17:15 〜 19:15
[SCG45-P37] Spatial Distribution of Slow and Fast Earthquakes Correlating to the Heterogeneity of Incoming Plate Structure along the Northern Japan Trench
キーワード：浅部テクトニック微動、微小地震、日本海溝
In the Japan Trench, slow and fast earthquakes are spatially complementary in their distribution. Furthermore, it is known that the characteristics of seismic activity differ significantly among the northern, central, and southern segments of this subduction zone (Nishikawa et al., 2023). These differences in seismic activity have been suggested to correlate with variations in the thickness of subducting sediment layers, with previous studies indicating that the sediment layer near the trench axis is thicker in the northern and southern segments, where tremors and very low-frequency earthquakes (VLFEs) are more active (Nakamura et al., 2023). Focusing on the northern segment, spatial variations are observed in both seismic activity and sediment layer thickness within this region. Does a correlation between seismic activity and sediment thickness exist even at such small spatial scales? This study aims to clarify the relationship between geological structure and seismic activity using a more accurate earthquake catalog than previously available, obtained by applying an improved detection and hypocenter determination method to dense offshore seismic observation data.
Traditionally, the Envelope Correlation Method (ECM, e.g., Obara, 2002) has been widely used for earthquake detection and location. This method takes advantage of the similarity of envelope waveforms among multiple stations, where time delays correspond to S-wave arrival time differences. However, there are limitations to improving the location accuracy due to two main issues: (1) Scattering waves generated by structural heterogeneities within the sediment layers distributed along the shallow plate boundary may degrade envelope similarity. (2) Constraining the focal depths using only S-wave arrival time differences is inherently difficult.
In this study, we introduced improvements in both (1) detection and (2) hypocenter location, leading to the following results:
(1) By incorporating a detection method that utilizes functions with high sensitivity to seismic wave arrivals (Hendriyana & Tsuji, 2021), the influence of later phases is expected to be suppressed. Compared with the conventional envelope-based approach, this method increased the number of detected events while also reducing travel time residuals in hypocenter locations.
(2) By relative hypocenter relocation using hypocenters of regular earthquakes with well-constrained depths, as references, focal depths of tremors are expected to be more accurately constrained. As a result, the tremor hypocenters were relocated along the plate boundary, suggesting that both regular earthquakes and tremors occur on the plate interface.
The earthquake catalog obtained in this study revealed variations in the duration of seismic waveforms along the strike of the trench, and these variations tended to correlate with the thickness of the incoming sediments (Nakamura et al., 2023). Regions with thick sediments were dominated by tremors with durations exceeding 30 seconds, whereas areas with average sediment thickness were dominated by events with relatively short durations (10-20 seconds). In addition, a comparison with reflection seismic survey records shows that areas with numerous long-duration tremors coincided with the graben structure on the incoming oceanic crust. This suggests that the graben-fill sediments present at the plate boundary may facilitate a slower rupture process compared to regular earthquakes or generate more scattered waves responsible for the observed long-duration waveforms.
Traditionally, the Envelope Correlation Method (ECM, e.g., Obara, 2002) has been widely used for earthquake detection and location. This method takes advantage of the similarity of envelope waveforms among multiple stations, where time delays correspond to S-wave arrival time differences. However, there are limitations to improving the location accuracy due to two main issues: (1) Scattering waves generated by structural heterogeneities within the sediment layers distributed along the shallow plate boundary may degrade envelope similarity. (2) Constraining the focal depths using only S-wave arrival time differences is inherently difficult.
In this study, we introduced improvements in both (1) detection and (2) hypocenter location, leading to the following results:
(1) By incorporating a detection method that utilizes functions with high sensitivity to seismic wave arrivals (Hendriyana & Tsuji, 2021), the influence of later phases is expected to be suppressed. Compared with the conventional envelope-based approach, this method increased the number of detected events while also reducing travel time residuals in hypocenter locations.
(2) By relative hypocenter relocation using hypocenters of regular earthquakes with well-constrained depths, as references, focal depths of tremors are expected to be more accurately constrained. As a result, the tremor hypocenters were relocated along the plate boundary, suggesting that both regular earthquakes and tremors occur on the plate interface.
The earthquake catalog obtained in this study revealed variations in the duration of seismic waveforms along the strike of the trench, and these variations tended to correlate with the thickness of the incoming sediments (Nakamura et al., 2023). Regions with thick sediments were dominated by tremors with durations exceeding 30 seconds, whereas areas with average sediment thickness were dominated by events with relatively short durations (10-20 seconds). In addition, a comparison with reflection seismic survey records shows that areas with numerous long-duration tremors coincided with the graben structure on the incoming oceanic crust. This suggests that the graben-fill sediments present at the plate boundary may facilitate a slower rupture process compared to regular earthquakes or generate more scattered waves responsible for the observed long-duration waveforms.