17:15 〜 18:45
[SEM12-P22] Three-dimensional resistivity structure of the epicenter area of the 1997 Kagoshima earthquake doublet, Japan
キーワード： magnetotelluric、1997年鹿児島県北西部地震、比抵抗構造
We estimated the 3-D resistivity structure around the epicenter of the 1997 Kagoshima Prefecture Northwestern Earthquake doublet in Japan by using broadband magneto-telluric (MT) data. The aim of this study is evaluations of the potential for large earthquakes (Aizawa et al., 2022) by investigating the spatial relationship between earthquake rupture and resistivity structure.
The 1997 Kagoshima northwest earthquake doublet is a couple of M6-class earthquakes. The earthquake created an F-shaped aftershock zone extending roughly east-west and north-south directions. The slip distribution of the two earthquakes was estimated from K-net strong-motion data (Horikawa 2001). Furthermore, the resistivity structure was estimated in this region (Umeda et al., 2014). According to the 3-D resistivity structure by Umeda et al., 2014, there exists a near-vertical low resistivity zone extending down to the base of the crust. Umeda et al. 2014 suggested that the invasion of mantle fluids into the crust is attributed to the occurrence of the 1997 earthquake doublet. However, Umeda et al. (2014) did not discuss the relationship between earthquake rupture and resistivity structure possibly because the spatial resolution of the resistivity structure is not enough for discussing the earthquake rupture scale ~5 km. Therefore, in this study, we estimate a finer resistivity structure in this region by adding new MT data.
We used the broadband MT data at 41 sites by Umeda et al. (2014) (obtained in 2013), plus data at 23 sites established by Kyushu University (obtained in 2016,2017, and 2022), for a total of 64 MT data points (see figure). The frequency bands of the MT response functions are 3*10^-4 ~ 320 Hz for Umeda et al. (2014) and 3*10^-4 ~ 320 Hz for Kyushu University. For the calculations of MT response functions by the Kyushu University, we used BIRRP code (Chave and Thomson, 2004) with remote reference processing by using the geomagnetic data at the Kirishima volcano (about 50 km east) and the Kakioka geomagnetic stations (about 1000 km northeast).
We used an unstructured tetrahedral mesh and FEMTIC code (Usui, 2015; Usui et al., 2017) for estimating a 3D resistivity structure. Input data are four components of impedance tensor and two components of Tipper at 20 frequencies (3.05 *10^-4 ~ 80 Hz). In the inversion, we gave the fixed air (10^8 Ωm), the fixed ocean (0.33 Ωm) and unfixed land (100 Ωm), respectively. We gave an error of 10% for the diagonal component of the impedance, 5% for the off-diagonal component, and 0.05 for the tipper. The final RMS value was 1.51 (the initial value: 14.12). The final resistivity structure shows low-resistivity zones near the eastern and western edge of the aftershocks at depths of 5-10 km. The hypocenters of the M6 earthquakes are located near the edge of the low-resistivity zones. This result is similar to those of the 2016 Kumamoto earthquake (Aizawa et al. 2022). It is interesting that a prominent high resistivity zone is sandwiched by two low-resistivity zones at a depth of 0-3 km, and its location coincides with the distribution of the granodiorite body.
Our resistivity structure is not final. In order to estimate the detailed resistivity structure near the hypocenters of M6 earthquakes, we will conduct MT survey on March, 2024. By including new data, we elucidate the relationship between the resistivity structure and the earthquake rupture.
The 1997 Kagoshima northwest earthquake doublet is a couple of M6-class earthquakes. The earthquake created an F-shaped aftershock zone extending roughly east-west and north-south directions. The slip distribution of the two earthquakes was estimated from K-net strong-motion data (Horikawa 2001). Furthermore, the resistivity structure was estimated in this region (Umeda et al., 2014). According to the 3-D resistivity structure by Umeda et al., 2014, there exists a near-vertical low resistivity zone extending down to the base of the crust. Umeda et al. 2014 suggested that the invasion of mantle fluids into the crust is attributed to the occurrence of the 1997 earthquake doublet. However, Umeda et al. (2014) did not discuss the relationship between earthquake rupture and resistivity structure possibly because the spatial resolution of the resistivity structure is not enough for discussing the earthquake rupture scale ~5 km. Therefore, in this study, we estimate a finer resistivity structure in this region by adding new MT data.
We used the broadband MT data at 41 sites by Umeda et al. (2014) (obtained in 2013), plus data at 23 sites established by Kyushu University (obtained in 2016,2017, and 2022), for a total of 64 MT data points (see figure). The frequency bands of the MT response functions are 3*10^-4 ~ 320 Hz for Umeda et al. (2014) and 3*10^-4 ~ 320 Hz for Kyushu University. For the calculations of MT response functions by the Kyushu University, we used BIRRP code (Chave and Thomson, 2004) with remote reference processing by using the geomagnetic data at the Kirishima volcano (about 50 km east) and the Kakioka geomagnetic stations (about 1000 km northeast).
We used an unstructured tetrahedral mesh and FEMTIC code (Usui, 2015; Usui et al., 2017) for estimating a 3D resistivity structure. Input data are four components of impedance tensor and two components of Tipper at 20 frequencies (3.05 *10^-4 ~ 80 Hz). In the inversion, we gave the fixed air (10^8 Ωm), the fixed ocean (0.33 Ωm) and unfixed land (100 Ωm), respectively. We gave an error of 10% for the diagonal component of the impedance, 5% for the off-diagonal component, and 0.05 for the tipper. The final RMS value was 1.51 (the initial value: 14.12). The final resistivity structure shows low-resistivity zones near the eastern and western edge of the aftershocks at depths of 5-10 km. The hypocenters of the M6 earthquakes are located near the edge of the low-resistivity zones. This result is similar to those of the 2016 Kumamoto earthquake (Aizawa et al. 2022). It is interesting that a prominent high resistivity zone is sandwiched by two low-resistivity zones at a depth of 0-3 km, and its location coincides with the distribution of the granodiorite body.
Our resistivity structure is not final. In order to estimate the detailed resistivity structure near the hypocenters of M6 earthquakes, we will conduct MT survey on March, 2024. By including new data, we elucidate the relationship between the resistivity structure and the earthquake rupture.