Use of volatile and trace elements to infer geodynamic and magmatic processes: Ontong Java Plateau (western Pacific) and Bishop Tuff (eastern California).

摘要

Quenched glassy pillow rims of submarine-erupted lavas and melt (glass) inclusions trapped in crystals in subaerial volcanic rocks preserve important information on pre-eruption volatile contents. In this work, such glasses were analyzed for volatiles and trace elements to infer submarine eruption depths, degassing during magmatic processes, and differentiation processes in a large body of silicic magma.; The Ontong Java Plateau (western Pacific) is anomalous compared to other oceanic large igneous provinces in that it never formed a large subaerial plateau. Paleo-eruption depths (at 122 Ma) inferred from H2O and CO2 in submarine pillow rims suggest that maximum uplift (2500-3600 m above the surrounding seafloor) and post-emplacement subsidence (1500 +/- 400 m) of the plateau were less than predictions from thermal models of oceanic lithosphere. Thus these results are inconsistent with formation of the plateau by a high-temperature mantle plume. Two controversial possibilities to explain this anomaly are: (1) the plateau formed as a result of a giant bolide impact, or (2) the plateau formed from a mantle plume but has a lower crust of dense garnet granulite and/or eclogite. This research shows that neither of these possibilities is fully consistent with all available geological and geophysical data.; In the second part of this research, the composition of matrix glass from rhyolitic pumice was compared with quartz-hosted melt inclusions to establish the origin and crystallization pressures of quartz in the Bishop Tuff (eastern California). The results suggest contrasting origins for quartz in different parts of the Bishop Tuff. In samples from the early to middle parts of the eruption, trace element compositions of matrix glasses are similar to, but slightly more evolved than, quartz-hosted melt inclusions. This result is consistent with in situ crystallization. In contrast, matrix glass in late Bishop Tuff samples is less evolved than quartz-hosted inclusions. The results for late erupted Bishop Tuff are compatible with either crystal settling of quartz from shallower levels within the magma body or late-stage mixing of a new batch of less differentiated rhyolitic magma in the deeper levels of the magma chamber.