Groundmass crystallization of Mount St. Helens dacite, 1980?1986: a tool for interpreting shallow magmatic processes

Abstract

The 1980–1986 eruption of Mount St. Helens volcano provides an unprecedented opportunity to observe the evolution of a silicic magma system over a short time scale. Groundmass plagioclase size measurements are coupled with measured changes in matrix glass, plagioclase and Fe−Ti oxide chemistry to document increasing groundmass crystallinity, and thus to better constrain proposed physical models of the post-May 18, 1980 magmatic reservoir. Measurements of plagioclase microlite and microphenocryst sizes demonstrate that relatively rapid growth (approximately 10-9 cm/s) of groundmass plagioclase occurred immediately subsequent to May 18. Relatively rapid plagioclase growth continued through the end of 1980 at an average rate of 3x10-11 cm/s; plagioclase growth rates then decreased to <1x10-11 cm/s through 1986. Changes in groundmass crystallinity are reflected in changes in both matrix glass and plagioclase microphenocryst-rim chemistry, although the matrix glass composition appears to have remained approximately constant from 1981–1986 after a rapid compositional change from May 18 until the end of 1980. Plagioclase microphenocrysts show increasingly more complex zoning patterns with time; microphenocryst-core compositions are commonly positively correlated with crystal size. Both of these observations indicate continuous groundmass plagioclase growth through 1986. Magmatic temperatures estimated from Fe−Ti oxide pairs are approximately constant through 1981 at eruption temperatures of ∼ 930°C and at log fO2 of -10.8; by 1985–1986 oxide temperatures decreased to ∼ 870°C. Chemical and textural changes can be explained by: (1) rapid degassing and crystallization in response to the intrusion of magma into a shallow (<4.5 km) reservoir toward the end of the May 18, 1980 eruption; (2) continued crystallization at a much reduced rate through 1986 due to slow cooling of the shallow magma reservoir. Growth rates (and consequent chemical changes) appear to decrease at the end of 1980—this is coincident with the change in eruption style from explosive eruptions, sometimes followed by dome growth, to solely extrusive (dome-building) events, and can be explained by the expected viscosity increase of both degassing and increasing crystallinity. The model of twostage crystallization of magma in a shallow reservoir is consistent with conclusions from gas studies (Casadevall et al. 1983; Gerlach and Casadevall 1986 a, b), patterns of crater deformation (Chadwkck et al. 1988) and post-1980 seismicity (Endo et al. 1990), although it does not explain the experimental data of Hill and Rutherford (1989) on the growth rate of amphibole reaction rims. Textural measurements on Mount St. Helens dacite can also be used to evaluate crystallization kinetics in silicic magmas, systems for which experimental data is almost non-existent. Plagioclase growth rates are 5–10 times slower than estimated plagioclase growth rates in basaltic systems, a result consistent with the higher viscosity of a more silicic melt. Furthermore, patterns of textural change (both average crystal size and number density) are similar to those observed during the 1984 Mauna Loa eruption by Lipman and Banks (1987), suggesting that the only modification to the crystallization behavior of plagioclase required in extrapolation from basaltic systems is a moderate decrease in rates, such that the rate of crystallization scales with the melt viscosity.