後藤 忠徳

Abstract Megathrust and slow earthquakes are known to occur in the area off Kumano, along the eastern part of the Nankai Trough, Japan. Pore fluids along the fault surface play an important role in earthquake occurrence, but the detailed fluid distribution remains unknown. In this study, based on seafloor electromagnetic field observations, we estimated the three-dimensional resistivity structure, which reflects the fluid distribution off Kumano. The optimal three-dimensional resistivity model showed a low-resistivity layer at shallow depths below the seafloor. The low-resistivity layer matches geological features observed in seismic reflection surveys, including forearc basin fill, the frontal prism, and underthrust sediments. Along the plate boundary (décollement), the resistivity increased markedly with increasing depth. Comparison with slow earthquakes along the plate boundary reveals that shallow very low-frequency earthquakes, shallow slow slip events, and tremor only occur at depths < 10 km and on the conductive zone of the plate boundary. Furthermore, the area of coseismic slip during the 1944 Tonankai earthquake was characterized by higher resistivity than the slow earthquake epicentral area. These findings indicate that the pore-fluid distribution along the plate boundary may influence earthquake occurrence and type. Graphical Abstract

Understanding the distribution of physical properties around shallow subducting plate interfaces, where both destructive and "slow" earthquakes occur due to rapid and slower fault slips, respectively, presents a major scientific and disaster mitigation challenge. Pore water is a key factor in understanding the different slip mechanisms and their spatial relationships; however, its distribution remains understudied. In this study, based on marine magnetotelluric survey in Hyuga-nada, southwestern Japan, we identified distinct resistive and conductive anomalies along the plate interface. These anomalies correspond to areas of scarce pore fluid and high concentration area of pore fluids sourced from subducting seamounts (Kyushu-Palau Ridge), respectively. The wet area corresponds to the slow slip area, whereas the dry and transition areas correspond to areas of fast fault slip. These findings provide clear observational evidence that pore fluid distribution correlates with fault rupture behavior.

Blurred resistivity boundaries resulting from smoothness-regularized inversions of electrical resistivity tomography (ERT) data can lead to inaccurate interpretations of sharp boundary structures. To address this issue, various ERT inversion algorithms have introduced localized adjustments (localized discontinuities) in the regularization operator at positions where sharp boundaries are anticipated. Current approaches rely on prior information about sharp boundary locations, obtained from complementary geophysical, geological, and drilling data, to determine the positions and weights for these regularization adjustments. However, such prior information is frequently insufficient, limiting the application of localized regularization adjustments. Accordingly, we developed a sharp boundary inversion (SBI) algorithm using the Akaike Bayesian Information Criterion (ABIC) that determines the optimal positions and weights for localized regularization adjustments by testing various configurations and selecting the one that minimizes ABIC. A synthetic modeling study demonstrated that the SBI algorithm correctly delineated the sharp boundaries of a conductor. Its application to field data demonstrated that it delineated the sharp boundaries of a utility tunnel, and the size and horizontal position of the recovered tunnel were consistent with the estimated dimensions from the blueprint. As it does not rely heavily on prior information, the SBI algorithm can be applied to a wide range of geophysical survey data, even when prior knowledge of sharp boundary locations is limited.

Abstract Near-seafloor concentrated gas hydrates (GHs) containing large amounts of methane have been identified at various gas chimney sites. Although understanding the spatial distribution of GHs is fundamental for assessing their dissociation impact on aggravating global warming and resource potential, the spatial distribution of GHs within gas chimneys remains unclear. Here, we estimate the subseafloor distribution of GHs at a gas chimney site in the Japan Sea using marine electrical resistivity tomography data. The resulting two-dimensional subseafloor resistivity structure shows high anomalies (10–100 Ωm) within seismically inferred gas chimneys. As the resistivity anomalies are aligned with high amplitude seismic reflections and core positions recovering GHs, we interpret the resistivity anomalies are near-seafloor concentrated GH deposits. We also detect various distribution patterns of the high resistivity anomalies including 100-m wide and 40-m thick anomaly near the seafloor and 500-m wide anomaly buried 50 m below the seafloor, suggesting that GHs are heterogeneously distributed. Therefore, considering such heterogeneous GH distribution within gas chimneys is critical for in-depth assessments of GH environmental impacts and energy resources.