) with water contents (1300 to 7700 ppm) show consistent variations with published data of worldwide mid-ocean ridges. The samples define rather constant regional ratios from 349 to 589 and from 524 to 1065 at 14 \deg and 34 \degN respectively. CO$_{2}$/Nb-ratios roughly correlate positively with enrichment indexes such K/Ti. It is worth noting that the little variable CO$_{2}$/Nb-ratios over a large range of CO$_{2}$, up to 44,000 ppm, (i.e. 12,000 ppmC) and Nb-contents, up to 41.3 ppm, confirm the accuracy of degassing corrections. From the broad inverse correlation between CO$_{2}$/Nb and TiO$_{2}$/H$_{2}$0 and the mean TiO$_{2}$/H$_{2}$0-ratio of worldwide MORB with 8 \pm1weight percent MgO, a MORB CO$_{2}$/Nb-ratio from 500 to 950 seems more appropriate than the value quoted by Saal et al. (2002). This corresponds to carbon flux from 4.0 to 7.6\times10$^{12}$ mol/y. However, in the light of the regional dependency of CO$_{2}$/Nb-ratios it is unclear yet whether it represents the best tool to constrain mantle carbon flux or should be more likely used to distinguish different mantle domains.
V31C-1454 0800h
Re-Os Systematics of Variably Altered Ultramafic Rocks, North Carolina.
Dunitic ultramafic bodies, presumed to be pieces of dismembered Ordovician ophiolites, crop out along strike within the Blue Ridge Province of the southern Appalachians. These rocks were metamorphosed under highly different conditions. Pristine olivine dominating some bodies may represent either primary mineralogy, or recrystallization under dry conditions. Highly serpentinized rocks, more common in other ophiolites, reflect either post-formation hydrous alteration of peridotite in the upper mantle or lower crust, or effects resulting from emplacement. All of the bodies contain chromian spinel. The presence of spinel with very low Re/Os and high Os concentration allows comparison of putative initial $^{187}$Os/$^{188}$Os for each body with initial isotopic compositions calculated for bulk rocks and olivine separates for the variably altered bodies. These results ultimately should allow a better understanding of alteration effects on the Os isotopic systematics of other ophiolite assemblages. The concentration of Os in spinel separates range from 1.7 to 340 ppb whereas bulk rock varies from 0.9 to 8.4 ppb. Whole rock and spinel separates have present day initial $^{187}$Os/$^{188}$Os that range from 0.1141 to 0.1327. Spinels have $^{187}$Re/$^{188}$Os of 0.0001 to 0.0832 and bulk rocks range from 0.0398 to 1.9712. Initial ratios calculated for 490 Ma range from 0.1027 to 0.1327. Data for some spinel-whole rock pairs indicate significantly different $^{187}$Os/$^{188}$Os at the time of formation on a very local scale. There is also evidence for late stage additions of Re to some whole rocks. Present day $^{187}$Os/$^{188}$Os as low as 0.1141 in spinel indicates melt depletion in precursor materials a minimum of 1.9 Ga ago.
V31C-1455 0800h
Pyroclastic Eruptions on the Northern East Pacific Rise
Glass sand-sized fragments were recovered from three sites along the northern East Pacific Rise using the ROV Tiburon. Abundant moderately-vesicular bubble-wall (like limu o Pele) fragments, and rare Pele's hair were recovered near the center of the Alarcon Rise at 2360-2400 m at 23\deg23'N to 23\deg26'N and just south of the Tamayo Fracture Zone at 3150-3190 m at 22\deg50'N. Less abundant, but morphologically similar, fragments were recovered at 2620 m at 20\deg50'N, near the 21\deg N hydrothermal vent sites. The samples were recovered in push cores and using a small 29-jar suction sampler designed to collect and filter glass particles. The particles in individual samples include numerous compositional types, suggesting widespread dispersal of the particles after their formation. Many of the bubble-wall fragments enclose abundant (10-15%) spherical to elongate, stretched vesicles. The particles provide stratigraphic information on the sequence of eruptions in an area and can therefore be used to determine eruption compositions in space and time. For example, at the site at 22\deg50'N, particles with a distinct composition were recovered from the flank of a small cone, establishing that the cone predates an eruption just to the south of the cone. Also at this site, glass particles collected on top of the cone have a range of compositions that, on average, are about 0.3% lower in MgO than pillow rind glasses from the same cone. This compositional difference indicates that the pillow basalt glasses average about 6\deg C cooler than the particulate glasses. Maicher and White (2002) proposed that bubble-wall fragments form from steam expansion where lava flows cover saturated sediment forming hydroclastic (secondary) eruptions. On the other hand, Clague et al. (2003) proposed that bubble-wall fragments from the Gorda Ridge were formed by discharge of coalesced magmatic gas bubbles through the erupting vent in pyroclastic eruptions. Four lines of evidence support the pyroclastic origin of the glass fragments on the East Pacific Rise: 1) bubble-wall fragments occur with Pele's hair fragments, 2) site at 22\deg50'N, like the sites on the Gorda Ridge, are at pressures greater than the critical pressure for sea water boiling so no vapor phase is produced upon heating, and 3) bubble-wall fragments occur on top of a monogenetic cone that should have been sediment-free when the glass particles formed, and 4) the abundant vesicles trapped in some bubble-wall fragments are evidence of high concentrations of magmatic gas in these lavas. Pyroclastic eruptions, most likely similar to strombolian eruptions, are ubiquitous along the slow-spreading Gorda Ridge and the fast-spreading northern East Pacific Rise demonstrating similar eruptive processes despite the different spreading rates and ridge morphologies.
V31C-1456 0800h
Li, Be, and B Concentrations in Mauna Loa and Mauna Kea Lavas from the HSDP-2 Drill Core
Continuous core samples collected during Phase 2 of the Hawaii Scientific Drilling Project (HSDP-2) provide the most complete time sequence stratigraphy for a hotspot volcano. 43 samples, representing 245 m of Mauna Loa (ML) tholeiitic lavas and 2853 m of Mauna Kea post shield (MKPS) and main shield (MKMS) alkaline and tholeiitic lavas, were analyzed for whole-rock Li, Be, and B abundances to determine how these elements vary as the volcano samples the Hawaiian plume. Anomalous Sr/MgO variations indicate that MKPS lavas have experienced alteration. Highly scattered B abundances (1.1-9.4 ppm), B/K ratios (0.0002-0.008), and B/Be ratios (1-25) suggest substantial remobilization of this H$_{2}$O-soluble element. Regular correlations between Be and MgO, Zr and Nd indicate that Be behaves incompatibly in HSDP lavas. Be ranges from 0.28-0.54 ppm in ML, 0.52-0.97 ppm in MKPS, and 0.23-0.71 ppm in MKMS lavas. Be/Nd = 0.03$\pm$0.005, slightly lower than MORBs or chondrites. Li ranges from 2.5-6.1 ppm and does not vary among lava types. Li/V ($\sim$0.02) and Li/Y ($\sim$0.18) ratios are similar to MORBs, but Li/Yb (2.3) ratios are higher than MORBs. Despite evidence for weathering in MKPS lavas, Li correlates well with Be, Nb, Zr, Y, Yb, and V, suggesting Li was not mobilized with alteration. However, such correlations do not occur in ML or MKMS lavas, indicating Li modification. The lack of change in Be/Nd, Li/V and Li/Yb shows abundances of these elements are governed largely by extent of melting and/or crystallization. Mantle compositions for these species do not seem to vary on the timescale of shield development. B and (to a lesser extent) Li are useful indicators for alteration and/or hydrothermal activity within shield edifices. Lentz et al. (2001) used variations in B and Li content in pyroxenes of some Martian meteorites to suggest these elements were lost (along with H$_{2}$O and CO$_{2}$) as the magma degassed during ascent. Since mantle plumes were the primary means for Mars volcanism, the Hawaiian plume may serve as a terrestrial analog for understanding geochemical variations in Martian meteorites.