Effect of the Earth’s surface topography on the quasi-dynamic earthquake cycle

Abstract:

For quasi-dynamic earthquake cycle simulations (ECSs) using BIEM, we have developed a method of calculating slip response function (SRF) in a homogeneous elastic medium with an arbitrary shaped Earth’s surface topography (Ohtani and Hirahara, 2013; Paper1). In this study, we report the improvement in our method.

Following Hok and Fukuyama (2011), we set the Earth’s surface as a free surface, in addition to the fault interface, in a homogeneous full-space medium. Then, using the analytic solution in full-space, we can calculate the Earth’s surface deformation, then the SRF change.

The surface cell setting determines the accuracy. For reducing the computational amount, we use the different sizes of the surface region and its divided subfault cells, depending on the fault depth. Paper1 used the uniform size for surface cells. Here, we improved our method where the Earth’s surface cells closer to the trench have the finer sizes for achieving more accuracy.

With such numerical SRF, we performed the quasi-dynamic ECS on a model, where the Earth’s surface is convex upward. Basically, with this topography, the slip behavior approaches the full-space case, from the half-space with flat surface case. This is because the distance from the Earth’s surface to the fault becomes large. When we set two asperities with negative A – B in the positive A – B background at 10km and 35km depths, the two asperities rupture independently. The recurrence time of the shallow asperity is Tr_s^half = 34.95, Tr_s^flat = 34.89, and Tr_s^actual =32.82 years, when using analytic SRF in half-space, and numerical SRF with flat surface and with actual topography, respectively. For each case, the recurrence time of the deep asperity is Tr_{1_d}^half = 26.80, Tr_{1_d}^flat = 26.89, and Tr_{1_d}^actual =26.69 years. Thus, the shallower asperity is more affected by the Earth’s surface topography than the deeper one, because the distance change rate from the surface to the fault is larger. On the other hand, when we set the entire seismogenic zone as negative A – B to produce coseismic slip, the slip behavior showed almost no change. This is because the rupture starts from the deep portion, and the deep area has less affected by the Earth’s surface topography.

We also examined the realistic case assuming the Nankai Trough, the subduction zone located in southeast, Japan.