We developed a travel-time tomography method to reveal general anisotropic media using the shortest path method. Introducing the local coordinate system fixed to the symmetric axis of an anisotropic medium, the rotation angles of the local coordinate system and elastic parameters in the local coordinate system are unknown parameters to be solved. The method is basically applicable to general anisotropic media. Using synthetic travel-time data, we applied our method to a hexagonal symmetric anisotropic medium and orthorhombic symmetric anisotropic medium to examine the method. The targeted symmetric anisotropic media were successfully restored, although the range of the rotation angle is limited.

Normal faulting earthquake sequence was induced by the 2011 off the Pacific coast of Tohoku Earthquake (M_w9.0) in Hamadori region of Fukushima prefecture and northern part of Ibaraki prefecture. T-axes derived from focal mechanisms of events occurred in Hamadori region of Fukushima prefecture and northern part of Ibaraki prefecture are perpendicular to each other. It should be noted that these normal faulting earthquakes are distributed in the Pacific coastal area of Abukuma belt. Based on the investigation about past activity of Tanakura Tectonic Line, Hatagawa Tectonic Line and Futaba fault, we found that Pacific coastal area of Abukuma belt should have been deformed as sinistral and dextral shear zones. According to Riedel shear experiment as performed by previous study, Riedel shears should have been developed by the sinistral and dextral shear deformation of Pacific coastal area of Abukuma belt. We concluded that normal faulting earthquakes occurred in Hamadori region of Fukushima prefecture can be regarded as reactivation of Riedel shears formed by dextral shear deformation of Pacific coastal area of Abukuma belt. In contrast, those events occurred in northern part of Ibaraki prefecture can be regarded as reactivation of Riedel shears formed by sinistral shear deformation of Pacific coastal area of Abukuma belt. Yunodake fault and western trace of Itozawa fault were ruptured by the occurrence of 2011 Fukushima Hamadori earthquake (M_j7.0). These normal faults can be regarded as ancient Riedel shears formed by sinistral shear deformation of Pacific coastal area of Abukuma belt.

This study reports the tsunami source modeling of an M_JMA 7.1 earthquake which occurred beneath Hyuganada Sea, off Miyazaki, Japan on 8 August 2024, using tsunami waveforms recorded by the recently-implemented offshore seafloor observation network called Nankai Trough Seafloor Observation Network for Earthquakes and Tsunamis (N-net) together with the Dense Oceanfloor Network system for Earthquakes and Tsunamis (DONET) installed in the Nankai Trough subduction zone, off the Pacific coast of southwestern Japan. We first processed the ocean-bottom pressure gauge data from the N-net offshore system and DONET to extract tsunamis, which showed tsunamis up to a few centimeters were recorded. The station NAE18 of N-net, nearest to the epicenter, observed the pressure offset change corresponding to the permanent seafloor vertical uplift of approximately 5 cm. We then forwardly simulated tsunamis based on centroid moment tensor solutions and finite fault models of the previous studies, indicating the good observation performance of the N-net ocean-bottom pressure gauges. We finally conducted the inversion analysis of the tsunami waveforms to estimate the tsunami source distribution, which extends in a dimension of approximately 40 km×20 km along the strike and dip directions, respectively. From the comparison with the past studies, the rupture area of the 2024 earthquake was likely to overlap with the southern half of the rupture area of the M_JMA 7.0 earthquake in 1961, while the rupture areas of the M_JMA 6.7 and 6.9 earthquakes in 1996 seemed separated from the 2024 event. This study demonstrates that the use of the near-field tsunami waveforms from N-net significantly enhanced the constraint on tsunami source estimation as well as the finite fault modeling off the coast of the western part of the Nankai Trough subduction zone. N-net also has the potential to improve tsunami monitoring and prompt evaluation for the “Nankai Trough Earthquake Extra Information.”

On August 8, 2024, the M_J 7.1 Hyuganada earthquake occurred at the southwestern edge of the Nankai Trough subduction zone. Near the source region of this earthquake, M_J 6.6 earthquake occurred on January 13, 2025 and two M 7-class earthquakes (M_J 6.9 and M_J 6.7) occurred in 1996. For the prediction of strong ground motions of the next M 8-9 Nankai Trough earthquake, the characterized source models of the four M 7-class earthquakes were estimated in this study. The empirical Green’s function method was used to synthesize the strong motions from the characterized source models, which were composed of strong motion generation areas (SMGAs). Two SMGAs for the M_J 7.1, M_J 6.9, and M_J 6.7 earthquakes and three SMGAs for the M_J 6.6 earthquake were estimated respectively, using the simulated annealing algorithm. The strong motion records over a frequency range of 0.2-10 Hz observed at N-net on the seafloor and K-NET, KiK-net, and F-net on land were used for the estimation. The SMGAs estimated for the four earthquakes showed little overlap with a subducted seamount. The Brune stress drops of the SMGAs near the leading edge of the subducted seamount were higher than the others. There was little overlap in the location of the estimated SMGAs for the four M 7-class earthquakes. Because the three SMGAs of the 2025 M_J 6.6 earthquake were scattered to fill in the gaps between the SMGAs of the other three earthquakes, the duration of the strong motion records of the M_J 6.6 earthquake was the longest among the four earthquakes. The locations of the SMGAs were then compared with the existing slip models of the three (M_J 7.1, M_J 6.9, and M_J 6.7) earthquakes estimated by long-period waveform and/or geodetic data. As a result, the SMGAs in proximity to the rupture initiation points were located near the main slip area, while the other SMGAs were found to be located at the periphery or outside of the main-slip area. This outcome suggests a frequency-dependent rupture process, a phenomenon that has been observed in M 7-class Miyagiken-oki earthquakes and the 2011 M_w 9.0 Tohoku earthquake in the Japan Trench. The relationships between the seismic moment and the short-period spectral level, which is defined as the constant value of the acceleration source spectral amplitudes at frequencies higher than the corner frequency, for the M_J 7.1, M_J 6.9, and M_J 6.6 earthquakes were identical to the empirical relationships for the M 7-8 class interplate earthquakes in the Japan Trench and the Kuril Trench. It was pointed out in previous studies that the relationship for the 2011 M_w 9.0 Tohoku earthquake was almost identical to the extrapolation of this empirical relationship. The findings in this study have the potential to contribute to strong ground motion predictions of the M 8-9 Nankai Trough earthquakes.

Hyuga-nada is among the most seismically active regions in Japan. Numerous major interplate earthquakes of approximately magnitude 7 have occurred in the southern part of Hyuga-nada off southern Miyazaki Prefecture at approximately 30-year intervals. The Mj 7.1 earthquake of August 2024 revealed new details about the occurrence patterns of large earthquakes in this region. Aftershock analysis of the 2024 earthquake conducted in this study, combined with source process investigations from previous studies, revealed that the rupture area of the 2024 earthquake did not overlap with those of the October (Mj 6.9) and December (Mj 6.7) 1996 earthquakes. In contrast, comparison of tsunami and strong-motion observation data indicates that the 2024 earthquake exhibits remarkable similarities to the 1961 earthquake in both tsunami and strong-motion waveforms. Furthermore, the hypocenters of the 1931 and October 1996 earthquakes were spatially close, and comparable foreshock activity preceded each event. These observations collectively suggest that two distinct rupture zones exist in the southern part of Hyuga-nada, each of which accumulates slip-deficit and releases it in M 7-class earthquakes at ~60-year intervals, offset from one another by ~30 years during the past 100 years. The occurrence of M 7-class earthquakes at ~60-year intervals can be explained by applying a hierarchical asperity model to understand the slip budget. Based on this model, we interpret that the asperity responsible for M 7-class events accumulates slip-deficit corresponding to the estimated aseismic slip at the plate boundary—which accounts for approximately half of the total slip-deficit expected from the plate subduction rate—and releases this slip-deficit through M 7-class earthquakes at 60-year intervals. This hypothesis also explains why the Mj 6.6 earthquake of January 2025 had less than half the magnitude of the 7.0 earthquake that had been expected in previous studies. Only ~30 years had elapsed since the previous event, approximately half the ~60-year intervals, resulting in roughly half the accumulated slip-deficit. Meanwhile, the remaining slip-deficit that was not released by the M 7-class earthquakes accumulates in the background over much longer periods. This unreleased slip-deficit could lead to a large M 8-class earthquake similar to the 1662 Hyuga-nada earthquake in the future. To better understand the occurrence patterns of major earthquakes, it is essential to further analyze historical M 7-class earthquakes in southern Hyuga-nada, particularly the 1931 and 1961 events, and more accurately evaluate the slip budget at the plate boundary including offshore slow earthquakes. Additionally, comprehensive numerical simulations are needed to reproduce and verify the various slip phenomena reported in Hyuga-nada. These efforts will contribute to clarifying the spatial and temporal properties of seismicity in this region and to advancing long-term seismic hazard assessment for earthquake disaster mitigation.

This study investigates earthquakes whose epicenters were possibly located in the Hyūga-nada region during early modern Japan, particularly the Edo period, based on the characteristics of seismic intensity distributions from both felt reports and instrumental observations compiled by the Japan Meteorological Agency. Historical earthquake records from southern Kyushu have traditionally been limited due to the scarcity of surviving documents, and their number has not been considered sufficient for detailed seismic analysis. To address this gap, newly uncovered historical records were utilized in the present study. Earthquakes were selected for analysis when seismic shaking was confirmed in at least three of the following locations: Miyazaki, Oita, Kagoshima, and Uwajima in Ehime Prefecture. The estimation of epicenters was carried out while taking into account the reliability of each historical source. In addition to the two well-known earthquakes of 1662 and 1769, it was confirmed that the earthquakes of 1698 and 1844 showed intensity distributions similar to those typically associated with Hyūga-nada earthquakes. Moreover, previously unclassified earthquakes, such as the 1843 event newly revealed through recent historical research, were examined. These findings help fill significant gaps in the seismic record for southern and central Kyushu. This study contributes to a more comprehensive reevaluation of earthquake activity during early modern Japan and is expected to provide a solid foundation for future seismic hazard assessments.

In the Izu Peninsula located about 100 km southwest of Tokyo, an earthquake swarm occurred in March 1997. The Earthquake Research Institute of the University of Tokyo installed its own developed deep borehole comprehensive crustal activity observation device in a deep borehole drilled in the Izu Peninsula in 1995 to conduct observations. The device was equipped with strainmeters, tiltmeters, seismometers, magnetometers, and a thermometer. The swarm started within about 5 km of the observation point from 0:00 on March 3, 1997. The analysis was carried out based on the hypocenter data by the Japan Meteorological Agency and the continuously observed strain and tilt data. Determined hypocenters of the swarm revealed the existence of an almost rectangular earthquake blank area. From 0:00 to 10:00 on March 3, 1997, microearthquakes of magnitude less than M 2 occurred near 10 km deep around the observation point. These hypocenters moved eastward over time and ended near the edge of the blank area. After that, at 10:00, a vigorous swarm of earthquakes, including earthquakes of M 2 and above, began. From changes observed by the tiltmeters and the strainmeters, it was considered that the microearthquakes were caused by the pressure of magma rising from the depths, and it was estimated that the magma generated the microearthquakes penetrated into the solid medium from near the edge of the blank area and formed an area containing magma (magma intrusion area) connected with magma chamber in the depth. Due to the presence of a pressure source in the solid medium, stress accumulation progressed in the medium, and earthquakes began to happen from the parts where a stress accumulation progressed beyond a fracture limit, resulting in the earthquake swarm. Then in response to changes of magma pressure and state, the number of earthquakes increased, and the strain and the tilt continued to change and accelerate. In this paper, the authors describe the path from magma rising to magma intrusion, the formation process of the magma intrusion area, the three-dimensional structure of the magma intrusion area, changes in earthquake hypocenters, strain and tilt in response to changes in magma intrusion over time. At 23:09, a M 5.5 earthquake occurred at a distance of about 4 km from the borehole observation point and near the east boundary of the magma intrusion area. Three earthquakes of M 3 level occurred near the M 5.5 between 22:15 and the occurrence. Large changes of the principal strains were recorded from around 22:20, and large changes in the area strain continued before the earthquake. The direction of the downward tilt changed and accelerated from around 22:20. Considering both observed and analyzed results, we believe that the comprehensive observation in a deep borehole will make an important contribution to future seismic research.

This study examines disaster preparedness education that has been continuously conducted in junior high schools in Miyazaki Prefecture, with a particular focus on the development of student disaster leaders. A questionnaire survey conducted after the Hyuga-Nada earthquake in August 2024 revealed that students who had taken on leadership roles in disaster preparedness responded to the event more calmly and appropriately. These findings suggest that the accumulation of daily disaster education may positively influence students’ behavior in actual emergencies. The results highlight the significance of continuous and experiential learning in helping students connect classroom-based knowledge with real-world disaster situations.

Coseismic uplift at Murotsu Port near Cape Muroto has been a well-known tectonic feature characterizing the past M 8-class megathrust earthquakes along the Nankai trough, southwest Japan. Although the coseismic uplift at Murotsu Port for the 1946 M 8.0 Showa Nankai earthquake was often believed to be 1.15 m, the measurement of the uplift was conducted at Tsuro Port, 3 km south of Murotsu Port. We utilized leveling and sea-level measurement data from past studies to re-examine the vertical displacement and its uncertainty at Murotsu Port. We obtained four uplift estimates and their uncertainties depending on the measurement data and correction methods for the relative displacement at different locations, preseismic and postseismic displacements, and a possible difference in high-tide level. By combining these estimates with equal weight in the probability density function, we concluded that the coseismic uplift was ~ 1.02 m with estimated uncertainties of ~ 0.06 m at Murotsu Port for the 1946 Nankai earthquake.

We found “normal-like” abnormal records observed with STS-2 broadband seismometer. The anomalous records have the three-dimensional particle motion on a certain plane, which might be due to some instrumental failures. We also proposed a monitoring index based on the scalar triple product of three unit vectors sampled from the three-component records with regular time intervals to detect the abnormal waveforms. The value of the index becomes zero when the particle motion lies only on a plane or a line. We applied the proposed method to STS-2 records of one JMA station and one F-net station (NIED), and confirmed that our method gives us a straightforward way to identify the anomalous records of the STS-2.