Intraslab reverse faulting adjacent to the hypocenter of the 1923 Kanto earthquake: The Mw 5.0 western Kanagawa earthquake in eastern Japan on 9 August 2024

Intraslab reverse faulting adjacent to the hypocenter of the 1923 Kanto earthquake: The Mw 5.0 western Kanagawa earthquake in eastern Japan on 9 August 2024 | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Intraslab reverse faulting adjacent to the hypocenter of the 1923 Kanto earthquake: The Mw 5.0 western Kanagawa earthquake in eastern Japan on 9 August 2024 Yasunori Sawaki, Takahiro Shiina, Takahiko Uchide This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6353744/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 25 Nov, 2025 Read the published version in Earth, Planets and Space → Version 1 posted 5 You are reading this latest preprint version Abstract Western Kanagawa in eastern Japan is a region under complex tectonics with seismic activity. The northward subduction of Philippine Sea plate has historically caused M-8 class megathrust earthquakes, known as Kanto earthquakes. On 9 August 2024, an Mw 5.0 reverse-fault earthquake occurred near the epicenter of the latest Kanto earthquake in 1923. For future evaluation of the earthquake generation in this area, it is imperative to determine the precise location of this earthquake’s faulting: whether it occurred on the plate interface, an active fault, or within the slab. To tackle this question, we conducted a machine-learning-based workflow of (1) phase picking using PhaseNetWC, (2) hypocenter relocation with phase picks and waveform cross-correlations, and (3) extraction of rectangular fault planes through hypocenter clustering of positions and point-cloud normal vectors. Our result exhibited five fault planes. We obtained a significant plane dipping steeply to the south, consistent with the steeper nodal plane of the mainshock focal mechanism. This extracted plane, being 2 km deeper than the slab surface, demonstrates that the Mw 5.0 earthquake occurred as an intraslab earthquake on a steep reverse fault, as opposed to thrusting on the PHS surface or reverse faulting on an active fault in the forearc crust. Given the high stress rate on the plate interface in this area, complex stress states may have given rise to the occurrence of the steep reverse intraslab earthquake beneath highly coupled plate interface. Intraslab reverse fault Fault geometry Hypocenter clustering Machine learning Active fault Sagami Trough The 1923 Kanto earthquake Figures Figure 1 Figure 2 Figure 3 Figure 4 Introduction An earthquake with the local magnitude (Mj) 5.3 occurred in western Kanagawa Prefecture, eastern Japan, on 9 August 2024 at 7:57:38 pm JST (UTC + 9), leading to a maximum Japan Meteorological Agency (JMA) seismic intensity scale of 5 Lower in this region. The F-net moment tensor (MT) solution indicated a reverse-fault-type focal mechanism of Mw 5.0 at a depth of 11 km, characterized by a north–south compressional axis (Fig. 1 ). Following the mainshock, numerous aftershocks occurred at depths > 10 km based on the JMA Unified Catalog and extended 5 km laterally in this region, with the largest aftershock (Mw 4.1) on 15 August at 8:20:34 pm JST in a similar mechanism with the mainshock. In this area, the Philippine Sea plate (PHS) subducts northward beneath the overriding North American plate along the Sagami Trough. The PHS surface depth is approximately 10 km (Hirose et al. 2008b , a ) around the aftershock area. Given that the compressional axis is roughly consistent with the PHS subduction, and that one nodal plane of these focal mechanisms aligns well with the geometry of PHS, interplate thrusting was considered though the JMA’s aftershock distribution did not show clear low-angle alignments. The Tokyo Metropolitan area, including Kanagawa Prefecture, has repeatedly experienced huge earthquakes, known as “Great Kanto earthquakes”. The latest earthquake was often called the Taisho Kanto earthquake (Mw 7.9–8.2) on 1 September 1923 (herein, the 1923 Kanto earthquake). The 1923 Kanto earthquake was a megathrust-type earthquake. The rupture area of the 1923 Kanto earthquake widely covered the Tokyo Metropolitan area (Kanamori 1971 ; Wald and Somerville 1995 ; Kobayashi and Koketsu 2005 ; Matsu’ura et al. 2007 ) (Fig. 1 a), which resulted in the most devastating earthquake disaster in the Japan’s history. The rupture of the 1923 Kanto earthquake was considered to initiate in the western Kanagawa region (Kanamori and Miyamura 1970 ), near the epicenter of this Mw 5.0 earthquake. Thus, investigating whether this Mw 5.0 earthquake occurred on the plate interface would help advance our understanding of mechanisms for future great earthquakes. The tectonics of this area is characterized not only by the subduction of PHS but the northward collision of the Izu Peninsula, called the Izu collision zone (Fig. 1 a). The upper surface of PHS and Izu Peninsula has been estimated through active source surveys (Sato et al. 2005 ; Arai et al. 2009 , 2014 ), seismic tomography (Hirose et al. 2008b , a ; Nakajima et al. 2009 ), azimuthal anisotropy (Ishise et al. 2021 ), and receiver functions (Abe et al. 2023 ). The Izu Peninsula is considered as a part of PHS or a separated microplate (e.g. Mazzotti et al. 1999 ; Nishimura et al. 2018 ). The tectonic boundary between the western collision and eastern subduction (Yoshida 1993 ) has been investigated approximately 139.1 \(\:^\circ\:\) E from focal mechanisms (Yukutake et al. 2012 ), crustal structures (e.g. Abe et al. 2023 ; Honda et al. 2023 ), and a geodetic study (Doke et al. 2021 ). Yukutake et al. ( 2012 ) investigated the seismicity and focal mechanisms beneath the Tanzawa Mountains and concluded that earthquakes typically occur within and the upper surface of PHS in the western and eastern parts, respectively. This area has experienced M7-class crustal earthquakes that are often referred to as Odawara earthquakes (e.g. Ishibashi 2004 ). Doke et al. ( 2021 ) indicated that northward detachment of the northern Izu Peninsula (Seno 2005 ) produces a shear zone, north of which situates the hypocenter of the Mw 5.0 earthquake (Fig. 1 a). These studies suggested that right-lateral slip on Western Sagami-Bay Fracture may have caused the Odawara earthquakes (Ishibashi 2004 ; Doke et al. 2021 ). On the contrary, Ishida and Kikuchi ( 1992 ) proposed from intraslab earthquakes at Odawara at the south of the Mw 5.0 event that Odawara earthquakes occurred in the PHS slab. Thus, possible sources of Odawara earthquakes are still under debate. These suggest that the area of interest is in complex stress states. Additionally, there are several known active fault zones in this area (Fig. 1 b). The left-lateral Kita-Izu fault zone and right-lateral Western Sagami-Bay Fracture give the counterclockwise shear zone (Doke et al. 2021 ). The Kōzu–Matsuda fault zone in the northeast of the shear zone is thought to be a branch fault for PHS interface (Headquarters for Earthquake Research Promotion (HERP) 2015). The north-dipping Matsuda-kita fault (MKF) in Hirayama–Matsuda-kita fault zone (HM-FZ) with a fault length of 4–8 km is situated around the epicenters of the Mw 5.0 earthquake sequence (Fig. 1 b). MKF is considered a north-dipping reverse fault in 40–50 degrees, whose loading rate is estimated 0.8 m/kyr or faster (HERP 2015). HERP (2015) declared that HM-FZ is classified as a group with a relatively high probability (0.09–0.6%) of experiencing an earthquake within the next 30 years among the major active faults in our country. Although MKF is considered in the forearc crust above the PHS slab, it is difficult to deny the possibility of faulting on the active fault or in the PHS slab from the initial aftershock distribution. These circumstances raise the scientific question: where and how the faulting occurred during this Mw 5.0 earthquake—interplate thrusting, active fault, or intraslab faulting. This is crucial for comprehending the mechanism of earthquake generation in the region of interest. To tackle this question, we relocated micro-earthquakes in the target area and clarified the fault geometry of the Mw 5.0 earthquake through hypocenter clustering analysis with machine learning techniques. Data and Methods We conducted machine-learning-based phase picking, hypocenter relocation, and hypocenter clustering analysis to extract rectangular fault planes (FaultNVC; Sawaki et al. 2025b ) for the Mw 5.0 earthquake in western Kanagawa on 9 August 2024. Below outlines the data processing and analysis performed in this study. First, we picked P- and S-wave arrivals for each event in the JMA Unified Earthquake Catalog using a deep-neural-network-based picker of seismic arrival times (PhaseNet; Zhu and Beroza 2019 ), which was re-trained on the JMA catalog (PhaseNetWC; Naoi et al. 2024 ). This aimed to increase the number of picks available for hypocenter relocation. Using these picked phase arrivals along with JMA pick data and computing waveform cross-correlations (CCs), we relocated hypocenters using hypoDD (Waldhauser and Ellsworth 2000 ). Subsequently, we extracted rectangular fault planes in the relocated aftershock using FaultNVC (Sawaki et al. 2025b ). Phase picking using PhaseNetWC We collected 867 earthquake events shallower than 25 km from the JMA Unified Earthquake Catalog, which occurred within the region specified in Fig. 1 b from January 2004 to October 2024. The minimum JMA magnitude was set to 0.5. We also retrieved JMA picks of the collected events. However, manual picks are limited; many include automatic picks. Consequently, there is a limit to the number of quality-controlled picks. Therefore, we applied PhaseNet to obtain more arrival times for P- and S-wave first arrivals. PhaseNet was trained with seismic events in California (Zhu and Beroza 2019 ). However, recent work by Naoi et al. ( 2024 ) re-trained PhaseNet using seismograms recorded around the Japanese archipelago. Thus, their model (PhaseNetWC) was suitable for our purpose of picking phases. We retrieved velocity waveforms recorded by Hi-net from the National Research Institute for Earth Science and Disaster Resilience (NIED) (NIED 2019), JMA, Earthquake Research Institute, the University of Tokyo (ERI), and Hot Springs Research Institute of Kanagawa Prefecture (HSRI) (Fig. S1 and Table S1 ). Seismograms with a sampling rate different from 100 Hz were decimated to 100 Hz. The maximum epicentral distance was set to 100 km. Theoretical travel times ( \(\:{t}^{TP},\:{t}^{TS}\) ) were calculated based on the JMA2001 travel-time table (Ueno et al. 2002 ) and used as a reference for a time window of seismograms applied to PhaseNetWC. We tried to find candidate phase arrival times by using PhaseNetWC as we aimed to increase the number of picks. JMA picks ( \(\:{t}^{DP},\:{t}^{DS}\) ) are available as a reference for time windows in some cases. Thus, three cases were considered: (A) seismograms without \(\:{t}^{DP}\) nor \(\:{t}^{DS}\) ; (B) without \(\:{t}^{DP}\) but with \(\:{t}^{DS}\) ; and (C) with \(\:{t}^{DP}\) . For case (A), we set a time window of [-15, 15] seconds centered on \(\:{t}^{TP}\) and ran PhaseNetWC to three-component waveforms for a tentative P-wave arrival time ( \(\:{\tau\:}^{\text{t}\text{n}\text{t}}\) ). No filter was applied to run PhaseNetWC. \(\:{\tau\:}^{\text{t}\text{n}\text{t}}\) was searched within ± 5 seconds of \(\:{t}^{TP}\) , maximizing the probability of the P-wave arrival ( \(\:{p}_{P}\) ). If the maximum \(\:{p}_{P}\) was 0.5 or greater, we used \(\:{\tau\:}^{\text{t}\text{n}\text{t}}\) instead of \(\:{t}^{TP}\) for another run of PhaseNetWC for an S-wave arrival. If the maximum \(\:{p}_{P}\) was less than 0.5, \(\:{t}^{TP}\) was used as \(\:{\tau\:}^{\text{t}\text{n}\text{t}}\) . Then, PhaseNetWC was applied again by setting a time window of [-5, 25] seconds from \(\:{\tau\:}^{\text{t}\text{n}\text{t}}\) . From an example case, PhaseNetWC was able to pick proper phase arrivals (Fig. S2 ). For cases (B) and (C), we tried to inspect JMA picks more precisely. For case (B), \(\:{t}^{DS}\) was given priority over \(\:{t}^{TS}\) , and we ran PhaseNetWC by setting a time window of [-20, 10] seconds from \(\:{t}^{DS}\) . For case (C), \(\:{t}^{DP}\) and \(\:{t}^{DS}\) were given priority over \(\:{t}^{TP}\) and \(\:{t}^{TS}\) , and we ran PhaseNetWC by setting a time window of [-5, 25] seconds from \(\:{t}^{DP}\) . Even though JMA picks are available, PhaseNetWC picks can correct proper arrivals (Fig. S2 ). Finally, we created a pick table as we executed PhaseNetWC for all cases. The pick times ( \(\:{\tau\:}^{P},\:{\tau\:}^{S}\) ), on which the probability was maximized, were searched. If the maximum probability ( \(\:{p}_{\text{m}\text{a}\text{x}}\) ) was 0.5 or greater, we took \(\:\tau\:\) for a phase pick. If \(\:{p}_{\text{m}\text{a}\text{x}}\) was less than 0.5, we checked if the JMA pick \(\:{t}^{D}\) was available. If available, we took it for a phase pick instead of \(\:\tau\:\) ; otherwise, we discarded the phase. In total, we obtained 87,154 picks from 157 stations: 47,284 picks for P and 39,870 picks for S. This is more than three times the picks obtained by JMA: 13,955 picks for P; 14,230 picks for S; and 28,185 picks in total from 77 stations. The increase in phase picks is more evident for stations distant from the target area and stations omitted by JMA (Fig. S3 ). Hypocenter relocation We relocated the hypocenters using hypoDD (Waldhauser and Ellsworth 2000 ) with the JMA2001 model. To create a differential travel-time table, we set the maximum inter-event distance of 3.0 km. The pick table yielded 352,608 and 329,632 differential times for P- and S-waves, respectively, when allowing 30 neighboring events. Additionally, we computed CC of seismograms from an event pair within 2 km both laterally and in depth, and extracted lapse times with maximized CC values as differential times. We applied the bandpass filter of 4–12 Hz and extracted a time window before and after 1.0 s of the pick time ( \(\:{\tau\:}^{P}\) or \(\:{\tau\:}^{S}\) ). For P-waves, the CC was computed between vertical components. For S waves, we considered calculating the particle motion of two horizontal components to obtain an average-like correlation of the two components, instead of adopting the horizontal component with better CC values (e.g. Yoshida et al. 2023 ) (Fig. S4 ). Although computed CC values tend to be smaller than those in the conventional way, this way is robust to the horizontal rotation of seismographs. To be exact, complex waveforms were generated from the two horizontal components (north and east), and the CC function was calculated between these complex waveforms: \(\:\begin{array}{c}{C}_{ijk}^{S}\left(t\right)={z}_{ik}\left(t\right)\star\:{z}_{jk}^{*}\left(t\right),\#\left(1\right)\end{array}\) where \(\:\star\:\) represents the calculation of CC, \(\:{C}_{ijk}^{S}\left(t\right)\) denotes the S-wave CC function between the i -th and j -th events at the k -th station. In Eq. 1, \(\:{z}^{\text{*}}\) denotes the conjugate of \(\:z\) . \(\:{z}_{ik}\left(t\right)\) represents a horizontal complex waveform for the i -th event at the k -th station: \(\:\begin{array}{c}{z}_{ik}\left(t\right)={u}_{ik}^{\text{E}}\left(t\right)+j{u}_{ik}^{\text{N}}\left(t\right),\#\left(2\right)\end{array}\) where \(\:{u}^{\text{E}}\left(t\right)\) and \(\:{u}^{\text{N}}\left(t\right)\) denote the east- and north-component waveforms, respectively. In Eq. 2, j denotes the imaginary unit. Note that we did not correct horizontal sensor azimuths because CCs were computed for waveforms at the same stations. We defined the S-wave CC value as the real part of Eq. 1. Finally, pairs with a CC of 0.8 or greater and an absolute differential time of 1.0 s or shorter were selected for each phase. We obtained 488,876 and 193,694 pairs for P- and S-waves, respectively. Using differential times from picks and CCs, we executed hypoDD under the JMA2001 one-dimensional velocity model (Ueno et al. 2002 ). We obtained the final result after 25 relocation iterations. In total, 864 out of 867 events were relocated from January 2004 to October 2024 (Table S2 ). Hypocenter clustering for fault planes (FaultNVC) As we focus on the earthquake sequence of the Mw 5.0 event, we used 364 events after 9 August 2024 at 6 pm JST for FaultNVC. We excluded the mainshock and events shallower than 10 km or deeper than 16 km. For a clustering purpose, these hypocenters in the geographic coordinate of World Geodetic System 1984 (the European Petroleum Survey Group (EPSG) code of 4326) were transformed into a local cartesian coordinate system. We used the EPSG code of 6677 as the local cartesian coordinate. We defined the center of the local cartesian coordinate as the longitude of \(\:139.160^\circ\:\text{E}\) and the latitude of \(\:35.410^\circ\:\text{N}\) by applying a parallel shift to EPSG 6677 for simplicity. For the transformed hypocenters in the Cartesian coordinate system, we applied FaultNVC, which extracts rectangular fault planes through hypocenter clustering (Sawaki et al. 2025b ). This method considers hypocenters as point-cloud data and performs simultaneous clustering of point-cloud normal vectors and hypocenter positions. Following their study, we conducted three steps: (STEP 1) calculating PCNVs; (STEP 2) performing feature-vector clustering using HDBSCAN (Campello et al. 2013 ) to extract rectangular fault planes; and (STEP 3) controlling the quality of these planes. We summarized the procedure in Text S1 and listed the parameter set in Table S3 . Results and discussion The relocated hypocenters are shown in Fig. 2 a. Hypocenter focal depths ranged from 12 to 14.5 km. Around the mainshock hypocenter, aftershocks occurred approximately 14 km depth. In contrast, several hundred meters to the north, the focal depth suddenly shallowed to 12.5 km. Furthermore, southwest of the mainshock, the focal depth approached 12 km. Compared to the original JMA catalog, hypocenters were aligned to show planar features though slightly shifted to east (Fig. S5 ). We extracted five rectangular fault planes (Planes #1, #2, #4, #5, and #8) through FaultNVC after quality control (Fig. 2 b), with fault parameters shown in Table 1 . The upper edge depth of these planes is nearly 12 km. Plane #4 in the northeastern area is a south-dipping plane, which consists of numerous aftershocks. The dip angle was estimated 68 \(\:^\circ\:\) . Plane #1 is situated several hundred meters south of Plane #4 and is similarly dipping to the south. The cluster of Plane #1 consists of 17 events. The fault parameters are also nearly identical to those of Plane #4. In contrast, Plane #5 to the west of Plane #4 is dipping towards the WNW direction. Plane #8, located to the west of Plane #1, is dipping towards the east direction. The extracted planes are consistent with hypocenter alignments characterized by point-cloud normal vectors (Fig. S6 ), though some clusters were rejected by quality control (Fig. S7). Table 1 Fault parameters extracted using FaultNVC (Fig. 2 b). The reference point is set to the upper left corner. ID Cluster size Latitude [ \(\:^\circ\:\) N] Longitude [ \(\:^\circ\:\) E] Depth [km] Strike [ \(\:^\circ\:\) ] Dip [ \(\:^\circ\:\) ] Length [km] Width [km] 1 17 35.4068 139.1598 12.53 81.7 67.9 1.10 0.83 2 11 35.4082 139.1469 12.40 312.9 24.9 0.62 0.62 4 130 35.4156 139.1575 12.51 84.3 67.5 1.57 2.12 5 17 35.4134 139.1606 12.66 201.0 70.0 0.54 0.53 8 21 35.3996 139.1532 12.38 350.8 29.1 0.48 0.54 We examined cross-sectional views for these extracted fault planes (Figs. 2 c– 2 f). Hypocenter alignments and steeply dipping structures both for Planes #4 and #1 were observed in the cross-sectional view along the P–p and Q-q lines (Figs. 2 c and 2 d). Plane #4 is consistent with the steeper nodal plane of the mainshock focal mechanism, which is different from the one parallel to the PHS surface. Considering the steep mainshock fault geometry and the PHS surface depth of 10 km, our results demonstrate that the mainshock occurred as an intraslab earthquake on a steep reverse fault, as opposed to thrusting on the PHS surface or reverse faulting on MKF in the forearc crust. The slab Moho depth was investigated to be 20–30 km from P-wave tomography or receiver functions (Ishise et al. 2021 ; Abe et al. 2023 ). Thus, this earthquake probably occurred in the PHS crust. We also focus on the largest (Mw 4.1) aftershock. For cross-sectional views along the R–r line, Plane #8 dipping east roughly is consistent with hypocenter alignments (Fig. 2 e). However, Plane #8 is not consistent with the nodal planes of the Mw 4.1 aftershock. The cross-section S–s revealed a clear depth difference between Planes #8 and #5: Plane #5 deeper than Plane #8. This is attributed to the seismic gap between these planes at the horizontal position of 1.4–1.6 km. In contrast to Plane #8, Plane #1 is likely associated with the steeper nodal plane of the Mw 4.1 aftershock, although this aftershock event does not belong to Plane #1. To demonstrate that, we examined the event occurrence time and showed a depth view around Plane #1 (Fig. 3 ). Event counts around Plane #3 are apparently smaller after the occurrence of the largest aftershock than before that (Fig. 3 b). In contrast, those around Plane #1 are relatively higher after the occurrence of the Mw 4.1 aftershock (Fig. 3 c). This suggests that aftershocks of the Mw 4.1 aftershock occurred near Plane #1. Additionally, the depth view of hypocenters along the \(\:{X}_{1}\) axis exhibited that one nodal plane is highly consistent with Plane #1 with a slight gap (Fig. 3 d). The fault scale of approximately 1 km for the focused area is comparable to the characteristic fault length of an Mw 4-class earthquake. Therefore, the fault slip of the Mw 4.1 aftershock probably occurred on Plane #1 or on its extension. The mainshock fault slip may not be so simple as the pure reverse faulting on Plane #4. The F-net MT solution has 50% non-double-couple components. Possible coseismic slip on extended planes such as Planes #5 and #8 other than #4 might have contributed to the high non-double-couple components. In the JMA catalog, six foreshocks with Mj of 0.7–2.1 were included, starting approximately one hour before the Mw 5.0 mainshock. The clustering analysis for the relocated catalog showed that all the foreshocks were included in Plane #4, shallower part of the mainshock hypocenter (Fig. 2 c). This may indicate that minor precursory slips had occurred on the principal plane (#4) before the mainshock slip and minor stress change or fluid migration occurred towards the mainshock hypocenter. Although we did not analyze the temporal variation of seismicity, the hypocenter clustering method can clarify the spatial relationships between foreshocks, mainshocks, and aftershocks. This will help in understanding the evolution of fault slip during an earthquake sequence. We discuss the possible causes of steep reverse intraslab earthquakes and their implications for future great earthquakes. The occurrence of steep and reverse intraslab earthquakes in the upper part of the slab is crucial key to understanding the stress state around the seismogenic zone. Generally, the faulting type of intraslab earthquakes reflects the state of stress and pore fluid pressure in the oceanic plate (e.g. Kirby 1995 ; Seno and Yoshida 2004 ; Wada et al. 2010 ). Outer rise normal faults are produced in extensional fields due to plate bending (Christensen and Ruff 1988 ) and can be re-activated as reverse faults in compressional fields due to slab compression along the upper layer of double seismic zones (e.g. Hasegawa et al. 1978 ; Yamasaki and Seno 2003 ; Ohta et al. 2011 ). In the Japan Trench, great reverse intraslab earthquakes have occurred around the downdip portion of coseismic slip area of large interplate earthquakes (Okada and Hasegawa 2003 ; Ohta et al. 2011 ). Ohta et al. ( 2011 ) investigated the coseismic fault model of the reverse intraslab event of M 7.1 on 7 April 2011 after the Mw 9.0 Tohoku-oki earthquake on 11 March 2011 and proposed that a pre-existing outer-rise normal fault was re-activated in a reversed sense with downdip compressional stress. The reverse faulting of this Mw 5.0 event may indicate a compressional stress field in the upper part of PHS. This area is in complex tectonics due to the subduction of PHS and the collision of the Izu Peninsula. The regional stress field, estimated by stress tensor inversion using micro-earthquake focal mechanisms, exhibited an NW–SE-trending maximum horizontal stress axis (Imanishi et al. 2019 ; Uchide et al. 2022 ). This is roughly consistent to cause reverse faulting on an E–W trending dip fault. Additionally, the plate coupling ratio is high or middle around the Odawara area, which was derived from GNSS data (Nishimura et al. 2018 ). This area is characterized by a high stress rate (Fig. 4 ), which can initiate the Taisho-type Kanto earthquake (Saito and Noda 2023 ). These studies may indicate that the PHS subduction has accumulated stress on the plate interface near the focal area of this Mw 5.0 earthquake since the last Taisho Kanto earthquake in 1923. Arai and Iwasaki ( 2015 ) pointed out that intraslab earthquakes tend to occur beneath strongly coupled plate interface and a reflective zone is located at the downdip end of it. Given that the coupled plate interface is willing to host intraslab earthquakes rather than interplate earthquakes and release the slab-loading stress, these complex and accumulated stresses may have given rise to the occurrence of the reverse intraslab earthquake near the hypocenter of the 1923 Kanto earthquake (Fig. 4 ). Monitoring intraslab stress states would help understand faulting properties around seismogenic zones. A seismic cycle simulation indicated that the occurrence of an interplate earthquake was advanced due to stress perturbations of an intraslab earthquake (Kato 2004 ). In the Nankai Trough, the stress orientation at the updip portion of deep slow-earthquake area, estimated from upper-plane reverse intraslab events, rotated after slow earthquake events of episodic tremor and slip (Kita et al. 2021 ). They inferred that fluid migration from slab to interface promoted slow earthquakes on the plate interface. In a similar manner, further research needs to investigate the influence of reverse intraslab slips on the plate interface conditions. Furthermore, the possible source of Odawara earthquakes might originate from the PHS slab (Ishida and Kikuchi 1992 ), despite undergoing debates and the spatial offset from the 2024 Mw 5.0 event. While Yukutake et al. ( 2012 ) suggested that the seismicity in the east of the tectonic boundary beneath the Tanzawa Mountains is induced by the PHS subduction process, it has not been clarified whether this seismicity in the north of our target area includes intraslab events. Although we did not perform local stress inversions, further studies that monitor the seismicity including these intraslab earthquakes and slab stress states are demanded to understand interplate faulting and crustal earthquakes, assessing the future occurrence of Kanto and Odawara earthquakes. Conclusions We identified fault planes of the Mw 5.0 earthquake in western Kanagawa region in eastern Japan on 9 August 2024 to understand its faulting mechanism. We conducted a machine-learning-based workflow of phase picking, hypocenter relocation, and fault plane identification through hypocenter clustering. The extracted fault planes included a south-dipping mainshock plane in a steep angle and a minor south-dipping plane parallel to the mainshock plane that probably consists of the largest (Mw 4.1) aftershock. This extracted plane, being 2 km deeper than the slab surface, demonstrates that the Mw 5.0 earthquake occurred as an intraslab earthquake on a steep reverse fault, as opposed to thrusting on the PHS surface or reverse faulting on an active fault in the forearc crust. Given the high stress rate on the plate interface in this area, complex stress states may have given rise to the occurrence of the steep reverse intraslab earthquake beneath highly coupled plate interface. Careful monitoring of intraslab earthquakes and slab stress states would help assess the future occurrence of Kanto earthquakes and great regional events. Abbreviations CC Cross correlation EPSG European Petroleum Survey Group ERI Earthquake Research Institute, the University of Tokyo GNSS Global Navigation Satellite System GSJ Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology (AIST), Japan HDBSCAN Hierarchical Density-Based Spatial Clustering of Applications with Noise HERP Headquarters for Earthquake Research Promotion, an extraordinary organ of MEXT HM-FZ Hirayama–Matsuda-kita fault zone HSRI Hot Springs Research Institute of Kanagawa Prefecture JMA Japan Meteorological Agency MEXT Ministry of Education, Culture, Sports, Science and Technology MKF Matsuda-kita fault MT moment tensor NIED National Research Institute for Earth Science and Disaster Resilience, Japan PCA principal component analysis Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Availability of data and materials The original hypocenter catalog is available on the website of the JMA Unified Earthquake Catalog (https://www.data.jma.go.jp/svd/eqev/data/bulletin/index_e.html), which was last accessed in December 2024. The NIED F-net MT solutions were retrieved from their page (https://www.fnet.bosai.go.jp/event/search.php?LANG=en) with ids of 20240809105600 and 20240815111900. We retrieved seismic waveforms from NIED Hi-net (NIED 2019), JMA, ERI, and HSRI through the NIED Hi-net’s website with a registration required (https://www.hinet.bosai.go.jp/?LANG=en). We used Python software FaultNVC available at the GSJ’s Open-File Report (Sawaki et al. 2025a). We used SeisBench (Woollam et al. 2022) to run PhaseNet (Zhu and Beroza 2019), available at GitHub (https://github.com/seisbench/seisbench). The weight for PhaseNetWC (Naoi et al. 2024) can be available at GitHub (https://github.com/mktnaoi/JMAuniPicker). We retrieved the computing code of hypoDD from the author’s page (https://www.ldeo.columbia.edu/~felixw/hypoDD.html), last accessed 17 January 2025. For visualizations, we used Matplotlib (Hunter 2007), colormaps provided by Colorcet (https://colorcet.holoviz.org/), and Cartopy v0.24 (Elson et al. 2024). The data generated or analyzed during this study, including the clustering result (Table S4 and S5), are contained in supplementary information files. Competing interests The authors declare that they have no competing interests. Funding This study was supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) Project for Seismology toward Research Innovation with Data of Earthquake (STAR-E) Grant Number JPJ010217. Authors' contributions YS developed methods, analyzed the data, discussed the results, wrote and revised the original manuscript, and created all figures and tables. TS and TU discussed the results. TU developed methods, obtained funding, and supervised YS as a principal investigator of our research topic in the STAR-E Project. All authors have read and approved the final manuscript. Acknowledgements We had fruitful discussions with Ryo Kurihara, Kodai Sagae, and Haruo Horikawa. Authors’ details YS, TS, TU: Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology (AIST), AIST Tsukuba Central 7, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8567, Japan Present address for YS: College of Science and Engineering, Ritsumeikan University, 1‑1‑1 Noji-Higashi, Kusatsu, Shiga 525‑8577, Japan References Abe Y, Honda R, Ishise M et al (2023) Relationship Between Crustal Structure and Plate Convergence Around the Izu Collision Zone in Central Japan. J Geophys Res Solid Earth 128. https://doi.org/10.1029/2022JB026314 . e2022JB026314 Arai R, Iwasaki T (2015) Transition from collision to subduction and its relation to slab seismicity and plate coupling. Earth Planets and Space 67:76. https://doi.org/10.1186/s40623-015-0247-6 Arai R, Iwasaki T, Sato H et al (2009) Collision and subduction structure of the Izu–Bonin arc, central Japan, revealed by refraction/wide-angle reflection analysis. Tectonophysics 475:438–453. https://doi.org/10.1016/j.tecto.2009.05.023 Arai R, Iwasaki T, Sato H et al (2014) Contrasting subduction structures within the Philippine Sea plate: Hydrous oceanic crust and anhydrous volcanic arc crust. Geochem Geophys Geosyst 15:1977–1990. https://doi.org/10.1002/2014GC005321 Campello RJGB, Moulavi D, Sander J (2013) Density-Based Clustering Based on Hierarchical Density Estimates. Advances in Knowledge Discovery and Data Mining. Springer, Berlin, Heidelberg, pp 160–172. https://doi.org/10.1007/978-3-642-37456-2_14 Christensen DH, Ruff LJ (1988) Seismic coupling and outer rise earthquakes. J Geophys Res Solid Earth 93:13421–13444. https://doi.org/10.1029/JB093iB11p13421 Doke R, Honda R, Harada M et al (2021) Deformation of the seismogenic zone in the northeastern part of the Izu Peninsula, Japan, inferred from GNSS observations. Characterization of Modern and Historical Seismic–Tsunamic Events, and Their Global–Societal Impacts. Geological Society of London, pp 111–129. https://doi.org/10.1144/SP501-2019-104 Elson P, de Andrade ES, Lucas G et al (2024) SciTools/cartopy: REL: v0.24.1. https://doi.org/10.5281/zenodo.13905945 . Zenodo Hasegawa A, Umino N, Takagi A (1978) Double-planed structure of the deep seismic zone in the northeastern Japan arc. Tectonophysics 47:43–58. https://doi.org/10.1016/0040-1951(78)90150-6 Headquarters for Earthquake Research Promotion (2015) Long-term evaluation of the Shiozawa fault zone, Hirayama-Matsuda North fault zone, and Kōzu-Matsuda fault zone (Kan’nawa-Kōzu-Matsuda fault zone) (Second Edition) [in Japanese]. https://www.jishin.go.jp/main/chousa/katsudansou_pdf/36_kannawa_kouzu_4.pdf Hirose F, Nakajima J, Hasegawa A (2008a) Three-dimensional seismic velocity structure and configuration of the Philippine Sea slab in southwestern Japan estimated by double-difference tomography. J Geophys Res Solid Earth 113:B09315. https://doi.org/10.1029/2007JB005274 Hirose F, Nakajima J, Hasegawa A (2008b) Three-Dimensional Velocity Structure and Configuration of the Philippine Sea Slab beneath Kanto District, Central Japan, Estimated by Double-Difference Tomography [in Japanese with English abstract]. Zisin (Journal Seismological Soc Japan 2nd ser) 60:123–138. https://doi.org/10.4294/zisin.60.123 Honda R, Abe Y, Doke R (2023) A Review of Recent Research on Crustal Structures and Seismo-tectonics of the Izu Collision Zone [in Japanese with English abstract]. Zisin (Journal Seismological Soc Japan 2nd ser) 76:135–148. https://doi.org/10.4294/zisin.2023-9S Hunter JD (2007) Matplotlib: A 2D Graphics Environment. Comput Sci Eng 9:90–95. https://doi.org/10.1109/MCSE.2007.55 Imanishi K, Uchide T, Ohtani M et al (2019) Construction of the Crustal Stress Map in the Kanto Region, central Japan [in Japanese with English abstract]. Bull Geol Surv Japan 70:273–298. https://doi.org/10.9795/bullgsj.70.273 Ishibashi K (2004) Seismotectonic modeling of the repeating M 7-class disastrous Odawara earthquake in the Izu collision zone, central Japan. Earth Planets and Space 56:843–858. https://doi.org/10.1186/BF03353091 Ishida M, Kikuchi M (1992) A possible foreshock of a future large earthquake near Odawara, center Japan. Geophys Res Lett 19:1695–1698. https://doi.org/10.1029/92GL01503 Ishise M, Kato A, Sakai S et al (2021) Improved 3-D P Wave Azimuthal Anisotropy Structure Beneath the Tokyo Metropolitan Area, Japan: New Interpretations of the Dual Subduction System Revealed by Seismic Anisotropy. J Geophys Res Solid Earth 126:e2020JB021194. https://doi.org/10.1029/2020JB021194 Kanamori H (1971) Faulting of the Great Kanto Earthquake of 1923 as Revealed by Seismological Data. Bulletin of the Earthquake Research Institute. Univ Tokyo 49:13–18. https://doi.org/10.15083/0000033257 Kanamori H, Miyamura S (1970) Seismometrical Re-Evaluation of the Great Kanto Earthquake of September 1, 1923. Bulletin of the Earthquake Research Institute. Univ Tokyo 48:115–125. https://doi.org/10.15083/0000033322 Kato N (2004) A possible effect of an intermediate depth intraslab earthquake on seismic cycles of interplate earthquakes at a subduction zone. Earth Planets and Space 56:553–561. https://doi.org/10.1186/BF03352516 Kirby S (1995) Interslab earthquakes and phase changes in subducting lithosphere. Rev Geophys 33:287–297. https://doi.org/10.1029/95RG00353 Kita S, Houston H, Yabe S et al (2021) Effects of episodic slow slip on seismicity and stress near a subduction-zone megathrust. Nat Commun 12:7253. https://doi.org/10.1038/s41467-021-27453-8 Kobayashi R, Koketsu K (2005) Source process of the 1923 Kanto earthquake inferred from historical geodetic, teleseismic, and strong motion data. Earth Planets and Space 57:261–270. https://doi.org/10.1186/BF03352562 Matsu’ura M, Noda A, Fukahata Y (2007) Geodetic data inversion based on Bayesian formulation with direct and indirect prior information. Geophys J Int 171:1342–1351. https://doi.org/10.1111/j.1365-246X.2007.03578.x Mazzotti S, Henry P, LePichon X, Sagiya T (1999) Strain partitioning in the zone of transition from Nankai subduction to Izu–Bonin collision (Central Japan): implications for an extensional tear within the subducting slab. Earth Planet Sci Lett 172:1–10. https://doi.org/10.1016/S0012-821X(99)00189-2 Nakajima J, Hirose F, Hasegawa A (2009) Seismotectonics beneath the Tokyo metropolitan area, Japan: Effect of slab-slab contact and overlap on seismicity. J Geophys Res 114:B08309. https://doi.org/10.1029/2008JB006101 Nakata T, Imaizumi T (2002) Digital active fault map of Japan (DVD-ROM) [in Japanese]. University of Tokyo Naoi M, Tamaribuchi K, Shimojo K et al (2024) Neural phase picker trained on the Japan meteorological agency unified earthquake catalog. Earth Planets and Space 76:150. https://doi.org/10.1186/s40623-024-02091-8 National Research Institute for Earth Science and Disaster Resilience (2019) NIED Hi-net. https://doi.org/10.17598/nied.0003 Nishimura T, Yokota Y, Tadokoro K, Ochi T (2018) Strain partitioning and interplate coupling along the northern margin of the Philippine Sea plate, estimated from Global Navigation Satellite System and Global Positioning System-Acoustic data. Geosphere 14:535–551. https://doi.org/10.1130/GES01529.1 Ohta Y, Miura S, Ohzono M et al (2011) Large intraslab earthquake (2011 April 7, M 7.1) after the 2011 off the Pacific coast of Tohoku Earthquake (M 9.0): Coseismic fault model based on the dense GPS network data. Earth Planets and Space 63:1207–1211. https://doi.org/10.5047/eps.2011.07.016 Okada T, Hasegawa A (2003) The M7.1 May 26, 2003 off-shore Miyagi Prefecture Earthquake in northeast Japan: Source process and aftershock distribution of an intra-slab event. Earth Planets and Space 55:731–739. https://doi.org/10.1186/BF03352482 Saito T, Noda A (2023) Mechanically Coupled Areas on the Plate Interface in the Kanto Region, Central Japan, Generating Great Earthquakes and Slow-Slip Events. Bull Seismol Soc Am 113:1842–1855. https://doi.org/10.1785/0120230073 Sato H, Hirata N, Koketsu K et al (2005) Earthquake Source Fault Beneath Tokyo. Science (1979) 309:462–464. https://doi.org/10.1126/science.1110489 Sawaki Y, Sato Y, Uchide T (2025a) FaultNVC: Earthquake Fault Plane Extractor through Point-Cloud Normal Vector Clustering for Hypocenter Distributions. Open-File Report of the Geological Survey of Japan, AIST 759. https://www.gsj.jp/publications/pub/openfile/openfile0759.html Sawaki Y, Shiina T, Sagae K et al (2025b) Fault Geometries of the 2024 Mw 7.5 Noto Peninsula Earthquake from Hypocenter-Based Hierarchical Clustering of Point-Cloud Normal Vectors. J Geophys Res Solid Earth 130:e2024JB030233. https://doi.org/10.1029/2024JB030233 Seno T (2005) Izu detachment hypothesis: A proposal of a unified cause for the Miyake-Kozu event and the Tokai slow event. Earth Planets and Space 57:925–934. https://doi.org/10.1186/BF03351872 Seno T, Yoshida M (2004) Where and why do large shallow intraslab earthquakes occur? Phys Earth Planet Inter 141:183–206. https://doi.org/10.1016/j.pepi.2003.11.002 Uchide T, Shiina T, Imanishi K (2022) Stress map of Japan: detailed nationwide crustal stress field inferred from focal mechanism solutions of numerous microearthquakes. J Geophys Res Solid Earth 127. https://doi.org/10.1029/2022JB024036 . e2022JB024036 Ueno H, Hatakeyama S, Aketagawa T et al (2002) Improvement of hypocenter determination procedures in the Japan Meteorological Agency (in Japanese with English abstract). Q J Seismology 65:123–134 Wada I, Mazzotti S, Wang K (2010) Intraslab Stresses in the Cascadia Subduction Zone from Inversion of Earthquake Focal Mechanisms. Bull Seismol Soc Am 100:2002–2013. https://doi.org/10.1785/0120090349 Wald DJ, Somerville PG (1995) Variable-slip rupture model of the great 1923 Kanto, Japan, earthquake: Geodetic and body-waveform analysis. Bull Seismol Soc Am 85:159–177. https://doi.org/10.1785/BSSA0850010159 Waldhauser F, Ellsworth WL (2000) A Double-Difference Earthquake Location Algorithm: Method and Application to the Northern Hayward Fault, California. Bull Seismol Soc Am 90:1353–1368. https://doi.org/10.1785/0120000006 Woollam J, Münchmeyer J, Tilmann F et al (2022) SeisBench—A Toolbox for Machine Learning in Seismology. Seismol Res Lett 93:1695–1709. https://doi.org/10.1785/0220210324 Yamasaki T, Seno T (2003) Double seismic zone and dehydration embrittlement of the subducting slab. J Geophys Res Solid Earth 108:2212. https://doi.org/10.1029/2002JB001918 Yoshida A (1993) Seismic Activities and Tectonics in and around Western Kanagawa [in Japanese]. J Geogr (Chigaku Zasshi) 102:407–417. https://doi.org/10.5026/jgeography.102.4_407 Yoshida K, Uchida N, Matsumoto Y et al (2023) Updip Fluid Flow in the Crust of the Northeastern Noto Peninsula, Japan, Triggered the 2023 M w 6.2 Suzu Earthquake During Swarm Activity. Geophys Res Lett 50:e2023GL106023. https://doi.org/10.1029/2023GL106023 Yukutake Y, Takeda T, Honda R, Yoshida A (2012) Seismotectonics in the Tanzawa Mountains area in the Izu-Honshu collision zone of central Japan, as revealed by precisely determined hypocenters and focal mechanisms. Earth Planets and Space 64:269–277. https://doi.org/10.5047/eps.2011.09.002 Zhu W, Beroza GC (2019) PhaseNet: A Deep-Neural-Network-Based Seismic Arrival Time Picking Method. Geophys J Int 216:261–273. https://doi.org/10.1093/gji/ggy423 Supplementary Files SawakiEPS2025SagamiSM.docx AdditionalFile2TableS1stations100kmtable.csv AdditionalFile3TableS2hypoddreloc.csv AdditionalFile4TableS415000505050000events.csv AdditionalFile5TableS515000505050000planes.csv graphicalabstract.png SupplementaryInformation.docx Cite Share Download PDF Status: Published Journal Publication published 25 Nov, 2025 Read the published version in Earth, Planets and Space → Version 1 posted Editorial decision: Major Revision 18 Jun, 2025 Reviewers agreed at journal 01 May, 2025 Reviewers invited by journal 01 May, 2025 Editor assigned by journal 04 Apr, 2025 First submitted to journal 02 Apr, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6353744","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":450568446,"identity":"022bb3bd-fe82-4fff-9141-7c8c1685974f","order_by":0,"name":"Yasunori Sawaki","email":"data:image/png;base64,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","orcid":"https://orcid.org/0000-0002-4043-3391","institution":"National Institute of Advanced Industrial Science and Technology Geological Survey of Japan: Sangyo Gijutsu Sogo Kenkyujo Chishitsu Chosa Sogo Center","correspondingAuthor":true,"prefix":"","firstName":"Yasunori","middleName":"","lastName":"Sawaki","suffix":""},{"id":450568447,"identity":"85d3c733-6294-4310-a880-0438b7d81eff","order_by":1,"name":"Takahiro Shiina","email":"","orcid":"","institution":"National Institute of Advanced Industrial Science and Technology Geological Survey of Japan: Sangyo Gijutsu Sogo Kenkyujo Chishitsu Chosa Sogo Center","correspondingAuthor":false,"prefix":"","firstName":"Takahiro","middleName":"","lastName":"Shiina","suffix":""},{"id":450568448,"identity":"22132016-04be-4451-9836-68890f6d8f3d","order_by":2,"name":"Takahiko Uchide","email":"","orcid":"","institution":"National Institute of Advanced Industrial Science and Technology Geological Survey of Japan: Sangyo Gijutsu Sogo Kenkyujo Chishitsu Chosa Sogo Center","correspondingAuthor":false,"prefix":"","firstName":"Takahiko","middleName":"","lastName":"Uchide","suffix":""}],"badges":[],"createdAt":"2025-04-01 13:48:56","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6353744/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6353744/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1186/s40623-025-02311-9","type":"published","date":"2025-11-25T15:58:32+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":82101157,"identity":"cd118f16-dfc0-4c2c-a0b4-dd9d62a1f1a9","added_by":"auto","created_at":"2025-05-06 18:52:17","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1033489,"visible":true,"origin":"","legend":"\u003cp\u003eMap showing the study area with seismicity. (a) Seismic stations (navy square) used in this study. The drawn area is indicated by the gray box in the inset plot. Black and gray dashed lines denote the top surface of PHS from Hirose et al. (2008b, a) and Nakajima et al. (2009), respectively. Thick dashed lines denote the Sagami or Suruga Troughs. Red triangles denote locations of active volcanoes. Red lines show active fault traces (Nakata and Imaizumi 2002). The red-hatched area is the shear zone implicated from GNSS data (Doke et al. 2021). The orange-hatched area shows the rupture zone for the 1923 Kanto earthquake (Wald and Somerville 1995), with its epicenter shown by a black star (Kanamori and Miyamura 1970). (b) Enlarged view for the red box in (a). Circular plots represent the seismicity from January 2004 to October 2024. Colored plots denote the seismic events analyzed in this study with depths coded by colors. Those without and with black edges show the events before and after 9 August 2024 at 6 pm (JST), respectively. Gray circular plots represent events that were not analyzed in this study. Navy and gray beachballs in lower hemisphere are the NIED F-net MT solutions for the mainshock (Mw 5.0) and the largest aftershock (Mw 4.1), respectively. HM-FZ and MKF represent the Hirayama–Matsuda-kita fault zone and Matsuda-kita fault, respectively. Gray rectangles denote the fault model of HM-FZ retrieved from the NIED J-SHIS’s page (https://www.j-shis.bosai.go.jp/en/), with top edges indicated by thick lines.\u003c/p\u003e","description":"","filename":"fig01map.png","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/561b831a64130627281bb947.png"},{"id":82101954,"identity":"33e745be-377d-4a68-9dd7-f56e6c7b486b","added_by":"auto","created_at":"2025-05-06 19:08:17","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":385807,"visible":true,"origin":"","legend":"\u003cp\u003eRelocated hypocenters and extracted fault planes. (a) Relocated hypocenters after 9 August 2024 at 6 pm (JST) with color coded by depth. The dashed line denotes the top surface of PHS at 10 km depth from Hirose et al. (2008b, a). (b) Map view of the extracted rectangular fault planes. Gray texts starting with # represent cluster IDs. A rectangle represents each fault plane with the top edge denoted by a thick line. Gray dots are the relocated hypocenters shown in (a). Black dots are the foreshocks approximately one hour before the mainshock. Red lines denote survey lines P–S. Navy and gray beachballs in lower hemisphere represent double-couple focal mechanisms of the mainshock (Mw 5.0) and the largest aftershock (Mw 4.1) extracted from NIED F-net MT solutions in Fig. 1b. (c–f) Cross-sectional views of the fault planes along the lines P–S, respectively. Relocated hypocenters within 0.4 km from the survey line are projected.\u003c/p\u003e","description":"","filename":"fig02planes.png","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/b6411bd6538fe6564d3807c8.png"},{"id":82101148,"identity":"533b148c-2bb8-4e30-964f-8b6fd0414fd7","added_by":"auto","created_at":"2025-05-06 18:52:17","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":223992,"visible":true,"origin":"","legend":"\u003cp\u003eTemporal variation of aftershocks around Planes #4 and #1. (a) Map view of epicenters with colors coded by days before/after 15 August 2024 at 8 pm JST. The double-couple focal mechanism is for the largest aftershock (Mw 4.1) on 15 August at 8:20:34 pm JST from the F-net MT solution. Dotted rectangles denote the regions of events of interest around Planes (b) #4 and (c) #1, respectively. (b, c) Histograms showing the occurrence time of events. Left and right vertical dashed lines correspond to the origin time of the mainshock (Mw 5.0) and the largest aftershock (Mw 4.1), respectively. Blue and red bins denote the occurrence time before and after the largest aftershock.\u003c/p\u003e","description":"","filename":"fig03aftaug15.png","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/820474507267edd666a02cd8.png"},{"id":82101952,"identity":"6cfb0c73-5131-4217-9995-14e11d664a0f","added_by":"auto","created_at":"2025-05-06 19:08:17","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":672048,"visible":true,"origin":"","legend":"\u003cp\u003eSchematic illustration of the mainshock fault plane in navy along with the PHS surface in black and the active fault in red dashed (MKF). In the upper panel, the blue contours indicate stress rates at intervals of 4 kPa/yr (Saito and Noda 2023). The orange-hatched area shows the rupture zone for the 1923 Kanto earthquake (Wald and Somerville 1995). The star and double-couple focal mechanism indicate the Mw 5.0 mainshock. The thick black line denotes the cross section for the lower panel. In the lower panel, brown corresponds to the forearc crust and light blue to PHS. Depth and lateral distance are not to scale.\u003c/p\u003e","description":"","filename":"fig04schematic2v2.png","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/0de9278687b307d458a391c4.png"},{"id":97178622,"identity":"08f8fddd-abe5-4ac2-9bc0-13d73081f7d5","added_by":"auto","created_at":"2025-12-01 16:11:41","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":2247160,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/2dac6db5-ae0e-4cce-b75c-90009273be4f.pdf"},{"id":82101160,"identity":"8c1bd417-b038-4ae8-86e1-f9bfc894f71e","added_by":"auto","created_at":"2025-05-06 18:52:17","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":2444169,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"SawakiEPS2025SagamiSM.docx","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/76fe41e19ac78216478efa95.docx"},{"id":82101734,"identity":"2fac18b5-4631-4851-8e6b-c4bf509a5817","added_by":"auto","created_at":"2025-05-06 19:00:17","extension":"csv","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":5814,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"AdditionalFile2TableS1stations100kmtable.csv","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/63e62ab5e07b22518df09e03.csv"},{"id":82101735,"identity":"0201dcd3-5696-428a-9b0e-f1d8fbb45c42","added_by":"auto","created_at":"2025-05-06 19:00:17","extension":"csv","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":54125,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"AdditionalFile3TableS2hypoddreloc.csv","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/f61613240c3efe8c9f18d646.csv"},{"id":82101951,"identity":"ca55b914-e579-4867-86c7-6df0e7258907","added_by":"auto","created_at":"2025-05-06 19:08:17","extension":"csv","order_by":4,"title":"","display":"","copyAsset":false,"role":"supplement","size":26271,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"AdditionalFile4TableS415000505050000events.csv","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/7ea01b9f4b883165b90acc21.csv"},{"id":82101739,"identity":"3c5def90-7c7b-4f00-8239-5f6a4337aabc","added_by":"auto","created_at":"2025-05-06 19:00:17","extension":"csv","order_by":5,"title":"","display":"","copyAsset":false,"role":"supplement","size":2121,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\u003c/p\u003e","description":"","filename":"AdditionalFile5TableS515000505050000planes.csv","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/2c5f0bf422447e1b2452430d.csv"},{"id":82101159,"identity":"3a42affb-7517-48ee-bd7b-dfce3fe76e19","added_by":"auto","created_at":"2025-05-06 18:52:17","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"supplement","size":99483,"visible":true,"origin":"","legend":"","description":"","filename":"graphicalabstract.png","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/db70c98b482a4913c587fc48.png"},{"id":82101152,"identity":"dcebfd17-0dbb-41fd-a3f2-b7e5eedf52e1","added_by":"auto","created_at":"2025-05-06 18:52:17","extension":"docx","order_by":7,"title":"","display":"","copyAsset":false,"role":"supplement","size":26393,"visible":true,"origin":"","legend":"","description":"","filename":"SupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-6353744/v1/2ed84d38667916f00f4bc5ad.docx"}],"financialInterests":"","formattedTitle":"Intraslab reverse faulting adjacent to the hypocenter of the 1923 Kanto earthquake: The Mw 5.0 western Kanagawa earthquake in eastern Japan on 9 August 2024","fulltext":[{"header":"Introduction","content":"\u003cp\u003eAn earthquake with the local magnitude (Mj) 5.3 occurred in western Kanagawa Prefecture, eastern Japan, on 9 August 2024 at 7:57:38 pm JST (UTC\u0026thinsp;+\u0026thinsp;9), leading to a maximum Japan Meteorological Agency (JMA) seismic intensity scale of 5 Lower in this region. The F-net moment tensor (MT) solution indicated a reverse-fault-type focal mechanism of Mw 5.0 at a depth of 11 km, characterized by a north\u0026ndash;south compressional axis (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). Following the mainshock, numerous aftershocks occurred at depths\u0026thinsp;\u0026gt;\u0026thinsp;10 km based on the JMA Unified Catalog and extended 5 km laterally in this region, with the largest aftershock (Mw 4.1) on 15 August at 8:20:34 pm JST in a similar mechanism with the mainshock. In this area, the Philippine Sea plate (PHS) subducts northward beneath the overriding North American plate along the Sagami Trough. The PHS surface depth is approximately 10 km (Hirose et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2008b\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003ea\u003c/span\u003e) around the aftershock area. Given that the compressional axis is roughly consistent with the PHS subduction, and that one nodal plane of these focal mechanisms aligns well with the geometry of PHS, interplate thrusting was considered though the JMA\u0026rsquo;s aftershock distribution did not show clear low-angle alignments.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe Tokyo Metropolitan area, including Kanagawa Prefecture, has repeatedly experienced huge earthquakes, known as \u0026ldquo;Great Kanto earthquakes\u0026rdquo;. The latest earthquake was often called the Taisho Kanto earthquake (Mw 7.9\u0026ndash;8.2) on 1 September 1923 (herein, the 1923 Kanto earthquake). The 1923 Kanto earthquake was a megathrust-type earthquake. The rupture area of the 1923 Kanto earthquake widely covered the Tokyo Metropolitan area (Kanamori \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e1971\u003c/span\u003e; Wald and Somerville \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e1995\u003c/span\u003e; Kobayashi and Koketsu \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Matsu\u0026rsquo;ura et al. \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2007\u003c/span\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea), which resulted in the most devastating earthquake disaster in the Japan\u0026rsquo;s history. The rupture of the 1923 Kanto earthquake was considered to initiate in the western Kanagawa region (Kanamori and Miyamura \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e1970\u003c/span\u003e), near the epicenter of this Mw 5.0 earthquake. Thus, investigating whether this Mw 5.0 earthquake occurred on the plate interface would help advance our understanding of mechanisms for future great earthquakes.\u003c/p\u003e \u003cp\u003eThe tectonics of this area is characterized not only by the subduction of PHS but the northward collision of the Izu Peninsula, called the Izu collision zone (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). The upper surface of PHS and Izu Peninsula has been estimated through active source surveys (Sato et al. \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2005\u003c/span\u003e; Arai et al. \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2009\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2014\u003c/span\u003e), seismic tomography (Hirose et al. \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2008b\u003c/span\u003e, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003ea\u003c/span\u003e; Nakajima et al. \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2009\u003c/span\u003e), azimuthal anisotropy (Ishise et al. \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2021\u003c/span\u003e), and receiver functions (Abe et al. \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). The Izu Peninsula is considered as a part of PHS or a separated microplate (e.g. Mazzotti et al. \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e1999\u003c/span\u003e; Nishimura et al. \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). The tectonic boundary between the western collision and eastern subduction (Yoshida \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e1993\u003c/span\u003e) has been investigated approximately 139.1\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:^\\circ\\:\\)\u003c/span\u003e\u003c/span\u003eE from focal mechanisms (Yukutake et al. \u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2012\u003c/span\u003e), crustal structures (e.g. Abe et al. \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Honda et al. \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), and a geodetic study (Doke et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Yukutake et al. (\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e2012\u003c/span\u003e) investigated the seismicity and focal mechanisms beneath the Tanzawa Mountains and concluded that earthquakes typically occur within and the upper surface of PHS in the western and eastern parts, respectively. This area has experienced M7-class crustal earthquakes that are often referred to as Odawara earthquakes (e.g. Ishibashi \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2004\u003c/span\u003e). Doke et al. (\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e) indicated that northward detachment of the northern Izu Peninsula (Seno \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2005\u003c/span\u003e) produces a shear zone, north of which situates the hypocenter of the Mw 5.0 earthquake (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). These studies suggested that right-lateral slip on Western Sagami-Bay Fracture may have caused the Odawara earthquakes (Ishibashi \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2004\u003c/span\u003e; Doke et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). On the contrary, Ishida and Kikuchi (\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e1992\u003c/span\u003e) proposed from intraslab earthquakes at Odawara at the south of the Mw 5.0 event that Odawara earthquakes occurred in the PHS slab. Thus, possible sources of Odawara earthquakes are still under debate. These suggest that the area of interest is in complex stress states.\u003c/p\u003e \u003cp\u003eAdditionally, there are several known active fault zones in this area (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb). The left-lateral Kita-Izu fault zone and right-lateral Western Sagami-Bay Fracture give the counterclockwise shear zone (Doke et al. \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). The Kōzu\u0026ndash;Matsuda fault zone in the northeast of the shear zone is thought to be a branch fault for PHS interface (Headquarters for Earthquake Research Promotion (HERP) 2015). The north-dipping Matsuda-kita fault (MKF) in Hirayama\u0026ndash;Matsuda-kita fault zone (HM-FZ) with a fault length of 4\u0026ndash;8 km is situated around the epicenters of the Mw 5.0 earthquake sequence (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb). MKF is considered a north-dipping reverse fault in 40\u0026ndash;50 degrees, whose loading rate is estimated 0.8 m/kyr or faster (HERP 2015). HERP (2015) declared that HM-FZ is classified as a group with a relatively high probability (0.09\u0026ndash;0.6%) of experiencing an earthquake within the next 30 years among the major active faults in our country. Although MKF is considered in the forearc crust above the PHS slab, it is difficult to deny the possibility of faulting on the active fault or in the PHS slab from the initial aftershock distribution.\u003c/p\u003e \u003cp\u003eThese circumstances raise the scientific question: where and how the faulting occurred during this Mw 5.0 earthquake\u0026mdash;interplate thrusting, active fault, or intraslab faulting. This is crucial for comprehending the mechanism of earthquake generation in the region of interest. To tackle this question, we relocated micro-earthquakes in the target area and clarified the fault geometry of the Mw 5.0 earthquake through hypocenter clustering analysis with machine learning techniques.\u003c/p\u003e"},{"header":"Data and Methods","content":"\u003cp\u003eWe conducted machine-learning-based phase picking, hypocenter relocation, and hypocenter clustering analysis to extract rectangular fault planes (FaultNVC; Sawaki et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2025b\u003c/span\u003e) for the Mw 5.0 earthquake in western Kanagawa on 9 August 2024. Below outlines the data processing and analysis performed in this study.\u003c/p\u003e \u003cp\u003eFirst, we picked P- and S-wave arrivals for each event in the JMA Unified Earthquake Catalog using a deep-neural-network-based picker of seismic arrival times (PhaseNet; Zhu and Beroza \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2019\u003c/span\u003e), which was re-trained on the JMA catalog (PhaseNetWC; Naoi et al. \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). This aimed to increase the number of picks available for hypocenter relocation. Using these picked phase arrivals along with JMA pick data and computing waveform cross-correlations (CCs), we relocated hypocenters using hypoDD (Waldhauser and Ellsworth \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2000\u003c/span\u003e). Subsequently, we extracted rectangular fault planes in the relocated aftershock using FaultNVC (Sawaki et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2025b\u003c/span\u003e).\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003ePhase picking using PhaseNetWC\u003c/h2\u003e \u003cp\u003eWe collected 867 earthquake events shallower than 25 km from the JMA Unified Earthquake Catalog, which occurred within the region specified in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb from January 2004 to October 2024. The minimum JMA magnitude was set to 0.5. We also retrieved JMA picks of the collected events. However, manual picks are limited; many include automatic picks. Consequently, there is a limit to the number of quality-controlled picks.\u003c/p\u003e \u003cp\u003eTherefore, we applied PhaseNet to obtain more arrival times for P- and S-wave first arrivals. PhaseNet was trained with seismic events in California (Zhu and Beroza \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). However, recent work by Naoi et al. (\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2024\u003c/span\u003e) re-trained PhaseNet using seismograms recorded around the Japanese archipelago. Thus, their model (PhaseNetWC) was suitable for our purpose of picking phases. We retrieved velocity waveforms recorded by Hi-net from the National Research Institute for Earth Science and Disaster Resilience (NIED) (NIED 2019), JMA, Earthquake Research Institute, the University of Tokyo (ERI), and Hot Springs Research Institute of Kanagawa Prefecture (HSRI) (Fig. \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e and Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e). Seismograms with a sampling rate different from 100 Hz were decimated to 100 Hz. The maximum epicentral distance was set to 100 km. Theoretical travel times (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{TP},\\:{t}^{TS}\\)\u003c/span\u003e\u003c/span\u003e) were calculated based on the JMA2001 travel-time table (Ueno et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2002\u003c/span\u003e) and used as a reference for a time window of seismograms applied to PhaseNetWC.\u003c/p\u003e \u003cp\u003eWe tried to find candidate phase arrival times by using PhaseNetWC as we aimed to increase the number of picks. JMA picks (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{DP},\\:{t}^{DS}\\)\u003c/span\u003e\u003c/span\u003e) are available as a reference for time windows in some cases. Thus, three cases were considered: (A) seismograms without \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{DP}\\)\u003c/span\u003e\u003c/span\u003e nor \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{DS}\\)\u003c/span\u003e\u003c/span\u003e; (B) without \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{DP}\\)\u003c/span\u003e\u003c/span\u003e but with \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{DS}\\)\u003c/span\u003e\u003c/span\u003e; and (C) with \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{DP}\\)\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eFor case (A), we set a time window of [-15, 15] seconds centered on \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{TP}\\)\u003c/span\u003e\u003c/span\u003e and ran PhaseNetWC to three-component waveforms for a tentative P-wave arrival time (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\tau\\:}^{\\text{t}\\text{n}\\text{t}}\\)\u003c/span\u003e\u003c/span\u003e). No filter was applied to run PhaseNetWC. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\tau\\:}^{\\text{t}\\text{n}\\text{t}}\\)\u003c/span\u003e\u003c/span\u003e was searched within \u0026plusmn;\u0026thinsp;5 seconds of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{TP}\\)\u003c/span\u003e\u003c/span\u003e, maximizing the probability of the P-wave arrival (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{p}_{P}\\)\u003c/span\u003e\u003c/span\u003e). If the maximum \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{p}_{P}\\)\u003c/span\u003e\u003c/span\u003e was 0.5 or greater, we used \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\tau\\:}^{\\text{t}\\text{n}\\text{t}}\\)\u003c/span\u003e\u003c/span\u003e instead of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{TP}\\)\u003c/span\u003e\u003c/span\u003e for another run of PhaseNetWC for an S-wave arrival. If the maximum \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{p}_{P}\\)\u003c/span\u003e\u003c/span\u003e was less than 0.5, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{TP}\\)\u003c/span\u003e\u003c/span\u003e was used as \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\tau\\:}^{\\text{t}\\text{n}\\text{t}}\\)\u003c/span\u003e\u003c/span\u003e. Then, PhaseNetWC was applied again by setting a time window of [-5, 25] seconds from \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\tau\\:}^{\\text{t}\\text{n}\\text{t}}\\)\u003c/span\u003e\u003c/span\u003e. From an example case, PhaseNetWC was able to pick proper phase arrivals (Fig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFor cases (B) and (C), we tried to inspect JMA picks more precisely. For case (B), \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{DS}\\)\u003c/span\u003e\u003c/span\u003e was given priority over \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{TS}\\)\u003c/span\u003e\u003c/span\u003e, and we ran PhaseNetWC by setting a time window of [-20, 10] seconds from \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{DS}\\)\u003c/span\u003e\u003c/span\u003e. For case (C), \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{DP}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{DS}\\)\u003c/span\u003e\u003c/span\u003e were given priority over \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{TP}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{TS}\\)\u003c/span\u003e\u003c/span\u003e, and we ran PhaseNetWC by setting a time window of [-5, 25] seconds from \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{DP}\\)\u003c/span\u003e\u003c/span\u003e. Even though JMA picks are available, PhaseNetWC picks can correct proper arrivals (Fig. \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eFinally, we created a pick table as we executed PhaseNetWC for all cases. The pick times (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\tau\\:}^{P},\\:{\\tau\\:}^{S}\\)\u003c/span\u003e\u003c/span\u003e), on which the probability was maximized, were searched. If the maximum probability (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{p}_{\\text{m}\\text{a}\\text{x}}\\)\u003c/span\u003e\u003c/span\u003e) was 0.5 or greater, we took \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\tau\\:\\)\u003c/span\u003e\u003c/span\u003e for a phase pick. If \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{p}_{\\text{m}\\text{a}\\text{x}}\\)\u003c/span\u003e\u003c/span\u003e was less than 0.5, we checked if the JMA pick \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{t}^{D}\\)\u003c/span\u003e\u003c/span\u003e was available. If available, we took it for a phase pick instead of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\tau\\:\\)\u003c/span\u003e\u003c/span\u003e; otherwise, we discarded the phase. In total, we obtained 87,154 picks from 157 stations: 47,284 picks for P and 39,870 picks for S. This is more than three times the picks obtained by JMA: 13,955 picks for P; 14,230 picks for S; and 28,185 picks in total from 77 stations. The increase in phase picks is more evident for stations distant from the target area and stations omitted by JMA (Fig. \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eHypocenter relocation\u003c/h3\u003e\n\u003cp\u003eWe relocated the hypocenters using hypoDD (Waldhauser and Ellsworth \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2000\u003c/span\u003e) with the JMA2001 model. To create a differential travel-time table, we set the maximum inter-event distance of 3.0 km. The pick table yielded 352,608 and 329,632 differential times for P- and S-waves, respectively, when allowing 30 neighboring events.\u003c/p\u003e \u003cp\u003eAdditionally, we computed CC of seismograms from an event pair within 2 km both laterally and in depth, and extracted lapse times with maximized CC values as differential times. We applied the bandpass filter of 4\u0026ndash;12 Hz and extracted a time window before and after 1.0 s of the pick time (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\tau\\:}^{P}\\)\u003c/span\u003e\u003c/span\u003e or \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\tau\\:}^{S}\\)\u003c/span\u003e\u003c/span\u003e). For P-waves, the CC was computed between vertical components. For S waves, we considered calculating the particle motion of two horizontal components to obtain an average-like correlation of the two components, instead of adopting the horizontal component with better CC values (e.g. Yoshida et al. \u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) (Fig. \u003cspan refid=\"MOESM4\" class=\"InternalRef\"\u003eS4\u003c/span\u003e). Although computed CC values tend to be smaller than those in the conventional way, this way is robust to the horizontal rotation of seismographs. To be exact, complex waveforms were generated from the two horizontal components (north and east), and the CC function was calculated between these complex waveforms:\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:\\begin{array}{c}{C}_{ijk}^{S}\\left(t\\right)={z}_{ik}\\left(t\\right)\\star\\:{z}_{jk}^{*}\\left(t\\right),\\#\\left(1\\right)\\end{array}\\)\u003c/span\u003e \u003c/span\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\star\\:\\)\u003c/span\u003e\u003c/span\u003e represents the calculation of CC, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{C}_{ijk}^{S}\\left(t\\right)\\)\u003c/span\u003e\u003c/span\u003e denotes the S-wave CC function between the \u003cem\u003ei\u003c/em\u003e-th and \u003cem\u003ej\u003c/em\u003e-th events at the \u003cem\u003ek\u003c/em\u003e-th station. In Eq.\u0026nbsp;1, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}^{\\text{*}}\\)\u003c/span\u003e\u003c/span\u003e denotes the conjugate of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:z\\)\u003c/span\u003e\u003c/span\u003e. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{z}_{ik}\\left(t\\right)\\)\u003c/span\u003e\u003c/span\u003e represents a horizontal complex waveform for the \u003cem\u003ei\u003c/em\u003e-th event at the \u003cem\u003ek\u003c/em\u003e-th station:\u003c/p\u003e \u003cp\u003e \u003cspan class=\"InlineEquation\"\u003e \u003cspan class=\"mathinline\"\u003e\\(\\:\\begin{array}{c}{z}_{ik}\\left(t\\right)={u}_{ik}^{\\text{E}}\\left(t\\right)+j{u}_{ik}^{\\text{N}}\\left(t\\right),\\#\\left(2\\right)\\end{array}\\)\u003c/span\u003e \u003c/span\u003ewhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{u}^{\\text{E}}\\left(t\\right)\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{u}^{\\text{N}}\\left(t\\right)\\)\u003c/span\u003e\u003c/span\u003e denote the east- and north-component waveforms, respectively. In Eq.\u0026nbsp;2, j denotes the imaginary unit. Note that we did not correct horizontal sensor azimuths because CCs were computed for waveforms at the same stations. We defined the S-wave CC value as the real part of Eq.\u0026nbsp;1.\u003c/p\u003e \u003cp\u003eFinally, pairs with a CC of 0.8 or greater and an absolute differential time of 1.0 s or shorter were selected for each phase. We obtained 488,876 and 193,694 pairs for P- and S-waves, respectively.\u003c/p\u003e \u003cp\u003eUsing differential times from picks and CCs, we executed hypoDD under the JMA2001 one-dimensional velocity model (Ueno et al. \u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2002\u003c/span\u003e). We obtained the final result after 25 relocation iterations. In total, 864 out of 867 events were relocated from January 2004 to October 2024 (Table \u003cspan refid=\"MOESM2\" class=\"InternalRef\"\u003eS2\u003c/span\u003e).\u003c/p\u003e\n\u003ch3\u003eHypocenter clustering for fault planes (FaultNVC)\u003c/h3\u003e\n\u003cp\u003eAs we focus on the earthquake sequence of the Mw 5.0 event, we used 364 events after 9 August 2024 at 6 pm JST for FaultNVC. We excluded the mainshock and events shallower than 10 km or deeper than 16 km. For a clustering purpose, these hypocenters in the geographic coordinate of World Geodetic System 1984 (the European Petroleum Survey Group (EPSG) code of 4326) were transformed into a local cartesian coordinate system. We used the EPSG code of 6677 as the local cartesian coordinate. We defined the center of the local cartesian coordinate as the longitude of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:139.160^\\circ\\:\\text{E}\\)\u003c/span\u003e\u003c/span\u003e and the latitude of \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:35.410^\\circ\\:\\text{N}\\)\u003c/span\u003e\u003c/span\u003e by applying a parallel shift to EPSG 6677 for simplicity.\u003c/p\u003e \u003cp\u003eFor the transformed hypocenters in the Cartesian coordinate system, we applied FaultNVC, which extracts rectangular fault planes through hypocenter clustering (Sawaki et al. \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2025b\u003c/span\u003e). This method considers hypocenters as point-cloud data and performs simultaneous clustering of point-cloud normal vectors and hypocenter positions. Following their study, we conducted three steps: (STEP 1) calculating PCNVs; (STEP 2) performing feature-vector clustering using HDBSCAN (Campello et al. \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2013\u003c/span\u003e) to extract rectangular fault planes; and (STEP 3) controlling the quality of these planes. We summarized the procedure in Text S1 and listed the parameter set in Table \u003cspan refid=\"MOESM3\" class=\"InternalRef\"\u003eS3\u003c/span\u003e.\u003c/p\u003e"},{"header":"Results and discussion","content":"\u003cp\u003eThe relocated hypocenters are shown in Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ea. Hypocenter focal depths ranged from 12 to 14.5 km. Around the mainshock hypocenter, aftershocks occurred approximately 14 km depth. In contrast, several hundred meters to the north, the focal depth suddenly shallowed to 12.5 km. Furthermore, southwest of the mainshock, the focal depth approached 12 km. Compared to the original JMA catalog, hypocenters were aligned to show planar features though slightly shifted to east (Fig. \u003cspan class=\"InternalRef\"\u003eS5\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eWe extracted five rectangular fault planes (Planes #1, #2, #4, #5, and #8) through FaultNVC after quality control (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eb), with fault parameters shown in Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e. The upper edge depth of these planes is nearly 12 km. Plane #4 in the northeastern area is a south-dipping plane, which consists of numerous aftershocks. The dip angle was estimated 68\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:^\\circ\\:\\)\u003c/span\u003e\u003c/span\u003e. Plane #1 is situated several hundred meters south of Plane #4 and is similarly dipping to the south. The cluster of Plane #1 consists of 17 events. The fault parameters are also nearly identical to those of Plane #4. In contrast, Plane #5 to the west of Plane #4 is dipping towards the WNW direction. Plane #8, located to the west of Plane #1, is dipping towards the east direction. The extracted planes are consistent with hypocenter alignments characterized by point-cloud normal vectors (Fig. \u003cspan class=\"InternalRef\"\u003eS6\u003c/span\u003e), though some clusters were rejected by quality control (Fig. S7).\u003c/p\u003e\n\u003cp\u003e\u003c/p\u003e\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eFault parameters extracted using FaultNVC (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003eb). The reference point is set to the upper left corner.\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eID\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eCluster size\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eLatitude [\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:^\\circ\\:\\)\u003c/span\u003e\u003c/span\u003eN]\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eLongitude [\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:^\\circ\\:\\)\u003c/span\u003e\u003c/span\u003eE]\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eDepth [km]\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStrike [\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:^\\circ\\:\\)\u003c/span\u003e\u003c/span\u003e]\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eDip [\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:^\\circ\\:\\)\u003c/span\u003e\u003c/span\u003e]\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eLength [km]\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eWidth [km]\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e35.4068\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e139.1598\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e12.53\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e81.7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e67.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.83\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e35.4082\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e139.1469\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e12.40\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e312.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e24.9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.62\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.62\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e130\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e35.4156\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e139.1575\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e12.51\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e84.3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e67.5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.57\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.12\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e35.4134\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e139.1606\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e12.66\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e201.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e70.0\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.54\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.53\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e35.3996\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e139.1532\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e12.38\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e350.8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e29.1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.54\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n\u003c/table\u003e\n\u003cp\u003e\u003c/p\u003e\n\u003cp\u003e\u003cbr\u003e\u003c/p\u003e\n\u003cp\u003eWe examined cross-sectional views for these extracted fault planes (Figs. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ec\u0026ndash;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ef). Hypocenter alignments and steeply dipping structures both for Planes #4 and #1 were observed in the cross-sectional view along the P\u0026ndash;p and Q-q lines (Figs. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ec and \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ed). Plane #4 is consistent with the steeper nodal plane of the mainshock focal mechanism, which is different from the one parallel to the PHS surface. Considering the steep mainshock fault geometry and the PHS surface depth of 10 km, our results demonstrate that the mainshock occurred as an intraslab earthquake on a steep reverse fault, as opposed to thrusting on the PHS surface or reverse faulting on MKF in the forearc crust. The slab Moho depth was investigated to be 20\u0026ndash;30 km from P-wave tomography or receiver functions (Ishise et al. \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e; Abe et al. \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). Thus, this earthquake probably occurred in the PHS crust.\u003c/p\u003e\n\u003cp\u003eWe also focus on the largest (Mw 4.1) aftershock. For cross-sectional views along the R\u0026ndash;r line, Plane #8 dipping east roughly is consistent with hypocenter alignments (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ee). However, Plane #8 is not consistent with the nodal planes of the Mw 4.1 aftershock. The cross-section S\u0026ndash;s revealed a clear depth difference between Planes #8 and #5: Plane #5 deeper than Plane #8. This is attributed to the seismic gap between these planes at the horizontal position of 1.4\u0026ndash;1.6 km.\u003c/p\u003e\n\u003cp\u003eIn contrast to Plane #8, Plane #1 is likely associated with the steeper nodal plane of the Mw 4.1 aftershock, although this aftershock event does not belong to Plane #1. To demonstrate that, we examined the event occurrence time and showed a depth view around Plane #1 (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e). Event counts around Plane #3 are apparently smaller after the occurrence of the largest aftershock than before that (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003eb). In contrast, those around Plane #1 are relatively higher after the occurrence of the Mw 4.1 aftershock (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003ec). This suggests that aftershocks of the Mw 4.1 aftershock occurred near Plane #1. Additionally, the depth view of hypocenters along the \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{X}_{1}\\)\u003c/span\u003e\u003c/span\u003e axis exhibited that one nodal plane is highly consistent with Plane #1 with a slight gap (Fig. \u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003ed). The fault scale of approximately 1 km for the focused area is comparable to the characteristic fault length of an Mw 4-class earthquake. Therefore, the fault slip of the Mw 4.1 aftershock probably occurred on Plane #1 or on its extension.\u003c/p\u003e\n\u003cp\u003eThe mainshock fault slip may not be so simple as the pure reverse faulting on Plane #4. The F-net MT solution has 50% non-double-couple components. Possible coseismic slip on extended planes such as Planes #5 and #8 other than #4 might have contributed to the high non-double-couple components.\u003c/p\u003e\n\u003cp\u003eIn the JMA catalog, six foreshocks with Mj of 0.7\u0026ndash;2.1 were included, starting approximately one hour before the Mw 5.0 mainshock. The clustering analysis for the relocated catalog showed that all the foreshocks were included in Plane #4, shallower part of the mainshock hypocenter (Fig. \u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003ec). This may indicate that minor precursory slips had occurred on the principal plane (#4) before the mainshock slip and minor stress change or fluid migration occurred towards the mainshock hypocenter. Although we did not analyze the temporal variation of seismicity, the hypocenter clustering method can clarify the spatial relationships between foreshocks, mainshocks, and aftershocks. This will help in understanding the evolution of fault slip during an earthquake sequence.\u003c/p\u003e\n\u003cp\u003eWe discuss the possible causes of steep reverse intraslab earthquakes and their implications for future great earthquakes. The occurrence of steep and reverse intraslab earthquakes in the upper part of the slab is crucial key to understanding the stress state around the seismogenic zone. Generally, the faulting type of intraslab earthquakes reflects the state of stress and pore fluid pressure in the oceanic plate (e.g. Kirby \u003cspan class=\"CitationRef\"\u003e1995\u003c/span\u003e; Seno and Yoshida \u003cspan class=\"CitationRef\"\u003e2004\u003c/span\u003e; Wada et al. \u003cspan class=\"CitationRef\"\u003e2010\u003c/span\u003e). Outer rise normal faults are produced in extensional fields due to plate bending (Christensen and Ruff \u003cspan class=\"CitationRef\"\u003e1988\u003c/span\u003e) and can be re-activated as reverse faults in compressional fields due to slab compression along the upper layer of double seismic zones (e.g. Hasegawa et al. \u003cspan class=\"CitationRef\"\u003e1978\u003c/span\u003e; Yamasaki and Seno \u003cspan class=\"CitationRef\"\u003e2003\u003c/span\u003e; Ohta et al. \u003cspan class=\"CitationRef\"\u003e2011\u003c/span\u003e). In the Japan Trench, great reverse intraslab earthquakes have occurred around the downdip portion of coseismic slip area of large interplate earthquakes (Okada and Hasegawa \u003cspan class=\"CitationRef\"\u003e2003\u003c/span\u003e; Ohta et al. \u003cspan class=\"CitationRef\"\u003e2011\u003c/span\u003e). Ohta et al. (\u003cspan class=\"CitationRef\"\u003e2011\u003c/span\u003e) investigated the coseismic fault model of the reverse intraslab event of M 7.1 on 7 April 2011 after the Mw 9.0 Tohoku-oki earthquake on 11 March 2011 and proposed that a pre-existing outer-rise normal fault was re-activated in a reversed sense with downdip compressional stress. The reverse faulting of this Mw 5.0 event may indicate a compressional stress field in the upper part of PHS.\u003c/p\u003e\n\u003cp\u003eThis area is in complex tectonics due to the subduction of PHS and the collision of the Izu Peninsula. The regional stress field, estimated by stress tensor inversion using micro-earthquake focal mechanisms, exhibited an NW\u0026ndash;SE-trending maximum horizontal stress axis (Imanishi et al. \u003cspan class=\"CitationRef\"\u003e2019\u003c/span\u003e; Uchide et al. \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e). This is roughly consistent to cause reverse faulting on an E\u0026ndash;W trending dip fault. Additionally, the plate coupling ratio is high or middle around the Odawara area, which was derived from GNSS data (Nishimura et al. \u003cspan class=\"CitationRef\"\u003e2018\u003c/span\u003e). This area is characterized by a high stress rate (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e), which can initiate the Taisho-type Kanto earthquake (Saito and Noda \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e). These studies may indicate that the PHS subduction has accumulated stress on the plate interface near the focal area of this Mw 5.0 earthquake since the last Taisho Kanto earthquake in 1923. Arai and Iwasaki (\u003cspan class=\"CitationRef\"\u003e2015\u003c/span\u003e) pointed out that intraslab earthquakes tend to occur beneath strongly coupled plate interface and a reflective zone is located at the downdip end of it. Given that the coupled plate interface is willing to host intraslab earthquakes rather than interplate earthquakes and release the slab-loading stress, these complex and accumulated stresses may have given rise to the occurrence of the reverse intraslab earthquake near the hypocenter of the 1923 Kanto earthquake (Fig. \u003cspan class=\"InternalRef\"\u003e4\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eMonitoring intraslab stress states would help understand faulting properties around seismogenic zones. A seismic cycle simulation indicated that the occurrence of an interplate earthquake was advanced due to stress perturbations of an intraslab earthquake (Kato \u003cspan class=\"CitationRef\"\u003e2004\u003c/span\u003e). In the Nankai Trough, the stress orientation at the updip portion of deep slow-earthquake area, estimated from upper-plane reverse intraslab events, rotated after slow earthquake events of episodic tremor and slip (Kita et al. \u003cspan class=\"CitationRef\"\u003e2021\u003c/span\u003e). They inferred that fluid migration from slab to interface promoted slow earthquakes on the plate interface. In a similar manner, further research needs to investigate the influence of reverse intraslab slips on the plate interface conditions. Furthermore, the possible source of Odawara earthquakes might originate from the PHS slab (Ishida and Kikuchi \u003cspan class=\"CitationRef\"\u003e1992\u003c/span\u003e), despite undergoing debates and the spatial offset from the 2024 Mw 5.0 event. While Yukutake et al. (\u003cspan class=\"CitationRef\"\u003e2012\u003c/span\u003e) suggested that the seismicity in the east of the tectonic boundary beneath the Tanzawa Mountains is induced by the PHS subduction process, it has not been clarified whether this seismicity in the north of our target area includes intraslab events. Although we did not perform local stress inversions, further studies that monitor the seismicity including these intraslab earthquakes and slab stress states are demanded to understand interplate faulting and crustal earthquakes, assessing the future occurrence of Kanto and Odawara earthquakes.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eWe identified fault planes of the Mw 5.0 earthquake in western Kanagawa region in eastern Japan on 9 August 2024 to understand its faulting mechanism. We conducted a machine-learning-based workflow of phase picking, hypocenter relocation, and fault plane identification through hypocenter clustering. The extracted fault planes included a south-dipping mainshock plane in a steep angle and a minor south-dipping plane parallel to the mainshock plane that probably consists of the largest (Mw 4.1) aftershock. This extracted plane, being 2 km deeper than the slab surface, demonstrates that the Mw 5.0 earthquake occurred as an intraslab earthquake on a steep reverse fault, as opposed to thrusting on the PHS surface or reverse faulting on an active fault in the forearc crust. Given the high stress rate on the plate interface in this area, complex stress states may have given rise to the occurrence of the steep reverse intraslab earthquake beneath highly coupled plate interface. Careful monitoring of intraslab earthquakes and slab stress states would help assess the future occurrence of Kanto earthquakes and great regional events.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cdiv class=\"DefinitionList\"\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eCC\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eCross correlation\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eEPSG\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eEuropean Petroleum Survey Group\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eERI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eEarthquake Research Institute, the University of Tokyo\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGNSS\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eGlobal Navigation Satellite System\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eGSJ\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eGeological Survey of Japan, National Institute of Advanced Industrial Science and Technology (AIST), Japan\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eHDBSCAN\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eHierarchical Density-Based Spatial Clustering of Applications with Noise\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eHERP\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eHeadquarters for Earthquake Research Promotion, an extraordinary organ of MEXT\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eHM-FZ\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eHirayama\u0026ndash;Matsuda-kita fault zone\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eHSRI\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eHot Springs Research Institute of Kanagawa Prefecture\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eJMA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eJapan Meteorological Agency\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMEXT\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eMinistry of Education, Culture, Sports, Science and Technology\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMKF\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eMatsuda-kita fault\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eMT\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003emoment tensor\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003eNIED\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eNational Research Institute for Earth Science and Disaster Resilience, Japan\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv class=\"DefinitionListEntry\"\u003e \u003cdiv class=\"Term\"\u003ePCA\u003c/div\u003e \u003cdiv class=\"Description\"\u003e \u003cp\u003eprincipal component analysis\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003ch3\u003e\u003cstrong\u003eEthics approval and consent to participate\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eConsent for publication\u0026nbsp;\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eNot applicable.\u003c/p\u003e\n\u003ch3\u003eAvailability of data and materials\u003c/h3\u003e\n\u003cp\u003eThe original hypocenter catalog is available on the website of the JMA Unified Earthquake Catalog (https://www.data.jma.go.jp/svd/eqev/data/bulletin/index_e.html), which was last accessed in December 2024. The NIED F-net MT solutions were retrieved from their page (https://www.fnet.bosai.go.jp/event/search.php?LANG=en) with ids of 20240809105600 and 20240815111900. We retrieved seismic waveforms from NIED Hi-net (NIED 2019), JMA, ERI, and HSRI through the NIED Hi-net\u0026rsquo;s website with a registration required (https://www.hinet.bosai.go.jp/?LANG=en). We used Python software FaultNVC available at the GSJ\u0026rsquo;s Open-File Report (Sawaki et al. 2025a). We used SeisBench (Woollam et al. 2022) to run PhaseNet (Zhu and Beroza 2019), available at GitHub (https://github.com/seisbench/seisbench). The weight for PhaseNetWC (Naoi et al. 2024) can be available at GitHub (https://github.com/mktnaoi/JMAuniPicker). We retrieved the computing code of hypoDD from the author\u0026rsquo;s page (https://www.ldeo.columbia.edu/~felixw/hypoDD.html), last accessed 17 January 2025. For visualizations, we used Matplotlib (Hunter 2007), colormaps provided by Colorcet (https://colorcet.holoviz.org/), and Cartopy v0.24 (Elson et al. 2024). The data generated or analyzed during this study, including the clustering result (Table S4 and S5), are contained in supplementary information files.\u003c/p\u003e\n\u003ch3\u003eCompeting interests\u003c/h3\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003ch3\u003eFunding\u003c/h3\u003e\n\u003cp\u003eThis study was supported by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) Project for Seismology toward Research Innovation with Data of Earthquake (STAR-E) Grant Number JPJ010217.\u003c/p\u003e\n\u003ch3\u003eAuthors\u0026apos; contributions\u003c/h3\u003e\n\u003cp\u003eYS developed methods, analyzed the data, discussed the results, wrote and revised the original manuscript, and created all figures and tables. TS and TU discussed the results. TU developed methods, obtained funding, and supervised YS as a principal investigator of our research topic in the STAR-E Project. All authors have read and approved the final manuscript.\u003c/p\u003e\n\u003ch3\u003eAcknowledgements\u003c/h3\u003e\n\u003cp\u003eWe had fruitful discussions with Ryo Kurihara, Kodai Sagae, and Haruo Horikawa.\u003c/p\u003e\n\u003ch3\u003eAuthors\u0026rsquo; details\u003c/h3\u003e\n\u003cp\u003eYS, TS, TU: Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology (AIST), AIST Tsukuba Central 7, 1-1-1 Higashi, Tsukuba, Ibaraki 305-8567, Japan\u003c/p\u003e\n\u003cp\u003ePresent address for YS: College of Science and Engineering, Ritsumeikan University, 1‑1‑1 Noji-Higashi, Kusatsu, Shiga 525‑8577, Japan\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAbe Y, Honda R, Ishise M et al (2023) Relationship Between Crustal Structure and Plate Convergence Around the Izu Collision Zone in Central Japan. J Geophys Res Solid Earth 128. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2022JB026314\u003c/span\u003e\u003cspan address=\"10.1029/2022JB026314\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. e2022JB026314\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArai R, Iwasaki T (2015) Transition from collision to subduction and its relation to slab seismicity and plate coupling. Earth Planets and Space 67:76. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s40623-015-0247-6\u003c/span\u003e\u003cspan address=\"10.1186/s40623-015-0247-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArai R, Iwasaki T, Sato H et al (2009) Collision and subduction structure of the Izu\u0026ndash;Bonin arc, central Japan, revealed by refraction/wide-angle reflection analysis. Tectonophysics 475:438\u0026ndash;453. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.tecto.2009.05.023\u003c/span\u003e\u003cspan address=\"10.1016/j.tecto.2009.05.023\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eArai R, Iwasaki T, Sato H et al (2014) Contrasting subduction structures within the Philippine Sea plate: Hydrous oceanic crust and anhydrous volcanic arc crust. Geochem Geophys Geosyst 15:1977\u0026ndash;1990. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1002/2014GC005321\u003c/span\u003e\u003cspan address=\"10.1002/2014GC005321\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCampello RJGB, Moulavi D, Sander J (2013) Density-Based Clustering Based on Hierarchical Density Estimates. Advances in Knowledge Discovery and Data Mining. Springer, Berlin, Heidelberg, pp 160\u0026ndash;172. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-642-37456-2_14\u003c/span\u003e\u003cspan address=\"10.1007/978-3-642-37456-2_14\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChristensen DH, Ruff LJ (1988) Seismic coupling and outer rise earthquakes. J Geophys Res Solid Earth 93:13421\u0026ndash;13444. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/JB093iB11p13421\u003c/span\u003e\u003cspan address=\"10.1029/JB093iB11p13421\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDoke R, Honda R, Harada M et al (2021) Deformation of the seismogenic zone in the northeastern part of the Izu Peninsula, Japan, inferred from GNSS observations. Characterization of Modern and Historical Seismic\u0026ndash;Tsunamic Events, and Their Global\u0026ndash;Societal Impacts. Geological Society of London, pp 111\u0026ndash;129. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1144/SP501-2019-104\u003c/span\u003e\u003cspan address=\"10.1144/SP501-2019-104\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eElson P, de Andrade ES, Lucas G et al (2024) SciTools/cartopy: REL: v0.24.1. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5281/zenodo.13905945\u003c/span\u003e\u003cspan address=\"10.5281/zenodo.13905945\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Zenodo\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHasegawa A, Umino N, Takagi A (1978) Double-planed structure of the deep seismic zone in the northeastern Japan arc. Tectonophysics 47:43\u0026ndash;58. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/0040-1951(78)90150-6\u003c/span\u003e\u003cspan address=\"10.1016/0040-1951(78)90150-6\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeadquarters for Earthquake Research Promotion (2015) Long-term evaluation of the Shiozawa fault zone, Hirayama-Matsuda North fault zone, and Kōzu-Matsuda fault zone (Kan\u0026rsquo;nawa-Kōzu-Matsuda fault zone) (Second Edition) [in Japanese]. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.jishin.go.jp/main/chousa/katsudansou_pdf/36_kannawa_kouzu_4.pdf\u003c/span\u003e\u003cspan address=\"https://www.jishin.go.jp/main/chousa/katsudansou_pdf/36_kannawa_kouzu_4.pdf\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHirose F, Nakajima J, Hasegawa A (2008a) Three-dimensional seismic velocity structure and configuration of the Philippine Sea slab in southwestern Japan estimated by double-difference tomography. J Geophys Res Solid Earth 113:B09315. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2007JB005274\u003c/span\u003e\u003cspan address=\"10.1029/2007JB005274\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHirose F, Nakajima J, Hasegawa A (2008b) Three-Dimensional Velocity Structure and Configuration of the Philippine Sea Slab beneath Kanto District, Central Japan, Estimated by Double-Difference Tomography [in Japanese with English abstract]. Zisin (Journal Seismological Soc Japan 2nd ser) 60:123\u0026ndash;138. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4294/zisin.60.123\u003c/span\u003e\u003cspan address=\"10.4294/zisin.60.123\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHonda R, Abe Y, Doke R (2023) A Review of Recent Research on Crustal Structures and Seismo-tectonics of the Izu Collision Zone [in Japanese with English abstract]. Zisin (Journal Seismological Soc Japan 2nd ser) 76:135\u0026ndash;148. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4294/zisin.2023-9S\u003c/span\u003e\u003cspan address=\"10.4294/zisin.2023-9S\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHunter JD (2007) Matplotlib: A 2D Graphics Environment. Comput Sci Eng 9:90\u0026ndash;95. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/MCSE.2007.55\u003c/span\u003e\u003cspan address=\"10.1109/MCSE.2007.55\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eImanishi K, Uchide T, Ohtani M et al (2019) Construction of the Crustal Stress Map in the Kanto Region, central Japan [in Japanese with English abstract]. Bull Geol Surv Japan 70:273\u0026ndash;298. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.9795/bullgsj.70.273\u003c/span\u003e\u003cspan address=\"10.9795/bullgsj.70.273\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIshibashi K (2004) Seismotectonic modeling of the repeating M 7-class disastrous Odawara earthquake in the Izu collision zone, central Japan. Earth Planets and Space 56:843\u0026ndash;858. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/BF03353091\u003c/span\u003e\u003cspan address=\"10.1186/BF03353091\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIshida M, Kikuchi M (1992) A possible foreshock of a future large earthquake near Odawara, center Japan. Geophys Res Lett 19:1695\u0026ndash;1698. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/92GL01503\u003c/span\u003e\u003cspan address=\"10.1029/92GL01503\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eIshise M, Kato A, Sakai S et al (2021) Improved 3-D P Wave Azimuthal Anisotropy Structure Beneath the Tokyo Metropolitan Area, Japan: New Interpretations of the Dual Subduction System Revealed by Seismic Anisotropy. J Geophys Res Solid Earth 126:e2020JB021194. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2020JB021194\u003c/span\u003e\u003cspan address=\"10.1029/2020JB021194\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKanamori H (1971) Faulting of the Great Kanto Earthquake of 1923 as Revealed by Seismological Data. Bulletin of the Earthquake Research Institute. Univ Tokyo 49:13\u0026ndash;18. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.15083/0000033257\u003c/span\u003e\u003cspan address=\"10.15083/0000033257\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKanamori H, Miyamura S (1970) Seismometrical Re-Evaluation of the Great Kanto Earthquake of September 1, 1923. Bulletin of the Earthquake Research Institute. Univ Tokyo 48:115\u0026ndash;125. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.15083/0000033322\u003c/span\u003e\u003cspan address=\"10.15083/0000033322\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKato N (2004) A possible effect of an intermediate depth intraslab earthquake on seismic cycles of interplate earthquakes at a subduction zone. Earth Planets and Space 56:553\u0026ndash;561. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/BF03352516\u003c/span\u003e\u003cspan address=\"10.1186/BF03352516\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKirby S (1995) Interslab earthquakes and phase changes in subducting lithosphere. Rev Geophys 33:287\u0026ndash;297. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/95RG00353\u003c/span\u003e\u003cspan address=\"10.1029/95RG00353\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKita S, Houston H, Yabe S et al (2021) Effects of episodic slow slip on seismicity and stress near a subduction-zone megathrust. Nat Commun 12:7253. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41467-021-27453-8\u003c/span\u003e\u003cspan address=\"10.1038/s41467-021-27453-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKobayashi R, Koketsu K (2005) Source process of the 1923 Kanto earthquake inferred from historical geodetic, teleseismic, and strong motion data. Earth Planets and Space 57:261\u0026ndash;270. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/BF03352562\u003c/span\u003e\u003cspan address=\"10.1186/BF03352562\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMatsu\u0026rsquo;ura M, Noda A, Fukahata Y (2007) Geodetic data inversion based on Bayesian formulation with direct and indirect prior information. Geophys J Int 171:1342\u0026ndash;1351. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1111/j.1365-246X.2007.03578.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1365-246X.2007.03578.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMazzotti S, Henry P, LePichon X, Sagiya T (1999) Strain partitioning in the zone of transition from Nankai subduction to Izu\u0026ndash;Bonin collision (Central Japan): implications for an extensional tear within the subducting slab. Earth Planet Sci Lett 172:1\u0026ndash;10. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/S0012-821X(99)00189-2\u003c/span\u003e\u003cspan address=\"10.1016/S0012-821X(99)00189-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNakajima J, Hirose F, Hasegawa A (2009) Seismotectonics beneath the Tokyo metropolitan area, Japan: Effect of slab-slab contact and overlap on seismicity. J Geophys Res 114:B08309. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2008JB006101\u003c/span\u003e\u003cspan address=\"10.1029/2008JB006101\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNakata T, Imaizumi T (2002) Digital active fault map of Japan (DVD-ROM) [in Japanese]. University of Tokyo\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNaoi M, Tamaribuchi K, Shimojo K et al (2024) Neural phase picker trained on the Japan meteorological agency unified earthquake catalog. Earth Planets and Space 76:150. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/s40623-024-02091-8\u003c/span\u003e\u003cspan address=\"10.1186/s40623-024-02091-8\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNational Research Institute for Earth Science and Disaster Resilience (2019) NIED Hi-net. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.17598/nied.0003\u003c/span\u003e\u003cspan address=\"10.17598/nied.0003\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNishimura T, Yokota Y, Tadokoro K, Ochi T (2018) Strain partitioning and interplate coupling along the northern margin of the Philippine Sea plate, estimated from Global Navigation Satellite System and Global Positioning System-Acoustic data. Geosphere 14:535\u0026ndash;551. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1130/GES01529.1\u003c/span\u003e\u003cspan address=\"10.1130/GES01529.1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOhta Y, Miura S, Ohzono M et al (2011) Large intraslab earthquake (2011 April 7, M 7.1) after the 2011 off the Pacific coast of Tohoku Earthquake (M 9.0): Coseismic fault model based on the dense GPS network data. Earth Planets and Space 63:1207\u0026ndash;1211. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5047/eps.2011.07.016\u003c/span\u003e\u003cspan address=\"10.5047/eps.2011.07.016\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOkada T, Hasegawa A (2003) The M7.1 May 26, 2003 off-shore Miyagi Prefecture Earthquake in northeast Japan: Source process and aftershock distribution of an intra-slab event. Earth Planets and Space 55:731\u0026ndash;739. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/BF03352482\u003c/span\u003e\u003cspan address=\"10.1186/BF03352482\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSaito T, Noda A (2023) Mechanically Coupled Areas on the Plate Interface in the Kanto Region, Central Japan, Generating Great Earthquakes and Slow-Slip Events. Bull Seismol Soc Am 113:1842\u0026ndash;1855. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1785/0120230073\u003c/span\u003e\u003cspan address=\"10.1785/0120230073\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSato H, Hirata N, Koketsu K et al (2005) Earthquake Source Fault Beneath Tokyo. Science (1979) 309:462\u0026ndash;464. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1126/science.1110489\u003c/span\u003e\u003cspan address=\"10.1126/science.1110489\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSawaki Y, Sato Y, Uchide T (2025a) FaultNVC: Earthquake Fault Plane Extractor through Point-Cloud Normal Vector Clustering for Hypocenter Distributions. Open-File Report of the Geological Survey of Japan, AIST 759. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.gsj.jp/publications/pub/openfile/openfile0759.html\u003c/span\u003e\u003cspan address=\"https://www.gsj.jp/publications/pub/openfile/openfile0759.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSawaki Y, Shiina T, Sagae K et al (2025b) Fault Geometries of the 2024 Mw 7.5 Noto Peninsula Earthquake from Hypocenter-Based Hierarchical Clustering of Point-Cloud Normal Vectors. J Geophys Res Solid Earth 130:e2024JB030233. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2024JB030233\u003c/span\u003e\u003cspan address=\"10.1029/2024JB030233\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSeno T (2005) Izu detachment hypothesis: A proposal of a unified cause for the Miyake-Kozu event and the Tokai slow event. Earth Planets and Space 57:925\u0026ndash;934. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1186/BF03351872\u003c/span\u003e\u003cspan address=\"10.1186/BF03351872\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSeno T, Yoshida M (2004) Where and why do large shallow intraslab earthquakes occur? Phys Earth Planet Inter 141:183\u0026ndash;206. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.pepi.2003.11.002\u003c/span\u003e\u003cspan address=\"10.1016/j.pepi.2003.11.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUchide T, Shiina T, Imanishi K (2022) Stress map of Japan: detailed nationwide crustal stress field inferred from focal mechanism solutions of numerous microearthquakes. J Geophys Res Solid Earth 127. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2022JB024036\u003c/span\u003e\u003cspan address=\"10.1029/2022JB024036\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. e2022JB024036\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eUeno H, Hatakeyama S, Aketagawa T et al (2002) Improvement of hypocenter determination procedures in the Japan Meteorological Agency (in Japanese with English abstract). Q J Seismology 65:123\u0026ndash;134\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWada I, Mazzotti S, Wang K (2010) Intraslab Stresses in the Cascadia Subduction Zone from Inversion of Earthquake Focal Mechanisms. Bull Seismol Soc Am 100:2002\u0026ndash;2013. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1785/0120090349\u003c/span\u003e\u003cspan address=\"10.1785/0120090349\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWald DJ, Somerville PG (1995) Variable-slip rupture model of the great 1923 Kanto, Japan, earthquake: Geodetic and body-waveform analysis. Bull Seismol Soc Am 85:159\u0026ndash;177. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1785/BSSA0850010159\u003c/span\u003e\u003cspan address=\"10.1785/BSSA0850010159\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWaldhauser F, Ellsworth WL (2000) A Double-Difference Earthquake Location Algorithm: Method and Application to the Northern Hayward Fault, California. Bull Seismol Soc Am 90:1353\u0026ndash;1368. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1785/0120000006\u003c/span\u003e\u003cspan address=\"10.1785/0120000006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWoollam J, M\u0026uuml;nchmeyer J, Tilmann F et al (2022) SeisBench\u0026mdash;A Toolbox for Machine Learning in Seismology. Seismol Res Lett 93:1695\u0026ndash;1709. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1785/0220210324\u003c/span\u003e\u003cspan address=\"10.1785/0220210324\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYamasaki T, Seno T (2003) Double seismic zone and dehydration embrittlement of the subducting slab. J Geophys Res Solid Earth 108:2212. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2002JB001918\u003c/span\u003e\u003cspan address=\"10.1029/2002JB001918\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoshida A (1993) Seismic Activities and Tectonics in and around Western Kanagawa [in Japanese]. J Geogr (Chigaku Zasshi) 102:407\u0026ndash;417. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5026/jgeography.102.4_407\u003c/span\u003e\u003cspan address=\"10.5026/jgeography.102.4_407\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYoshida K, Uchida N, Matsumoto Y et al (2023) Updip Fluid Flow in the Crust of the Northeastern Noto Peninsula, Japan, Triggered the 2023 M w 6.2 Suzu Earthquake During Swarm Activity. Geophys Res Lett 50:e2023GL106023. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1029/2023GL106023\u003c/span\u003e\u003cspan address=\"10.1029/2023GL106023\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYukutake Y, Takeda T, Honda R, Yoshida A (2012) Seismotectonics in the Tanzawa Mountains area in the Izu-Honshu collision zone of central Japan, as revealed by precisely determined hypocenters and focal mechanisms. Earth Planets and Space 64:269\u0026ndash;277. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.5047/eps.2011.09.002\u003c/span\u003e\u003cspan address=\"10.5047/eps.2011.09.002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhu W, Beroza GC (2019) PhaseNet: A Deep-Neural-Network-Based Seismic Arrival Time Picking Method. Geophys J Int 216:261\u0026ndash;273. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1093/gji/ggy423\u003c/span\u003e\u003cspan address=\"10.1093/gji/ggy423\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":true,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"earth-planets-and-space","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"epsp","sideBox":"Learn more about [Earth, Planets and Space](http://earth-planets-space.springeropen.com)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/epsp/default.aspx","title":"Earth, Planets and Space","twitterHandle":"@SpringerOpen","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Intraslab reverse fault, Fault geometry, Hypocenter clustering, Machine learning, Active fault, Sagami Trough, The 1923 Kanto earthquake","lastPublishedDoi":"10.21203/rs.3.rs-6353744/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6353744/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eWestern Kanagawa in eastern Japan is a region under complex tectonics with seismic activity. The northward subduction of Philippine Sea plate has historically caused M-8 class megathrust earthquakes, known as Kanto earthquakes. On 9 August 2024, an Mw 5.0 reverse-fault earthquake occurred near the epicenter of the latest Kanto earthquake in 1923. For future evaluation of the earthquake generation in this area, it is imperative to determine the precise location of this earthquake\u0026rsquo;s faulting: whether it occurred on the plate interface, an active fault, or within the slab. To tackle this question, we conducted a machine-learning-based workflow of (1) phase picking using PhaseNetWC, (2) hypocenter relocation with phase picks and waveform cross-correlations, and (3) extraction of rectangular fault planes through hypocenter clustering of positions and point-cloud normal vectors. Our result exhibited five fault planes. We obtained a significant plane dipping steeply to the south, consistent with the steeper nodal plane of the mainshock focal mechanism. This extracted plane, being 2 km deeper than the slab surface, demonstrates that the Mw 5.0 earthquake occurred as an intraslab earthquake on a steep reverse fault, as opposed to thrusting on the PHS surface or reverse faulting on an active fault in the forearc crust. Given the high stress rate on the plate interface in this area, complex stress states may have given rise to the occurrence of the steep reverse intraslab earthquake beneath highly coupled plate interface.\u003c/p\u003e","manuscriptTitle":"Intraslab reverse faulting adjacent to the hypocenter of the 1923 Kanto earthquake: The Mw 5.0 western Kanagawa earthquake in eastern Japan on 9 August 2024","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-06 18:52:12","doi":"10.21203/rs.3.rs-6353744/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Major Revision","date":"2025-06-18T09:51:21+00:00","index":"","fulltext":""},{"type":"reviewerAgreed","content":"","date":"2025-05-01T08:53:54+00:00","index":0,"fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-05-01T08:51:08+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-04-04T04:51:06+00:00","index":"","fulltext":""},{"type":"submitted","content":"Earth, Planets and Space","date":"2025-04-03T03:07:46+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"earth-planets-and-space","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"epsp","sideBox":"Learn more about [Earth, Planets and Space](http://earth-planets-space.springeropen.com)","snPcode":"","submissionUrl":"https://www.editorialmanager.com/epsp/default.aspx","title":"Earth, Planets and Space","twitterHandle":"@SpringerOpen","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"BMC/SO AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"4f378f06-b48b-4003-88e0-f2e554ad3fdb","owner":[],"postedDate":"May 6th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"published-in-journal","subjectAreas":[],"tags":[],"updatedAt":"2025-12-01T16:04:45+00:00","versionOfRecord":{"articleIdentity":"rs-6353744","link":"https://doi.org/10.1186/s40623-025-02311-9","journal":{"identity":"earth-planets-and-space","isVorOnly":false,"title":"Earth, Planets and Space"},"publishedOn":"2025-11-25 15:58:32","publishedOnDateReadable":"November 25th, 2025"},"versionCreatedAt":"2025-05-06 18:52:12","video":"","vorDoi":"10.1186/s40623-025-02311-9","vorDoiUrl":"https://doi.org/10.1186/s40623-025-02311-9","workflowStages":[]},"version":"v1","identity":"rs-6353744","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6353744","identity":"rs-6353744","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}