V33A-1152
Inference of mantle viscosity from gravity data: a genetic algorithm inversion method
Relative radial variations in the viscosity of the mantle can in principle be determined from surface gravity measurements: an analytical theory of mantle flow provides geoid kernels relating density maps and viscosity profiles to the Earth's gravity field. A scaled global tomographic map of seismic wave speeds can be used as an estimate of the Earth's density distribution. A linear inverse problem can then be set up, with gravity observations as data, and the viscosity profile as the unknown. This method has the limit of constraining only the ratios between viscosity values at the different depths, rather than the viscosity values themselves. Additionally, the solution to this inverse problem is strongly non-unique. Last, seismic velocities in the mantle are known only approximately, and establishing an appropriate velocity-to-density scaling for the mantle is, likewise, not trivial. We attempt to account for non-uniqueness in the inverse problem by exploring the solution space, formed of all possible radial profiles of Earth viscosity, by means of a non-deterministic global optimization method: the genetic algorithm. For each sampled point of the solution space, a forward calculation is conducted to determine a map of gravity anomalies, and its similarity to GRACE is then measured; the procedure is iterated to convergence, according to genetic algorithm criteria.
V33A-1153
Significance of Picritic and Tholeiitic Lavas Within Wrangellia Flood Basalts on Vancouver Island for the Melting History and Magmatic Evolution of a Major Oceanic Plateau
Geochemical studies of lavas from the accreted Wrangellia oceanic plateau forming a large part of Vancouver Island (~20,000 km2) in the Pacific Northwest of North America offer a view of the melting history of plume-derived magmas that does not involve continental lithosphere and where source heterogeneity does not play a major role. The Late Triassic Wrangellia flood basalts (~229-226 Ma) are predominantly homogeneous tholeiitic basalt. However, the lower submarine part of the 6 km-thick stratigraphy, on northern Vancouver Island, contains picritic pillow basalts. These high-MgO (9-20 wt%) lavas are depleted in LREE (La/YbCN= 0.5 ± 0.2), whereas the tholeiitic lavas (6-8 wt% MgO) are LREE-enriched (La/YbCN= 2.2 ± 0.3). Both lava groups have overlapping initial εHf (+10.3 ± 2.1) and εHf Nd (+7.7 ± 1.3), indicating a common, depleted (but not MORB) Pacific mantle source similar to the source of basalts from the Ontong Java and Caribbean Plateaus. The presence of picritic lavas indicates melting of anomalously hot mantle and, along with the high degree of melting and high melt production rate, provides strong evidence for a mantle plume origin for this oceanic plateau. Estimated primary melts, using PRIMELT1 [Herzberg et al. (2007) G3], contain 15-17 wt% MgO and ~10 wt% CaO and they represent 23-27% melting of peridotite with a potential temperature of ~1490°C. Differences in trace element signatures between the high-MgO and tholeiitic lavas primarily reflect differences in depth of melting. The high- MgO lavas formed without involvement of garnet, whereas the tholeiitic basalts involved melting of both garnet and spinel lherzolite. The evolved tholeiitic basalts underwent significant fractional crystallization (>50%) and the fractionated residues are likely represented by high-velocity rocks beneath Vancouver Island identified from seismic reflection studies [Clowes et al. (1995) CJES].
V33A-1154
Pb, Hf, Nd, and Sr Isotopic Variations of Hualalai Shield Stage Tholeiites from the Submarine North Kona Region
We present the isotope and trace element compositions of tholeiitic lavas collected from deep submarine portions of North Kona region, the west flank of Hualalai volcano. The samples were collected from the lower section of the North Kona bench (dives K218 and K219), a submarine section at Hualalai volcano's northwest rift zone (dive S690), and an elongate ridge below the central section of the bench (dive S692) during 2001 and 2002 JAMSTEC Hawaii cruises. Hualalai volcano is presently in the post-shield alkalic stage and most of its subaerial surface is covered by alkalic basalt. All analyzed samples of the pillow lavas are tholeiites that erupted during Hualalai shield stage. It is important to identify source materials involved in the volumetrically dominant stage of Hualalai volcano in order to provide constraints for the size and distribution of compositional heterogeneities of Hawaiian plume. The isotopic compositions of the submarine North Kona tholeiites are similar to the data previously reported for Mauna Loa tholeiites. The data trends define clear mixing relationships that require at least three mantle source components. The mixing is dominated by a Koolau-like enriched component and a Kea-like depleted component. The K219 data trend toward higher epsilon Hf and 87Sr/86Sr relative to the K218, S690, and S692 arrays, requiring another component similar to that observed in post-shield lavas of Hualalai volcano. Pb isotopic compositions for samples from dive K218 and K219 form distinct non-overlapping Pb-Pb arrays suggesting further source heterogeneity. Samples from dive S690 and S692 plot on both of these trends. These findings suggest small-scale compositional heterogeneity in the source regions that can be attributed to anomalous irregular gblobsh involved in late shield stage magmatism of Mauna Loa reported in previous works.
V33A-1155
A broad Galápagos hotspot melting anomaly and disturbance of the underlying core- mantle boundary: A natural laboratory for studying interaction between the core-mantle boundary and overlying lithosphere?
New evidence from direct isotopic dating of the oceanic hotspot record is suggesting that hotspot melting anomalies might be much broader than commonly inferred from the dimensions of individual seamount chains and aseismic ridges and their associated active `volcanic` hotspots. Such an inference is supported by recent thermo-chemical numerical modelling exploring scenarios where upwelling structures are more irregular in shape and behaviour compared to a classic thermal plume `head-tail` (e.g., Farnetani and Samuel, 2006). New age data from the Galápagos Volcanic Province suggest that it developed via the progression of broad regions of widespread, long-lived and possibly concurrent volcanism resulting from tectonic plate motion over a broad Galápagos hotspot melting anomaly (O'Connor et al., 2007). Seismic imaging of the core-mantle boundary under the Cocos plate shows a 100-km vertical step occurring in an otherwise flat D" shear velocity discontinuity (Thomas et al., 2004, Hutko et al., 2006, Kito et al., 2007). One possible explanation is that folding and piling of a cold subducted slab on reaching the core-mantle boundary might account for this lateral variation in terms of a post-perovskite phase change (Thomas et al., 2004, Hutko et al., 2006, Kito et al., 2007). Low velocities inferred at the edge of this proposed slab material may result from the lateral displacement of a thin hot thermal boundary layer leading to upwelling at the tip of the slab, (Thomas et al., 2004, Hutko et al., 2006, Kito et al., 2007), which in turn might possible be connected to our inferred broad hotspot melting anomaly. The combination of the recent imaging of an anomaly at the D"-discontinuity and the inference of a broad overlying Galápagos hotspot melting anomaly suggest that the Galápagos region is an ideal natural-laboratory for studying the possibility of interaction between the core-mantle boundary and overlying lithosphere.
V33A-1156
Lack of Correlated Isotopic and Compositional Variations in Mauna Loa Lavas: A Serious Problem for Pyroxenite/Eclogite Plume Source Models
The long held notion that basaltic magmas are produced by decompressional melting of peridotite is under challenge. Recent models for the Hawaiian and other plumes argue that they consist of a heterogeneous mix of peridotite and discrete eclogite blobs, the latter derived from recycled subducted crust. Eclogite melting produces relatively siliceous magmas (dacite to andesite) which either mix with picritic melts from the peridotite, or, more plausibly, react with the peridotite to produce pyroxenite. Melting of varying proportions of the peridotite/pyroxenite mix is thought to produce the correlated compositional and isotopic characteristics of Hawaiian volcanoes. Magmas from Mauna Loa and Koolau volcanoes are thought to contain more of the recycled component; those from Loihi and Kilauea volcanoes contain less. A simple test of these mixed source models examines whether isotopic changes within the long magmatic history of a single volcano are accompanied by corresponding changes in major and trace element characteristics. Mauna Loa, where we have sampled around 400 - 500 ka of the volcano's eruptive history, provides an excellent opportunity for such a test. During this time, Mauna Loa will have traversed almost half the Hawaiian plume. According to the models, it should have erupted magmas produced from a range of pyroxenite/peridotite mixes with corresponding differences in both isotopic ratios and major and trace elements. Our data show that there is only minor isotopic (Sr, Pb, Nd, Hf) diversity in young lavas (<100 ka), but older lavas are highly diverse, ranging from modern values to those that are close to, and overlap with, those of Loihi volcano. If this isotopic diversity is a consequence of different proportions of pyroxenite and peridotite in the plume source, as the new models predict, we should expect to see correlated changes in bulk composition, particularly. in normalized SiO2, CaO/Al2O3, FeO/MgO and Ni - MgO relationships, as well as changes in Ni - Sc - V relationships. We do not. These parameters remain remarkably uniform over the 400 to 500 ka magmatic history of the volcano, with no correlated variation with isotopic ratios. We conclude that the isotopic heterogeneity within the Hawaiian plume is intrinsic to the peridotite plume source and not dependent on variable contributions from entrained, lithologically-discrete units.
V33A-1157
Small-scale Heterogeneity in Pacific Upper Mantle Unrelated to Plumes Evidenced by Sr-Nd- Pb Isotopic Composition of Basalt from Petit-spot Volcanoes
Petit-spot is a new type of volcano that is not related to plumes on the northwestern Pacific plate. An important aspect of petit-spot is that this provides a unique window for mantle source material beneath the oldest Pacific plate. To investigate the geochemical features of upper mantle unrelated to plumes in northern hemisphere, we have determined Sr, Nd, and Pb isotopic compositions of basalts collected from petit-spots. The isotope compositions of petit-spot lava fall within the range of compositional variations in MORB. However, the lava is higher in 87Sr/86Sr, lower in 206Pb/204Pb and 143Nd/144Nd, and higher in 207Pb/204Pb and 208Pb/204Pb at a given 206Pb/204Pb than those of Pacific MORB or Depleted MORB Mantle (DMM). Remarkably, the isotopic compositions are equivalent to Indian and southern Atlantic MORB with strong Dupal signature (MORB from the southwest Indian ridge at 40°E, 43.4°S, and the southern mid-Atlantic ridge at 10°-14°W, 46°-49°S), and are almost identical to the composition of the EM-1 endmember. We propose that recycled material from an ancient subduction system or crustal delamination resides throughout the upper mantle of the Pacific Ocean as minor blobs. Our model of 'small-scale recycled material melting' consistently explains geochemistry and noble gas isotopic composition of magma, and volume of petit-spot volcanism. Furthermore, our results strongly suggest that the Earth's upper mantle is essentially heterogeneous, but its gross composition is controlled by abundances of blobs of recycled material. Upper mantle located beneath (such as the Indian, Atlantic, and possibly Arctic oceans) or on the periphery (such as the Philippine Sea) of the Pangea supercontinent includes abundant recycled material, in contrast to the Pacific Ocean. Because the Dupal signature was discovered in the northern hemisphere, the nature of the Dupal signature must be reconsidered as involving recycling of the entire solid Earth.
V33A-1158
Thermal and compositional contributions to mantle heterogeneity
We have integrated seismic, geodynamic and mineral physics constraints to obtain models of mantle heterogeneity. The seismic constraints primarily consist of teleseismic shear body wave travel times and the geodynamic constraints include the global free-air gravity field, tectonic plate divergences, dynamic surface topography and the excess ellipticity of the core-mantle boundary. The geodynamic observations are interpreted with viscous flow response kernels and velocity-to-density scaling relationships for thermally-varying mantle material. Considering the viscosity model from Mitrovica & Forte (2004) and an optimal thermal density-velocity scaling relationship, we have found a single velocity/density model capable of satisfying the combined dataset to high degrees. The implication is that thermal variations dominate and compositional contributions are secondary throughout most of the non-cratonic mantle (Simmons et al. 2007). This modeling approach inherently minimizes potential non-thermal contributions to the density field and thus establishes a minimal estimate of the influence of composition needed to reconcile the observations. This is due to the fact that we use optimal viscosity and density-velocity scaling relationships to model the data with the initial assumption that all heterogeneity is generated by thermal variations. Therefore, we test other possibilities including simplified viscosity profiles and alternative thermal density-velocity relationships in order to evaluate how these input parameters increase the level of compositional influence required to satisfy the combined dataset. We also demonstrate the potential downfall of scaling a purely seismically-derived shear velocity model to obtain density heterogeneity in the mantle. This work was performed under the auspices of the U.S. Department of Energy by the University of California, Lawrence Livermore National Laboratory under Contract No. W-7405-Eng-48. UCRL-ABS-233968
V33A-1159
Geochemistry of Post-Shield and Secondary Volcanism on and Around the Island of Kaua`i
The Koloa Volcanics are the most voluminous (~ 58 km3) and long-lived (2.4 m.y.) example of Hawaiian rejuvenated volcanism. The recent discovery of extensive Koloa volcanism offshore heightens their importance in helping to determine the mechanism(s) responsible for downstream plume-related volcanism, as do post-shield lavas. We collected a suite of K-Ar dated, post-shield and Koloa lavas and new offshore samples for major and trace element, and Pb, Sr, and Nd isotope characterization for comparison with well characterized shield lavas. Previous studies of Koloa lavas were compromised by rock alteration. X-ray fluorescence analyses of new, unaltered to weakly altered rocks indicate the Koloa Volcanics are exclusively alkalic, ranging from alkali basalts to foidites, while the post-shield stage lavas range from tholeiites to hawaiites and basanites. Preliminary results show that four of five offshore blocks are tholeiitic, and may be related to the Kaua`i shield. The other is geochemically similar to Koloa Volcanics. The post-shield stage is primarily defined by its age range, rather than by its geochemistry. These lavas were erupted between 4.0-3.6 Ma vs. 5.14-4.0 Ma for shield lavas, and were followed by the rejuvenated stage after a 1.2 Ma hiatus. Major and trace element data show overlapping compositions for the post-shield stage and the shield or rejuvenated stages. Based on trace element ratios, two geochemically distinct groups can be distinguished within the post-shield stage. This distinction is shown clearly on plots of Ba/Ce vs. Zr/Nb and Ba/Sr vs. Zr/Nb, where the alkalic samples plot near to, but not within, the Koloa lavas field and the tholeiitic samples, plot in a separate field, within the tholeiitic shield field. No temporal gradational trend exists in trace element ratios from shield stage through post-shield to rejuvenated stages. Whereas the shield and rejuvenated stages form separate fields, post-shield stage lavas do not form a transitional field; rather it is sharply divided into two groups indicating the source of the rejuvenated lavas is probably related to the Hawaiian plume rather than the lithosphere. Preliminary Sr and Nd data for the tholeiitic post-shield stage lavas are typical of shield lavas. The forthcoming isotopic data for the alkalic post-shield lavas will supplement this characterization and possibly allow for a geochemical distinction between the post-shield and rejuvenated stages.
V33A-1160
Towards an Improved Eruptive History for Mauritius
The volcanic island of Mauritius (20°20' S, 57°30' E) is located in the western Indian Ocean, and is understood to be the product of the Réunion mantle plume. Mauritius has a well-established history of episodic volcanism and erosional hiatus, traditionally characterized as three chemically and temporally distinct eruptive phases: 1) the voluminous shield-building lavas of the Older Series (7.8 to 5.5 Ma), 2) the Intermediate Series (3.5 to 1.9 Ma), and 3) the Younger Series (1.00 to 0.00 Ma; Duncan, unpublished data). The Older Series is exposed in remnants of the old shield volcano and is hence analogous to the Hawaiian shield-building phase. Lavas of the Intermediate and Younger Series cover 90% or more of the surface of the island, but are thought to be volumetrically minor. The Intermediate and Younger Series are enriched in incompatible elements, and but have more depleted isotopic signatures than the Older Series and in this and other ways are analogous to Hawaiian post-erosional volcanism. Recent collaboration with the Mauritian Water Resources Unit has permitted the study of a series of newly available drill cores. Dense sampling of a 220-meter core section (#89, Beau Bois, 20°12'25.9'' S, 57°33'50.8'' E) has revealed the presence of nine flow sequences, separated by highly weathered and altered basalt and lateritic soils varying in thickness up to 20 meters. The drill site is located on Intermediate Series lavas, approximately one kilometer inside the topographic rim of the large central caldera complex. Isotopic analysis shows relatively uniform 87Sr/86Sr (0.70377 ± 0.00018) and \varepsilonNd (+5.1 ± 0.3) throughout the upper 180 meters of the core. These compositions match those previously reported for the rejuvenescent Intermediate and Younger Series. The deepest sample has more enriched 87Sr/86Sr (0.70422) and \varepsilonNd (+4.1), values that fall well within previously established fields for the Older Series. In contrast to recent work ( Paul et al, 2007, Compositional diversity among primitive lavas of Mauritius, Indian Ocean: Implications for mantle sources, JVGR, 164, 76-94) that assumed the sole presence of Older Series lavas at depth and interpreted variation as diversity within the Older Series, we interpret this core to be predominantly Intermediate Series with only the deepest sample classified as Older Series. These results suggest that volumes of rejuvenescent lavas on Mauritius are greater than had been realized. Rejuvenescent volcanism on Mauritius thus contrasts strongly with that occurring on Hawaii and other Pacific islands, both in erupted volumes and in duration. Rejuvenescent volcanism on oceanic islands remains enigmatic in the context of models and intraplate volcanism, and nowhere more so than on Mauritius.
V33A-1161
Mass and Composition of the Continental Crust Estimated Using the CRUST2.0 Model
The mass, age, and chemical composition of the continental crust are fundamental data for understanding Earth differentiation. The inaccessibility of most of the volume of the crust requires that inferences be made about geochemistry using seismic and heat flow data, with additional constraints provided by scarce lower crustal samples (Rudnick and Fountain, Rev. Geophys., 1995; Rudnick and Gao, Treatise on Geochem., 2003). The global crustal seismic database CRUST2.0 (Bassin, et al., EOS, 2000; Mooney, et al., JGR, 1998; hereafter C2) provides a useful template with which the size and composition of the continents can be assessed, and may be a useful vehicle to organize and analyze diverse geochemical data. We have used C2 to evaluate the modern mass and composition of the continental crust and their uncertainties, and explored our results in the context of global mass balances, such as continents versus depleted mantle. The major source of uncertainty comes from the definition of "continent." The ultimate constraint is the total mass of Earth's crust (oceanic + continental), which, from C2, is 2.77 (in units of 1022 kg). Using crustal thickness as a definition of continent, the mass of continental crust (CC) is 2.195 if the minimum thickness is 12-18km, 2.085 for 22.5km, 2.002 for 25km, and 1.860 for 30km. These numbers include all sediment as continental crust. Using C2 definitions to distinguish oceanic and continental crust (and including oceanic plateaus which contain some continental crust), we calculate the CC mass as 2.171. To estimate chemical composition, we use the C2 reservoir masses. For minimum thickness of 22.5km, C2 yields the proportions 0.016 oceanic sediment, 0.038 continental sediment, 0.321 upper crust, 0.326 middle crust, 0.299 lower crust. Upper, middle, and lower crust are assigned compositions from Rudnick and Gao (2003), continental sediments are assigned upper crust composition, and oceanic sediments are assigned GLOSS composition (Plank and Langmuir, Chem. Geology, 1998). Because the C2 model gives a larger proportion of upper versus lower crust, the resulting CC composition is enriched in P2O5, K2O, Rb, Pb, Th, U, Hf and LREE, and depleted in MgO and CaO, relative to Rudnick and Gao (2003). The mass and chemical composition of CC figure prominently in calculations of the mass of depleted mantle (DM) that is required to balance the crustal enrichments in incompatible elements. However, because thin crust also tends to be young crust, and is less enriched in incompatible elements, uncertainties in the mass of CC have little effect on the calculated size of DM, which must be considerably smaller than the whole mantle. Using different approaches to estimating crustal composition, and adding age provinces as well as other types of geochemical and geochronological data, could make CRUST2.0 an important geochemical resource.
V33A-1162
Sulfur Earth
Variations in surface tension affect the buoyancy of objects floating in a liquid. Thus an object floating in water will sink deeper in the presence of dishwater fluid. This is a very minor but measurable effect. It causes for instance ducks to drown in aqueous solutions with added surfactant. The surface tension of liquid iron is very strongly affected by the presence of sulfur which acts as a surfactant in this system varying between 1.9 and 0.4 N/m at 10 mass percent Sulfur (Lee & Morita (2002), This last value is inferred to be the maximum value for Sulfur inferred to be present in the liquid outer core. Venting of Sulfur from the liquid core manifests itself on the Earth surface by the 105 to 106 ton of sulfur vented into the atmosphere annually (Wedepohl, 1984). Inspection of surface Sulfur emission indicates that venting is non-homogeneously distributed over the Earth's surface. The implication of such large variation in surface tension in the liquid outer core are that at locally low Sulfur concentration, the liquid outer core does not wet the predominantly MgSiO3 matrix with which it is in contact. However at a local high in Sulfur, the liquid outer core wets this matrix which in the fluid state has a surface tension of 0.4 N/m (Bansal & Doremus, 1986), couples with it, and causes it to sink. This differential and diapiric movement is transmitted through the essentially brittle mantle (1024 Pa.s, Lambeck & Johnson, 1998; the maximum value for ice being about 1030 Pa.s at 0 K, in all likely hood representing an upper bound of viscosity for all materials) and manifests itself on the surface by the roughly 20 km differentiation, about 0.1 % of the total mantle thickness, between topographical heights and lows with concomitant lateral movement in the crust and upper mantle resulting in thin skin tectonics. The brittle nature of the medium though which this movement is transmitted suggests that the extremes in topography of the D" layer are similar in range to those observed on the Earth's surface and are mimicked by lows under the oceans and highs under the altiplanos. Careful and area selective S wave core mantle ellipsometry might be able to discern these core-mantle topographic variations. As such this process demonstrates the validity of the Gaia hypothesis enunciated by Baas Becking(1931) that no ecological niche on our planet is closed off from other niches "nothing in the world is single".
V33A-1163
The importance of magnesite and CO2 in the initiation and subsequent evolution of mantle plumes
To better understand the nature of the processes during the generation and evolution of mantle plumes it is imperative that geochemical, geophysical and mineral experimental observations are combined to generate realistic computational models. In an axi-symmetrical finite element model we make use of observations that; (1) some mantle plumes can be imaged to D"; (2) mantle plumes have high contents of CO2 (300 to 1000 ppm; [1]) of which only 12 ppm can partition into olivine [2]; (3) high-pressure experimental studies show that although carbonate is present as a liquid at upper mantle conditions [3], it is stable as a magnesite solid solution (Mg,Ca CO3) at <14 GPa and throughout the lower mantle [4]. However at D" it disassociates to MgO and CO2 [5]. CO2 is ubiquitous in mantle materials. CO2 is the dominant phase that degasses in purely mantle derived melts both mid ocean ridge basalts and ocean island basalts. That mantle plumes have high CO2 contents and are derived from the lower mantle requires that magnesite plays an integral role in plume evolution. The stability of magnesite varies dramatically in the Earth from D", where deep sourced plumes originate, to the upper mantle, where generation of basaltic melts occurs. In our simulations we attempt to address the scenario where CO2-bearing liquids can be generated at the D", can freeze in the lower mantle and again melt as a thermal upwelling passes into the upper mantle. Our computational study focuses specifically on the dynamics of the generation and evolution of a mantle plume containing magnesite in its source at the D". Our initial results centre purely on the affect of plume migration across the stability field of magnesite. This allows investigation of migration and accumulation of carbonate melt both during plume initiation and as plumes reach steady state. Preliminary results provide constraints for the dynamics and interaction of carbonate and silicate melts in mantle plumes. We propose that carbonate melting plays an extremely important role in plume evolution. 1 Trull et al. 1993; EPSL, 118; 43-64 2 Shcheka et al. 2006; EPSL, 245; 730-742 3 Dasgupta et al. 2007, Am Min, 92; 370-379 4 Fiquet et al. 2002, Am Min, 87; 1261-1265 5 Isshiki et al. 2004; Nature; 427 60-63
V33A-1164
Heterogeneity in the Oceanic Mantle From Nd Isotopic Composition of the Abyssal Peridotites
Abyssal peridotites (APs) are believed to be the residue of the mid ocean ridge basalt (MORB) melting event, and thus provide information complementary to the basalts. We present the trace element concentrations, and Sr and Nd isotopic compositions of clinopyroxene (cpx) separates from the APs from three different slow to ultra-slow spreading ridges -Southwest Indian Ridge (SWIR), Mid Cayman Rise (MCR) and Mid Atlantic Ridge Kane Fracture Zone (MARK). These areas were chosen because they are thought to represent relatively low degrees of melting (based on ridge depth and Na8), which thus would enhance the signature of an enriched (low solidus) components in the melt and thus provide maximum contrast between peridotites and basalts. At SWIR we have analyzed 15 peridotites from three dredges from 90 to 160 oblique segment. Dredge 85 peridotites, recovered from avolcanic segment, are harzburgitic (<5% modal cpx) and are LREE depleted (Ce/Yb)N=0.001-0.035 and CeN=0.003-0.11. Dredge 78 peridotites have up to 5% of modal cpx, and are moderately depleted in REEs (CeN = 2-5). In contrast to other two dredges, dredge 96 peridotites are lherzolitic and two samples some show a pyroxenite vein. These APs have approximately 10% modal cpx indicating either a low degree of melting or refertilization. The light REE pattern of the cpx (CeN=8) is also less depleted and these are the only cpx whose REE pattern can be in equilibrium with MORB. Nd-isotopic compositions of the cpx from dredge 78 and 96, 143Nd/144Nd=0.513003-0.513046 and 0.512885- 0.512933 respectively, which falls within the Nd-isotopes ratios of the associated basalts. In contrast, peridotites from dredge 85 display a large range in Nd-isotopic composition(0.513086-0.513320), and their Nd isotopic compositions are more radiogenic than the associated ridge basalts. At MCR, peridotites from two dredges are dominated by lherzolite with relatively high modal abundance of cpx (up to 9 modal%). 13 peridotites analyzed from MCR are characterized by strong depletion in light REEs with (Ce/Sm)N as low as 0.002 and by large range of variation in heavy REEs with LuN ~0.10-0.47. 143Nd/144Nd isotope ratios of the cpx range from 0.513111-0.513425 which partially overlaps with the associated basalts but extends to more radiogenic compositions. Peridotites from the MARK area are harzburgitic with depletions in light REEs with CeN as low as 0.004 and (Ce/Yb)N=0.002-0.20. In comparison to SWIR and MCR, the cpx from MARK area are characterized by low abundance of middle to heavy REEs (DyN=0.8,LuN=1), indicating an overall larger degree of depletion. However, the Nd-isotopic compositions (0.513011-0.513190) of the cpxs are indistinguishable from the associated basalts. These results lead to several constraints on the ridge process: 1. Assuming that the peridotites traveled along similar pressure-temperature paths, the large range of variation in trace element ratios at a single dredge in a limited geographic area can not be explained by the degree of melting alone, requiring significant source variation. 2. The discrepancy between peridotites and basalts are too large to explain by radiogenic ingrowth during ascent (after melt extraction). 3. Although a low solidus component is required to explain the isotopic differences at the SWIR and MCR, the similarity of the Nd-isotopic composition of the basalts and peridotites at the MARK area indicates this low solidus component is not necessarily a ubiquitous feature of the depleted mantle.
V33A-1165
Not all Primordial Noble Gas Signatures are Associated with OIBs and Mantle Plumes – Mantle Heterogeneity, Primordial Shallow Sources and a Solar-like He, Ne Signature in an Ancient North American Craton
The presence of primordial He and Ne components in ocean island basalts (OIBs) as well as a mantle He/heat flux ratio lower than the production ratio near mid-ocean ridges have historically been used to support the existence of a two-layer mantle convection model. This would comprise a lower, primordial, undegassed reservoir from which He removal to the upper degassed mantle would be impeded. Arguments based on He and heat transport have been recently invalidated by Castro et al. (2005) and should no longer be used to justify the presence of two such distinct mantle reservoirs. Indeed, it was shown that such low He/heat flux ratios are expected and do not reflect a He deficit in the original crust or mantle reservoir. By contrast, the occurrence of a He/heat flux ratio greater than the radiogenic production ratio can only result from a past mantle thermal event in which the released heat has already escaped while the released He remains, and is slowly rising to the surface. Such a high He/heat flux ratio is present in shallow groundwaters of the Michigan Basin. We now present results of a new noble gas study conducted in the Michigan Basin, in which 38 deep (0.5-3.6km) brine samples were collected and analyzed for all noble gas abundances and isotopic ratios. As expected from previously computed shallow high He/heat flux ratios, both He and Ne isotopic ratios clearly indicate the presence of a mantle component. Of greater significance is the primordial, solar-like signature, of this mantle component. It is also the first primordial signature ever recorded in crustal fluids in a continental region. Because no hotspots or hotspot tracks are known in the area, it is highly unlikely for such primordial, solar-like signature to result from a mantle plume-related mechanism originating deep in the mantle. We argue that such a primordial signature can be explained by a shallow noble gas reservoir in the subcontinental lithospheric mantle (SCLM) beneath the Michigan Basin, possibly created by a mechanism similar to that proposed by Anderson (1998) for oceanic regions. Indeed, the Michigan Basin, located within the ancient North American craton (~1.1->2.5Ga), lies on a very thick U-Th depleted SCLM, possibly allowing preservation of a primordial, residual, mantle reservoir beneath the continental crust. Recent reactivation of the old mid-continent rift transecting the crystalline basement is likely responsible for the release of this primordial signature into the basin. The solar-like He and Ne signatures present in the Michigan Basin fluids not only suggest that a deep primordial mantle reservoir is not required to explain the presence of such components, they also point to a very heterogeneous mantle as previously suggested by Anderson (1998), Albarede (2005), and others. Consequently, the presence of a primordial noble gas signature, at least if observed in a continental region, should not be used to conclude at the existence of a deep mantle source and thus, of a hotspot as typically defined. The SCLM underneath ancient cratons is a great candidate for hosting primitive ancient mantle reservoirs. Arguments based on He/heat flux ratios as well as the presence of a primordial noble gas signature should not be used to support the existence of a lower, primordial, versus an upper, degassed mantle reservoir. Our study provides the first observational case for long-term primordial lithospheric storage. Anderson, 1998, Proc. Natl. Acad. Sci. USA, 95, 9087-9092. Albarede, 2005, AGU Monograph, 160, 27-46. Castro et al., 2005, EPSL, 237, 893-910.
V33A-1166
The Afar Plume and Ethiopian Lithosphere: A Hf and He Isotope Study of Primitive Quaternary Rift Lavas.
Quaternary magmatism in the Ethiopian rift records the interaction of a deeply sourced mantle plume with the ambient upper mantle and the continental lithosphere. We present the first hafnium isotope values for the region and utilize new helium data from these well-characterized samples to gain insight into the distribution of Afar plume material in the Ethiopian mantle. Modern volcanism in the region is focused in two distinct NE-SW trending linear zones of extension within the rift. Chains of cinder cones erupting between large silicic centers typify both zones. The Wonjii Fault Belt (WFB) is the larger of these zones both in length and eruptive volume, extending into the Afar Depression. The Silti-Debre Zeyit Fault Zone (SDFZ) is the smaller, extending only to 8.5 degrees north, and ending against a protrusion of the rift wall. We present 12 new 3He/4He and 18 new 176Hf/177Hf results on samples that have been previously analyzed for Sr, Nd and Pb isotopic ratios, making these samples among the best geochemically characterized in the region. We observe that samples within the WFB consistently have higher 3He/4He (R/Ra = 12.2 to 15.1) than those of the SDFZ (R/Ra = 5.6 to 8.8) and extend to more radiogenic Hf isotope values. Hf isotope values correlate positively with 3He/4He and 206Pb/204Pb. Distinct linear Pb isotope arrays for Butajira, WFB, and Debre Zeyit, suggest 3-component mixing between the "C"-like Afar Plume, asthenosphere and lithosphere. In this scenario the proportions of Afar plume and ambient asthenospheric mantle are relatively fixed, suggesting that they interacted prior to mixing with heterogeneous lithosphere. The WFB, Debre Zeyit, and Butajira lavas define Pb-Pb trends that radiate away from the Afar plume-ambient mantle mix towards continental lithosphere-like compositions. The WFB has a greater proportion of Plume plus asthenosphere source material while lavas along the SDFZ have an increased lithospheric source contribution. We suggest that plume material may be channeled along the rift's thinned lithosphere southward from its hypothesized current location in the Afar Depression, mix with the ambient upper mantle, and subsequently interact with the continental lithosphere to form the rift lavas.
V33A-1167
Chemical and Isotopic Heterogeneities in the Deep Earth:Importance of Lower Mantle Carbonate-rich Melts
Evolution of mantle chemical heterogeneity reflects a spectrum of processes. Nature of reservoirs has been inferred from radiogenic isotope and trace element systematics of mid-ocean ridge basalts (MORB) and ocean island basalts (OIB) [1]. Carbonatites, kimberlites and lamproites [2-4] also sample depleted and enriched reservoirs, however, their origin remains equivocal. Secular decrease in Th/U ratio in MORB mantle (DMM), homogeneity of Th/U inferred from Pb-isotopic data, and systematic variation in Nb/Th and Nb/U ratios in MORBs [5], show that recycled components in DMM are well mixed. Thus isotopically hererogeneous domains in DMM must be transient features and are unlikely to yield HIMU and EM chemistries. Explanations for HIMU and EM OIB chemistries include involvement of: (1) subcontinental lithospheric mantle; (2) subducted oceanic lithosphere; (3) subducted sediment; or (4) an enigmatic lower mantle (LM) "plume component". Elevated 3He/4He in OIBs and kimberlites [6] and excess 129Xe and high 40Ar/39Ar [e.g., 7-8] and solar 20Ne/22Ne [9] in carbonatites indicate that they were derived from a primitive, isolated, and less degassed source than MORB. Primordial compositions show that this reservoir escaped atmospheric contamination by Ar, Xe, and Ne and pollution by 4He-rich material (from recycled 238U) during subduction. This primitive reservoir likely exists below the depth subducted slabs obviously penetrate (ca. 1700 km) e.g., [10]. That kimberlites are deeply sourced is also shown by lower mantle inclusions in diamond, e.g., [11]. Importantly, Gp. 1 and 2 kimberlites are isotopically similar to HIMU and EM-1 OIBs [4]. We interpret Gp 1 kimberlites as mixtures of HIMU and EM sources, while Gp. 2 kimberlites (close to EM-1) are interpreted as melts of a Ca perovskite-rich reservoir, possibly from slabs in the LM. We model melting of LM phases to simulate evolution of EM1 and HIMU 87Sr/86Sr, 143Nd/144Nd, 176Hf/177Hf, 207Pb/204Pb, 206Pb/204Pb and 208Pb/204Pb isotopic compositions. HIMU appears to be the residue of LM melting, and thus is indicative of a buoyant deeply-derived component. Thus, U/Pb, Th/Pb, Rb/Sr, and Sm/Nd ratios of kimberlite and carbonatite likely reflect mode of the LM mineral assemblage being melted, as well as the buoyancy of melts and residues. The CO2-rich nature of kimberlite and carbonatite magmas is clearly important for melt transport in LM- derived plumes. Carbonate phases are stable at high pressures [e.g., 12] and available elastic and density data on carbonate and C-bearing melts show that LM CO2-rich melts are less dense than their source, and are able to migrate, in contrast to LM silicate melts. C-enrichment thus allows extraction of deep mantle melts. Escape of these LM-derived melts provides a mechanism to explain transfer of mantle isotope heterogeneity, as well as the rare gas isotope systematics of plume magmas. [1] Zindler, A. & Hart, S., 1986. AREPS. 14: 493-571. [2] Bell, K. & Tilton, G.R., 2001. J. Pet. 42: 1927-1945. [3] Murphy, D., Collerson, K.D. & Kamber, B.S., 2002. J. Pet. 43: 981-1001. [4] Smith, C.B., 1983. Nature, 304: 51-54. [5] Kamber, B.S. & Collerson, K.D., 1999. JGR, 104: 25479-25491. [6] Tachibana, Y. et al., 2006. Geology, 34:273-276. [7] Sasada, T. et al., 1997. GCA, 61: 4219-4228. [8] Tolstikhin, I.N. et al., 2002. GCA, 66: 881-901. [9] Yokochi, R. & Marty, B. 2004. EPSL, 225, 77-88. [10] Fukao, Y. et al., 2001. Rev. Geophys., 39: 291-323. [11] Hayman, P.C. et al., 2005. CMP. 149 430-445 [12] Isshiki M. et al., 2003. Nature, 427, 60-63.
V33A-1168
Ultra-Refractory Domains in the Oceanic Lithosphere: Evidence from Major Element and Modal Relationships in Mantle Xenoliths from Ocean Islands
We have compiled a database for major element and modal data of mantle harzburgite and lherzolite xenoliths from different ocean islands. The xenoliths fall in two main categories. Xenoliths showing no petrographic evidence of metasomatism (OI1) from the Canary Islands, Kerguelen, Cape Verde and Samoa are ultra-depleted spinel harzburgites. These xenoliths formed by high degrees of partial melting leading to total exhaustion of cpx. Small amounts of cpx observed in these rocks (mainly <3 vol. percent) have partly formed through exsolution from opx followed by recrystallization, and partly through reactions between mantle wall-rocks and LREE-enriched metasomatic melts. Peridotites from the same islands (and the Azores) which show petrographic evidence of metasomatism (OI2) partly fall within the major element and modal range of OI1 xenoliths, partly they are less refractory. On average, the OI1 xenoliths are significantly more refractory than peridotites collected along mid- ocean ridges and fracture zone (MORP), although the most refractory mid-ocean ridge peridotites fall within the OI1 range. Although the OI1 xenoliths show indications of metasomatism in trace elements, Na2O, and radiogenic isotopes, their major element and modal compositions seem to be essentially unaffected by metasomatic processes. The ultra-depleted nature of the OI1 xenoliths is also reflected in the Fo-contents in ol, and in the HREE and MREE in opx. OI2 xenoliths are OI1 rocks, many of which have been metasomatized to the degree that major element and modal phase compositions have been modified. The other main category, consisting of harzburgite and lherzolite xenoliths collected in Hawaii, Tahiti and Grande Comore (OIcpx) are significantly less refractory that the OI1 xenoliths and show similarities to MORP. Many of these xenoliths contain significant amounts of cpx (up to 20 vol. %), and contain primary cpx, indicating only moderate amounts of partial melting (or extensive re-fertilization). Unradiogenic Os isotopes and high Re-Os model ages suggest that OI1 and OI1cpx rocks from some islands are much older than the abyssal lithosphere in which they are found. P-T estimates of 0.7-1.3 GPa and 850- 1200„aC indicate that the OI1 xenoliths last equilibrated at depths corresponding to >80 Ma old oceanic lithospheric mantle. We interpret the ultra-depleted xenoliths as fragments of ¡§exotic¡¨ material trapped in the oceanic mantle lithosphere. Major element similarities to some series of oceanic sub-arc mantle, and significant differences from continental xenolith series, makes it most likely that the ultra-refractory OI1 peridotites represent recycled abyssal mantle. The large proportion of ultra-refractory peridotite xenoliths in ocean islands, and the presence of such rocks also along some mid-ocean ridges and in sub-arc mantle, suggest that ultra-refractory material may be important ingredients in the convecting mantle. Because of their ultra-depleted chemistry the OI1 type harzburgites will have relatively low densities and are therefore buoyant relative to less refractory mantle rock types. Their high solidus temperatures make the OI1 xenoliths immune to further partial melting (sterile).
V33A-1169
Correlation of Sr, Nd, Pb and He Isotopic Variation With Ridge Segmentation: Interaction Between the Southeast Indian Ridge (77-88°E) and the Amsterdam/St. Paul Hotspot
The Southeast Indian Ridge (SEIR) bisects the Amsterdam-St. Paul (ASP) Plateau, which is a submarine, bathymetric high that includes the islands of Amsterdam and St. Paul and which rises two km above the surrounding seafloor. The origin of this volcanic plateau is attributed to a mantle hot spot. We obtained Sr, Nd, and Pb (n=37) and He isotope ratios for basalt glasses from 11 SEIR segments on and adjacent to the plateau and from three plateau seamounts. The results show systematic spatial variations in these isotopic ratios that correlate with physical segmentation of the ridge. Specifically, lavas from the four ridge segments on the ASP Plateau have higher 208Pb/204Pb at a given 206Pb/204Pb, than SEIR MORB (mid-ocean ridge basalt) distant from ASP. Surprisingly, lavas from the ridge segment 100 km north of the ASP Plateau are distinguished by the most radiogenic 206Pb/204Pb (up to 19.5) and highest 3He/4He ratios (up to 14.1 RAir). These are characteristics of lavas erupted at Amsterdam and St Paul Islands; consequently. Overall, the Pb-Nd-Sr-He isotope variations within this dataset are explained well by the interaction of three mantle end-members: 1) depleted mantle with relatively low 206Pb/204Pb and 87Sr/86Sr, and high 143Nd/144Nd, which has been variably mixed with 2) material with relatively high 208Pb/204Pb and 87Sr/86Sr and low 143Nd/144Nd at a given 206Pb/204Pb, a signature commonly ascribed to detached or eroded metasomatized, continental lithosphere, and 3) hotspot-related mantle with elevated 3He/4He and 206Pb/204Pb but intermediate 87Sr/86Sr and 143Nd/144Nd, similar to the common or C material observed in hotspots globally. These results suggest either that the ASP hotspot is isotopically heterogeneous or that the lithosphere or shallow asthenosphere beneath the ASP Plateau contains more continentally-derived material than the SEIR mantle > 500 km away. Perhaps rafts of continental material lie stranded within a local °tectonic corridor° beneath the ASP Plateau, possibly since the opening of the Indian Ocean basin.
V33A-1170
Isotopic variations within upper oceanic crust at IODP Site 1256: Implications for crustal recycling and the formation of ocean island basalts
The origin of ocean island basalts (OIBs) is a fundamental question facing Earth scientists. It is commonly agreed that lithospheric material recycled in the mantle is involved in the magma source of OIBs. The relative importance of 1) subducted altered oceanic basaltic crust (AOC), 2) subducted marine sediments and/or 3) delaminated metasomatised subcontinental lithosphere and continental lower crust remains to be resolved. We examine the geochemical composition of a complete in situ section of oceanic crust drilled at Site 1256 during IODP Expeditions 309 and 312. It includes the extrusive layer, sheeted dikes and gabbros of ca. 15 Ma old oceanic crust of the Cocos Plate formed during a period of superfast spreading at the East Pacific Rise. Modeling in the Sr-Nd-Pb-isotope space and comparison with present day radiogenic isotope ratios of OIBs provides constraints on the significance of recycled oceanic crust in the OIB mantle source(s). Our study shows that the generation of sulphides during low- and high-temperature alteration of oceanic crust has a strong influence on U/Pb and Th/Pb ratios and whether an AOC domain evolves relatively low or high Pb-isotope ratios over geological timescales. The model suggests that AOC as the sole precursor material, modified during the subduction process, and after relatively low to moderate recycling ages of ca. 300-800 Ma, is sufficient to explain the Sr-Nd-Pb-isotopic composition of OIBs with Pb-isotopic compositions along or below the Northern Hemisphere Reference Line (NHRL) and relatively high Nd-isotope ratios (e.g. Canaries, Galapagos, Iceland, Madeira). This indicates that additional EM-components, potentially associated with recycled lithospheric material such as subducted sediments, lower continental crust or subcontinental lithosphere, are not required for an array of OIBs, but are only necessary to explain OIBs with Pb-isotope ratios above the NHRL and relatively low Nd- isotope ratios (e.g. Pitcairn, Tristan, Samoa).
V33A-1171
Reduced Seismic Anomalies in the Lower Mantle: Role of the High-Spin to Low-Spin Electronic Transition
We determined the pressure dependence of the single crystal elastic constants of ferropericlase (6% iron) to a static pressure of 63 GPa. The velocities of the quasi-longitudinal and interfacial elastic wave were measured in the (100) crystal plane using impulsive stimulated light scattering. We observed anomalies in measured velocities over the pressure range of 40 to 60 GPa and associated them with the high-spin to low-spin (HSLS) transition in the iron cations of ferropericlase. We find that all three elastic elements (c11, c12, c44) smoothly pass through minima with respect to an extrapolation of the pure high-spin state. This behavior is consistent with a macroscopic thermodynamic description for the HSLS transition that predicts no sharp changes in spin fractions and dependent physical properties. The associated enthalpy and volume changes scale correctly between these data and literature compression data obtained on a sample with higher iron concentration. A significant geophysical consequence of the observed behavior is that in the spin transition pressure regime, normal temperature derivatives of elasticity are reduced. Within Earth, a minimum in the temperature derivatives caused by the HSLS transition is predicted to occur at a depth of 1500 km. This correlates well with the depth where seismic tomographic results show minimum structure and model power. Thus, the HSLS transition may serve to mask velocity anomalies associated with lateral temperature differences.
V33A-1172
Hidden Records Of Critical Melt Fraction? Constant Intergranular Component In Mantle Xenoliths
We recently discovered that the chemical composition of grain-grain interfaces (i.e., grain and interphase boundaries) relative to that of the grain interiors can be compared to element partitioning between melt and grains if the interface thickness is taken to be 2 mono-atomic layers. This analogy allows precise investigation of the composition of the intergranular component present in a rock based on a comparison of the concentrations of elements in a "whole rock" with those in a "clean rock" combined with the mineral mode, grain size, and temperature of equilibration. In this analysis, a clean rock is one reconstructed from the composition and mode of the minerals. In this paper, we have extracted information about intergranular components from data reported for 20 mantle xenoliths collected from 5 localities worldwide. We previously demonstrated that, at chemical equilibrium, interfaces are enriched in incompatible elements and that the enrichments are positively correlated with excess energies required for the solution of the elements into the crystal lattice [1]. The intergranular components in all the samples are consistently ~103 larger than expected from equilibrium chemical segregation at grain-grain interfaces. We argue that this intergranular component, which is common to all mantle xenoliths, corresponds to an unrecognized ~0.2 vol% of melt, now possibly converted into micro-glass or micro- crystalline phases. Based on plausible upwelling rates for mantle xenoliths, it is unlikely that this melt formed during ascent. Instead, we conclude that this small amount of melt is a remnant of previous melting of the mantle. Consequently, it corresponds to the critical melt fraction below which melt is unable to escape from the rock by porous flow. [1] Hiraga T, Kohlstedt DL (2007) Geochimica et Cosmochimica Acta 71: 1266-1280
V33A-1173
High-Precision Nd isotopes in Picrites from Hawaii and Iceland - No Evidence for an Early- Formed Enriched Reservoir
New high-precision Nd isotope measurements have shown that the present-day 142Nd/144Nd for average chondrites is approximately 20 ppm lower than that for the terrestrial standard and all recent mantle- derived samples measured to date. One explanation for these differences is that an enriched ‘missing' reservoir with lower 142Nd/144Nd resides in the mantle to balance the bulk Earth to chondritic. Data from Archean Greenland rocks show anomalies in 142Nd/144Nd of up to 40 ppm higher than the proposed average for chondrites. This difference between the Archean Greenland rocks and present-day mantle-derived samples has been interpreted to result from remixing of an early-formed enriched reservoir into the convecting mantle during the Archean. If so, the implication from this is that remixing of the enriched reservoir largely shut down some time in the past such that it must at present reside in a region in the mantle that infrequently participates dynamically and leading to volcanism at the surface. Several studies have suggested that the missing reservoir may be present just above the core-mantle boundary (CMB). Depending on the size of this reservoir it could potentially make up all of D" or exist as distinct domains within it. If such a reservoir does exist, then mantle-derived samples from plume systems that are thought to come from the CMB may be the best opportunity to identify this component using high-precision Nd isotope measurements. To test this, picrites from Hawaii with coupled enrichments in 186Os-187Os that has been proposed to be a signature of core-mantle interaction were measured on the JSC Triton for high-precision 142Nd/144Nd. For comparison, picrites from Hawaii and Iceland that do not show coupled enrichments in 186Os-187Os were measured. We have established an external precision for 142Nd/144Nd of ±3.6 ppm (2σ) during the analytical campaign. The Iceland (n=5) and Hawaiian data (n=9) have ε142Nd ranging from -0.01±0.03 to +0.03±0.03 (2σ) and there is no resolvable difference between samples with Os isotope enrichments versus those without. The average ε142Nd of +0.011±0.034 (2σ) for all of the samples (n=14) is not resolvable from the Ames and La Jolla standards with ε142Nd of 0.000±0.036 (n=16, 2σ). These data confirm recent measurements on lavas for the absence of an ancient enriched 142Nd isotopic signature in plume systems likely derived from D" and indicates that such a reservoir, if existing, must now reside in areas of the deep mantle that plumes do not sample.
V33A-1174
Penguin Bank: A Loa-Trend Hawaiian Volcano
Hawaiian volcanoes along the Hawaiian Ridge from Molokai Island in the northwest to the Big Island in the southeast, define two parallel trends of volcanoes known as the Loa and Kea spatial trends. In general, lavas erupted along these two trends have distinctive geochemical characteristics that have been used to define the spatial distribution of geochemical heterogeneities in the Hawaiian plume (e.g., Abouchami et al., 2005). These geochemical differences are well established for the volcanoes forming the Big Island. The longevity of the Loa- Kea geochemical differences can be assessed by studying East and West Molokai volcanoes and Penguin Bank which form a volcanic ridge perpendicular to the Loa and Kea spatial trends. Previously we showed that East Molokai volcano (~1.5 Ma) is exclusively Kea-like and that West Molokai volcano (~1.8 Ma) includes lavas that are both Loa- and Kea-like (Xu et al., 2005 and 2007).The submarine Penguin Bank (~2.2 Ma), probably an independent volcano constructed west of West Molokai volcano, should be dominantly Loa-like if the systematic Loa and Kea geochemical differences were present at ~2.2 Ma. We have studied 20 samples from Penguin Bank including both submarine and subaerially-erupted lavas recovered by dive and dredging. All lavas are tholeiitic basalt representing shield-stage lavas. Trace element ratios, such as Sr/Nb and Zr/Nb, and isotopic ratios of Sr and Nd clearly are Loa-like. On an εNd-εHf plot, Penguin Bank lavas fall within the field defined by Mauna Loa lavas. Pb isotopic data lie near the Loa-Kea boundary line defined by Abouchami et al. (2005). In conclusion, we find that from NE to SW, i.e., perpendicular to the Loa and Kea spatial trend, there is a shift from Kea-like East Molokai lavas to Loa-like Penguin Bank lavas with the intermediate West Molokai volcano having lavas with both Loa- and Kea-like geochemical features. Therefore, the Loa and Kea geochemical dichotomy exhibited by Big Island volcanoes existed at ~2.2 Ma when the Molokai Island volcanoes formed and has persisted until the present. References: Abouchami et al., 2005 Nature, 434:851-856 Xu et al., 2005 G3, doi: 10.1029/2004GC000830 Xu et al., 2007 G3, doi: 10.1029/2006GC001554
V33A-1175
Geochemical study on the seamounts along the Japan Trench: implications for the origin of HIMU magmas
The Northwestern Pacific plate is characterized by existence of a large number of seamounts that is defined as Western Pacific Seamount Province (WPSP) (Koppers et al., 1998). These seamounts are considered to be produced by large-scale mantle upwelling (superplume) that is now active in the South Pacific region. As the magmatism in Cretaceous is much more active than that in present, research on the past magmatism in South Pacific region could provide a key constraint on the large-scale mantle upwelling. The seamounts that are situated along the Japan trench (Takuyo, Erimo and Ryofu-Daini Seamounts, and Joban Seamount Chain) are old seamounts in WPSP of which ages are Early to Late Cretaceous (Ozima et al. 1970; Takagi et al., 1989; Masalu et al., 1997; Nakanishi and Winterner 1998; Kobayashi et al., 1998), suggesting that these seamount would be suitable examples to study the past superplume activity. However, only a few geochemical studies have been made for the seamounts near the NE Japan (e.g., Izu, 1985). In order to evaluate the origin of seamounts near Japan trench, we collected samples during Geological Survey of Japan (GSJ) Daini-Hakureimaru cruise 05GH. Drilling was conducted at 10 sites, i.e., Smetanin seamounts (4 sites), Joban Seamount chain (6 site) including; Bousei seamount (2 sites), Daisan-Kahisma Northwest seamount (3 site) and Daisan Kashima seamount (1 site). The samples from the Hotta seamountt were also collected by dredging (2 sites). Among these sites, fresh samples were recovered from Bousei, Hotta and Smetanin seamounts. These samples are alkali basalts and andesites that are porphyritic and poorly to moderately vesiculated. The phenocrysts are mainly plagioclase and pyroxene with/without hornblende. The geochemical character of the rocks from the Bosei and Hottat is HIMU-type. However, those from the Smetanin show less HIMU signature. In this presentation, we will discuss the origin of Seamounts along Japan Trench and constrain the origin of HIMU-type magmas
V33A-1176
New Insights Into Petrogenetic Processes At Kula Volcanic Province, Western Turkey
Three episodes of alkaline volcanism of Quaternary age have been recognized in the Kula Volcanic Province (KVP) in Western Turkey. The alkaline volcanic rocks of the KVP vary in composition from basanite to tephriphonolite and from trachybasalt to basaltic trachyandesite. Measured values for 87Sr/86Sr and 143Nd/144Nd in the rocks of the KVP range from 0.703029 to 0.703534 and 0.512773 to 0.512998, respectively, (Gulec, 1991; Alici, et al., 2002) suggesting that an isotopically depleted mantle component is involved in the genesis of the Kula lavas. This mantle component is also enriched in the most incompatible elements, as shown by OIB-like primitive mantle normalized multi-element patterns, this indicates that enrichment of the mantle source is probably a recent event. Generation and evolution of the Kula magmas were simulated successfully by using pMELTS algorithm (Ghiorso, et al., 2002). The fractional melting of an average spinel peridotite mantle was simulated at 23 kbar in the presence of 0.1% H2O at fO2 = QFM-1. The degree of partial melting is 8% for the KVP rocks and the primary magma composition is then subjected to polybaric fractional crystallization from 23 to 4 kbar. The compositional diversity displayed by the Kula volcanic rocks is reproduced successfully by the pMELTS polybaric fractionation calculations when the dP/dT gradient is set to 42 and 25 (bar/° C)in the mantle and crust, respectively, and the oxygen fugacity is increased from one log unit below to one log unit above the QFM buffer at 13 kbar.