V51B-0522

Melting Behaviors of Recycled Subducted Crust in the Mantle: Constraints from Melting Phase Relations of Pyroxenites

* Kogiso, T (kogisot@jamstec.go.jp) , JAMSTEC, 2-15 Natsushima, Yokosuka, 237-0061 Japan
Hirschmann, M M (Marc.M.Hirschmann-1@umn.edu) , Univ. Minnesota, 310 Pillsbury Dr SE, Minneapolis, MN 55455 United States

Ocean island basalts (OIB) have isotopic and trace element signatures that are considered to originate in subducted mafic oceanic crust. Recent experimental studies on mafic lithologies (pyroxenites) revealed that diversity in major element composition of OIB can also be explained by partial melting of a variety of pyroxenites derived from subducted mafic crust. Most part of subducted oceanic crust has MORB-like (silica-excess) composition, but the uppermost part of the subducted oceanic crust suffers extraction of small degree partial melts or siliceous hydrous fluids during subduction, resulting in formation of bimineralic pyroxenite which consists solely of garnet and clinopyroxene. Partial melts from MORB-like pyroxenite at high pressures have silica-saturated compositions, which have similarities to silica-rich tholeiitic OIB [1], whereas partial melts from bimineralic pyroxenite have silica-undersaturated compositions similar to alkalic OIB lavas [2]. On the other hand, subducted oceanic crust is hybridized with surrounding mantle peridotite through mechanical and diffusive interactions with peridotite, yielding olivine-bearing pyroxenite. To investigate melting behaviors of such olivine-bearing pyroxenite, we have conducted partial melting experiments on a mixture of bimineralic eclogite (B-ECL1) with 5 % Fo60 olivine (bulk Mg\# = 61) at 5 GPa. The experimental results show that addition of olivine to B-ECL1 expands the stability of garnet in the residue, resulting in producing liquids richer in SiO$_{2}$ and alkalis and poorer in Al$_{2}$O$_{3}$ than those of B-ECL1. At $1600\deg$C, the degree of partial melting of the B-ECL1 + olivine mixture is higher than that of B-ECL1, suggesting that the solidus temperature is lowered by addition of olivine. However, experiments on a Mg-rich pyroxenite (MIX1G: bulk Mg\# = 79), which has more olivine component, have shown that the solidus of MIX1G is higher than $1600\deg$C [3]. These results suggest that bimineralic pyroxenite can produce larger amounts of partial melts when it is mixed with smaller amounts of olivine. This further indicates that upwelling mantle containing recycled crust with bimineralic composition would begin to melt at the boundary between pyroxenite and peridotite, producing strongly alkalic liquids first. [1] Pertermann and Hirschmann 2003, J.Petrol. 44, 2173. [2] Kogiso and Hirschmann 2002, EOS 83, V71C-12. [3] Kogiso et al. 2003, EPSL 216, 603. Back

V51B-0523

The Role of Garnet Pyroxenite in High-Fe Mantle Melt Generation: High Pressure Melting Experiments

* Tuff, J (jtuf02@esc.cam.ac.uk) , Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge, CB2 3EQ United Kingdom
Takahashi, E (etakahas@geo.titech.ac.jp) , Department of Earth and Planetary Sciences, Tokyo Institute of Technology, Meguro, Tokyo, 152-8551 Japan
Takahashi, E (etakahas@geo.titech.ac.jp) , Institute for Research on Earth Evolution, Jamstec, 2-15 Natsushima-Cho, Yokosuka, 237-0061 Japan
Gibson, S (sally@esc.cam.ac.uk) , Department of Earth Sciences, University of Cambridge, Downing Street, Cambridge, CB2 3EQ United Kingdom

Evidence for the existence of heterogeneous or 'marble cake' convecting mantle$^{1}$ is provided recently by rare, high MgO ($\sim$ 15 wt.%) primitive magmas with anomalously high abundances of FeO* ($\sim$ 13.5 to 16 wt. %$^{2,3}$; where FeO* = total Fe as FeO). These high-Fe mantle melts show a limited occurrence in the initial stage of magmatism in large igneous provinces (e.g. Deccan, Ethiopia and Paran\'{a}-Etendeka) and some have incompatible trace-element and radiogenic-isotopic ratios (Sr, Nd and Pb) that resemble those of ocean-island basalts. This suggests that they are predominantly derived from the convecting mantle$^{2}$. The ferropicrites are mildly- to sub-alkaline and have low contents of Al$_{2}$O$_{3}$ ($<$ 10 wt.%) and heavy rare-earth elements (e.g. Lu $<$ 0.18ppm) that are consistent with the increased stability of garnet, due to the high FeO* content in the ferropicrite mantle source. It has been proposed that the source of the high FeO* may be garnet-pyroxenite streaks derived from subducted mafic oceanic crust$^{2}$. We have undertaken melting experiments between 1 atmosphere and 7 GPa in order to determine the anhydrous phase relations of an uncontaminated ferropicrite lava from the base of the Early-Cretaceous Paran\'{a}-Etendeka continental flood-basalt province. The sample has high contents of MgO ($\sim$ 14.9 wt.%), FeO* (14.9 wt.%) and NiO (0.07 wt.%). Olivine phenocrysts have maximum Fo contents of 85 and are in equilibrium with the host rock, assuming a K$_{d}$ of 0.32 and we believe that the sample is representative of a primary Fe-rich mantle plume derived melt. In total, 75 experimental runs were carried out. Melting phase relations as well as compositions and modal proportions of all coexisting phases were successfully determined in 60 run products. Phase relations indicate that the ferropicrite melt was generated either at $\sim$ 2.2 GPa from an olivine-pyroxene residue or $\sim$ 5 GPa from a garnet-pyroxene residue. A low bulk-rock Al$_{2}$O$_{3}$ content (9 wt.%) and high [Gd/Yb]$_{n}$ ratio (3.1) are consistent with residual garnet in the ferropicrite melt source and favour high-pressure melting of garnet-pyroxenite. The garnet pyroxenite may represent subducted oceanic lithosphere entrained by the upwelling Tristan mantle plume starting-head. During adiabatic decompression, intersection of the garnet pyroxenite solidus at $\sim$ 5 GPa would occur at mantle potential temperatures of $\sim$ 1550$\deg$C. Subsequent melting of peridotite at $\sim$ 4.5 GPa may be restricted by the thick overlying sub-continental lithosphere such that dilution of the garnet-pyroxenite component would be significantly less than in intra-plate oceanic settings. This model accounts for the limited occurrence of ferropicrite magma in the initial stage of continental large igneous provinces and its absence in ocean-island basalt successions. $^{1}$ All\`{e}gre {\it et al}., Philosophical Transactions of the Royal Society of London A297, 447-477 (1980). $^{2}$ Gibson {\it et al}., Earth and Planetary Science Letters 174, 355-374 (2000). $^{3}$ Gibson, Earth and Planetary Science Letters 195, 59-74 (2002). Back

V51B-0524

A Global Dataset of Noble gas Concentrations and Their Isotopic Ratios in Volcanic Areas

* Abedini, A A (aabedini@usgs.gov) , U.S. Geological Survey, 345 Middlefield Rd., Menlo Park, CA 94025 United States
Hurwitz, S (shaulh@usgs.gov) , U.S. Geological Survey, 345 Middlefield Rd., Menlo Park, CA 94025 United States

The extent to which ocean islands are derived from the deep mantle (mantle plumes) or from chemical heterogeneities embedded within the mantle convective flow has long been debated. Noble gases have unique properties that provide significant information regarding this debate and make them important geodynamic tracers. The study of noble gas isotopic compositions in active tectonic and volcanic areas is central to understanding the origins of major volcanic anomalies. For example, helium isotope composition is considered to be the most unambiguous geochemical indicator of a lower mantle plume component in surface rocks, and its variability is often taken as the strongest evidence for a layered mantle. We have compiled all the published global noble gas data from MOR, ocean islands, seamounts, and volcanic arcs, extending existing datasets which are limited to samples from oceanic rocks (Farley and Neroda, 1998; Graham, 2002). Our data set contains information on helium, neon, argon, and xenon concentrations, as well as their isotopic ratios. Where available we also included the isotopic ratios of lead, strontium, neodymium, and carbon. Overall, there are more than 5,000 entries in the database, which is sub-divided both by material sampled (e.g., volcanic glass, different minerals, fumarole, spring) and by tectonic setting (MOR, ocean islands, volcanic arcs). Our extended dataset is consistent with earlier studies and shows that helium isotope ratios in MORB glass in the Pacific, Indian and Southern Atlantic Oceans have a sharp peak between 7 and 9 RA. The large peak in MORB samples correlates with the maxima in samples from volcanic arcs, probably implying that the upper mantle has a uniform helium isotope ratio. In the North Atlantic there is a broader distribution with a maximum at 11 Ra. In contrast, helium isotope ratios in ocean-island basalts (OIB) are highly variable. As noted by previous workers, the helium isotope ratios in OIB have a large peak at 8 RA and a second peak at 13 RA, separated by a pronounced minimum at about 10 RA. In most cases, higher-than-MORB He-isotope ratios coincide with deep mantle plumes revealed by seismic tomography (Montelli et al., 2004). The database will be available through the World Wide Web and will allow examination of some unresolved scientific problems. Farley K.A., Neroda E., Ann. Rev. Earth Planet. Sci., 1998. Graham, D. W., Rev. Mineral. Geochem., 2002. Montelli, R. et al., Science. 2004. Back

V51B-0525

THE CHEMICAL SURFACE SIGNATURE CREATED BY UPWELLING PLUMES

* Schmalzl, J (joergs@uni-muenster.de) , University Muenster Institute of Geophysics, Corrensstr. 24, Muenster, NRW 48149 Germany
Hansen, U (hansen@earth.uni-muenster.de) , University Muenster Institute of Geophysics, Corrensstr. 24, Muenster, NRW 48149 Germany

Convective flows govern much of the dynamics of the Earth. Examples of such flows are convection in the Earth's mantle, convection in magma chambers and much of the dynamics of the world oceans. Nowadays these time-dependent flows are often studied by means of three dimensional (3D) numerical models which solve the equations for the transport of heat and momentum alternatingly. These flows are often driven by a temperature difference. But for many flows there is also an active or passive chemical component that has to be considered such as entrainment of material from an underlying layer. One characteristics of these flows is that the chemical diffusivity is very small. Implementing such a chemical field with a very low diffusivity into a numerical model using a field approach is difficult due to numerical diffusion introduced by the Eulerian schemes. We have implemented a tracer-mesh method which tracks only the position of the interface between the two different components. While the thermal shape of a plume for conditions as appropriate for the Earth's mantle is cylindrical in an first order approximation the shape of the entrained chemical material also has a very strong sheet-like component. This can be important for the understanding of the observed isotopic variability of plumes. Back

http://earth.uni-muenster.de/~joergs

V51B-0526

Limitations on the Estimation of Parental Magma Temperature Using Olivine-melt Equilibria: Hotspots Not So Hot

* Natland, J H (jnatland@rsmas.miami.edu) , RSMAS/MGG University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33149 United States

Estimates of temperatures of magmas parental to picritic tholeiites using olivine-melt equilibria and FeO-MgO relationships depend strongly on the assumption that a liquid composition, usually a glass, is related to the most magnesian olivine in the rock, or to an olivine composition in equilibrium with mantle peridotite, along an olivine-controlled liquid line of descent. The liquid Fe$^{2+}$/Fe$^{3+}$ also has to be known; where data exist, average values from wet chemical determinations are used. Crystallization histories of tholeiitic picrites from islands, spreading ridges, and large igneous provinces, however, usually reveal them to be hybrid rocks that are assembled by two types of magma mixing: 1) between a) differentiated magmas that are on olivine-plagioclase or olivine-plagioclase-clinopyroxene cotectics and b) crystal sludges with abundant olivine that may have accumulated from liquids crystallizing olivine alone; and 2) between primitive magma strains in which olivine crystallized either alone or with other silicate minerals at elevated pressure on separate liquid lines of descent. Many picrites give evidence that both types of mixing have occurred. If either type has occurred, the assumption of olivine-control linking a glass and an olivine composition can only circumstantially be correct. Oxidation state can also be underestimated and therefore FeO contents overestimated if basalts have degassed S, as at Hawaii. In Case 1, hybrid host glass compositions often have higher FeO at given MgO content than liquids which produced many olivine crystals in the rock. In Case 2, the separate parental melt strains are revealed by diversity of compositions of both melt inclusions and Cr-spinel and are most often interpreted to mean local heterogeneity of the mantle source. The inclusions do not always affirm an olivine-controlled liquid line of descent. Instead, inclusions with $<$13% Al$_{2}$O$_{3}$ are increasingly interpreted from both major oxides and trace elements to be derived from melt strains produced by partial melting of both depleted and enriched pyroxenite or recycled ocean-crust (eclogite) (e.g., refs.1 and 2). Some Icelandic picrites also contain large phenocrysts of plagioclase and clinopyroxene; their abundant olivine evidently resulted from mechanical processes of concentration of olivine such as flowage differentiation. Using compositions of low-Al2O3 melt inclusions and host liquids to estimate spinel compositions (ref. 3) reveals many instances of crystallization at higher oxidation states than occur during MORB crystallization, and successfully predicts presence of spinel with Cr/(Cr+Al) = 60-75 actually found in picrites from Hawaii, Iceland, elsewhere in the North Atlantic Igneous Province, and the komatiites of Gorgona, but not in MORB. Where fresh glass is lacking (e.g., Gorgona), bulk-rock compositions have been used to reconstruct conditions of crystallization of parental liquids; but this is greatly complicated by the type and extent of alteration of the rocks. The consequence of all of these factors is that FeO in presumed olivine-controlled liquids is often overestimated, thus many estimated temperatures of crystallization of primitive magnesian liquids are too high by as much as 50-100$^{o}$ absolute, and derived potential temperatures consequently are too high by more than this. (1) Hansteen, T., 1991. Contrib. Mineral. Petrol. 109, 225. (2) Sobolev, A., Hofmann, A., and Nikogosian, I., 2000. Nature, 404, 986. (3) Poustovetov, A., and Roeder, P., 2001, Canad. Mineral. 39, 309. Back

V51B-0527

Two-Stage Melting Of Mantle Plumes And The Origin Of Rejuvenescent Volcanism On Oceanic Islands

* White, W M (white@geology.cornell.edu) , Cornell University, Dept. of Earth and Atmospheric Sciences, Ithaca, NY 14853 United States
Morgan, J P (jpm@geology.cornell.edu) , Cornell University, Dept. of Earth and Atmospheric Sciences, Ithaca, NY 14853 United States

Many mid-plate oceanic volcanoes experience a rejuvenescent, or "post-erosional" phase of volcanism that occurs hundreds of thousands or million years after the main shield-building phase of volcanism has ended. The Hawaiian Islands are the best-documented example, but rejuvenescent volcanism also occurs on the Society Islands, the Marquesas, the Australs, Samoa, and Mauritius. It does not occur on near-ridge islands such as the Galapagos, the Azores, and Iceland. Rejuvenescent lavas have a number of features in common: they are erupted in small volumes, they are highly enriched in incompatible elements, and they are highly alkalic, typically basanitic to nephelenitic. All these features suggest they are quite small degree melts. In addition, rejuvenescent magmas have more depleted isotopic signatures, implying they are melts of more depleted sources, despite their strong incompatible element enrichment. Although isotopic signatures of these lavas are more depleted that those of the corresponding shield stage lavas, they are nevertheless not as depleted as MORB. Furthermore, the isotopic compositions of the rejuvenescent magmas rule out their sources being mixtures of plume material and MORB-source material. Thus geochemical considerations rule out both the lithosphere and the asthenosphere surrounding the plume as the source of rejuvenescent magmas; this implies the plume itself must be the source of rejuvenescent magmas. This conclusion is consistent with geophysical models of plumes. Finite difference numerical models of plume-lithosphere interaction that include both temperature and compositional viscosity dependence reveal that while most melting is concentrated above the hot core of the plume, a melting "tail" extends hundreds of km downstream. In this tail region, lateral spreading of the plume results in a slight rising motion of the plume, and consequently, small extents of melting. The problem thus becomes that of deciphering why melts produced in this tail region are isotopically distinct from those produced in the main melting region. We propose the following model to explain this difference: Mantle plumes are lithologically heterogeneous, consisting of eclogite or pyroxenite "plums" that have a solidus temperature several tens of degrees lower than the more refractory peridotite "pudding" in which they are embedded. Complete isotopic equilibrium is not achieved during melting - either because the plums are large enough ($>$10-100m) or the extraction of plum melts is rapid after their generation. Both the plums and the peridotite are incompatible-element enriched relative to the average depleted upper mantle, but the plums are substantially more enriched. The plums melt entirely in the base of the main melting region and the heat so consumed initially suppresses melting of the peridotite pudding. Plum-derived melts mix as they rise with melts of the peridotite pudding produced higher in the main melting region. This mixture of eclogitic and peridotitic melts form the shield stage magmas. Material in the melting "tail" has had the plums melted out of it in the main melting region. Low degree melting of the plum-free peridotite in the melting tail gives rise to rejuvenescent magmas. Melt production in the tail is more or less continuous, but rejuvenescent volcanism is not. This suggests that some other factor is involved, such as lithospheric loading by adjacent volcanoes, that provides pathways to the surface for small degree tail melts. Back

V51B-0528

How Many Plumes In Africa ? The Geochemical Point of View

PIK, R (rpik@crpg.cnrs-nancy.fr) , CRPG-CNRS, BP20, Vandoeuvre-Les-Nancy, 54501 France
* Marty, B (bmarty@crpg.cnrs-nancy.fr) , CRPG-CNRS, BP20, Vandoeuvre-Les-Nancy, 54501 France
* Marty, B (bmarty@crpg.cnrs-nancy.fr) , ENSG, BP40, Vandoeuvre-Les-Nancy, 54501 France
Hilton, D (drhilton@ucsd.edu) , SCRIPPS, 8675 Discovery Way, La Jolla, CA 92093 United States

Since the Oligocene, volcanic activity in Africa was particularly important in the Horn of Africa where ~ 1 million km3 of continental flood basalts (the Ethiopian CFB) erupted 30 Ma ago in a time interval of 1-2 Ma or less. The Afar volcanic province which is still magmatically active is thought to represent the surface expression of a deep mantle plume, a view consistent with ultra-low velocity anomalies at the base of the mantle beneath the African superswell and the Ethiopia-Afar volcanic province. This plume origin is also supported by the occurrence of 3He/4He ratios up to 20 Ra (Ra is the 3He/4He ratio of atmospheric helium) much higher than those of mid-ocean ridge basalts (on average, 8,b1 Ra) and thought to characterize mantle material originating from below the 660 km discontinuity. However, a deep mantle origin for "high 3He" material is currently questioned by some models which rather ascribe a lithospheric or shallow asthenospheric origin for such He component. The origin of this signal can be tested with the distribution of He isotopic signatures and other geochemical tracers among different African volcanic provinces. All these other provinces exhibit 3He/4He ratios that are equal to, or lower than, the mean MORB ratio of 7-9 Ra (Cameroon line: 5-7 Ra; Hoggar: 8 Ra, this work; Darfur 5.4-7.5 Ra; West African rift: 5-8.5 Ra, this work; Comores, 6.5 Ra, this work). Although low 3He/4He ratios in intraplate volcanic provinces could result from crustal recycling in the mantle and remobilisation of recycled crust during plume uprise, the upper range of 3He/4He values within the field of MORB values points to the strong involvement of asthenospheric mantle and limited interactions of magmas with the aged African crust. Furthermore, these "low-3He" volcanic provinces are characterized by strongly alkaline to undersaturated volcanism indicative of low degrees of partial melting and a thermal regime of the asthenosphere cooler than the one that gave rise to transitional to tholeiitic Ethiopian CFBs. These geochemical observations also conflict with models that advocate channelling of the Afar hotspot material by pre-existing tectonic features to account for all these African volcanic provinces. Back

V51B-0529

Recycling and Mantle Stirring Determined by $^{142}$Nd/144Nd Isotopic Ratios

* Jacobsen, S B (jacobsen@neodymium.harvard.edu) , Harvard University, Department of Earth and Planetary Sciences, Cambridge, MA 02138 United States
Ranen, M C (ranen@fas.harvard.edu) , Harvard University, Department of Earth and Planetary Sciences, Cambridge, MA 02138 United States

It is now well established that $^{146}$Sm was live in the early solar system with an initial uniform $^{146}$Sm/$^{144}$Sm ratio of ~0.008. Harper and Jacobsen (1992) discovered that a sample from Isua (~3.8 Ga old) had a positive $^{142}$Nd/$^{144}$Nd anomaly of 33 ppm when compared to normal terrestrial and chondritic Nd. Furthermore, Jacobsen and Harper (1996) reported results from other Isua as well as Acasta (~4 Ga old) samples. Three other Isua samples had a possible small range (about -15 to +15), while two Acasta samples had no anomalies (normal to within 5 ppm). The presence of $^{142}$Nd anomalies at Isua has recently been confirmed by two other groups (Boyet et al. 2003; Caro et al. 2003). The available data demonstrate both the existence of early depleted mantle and that the early mantle was isotopically heterogeneous. As discussed by Jacobsen and Harper (1996), the recycling rate can be determined by tracing the decay of the average $^{142}$Nd/$^{144}$Nd value of the depleted mantle. In addition, by using the $^{142}$Nd/$^{144}$Nd heterogeneity in the depleted mantle through time we can determine the stirring rate of the mantle (Kellogg, Jacobsen and O'Connell, 2002) as a function of time. For this project our goal is to obtain a resolution in $^{142}$Nd/$^{144}$Nd measurements of ~1 ppm. We have thus compared results obtained for the Nd isotope composition and $^{142}$Nd enriched standards for three different TIMS instruments: The Finnigan MAT 262 at Harvard, the Isoprobe-T and Finnigan TRITON mass spectrometers in GV Instrument's and Thermo Electron's demo laboratories in Manchester and Bremen, respectively. The Finnigan TRITON was designed in response to a request from the senior author for such an instrument. The results obtained so far demonstrate that all three instruments yield the same $^{142}$Nd/$^{144}$Nd, $^{143}$Nd/$^{144}$Nd and $^{145}$Nd/$^{144}$Nd isotopic ratios to within a few ppm, while $^{148}$Nd/$^{144}$Nd and $^{150}$Nd/$^{144}$Nd ratios agree to within 10-20 ppm, when all ratios are normalized to $^{146}$Nd/$^{144}$Nd using the exponential law. Due to the excellent agreement between results from three different instruments we conclude that other reports that claimed that such measurements could not be reproduced at the 5 ppm level for either our 15 year old MAT262 or the newer instruments must be in error. Acknowledgements: We thank GV Instruments and Thermo Electron Corporation for making measurements of our $^{142}$Nd enriched standards. Back

V51B-0530

The Importance of Being Plumes: Entrainment, Isotopes, Melting and the Vertical Structure of the Earth's Mantle

* DePaolo, D J (depaolo@eps.berkeley.edu) , University of California, Dept. of Earth and Planetary Science, Berkeley, CA 94720 United States
Weaver, K L (karrie@eps.berkeley.edu) , University of California, Dept. of Earth and Planetary Science, Berkeley, CA 94720 United States

The standard approach to hot spots has been to associate their geochemical characteristics with the lower mantle. This, along with arguments about material balance, has led to the notion that the composition of the mantle beneath some depth horizon (e.g. 670 km) is different from that above. This particular picture is difficult to reconcile with ideas of whole mantle convection. Plumes, however, are generally imagined as originating from a thermal/compositional boundary layer, and probably at or near the base of the mantle. Thus the fact that plume volcanic rocks are different from, e.g. MORB, may mean only that the base of the mantle is distinct from the rest of the mantle, and that there need be no major compositional discontinuity suspended somewhere in the mid-mantle. This point can be generalized - we should be able to get more information from hot spots with better models and more systematic sampling of lava flows. Theoretically, if plumes originate near the bottom of the mantle, and then entrain some ambient surrounding mantle on the way to the surface, the radial structure of a plume should mimic the vertical structure of the mantle. The full cross-section of the plume however, does not necessarily pass through the melting zone, especially under thick lithosphere, so the lavas represent only a fraction of the plume: the hottest part. In areas where the lithosphere is thin, e.g. where plumes impinge on very thin lithosphere, almost the full cross section of the plume (and hence the full vertical section of the traversed mantle) should be represented in the lavas. We use this concept along with available geodynamic models to compare the observations at Hawaii (thick lithosphere) with Iceland (thin or no lithosphere along the ridge). In Hawaii the highest value of 87Sr/86Sr is about 0.7037 near the plume axis (although it varies somewhat with time). In Iceland the value is not much different at 0.7036. However, in Hawaii the lowest values in the shield stage tholeiites are only slightly lower: 0.7035. In Iceland the values extend continuously all the way down to the local MORB values of 0.7027. In Hawaii, the width of the sampled region is about 100 km, whereas in Iceland the width of the sampled region along the ridge is about 600 km. There is consistency between these patterns that is explained by the effect of lithosphere thickness. Other relationships like this apply for Nd, He, and Pb, and the length scales are not all the same. The details of these distributions, if properly placed into a context of plumes, entrainment and melting beneath the lithosphere, could provide more information about the vertical structure of the mantle, and detailed information about the structure of the base of the mantle. A limiting factor at present is geodynamic models to describe the plumes and melting. Back

V51B-0531

MiFil: a method to characterize seafloor swells with application to the South Central Pacific

* Bonneville, A (bonnevil@ipgp.jussieu.fr) , Institut de Physique du Globe - CNRS, 4, place Jussieu, Paris, 75252 France
Adam, C (adam@ipgp.jussieu.fr) , Institut de Physique du Globe - CNRS, 4, place Jussieu, Paris, 75252 France
Vidal, V (vidal@ipgp.jussieu.fr) , Institut de Physique du Globe - CNRS, 4, place Jussieu, Paris, 75252 France

We propose a new filtering method to characterize large-scale depth anomalies such as seafloor swells associated to intraplate volcanism. Young hotspot volcanoes within plate interiors are frequently surrounded by smooth, broad regions of shallow seafloor termed midplate swells. These swells are typically hundreds of kilometers wide and can be more than a kilometer in elevation. The most frequently invoked explanation for these swells is that they represent the thermal and dynamic surface uplift from rising mantle plumes but they can also be caused by underplating. Wathever their origin be, these swells need to be precisely characterized and we present here a simple method to do the job. This method that we called MiFil, for Minimization and Filtering, requires two stages: a first one to roughly remove the volcano component by minimizing the depth anomaly; a second one to smooth the shape and totally remove the small spatial length scale remaining topography using a median filter. The strength of this method, directly applicable on bathymetry or seafloor depth anomaly grids is that it does not require any assumption on the location, amplitude or width of the large-scale feature to characterize, except its minimal width. Application to hotspot volcanic chains of the South Central Pacific is presented and the results lead to a better understanding of the tectonics and volcanism emplacement of the zone. For each chain, we determine the associated seafloor swell and its main characteristics : (1) the Society is the only 'classical' hotspot that corresponds to the simple interaction of a plume with the lithosphere and for which a buoyancy flux of 1.58$\pm$0.15~Mg~s$^{-1}$ can be obtained; (2) the Marquesas volcanic chain, although quite comparable, presents a swell morphology that prevents such interpretation and quantification; (3) for the Tuamotu, Pitcairn and Cook-Austral volcanic chains, no reliable quantification can be made because the depth and geoid anomalies are caused by several phenomena occurring at different depths that cannot be separated. Back

V51B-0532

On the Zoology of Mantle Upwellings : the case of the Pacific

Davaille, A (davaille@ipgp.jussieu.fr) , Institut de Physique du Globe, 4 Place Jussieu, PARIS cedex 05, 75 252 France
* Bonneville, A (bonnevil@ipgp.jussieu.fr) , Institut de Physique du Globe, 4 Place Jussieu, PARIS cedex 05, 75 252 France
Stutzmann, E (stutz@ipgp.jussieu.fr) , Institut de Physique du Globe, 4 Place Jussieu, PARIS cedex 05, 75 252 France

The characteristics of intraplate volcanism in the Pacific are very diverse. There is probably more than $10^6$ seamounts on the Pacific seafloor created by past or present volcanism. Among them, a number are organized in alignments, some of which showing increasing ages along the track in the direction of plate motion. Sorting the latter by their track duration further shows two populations : a big peak around 20 Myr and a few tracks longer than 50 Myr, sometimes originating at an oceanic basaltic plateau. The Pacific presents also a number of isolated oceanic plateaus, as well as the Polynesian superswell, a region of anomalously shallow sea-floor several thousands of kms in extent with an unusually dense concentration of alignments. Tomographic images of the Pacific mantle reveal several 3D structures of slow velocities, presumably indicating hot material. Moreover, paleo-reconstructions show replication of volcanism over certain areas on a 100 Myr time-scale. In a heterogeneous viscous fluid like the mantle, several kinds of upwellings may develop, from the classical, mushroom-shaped, thermal plume to more complicated thermo-chemical structures. Fluid Mechanics studies give definite constraints on the necessary conditions for these upwellings existence, characteristics (spacing, recurrence time, temperature anomaly,.), and ability to reach the lithosphere. We therefore use those constraints to identify : 1) what types of upwellings could develop in the Earth's mantle, 2) what observations, if any, CAN be explained by a mantle upwelling, 3) what observations can NOT be explained by a mantle upwelling. Back

V51B-0533

Large Igneous Provinces, Mantle Plumes and Uplift: A Sedimentological perspective

Mazumder, R (mrajat2003@yahoo.com) , Asutosh College, Dept. Geology, Kolkata, 700026 India
* Foulger, G R (g.r.foulger@durham.ac.uk) , University of Durham, Science Laboratories, South Rd., Durham, Durham United Kingdom

Significant pre-volcanic uplift of the lithosphere is one of the expected consequences of mantle plume upwelling. Many major mafic/ultramafic lavas have been attributed to mantle plumes that are expected to produce crustal uplift (doming) preceding the major phase of volcanism. Such pre-volcanic uplift would have significant consequences on the regional sedimentation pattern. Subsequent erosion may remove much of the volcano-sedimentary record of domal uplift. However, sedimentologists can identify progressive shallowing of palaeogeography prior to volcanism and distinctive palaeocurrent patterns if these exist, and this may provide constraints on plume interpretations of the volcanic episode. It is difficult to defend a LIP-mantle plume connection where pre-volcanic lithospheric uplift is absent. This is particularly true for continental flood basalts. The sedimentological and stratigraphic criteria proposed for tracing LIP-plume connections are somewhat generalized, and will not all be useful at a given location. For example, progressive shallowing of palaeogeography as a consequence of plume-induced lithospheric uplift should be clear in a marine depositional setting (e.g., in the transition from deep to shallow marine, or marine to terrestrial palaeogeography). In a continental depositional setting, such evidence is less likely to be found, however. This is because the depositional surface is already well above mean sea level (the base level of erosion) and its preservation potential is small. Unlike the case of marine depositional settings, an erosional unconformity/sequence boundary will thus not develop because the depositional surface is already subaerially exposed. The claim that the sedimentary record provides independent supporting evidence for mantle plume influence on the generation of LIPs, is not always true. The genetic linkage between CFBs and mantle plumes is at best difficult to establish from sedimentological analysis alone. Back

http://www.mantleplumes.org

V51B-0534

Heat Flow on Hotspot Swells Reflect Fluid Flow Masking Potential Variations in Mantle Heat Flux

* Harris, R N (rnharris@mines.utah.edu) , University of Utah, Dept. of Geology and Geophysics, 135 S 1460 E, Rm 719, Salt Lake City, UT 84112-0111 United States
McNutt, M K (marcia@mbari.org) , Monterey Bay Aquarium Research Institute, 7700 Sandholdt Road, Moss Landing, CA 95039 United States

The origin of hotspot swells is poorly understood. Heat flow data collected on hotspot swells have been used to argue for and against sublithospheric thermal anomalies. The presence of sublithospheric thermal anomalies has been inferred from interpretations of anomalously high heat flow determinations, whereas the contention that hotspot swells result from normal melting processes within the lithosphere is based on `normal' heat flow values. These arguments depend in part on the choice of a thermal reference model, but more importantly assume conductive heat transfer through the lithosphere. We provide evidence that heat flow measurements collected on hotspot swells likely reflect shallow fluid flow rather than deep thermal variations within or at the base of the lithosphere. Discriminating between environments where heat is transferred conductively or convectively requires closely spaced heat flow determinations (1-2 km) collocated with seismic reflection profiles. Only Hawaii and Reunion have surveys meeting these requirements. The Hawaiian survey consists of two profiles, one north of Oahu and one north of Maro Reef. The Reunion survey also consists of two profiles, both north of Mauritius. These heat flow profiles reveal greater scatter than anticipated with spectral peaks on the order of 10 km consistent with fluid flow. Root mean square variations along the Oahu and Maro Reef profiles are 15 and 5 mW m$^{-2}$, respectively, and along both Reunion profiles are about13 mW m$^{-2}$. Coupled heat and fluid flow models demonstrate that thermal buoyancy due to bathymetric relief is capable of driving significant fluid flow that may suppress the background thermal field. These models are consistent with heat flow patterns observed at individual seamounts and oceanic basement highs that are more easily sampled and characterized than large hotspot swells. We caution that the ability of fluid flow to mask variations in sublithospheric heat flux make surface heat-flow values a poor discriminator between geodynamic models for hotspot swells. Back

V51B-0535

Cratonal lithosphere a potential recorder of ancient mantle plumes

* Sleep, N H (norm@pangea.stanford.edu) , Department of Geophysics, Stanford University, Stanford, CA 94305 United States

In the conventional mantle plume hypothesis, hot plume material ponds at the base of the lithosphere. The deep cratonal lithosphere provides a record and a potential test of this aspect of plumes. First uplift and subsidence in the geological record give lthe more likely times of plume impingement. Second studies of pressure-temperature-time histories from zoned diamonds may make this process observable. I make analytical and numerical calculations relevant the the thermal history of the deep lithosphere. Physically stagnant-lid convection subsequent to the ponding of plume material heats the overlying lithosphere and cools the plume material. It thins the rheological boundary layer of ordinary mantle between overlying chemically buoyant cratonal lithosphere and the underlying plume material. Convection moves the chemically buoyant material producing a potentially observable pressure change. Loitering plume tails have a stronger effect than plume head episodes. They can occur only at times that the plate moves slowly, which provides another potential test. The xenolith geotherm provides a further constraints. If a thick layer of plume material is present, the scaling relationship for the thickness of the rheological boundary layer beneath a convecting lid is inverse to the viscosity of the underlying hot material to the fourth power at a given heat flow. Back

V51B-0536

Diagnostic Features of Mantle Plume Penetration Into the Lithosphere.

* Jurine, D (jurine@ipgp.jussieu.fr) , Institut de Physique du Globe de Paris, 4, place Jussieu, Paris, 75005 France
Jaupart, C (cj@ccr.jussieu.fr) , Institut de Physique du Globe de Paris, 4, place Jussieu, Paris, 75005 France
Brandeis, G (brandeis@ipgp.jussieu.fr) , Institut de Physique du Globe de Paris, 4, place Jussieu, Paris, 75005 France

The success or failure of the mantle plume hypothesis rests in part on its predictions for the manner of plume penetration through the lithosphere. We study how thermal plumes interact with viscous and buoyant boundary layers which serve as analogs for oceanic and continental lithospheres of various ages. Laboratory experiments and numerical simulations have been performed over a large range of values of plume buoyancy flux, lithosphere density, viscosity and thickness. Plume penetration proceeds in two distinct phases. In an initial phase, the lithosphere thins and is heated by the plume. The extent of plume penetration depends on lithosphere viscosity and plume buoyancy, and is not sensitive to plume dimensions. In a second phase, heated lithosphere becomes unstable and generates a secondary upwelling of larger dimensions than the plume. The shape and dimensions of the intrusion into the lithosphere vary significantly as a function of the control parameters. Scaling laws have been derived for the initial penetration velocity and for the critical time for lithosphere instability. Application to the Earth leads to a critical time equal to a few tens of My. The two penetration phases are associated with different styles and rates of upwelling, and hence different melting rates and magma compositions involving mixtures of asthenospheric and lithospheric melts. Back

V51B-0537

Revised Models of Global Plate Motions and Mantle Flow Successfully Predict the Emperor-Hawaii and Other Hotspot-related Seamount Chains

* Sutherland, R (r.sutherland@gns.cri.nz) , Institute of Geological and Nuclear Sciences (GNS), PO Box 30-368, Lower Hutt, Wellington New Zealand
Steinberger, B (bernhard.steinberger@ngu.no) , Institute for Frontier Research on Earth Evolution (IFREE), Japan Marine Science & Technology Center (JAMSTEC) 2-15 Natsushima-cho, Yokosuka, 237-0061 Japan
O'Connell, R J (oconnell@geophysics.harvard.edu) , Department of Earth and Planetary Sciences, Harvard University, 20 Oxford Street, Cambridge, MA 02138 United States

The bend in the Hawaiian-Emperor chain is a prominent feature usually attributed to a change in Pacific plate motion. However, global plate motion chains fail to predict that bend. We show how the geometry of the Hawaiian-Emperor chain and other hotspot tracks can be explained by a combination of hotspot motion and intraplate deformation. Global mantle flow models predict a southward motion of the Hawaiian hotspot. This, in combination with a plate motion chain connecting Pacific and African plates through Antarctica, predicts the Hawaiian track correctly since the bend, but too far west prior to it. If a chain through Australia - Lord Howe Rise is used instead, the track is predicted correctly back to 65 Ma. The difference between the two predictions indicates the effect of intraplate deformation not yet recognized or else not recorded on the ocean floor. The remaining misfit prior to 65 Ma can be attributed to additional intraplate deformation of similar magnitude. Back

V51B-0538

Intermittent Volcanism in the S Pacific: Tracking Persistent Geochemical Sources

* Konter, J G (jkonter@ucsd.edu) , SIO-UCSD, 9500 Gilman Dr, La Jolla, CA 92093-0225 United States
Koppers, A A (akoppers@ucsd.edu) , SIO-UCSD, 9500 Gilman Dr, La Jolla, CA 92093-0225 United States
Staudigel, H (hstaudigel@ucsd.edu) , SIO-UCSD, 9500 Gilman Dr, La Jolla, CA 92093-0225 United States
Hanan, B B (bhanan@sdsu.edu) , SDSU, 5500 Campanile Dr, San Diego, CA 92182 United States
Blichert-Toft, J (jblicher@ens-lyon.fr) , ENS, 46 allee dItalie, Lyon, 7 France

South Pacific ocean intraplate volcanoes (OIV's) have formed relatively short, discontinuous chains over the last 140 Ma, in contrast to classic continuous hot spot chains like Hawaii or Louisville. Moreover, its volcanism stands apart by very diverse and radiogenic mantle source regions, defining the South Pacific Isotopic and Thermal Anomaly (SOPITA). Discontinuous SOPITA OIV's form a complex array of crossing lineaments and isolated seamounts that can be back-tracked in time to the western Pacific. We studied the Hf and Pb isotopic composition of the prominent western Gilbert's and Tokelau chain that mostly correspond to the plate motion stage expressed by the Emperor Seamounts. The Gilbert's chain (to the West of Tokelau) has an age range of 74-65 Ma, and back-tracks to an OIV source origin near Rurutu. Notably, this chain shows an isotopic signature similar to the Rurutu chain, from Burtaritari ($^{176}$Hf/$^{177}$Hf: 0.282900, $^{206}$Pb/$^{204}$Pb: 20.582) to Tuba (0.282938, 19.691). The parallel-running Tokelau chain ranges from 76-58 Ma, projecting current activity around Macdonald seamount. Its isotopic composition is also similar to Macdonald seamount, as shown by Howland (0.282951, 20.773) to Polo (0.283050, 19.390). However, both chains also contain seamounts with ages and isotopic signatures that are outliers, inconsistent with their general age range and isotopic trends (e.g. Ava; 0.282641, 17.729 and Sakau; 0.282880, 18.997). Our results show that the discontinuous magmatic provinces within SOPITA can be tracked back in time by geochronology and geochemical fingerprinting. This suggests that the SOPITA OIV's are fed by distinct, long-lived mantle source regions that may turn "on" or "off" for distinct periods during their life-times. This intermittent volcanism is inconsistent with the "standard" fixed hot spot model, since it does not produce continuous chains with linear age progressions. Furthermore, our results demonstrate the utility of geochemical fingerprinting for relating the products of discontinuous mantle melting anomalies. Such combined evidence for long-lived but discontinuous OIV lineaments allows for their use in plate and/or hot spot motion models, and might offer a new understanding of the processes responsible for the formation of OIV's. Back

V51B-0539

Improved Absolute Plate Motion Modeling in the Pacific

* Wessel, P (pwessel@hawaii.edu) , School of Ocean and Earth Science and Technology, University of Hawaii at Manoa, 1680 East-West Rd, Honolulu, HI 96822 United States
Harada, Y (harada@scc.u-tokai.ac.jp) , School of Marine Science and Technology, Tokai University, 3-20-1 Orido Shimizu, Shizuoka, 424-8610 Japan
Kroenke, L W (kroenke@soest.hawaii.edu) , School of Ocean and Earth Science and Technology, University of Hawaii at Manoa, 1680 East-West Rd, Honolulu, HI 96822 United States

In studies of Relative Plate Motion (RPM), the model constraints are conjugate magnetic isochrons identified in marine magnetic anomalies. The model is a finite rotation that rotates an isochron on plate A such that the rotated segment matches the conjugate isochron on plate B. Chang (1987; 1988) solved for such rotations using nonlinear spherical regression and developed statistical confidence regions for the resulting rotations. Because conjugate data can be optimally superimposed using a single, finite rotation it was natural to define the model in terms of total reconstruction rotations. In studies of Absolute Plate Motion (APM), the constraints are the surface expressions of hotspot seamount chains and their measured ages. The traditional approach is to model coeval segments of seamount chains as small circles about stage poles of rotation found by minimizing the distances from each seamount to its locally best-fitting, small circle about a candidate pole. The opening angles are typically found by trial and error. Given the age range of a particular set of copolar segments, opening rates can be determined. Because the data portray small circles, it was natural to define the model in terms of stage rotations. The traditional APM modelling approach has many limitations, including (1) shorter segments, possibly reflecting APM changes, are difficult to identify and correlate across several chains; (2) short small-circle segments become indistinguishable from great circles and hence reliable poles cannot be determined; (3) without easily identifiable kinks between chain segments, ages are needed to make the correlation and these are often lacking; and (4) unlike RPM modelling, no rigorous approach for estimating APM uncertainties exists. However, Wessel and Kroenke (1997) developed a method to derive optimal hotspot locations from seamount data if the APM is known, whereas Harada and Hamano (2000) introduced a technique to determine total reconstruction rotations if hotspot locations are known. We improve the modelling of APM by combining these two complimentary methods into a self-consistent hybrid technique. The hybrid technique allows us to determine (1) the best location for hotspots, (2) a high-resolution APM model, and (3) covariance matrices for each rotation. We present the first self-consistent Pacific APM with confidence regions for each rotation pole and reconstructed points. The new model is contrasted with traditional models, and the implications of the model for drift within the Pacific hotspot group and the origin of the Hawaii-Emperor bend is addressed. Back

V51B-0540

Paleomagnetic Tests of Global Plate Reconstructions with Fixed and Moving Hotspots

* Andrews, D L (dandrews@jhu.edu) , Department of Earth Science, Rice University, 6100 Main St., Houston, TX 77005 United States
Gordon, R G (rgg@esci.rice.edu) , Department of Earth Science, Rice University, 6100 Main St., Houston, TX 77005 United States
Horner-Johnson, B C (ben@esci.rice.edu) , Department of Earth Science, Rice University, 6100 Main St., Houston, TX 77005 United States

Three distinct approaches have been used in prior work to estimate the motion of the Pacific basin plates relative to the surrounding continents. The first approach is to use the global plate motion circuit through Antarctica (e.g., the Pacific plate to the Antarctic plate to the African plate to the North American plate). An update to this approach is to incorporate the modest mid-Tertiary motion between East and West Antarctica estimated by Cande et al. (2000). A recently proposed second approach is to take an alternative circuit for the early Tertiary of the Pacific plate to the Australian plate to the East Antarctic plate to the African plate to the North American plate (Steinberger et al. 2004). The third approach is to assume that the hotspots in the Pacific Ocean are fixed relative to those in the Atlantic and Indian Oceans (e.g., Engebretson et al., 1986), which we recently showed indicates motion between East and West Antarctica of 800 $\pm$ 500 km near the Ross Sea Embayment. The first approach (global plate motion circuit through Antarctica) indicates very rapid motion between Pacific and Indo-Atlantic hotspots during the early Tertiary (e.g., Raymond et al. 2000). The second approach (global plate motion circuit through Australia) indicates slower, but still substantial, motion between Pacific and Indo-Atlantic hotspots (Steinberger et al. 2004). Because each of the three approaches predicts distinctly different motion between the Pacific plate and the continental plates, they can be tested with paleomagnetic data. The results of such tests indicate that the first approach leads to systematic and significant misfits between Pacific and non-Pacific early Tertiary and Late Cretaceous paleomagnetic poles. The second approach leads to slightly smaller misfits. In contrast, the circuit based on fixed hotspots brings the Pacific and non-Pacific paleomagnetic poles into consistency. Thus the paleomagnetic data decisively favor fixed hotspots over the alternative approaches and suggests that motion between hotspots is substantially less than inferred by Steinberger et al. (2004). Back

V51B-0541

Latitudinal Shift of the Hawaiian Hotspot: Motion Relative to Other Hotspots or True Polar Wander?

* Gordon, R G (rgg@rice.edu) , Rice University, 6100 Main St. Earth Science--MS 126, Houston, TX 77005 United States
Horner-Johnson, B C (ben@esci.rice.edu) , Rice University, 6100 Main St. Earth Science--MS 126, Houston, TX 77005 United States

Recent results from deep sea drilling confirm a large southward drift of the Hawaiian hotspot since Campanian and Maastrichtian time (ca. 70 to 83 Ma), as was previously found from prior paleomagnetic results from drilling (Kono, 1980; Jackson et al. 1980), from skewness analysis of Pacific magnetic anomalies (Gordon 1982, Petronotis & Gordon 1989, 1999; Petronotis et al. 1992, 1994; Acton & Gordon, 1991; Vasas et al. 1994; Horner-Johnson & Gordon 2003), and from other paleomagnetic and paleolatitude data (Gordon & Cape 1981; Sager & Bleil 1987). This southward drift could have been the result of motion of the Hawaiian hotspot relative to some other hotspots, or of true polar wander, or of both. Tarduno et al. (2003) have recently presented an extreme interpretation of these results as being entirely due to southward motion of the Hawaiian hotspot through the mantle. Here we show that this extreme interpretation is not supported by available data. While the Pacific plate paleomagnetic data are sufficient to show that the Hawaiian hotspot has moved southward relative to the spin axis, alone they cannot be used to demonstrate motion relative to the mantle or relative to other hotspots. To do so, coeval paleomagnetic poles are needed from the continents bordering the Atlantic and Indian Oceans. Here we show that few, if any, of the coeval paleomagnetic poles from the continents incorporated into widely used reference paths pass minimum reliability criteria. Thus, the inference of rapid motion of the Hawaiian hotspot relative to the mantle is surely premature and probably incorrect. We further show that other earlier studies purporting to show motion between hotspots from paleomagnetic data are now invalid because of revisions to paleomagnetic poles from the continents or because of flaws in analysis. Updated paleomagnetic analyses indicate that little motion has occurred between Pacific hotspots and non-Pacific hotspots. Instead, available data are consistent with the hypothesis that the southward motion of the Hawaiian hotspot relative to the spin axis is mainly caused by true polar wander. Back

V51B-0542

Fifty Million Years of Fixed Hotspots: A New Self-Consistent Global Set of Plate Rotations

* Kumar, R R (rkumar@rice.edu) , Rice University, 6100 Main St. Earth Science--MS 126, Houston, TX 77005 United States
Andrews, D L (dandrews@jhu.edu) , Rice University, 6100 Main St. Earth Science--MS 126, Houston, TX 77005 United States
Gordon, R G (rgg@rice.edu) , Rice University, 6100 Main St. Earth Science--MS 126, Houston, TX 77005 United States

We have developed new methods for appropriately estimating the uncertainties in plate rotations relative to hotspots. Using these new methods and the latest available relative plate rotations, and also incorporating uncertainties in relative plate motions, we showed that there is no significant motion between Pacific hotspots and Indo-Atlantic hotspots for the past ca. 50 Myr. Here we take the next step and seek a sequence of rotations based on simultaneous inversion of hotspot track data in the Pacific, Atlantic, and Indian Oceans. The resulting set of reconstructions are thus optimal estimates for global plate motion relative to the hotspots for the past 50 Myr. Evidence for significant changes in plate motion, including changes in pole of rotation and in rates of rotation, will be presented. Implications for true polar wander will be discussed. Back

V51B-0543

Is the Hawaiian-Emperor Bend Coeval for all Pacific Seamount Trails?

* Koppers, A A (akoppers@ucsd.edu) , IGPP, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093-0225 United States
Staudigel, H (hstaudig@ucsd.edu) , IGPP, Scripps Institution of Oceanography, University of California, San Diego, La Jolla, CA 92093-0225 United States

By far the largest number of hotspots can be found in the South Pacific Thermal and Isotopic Anomaly (SOPITA). Its Cretaceous counterpart is preserved in a large range of seamounts and guyots found in the West Pacific Seamount Province (WPSP). The seamounts in these regions display very distinct and long-lived isotopic signatures (Staudigel et al., 1991; Koppers et al., 2003) that can be used to combine source region chemistry and seamount geochronology to map out mantle melting anomalies over geological time. These mappings may resolve many important questions regarding the stationary character, continuity and longevity of the hotspots in the South Pacific mantle. Most importantly, it may also answer the question whether the Hawaiian-Emperor Bend (HEB) is coeval for all Pacific Seamount trails at 47 Ma? Fixed hotspots should be expressed in volcanic trails on the lithospheric plates revealing absolute rates of motion from their age progressions and the direction of motion based on their azimuths. By definition, bends in these hotspot trails thus should give an indication of changing plate motion happening simultaneously across individual lithospheric plates. Based on the morphology of seamounts in the Pacific, the Hawaiian-Emperor, Louisville, Gilbert Ridge and Tokelau seamount trails may be identified as the only hotspot trails to exhibit a clear HEB-type bend (Kroenke et al. 2004). Of these, the Louisville seamount trail only displays a faint bend that may be coeval with the sharp 60 degree bend in the Hawaiian-Emperor trail (Koppers et al. 2004) at 47 Ma. However, new 40Ar/39Ar analyses indicate that the HEB-type bends in the Gilberts Ridge and Tokelau seamount trails are asynchronous around 67 Ma and 57 Ma, respectively. We argue, therefore, that plate motion alone cannot explain these age systematics, but that both hotspot motion and changing lithospheric stress regimes may play an important role in their creation. The simple and elegant hotspot model that (almost without difficulty) may explain primary hotspots such as Hawaii and Louisville, seems unsatisfactory to explain the age distributions of the short-lived Gilbert Ridge and Tokelau hotspots. To explain intra-plate volcanism in the South Pacific, we argue for a combination of processes: one that forces regional magmatism from a large-scale source of buoyancy from below (like the rise of plumelets shooting off the top of a superplume that die-off after a short life-cycle) and one process that acts from above, as intra-plate extension opens up pathways that allow the lithosphere to be penetrated by magma. Back

http://earthref.org/databases/SC/

V51B-0544

Critical Evaluation of Radiometric Ages Used for Tracking Hotspots in the Pacific Ocean

* Baksi, A K (abaksi@geol.lsu.edu) , Department of Geology, Louisiana State University, Baton Rouge, LA 70803 United States

One of the pillars supporting the plume hypothesis, is the progression of ages for numerous hotspot tracks in oceans. These ages should be based on radiometric measurements. The argon dating methods have been the tool most commonly used. Since most of the rocks selected for dating have suffered (considerable) alteration, K-Ar dates should not be used as accurate measures of the age of crystallization. 40Ar/39Ar total fusion ages, though better than K-Ar dates in general, do not pinpoint samples that (a) contain excess argon or (b) have suffered partial loss of 40Ar* due to alteration. Hence 40Ar/39Ar incremental heating studies remain as the (only) tool of choice. From such experiments, at a minimum, ages must be based on plateau and/or isochron sections that meet the necessary statistical requirements to be considered crystallization ages. Earlier (Baksi, 1999, Jour. Geol.) it has been shown that almost all the purported crystallization ages for hotspot tracks in the Atlantic and Indian Oceans, are invalid (see also www.mantleplumes.org/ArAr.html). Herein, I apply the tests outlined therein, to evaluate ages available in the literature for hotspot tracks in the Pacific Ocean. These can be divided into five main groups. (1) Those with reliable age data (e.g. Dalrymple and Garcia,1980; Dalrymple et al., 1980, DSDP 55, Hawaiian-Emperor Chain); the authors use care in selecting valid ages from their data sets. (2) Others (e.g. Pringle, 1993, AGU Monograph 77, Musicians Seamounts), most ages are statistically valid, though some fail the requisite test. In addition, many samples show high levels of atmospheric argon, suggesting the samples are (quite) altered; this could lead to incorrect plateau ages. (3) The next set (e.g. Winterer et al., 1993, AGU Monograph 77, Cretaceous guyots in the Northwest Pacific; Ozima et al., 1977, JGRAS, Western Pacific guyots; Saito and Ozima, 1977, EPSL, Western Pacific area) have few, if any, valid ages. Most plateaux/isochrons clearly fail the statistical test of reliability; many steps show high levels of atmospheric argon - the samples are (badly) altered. (4) A set of papers (e.g. Gripp and Gordon, 2002, Geophys. J. Int., young hotspot tracks; Duncan, 1985 - New Hebrides-Samoa lineament) make use of K-Ar dates, wholly or in the main. These dates should be treated as minimum estimates of the crystallization age. (5) A final set of papers (Sager et al., 1993, AGU Monograph 77, Japanese and Marcus-Wake Seamounts; Lincoln et al., 1993, AGU Monograph 77, Marshall Islands), quote ages without listing the relevant analytical data. These results are to be treated as suspect, and not used for quantitative tracking of hotspot trails. In conclusion, the purported progression of ages for numerous hotspot tracks in the Pacific Ocean does not withstand critical scrutiny. Back

V51B-0545

40Ar/39Ar Geochronology of the Sylhet Traps, Eastern India, and their relationship to the Kerguelen Plume related magmatism

* Ray, J S (jsray@prl.ernet.in) , Physical Research Laboratory, Navrangpura, Ahmedabad, 380009 India
Pande, K (kanchanpande@iitb.ac.in) , Physical Research Laboratory, Navrangpura, Ahmedabad, 380009 India

It is now an accepted view that the Rajmahal-Bengal flood basalts of eastern India belong to the Kerguelen plume generated Large Igneous Province (LIP) that encompasses the Bunbury-Naturaliste Plateau basalts of western Australia, volcanism on the southern and central Kerguelen Plateau, Elan Blank, and Broken Ridge in the Indian Ocean (e.g., Kent et al., J. Petrol., 43, 2002). The lesser-known Sylhet Traps (25.5$^{o}$ N, 91.8$^{o}$ E), exposed ${\sim}$400 km east of the Rajmahal Traps, are assumed to be a part of the ${\sim}$118 Ma old Rajmahal-Bengal flood basalt province (e.g., Bakshi, Chem. Geol., 121, 1995). However, absence of physical continuity between these volcanic activities, and more importantly lack of precise age data always created difficulty in correlations. In an attempt to test the hypothesis of their cogenesis we dated the Sylhet Traps using $^{40}$Ar/$^{39}$Ar incremental-heating technique. Two whole rock samples yielded good plateau ages: 115.9{$\pm$}4.1 (2{$\sigma$}) Ma, and 115.5{$\pm$}5.4 (2{$\sigma$}) Ma, calibrated against an age of 523.2{$\pm$}0.9 Ma for the Minnesota Hornblende (MMhb-1; Spell and McDougall, Chem. Geol., 198, 2003). The concordant plateau, isochron and inverse isochron ages, the atmospheric value for the trapped $^{40}$Ar/$^{36}$Ar component, and good MSWD values for the isochrons suggest that 115.8{$\pm$}3.2 Ma (the weighted mean of both the ages) can be considered as the age of crystallization. This age for the Sylhet Traps clearly falls within the range of ages reported for various groups of lavas in the Rajmahal-Bengal province (118-115), and is consistent with the idea that magmatism in this part of India continued well beyond the major pulse at ${\sim}$118 Ma (e.g., Kent et al., J. Petrol., 43, 2002). It also supports the proposal based on geochronology and geochemistry that the Kerguelen hotspot was located close to the eastern Indian margin during its peak activity (e.g., Kumar et al., Geophys. Res. Lett., 30, 2003). Back

V51B-0546

High precision Pb, Sr, and Nd isotope geochemistry of alkalic early Kilauea magmas from the submarine Hilina bench region, and the nature