11:00 AM - 1:00 PM
[SCG54-P03] Modeling of harmonic volcanic tremor occurred in the deep part of Hakone volcano, central Japan.
Keywords:Nonlinear excitation, volcanic tremor
In this study, we modeled the observed waveform of a harmonic tremor that occurred at the deep part of Hakone volcano, central Japan, on May 26, 2019. This tremor signal continued for about 10 minutes, and the latter part of the signal exhibited the harmonic feature, having a fundamental mode of approximately 1 Hz. In Hakone, volcano tectonic (VT) earthquakes occur above the depth of 6km. A low-velocity region suggesting the existence of magmatic fluid was estimated beneath the hypocenter distribution of the VT earthquakes. Deep low-frequency earthquakes (DLFEs) also occur at depths of 20~30km. The tremor source was estimated around the depth of 20km~40km beneath Hakone volcano, corresponding to the same region (or deeper extension) of the hypocenter distribution of DLFEs (Yukutake et al., 2021). Since the harmonic tremor at the volcanic root is rarely observed, we tried to discuss the mechanism of the tremor based on the theoretical modeling assuming fluid flow.
Julian (1994) proposed tremor model made of elastic plates, which viscous fluid vibrates the plates. Fluid flows in a dike with infinite long in the model, which means it doesn’t have limitation of length and moment. In this model, viscos fluid flows in ellipse tube consistent of elastic surface and solve to time integration of cross-sectional area. This change constricts length of dike and made relationship to moment. Furthermore, by introducing constant of connecting physical property, we challenged comparing to the observation. Also we extend model into turbulent flow.
We adjusted the model into two patterns of fluid, magma and supercritical water. We detected that both fluids can cause tremor which frequency is same to observation.
If magma flows, tube of 3 m height, 2 m long, 0.15m width vibrates by flowing 6m/s with 300kPa effective fluid pressure. On the other hand, tube of flowing supercritical water with 0.6 mole fraction CO2, 2.5 m height, 2 m long, 0.04m width oscillates with 43kPa effective fluid pressure. Flow of supercritical water is turbulent flow and effective fluid pressure is same order of its buoyancy. Both, we must multiply bulk modulus by coefficients of about 10-6 in order to recreate observed frequency.
We compared the observed seismic waveform and modeled waveform that was obtained by the convolution of Green’s function. We calculated the green function by using the three-dimensional differential method (Maeda et al., 2017; EPS), assuming the one-dimensional velocity structure beneath Tanzawa Mountains. We could reproduce several features of the tremor signal, such as the one-side oscillation in the vertical component and the sharp peaks of waveform records, by using this model. However, the amplitude of the model wave was higher than the observation, which may suggest the effect of attenuation of the seismic wave that was not fully considered in this model. Moreover, the effective bulk modulus could be smaller than the geometrically expected value for the elliptic tube due to the partial melting of surrounding rocks.
Julian (1994) proposed tremor model made of elastic plates, which viscous fluid vibrates the plates. Fluid flows in a dike with infinite long in the model, which means it doesn’t have limitation of length and moment. In this model, viscos fluid flows in ellipse tube consistent of elastic surface and solve to time integration of cross-sectional area. This change constricts length of dike and made relationship to moment. Furthermore, by introducing constant of connecting physical property, we challenged comparing to the observation. Also we extend model into turbulent flow.
We adjusted the model into two patterns of fluid, magma and supercritical water. We detected that both fluids can cause tremor which frequency is same to observation.
If magma flows, tube of 3 m height, 2 m long, 0.15m width vibrates by flowing 6m/s with 300kPa effective fluid pressure. On the other hand, tube of flowing supercritical water with 0.6 mole fraction CO2, 2.5 m height, 2 m long, 0.04m width oscillates with 43kPa effective fluid pressure. Flow of supercritical water is turbulent flow and effective fluid pressure is same order of its buoyancy. Both, we must multiply bulk modulus by coefficients of about 10-6 in order to recreate observed frequency.
We compared the observed seismic waveform and modeled waveform that was obtained by the convolution of Green’s function. We calculated the green function by using the three-dimensional differential method (Maeda et al., 2017; EPS), assuming the one-dimensional velocity structure beneath Tanzawa Mountains. We could reproduce several features of the tremor signal, such as the one-side oscillation in the vertical component and the sharp peaks of waveform records, by using this model. However, the amplitude of the model wave was higher than the observation, which may suggest the effect of attenuation of the seismic wave that was not fully considered in this model. Moreover, the effective bulk modulus could be smaller than the geometrically expected value for the elliptic tube due to the partial melting of surrounding rocks.