11:00 AM - 1:00 PM
[SVC32-P09] Effects of magma properties and conduit geometry on the conduit flow dynamics during the 1914 Sakurajima eruption
Keywords:Conduit flow, Sakurajima volcano, Crystallization experiment
During the 1914 eruption at Sakurajima volcano, Japan, two types of eruptions occurred with andesitic magma: a Plinian eruption and a lava flow eruption. It is important to understand the dynamics of conduit flow to investigate the origin of these eruption styles. However, the effects of physical properties of andesitic magma, such as crystallization and viscosity, on the conduit flow have not been fully investigated. It is also necessary to investigate the effects of magma ascent in a dyke-shaped volcanic conduit at Sakurajima. The purpose of this study is to investigate the volcanic flow dynamics during the 1914 Sakurajima eruption using a numerical model with the constraints of the magmatic properties and the conduit geometry for the Sakurajima eruption.
We investigated the conduit flow during the 1914 Sakurajima eruption using a 1-dimensional steady conduit flow model. We applied the magma properties of the Sakurajima eruption to the conduit flow model by considering (1) pressure dependence of equilibrium crystallinity and crystal growth rate based on crystallization experiments, and (2) liquid-phase viscosity depending on the chemical composition of the magma. In addition, the ellipsoidal cross-section of the conduit was set up to simulate a dyke-like conduit, which enables us to compare the cylindrical and the dyke-like conduits. In the model for the Plinian eruption, we assumed that magma fragmentation occurs when the volume fraction of the gas exceeds a critical value. In the model for the lava flow eruption, we considered vertical and lateral gas escapes from the magma.
Based on this model, we investigated the relationship between chamber pressure and mass flow rate (pch-Q) in the steady conduit flow, which is important for describing the macroscopic dynamics of the conduit flow. When there is a positive correlation in the pch-Q relationship, the conduit flow is stable, which corresponds to eruptive phases such as the Plinian eruption and the lava flow eruption. In the analysis of the Plinian eruption, we found that a stable region in the pch-Q relationship can exist in the range of the estimated mass flow rate by considering the dyke-like conduit geometry. In the analysis of the lava flow eruption, we found that the equation and the parameters related to the crystallization are important to generate the stable region of the pch-Q relation at the low mass flow rate both in the cases of the cylindrical and the dyke-like conduit geometry. When we use the equilibrium crystallization equation based on our experiments, a maximum mass flow rate of the stable region at the low-flow rate side (Qmax) becomes smaller than the case of the equation for more felsic melt in the previous studies. Furthermore, we found that when the initial crystallinity is set to 25 vol.% based on the petrological data at Sakurajima, the critical crystallinity at which magma viscosity rapidly increases should be set below about 60 vol.% to generate the stable region in the low-flow rate range. These features originate from the fact that the lower viscosity of the andesitic magma induces a smaller friction between the conduit wall and the liquid and a smaller magma overpressure, leading to the suppression of the gas escape. We also recognized that Qmax is smaller than the estimated mass flow rate during the 1914 lava flow eruption. It is necessary to consider additional effects such as a depth-varying conduit geometry in the future study.
We investigated the conduit flow during the 1914 Sakurajima eruption using a 1-dimensional steady conduit flow model. We applied the magma properties of the Sakurajima eruption to the conduit flow model by considering (1) pressure dependence of equilibrium crystallinity and crystal growth rate based on crystallization experiments, and (2) liquid-phase viscosity depending on the chemical composition of the magma. In addition, the ellipsoidal cross-section of the conduit was set up to simulate a dyke-like conduit, which enables us to compare the cylindrical and the dyke-like conduits. In the model for the Plinian eruption, we assumed that magma fragmentation occurs when the volume fraction of the gas exceeds a critical value. In the model for the lava flow eruption, we considered vertical and lateral gas escapes from the magma.
Based on this model, we investigated the relationship between chamber pressure and mass flow rate (pch-Q) in the steady conduit flow, which is important for describing the macroscopic dynamics of the conduit flow. When there is a positive correlation in the pch-Q relationship, the conduit flow is stable, which corresponds to eruptive phases such as the Plinian eruption and the lava flow eruption. In the analysis of the Plinian eruption, we found that a stable region in the pch-Q relationship can exist in the range of the estimated mass flow rate by considering the dyke-like conduit geometry. In the analysis of the lava flow eruption, we found that the equation and the parameters related to the crystallization are important to generate the stable region of the pch-Q relation at the low mass flow rate both in the cases of the cylindrical and the dyke-like conduit geometry. When we use the equilibrium crystallization equation based on our experiments, a maximum mass flow rate of the stable region at the low-flow rate side (Qmax) becomes smaller than the case of the equation for more felsic melt in the previous studies. Furthermore, we found that when the initial crystallinity is set to 25 vol.% based on the petrological data at Sakurajima, the critical crystallinity at which magma viscosity rapidly increases should be set below about 60 vol.% to generate the stable region in the low-flow rate range. These features originate from the fact that the lower viscosity of the andesitic magma induces a smaller friction between the conduit wall and the liquid and a smaller magma overpressure, leading to the suppression of the gas escape. We also recognized that Qmax is smaller than the estimated mass flow rate during the 1914 lava flow eruption. It is necessary to consider additional effects such as a depth-varying conduit geometry in the future study.