17:15 〜 19:15
[SEM15-P06] Three-dimensional electrical resistivity structure beneath Izu-Oshima Island estimated by combining the onland and ocean bottom electromagnetic data
キーワード：伊豆大島、比抵抗構造、陸海合同探査
Izu-Oshima is an active volcanic island that erupts about every 35 years. The latest eruption occurred in 1986, and the next eruption is expected in the near future. Thus, it is important to understand the subsurface structure, especially the magma supply system of Izu-Oshima. Previous geophysical studies such as seismic, geodetic, gravity and electromagnetic studies revealed that a shallower structure than a couple of kilometers beneath Izu-Oshima, revealing a hydrothermal system, a depth of shallow magma reservoir and a pressure source. However, a subsurface structure as deep as 10 km depth is still not well known so far. This study shows the first result of the deep subsurface resistivity structure beneath Izu-Oshima and discusses the shape of magma reservoir and hydrothermal system.
In 2021-2022, Earthquake Research Institute, the University of Tokyo (ERI) and Japan Agency for Marine-Earth Science and Technology (JAMSTEC) conducted joint land and marine magnetotelluric (MT) surveys in and around Izu-Oshima. The onland data that measured 2 horizontal components of electric field and 3 components of magnetic field were obtained at 11 sites on Izu-Oshima for about 40 days, with sampling frequencies of 32 Hz continuously and 1024 Hz only for four hours per night. The marine data were obtained at 7 sites (although we initially deployed OBEMs at 10 sites) on the seafloor around Izu-Oshima for about 5 months, with sampling frequencies of 0.1 Hz continuously and 8 Hz for about 1 month.
First, we performed a time-series analysis by using BIRRP (Bounded Influence Remote Reference Processing) to estimate the MT response functions and geomagnetic transfer functions in frequency domain (Chave and Thomson, 2004). By analyzing land and marine data, MT response functions and geomagnetic transfer functions were estimated at the frequency range of 384 - 9.77 x 10-5 Hz.
Second, we performed a 3D inversion by using FEMTIC inversion code (Usui, 2015; Usui et al., 2017; Usui et al., 2024). We used all components of the MT impedance tensors and geomagnetic transfer functions as the input data and assigned error floors of 5% for all impedance components and 10% for all tipper components. An electrically conductive sediment layer model with 500 m thickness beneath seafloor was embedded as conductive as 1Ωm as a priori information to stabilize the inversion, and some different cases of the starting resistivity models and trade-off parameters which balance the weight between data-misfit and model roughness were tested to obtain an optimal model.
As a result of the inversion, we were able to estimate the wide and deep subsurface resistivity structure of Izu-Oshima (Figure 1). The subsurface model shows that the higher resistive layer than 100Ωm are in the upper region than sea level, and beneath the sea level of the island, the low resistive body are vertically elongated to 15 km depth.
The shallow part of our result can be interpreted that the upper high resistive layer is unsaturated, and the uppermost low resistive body is saturated by water.
The hypocenters are distributed in the relatively high resistivity region and the resistivity boundary region beneath Izu-Oshima. This may also suggest that the low resistive body is associated with volcanic fluid such as hydrothermal fluid and melt that reduces the stiffness, and prevents brittle rupture. In addition, the low resistivity region extending to 2 km depth is also consistent with the low velocity region of the seismic wave velocity structure reported in previous studies.
Beneath a couple of kilometers depth, considering the temperature and pressure conditions, it may be difficult to explain the low resistive body as low as 10Ωm only by the hydrothermal water, but some portion of melt might be expected.
This study provides the new constraints and insights on the deep subsurface structure of Izu-Oshima from the electromagnetic perspective.
Figure 1: Cross sections of resistivity model. (left) west-east profile (right) south-north profile. The horizontal origin of the coordinate is the summit of Mt. Mihara. Black dots indicate hypocenters, provided by the JMA.
In 2021-2022, Earthquake Research Institute, the University of Tokyo (ERI) and Japan Agency for Marine-Earth Science and Technology (JAMSTEC) conducted joint land and marine magnetotelluric (MT) surveys in and around Izu-Oshima. The onland data that measured 2 horizontal components of electric field and 3 components of magnetic field were obtained at 11 sites on Izu-Oshima for about 40 days, with sampling frequencies of 32 Hz continuously and 1024 Hz only for four hours per night. The marine data were obtained at 7 sites (although we initially deployed OBEMs at 10 sites) on the seafloor around Izu-Oshima for about 5 months, with sampling frequencies of 0.1 Hz continuously and 8 Hz for about 1 month.
First, we performed a time-series analysis by using BIRRP (Bounded Influence Remote Reference Processing) to estimate the MT response functions and geomagnetic transfer functions in frequency domain (Chave and Thomson, 2004). By analyzing land and marine data, MT response functions and geomagnetic transfer functions were estimated at the frequency range of 384 - 9.77 x 10-5 Hz.
Second, we performed a 3D inversion by using FEMTIC inversion code (Usui, 2015; Usui et al., 2017; Usui et al., 2024). We used all components of the MT impedance tensors and geomagnetic transfer functions as the input data and assigned error floors of 5% for all impedance components and 10% for all tipper components. An electrically conductive sediment layer model with 500 m thickness beneath seafloor was embedded as conductive as 1Ωm as a priori information to stabilize the inversion, and some different cases of the starting resistivity models and trade-off parameters which balance the weight between data-misfit and model roughness were tested to obtain an optimal model.
As a result of the inversion, we were able to estimate the wide and deep subsurface resistivity structure of Izu-Oshima (Figure 1). The subsurface model shows that the higher resistive layer than 100Ωm are in the upper region than sea level, and beneath the sea level of the island, the low resistive body are vertically elongated to 15 km depth.
The shallow part of our result can be interpreted that the upper high resistive layer is unsaturated, and the uppermost low resistive body is saturated by water.
The hypocenters are distributed in the relatively high resistivity region and the resistivity boundary region beneath Izu-Oshima. This may also suggest that the low resistive body is associated with volcanic fluid such as hydrothermal fluid and melt that reduces the stiffness, and prevents brittle rupture. In addition, the low resistivity region extending to 2 km depth is also consistent with the low velocity region of the seismic wave velocity structure reported in previous studies.
Beneath a couple of kilometers depth, considering the temperature and pressure conditions, it may be difficult to explain the low resistive body as low as 10Ωm only by the hydrothermal water, but some portion of melt might be expected.
This study provides the new constraints and insights on the deep subsurface structure of Izu-Oshima from the electromagnetic perspective.
Figure 1: Cross sections of resistivity model. (left) west-east profile (right) south-north profile. The horizontal origin of the coordinate is the summit of Mt. Mihara. Black dots indicate hypocenters, provided by the JMA.