Numerical simulation of magma plumbing system associated with the eruption at the Showa crater of Sakurajima inferred from ground deformation

We studied the ground deformation associated with the eruption at the Showa crater of Sakurajima, which has been active since 2006. Using the Mogi’s spherical pressure model, a volume change of magma chambers can be estimated from the displacement, tilt, or strain observations near the ground surface. After the application of the Mogi’s model to data in the past observations, the existence of two magma chambers has been inferred beneath the Sakurajima down to a depth of 5 km. The tilt and strain data in 2 underground tunnel sites observed 36 hours before an eruption in April 9, 2009, are analyzed to reveal the behavior of magma leading to eruption. From these data, there seems to be a time lag in the inflation between the two magma chambers at a depth of 4km and 0.1km, respectively. In addition, the order of the volume change of the shallow source is about one tenth of that of the deep one. A system which consists of shallow and deep magma chambers and a vertical conduit connecting them is numerically modeled to investigate the mechanism of the time lag and why the difference in the magnitude of the volumetric changes in the two chambers appears as described above. The initial values of magma properties in the deep magma chamber are assumed from the volcanic ejecta of Sakurajima volcano. We assumed that magma is supplied with a constant rate to the deep magma chamber. The two different pressure limits are assigned to the deep and shallow chamber, respectively: (1) one triggers the magma uprise from the deep to the shallow, and (2) the other to start to erupt. In a one-dimentional steady flow model of a magma conduit, we consider the vesiculation of volatile-bearing magma, gas escape and overpressure in the bubble due to the viscous resistance, which largely influences the physical properties of magma. Although our simulation results cannot exactly describe the data, we confirmed that our hypothetical model with two triggers could explain the time lag of the inflation and the difference in the order of the volume change. We would like to propose that the numerical simulation could be a powerful tool for understanding the behavior of magma before eruption once ground deformations are well observed.