5:15 PM - 6:45 PM
[SVC31-P07] Pre-eruptive process of the smallest caldera-forming eruption: a case study of Ikeda caldera, South Kyushu, Japan
Keywords:Ikeda caldera, Caldera-forming eruption, Pre-eruptive process, Funnel-type caldera, Caldera collapse, Pumice
The 6.4ka Ikeda caldera eruption at Ikeda volcano, south Kyushu, is the smallest one among the caldera-forming eruptions. The total volume of the erupted materials during the eruption is estimated to be ~5km3(VEI5). Understanding the pre-eruptive process of the smallest caldera-forming eruption is important to constrain the threshold condition to induce caldera collapse. In this study, we performed a stratigraphic description of the pumice fall (Ikeda pumice fall) and pyroclastic flow (Ikeda pyroclastic flow) deposits and textural and chemical analyses of the white pumices, which is the most abundant ejecta of the deposits. Based on the results, we discuss the pre-eruptive process of the Ikeda caldera eruption and the threshold condition of caldera collapse.
The Ikeda pumice fall deposit is subdivided into three units based on the grain size of the pumice clast and the amount of lithic fragments: Unit-1 is composed of slightly fine-grained pumice, Unit-2 is characterized by coarse-grained lithic fragments, and Unit-3 is thick deposits of coarse-grained pumice.The white pumice contains plagioclase (~15vol%), quartz (~3vol%), hornblende (~2vol%), orthopyroxene (~1vol%), and Fe-Ti oxides (~1vol%) as phenocryst; the modal abundance is constant through the three units. The groundmass (~78vol%) is glassy without microlite. Chemical compositions of Fe-Ti oxides, amphibole, and groundmass glass are almost homogeneous through all the deposits.
Xusp [=2Ti/(Ti+Fe3+)] of magnetite and Xilm [=2Ti/(Ti+Fe3+)] of ilmenite were ~0.21 and ~0.85, respectively. Mn/Mg partitioning between the two phases indicates they are in equilibrium. 96% of amphibole grains are homogeneous with Si~7 (apfu) and Mg/(Mg+Fe2+) ~0.85 and classified into magnesiohornblende. Groundmass glass and quartz-hosted melt inclusions show identical rhyolitic composition with SiO2 content ~79wt%.Fe-Ti oxide thermometry-oxybarometry indicates equilibrium T-fO2 conditions of the rhyolite magma at the eruption are ~773°C and ΔQMF ~2.2, respectively. Assuming H2O-saturation, the pre-eruptive pressure and melt H2O content conditions are estimated to be ~220 MPa and ~6.7wt%, respectively, based on the plagioclase liquidus combined with the estimated temperature. The estimated T-H2O content conditions are consistent with the results of amphibole thermometry and melt-SiO2-metry. The estimated pressure corresponds to a depth of about 8.3km, assuming a crustal density of ~2700 kg/m³. The results suggest that the pre-eruptive magmatic conditions were constant throughout the eruption and the conduit ascent process controlled the style and intensity of the eruption.
The threshold volume of erupted magma required for caldera collapse is calculated for Ikeda caldera eruption using the model of Geshi et al. (2014). However, the calculated volume was more than twice the previously estimated total magma volume of Ikeda eruption. The possible causes of this contradiction may be (1) the underestimation of the erupted magma volume and/or (2) the difference in calder-forming mechanism. To examine the latter possibility, we examined the type of Ikeda caldera based on the ratio of caldera width/magma chamber depth (Roche, 2000), and the result indicates that the Ikeda caldera is a funnel-type. In addition, we investigated the relations between the erupted volumes and the caldera types for a dozen calderas and found that the total magma volumes for funnel-type calderas are lower than the threshold ones calculated from the model of Geshi et al. (2014). This suggests that funnel-type calderas may cause a caldera collapse with a smaller amount of magma volume than piston-cylinder-type calderas.
The Ikeda pumice fall deposit is subdivided into three units based on the grain size of the pumice clast and the amount of lithic fragments: Unit-1 is composed of slightly fine-grained pumice, Unit-2 is characterized by coarse-grained lithic fragments, and Unit-3 is thick deposits of coarse-grained pumice.The white pumice contains plagioclase (~15vol%), quartz (~3vol%), hornblende (~2vol%), orthopyroxene (~1vol%), and Fe-Ti oxides (~1vol%) as phenocryst; the modal abundance is constant through the three units. The groundmass (~78vol%) is glassy without microlite. Chemical compositions of Fe-Ti oxides, amphibole, and groundmass glass are almost homogeneous through all the deposits.
Xusp [=2Ti/(Ti+Fe3+)] of magnetite and Xilm [=2Ti/(Ti+Fe3+)] of ilmenite were ~0.21 and ~0.85, respectively. Mn/Mg partitioning between the two phases indicates they are in equilibrium. 96% of amphibole grains are homogeneous with Si~7 (apfu) and Mg/(Mg+Fe2+) ~0.85 and classified into magnesiohornblende. Groundmass glass and quartz-hosted melt inclusions show identical rhyolitic composition with SiO2 content ~79wt%.Fe-Ti oxide thermometry-oxybarometry indicates equilibrium T-fO2 conditions of the rhyolite magma at the eruption are ~773°C and ΔQMF ~2.2, respectively. Assuming H2O-saturation, the pre-eruptive pressure and melt H2O content conditions are estimated to be ~220 MPa and ~6.7wt%, respectively, based on the plagioclase liquidus combined with the estimated temperature. The estimated T-H2O content conditions are consistent with the results of amphibole thermometry and melt-SiO2-metry. The estimated pressure corresponds to a depth of about 8.3km, assuming a crustal density of ~2700 kg/m³. The results suggest that the pre-eruptive magmatic conditions were constant throughout the eruption and the conduit ascent process controlled the style and intensity of the eruption.
The threshold volume of erupted magma required for caldera collapse is calculated for Ikeda caldera eruption using the model of Geshi et al. (2014). However, the calculated volume was more than twice the previously estimated total magma volume of Ikeda eruption. The possible causes of this contradiction may be (1) the underestimation of the erupted magma volume and/or (2) the difference in calder-forming mechanism. To examine the latter possibility, we examined the type of Ikeda caldera based on the ratio of caldera width/magma chamber depth (Roche, 2000), and the result indicates that the Ikeda caldera is a funnel-type. In addition, we investigated the relations between the erupted volumes and the caldera types for a dozen calderas and found that the total magma volumes for funnel-type calderas are lower than the threshold ones calculated from the model of Geshi et al. (2014). This suggests that funnel-type calderas may cause a caldera collapse with a smaller amount of magma volume than piston-cylinder-type calderas.