Evidence of Phenocryst Growth and Magma Mixing during Ascent along Fractures in Volcanic Rocks from Basalt to Rhyolite

Abstract

Phenocrysts in volcanic rocks offer the unique opportunity to constrain the temperature, pressure, water content and oxidation state of their host melt at the time of their growth. This thesis presents several lines of evidence that phenocryst growth and magma mixing in erupted magmas, ranging from basalt to rhyolite, largely occurred during ascent along fractures and not in stalled magma reservoirs, and that most erupted magmas begin their ascent to the surface as largely crystal-free melts. The results from this thesis show that the most successful application of mineral-melt thermometers and hygrometers is at the liquidus of ascending melts. In Chapter 2, ten intraplate, alkaline basalts, which erupted within a rift zone in the western Mexican arc, are examined. Despite a wide range of MgO contents (5-10 wt%), all contain Mg-rich olivine phenocrysts (~Fo89-90) with diffusion-limited, rapid growth textures. Application of olivine-melt thermometry and hygrometry at the liquidus is only successful for basalts with >9 wt% MgO. For the remaining samples, all available evidence points to magma mixing between high-MgO (>9 wt%) and low-MgO (<5 wt%) melts immediately prior to eruption, most likely during ascent along fractures. Thus, it is not only phenocryst growth, but also mixing of distinctly different melts, that can occur during magma ascent along intersecting fractures. Chapter 3 examines the only known dike swarm of high-SiO2 (75-77 wt%) rhyolite, with near-eutectic compositions (i.e., liquidus-solidus intervals of ≤ 50°C), that successfully escaped its granitoid source region. The dike swarm (4 x 27 km) crosscuts metasedimentary hornfels units in eastern California. Most dikes range from 2-3 m across and can be individually traced for up to 4 km. The dikes contain large, sparse tectosilicate phenocrysts, often with granophyric and/or diffusion-limited growth textures, in a fine-grained groundmass. Application of quartz-melt thermometry and plagioclase-melt hygrometry, respectively, gives temperatures of ~710-770°C and water contents of ~7.0-5.4 wt%. This enables the average density (~2.2 g/cm3) and viscosity (~4.9 log10 Pa-s) of the melts to be calculated, along with their velocity during transit through the crust. On average, for dikes of ~2-3 m width, ascent rates are ~2-4 km/day, which leads to >10 km of transport through the upper crust within < 1 week. All evidence precludes phenocryst growth in these near-eutectic, hydrous melts in a stalled magma reservoir, and instead requires that it took place during their ascent as dikes through the upper crust. In Chapter 4, an experimental study was conducted on a low-MgO basalt with large (≤8 mm) plagioclase phenocrysts to examine the effect of rapid phenocryst growth on plagioclase-melt disequilibrium. There is growing evidence in the literature that most phenocrysts in erupted basalts grew rapidly, following a kinetic delay in nucleation, and the question is whether disequilibrium develops that affects the application of plagioclase-liquid hygrometry at the liquidus. Both equilibrium and dynamic crystallization experiments were performed, where the latter involved a kinetic delay to nucleation, followed by high crystal growth rates (~1-4 mm/day). Application of the experimental results to the natural basalt confirms that its plagioclase phenocrysts must have crystallized after a kinetic delay to nucleation, and that growth of its largest phenocrysts occurred within <2-8 days. Application of the plagioclase-liquid hygrometer at the liquidus of this natural basalt, owing to the kinetic delay in nucleation and subsequent undercooling, slightly overestimates melt water content by <0.7 wt%.