The Effect of Splay Faults on Earthquake Size and Rupture Behavior at Subduction Zones: Insights from Discrete Element Simulations

Deep water megasplay faults may play an important role in the generation of large coseismic slip and transoceanic tsunami. However, how they affect earthquake sizes is still not clear. Whether the megasplay faults promote or obstruct the earthquake rupture may depend on their fault type. One example is the apparent reactivation of a normal fault within the upper plate during the 2011 Tohoku-oki earthquake. The uplift and seaward displacement of the footwall of the splay normal fault may have contributed to the large horizontal displacement, which may explain why the earthquake rupture propagated all the way to the trench, and the peak coseismic slip patch was very close to the trench. In contrast, a megasplay thrust fault may act as a barrier, limiting the rupture area and the magnitude of megathrust earthquakes, for example the Santa Maria fault system (SMFS) along the Chile Margin. Motions along the SMFS may have slowly released the elastic strain accumulated in the forearc before the 2010 Maule earthquake. Thus, the SMFS may have contributed to constraining coseismic slip during the 2010 earthquake.

To better understand how megasplay faults affect earthquake sizes and ruptures, we build on recent modeling efforts based on observations of distributed extensional deformation in the Japan Trench forearc and Chile Margin, to test the effect of different types of splay faults. We model the upper plate as a wedge that is partitioned into an inner (velocity-weakening) wedge and outer (velocity-strengthening) wedge, with the splay fault marking the boundary between the two. We vary fault dip angle, dip direction, and relative distance to the downdip limit of outer (velocity-strengthening) wedge, and examine effects of changing friction conditions during seismic cycles.

Our models demonstrate that a splay normal fault can initiate during a great subduction earthquake and may then evolve during multiple seismic cycles. Moreover, our preliminary results show that the splay normal faults can facilitate the coseismic slip caused by the later megathrust earthquake, but may have less effect on the slip magnitude than the presence of a velocity-strengthening outer wedge. Furthermore, splay thrust faults can either promote coseismic slips or behave as a barrier to coseismic ruptures, depending on fault dips.