Cycles of slow slip events on non-planar subduction faults and their implications on megathrust earthquakes
Duo Li, Ludwig-Maximilians-Universität München

Slow slip and slow earthquakes (SSEs), driven by transient aseismic deformation along the subduction interface, are proposed to affect the initiation of megathrust earthquakes. Prominent examples, as inferred from geodetic and seismic observations, include the 2011 M9.1Tohoku-Oki,, the 2014 Mw8.1 Iquique (Chile), and the 2014 Mw7.3 Guerrero (Mexico) earthquakes. Understanding the conditions, such as 3D geometry, governing the dynamics of SSEs and, specifically, its interaction with megathrust earthquake initiation has the potential to fundamentally advance our understanding of subduction zone faulting. However, since observational coverage of subduction zones is limited and direct monitoring of earthquake nucleation phases is currently impossible, the physics underlying deep stress buildup remain elusive. We here present the numerical simulations of long-term quasi-dynamic slow slip cycles in the so-called Guerrero seismic gap to explore the physics of several Mw>7.0 transient slow slip events and their significant role in the long-term plate coupling (Perez-Silva et al.,2021). Our model suggests that the flat-slab segment of the Cocos plate aids the SSE large magnitudes and long recurrence intervals and also leads to significant shear traction concentration at the base of the seismogenic zone. We then use these transient shear traction as the initial condition for a 3D dynamic rupture scenario of the 2014 Mw7.3 Guerrero earthquake. We show that stress concentrations caused by the deep SSEs are significant enough to affect the initiation of spontaneous earthquake dynamic rupture.

Modeling spontaneous and triggered slow-slip events at the Hikurangi subduction plate interface
Bunishiro Shibazaki, Building Research Institute;

In this talk, I present the models of spontaneous slow slip events (SSEs) along the Hikurangi subduction zone and triggered SSEs by static and dynamic stress kicks by the Kaikoura earthquake (Shibazaki et al., 2019). Recent studies revealed the occurrence of various slow-slip events (SSEs) along the Hikurangi subduction plate interfaces (e.g., Wallace and Beavan, 2010). Long-term SSEs with a duration of 1.5 years (Manawatu and Kapiti SSEs) occur in the deeper portion of the Hikurangi subduction zone, and shallow, short-term SSEs with a duration of 1–3 weeks occur along with the northern and central parts of the subduction zone. On 14 November 2016, Mw 7.8 Kaikoura earthquake occurred in the northeastern part of the South Island of New Zealand (Hamling et al., 2017). The Kaikoura earthquake triggered a large, shallow SSE and a deep Kapiti SSE of the Hikurangi subduction zone. The earthquake did not trigger a deep Manawatu SSE. The dynamic stress changes in the shallow SSE zone are calculated to be on the order of 200–700 kPa, though the static stress change in this zone is very small (0.2–0.7 kPa) (Wallace et al, 2017). Therefore, dynamic triggering is considered to have caused the shallow SSE. On the other hand, static stress change in the Kapiti SSE is large (500 kPa), because this zone is very close to the source region of the Kaikoura earthquake.

We consider a rate-and-state friction law with a cutoff velocity to the evolution effect that exhibits a transition from velocity weakening to velocity strengthening with increasing slip velocity, following on from the experimental studies of Rabinowitz et al. (2018). We consider a realistic configuration of the plate interface. We set the effective stress and critical displacement of long-term SSE zones to be much larger than those of shallow short-term SSE zones. We consider shear and normal stress changes calculated by the source model of the Kaikoura earthquake by Hamling et al. (2017). In the southern part of the Kapiti SSE zone, the shear stress increase exceeds 0.2 MPa. This value is larger than the theoretically estimated stress drop of 0.13 MPa for Kapiti SSEs. We give stress changes at several time steps to examine whether or not SSEs are triggered. Kapiti SSEs are triggered in almost all cases. However, in the main slip zone of the Manawatu SSE, SSEs are not triggered. This result is consistent with the observed results.

We also try to model SSEs caused by dynamic triggering. For simplicity, we consider only a region of shallow SSEs and give a stress perturbation that is a sine function of time and that propagates from south to north. Just after the perturbation, slip velocity increases, and slips continue to occur for a few days. Our results suggest that dynamic shear stress changes caused by wave propagation along the fault can trigger shallow short-term SSEs.

Numerical modeling of deep long- and short-term SSEs in the Nankai and Hyuganada region
Takanori Matsuzawa, National Research Institute for Earth Science and Disaster Resilience, Japan
Bunishiro Shibazaki, Building Research Institute, Japan

Slow slip events (SSEs) are often found in the vicinity of the region where large earthquakes are expected. Recurrences of long-term SSEs are found in the Tokai and Bungo Channel region, southwestern Japan. In recent years, long-term SSEs have been also reported in the Hyuganada, Kii Channel, and northern Kii region (e.g., Ozawa et al., 2017; Kobayashi, 2017; Kobayashi and Tsuyuki, 2019). Takagi et al. (2019) suggests that long-term SSEs in Hyuganada occur at the intervals of 2-3 years, which is more frequent than the SSEs in the other regions (e.g. Bungo Channel). These studies suggest that long-term SSEs are common phenomena and not localized at a specific region, while periodical recurrence of the SSEs seems to be more clearly found in the Bungo Channel, Tokai, and Hyuganada. In this study, we aim to reproduce such long- and short-term SSEs in the Nankai and Hyuganada region, within a single comprehensive model.

We assume a rate- and state-dependent frictional law (RS-law) with a cut-off velocity to simulate SSEs. The numerical modeling is similar to our previous study (Matsuzawa et al., 2013). The modeled region is spanned from the off Tanegashima and the Tokai region. In our model, we assumed short-term SSE region based on the actual distribution of tremor, and set negative (a-b) in RS-law within this region. We posed long-term SSE region changing the width of low effective normal stress region, as assumed in the previous model of the Shikoku region (Matsuzawa et al., 2013). The width of long-term SSE region in Tokai is similar to the Bungo Channel. In addition, the width in Hyuganada is slightly narrower than those in Tokai and the Bungo Channel.

Our numerical result reproduced recurring long-term SSEs in the most regions at the depth from 25 to 29 km, which is typical depth of observed long-term SSEs. In the Hyuganada region, long-term SSEs recur at the interval of about two years, as in the observation (Takagi et al., 2019). Our model reproduced that long-term SSEs can commonly occur in the Nankai region. In addition, our model also implies that there are two types of long-term SSE regions. One is the region where long-term SSEs have similar size during seismic cycles, as found in the Bungo Channel, Tokai, and Hyuganada region. The other is the region where slip area of long-term SSEs becomes larger at the later stage of the seismic cycles, and recurrence intervals are usually longer than several years. In our result, such long-term SSEs with longer recurrence intervals are found in the Central Shikoku, and Kii Peninsula. Further long-term observation data would validate our modeling approach.

REGISTER TO ATTEND [register]