Recharge, decompression, and collapse : dynamics of volcanic processes

Non-linear volcanic and magmatic processes control the occurrence and behavior of volcanic eruptions. Consequently, understanding the responses of volcanic systems to processes of different length scales, timescales, and magnitudes is critical to interpreting ancient deposits, understanding current eruption dynamics, and predicting future activity. Here I present the results of three studies wherein analytical geochemistry, experimental petrology, and turbulent flow analysis describe otherwise obscured volcanic processes. Injections of new magma are common events in magma chambers. Recharging magma can change the chamber composition and temperature and may facilitate assimilation of country rock. Plagioclase phenocrysts provide an opportunity to examine recharge and assimilation processes, because their compositions are sensitive to temperature and their Sr isotopic ratios can record compositional variations in the chamber. Chemical and isotopic microanalyses of crystals from 7 eruptions of El Chichón Volcano, Mexico, reveal that recharge and assimilation events are very common and mixing is efficient, but individual events seldom affect the entire chamber. During every eruption, magma decompresses and ascends through a conduit from a chamber at depth to a vent at the surface. Changes in pumice textures during the 1800 ¹⁴C yr BP eruption of Ksudach Volcano, Kamchatka, suggest that conduit structure changed following caldera collapse. Decompression experiments show that the post-collapse pumice decompressed at ~0.0025 MPa/s, compared to pre-collapse decompression rates of >0.01 MPa/s. By balancing those results with eruptive mass fluxes I quantify the effects of caldera collapse on a conduit, and show that collapse resulted in a conduit with a very broad base and narrow vent. Turbulent air entrainment controls whether an eruption column rises buoyantly or collapses to generate pyroclastic flows. Through extensive re-evaluation of video and photographs of the 18 May 1980 eruption of Mount St. Helens, I report the first measurements of the turbulent velocity field of a volcanic column and show that changes in its turbulence reflect changes in eruption behavior. Those results indicate collapse was caused by a reduction in eddy size and turbulent air entrainment initiated by an increased vent size and the development of a buoyant annulus surrounding a dense, collapsing core.