V21B-0600 0800h
Can kimberlites result from melting of ``normal'' mantle?
The extreme enrichments of kimberlitic magmas in incompatible elements (e.g., compared to a typical undepleted mantle) are usually attributed to low degrees of melting and/or metasomatized source compositions. We present a model explaining the observed enrichment of kimberlitic magmas with rare earth elements (REE) in terms of melt migration through source rocks having the composition of ``normal'', primitive mantle. The model considers chemical evolution of a batch of melt that interacts with the solid matrix along its propagation path. This chromatographic process leads to abundances of trace elements that approach $C_0/D$ asymptotically, where $C_0$ is the concentration of a trace element in the primitive mantle and $D$ is the partition coefficient for this element between primitive mantle and the kimberlitic melt. This final, ``fully saturated'' melt composition is independent of the initial composition of the melt input and of the melt fraction. The fully saturated REE spectra that would result from this process are in excellent agreement with the maximum observed REE enrichments of kimberlites from South Africa, Siberia and India. This result may reconcile relatively high degrees of melting (i.e., $\sim$1 volume %) thought to be required for efficient melt segregation from the matrix with the extreme REE enrichments of kimberlitic magmas. This chromatographic process can also explain the low La/Yb ratios typical of many kimberlitic magmas if the length of the chromatographic column is shorter than that required for a full saturation of the most incompatible elements. The observed low La/Yb ratios may be consistent with porosities in excess of $\sim$1% since lower melting degrees should result in full saturation and high La/Yb. An important feature of our model is that the resulting saturated REE spectrum is practically independent of a source mineral composition, which may explain the observed similarity of the REE spectra of kimberlites from different geographic localities.
V21B-0601 0800h
Peculiar Feldspar And Quartz Inclusions Within Zircons From Anorthosites, North Eastern Desert, Egypt
Zircons from three anorthosite outcrops along Wadi Dib area, north Eastern Desert of Egypt contain abundant and conspicuous inclusions of quartz, feldspar, amphibole and apatite. These anorthosites, as (50-100m thick) layers, represent the top of mafic-ultramafic intrusions exhibiting rhythmic layering visible by reputation of melanocratic and leucocratic layers. Field and microscopic studies exhibit that these anorthosites were affected by the action of residual magmatic solutions associated with the late stage crystallization of the younger granites, which modified their mineralogical composition. They are composed totally of plagioclase with subordinate amount of clinoenstatite, augite, amphibole, biotite, K-feldspar, and quartz. Accessories are magnetite, ilmenite, apatite and zircon. The abundance and the mode of occurrence of K-feldspar, quartz, and biotite with apatite and zircon among the megacrysts suggest their formation is ascribed to the interaction with the residual solutions. The microprobe data exhibit difference between feldspar and amphiboles contained herein zircons and those as anorthosite mineral constituents. The genetic relationship between zircons and their inclusions suggests later growth of zircons than inclusions and most probably at the final stage of rock modification. Zircons are magmatic and found in the interstitial feldspar and quartz among plagioclase megacrysts in aggregates or as individual grains. The microscopic and SEM images investigation exhibit that most zircons are subhedral to euhedral equant and prismatic crystals. Most zircons have same range of crystal morphologies and internal growth structures with predominance of prism /{100/} and pyramid /{101/} and occasionally prism /{110/} and pyramid /{111/}. No evidences for poly-faceted grains, inherited cores or later overgrowths were detected. CL images distinguished zircons with visible core-rim structures and others with regular and continuous growth zones contained herein various inclusions. The dark CL cores in the core-rim structured zircons are higher in U, Y and sometimes Hf relative to the CL bright rims. Microprobe data and x-ray chemical mapping of various zoned zircons suggest that U and Y with sometimes Hf have a negative correlation to the CL brightness, while Th doesn't exhibit any significant correlation.
V21B-0602 0800h
Making metamorphic monazite - major silicates are not major players
Monazite is commonly assumed to derive its LREEs from the breakdown of LREE-rich accessory phases such as allanite. However, it is also possible that significant LREEs are scavenged from the major silicates. In this study, we tested the latter hypothesis via high-precision ($\pm$3-5 ppb), {\it in situ} LA-ICP-MS analysis of LREE zoning patterns and concentrations in all major silicates from garnet and amphibolite facies pelitic rocks of the Great Smoky Mountains, North Carolina. Previous research on schists from the Smokies suggests that monazite formed during the staurolite-in reaction (Kohn and Malloy, 2004, Geochim. Cosmochim. Acta). Garnets from below the staurolite- and monazite-in isograds contain 100s of ppb LREEs in the core, with a gradual decrease rimward to a concentration that is about one half that in the core. Plagioclase from garnet-zone rocks shows an increase in LREE contents from 1-2 ppm in the core to about 4 ppm at the rim. Above staurolite-in and monazite-in, LREEs in garnet typically show core-rim zoning patterns characterized by 1) a gradual decrease outward from a core value of 10-150 ppb, 2) a spike to approximately 1.5x the core concentration, coincident with a high-Y annulus, 3) a decrease by at least an order of magnitude towards the rim, and 4) a very slight rise at the rim. Monazite occurs external to the Y annulus, indicating monazite formation was accompanied by a large drop in the LREE content of garnet and presumably all other major silicates. This inference is consistent with zoning in plagioclase from st-zone rocks, which shows a decrease from 3-6 ppm in the cores to about 1 ppm on the rims. LREE contents in muscovite and biotite average less than 50 ppb each, and staurolite and kyanite contain about 10 ppb for all rocks. The large and immediate decreases in LREEs in the garnet and plagioclase at the onset of monazite formation show that monazite indeed scavenges LREEs from major silicates; nonetheless, the absolute concentration of LREEs in the silicates (no more than a few ppm) is not sufficient to produce monazite in the abundance witnessed by these rocks. A LREE-rich accessory phase must have been present prior to monazite formation to account for the quantity of monazite formed. Single allanite grains in three samples (<5% of samples investigated), and possible pseudomorphs of plagioclase after allanite indicate that allanite must have been a common source of LREEs, despite the lack of direct evidence in a vast majority of the samples.
V21B-0603 0800h
Trace Element Evidence for a Hydrothermal/Magmatic Origin of Stratiform Magnetite Deposits in the Crystal Spring Formation
The Beck Spring Magnetite Deposit is a stratiform body located within the stromatrolitic limestone deposits of the Neoproterozoic Crystal Springs Formation in San Bernardino County, California. There is still some debate regarding its origin and age. Its stratiform nature suggests that it could be a product of metamorphism of a Fe-rich sedimentary layer. Alternatively, it may have formed through hydrothermal processes associated with the 1.0 Gy diabase intrusions or perhaps even younger magmatism, such as the ~14 Ma magmatism responsible for the nearby Kingston Peak pluton. The magnetite deposits are two layers radiating from an ore body characterized by an intergrowth of massive fine-grained magnetite, calcite and pyrite, while the surrounding country rock is a limestone consisting of an orangish white calcite. Here we report the trace element abundances of the magnetite deposits and surrounding limestone in an effort to constrain its origin. We find that the magnetite and limestone wallrock have distinctive primitive-mantle normalized trace-element abundances. In particular, the magnetite has a distinct negative Eu anomaly, anomalously low Th/U ratios, general enrichments in REEs and HFSEs and extreme depletions in Ni and Cr. These features are unlike Fe-rich sedimentary deposits. Instead, the negative Eu anomaly strongly suggests that fractional crystallization of plagioclase took place, while depletions in Ni and Cr are indicative of a highly evolved compositions. The low Th/U ratios also imply that transport of these elements may have occurred via aqueous fluids. These characteristics suggest a magmatic or hydrothermal origin for the magnetite. This is supported by the fact that the immediate surrounding limestone has trace-element systematics very similar to that of the magnetite layers although there is little or no petrographic evidence for sulfide and magnetite infiltration into the limestone. The trace-element systematics of the limestones thus indicates that the fluids associated with the formation of the magnetite deposits metasomatized the surrounding limestone. Thus, we propose that magmatism in the area caused the formation of the magnetite ore deposits and that associated hydrothermal fluids metasomatized the surrounding limestone resulting in the trace element characteristics we have reported.
V21B-0604 0800h
Quartz Hydration Dating in a Geological Perspective
Quartz hydration dating (QHD) is a new direct geochronological technique. The technique measures the lapse time since the fracture of coarse grained (>3mm) quartz crystals. It relies on the phenomenon of water diffusion into quartz leading to the formation of a hydration layer at that can be measured by a nuclear reaction of hydrogen by particle acceleration or SIMS. Diffusivity coeffients for temperature, crystallographic orientation due to anisotrophy of quartz and the initial H content provide a general equation for calculation of age. The activation energy is 57 kJ/mole in the temperature range from ambient to 200C. The time range of QHD is 100-100,000 years. The error is 20-35% depending on the defects in the quartz structure and exposure temperature. The dating of fractured quartz crystals has numerous geological applications, applied to earthquake recurrence intervals, chronology of volcanic eruptions, meteorite impacts, mass wasting, weathering and fracture during transport and deposition. The technique has been applied in geoarchaeology. QHD provides a physical model for the development of other mineral dating techniques. The experimental protocol outlined in this study can be adopted for future studies.
V21B-0605 0800h
Source-Related Chemical And Isotopic Heterogeneities In Granitoid Magmas
A survey of a large number of post-orogenic plutons, shows that regional-scale, intraplutonic isotope heterogeneities (initial 87Sr/86Sr and/or eNd) are common in crustally derived granitic rocks. This is expected since the source materials, from which the granitic magmas were created by partial melting, are themselves heterogeneous, especially from metasedimentary sources. If we accept models of granite emplacement involving the rapid ascent of magma along dykes, there must be potential for the preservation of source-inherited chemical and isotopic heterogeneities. However, their preservation will depend on the extent of post-emplacement homogenisation processes, such as diffusion and convective mixing, in the magma chambers. Micro-sampling of feldspar crystals in granites has revealed complex internal Sr and Nd isotope variations thought to reveal subtle variations in the isotopic composition of the melt from which particular zones crystallised. We therefore have evidence of both macro-scale (regional) and micro-scale isotopic variation within granitic magmas. What is unknown is the extent of chemical and isotopic variation on the meso-scale. We have obtained high-precision geochemical and isotopic (Sr, Nd, and O) analyses for spatially well constrained samples, from two granitic intrusions from the UK, to study the scales over which isotope heterogeneities are preserved, and the maximum volumes of magma over which isotope and elemental homogenisation may have been achieved. These data provide important constraints on the physical and chemical characteristics of processes that occur during magma genesis, ascent, and emplacement. The Criffell pluton (SW Scotland) has previously been shown to be isotopically heterogeneous on the regional scale, with initial 87Sr/86Sr varying from 0.70521 to 0.70728, generally increasing inward, towards the centre of the pluton. This was interpreted as representing isotopic heterogeneity within the source region. Our study has revealed 87Sr/86Sr and eNd homogeneity on the 10 to 100 m scale. This scale therefore represents either the minimum scale over which homogenisation was achieved, or alternatively, the size of a single isotopically distinct batch of magma, within a pluton composed of many such batches. In contrast, the Dartmoor granite (SW Britain) shows marked variations in geochemistry, mineralogy, and texture within samples collected from the same location. Modelling suggests that geochemical variations cannot be explained by fractional crystallisation alone, implying the presence of more than one magma. Initial 87Sr/86Sr ratios (0.70961 to 0.71291 ± 0.0003) and d18O values (10.01 to 10.98 ± 0.1) show that isotopic variation exists on a scales at least as small as 0.5 m. The scale of isotopically distinct domains in the intrusion may represent the sizes of magma batches that were emplaced as successive magma pulses that coalesced to form the pluton. Alternatively these heterogeneities may represent volumes of magma that escaped homogenisation by mixing and diffusion. The physical causes of the preservation of such heterogeneities are under investigation. Chemically, it seems certain that diffusional equilibration (mixing) has been quite inefficient, at least in Dartmoor.
V21B-0606 0800h
Generation of Magmas Within the Southwest Nevada Volcanic Field: Constraints Based on Trace Element Concentrations in Melt Inclusions and Sanidine
The southwest Nevada volcanic field contains four large compositionally zoned ash-flow tuffs, which are among the best studied in the world. This study presents trace element data from melt inclusions and sanidines in order to evaluate interpretations that the compositional zoning is due to the emplacement of discrete magma batches. The earlier ash-flows are the Topopah Spring (13.4 Ma, 1200 km$^{3}$) and the Tiva Canyon (12.9 Ma, ~900 km$^{3}$) tuffs. The later ash-flow tuffs are the Rainier Mesa (11.6 Ma, 1200 km$^{3}$) and the Ammonia Tanks tuff (11.4 Ma, 900 km$^{3}$), which erupted following a period of major extension that occurred within this region of the southern Great Basin. Each unit consists of a lower portion, dominated by rhyolitic pumice fragments and an upper portion dominated by more mafic pumice fragments - each of these contain distinct Sr, Nd and $\delta$$^{18}$O isotopic compositions. These data are inconsistent with fractional crystallization relating the rhyolitic and mafic portions within each ash-flow tuff. In addition, the compositional variation among ash-flow units cannot be related by fractional crystallization. All melt inclusions within phenocrysts from the more mafic pumice fragments of each tuff have identical trace element concentrations. Furthermore, melt inclusions within the high-silica pumice fragments of each tuff also have identical trace element concentrations, except for a group of high-Rb, high-Nb inclusions from Rainier Mesa. Magma mixing occurred during evolution of each tuff, and is recorded in both the melt inclusions and in the host sanidines. These mixing events are subtle in the Topopah Spring tuff, but extensive mixing produced magmas of intermediate composition within the Tiva Canyon, Rainier Mesa, and Ammonia Tanks tuffs. The chemical analyses of pumice fragments from the Topopah Spring, Tiva Canyon, and Ammonia Tanks tuffs are consistent with their generation from a common source. However, Rainier Mesa magmas are different, having been formed from a more Th-enriched, La-depleted source region with a higher $\delta$$^{18}$O signature, and the existence of melt inclusions with unusual compositions. Comparison of isotopic compositions among ash flows indicates an increase in mantle melt contribution over time. However, the Rainier Mesa source region was sufficiently high in $\delta$$^{18}$O so that this signature was dominant.
V21B-0607 0800h
Orthopyroxene/Clinopyroxene Trace Element Partitioning Systematics in the Natural Sample
It is well established that trace element partition coefficients (D) vary as a function of pressure, temperature (T) and composition, but experimental partitioning studies are expensive, time consuming and often produce a limited number of mineral partitioning data. However, fresh spinel peridotite xenoliths where opx and cpx are in major element equilibrium can provide abundant mineral/mineral partitioning data that allows exploring some aspects of partitioning systematics in the upper mantle. Here we present major and trace element data on pure mineral separates of spinel peridotites from the oceanic lithosphere (Oahu, Hawaii) and continental settings (Portillo Maar, New Mexico, N.E. Queensland, Australia). Our data, along with literature data, shows that D$^{opx/cpx}$ for the HREE, Y, Zr, Hf, Ti, Sc, V decrease with decreasing temperature (and, therefore, increasing Ca content in cpx), while Co, Mn, and Ni increase. Interestingly, where experimental data exists, the natural D$^{opx/cpx}$ are always lower and fall at the extension of the experimental data on an lnD$^{opx/cpx}$ vs. T plot, with good correlations between temperature and lnD$^{opx/cpx}$, e.g. T=355.2lnD$^{opx/cpx}$$_{Y}$+2016 with r$^{2}$ = 0.91, T=427.8lnD$^{opx/cpx}$$_{Er}$+2107 with r$^{2}$ = 0.9. These empirical thermometers may be used in the absence of major element data. Also, these systematic variations in the natural sample can be used to predict D$^{opx/cpx}$ for melting appropriate temperatures in the upper mantle where such Ds are not experimentally available, e.g. Co and Mn. The systematic difference between experimental and natural D$^{opx/cpx}$ also suggests that it is inappropriate to use experimental Ds to reconstruct peridotite bulk rock trace element contents, since this leads to overestimating these concentrations. This is especially true for abyssal peridotites where no reliable opx or bulk rock trace element data is available.
V21B-0608 0800h
Geochemistry of Basaltic Dikes From the Northwestern Deccan Volcanic Province: Chemostratigraphic Implications
Basaltic dikes occurring in the northwestern part of the Deccan Volcanic Province (DVP) remain poorly studied; yet, their position within the province is important since this region has been postulated to be the source for most of the Deccan lavas. In the Sangamner region of the northwestern DVP, more than 30 dikes and at least 1 sill intrude the Thakurvadi Formation, which is part of the oldest Kalsubai subgroup. This study focuses on the geochemistry of these intrusives and seeks to understand their relationship to the flows that they intrude, and to other flows in the DVP. The ultimate objective is to attempt to place the instrusive suite into a chemostratigraphic context in order to address broader issues such as the location of potential feeders and the evolution of the magmatic plumbing system. Most dikes trend NE-SW or E-W, and range in thickness from 5-18 m. 25 dikes and a sill were analyzed for major and selected trace elements in order to relate them to existing DVP chemostratigraphy, as well as to identify chemical groups within the dikes. High MgO ($>$ 8%) and low MgO ($<$ 7%) groups are readily distinguished. Within this context, further subgroups are identified based primarily on concentrations and ratios of elements such as Fe, Ti, Nb, Y, and Zr. A combination of geochemical characteristics was used to assign the intrusives to particular subgroups within the established DVP chemostratigraphy and to individual formations in some cases. The high MgO group (including the picritic sill) is assigned to either the Thakurvadi or the Bushe formations. Most dikes belonging to the low MgO group are assigned to either of the Poladpur, Bhimashankar, or Khandala formations. Three dikes belonging to this group have chemical characters similar to the Giant Plagioclase Basalts, which separate each of the older formations (although they lack the typical plagioclase mega-phenocrysts). These data highlight the compositional diversity of the intrusives in this region. They suggest that while some dikes could have been feeders to the flows that they intrude, most dikes probably fed younger formations or were simply late stage intrusions of magmas representing the younger formations. The dikes are presently under investigation for radiogenic isotopes and REE in order to better constrain their stratigraphic affiliation as well as to understand their petrogenesis.
V21B-0609 0800h
Laser Raman Spectroscopic Characterization of Shocked Plagioclase from the Lonar Impact Crater, India.
We report Raman spectra of shocked plagioclase grains from the Lonar impact Crater of India. The Lonar Crater, located in the Buldana district of Maharashtra, India ($19\deg$58'N, $76\deg$31'E), is an almost circular depression in the 65Ma old basalt flows of the Deccan Traps. Age estimates of this impact crater range from 10-50ka. Tektite and basalt samples were collected for this study from the rim of the crater, which is raised about 20 meters above the surrounding plains. For comparison, a Manicouagan maskelynite and an unaltered mid-oceanic ridge basalt with plagioclase laths were also analyzed. Polished thin sections of all these samples were first petrographically studied. The MORB plagioglase as well as the plagioclase from Lonar host-basalts show first order interference colors and distinct multiple lamellar twinning. The Manicouagan maskelynite is isotropic under crossed-polars. The Lonar tektite samples characteristically demonstrate spherules which are identified by their perfectly circular cross-section and isotropic nature. The spherules also contain fragments of the host basalt with plagioclase laths showing lamellar twinning. The groundmass within the spherules shows lath shaped plagioclase grains, most of which show varying degrees of isotropism due to maskelynitization. Raman scattering measurements were performed using the 514.5 nm line of an argon ion laser at an intensity of 40 kW/cm$^{2}$. An inverted microscope (Nikon TE3000) with 50x objective (NA 0.55) was used for confocal imaging. A holographic notch filter removed residual laser scatter and the Raman scattering was detected by a silicon CCD at -$90\deg$C (Princeton Instruments Spec10-400R). Raman spectra were collected from ~250 cm$^{-1}$ through 2000 cm$^{-1}$. Raman spectra of crystalline unshocked plagioclase feldspars from the MORB and the Lonar host basalt show strongest peaks at 265 cm$^{-1}$, 410 cm$^{-1}$, 510 cm$^{-1}$ and 1110 cm$^{-1}$. The results remain the same for different points in a single grain but vary slightly from one grain to another, perhaps due to difference in composition. This observation is consistent with previously reported analyses of unshocked plagioclase feldspars. The Raman spectra of the maskelynites from the tektite samples show more flattened-out patterns. Earlier studies have reported the "disappearance" of peaks due to an increased luminescent background in experimentally shocked single plagioclase grains (Heymann and Herz, 1990, Cont. Min. Petr. 17, 38-44, 1990). Our study in multiple variably shocked plagioclase grains of the Lonar basalt impact breccia further corroborates these previous observations and reinstates the importance of Raman Spectroscopy in identifying shocked plagioclase grains.
V21B-0610 0800h
Evidence for Fractionation and Recharge in Shallow Basaltic Magma Chambers: Kimama Butte, Snake River Plain, Idaho
Small, monogenetic shield volcanoes are the most prominent feature of the Snake River Plain, Idaho. Even though monogenetic shields are very short lived, it is not uncommon for compositional heterogeneity to be present within a single volcano. Kimama Butte shows distinct changes in major and trace element compositions and in olivine and plagioclase from the earliest, most distal flows to the late-stage summit eruptions. The rocks from Kimama Butte display a strong Fe-enrichment trend typical of fractionating tholeiitic basalt. Major and trace element compositions range from: 6.9- 5.5% MgO, 15.6-18% Fe2O3, 15-13% Al2O3, 3.2-4.3% TiO2, 9.5-10.1% CaO, 0.6-0.8% K2O, and P2O5 0.6 -1.1%. P, the most incompatible element, increases by a factor of 1.8 and Ce increases by 1.7 times. Trace elements (Ni$<$80 and Cr$<$200 ppm) , olivine compositions (about Fo65), and low Mg\# ($<$65) show that none of the lavas are primary melts from the mantle. High iron concentrations may be due to differentiation not great depth of origin. Element variations are best explained by the fractionation of plagioclase and olivine. For example, Al2O3 declines with increasing differentiation and Ca varies only slightly across the entire range of lavas. Moreover, normalized trace element patterns have a prominent negative Sr anomaly as a result of plagioclase fractionation. MELTS simulations were performed with a constant parent but varying pressure, water content, and fO2 to determine the conditions which best fit the observed lava compositions. The phase assemblage (no pyroxenes or oxides) and major element variations are best explained by approximately 25-30% crystallization of olivine and plagioclase at low-pressure ($<$3 kb to as low as 100 b). Electron microprobe analyses of plagioclase (An72) and olivine (Fo75) phenocrysts agree with the predicted mineral compositions from MELTS. However, low-pressure fractionation alone cannot explain the observed the strong enrichments of incompatible elements. Instead, they can be modeled as the result of magma recharge into the evolving chamber. Calculations suggest that recharge rate was about 0.6 the crystallization rate. Assimilation of ferrogabbro intrusions or of the Phosphoria Formation at great depth may not be necessary to explain the incompatible element enrichments.
V21B-0611 0800h
The Possible Effect of Benches and Convective Flow Segregation in Creating Mineral Layers in Intrusive Sheets: A Model for the Palisades Sill
Recent mapping of the Palisades eruptive sheet of New York and New Jersey along strike defines a mafic horizon with MgO\# (56-60), Cr ($>$500 ppm), and an internal quench horizon marked by a decrease in grain size of approximately 20%. The mafic horizon pinches northward leading to a quenching of a hypersthene-phyric horizon at the level of Piermont, New York. Analyses of the sandwich horizon are consistent with fractionation of pyroxene and plagioclase and lead to a subdivision of the more felsic granophyre into a sediment-derived partial melt and basic pegmatite fractions (crystals to 2 cm). Lack of a distinctive positive Eu anomaly separate the bulk of the Palisades sections from the section at the George Washington Bridge. This is consistent with flow-injection models that create a pre-or syn-intrusion segregation of mafic materials followed by convection overturns. We suggest that basic and ultramafic igneous sheets may become stratified by encountering benches or simply through ordinary internal mechanical differentiation processes driven by flow segregation. Mechanical differentiation processes in lavas and magmas are modeled using the unrestricted Navier Stokes equations in both 2D and 3D as applied successfully to the formation of zoning in the Bushveld complex and in Komatiite flows. In contrast to the Bushveld complex that spans almost 7000 to 10000 meters, the Palisades Sill averages approximately 300 meters in depth and the classical range along strike of the olivine zone within a sheet-like body (from the George Washington Bridge Section to approximately Haverstraw New York) is almost 25 kilometers. Previous assessment that the hyalosiderite dolerite portion appears to occur in a topographic low leads to the suggestion that a perturbation in the wall rock of the intrusion, such as downfaulted block, may trigger the development of a mineralogical layer as a manifestation of eddying behind the block during the flow period in which the magma is emplaced. Arguments for a centered olivine zone, i.e., a D shaped distribution, have given no theoretical structure to date. There is field evidence for D-shape, but also for S-shape distributions. The two models are not inconsistent. The D-Shape model suggests that the suspended olivine may be driven by dispersive pressure to the center of the conduit. Some of the S-shape intrusions may reflect variations on a suspended load. Both situations may well be manifestations of the flow regimes that attended the emplacement of the magma as they both arise in the mathematical modeling.
V21B-0612 0800h
Layer Formation in Convective Magma Chambers
The dynamics of a convective magma chamber is crucially influenced by the competetion between sedimentation and convective suspension of crystals. Crystal settling combined with the crystal's density contribution is a possible mechanism leading to differentiation and layer formation. Here we address the question whether crystals can remain suspended or whether they are able to dynamically form a layered structure within the convective lifetime of a magma chamber. We employ an existing numerical method that, by means of a finite volume scheme, discretizes the equations for thermally driven convection in an infinite Prandtl-number Boussinesq fluid in Cartesian geometry. We implement a newly developed settling algorithm for the numerical study of finite-sized-particle settling in a non-dilute convective suspension. Our approach considers a consistent settling velocity and the density contribution due to particle mass. The buoyancy ratio $B$, which is the ratio of the density variation due to crystal mass to the thermal density variation, is varied for five different Rayleigh numbers, covering a range of four orders of magnitude. We find $B$ to be a critical parameter and its critical value to depend on the Rayleigh number. For subcritical values we observe that the presence of a crystal phase reduces convective vigor and most crystals stay suspended. When a critical buoyancy ratio is exceeded, the presence of crystals can significantly alter convective motion. For all investigated Rayleigh numbers we find a critical buoyancy ratio, above which layering can be achieved from an initially unstratified fluid. Most of the crystal mass collects in the dynamically created bottom layer, even for cases where the average settling velocity is three orders of magnitude smaller than the root mean square convective velocity. The time it takes a crystal to travel across the height of the cell with the full settling velocity in the absence of a thermal gradient defines the settling timescale. Layer formation in all observed layering cases occurs on this time scale, even though the average settling velocity is reduced by at least one order of magnitude due to hindered settling. In many cases (e.g. basaltic magma chambers) the settling time is short compared to the time that magma chambers take to solidify. We conclude that dynamical layer formation that is connected to crystal settling and the crystals' density contribution is a likely mechanism for creating layered structures within the convective lifetime of a magma chamber.
V21B-0613 0800h
Carbon-Bearing Volcanic Rocks in the Linshan, Sichuan, China
The Linshan carbon-bearing (carbonite-bearing) igneous rocks are sparsely distributed in northwest of Sichuan province, China. On the basis of petrogeochemical data, the basalt lavas can be classified into two major magma types. These are (1) a low-Ti/Y type situated in central and eastern Linshan that exhibits low Ti/Y ($<$900), Ce/ Y ($<$10), and SiO$_{2}$ (40-55 wt%) and relatively high Fe$_{2}$O$_{3}$ (Fe$_{2}$O$_{3}$ as total Fe; 3.4-11.5 wt%); and (2) a high-Ti/Y type situated in western Linshan that has high Ti/Y ($>$900), Ce/Y ($>$10), and SiO$_{2}$ (42-55 wt%) and relatively low Fe$_{2}$O$_{3}$ (3.8-7.8 wt%). Elemental data suggest that the chemical variations of the low-Ti/Y and high-Ti/Y lavas cannot be explained by crystallization from a common parental magma. The Linshan Carboniferous basic lavas most likely originated from an asthenospheric oceanic-island-basalt-like mantle source (Sr-87/Sr-86((t)) approximate to 0.60-0.72, is an element of(Nd(t)) approximate to +4 to +7). Crustal contamination and continental lithospheric mantle have also contributed significantly to the formation of the basic lavas of the Linshan Carboniferous rift. Our data show that spatial petrogeochemical variations exist in the volcanic rocks of the Linshan large igneous province. The location of the thickest volcanic succession, which has dominantly tholeiitic lavas, in the eastern Linshan may have been centered over the melting anomaly in the mantle. The eastern Linshan basic magmas were generated by a higher degree of partial melting in the spinel-garnet transition zone of the mantle compared to the alkaline basaltic lavas that are the dominant magma type in the western Linshan. The lower degree of partial melting in the garnet stability field of the mantle, as is characteristic of the western Linshan basic lavas, may be the result of a relatively thicker lithosphere and lower geotherm.
V21B-0614 0800h
Modal Petrology and Geostatistics of the Blue Hills Igneous Complex, Boston, Massachusetts, by Rietveld X-ray Diffraction: Multi-scalar Investigation of Volcanic and Intrusive Relationships
The Blue Hills Igneous Complex of eastern Massachusetts consists of mildly peralkaline volcanic and intrusive units including the Quincy Granite, the Blue Hills Porphyry, and a set of mainly pyroclastic rhyolite flow units traditionally called the Aporhyolite. Similar whole-rock chemistry has led most workers to assume that they are related rocks, despite some unclear field relationships. Kaktins (1976) divided the volcanic rocks into six units, but buried contacts do not permit confidence in either their number or stratigraphic position. To test a new method of modal analysis of these rocks, thirty-five samples were crushed, ground to approximately 5 micrometers, spray-dried to produce randomly oriented powder, and analyzed by x-ray diffraction. A constant eleven-phase Rietveld starting model was applied to the x-ray spectra, and then refined to produce a modal database of phase proportions in each sample. Geostatistical analysis with GIS software delineates a number of trends, with statistical measures of uncertainty. Aegirine in volcanics decreases in abundance with distance south from the E-W contact of volcanic rocks and granite. Riebeckite is found in the granite (both as veins and as apparently magmatic crystals) and the porphyry, but is less abundant or absent among the volcanic rocks. Where both amphibole and pyroxene are present, they are negatively correlated. The goal is to develop an additional tool for correlation of volcanic rocks, one based on mineral proportions in both aphanitic and phaneritic rocks.
V21B-0615 0800h
Genesis of a Chemically Enriched Olivine Tholeiite from the Eastern Snake River Plain, Idaho
The eastern Snake River Plain (ESRP) volcanic province is overlain by 1-2 km of dominantly Quaternary olivine tholeiite lava flows associated with the Yellowstone hotspot. A second highly evolved geochemical basalt suite has been identified in previous studies, typified by the lava flows of Craters of the Moon National Monument (COM). Geochemical data from three new coreholes drilled on the ESRP was collected to correlate chemically distinct basaltic flow groups and to evaluate chemical changes in ESRP basaltic magmatism over time. The coreholes extend to a depth of approximately 800 ft and penetrate lava flows from 200,000 to 640,000 years old, allowing the temporal geochemical variability of the ESRP to be studied in this area. One flow group, known as the "high-K" flow, identified in two of these cores, is unusually enriched in incompatible major and trace elements relative to ESRP olivine tholeiites. Isotopic data for this flow group ($^{87}$Sr/ $^{86}$Sr = 0.70599 - 0.70604; $^{143}$Nd/ $^{144}$Nd = 0.512431 - 0.512446) suggest derivation from a slightly different, isotopically less enriched magma reservoir than most ESRP olivine tholeiites ($^{87}$Sr/ $^{86}$Sr = 0.70640 - 0.70765; $^{143}$Nd/ $^{144}$Nd = 0.512261 - 0.512474). Major and trace element data suggest that a combination of lower-degree partial melting and more extensive fractional crystallization contribute to the evolved chemistry of the high-K flow. The geochemistry of the high-K flow reflects differences in source chemistry and petrogenetic processes compared to both the ESRP olivine tholeiites and the highly evolved basalt suite (COM), and may represent an intermediate chemical suite of ESRP basalts.