V11A-01 INVITED
On the Integration of Micro-textural and Crystal-Chemical Data With Magma Chamber Dynamics
Many magmatic systems display a dual nature: compositional uniformity at the macro-to-meso scales, and great complexity of provenance, textures and age at the micro-to-meso scale. The integration of microtextural and crystal-chemical data with dynamic models and U-series age dating has tremendous potential to reveal the inner-workings of magmatic systems. However, a means of extending that crystal-scale data back to the largest scales is needed to evaluate the statistical significance of any conclusions drawn on the basis of a few crystals or thin sections. We will address two aspects of this: 1) The application of various statistical tools for the ordering of geochemical information to yield crystal populations who share distinct chemical environments and, 2) exemplify how dynamic templates of crystal dispersal and gathering, linked with kinetics, can reveal the timescales of the emergence, persistence and decay of regions of distinct chemical potential and the corresponding crystal-cargo. We will reprise Wallace and Bergantz (2005), which introduced the concept of the shared characteristics diagram as an organizational framework for crystal zoning that compares information from different phases and chemical tracers, in a common framework. It combines elements of cluster and structured data analysis to create a crystal phylogeny. This allows one recognize the progression of shared environments in the crystal-cargo at any level of statistical significance. We will then use numerical experiments of multiphase reacting flow to illustrate how various regions of distinct compositional character arise, and the controls on how that information is propagated to resident crystals. We will illustrate the diversity of outcomes that even simple open-systems produce, and how the ratio of the Damkohler (kinetic reaction time scale to diffusion time) and Peclet (convection time scale to diffusion time) numbers can be used to characterize both simple and competitive-consecutive reactions. This ratio of dimensionless numbers yields the characteristic time for reaction to that of convection. To first order, this allows one to explore and rationalize the interplay between equilibrium as dictated by the (changing) far-field environment at the largest scales, mediated by the kinetics at the crystal interface. Wallace, G., Bergantz, G.W., 2005, "Reconciling heterogeneity in crystal zoning data: An application of shared characteristic diagrams at Chaos Crags, Lassen Volcanic Center, California," Contrib. Min. Pet., v. 149, p. 98-112.
V11A-02
The Reactive Magma Concept, Local Equilibrium, and Experimental Studies of Silicic to Intermediate Magmas
Experimental phase equilibria are being increasingly used for the determination of pre-eruptive conditions of subvolcanic igneous bodies. Yet, numerous petrologic studies of silicic to intermediate rocks have documented evidence for disequilibrium either between phenocrysts or between phenocrysts and melt. Thus, the applicability of equilibrium experiments requires clarification. Although bulk equilibrium is generally not attained in magmas, in all cases it is possible to identify a subsystem of the magma at local chemical equilibrium (hereafter designated as the reactive magma). This is because magmas, as other natural systems, tend to approach chemical equilibrium. Local equilibrium can be proved from tests of element partitioning between coexisting phases, thermobarometric calculations, or comparing compositions of phenocrysts and melts with phases from equilibrium experimental studies. Reactive magma volumes can vary by orders of magnitude between magmatic systems, depending on physical processes and their timescales, and on equilibration kinetics. Because equilibration kinetics are much faster in melts than in crystals, sub-aphyric chemically homogeneous (on a whole-rock basis) igneous bodies (eg, Novarupta 1912 high silica rhyolite) have large reactive magma volumes, and are probably close to bulk equilibrium. By opposite, smaller reactive magma volumes are to be expected for crystal-rich bodies. The recently erupted Mt. Unzen dacite illustrates one extreme case of phenocrysts being all out of equilibrium with the interstitial melt. Another indication for small reactive magma volumes is chemical zonation of matrix glasses (eg, Laacher See Tephra). Reactive magma volumes also change with perturbations (of either physical or chemical nature) imparted to the igneous body. For pre-eruptive conditions of equilibration to be determined experimentally, starting materials must closely approach the reactive magma composition appropriate for the pre-eruptive history of the igneous body. Certain discrepancies between experimental phase and phenocryst assemblages in recent studies likely result from a poor matching between starting compositions and the appropriate reactive magma. The reactive magma concept is useful for the planning and interpretation of experimental studies on silicic to intermediate magmas.
V11A-03
A new Model for Crystallization and Differentiation in Magma Chambers
How magma reservoirs cool and crystallize is intimately related to how they form and grow. We present a new model of magma accumulation and crystallization due to successive magma injections which accounts for the kinetics of crystal nucleation and growth. In this model, magma reservoirs develop out of sill or dyke complexes. For a thin dyke or sill, cooling proceeds faster than crystal nucleation and growth, implying glass formation. Successive injections thus lead to the accumulation of uncrystallized magma. Temperatures in the sill or dyke complex increase gradually with each injection until they become large enough to allow crystal nucleation and growth. In such a system, crystals are generated as temperatures rise. Our calculations show that crystallization is sudden and massive, affecting most of the stored magma, due to a positive feedback between latent heat release and crystallization (i.e. opposite to what occurs in a cooling sequence). The crystal population includes early crystals of primitive composition generated at each injection and late crystals spanning a large compositional range due to the massive crystallization event. At Earth's surface, the eruption record would thus be characterized by a series of primitive lavas, corresponding to the discrete injection events, followed by a large volume of evolved lava, corresponding to the residual melt generated by the massive crystallization event. Following the massive crystallization event, subsequent magma injections lead to an episodic crystallization evolution, with a different compositional trend.
V11A-04
Sector and/or Concentric Zoned Augite Phenocryts in Mafic Arc Lavas: Evidence for Decompression Induced Crystallization With Implications for the Mechanism of Andesite Fractionation
The assumption that heat loss to the surrounding country rock is the primary driving force for crystallization and crystal-liquid fractionation underlies nearly every model of magmatic differentiation. In calc-alkaline magmas, which contain large amounts of dissolved H2O, another driving force for crystallization exists, that being decompression and H2O exsolution during magma ascent. Numerous mafic (basalt to basaltic andesite) lavas from volcanic arcs throughout the world contain large (2 to 10 mm diameter) augite phenocrysts that are commonly sector zoned and invariably concentrically zoned, where the concentric zoning consists of zones of optically clear augite separated by thin zones rich in mineral and melt inclusions. In many crystals the melt inclusions display a systematic core-to-rim decrease in H2O concentration. Together these observations are evidence for rapid crystallization during decompression-induced H2O exsolution. Starting with this assumption, detailed analysis of a single hydrous (2 wt. % H2O) augite bearing Aleutian basalt yielded the following results: (1) crystallization commenced at a depth of around 1 km; (2) augite crystal growth rates are on the order of 10$^{-7}$ cm/s; (3) augite residence times was on the order of 40 days; which (4) yields an average magma ascent rate of 0.03 cm/s. This ascent rate is several orders of magnitude greater than calculated crystal settling rates implying that most such mafic magma bodies are likely to become congested and stall within the upper most crust. If and when this happens, extraction of the fractionation-generated interstitial andesitic liquid followed by cumulate entrainment and/or mixing with more mafic liquids could explain many of the chemical, mineralogic and textural features commonly observed in andesitic lavas throughout the world.
V11A-05
Seeking the Spoor of Reactive Bulk Assimilation.
The integration of crustal solids and liquids into mantle-derived magmas as a consequence of reactive bulk assimilation is a key process in the petrogenesis of most large-volume intermediate to silicic magmas. Feldspars, especially plagioclase, are ideal monitors for solid input into magmatic systems because they commonly preserve a chemical and isotopic record of their history. However, much of the solid material input during bulk assimilation is in the form of less robust phases, especially quartz, amphiboles, micas, pyroxenes and Fe-Ti oxides. In hydrous systems (e.g. arcs and any region where assimilation of hydrated crust occurs), incongruent melting and crystallization reactions involving amphiboles and micas are the principal controls on the redistribution of elements and isotopes amongst solids and melts during bulk assimilation. These reactions convert hydrous silicate assemblages (e.g. biotite gneiss or amphibolite) to an anhydrous (pyroxene granulite) solid assemblage coexisting with a siliceous (trondjhemitic to granitic) melt during assimilation and produce amphibole and mica via equilibrium (essentially) reaction between solid pyroxenes and oxides and evolved melts during late crystallization. The spoor of these reactions is likely to be subtle and commonly either avoided or ignored during geochemical study. Reactive crystallization features are relatively familiar and include replacement of pyroxenes and oxides by amphibole and biotite. The subtleties of chemical redistribution that can accompany these reactions, however, are less well appreciated. For example, solid reactants may carry the chemical and isotopic signatures of crustal xenoliths, mantle phenocrysts, or both, while the chemistry of the reactant melt likely reflects complex AFC (as well as low-T, equilibrium reaction crystallization) processes. Remnants of xenoliths that have undergone extensive dehydration melting may be more difficult to recognize. The presence of single plagioclase crystals that retain a crustal signature suggests disaggregation of some xenoliths down to the scale of individual crystals. In such cases, remnants of xenolithic solids will be difficult or impossible to recognize petrographically. On the other hand, if disaggregation is less complete, ghosts of individual xenoliths may still be found, even if they are microscopic or represented only by trains of xenocrysts. Easily recognizable crustal xenoliths probably contribute little to the chemical modification of the magma as a whole, either because they are refractory/non-reactive or because they were incorporated when the magma was very close to its solidus.
V11A-06
Evidence for Magma-Mixing and Disequilibrium in 'Primitive' Basaltic Andesites From Mount Shasta, Northern California
High-Mg basaltic andesites near Mt. Shasta volcano have been considered fundamental to establishing the existence of exceptionally water-rich primary magmas in this system, implying significant slab-derived fluid fluxes into the underlying mantle wedge (Grove et al., 2002). This notion was reinvestigated via new mineralogical and geochemical studies of fresh scoria blocks from the Whaleback volcano (loc. S17; Anderson,1979). These high-Mg andesites (58% SiO2, 8.5% MgO, Mg\# = 76, 120 ppm Ni, 550 ppm Cr) carry small dunitic xenoliths and xeno/phenocrysts (ol+opx+cpx). Plagioclase is not a liquidus phase. Electron microprobe traverses and back-scattered images show that mafic silicates, particularly pyroxenes, have complex histories. Olivine compositions of larger crystals and interiors are often above Fo$_{90}$ up to Fo$_{94}$ whereas microphenocrysts and rims of larger crystals are ~Fo$_{87}$. Complexities among pyroxenes include: (a) Cores of opx and cpx with low Mg\# (~67) containing melt inclusions; this evidence indicates these pyroxenes crystallized from magma of roughly dacitic composition; (b) Virtually all low Mg\# grains are resorbed and have overgrowths (~20 microns) of high Mg\# (87-92) that may be internally zoned arriving at a Mg\# near 80 at the outermost euhedral rim; (c) Another variant is orthopyroxene with 'wormy' texture and either a thin (~15 microns) euhedral overgrowth or anhedral outline; compositions of resorbed interiors and overgrowth are similar ( Mg\# range: 80 to 90), but distribution of lower and higher Mg\# in resorbed areas is patchy whereas any compositional zoning of overgrowth follows crystal shape and arrives again at a Mg\# of ~80 at the outermost rim. These data record mixing of diverse magmas (dacite and one or more basaltic liquids) combined with entrainment of ultramafic crystal debris during wall rock contamination, and eventual cooling and equilibration. Low Al2O3 contents in the pyroxenes imply that these minerals grew at relatively low pressure. Given these relations, the erupted hybrid magma is a product of open system processes and it is unlikely to retain a high fidelity record of its deeper origins. The inferred high water contents are associated with the highest Mg\# ol and opx (Anderson, 1979), but the origin of these signatures remains enigmatic. Low B contents and δ11B (ca. -5 permil) in this and other Shasta lavas are inconsistent with significant involvement of slab-derived fluids.
V11A-07
Magma Evolution and Open System Processes at Shiveluch Volcano: Insights From Phenocryst Zoning and Melt Inclusions
This study identifies open processes relating to magma evolution and mixing at Shiveluch Volcano, Kamchatka, from phenocryst zoning and melt inclusion chemistry. Shiveluch lavas are hornblende-plagioclase andesite with average pre-eruptive temperatures of 857 °C and high fO2 (NNO+1.5 to NNO+2). Phenocryst zoning in plagioclase, hornblende and apatite is diverse, with oscillatory, simple, multiple and patchy zoning and sieve textures identified. Each phenocryst type records a distinct magmatic history, and different populations are spatially and temporally separated during crystallisation. In patchy plagioclase, melt inclusions are associated with low-An patches, while in patchy hornblende, they are associated with Al-rich patches. Ba, Sr systematics indicate that patchy phenocryst cores are derived from cumulate material. The patchy texture forms by resorption during H2O-Undersaturated decompression. No external change in melt chemistry occurs, but decreasing XAn in plagioclase results from decreasing pH2O and drives the compositional changes in hornblende. Abundant evidence for magma mixing events is also observed. Mixing with hotter, mafic magma results in olivine and orthopyroxene xenocrysts with reaction rims and strong normal zoning. Mixing with cooler, felsic magma results in sieve plagioclase with Mg, Fe-rich rims. Cryptic mixing, between magmas of similar compositions, can result in phenocryst zoning if there are differences in volatile contents. Sulphur-rich magma batches release SO2 and O2 as anhydrite breaks down, oxidising the surrounding melt and causing simple zoning in hornblende and apatite. Cryptic mixing reflects the intermittent ascent and assimilation of small magma pulses, which coalesce to form larger subvolcanic magma bodies. At shallower levels, oscillatory plagioclase and rims on patchy plagioclase crystallise during ascent in the conduit. Crystallisation is accompanied by oxidation, and by a temperature increase of ~80 °C as latent heat is released.
V11A-08
Crystalline Debris, Not Phenocrysts: Mineralogical Evidence for Partial Assimilative Recycling of the Mafic-Ultramafic Plutonic Roots of a Continental Arc
Phenocryst proportions in many mafic lavas of the Quaternary Tatara-San Pedro complex (TSPC; 36°S, Chilean Andes) are minor compared to modal abundances of coarse olivine, clinopyroxene, and plagioclase xenocrysts derived from disaggregated mafic-Ultramafic xenoliths. The major and trace element imprint of this process is significant, but compatible and incompatible elements are poorly anti-correlated due to irreversible blending of grain-boundary melts and variable retention of xenocrysts. Associated isotopic variations are minor and decoupled from the elemental signal, apparently due to the young age of the assimilated lithologies. Healed microfractures in embayed olivine xenocrysts (Fo$_{85-75}$), which are rampant in mafic and ultramafic xenoliths at the TSPC, are proof of subsolidus histories. Evidence for the former presence of plagioclase, othopyroxene, amphibole, and phlogopite as replacement-reaction products after olivine and other phases is preserved as secondary mineral assemblages which are the crystallization products of Mg- and alkali-rich grain-boundary melts that were encapsulated in partly replaced olivines: (1) F-phlogopite with widely varying compositions and modal abundances within and among olivines in the same thin section, (2) multiple spinel populations that require multiple origins, (3) relatively sodic plagioclase (An$_{40-60}$) with very high ferric iron, and (4) shifts in oxygen fugacity of up to +1.4 log-Units in contaminated magmas relative to uncontaminated parental magmas (NNO buffer). Residence times of olivine xenocrysts from diffusion-rate modeling are 0.2-24 yrs, with a mode at 3-5 yrs. Short residence times correlate with the preservation of phlogopite-rich assemblages that crystallized from trapped grain-boundary melts, whereas those in olivines with long residence times are partially equilibrated with host magmas. Stoping and digestion of such xenoliths is fast and efficient with regard to modification of basalt compositions, but the mineralogical record may be partly obscured by magma-storage times as short as 5-25 years (i.e., much shorter than typical repose periods at arc volcanoes).