講演情報
[S09-09]Spatiotemporal Characteristics of Fast and Slow earthquakes in the Hyuga-nada Region from OBS Observations
*胡 靚妤1,2、篠原 雅尚2、山下 裕亮3、利根川 貴志4、山田 知朗2、悪原 岳2、望月 公廣2 (1. 東京大学大学院理学系研究科、2. 東京大学地震研究所、3. 宮崎公立大学人文学部地域連携・防災研究センター、4. 国立研究開発法人海洋研究開発機構)
The Hyuga-nada region lies at the westernmost Nankai Trough, where the Philippine Sea Plate subducts northwestward beneath Kyushu. It is characterized by the subducting Kyushu-Palau Ridge (KPR) and represents a transition zone from strong interplate coupling in the Nankai Trough to weak coupling in the Ryukyu Trench. Although no earthquakes larger than M8 have been documented here, the region is seismically active, with M6–7 events recurring every 10–20 years. Hyuga-nada is also notable for diverse slow earthquakes, including low-frequency tremors (LFTs), very low-frequency earthquakes (VLFEs), and slow slip events, making it an ideal setting to investigate fast–slow earthquake interactions.
This study aims to clarify seismic activity in Hyuga-nada by precisely relocating fast earthquakes and LFTs from long-term marine observations, and by comparing them with previously studied VLFEs. The relationship between slow and fast events remains poorly understood due to limited spatiotemporal data, but resolving it may help explain the absence of great earthquakes in this region. Marine seismic arrays of long-term ocean bottom seismometers (LT-OBSs) were deployed during five observation periods (2015–2016, 2017–2018, 2022, 2023, and 2024), providing high-quality seafloor records.
For fast earthquakes, P- and S-wave arrivals and first-motion polarities were manually picked. Initial locations were determined using a 1-D velocity model with station corrections, followed by double-difference relocation with absolute and differential times in a 3-D velocity structure. Focal mechanisms were estimated through grid search using P-wave polarities and S/P amplitude ratios.
LFTs were re-analyzed from previously obtained differential S-wave arrivals between station pairs. These were converted into estimated arrivals, and events with ≥5 picks were located using the same relocation workflow as fast events. Events with horizontal or vertical errors <2 km were further refined using the double-difference method. The relocated events are referred to as low-frequency events.
Results show that fast earthquakes generally occurred beneath the plate boundary, distributed uniformly in time and space except the swarms. Focal mechanisms indicate strike-slip and normal faulting as dominant. Low-frequency events were shallower, typically along the plate boundary, and no assumptions about their depths were made before relocation.
Swarms were identified as tightly clustered fast earthquakes, consistent with JMA catalog data. Comparison with VLFEs revealed that swarm areas coincided with VLFE activity, while low-frequency events were distributed more broadly. Temporal sequencing showed low-frequency events became active days before swarms, and VLFEs occurred hours before fast earthquakes during swarm periods. Swarm focal mechanisms exhibited a higher proportion of thrust faulting. These patterns suggest that tremor and VLFE activity along the plate interface facilitated fluid migration through slab-bending-related fractures, promoting pore-pressure changes and triggering swarms.
The distribution of low-frequency events indicates control by plate boundary heterogeneity. A low-seismicity zone of slow events above the trenchward summit of the subducting KPR is consistent with a stress shadow generated by the seamount. Furthermore, the scarcity of intraslab fast earthquakes north of the seamount, compared with their abundance to the south, may reflect differences in the age and thermal state of the subducting plate across the ridge.
This study aims to clarify seismic activity in Hyuga-nada by precisely relocating fast earthquakes and LFTs from long-term marine observations, and by comparing them with previously studied VLFEs. The relationship between slow and fast events remains poorly understood due to limited spatiotemporal data, but resolving it may help explain the absence of great earthquakes in this region. Marine seismic arrays of long-term ocean bottom seismometers (LT-OBSs) were deployed during five observation periods (2015–2016, 2017–2018, 2022, 2023, and 2024), providing high-quality seafloor records.
For fast earthquakes, P- and S-wave arrivals and first-motion polarities were manually picked. Initial locations were determined using a 1-D velocity model with station corrections, followed by double-difference relocation with absolute and differential times in a 3-D velocity structure. Focal mechanisms were estimated through grid search using P-wave polarities and S/P amplitude ratios.
LFTs were re-analyzed from previously obtained differential S-wave arrivals between station pairs. These were converted into estimated arrivals, and events with ≥5 picks were located using the same relocation workflow as fast events. Events with horizontal or vertical errors <2 km were further refined using the double-difference method. The relocated events are referred to as low-frequency events.
Results show that fast earthquakes generally occurred beneath the plate boundary, distributed uniformly in time and space except the swarms. Focal mechanisms indicate strike-slip and normal faulting as dominant. Low-frequency events were shallower, typically along the plate boundary, and no assumptions about their depths were made before relocation.
Swarms were identified as tightly clustered fast earthquakes, consistent with JMA catalog data. Comparison with VLFEs revealed that swarm areas coincided with VLFE activity, while low-frequency events were distributed more broadly. Temporal sequencing showed low-frequency events became active days before swarms, and VLFEs occurred hours before fast earthquakes during swarm periods. Swarm focal mechanisms exhibited a higher proportion of thrust faulting. These patterns suggest that tremor and VLFE activity along the plate interface facilitated fluid migration through slab-bending-related fractures, promoting pore-pressure changes and triggering swarms.
The distribution of low-frequency events indicates control by plate boundary heterogeneity. A low-seismicity zone of slow events above the trenchward summit of the subducting KPR is consistent with a stress shadow generated by the seamount. Furthermore, the scarcity of intraslab fast earthquakes north of the seamount, compared with their abundance to the south, may reflect differences in the age and thermal state of the subducting plate across the ridge.
