日本地球惑星科学連合2021年大会

講演情報

[J] 口頭発表

セッション記号 S (固体地球科学) » S-SS 地震学

[S-SS08] 地震発生の物理・断層のレオロジー

2021年6月3日(木) 15:30 〜 17:00 Ch.20 (Zoom会場20)

コンビーナ:金木 俊也(京都大学防災研究所)、大谷 真紀子(東京大学地震研究所)、岡崎 啓史(海洋研究開発機構)、吉田 圭佑(東北大学理学研究科附属地震噴火予知研究観測センター)、座長:大谷 真紀子(東京大学地震研究所)、吉田 圭佑(東北大学理学研究科附属地震噴火予知研究観測センター)

15:45 〜 16:00

[SSS08-02] Rupture segmentation caused by fault bends in simulated earthquake sequences

*小澤 創1、安藤 亮輔1 (1.東京大学大学院理学系研究科)


キーワード:地震サイクル、断層セグメント化、数値シミュレーション、速度状態依存摩擦、バリア、断層形状

Observations show, in several faults, earthquake ruptures are segmented by barriers - locations that tend to stop rupture propagations. One plausible mechanism of barriers is geometrical complexity on the fault. For example, Elliott et al. 2018 shows geomorphological evidence for persistent rupture terminations at a restraining bend in the Altyn Tigh fault.

Here, we study how earthquake ruptures are arrested at bends from physics-based modeling and numerical simulations. Specifically, we consider two fault segments connected each other via a restraining/releasing bend as a simple and instructive example. An important mechanical result of slip on a nonplanar fault is the change of normal stresses. Normal stress increases at restraining bends and decreases at releasing bends. The rupture termination occurs when the fracture energy exceeds the static energy release rate at the rupture front. Because the fracture energy is proportional to normal stress, we expect that ruptures is arrested around the center of a restraining bend.

To verify our hypothesis, we perform quasi-dynamic earthquake sequence simulations on two fault segments connected by a restraining/releasing bend. The use of earthquake sequence simulations allows us to obtain the initial condition of ruptures in a self-consistent manner, and to ask how often the bend acts as a barrier. In the case of restraining bends, our simulation shows a clear segmentation due to the bend, and the ratio of partial ruptures increases with increasing cumulative slip. Hypocenters are concentrated at both edges of the bend, where normal stress takes local minimum. Although much less effective than restraining bends, releasing bends could also act as barriers.

We perform systematic geometrical parameter investigation; the length and slope of the bend. The ratio of full ruptures decreases with increasing the slope of the bend. The effect of the bend length on the ratio of full ruptures is more complex because it interacts with the process zone size. Lastly, we compare geometrical barriers and rheological barriers in concept, and discuss unmodeled effects, such as inertia and off-fault plasticity.