|Abstract or Summary
- Crystal-rich (40-50 vol.%) intermediate lava has been the primary eruptive product of several recent hazardous eruptions: Mt. Pinatubo, Philippines (1991), Soufriere Hills, Montserrat (1995-present), and Unzen, Japan (1990-1995). Despite this association with such devastating eruptions, the formation, timing, and evacuation of such magma is not well understood: do such eruptions tap a long-lived, multi-cycle crystal mush, or, is it generated in a single magmatic cycle prior to eruption?
This thesis explores this question through research at the Unzen Volcanic Complex (UVC), southwestern Japan, where a 500 ka history of crystal-rich dacitic dome eruptions has built the Unzen Volcanic Complex. Previous studies (e.g. Nakamura, 1995) have demonstrated the role of mafic recharge in rejuvenating crystal mush zone a few months prior to the most recent eruptions, but - until now - its formation and evolution have not been investigated using detailed geochronology. Now, this gap has been addressed through U/Th zircon geochronology and zircon chemistry studies (zircon chronochemistry) on several eruptions, spanning in age from the most recent (Heisei-Shinzan; 1990-1995) to lavas of the Older Unzen period (≥200 ka).
Zircon geochronology has revealed the protracted history and evolution of the crystal mush zone at the Unzen Volcanic Complex. Individual zircon surface-interior age pairs, together with zircon age spectra, suggest that portions of this crystal mush have been present in the crust since at least the late Older Unzen period (≥200 ka). Significant zircon growth appears to occur throughout the eruptive hiatus (200-100 ka), suggesting that magmatic activity continued uninterrupted through this period.
Zircon chemistry shows that UVC crystals are typical of unaltered, igneous zircon, and - overall - zircon crystals show a restricted range in trace element chemistry in all eruption samples. Ti-in-zircon temperatures show that zircon crystalizes from a low temperature (≤790˚C) magma, in agreement with previous temperatures estimates for UVC upper-crustal magma (e.g. Venezky and Rutherford 1999). Subtle variations in trace element chemistry are observed in titanium and hafnium content through time: titanium contents show a moderate negative correlation with time, while hafnium shows a moderate positive correlation. This suggests that the crystal mush zone has evolved towards a more mature, heterogeneous system. Additionally, Older Unzen eruption samples show restricted chemistry, suggesting that crystallization occurred in a more homogenous, less evolved and possibly warmer crystal mush, compared to zircon of Younger Unzen (100 ka - present) that record variable storage conditions skewed towards a more evolved, poorly-mixed crystal mush. Complex age populations, particularly from samples of the Younger Unzen period, suggest that localized regions within the crystal mush may have different thermal histories. These results lead to the conclusion that eruptions at Unzen are tapping a mature and long-lived, multi-cycle mush of significant longevity.