2:15 PM - 2:30 PM
[SVC31-02] Factors affecting the sequence of Asama volcano 1783 eruption inferred from rock texture
Keywords:Asama Volcano, eruptive style, magma ascent process
Plinian eruption is characterized by a large, sustained column, and often changes into pyroclastic flow eruptions. Determining how the eruptive style changes is important for predicting such volcanic activity. Processes during magma ascent such as crystallization, degassing, and vesiculation are reflected in rock texture, thus analyses of rock texture are important to constrain magma ascent process. In this study, we focus on the 1783 eruption of Asama Volcano, and discuss the mechanism of eruption transition through analysis of chemical composition and observation of rock texture. In the climactic phase of 1783 eruption, Plinian column was generated and pumice-fall deposits are formed at first, then eruptive style changed to pyroclastic flow later. Pumice-fall deposits are divided into 14 layers. Upper half of deposits were formed in climactic phase. In the middle level of deposits, there are some reddish-brown silt layer which might come from pyroclastic flows before climactic phase. Each layer in the middle level contain more than 10 wt.% of juvenile lava fragments.
The 14 layers of Pumice-fall deposits and two types of pyroclastic flow deposits are chosen for measurement of density, microscopy, and chemical analysis. The average apparent densities range 0.7-0.8 g/cm3 for fall deposits, 0.8-0.9 g/cm3 and >1.20 g/cm3 for pyroclastic flow deposits. Groundmass glass is heterogeneous in some cases, and SiO2 content shows bimodal distribution around 67 wt.% and 73 wt.%. Pyroclastic flow deposits have lower SiO2 glass than pumice-fall deposits. Magma temperature (T) and water content (Cw) are estimated based on the compositions of plagioclase and glass matrix (Putirka, 2008; Waters and Lange, 2015). For pyroclastic flow deposits (67 wt.% in SiO2), T = 1030-1090℃ and Cw = 1.6-2.0 wt.%. For pyroclastic fall deposits (73 wt % in SiO2), T = 960-1000℃ and Cw = 1.9-2.3 wt.%. Bubble size distribution is basically bimodal, which is characterized by bubbles smaller than 100 μm that were formed just before the eruption and bubbles larger than 100 μm that were strongly affected by coalescence or formed in the magma reservoir. Bubbles smaller than 5 μm are significantly included in pumices from the pyroclastic flow deposits. The bubble number density (BND) is about 1014-1015 (bubbles/m3). Using the method of Toramaru (2006) and BND, SiO2 content, water content, and temperature, decompression rates were estimated to be 4.2-12 MPa/s for fall deposits, 16-21 MPa/s for fall deposit with reddish ash, and 15-36 MPa/s for pyroclastic flow deposits, respectively. These values are significantly larger than the decompression rate estimated in assumption of steady-state ascent, suggesting that the magma ascent for the 1783 eruption cannot be explained by steady state process.
Increase of juvenile lava fragments in the deposits of pyroclastic flow phases suggests that the lava-plug was formed onto the upper conduit. In this situation, when the lava-plug was ruptured, the rapid decompression may cause bubble nucleation effectively. In the pyroclastic flow deposit, small size bubbles less than 5 μm are remarkable and they show high number density. Such pyroclasts also have low vesicularity. In the pyroclastic flow phase, bubble growth and coalescence were not promoted. This suggests that the time between bubble nucleation and magma fragmentation might be shorter due to deeper fragmentation front. Based on the observation that amount of non-juvenile fragments are not significantly changed, a significant increase of the vent radius and decrease in the exit velocity are unlikely. If the time between decompression (thus bubble nucleation) and magma fragmentation was shorter, the amount of gas precipitated into the gas phase will decrease and may not be easily separated from pyroclastic particles to outside without promoting coalescence of bubbles. This process will reduce the amount of gas supplied to the plume and is possible to contribute to generate pyroclastic flows.
The 14 layers of Pumice-fall deposits and two types of pyroclastic flow deposits are chosen for measurement of density, microscopy, and chemical analysis. The average apparent densities range 0.7-0.8 g/cm3 for fall deposits, 0.8-0.9 g/cm3 and >1.20 g/cm3 for pyroclastic flow deposits. Groundmass glass is heterogeneous in some cases, and SiO2 content shows bimodal distribution around 67 wt.% and 73 wt.%. Pyroclastic flow deposits have lower SiO2 glass than pumice-fall deposits. Magma temperature (T) and water content (Cw) are estimated based on the compositions of plagioclase and glass matrix (Putirka, 2008; Waters and Lange, 2015). For pyroclastic flow deposits (67 wt.% in SiO2), T = 1030-1090℃ and Cw = 1.6-2.0 wt.%. For pyroclastic fall deposits (73 wt % in SiO2), T = 960-1000℃ and Cw = 1.9-2.3 wt.%. Bubble size distribution is basically bimodal, which is characterized by bubbles smaller than 100 μm that were formed just before the eruption and bubbles larger than 100 μm that were strongly affected by coalescence or formed in the magma reservoir. Bubbles smaller than 5 μm are significantly included in pumices from the pyroclastic flow deposits. The bubble number density (BND) is about 1014-1015 (bubbles/m3). Using the method of Toramaru (2006) and BND, SiO2 content, water content, and temperature, decompression rates were estimated to be 4.2-12 MPa/s for fall deposits, 16-21 MPa/s for fall deposit with reddish ash, and 15-36 MPa/s for pyroclastic flow deposits, respectively. These values are significantly larger than the decompression rate estimated in assumption of steady-state ascent, suggesting that the magma ascent for the 1783 eruption cannot be explained by steady state process.
Increase of juvenile lava fragments in the deposits of pyroclastic flow phases suggests that the lava-plug was formed onto the upper conduit. In this situation, when the lava-plug was ruptured, the rapid decompression may cause bubble nucleation effectively. In the pyroclastic flow deposit, small size bubbles less than 5 μm are remarkable and they show high number density. Such pyroclasts also have low vesicularity. In the pyroclastic flow phase, bubble growth and coalescence were not promoted. This suggests that the time between bubble nucleation and magma fragmentation might be shorter due to deeper fragmentation front. Based on the observation that amount of non-juvenile fragments are not significantly changed, a significant increase of the vent radius and decrease in the exit velocity are unlikely. If the time between decompression (thus bubble nucleation) and magma fragmentation was shorter, the amount of gas precipitated into the gas phase will decrease and may not be easily separated from pyroclastic particles to outside without promoting coalescence of bubbles. This process will reduce the amount of gas supplied to the plume and is possible to contribute to generate pyroclastic flows.