Magma recharge events and subsequent mixing processes are understood to precede volcanic eruptions. Textural evidence of intrusion of hot, mafic magma into a cooler, rheologically locked silicic magma is commonplace. Solidified "blobs" of injected magma, called enclaves, are evidence of magma mixing, but the petrological and mechanical conditions during their formation are debated. Mount Hood, Oregon consistently erupts andesite bearing compositionally similar enclaves. These enclaves are evidence of mingling and mixing of two magmas. However, due to the compositional similarity between enclave and host lava (e.g. ~1-5 wt.% difference in SiO2), it is unclear whether the preserved enclaves represent; 1) partially hybridized mafic melt remaining after mixing with significant crystal exchange with the host magma or 2) the preserved remnants of the intruding magma during recharge, with no homogenization or crystal exchange with the host magma. The aim of this study is to understand how and why enclaves form in compositionally similar host magmas, such as those at Mount Hood. Building off previous research, we utilize a combination of field observations, chemical analyses, and numerical modeling to constrain the rheology of the magmas prior to and during mixing. The degree of magma mixing is dependent on the viscosity contrast between the host and intruding magmas. Since these magmas are similar compositionally, variations in other magmatic properties such as crystallinity, and therefore temperature, and density may drive the viscosity differences between the host and intruding magmas needed for enclave formation. The enclaves at Mount Hood are vesicular (13-28%), coarse-grained; made up of mainly groundmass crystals (200-450 µm) with sparse microlites (< 200 µm), glass (450 µm) proportions, and rarely contain quenched margins. Additionally, crystals within the host magma show preferential alignment along the margins between host and enclave, suggesting a fluid behavior of the host magma during mixing. Based on textural and compositional evidence, we hypothesize that the intruding magma was buoyant, viscous, and crystalline, due to decompression-induced crystallization and exsolution of volatiles, during recharge and ascent to the shallow magma reservoir. Injection and underplating of the viscous crystalline intruding magma into a hot convecting host magma induces enclave formation. Crystallization temperatures differ by only 6-15 °C between host and enclave lavas, derived by the two pyroxene geothermometry method by Putrika (2008). These crystallization temperatures are consistent with crystallization in compositionally similar magmas. However, with such similar crystallization and liquidus temperatures, maintaining a viscosity contrast between the mixing magmas for enclave survival after formation suggests other properties, apart from temperature, must explain the viscosity contrast needed for enclave survival after enclave dispersal and thermal equilibration occurs. The presence of bubbles, from exsolution during crystallization, within the enclave magma increases the viscosity while simultaneously decreasing the density. Therefore, the presence of bubbles increases the viscosity of the intruding magma and maintains the viscosity contrast during the mixing process after thermal equilibration occurs. Additionally, if degassing occurs, rapid crystallization maintains the high viscosity of the enclaves. The enclaves observed at Mount Hood represent the solidified remnants of the last recharge event prior to eruption. The presence of compositionally similar enclaves and host lavas suggest a transient precursor event just prior to eruption at Mount Hood and can be applied to other recharge-driven arc volcanic systems.

Large-volume silicic ignimbrites erupt from reservoirs that vary in composition, temperature, volatile content and crystallinity. The 9.7 Ma Devine Canyon Tuff (DCT) of eastern Oregon is a large-volume (>250 km3), compositionally zoned and variably welded ignimbrite. The ignimbrite exhibits heterogeneous trace element compositions, variable volatile content and crystallinity. These observations were utilized in the investigation into the generation, accumulation and evolution of the magmas composing the DCT. Building off previous research, pumices were selected from the range of trace element compositions and analyzed with respect to crystallinity, mineral abundances and assemblages. The DCT displays a gradational trace element enrichment and decrease in crystallinity from least evolved, dacite, at ~22% crystals to most evolved high-silica rhyolite at 3% crystals. Two distinct mineral populations of feldspar and clinopyroxene were identified in previous work, one belonging to the rhyolitic magma and the other to the dacitic magma. Volatile content derived from melt inclusion Fourier Transform Infrared (FTIR) spectrometer analysis revealed an increase in water content from 1.2 to 3.7 wt.% in the most evolved rhyolite. The DCT exhibits low and variable δ18O signatures, 4.52‰ to 5.76‰ , based on δ18O values measured on quartz and sanidine. Low δ18O signatures of all DCT rhyolites suggest the incorporation of hydrothermally altered crust into the melt. Furthermore, quartz phenocrysts from all high-silica rhyolite groups display dark oscillatory zoned cores and Ti-rich bright rims. These data provide insight into how these magmas were generated and subsequently stored in the crust. Commonalities of petrographic and compositional features among rhyolites, especially the zoning characteristics of quartz phenocrysts, exclude the possibility of storage and evolution in multiple reservoirs. Envisioning a scenario where all magmas are stored within a single reservoir prior to eruption and assuming rhyolites A and D are the product of partial melting. The mixing of A and D rhyolites produced rhyolite B, and subsequent mixing of intermediate rhyolite B and end-member rhyolite D generated rhyolite C. However, some trace element inconsistencies, between mixing model and observed intermediate rhyolites suggest a secondary process. Post mixing, rhyolites B and C require some modification by fractional crystallization to account for LREE and other inconsistencies between mixed models and observed rhyolites. Finally, the origin of the dacite is likely through mixing of group D rhyolite and an intrusive fractionated basalt, which could have led to the eruption of the Devine Canyon Tuff.

This study is an effort to characterize the magma sources, plumbing system, and eruptive behavior of Mount Hood, a low-explosivity recharge-dominated volcano in the Oregon Cascades. The three manuscripts in this dissertation make use of melt inclusion data, phenocryst compositions, and whole rock petrology and geochemistry to build a schematic model of plumbing, mixing, and eruption at Mount Hood. Volatile contents in melt inclusions were measured by Fourier Transform Infrared Spectroscopy (FTIR) and Secondary Ion Mass Spectometry (SIMS). These measurements indicate that the pre-eruptive volatile contents at Mount Hood are comparable to concentrations in more explosive volcanoes, and do not sufficiently explain the low explosivity of Mount Hood. Measured H₂O contents were also used to test the validity of multiple different hygrometers. Various geothermobarometers were applied to the melt inclusions and phenocrysts from Mount Hood, and demonstrate that pre-eruptive temperatures increase by 100-150 ̊C immediately after mafic recharge, which occurs days to weeks prior to eruption and is accompanied by a 5-10 fold decrease in magma viscosity. Numerical simulations of magma ascent indicate that magma fragmentation is significantly delayed with this magnitude of pre-eruptive heating, which reduces the likelihood of explosive eruption. Analyses of amphibole demonstrate two markedly different populations, which correspond to different magma compositions, temperatures, and pressures. Pressure and temperature calculations were compared to other geothermobarometers to crosscheck the validity of these results and generally agreed well. Trace element concentrations in lavas, enclaves, and inclusions from Mount Hood confirm previous models for simple binary mixing at Mount Hood. A linear regression technique for extrapolating the major element contents of the mixing endmembers works acceptably well to characterize the trace element budgets of these endmembers. Additionally, we observe that the "recharge filter" that is responsible for the compositionally monotonous lavas at Mount Hood is also the likely cause of long-term low explosivity, and is indicative of a two-part magma plumbing system that may be a general model for a number of other recharge-dominated arc volcanoes. The results presented in this dissertation, in concert with previous results by other authors, converge on a generally consistent model for the production, hybridization, and eruption of intermediate lavas at Mount Hood and elsewhere. / Graduation date: 2012 / Access restricted to the OSU Community at author's request from Sept. 16, 2011 - March 16, 2012