著者
上田 拓 山谷 里奈 尾形 良彦 加藤 愛太郎
雑誌
JpGU-AGU Joint Meeting 2020
巻号頁・発行日
2020-03-13

On June 18, 2019, a Mj6.7 earthquake occurred at Yamagata-oki. The source region of this earthquake is adjacent to that of the Mj7.5 earthquake which occurred on June 16, 1964, and in this region, there are few aftershocks right after the 1964 earthquake, and the seismicity rate in recent years is extremely low (Earthquake Research Committee, 2019). This observation suggests that the source region of the 2019 Yamagata-oki earthquake was not ruptured by the Niigata earthquake, but the cause has not been revealed. In order to elucidate the relationship between these two areas, this study compared the characteristics of seismicity between the two areas.We used the JMA catalog constructed by Japan Meteorological Agency (the Preliminary Determination of Epicenters). We applied HIST-ETAS (Hierarchical Space Time Epidemic Type Aftershock Sequence) model (e.g., Ogata, 2004) considering the spatial dependence of each parameter of the Space Time ETAS model (e.g., Ogata, 1998), to the hypocenter catalog (M1.8) from 1998 through 2019 in order to estimate the spatial distribution of background seismicity rate μand number of aftershock occurrences K. As a result, we find that μ-value is higher and K-value is lower in the source region of Yamagata-oki earthquake than in that of Niigata earthquake.In addition to these differences, we find that the b-value, which is one of the characteristics of the seismicity, is lower in the source region of Yamagata-oki earthquake than in that of Niigata earthquake. Moreover, comparing the seismic wave velocity structure obtained by Matsubara et al. (2019), the P wave velocity is lower in the source region of Yamagata-oki earthquake than in that of Niigata earthquake.The difference in seismic wave velocity and characteristics of seismicity between these two areas suggests that the macroscopic behavior in the source region of Yamagata-oki earthquake is more ductile than in that of Niigata earthquake. In more ductile area, microfracture is likely to proceed and it decreases seismic wave velocity. In addition, background seismicity rate (μ) decreases in more ductile area because of low brittleness. Moreover, the results of rock experiments and numerical simulation by Amitrano (2003) imply the increase in aftershock productivity (K) and the decrease in b-value in more ductile area. Focusing on the short-wavelength component of the linear strain rate distribution in the east-west direction (Meneses-Gutierrez and Sagiya , 2016), the different response for the 2011 off the pacific coast of Tohoku earthquake between the source regions of Yamagata-oki earthquake and Niigata earthquake is appeared. These differences may reflect different deformation styles between the two regions.
著者
山谷 里奈 望月 公廣 悪原 岳 西田 究 市村 強 藤田 航平 山口 拓真 堀 高峰
雑誌
JpGU-AGU Joint Meeting 2020
巻号頁・発行日
2020-03-13

Off Ibaraki region is located at the southern end of the focal area of the 2011 off the Pacific coast of Tohoku Earthquake (Tohoku Earthquake). A dense network of 32 ocean bottom seismometers (OBSs) was deployed at this region with a station interval of about 6 km from October 2010 (11 OBSs started from February 2010) to October 2011. A large number (> 10,000) of aftershocks following the 2011 Tohoku earthquake were detected by this network. However, precise determination of these hypocenters and focal mechanisms is challenging due to uncertainties of seismic properties of thick sediment layers beneath the seafloor. The P-wave velocity structure has been reasonably constrained by active-source seismic surveys (Mochizuki et al., 2008), but the S-wave velocity structure is still unrevealed despite its importance.To constrain the S-wave velocity of the shallower portion, we apply the ambient noise interferometry to the short-period OBS data in this study. After dividing the data into ten-minute segments, we deconvolve the data with instrumental response function, remove trends, and discard data dominated by seismic events. Then, we apply a one-bit normalization and spectrum whitening. Finally, we calculate cross-correlations for vertical-vertical, radial-radial, and transverse-transverse components to retrieve Green's functions.We measure average phase velocity in the array using spatial auto-correlation method (Aki, 1957; Nishida et al., 2008). The phase velocities of the fundamental Rayleigh, the first-higher Rayleigh, and the fundamental Love modes are 0.5 to 2.5 km/s (in the frequency range of 0.1 to 0.3 Hz), 0.8 to 1.5 km/s (0.17 to 0.3 Hz), and 0.5 to 2.0 km/s (0.25 to 0.1 Hz), respectively. Next, we infer the 1-D average S-velocity isotropic structure by non-linear inversion, whose sensitivity is mainly ~5 km. The results show ~1000 m thick sediment with S-wave velocity of 300–1000 m/s immediately beneath the seafloor. At last, we apply band-pass filter with frequency range of 0.125 Hz and measure travel-time anomaly of the phase velocity in each frequency range, following Nagaoka et al. (2012). We apply non-linear inversion (Rawlinson & Sambridge, 2003) and find low-velocity anomalies in the deeper of the northern part and in the shallower of the center part.