9:30 AM - 9:45 AM
[SVC31-09] A global sensitivity analysis of a conduit flow model for andesitic lava effusion
Keywords:conduit flow, lava effusion, Sakurajima volcano
A high-flux lava effusion (106–107 kg s-1) is occasionally observed during andesitic eruptions, such as the 1914 (Taisho) eruption at Sakurajima volcano in southern Kyushu, Japan, and the 1783 (Tenmei) eruption at Asama volcano in central Honshu, Japan. The underlying factors contributing to this phenomenon and a temporal change in lava-effusion rate have not been fully investigated. Generally, the relationship between eruption styles and the magma discharge rate is controlled by the magma ascent process through a volcanic conduit, which is known as a conduit flow. In this study, we investigated the conduit flow dynamics during the andesitic lava effusion and tried to reveal the effects of magma properties and geological conditions on the discharge rate during the andesitic lava effusion event at Sakurajima volcano. This investigation was conducted through a global sensitivity analysis of the numerical conduit flow model.
The progress of lava flow during the 1914 lava effusion from the western crater of Sakurajima volcano was observed in aerial photography. Based on this data, a maximum discharge rate and a decrease in the discharge rate were estimated (Ishihara et al, 1985). To investigate the conduit flow dynamics during the lava effusion based on these estimated data, we constructed a 1-dimensional steady conduit flow model that considers crystallization kinetics as well as vesiculation and gas escape during the magma ascent in the volcanic conduit. We also applied the magma properties of the Sakurajima eruption to the model by considering equilibrium crystallinity, crystal growth rate, and liquid-phase viscosity based on the experimental data. This model can be used to estimate the discharge rate when the input parameters of the magma system (magma properties and geological conditions) are specified. We performed a global sensitivity analysis of the conduit flow model to a selection of model inputs, and the results were analyzed using parallel coordinate plots for visualization to quantify the sensitivity of model outputs (i.e., discharge rate).
Based on the global sensitivity analysis, we derived the magma properties and geological conditions for reproducing the maximum discharge rate during the 1914 lava effusion. There are two competing effective parameters. First, the maximum discharge rate is strongly controlled by the maximum packing fraction of crystals which is the parameter controlling magma viscosity. A lower maximum packing fraction leads to higher magma viscosity and more efficient gas escape. As a result, the conduit flow becomes stable, leading to a higher maximum discharge rate. The value of this parameter should be set below 0.4 to reproduce the actual maximum discharge rate. Second, as the permeability for lateral gas escape increases, the maximum discharge rate increases effectively. This result implies that the geological condition of the host rock surrounding the conduit is also an effective factor in the high-flux lava effusion. Furthermore, we investigated a temporal change in the discharge rate by combining the conduit flow model with an equation governing a pressure change in an elastically deformable magma chamber in response to magma outflux to the conduit. We found a minimum magma chamber volume required to achieve the observed temporal change in the discharge rate. These findings enable us to evaluate how the spatial and temporal scale of andesitic lava effusion depends on the magmatic and geological factors in the conduit-chamber system.
The progress of lava flow during the 1914 lava effusion from the western crater of Sakurajima volcano was observed in aerial photography. Based on this data, a maximum discharge rate and a decrease in the discharge rate were estimated (Ishihara et al, 1985). To investigate the conduit flow dynamics during the lava effusion based on these estimated data, we constructed a 1-dimensional steady conduit flow model that considers crystallization kinetics as well as vesiculation and gas escape during the magma ascent in the volcanic conduit. We also applied the magma properties of the Sakurajima eruption to the model by considering equilibrium crystallinity, crystal growth rate, and liquid-phase viscosity based on the experimental data. This model can be used to estimate the discharge rate when the input parameters of the magma system (magma properties and geological conditions) are specified. We performed a global sensitivity analysis of the conduit flow model to a selection of model inputs, and the results were analyzed using parallel coordinate plots for visualization to quantify the sensitivity of model outputs (i.e., discharge rate).
Based on the global sensitivity analysis, we derived the magma properties and geological conditions for reproducing the maximum discharge rate during the 1914 lava effusion. There are two competing effective parameters. First, the maximum discharge rate is strongly controlled by the maximum packing fraction of crystals which is the parameter controlling magma viscosity. A lower maximum packing fraction leads to higher magma viscosity and more efficient gas escape. As a result, the conduit flow becomes stable, leading to a higher maximum discharge rate. The value of this parameter should be set below 0.4 to reproduce the actual maximum discharge rate. Second, as the permeability for lateral gas escape increases, the maximum discharge rate increases effectively. This result implies that the geological condition of the host rock surrounding the conduit is also an effective factor in the high-flux lava effusion. Furthermore, we investigated a temporal change in the discharge rate by combining the conduit flow model with an equation governing a pressure change in an elastically deformable magma chamber in response to magma outflux to the conduit. We found a minimum magma chamber volume required to achieve the observed temporal change in the discharge rate. These findings enable us to evaluate how the spatial and temporal scale of andesitic lava effusion depends on the magmatic and geological factors in the conduit-chamber system.