[S-CG40] Science of slow-to-fast earthquakes

[SCG40-33] 豊後水道における長期的スロースリップの発生に伴うフィリピン海プレート内での地震活動・応力変化

*北 佐枝子¹、ヒューストン ハイジ²、矢部 優³、田中 佐千子⁴、浅野 陽一⁴、木村 武志⁴ (1.建築研究所、2.南カリフォルニア大学、3.産業技術総合研究所、4.防災科学技術研究所)

キーワード：スロースリップ、スラブ内地震、応力変化、南海トラフ、プレートの引っ張り、プレート間固着

Relationships between in-slab events and two slow slip events (SSEs) were examined in the Bungo regions, where there are in-slab earthquakes and long-term SSEs (L-SSEs) with recurrence times of ~7 years and ~1 year durations. We applied a stress inversion method to focal mechanisms of in-slab events for ~20 years and examined temporal changes of stress axes relative to the occurrence times of the three L-SSEs. In general, the Sigma-3 axes from the inversions strike in the east-west direction, whereas the Sigma-1 axes are close to vertical (60 to 90 degrees) and almost parallel to the direction of plate motion of the subducting plate. During the 1.5 years before the start of the L-SSEs and during the occurrence of the L-SSEs, the plunges of the Sigma 1 axes decreased, becoming less vertical over the entire fault plane. The plunges of Sigma 1 rotate back to the original position during the 1.5 years after the end of the L-SSEs over the entire fault plane. But in the eastern portion of the fault plane, the plunges of Sigma 1 start to decrease later than in western portion, at 1 year before the start of the L-SSEs. This result is consistent with the propagation of the L-SSEs from west to east, which was also detected by a geodetic study (Seshimo et al. 2022). The stress ratio R in the western portion (0.2 to 0.7) is generally larger than that in the eastern portion (0.5 to 0.8), and the ratios in both portions generally decrease during the L-SSEs. This result suggests that locking of the plate boundary in the eastern portion is larger than that in the western portion because the locked region of the Nankai trough is relatively close to the eastern portion of the fault plane. The R in the southern portion is 0.1 to 0.8 and it decreases during the period of the L-SSEs. On the other hand, the R in the northern portion is generally ~0.5. The effect by slab pull in the northern portion is larger than that in the southern portion, and the slab pull in the oceanic plate in the northern portion grows larger during the period of the L-SSE due to unlocking of the plate boundary. These results suggest the promise of our methodology, using in-slab events to infer the timing of L-SSEs, to monitor ongoing spatiotemporal changes in locking of the plate boundary by aseismic slip and slab pull.