[S-VC32] Dynamics of volcanic eruptions and their physical and chemical processes

[SVC32-P03] Shallow conduit process of mafic subplinian eruption inferred from the groundmass texture of the B scoria of Izu-Oshima 1986 eruption

*Akinori Miyashita¹, Hidemi Ishibashi², Natsumi Hokanishi³, ATSUSHI YASUDA³ (1.Shizuoka University, 2.Faculty of Science, Shizuoka University , 3.Earthquake Research Institute, University of Tokyo)

Keywords:Mafic magma, Subplinian eruption, Microlite, Shallow conduit process, Izu-Oshima volcano

The 1986 eruption is the latest VEI-3 eruption at Izu-Oshima volcano. In this eruption, a subplinian eruption of aphyric basaltic andesite magma occurred in the newly opened B-vents. In this study, optical observation, and textural and chemical analyses of the groundmass were carried out for the scoria from the B vents (referred to as the B scoria) to investigate the shallow conduit process which induced the subplinian eruption. The groundmass of the glassy B scoria consists of plagioclase and clinopyroxene microlites and glass, and the total crystallinity varied in the range of ~11-50 vol%. The microlite number densities of both plagioclase and clinopyroxene increase as the crystallinity of each mineral increases. No preferred orientation of microlites was observed. The residual glasses in the groundmass have a basaltic andesite composition with SiO₂ content of ~54.8-58.3 wt%. As crystallinity increases, the SiO₂, TiO₂, and FeO* contents increase while the contents of Al₂O₃, MgO, and CaO decrease. Assuming an eruption temperature of 1100 °C for the 1986 B eruption, the water content and viscosity of the residual melt are estimated to be ~1.0-1.9 wt% and ~10^0.3-10^5.7 Pa s, respectively. Furthermore, the viscosity of magma considering the influence of microlite was estimated to be ~ 10^2.2-10^8.3 Pa s. The relationship between the melt water content and the crystallinity is distinct between the samples with the crystallinity < 25 % (microlite-poor) and those with the crystallinity > 25 vol% (microlite-rich), but negative correlations were observed in both samples. The magma decompression rate of the B eruption was estimated to be ~ 10^-1.6-10^-0.1 MPa s^-1 and showed a positive correlation with the melt water content. The microlite-rich and poor samples show a different relationship between the decompression rate and the crystallinity, but the decompression rate decreases as the crystallinity increases in both samples.
The total crystallinity of scoria varies from lower to higher values compared to the critical crystallinity at which the liquid-solid transition occurs. In addition, the viscosity of plagioclase-bearing magma is estimated to be 10⁶ Pa s at the crystallinity of ~30 vol%. These results suggest that the fragmentation mechanism is different between the microlite-poor and rich scoria. For the microlite-poor magma, inertia fragmentation occurred and it erupted as a liquid. On the other hand, degassing-driven crystallization induced the liquid-solid transition, and as a result, the solid-like fragmentation occurred in the microlite-rich magma. This is supported by the observation that bubbles in microlite-rich samples are distorted, while bubbles in the microlite-poor samples are round. The lava fountains and the secondary lava observed in the 1986B eruption were attributed to the magma erupted as a liquid, whereas the magma erupted as a solid may have contributed to the formation of eruption columns. The relationship between the melt water content, decompression rate, and crystallinity is consistent with the interpretation that the fragmentation mechanism may have changed at the crystallinity ~25 vol%. In addition, the relationship suggests that the crystallinity increases as the magma ascent and that magma fragmentation occurred at shallower depth as the magma ascent rate decrease.