The Growth and Evolution of Large Silicic Magma Bodies III
Presiding: O Bachmann, University of Geneva; C F Miller, Vanderbilt University
V22A-01 10:30h
Insights into the near-solidus evolution of an erupted batholith: the trace element record in the Fish Canyon magma body, Colorado
Highly evolved rhyolite glass plus near-solidus mineral assemblages in voluminous, dacitic, crystal-rich ignimbrites provide an opportunity to evaluate the late magmatic evolution of granodiorite batholiths. This study reports laser-ablation ICP-MS analyses of trace element concentrations in feldspars, hornblende, biotite, titanite, zircon, magnetite, and interstitial glass of the crystal-rich Fish Canyon magma. The high-silica rhyolite glass is characterized by relatively high concentrations of feldspar-compatible elements (e.g., 100 ppm Sr and 500 ppm Ba) and low concentrations of Y (<7 ppm) and HREE (~1 ppm Yb), hence high LREE/HREE (Ce/Yb>40) compared to many well-studied high-silica rhyolite glasses and whole-rock compositions. Most minerals record some trace element heterogeneities, with, in particular, one large hornblende phenocryst showing 4-6 fold core-to-rim increases in Sr and Ba coupled with a decrease in Sc. The depletions of Y and HREE in the Fish Canyon glass relative to the whole-rock composition (concentrations in glass ~30% of those in whole-rocks) reflect late crystallization of phases wherein these elements were compatible. As garnet is not stable at the low-P conditions at which the Fish Canyon magma crystallized, we show that a combination of modally abundant hornblende (~4%) + titanite (~0.5-1%) and the highly polymerized nature of the rhyolitic liquid led to Y and HREE depletions in melt. Relatively high Sr and Ba contents in glass and rimward Sr and Ba increases in euhedral, concentrically-zoned hornblende suggest partial feldspar dissolution and a late release of these elements to the melt as hornblende was crystallizing, in agreement with textural evidence for feldspar (and quartz) resorption. Both observations are consistent with thermal rejuvenation of the magma body prior to eruption, during which the proportion of melt increased via feldspar and quartz dissolution, even as hydrous and accessory phases were crystallizing. These thermal oscillations may be common in batholithic chambers, and could provide a mechanism to stir the magma and induce the whole-rock homogeneity that characterises these bodies.
V22A-02 10:45h
Eruptive and Intrusive History of Mount Mazama and the Crater Lake Region Explains Growth of a Young Silicic Magma Body
Crater Lake caldera (subsidence area >25 km2) collapsed into a shallow long-lived magmatic system during the ~7.7 ka climactic eruption of Mount Mazama that ejected ~50 km3 of magma dominated by homogeneous crystal-poor rhyodacite pumice. Geologic mapping, Ar and U-Th geochronology, and geochemistry provide volume-time-composition information on the growth and evolution of the Mazama system. Granodiorite and related plutonic rocks in the climactic ejecta give a window into the intrusive component of that system. We interpret the eruptive and intrusive history in terms of the competition between (1) crystallization driven by degassing and hydrothermal cooling and (2) thermal input from a regional magma flux focused at Mazama. Throughout its ~400 kyr history, the andesite-dacite volcano known as Mount Mazama, built astride an earlier silicic dome field, was accompanied by nearby basaltic to andesitic effusive volcanism representing a range of differentiation of melts of depleted mantle containing varied amounts of water and subduction-related fluid-mobile elements. Similar parent magmas appear to have differentiated to form the products of Mount Mazama. The Mazama edifice was constructed in many comparatively brief episodes, some of the more voluminous being approximately coeval with volcanic pulses in the surrounding region. Magmas as evolved as dacite erupted many times, commonly associated with or following voluminous andesite effusion. It was not until ~30 ka, after 400 kyr of preconditioning, that a shallow rhyodacite-dominated magma reservoir began to grow. Between ~30 ka and the climactic eruption, rhyodacitic magmas vented mainly north of the edifice, unaccompanied by other compositions. From ~35 ka into postglacial time unusually primitive lavas (magnesian basaltic andesite and tholeiitic basalt) erupted west of Mazama suggest increased thermal input. Reconstruction of the Mazama climactic magma chamber from preclimactic rhyodacites and their enclaves and from andesitic scoriae and gabbroic cumulates ejected near the end of the climactic eruption suggests incremental generation of an average of 2.5 km3/kyr of rhyodacite dominantly by crystallization differentiation of basaltic to andesitic magma that was repeatedly intruded between cumulate mush and overlying silicic derivative magma. The mineralogy and elemental and radiogenic isotopic compositions of the granodiorite blocks are similar to those of volcanic rocks of Mount Mazama. Zircon SHRIMP U-Th geochronology gives a range of crystallization ages for granodiorites that correspond to times of dacite effusion. Because erupted silicic magmas were zircon undersaturated, zircon apparently grew late when magmas were mostly crystallized. Variation in crystallization age of blocks from different localities around the caldera suggests a composite pluton underlies Mount Mazama at shallow depth. The granodiorite blocks commonly are partially melted up to 50%, showing that the climactic magma chamber was melting its walls. Retention of zircon antecrysts in pre-35-ka magmas of the granodiorites implies recycling of zircon into high-crystallinity magmas whereas climactic rhyodacite appears to have consumed virtually all zircon from any assimilated wallrock. We conclude that (1) the composite pluton formed from relatively small volumes of derivative magma that crystallized before the upper crust had been heated sufficiently to sustain voluminous convecting crystal-poor melt and (2) that a large volume of eruptible silicic magma accumulated because of heating associated with input of relatively primitive magma to the roots of the system.
V22A-03 11:00h
Tracking Development of the Taupo (New Zealand) Rhyolitic Magma Chamber Through Melt Inclusions
The Taupo Volcanic Centre is one of the most globally productive in terms of magma volumes, and is located in the North Island of New Zealand, in-land from the Hikurangi Trench and along strike from the Kermadec Arc. The largest eruption in the last 30 thousand years was the Oruanui (~300 km3). On the basis of whole-rock and mineral compositions, Sutton et al. (2000; Jl. Geol. Soc. London, 157, 537-552) recognize four post-Oruanui (~26.5 ka) magma types erupted from the Taupo Volcanic Centre; the largest (about 35 km3) of these erupted at 1.77 ka had a short magma chamber residence time (less than 103 years), possibly resulting in a lack of compositional zonation. We have analysed phenocrysts, matrix glasses, and glass (formerly melt) inclusions trapped within the dominant plagioclase, pyroxene, and Fe-Ti oxide phenocryst assemblages of 3 (rhyolitic) of these groups, using electron microprobe (major elements; 5 to 20 micron-diameter spots) and laser ablation, inductively coupled plasma mass spectrometry (LA-ICP-MS) (trace elements; 30 to 40 micron spots). The most significant of these results are: 1. for the trace alkali (Rb, Cs) and alkaline earth elements (Sr, Ba), there are large abundance ranges both within matrix and melt inclusions, by up to factors of two; 2. the range of individual LA-ICP-MS spot analyses encompasses the compositional range (by X-ray fluorescence) of bulk pumices; 3. distinctive (by individual eruption) and positive correlations between the alkalis, alkaline earths and light rare earth elements; 4. even within a specific eruption, there is more than one trace element correlation trend; 5. individual phenocrysts are compositionally zoned requiring some heterogeneity of former host melts. Using the experimentally-constrained criterion of Mn partitioning for equilibrium between coexisting ilmenite-magnetite solid solutions, it is possible to calculate sequentially: the equilibrium T-fO2 of homogenised host magmas (temperature ranges ~ 760 to 860oC; -log10fO2 ~ 15.8 to 12); the Fe2+/Fe3+ and hence Mg/Fe2+ of the host magma; and equilibrium orthopyroxene composition. For the 1.77 ka eruption, the equilibrium orthopyroxene Mg/(Mg+Fe2+) = 0.5 (cf. 0.55 to 0.45 phenocryst range). We interpret these results to indicate the preservation within the melt inclusion population of the post-Oruanui, Taupo eruption sequence, of the compositional record of individual melt components that became homogenized within the larger magmatic system feeding this eruption sequence. Multiple crustal sources must be involved in Taupo Zone rhyolite generation driven by mantle-derived basalt ingress. A physical issue is how the diverse melt inclusion-hosting phenocrysts can be so uniformly distributed in otherwise chemically unzoned bulk (host) rhyolitic magma.
V22A-04 INVITED 11:15h
Do large silicic eruptions leave behind even larger plutons?
Granitic intrusions are widely regarded as the intrusive roots of felsic volcanic systems, but the precise nature of the relationship between intrusive and extrusive magmatism is difficult to determine by direct observation. We recently argued that large plutons commonly form by aggregation of many small intrusive pulses rather than by solidification of large bodies of magma. Plutons with volumes >10,000 km3 and ignimbrites with volumes >1000 km3 each are widespread in the geologic record, and this appears consistent with the widely embraced hypothesis (e.g., Smith, 1979, GSA Spec. Paper 180) that a silicic volcanic eruption leaves behind 10 times its mass in unerupted magma. However, most estimates of the intrusive:extrusive ratio (e.g., Crisp, 1984, JVGR) are based on the integrated geologic histories of long-lived igneous complexes. Such estimates probably should not be applied to individual components of such a system, particularly if the plutons commonly represent several million years of incremental growth. If a magma must be dominantly liquid in order to erupt, the intrusive:extrusive ratio is likely to be greatest during times when the amount of magma in the system is least, and the ratio then decreases as the magma volume increases. When a large silicic crustal magma body does form, it may typically erupt as an ignimbrite rather than solidify to form a pluton. Whether a pluton or an ignimbrite forms mainly would depend on power input (in the form of mantle-derived mafic magma) into the system. When power input is low, only small magma bodies can form and, although probably feeding small eruptions, much of the magma is likely to solidify in situ to form plutons. When power input is high, a larger magma body forms rapidly. The roof of such a body is intrinsically unstable, resulting in catastrophic eruption of a voluminous ignimbrite that is the main geologic record of the magma body. Plutons thus may primarily record processes that operate when a large magma body is absent from a magmatic system.
V22A-05 11:30h
Chemical Variations in Granitic Magmas: Source-inherited or Products of Magmatic Processes?
Though exhibiting considerable scatter in the data, geochemical variations in granitic plutons and silicic volcanic deposits are commonly modelled as the products of differentiation of originally homogeneous magmas. However, many silicic igneous bodies are internally heterogeneous in their mineralogy, geochemistry and isotopic ratios, on a variety of scales, down to 1 m or less, and this variation is mainly inherited from the magma source regions. Granitic plutons and silicic volcanic complexes are therefore commonly constructed by the accumulation of numerous (sometimes quite small) batches (pulses) of magma derived from contrasting source rocks. Such pulses result from the nature of the melting reactions and the physical structure of magma source regions. Internal differentiation of these batches can occur, but most probably not on the scales of whole magma chambers. Rather than being created through differentiation or hybridisation processes, at or near emplacement levels, much of the variation observed within such magmatic bodies (e.g., trace-element or Mg&35; variation with SiO2 or isotopic ratios) is therefore a primary feature. At emplacement levels, the relatively high viscosities and slow diffusion rates of many chemical components in silicic melts probably inhibit processes that would lead to magma homogenisation; this permits at least partial preservation of the primary heterogeneities. These features create difficulties for petrologists trying to date rocks and model chemical variations in igneous bodies. However, these same heterogeneities are a rich source of data on the charcteristics of and processes within the magma source regions.
V22A-06 11:45h
Discovering Eruption Dynamics via Inversion of Tephra Fallout Data
Reconstructing the dynamics of an eruption helps us to understand the eruptive behavior of a specific volcano. When a volcanic eruption cannot be directly or remotely observed and the resulting tephra deposit is not rapidly sampled, very basic parameters, such as plume height, grainsize distribution, erupted mass, and the prevailing wind speed and direction, cannot be determined. These crucial eruption parameters are very difficult to estimate for recent eruptions and practically impossible to resolve for older eruptions. This issue has been previously addressed by interpreting and extrapolating existing data, but recent studies have shown that such data extrapolation is not reliable for the calculation of such important parameters as the erupted mass and the total grainsize distribution. A straightforward answer to this problem is to invert observed physical volcanological data to solve for the eruption parameters directly. This set of modeled eruption parameters is continually adjusted by using the downhill simplex algorithm. Tephra accumulation must be recalculated for multiple geographical locations using a forward solution each time the set of assigned properties is changed. Calculated values are continually compared to the observed data using a goodness-of-fit test until the model parameters no longer change within a specified tolerance between successive iterations. A set of eruption parameters producing a grainsize distribution mimicking the observed distribution results. Parallelizing the inversion problem specifies a more complete solution space of possible eruption dynamics.