17:15 〜 18:45
[SCG40-P01] Slow slip event nucleation by tremor activities unpinning locked Episodic Tremor and Slip zone
キーワード：スロースリップイベント、微動、スロー地震、GNSS、Episodic tremor and slip
Slow slip events (SSEs) and tremors are aseismic and seismic components of slow earthquakes, which radiate no or feeble seismic waves. They are transient shear faulting like ordinary earthquakes but occur down-dip of the seismogenic locked zone, where the megathrust rheology is in transition from brittle to ductile regime. Geological outcrops of the ETS zone fault support the presence of brittle patches embedded in the ductile background fault, which likely host tremors and SSEs, respectively. Although ductile fault rheology is considered a mechanism keeping SSEs slow, the spontaneous occurrence of SSEs would not be realized in such cases. Various mechanisms have been proposed to reconcile them through mechanical numerical simulations, but key observations are still lacking to examine these mechanisms.
Known as the term Episodic Tremor and Slip (ETS), the spatiotemporal proximity of the short-term SSEs is one of the fundamental observational characteristics. However, the daily GNSS time series, widely used for SSE characterization, is much coarser than the temporal resolution of tremor evolution. Hence, the temporal relationship between tremors and SSEs has not been revealed in detail. Furthermore, temporal smoothing of slip inversions should distort the estimated temporal evolution of SSE moments, particularly their onset. Hence, the detailed nature of ETS nucleation has been hardly resolved. In this study, we attempt to provide a new observational constraint to the nucleation stage of ETS by analyzing the 5-minute subdaily GNSS time series without slip inversions.
We selected major short-term SSEs in Northern and Central Cascadia from Michel et al. (2019, PAGEOPH) and defined spatiotemporal windows for their nucleation stages based on the along-strike migration pattern of corresponding PNSN tremors. Then, we extracted the temporal evolution of the SSE moment during the nucleation stage by stacking GNSS coordinates processed by Nevada Geodetic Laboratory. For the stacking, we weighted time series at each site according to static displacements expected from an interface slow slip, computed from a rectangle fault enclosing tremors at the nucleation stage. As a tremor proxy, we prepared synthetic displacement time series by computing point-source displacements due to tremors at each GNSS site and then stacked them as we did for the GNSS stack. We estimated the time lag between the SSE and tremor proxies using cross-correlation between them. The results for 13 nucleation stages show that the slip proxy is usually delayed by the tremor proxy by about 0 – 2 days. We found that the SSE moment release rate is on average lower in the earlier than in the later stages, suggesting that the development of the SSE moment is nonlinear to the tremor activity during the ETS nucleation stage. Based on these results, we propose that the nucleation of SSEs is a recurring feedback process between a collective rupture of brittle tremor patches and the passive activation of aseismic creep of bulk ductile fault. The inter-ETS full locking at the ETS zone (Saux et al., 2022, JGR) suggests that spatially distributed tremor patches act as pins coupling the slab and upper plate. A significant number of these patches should need breaking (unpinning the interface) to develop observable moment release due to slip, which we observed as the time lag between slip and tremors. The apparent spontaneous nucleation of SSE can be hence understood as the passive response of ductile or velocity-strengthening background fault to the unpinning of the tremor patches. Unpinning at the very beginning of each ETS to launch the feedback process could be driven externally, for example, by fluid pressure perturbation and/or tectonic loading due to the stable sliding zone. The presence of an unpinning threshold to activate observable slip moment explains why short-lasting tremor bursts during inter-ETS stages are not accompanied by observable slip signals.
Known as the term Episodic Tremor and Slip (ETS), the spatiotemporal proximity of the short-term SSEs is one of the fundamental observational characteristics. However, the daily GNSS time series, widely used for SSE characterization, is much coarser than the temporal resolution of tremor evolution. Hence, the temporal relationship between tremors and SSEs has not been revealed in detail. Furthermore, temporal smoothing of slip inversions should distort the estimated temporal evolution of SSE moments, particularly their onset. Hence, the detailed nature of ETS nucleation has been hardly resolved. In this study, we attempt to provide a new observational constraint to the nucleation stage of ETS by analyzing the 5-minute subdaily GNSS time series without slip inversions.
We selected major short-term SSEs in Northern and Central Cascadia from Michel et al. (2019, PAGEOPH) and defined spatiotemporal windows for their nucleation stages based on the along-strike migration pattern of corresponding PNSN tremors. Then, we extracted the temporal evolution of the SSE moment during the nucleation stage by stacking GNSS coordinates processed by Nevada Geodetic Laboratory. For the stacking, we weighted time series at each site according to static displacements expected from an interface slow slip, computed from a rectangle fault enclosing tremors at the nucleation stage. As a tremor proxy, we prepared synthetic displacement time series by computing point-source displacements due to tremors at each GNSS site and then stacked them as we did for the GNSS stack. We estimated the time lag between the SSE and tremor proxies using cross-correlation between them. The results for 13 nucleation stages show that the slip proxy is usually delayed by the tremor proxy by about 0 – 2 days. We found that the SSE moment release rate is on average lower in the earlier than in the later stages, suggesting that the development of the SSE moment is nonlinear to the tremor activity during the ETS nucleation stage. Based on these results, we propose that the nucleation of SSEs is a recurring feedback process between a collective rupture of brittle tremor patches and the passive activation of aseismic creep of bulk ductile fault. The inter-ETS full locking at the ETS zone (Saux et al., 2022, JGR) suggests that spatially distributed tremor patches act as pins coupling the slab and upper plate. A significant number of these patches should need breaking (unpinning the interface) to develop observable moment release due to slip, which we observed as the time lag between slip and tremors. The apparent spontaneous nucleation of SSE can be hence understood as the passive response of ductile or velocity-strengthening background fault to the unpinning of the tremor patches. Unpinning at the very beginning of each ETS to launch the feedback process could be driven externally, for example, by fluid pressure perturbation and/or tectonic loading due to the stable sliding zone. The presence of an unpinning threshold to activate observable slip moment explains why short-lasting tremor bursts during inter-ETS stages are not accompanied by observable slip signals.