11:00 AM - 1:00 PM
[SVC32-P01] Experimental constraints on the H2O-saturated plagioclase liquidus and the storage depth of the Izu-Oshima 1986B basaltic andesite melt
Keywords:Izu-Oshima volcano, pre-eruptive condition, plagioclase liquidus, magma chamber, high P-T experiment
The latest eruption of VEI > 3 occurred in AD 1986 at Izu-Oshima volcano. The notable feature of this eruption event is that a sub-Plinian eruption occurred at the newly opened vents on the caldera floor (B vents). The magma erupted from the B vents (B magma) is almost aphyric with small abundances of plagioclase and pyroxene phenocrysts. Most of the magma has basaltic andesite composition with SiO2~54.5 wt%. The composition of the B magma is different from the major basaltic magmas erupted in this volcano, including the A magma erupted from the summit vent. In addition, the compositional gap is found between the B and the A magmas. Therefore, the chamber of the B magma is thought to be separated from the magma plumbing system of the A magma. Previous geophysical studies proposed that the chamber of the B magma is located at the depth of ~3-5km. However, the petrological constraint on the storage condition of the B magma is not enough. The H2O-saturated plagioclase liquidus (HSPL) is one of the most useful tools to constrain the pre-eruptive storage condition of the magma. Using the HSPL, the H2O content (or the pressure) condition can be constrained from the temperature and the major element composition of a melt. However, the HSPL is model-dependent; we need a reliable model to constrain the pre-eruptive storage condition of the B magma. In this study, high-temperature equilibrium experiments were carried out at pressures of 1-atm to 196 MPa and under H2O-saturated conditions on the B melt of the Izu-Oshima 1986 eruption to constrain the HSPL and the storage pressure condition of the melt.
We used the powder of the B lava as the starting material of our experiments. High-temperature equilibrium experiments were carried out for the sample at 1-atm and high-pressure conditions (49, 98 and 196MPa) using a 1-atm fO2-controlled furnace and internally heated pressure vessel at the GSJ, AIST, respectively. For the 1-atm experiments, the fO2 condition was controlled at Ni-NiO buffer using the H2-CO2 mixed gas, and the Pt-wire method was adopted. The samples were heated at the experimental temperatures (1150, 1165, 1180, 1200ºC) for 3 hours and then quenched. For the high-pressure experiments, H2O was saturated in the sample, but the fO2 condition was not controlled. The samples enclosed with H2O in the Ag50Pd50 capsules were heated in the pressure vessel for 3 hours at the experimental temperatures ranging from 1050 ºC to 1130 ºC and then quenched. The quenched samples were polished to thin sections and petrological observations and chemical analyses of glasses were performed using EPMA and FE-EPMA at ERI, University of Tokyo, Japan.
Plagioclase is the liquidus phase at 1 atm, whereas early saturation of Fe-Ti oxide above the plagioclase liquidus occurred in the high-pressure experiments due to the elevated fO2 conditions. The HSPL temperature decreases from 1172 ± 8ºC to 1030 ± 20ºC as the pressure increases from 1atm to 196 MPa. A combination of previously proposed models for the plagioclase liquidus and melt H2O-solubility can predict the experimentally determined HSPL temperatures, even if oxidation-induced magnetite crystallization occurs. On the other hand, the thermodynamic equilibrium calculation program “rhyolite-MELTS” failed to predict the HSPL at high-pressure conditions although it is successfully predicted at 1 atm.
Using the combination of previously proposed models for the plagioclase liquidus and melt H2O-solubility models and the previously reported eruption temperature of ~1100 ± 30ºC, we estimate the pre-eruptive pressure conditions of the B melt to be 42 (+48/-32) MPa, which correspond to depths of 1.9 (+1.9/-1.4) km. The estimated depth is consistent with that of the shallow active dikes previously identified from geophysical studies. The result suggests that the B melt was derived from a small, shallow magma chamber formed in the shallow dike region that is shallower than 3-5 km if the assumed temperature condition is valid.
We used the powder of the B lava as the starting material of our experiments. High-temperature equilibrium experiments were carried out for the sample at 1-atm and high-pressure conditions (49, 98 and 196MPa) using a 1-atm fO2-controlled furnace and internally heated pressure vessel at the GSJ, AIST, respectively. For the 1-atm experiments, the fO2 condition was controlled at Ni-NiO buffer using the H2-CO2 mixed gas, and the Pt-wire method was adopted. The samples were heated at the experimental temperatures (1150, 1165, 1180, 1200ºC) for 3 hours and then quenched. For the high-pressure experiments, H2O was saturated in the sample, but the fO2 condition was not controlled. The samples enclosed with H2O in the Ag50Pd50 capsules were heated in the pressure vessel for 3 hours at the experimental temperatures ranging from 1050 ºC to 1130 ºC and then quenched. The quenched samples were polished to thin sections and petrological observations and chemical analyses of glasses were performed using EPMA and FE-EPMA at ERI, University of Tokyo, Japan.
Plagioclase is the liquidus phase at 1 atm, whereas early saturation of Fe-Ti oxide above the plagioclase liquidus occurred in the high-pressure experiments due to the elevated fO2 conditions. The HSPL temperature decreases from 1172 ± 8ºC to 1030 ± 20ºC as the pressure increases from 1atm to 196 MPa. A combination of previously proposed models for the plagioclase liquidus and melt H2O-solubility can predict the experimentally determined HSPL temperatures, even if oxidation-induced magnetite crystallization occurs. On the other hand, the thermodynamic equilibrium calculation program “rhyolite-MELTS” failed to predict the HSPL at high-pressure conditions although it is successfully predicted at 1 atm.
Using the combination of previously proposed models for the plagioclase liquidus and melt H2O-solubility models and the previously reported eruption temperature of ~1100 ± 30ºC, we estimate the pre-eruptive pressure conditions of the B melt to be 42 (+48/-32) MPa, which correspond to depths of 1.9 (+1.9/-1.4) km. The estimated depth is consistent with that of the shallow active dikes previously identified from geophysical studies. The result suggests that the B melt was derived from a small, shallow magma chamber formed in the shallow dike region that is shallower than 3-5 km if the assumed temperature condition is valid.