CO2 contents of basaltic arc magmas from the southern Cascades: Corrections for shrinkage bubble effects and implications for crustal storage

Friday, 18 December 2015: 16:30
308 (Moscone South)
Kristina J Walowski1,2, Paul J Wallace1, Ellen Marie Aster1 and Michael A Clynne3, (1)University of Oregon, Eugene, OR, United States, (2)University of Edinburgh, School of Geosciences, Edinburgh, United Kingdom, (3)USGS California Water Science Center Menlo Park, Menlo Park, CA, United States
Volatiles such as H2O and CO2 play an important role in a variety of magmatic processes from magma generation to eruption, and melt inclusions (MI) – small volumes of melt trapped inside phenocrysts – have been used to measure their pre-eruptive concentrations. In particular, the volatile contents of MI from basaltic arc magmas have been used to track the role of dehydrating subducted oceanic lithosphere in magma formation in subduction zones. However, recent studies have shown that MI are imperfect storage containers and can lose H by diffusion through the mineral host and CO2 due to formation of a vapor bubble in the inclusion. Such results suggest that even the least degassed melt inclusions from a volcano may have volatile concentrations that underestimate the initial volatile contents of the magma. Thus, recognizing pre- and post-entrapment processes that influence MIs is important for interpreting magmatic processes at depth. Recent studies have developed methods that can be used to distinguish and correct for H diffusive loss (Bucholz et al., 2013) and CO2 loss to vapor bubbles (Wallace et al., 2015). Here, we focus on MI from eight cinder cones that erupted primitive basaltic magmas in the Lassen region of the Cascade arc, where H2O and Cl concentrations have been shown to relate to the amount of a subduction component added to the mantle wedge (Walowski et al., 2015). Using methods of Aster (2015), we correct for the loss of CO2 to a vapor bubble formed within a melt inclusion as the result of post-entrapment crystallization and thermal contraction. The results of the CO2 restoration calculations suggest that ~25-75% of the initial dissolved CO2 in the melt inclusions at the time of trapping was lost to a vapor bubble after entrapment. Trapping pressures for the restored CO2 and maximum H2O contents calculated using methods of Iacono-Marziano et al. (2012) range from ~2-5 kbar, equivalent to entrapment depths of ~7-18 km below the surface. The results indicate that primitive magmas in the southern Cascades stall in the middle crust, where they crystallize olivine and trap melt inclusions. Lastly, we consider the relationships of restored CO2 contents to highly incompatible elements like Nb and compare our results to data for basaltic magmas from other tectonic environments.