ИСТИНА

Zoned plagioclase phenocrysts are ubiquitous in arc volcanic rocks. Their zoning patterns are hold clues to the magmatic processes that shaped their host rocks. An enduring question is whether this information pertains to kinetic processes, changes in intensive variables, open system behaviour, or combinations thereof. The challenge is to unpick the testimony of zoned plagioclase as an archive of pre-eruptive processes occurring in the sub-volcanic magma reservoir. For Mount St Helens (MSH) dacites from the 1980-86 eruption, we have approached the problem using a combination of ion microprobe (SIMS) analysis of zoned plagioclases, experimentally-derived models for plagioclase-melt equilibrium and element partitioning, and numerical models of intracrystalline diffusion. Using phase equilibrium experiments on MSH dacites to parameterise plagioclase composition as a function of pressure (P), temperature (T) and melt fraction (F), we have developed a numerical method to invert zoned MSH plagioclase phenocrysts for their core-rim evolution in P-T space prior to eruption. We remove ambiguity in the family of plausible P-T paths by solving simultaneously for F using the Sr and Ba contents of the plagioclase as measured by SIMS. This approach is valid at MSH because of the relative monotony of magma composition and the absence of significant mafic magma inputs to perturb the bulk compsition. Data from a representative set of crystals from the 1980-86 eruption record the same evolution – an abrupt change from plagioclase core crystallisation at ~12 km depth to rim growth at ~4 km. Simultaneous with crystal growth, trace element diffusion occurs in an attempt to restore chemical potential equilibrium across the zoned plagioclase. Using correlations between anorthite and Sr we show that the cores are >10,000 yrs old, whereas the rims pre-date eruption by ≤3 yrs. The core-rim interface, that is related to the time of magma ascent from 12 to 4 km, can be precisely dated using high spatial resolution SIMS Sr profiles (2 µm spacing). A picture emerges of a long-lived, vertically extensive, mushy, magmatic system that became abruptly destabilised, probably gravitationally, in the months to years before eruption, a timescale commensurate with volcano monitoring. Mush destabilisation appears to be a key process in volcano rejuvenation and eruption triggering.