VARIATIONS OF GLASS COMPOSITION AT THE START OF THE 2006 ERUPTION, AUGUSTINE VOLCANO, ALASKA

Augustine Volcano, one of the most active volcanoes of Alaska, is located in the southwestern part of Cook Inlet. The last major eruption of Augustine occurred in 1986, after which the volcano was quiet for nearly 20 years. Beginning in mid 2005, the number of earthquakes beneath the volcano started to increase. Concurrently, the local GPS network recorded deformation of the volcanic edifice consistent with a gradual inflation with a pressure source centered at shallow depth below the volcano. The first phreatic explosions occurred in mid-December 2005, after which the activity escalated and eventually progressed into a major explosive phase in January 2006.

We studied composition of the volcanic ash erupted in mid December 2005 and in January 2006 using EPMA analysis, SEM, as well as optical petrography. December 2005 ash consists of dense, microlite-rich particles (80-90 vol.%) and microlite-poor, vesicular ash particles (10-20 vol.%). Glass in the vesicular ash particles commonly exhibits signs of devitrification. Glass composition is rhyolitic but highly variable in terms of major oxides, particularly in the concentrations of alkalis, which display a negative correlation with silica. Within this population are some glasses matching the composition of the 1986 glass.

Volcanic ash erupted on January 13, 2006 consists of vesicular ash particles and shards (60-70 vol. %), individual crystals of plagioclase, orthopyroxene (20-30 vol.%), and dense microlite-rich clasts (less than 10 vol. %). The vesicular particles are heterogeneous in terms of microlite content, color of the matrix glass (brown and white), and concentrations of major oxides. The composition of brown matrix glass varies from 68 to 70 wt.% SiO2, whereas the composition of white glass varies from 73 to 78 wt.% SiO2 and matches that of the 1986 volcanic glass.

Our preliminary interpretation is that the explosions in mid-December were mainly phreatic and contained little, if any juvenile component. The chemical trend reflects variable devitrification of 1986 or earlier glass. The ash erupted in January 2006 was mainly juvenile, yielding true melt compositions. The bimodal heterogeneity may be due to introduction of more mafic magma into 1986-like magma just prior to the eruption. This hypothesis will be tested further when lava and tephra samples from the 2006 deposits can be collected and studied.