5:15 PM - 6:45 PM
[SVC26-P16] Extensometer records of short-term ground deformation prior to the Aso 2021 eruption
Keywords:strainmeter, Aso volcano, phreatic eruption, ground deformation
At Aso Volcano, south of Japan, phreatic eruptions occurred on the 14 and 20, October 2021, with a maximal eruptive plume height of ~ 4 km. From 8 October, clear ground deformation began, probably implying pressurization of shallow depth. This deformation was recorded with precise extensometers near the recent active crater, Nakadake first crater. Here, we present temporal characteristics of extensometer records and preliminary results of geodetic modeling.
Hondo tunnel station, 1 km southwest of the crater, has three extensometers of ~ 30 m length, heading different azimuths (E1: N05E, E2: S85E, E3: N50E). Based on temporal characteristics of the strain changes, the strain changes can be divided into three periods: from October 8 to October 13, 12:00 (period 1), from October 13, 12:00 to October 14, 4:43 (period 2), and 3: from October 14, 4:43 to October 13, 13:00 (period 3). In period 1, there was a rapid shortening of the radial component (E3) with a cumulative change of 0.4 μstrain, while no apparent change in E1 and E2 was observed. In period 2, E1 and E3 showed extension with their cumulative changes of 0.1 μstrain. In period 3, at the time of the eruption at 4:43 am on October 14, there was a stepwise extension of 0.05 and 0.10 μstrain in E1 and E3, respectively, while E2 remained almost unchanged. After that, the extensions of E1 and E3 continued until 13:00, and the cumulative strain change during period 3 was 0.07 and 0.20 μstrain, respectively. During the same period, except for the eruption timing, two tiltmeters at the Hondo tunnel showed no apparent tilt change, suggesting that the pressure source of the strain changes was the shallow part of the volcanic hydrothermal system.
To estimate the pressure source model of the hydrothermal system, we assumed a cylindrical poro-elastic source and calculated the strain fields by applying a pore-pressure change to the source. The horizontal position of the source was fixed at the crater. The depth and aspect ratio of the cylinder were assumed to be unknown parameters and estimated by a simple grid search. For this modeling, we used the finite element model of COMSOL Multiphysics software and incorporated a 10-m mesh digital elevation model to consider volcanic topography. Our modeling results show that the almost same source located 300-500 m below the crater contributed to the strain changes in periods 1 and 2. However, the polarity of the pressure change was reversed: pressurization in period 1 and depressurization in period 2. The timing of the pressure reversal was almost consistent with the time of a rapid increase in continuous tremor amplitudes. As one possibility, this may suggest that an over-pressurized source started to fail and switched to depressurization due to fluid evacuation. In period 3, the depressurization source was estimated, and its depth was 200-400 m, slightly shallower than during periods 1 and 2. These sources are located at the upper end of the crack-like conduit estimated from the seismicity of long-period tremors. The pressure build-up and depressurization process can be preliminarily interpreted as follows: magmatic fluid rising rapidly from a magma reservoir passed through the cracks with relatively high permeability at high velocity, but the flow stagnated at the exit of the crack due to the relatively low permeability of surrounding host rock. As a result, the pressure accumulated, and rock failure occurred, resulting in fluid leakage from the source. It is speculated that this depressurization may make the hydrothermal system unstable and may lead to eruption and subsequent rapid depressurization. In other words, depending on the ratio between the host rock permeability and the fluid supply rate, the system temporarily behaved like a closed system, with a sudden increase in the fluid supply rate.
Hondo tunnel station, 1 km southwest of the crater, has three extensometers of ~ 30 m length, heading different azimuths (E1: N05E, E2: S85E, E3: N50E). Based on temporal characteristics of the strain changes, the strain changes can be divided into three periods: from October 8 to October 13, 12:00 (period 1), from October 13, 12:00 to October 14, 4:43 (period 2), and 3: from October 14, 4:43 to October 13, 13:00 (period 3). In period 1, there was a rapid shortening of the radial component (E3) with a cumulative change of 0.4 μstrain, while no apparent change in E1 and E2 was observed. In period 2, E1 and E3 showed extension with their cumulative changes of 0.1 μstrain. In period 3, at the time of the eruption at 4:43 am on October 14, there was a stepwise extension of 0.05 and 0.10 μstrain in E1 and E3, respectively, while E2 remained almost unchanged. After that, the extensions of E1 and E3 continued until 13:00, and the cumulative strain change during period 3 was 0.07 and 0.20 μstrain, respectively. During the same period, except for the eruption timing, two tiltmeters at the Hondo tunnel showed no apparent tilt change, suggesting that the pressure source of the strain changes was the shallow part of the volcanic hydrothermal system.
To estimate the pressure source model of the hydrothermal system, we assumed a cylindrical poro-elastic source and calculated the strain fields by applying a pore-pressure change to the source. The horizontal position of the source was fixed at the crater. The depth and aspect ratio of the cylinder were assumed to be unknown parameters and estimated by a simple grid search. For this modeling, we used the finite element model of COMSOL Multiphysics software and incorporated a 10-m mesh digital elevation model to consider volcanic topography. Our modeling results show that the almost same source located 300-500 m below the crater contributed to the strain changes in periods 1 and 2. However, the polarity of the pressure change was reversed: pressurization in period 1 and depressurization in period 2. The timing of the pressure reversal was almost consistent with the time of a rapid increase in continuous tremor amplitudes. As one possibility, this may suggest that an over-pressurized source started to fail and switched to depressurization due to fluid evacuation. In period 3, the depressurization source was estimated, and its depth was 200-400 m, slightly shallower than during periods 1 and 2. These sources are located at the upper end of the crack-like conduit estimated from the seismicity of long-period tremors. The pressure build-up and depressurization process can be preliminarily interpreted as follows: magmatic fluid rising rapidly from a magma reservoir passed through the cracks with relatively high permeability at high velocity, but the flow stagnated at the exit of the crack due to the relatively low permeability of surrounding host rock. As a result, the pressure accumulated, and rock failure occurred, resulting in fluid leakage from the source. It is speculated that this depressurization may make the hydrothermal system unstable and may lead to eruption and subsequent rapid depressurization. In other words, depending on the ratio between the host rock permeability and the fluid supply rate, the system temporarily behaved like a closed system, with a sudden increase in the fluid supply rate.