[S-CG50] Dynamics in mobile belts

[SCG50-P03] Earthquake simulation based on Coulomb stress change on inland faults during megathrust earthquake cycles in southwest Japan

*Tsukasa MITOGAWA¹, Takuya NISHIMURA² (1.Graduate School of Science, Kyoto University, 2.Disaster Prevention Research Institute, Kyoto University)

Keywords:Megathrust Earthquake Cycles, Southwest Japan, Inland Earthquake, Earthquake Simulation, Active Period

In the subduction zone, the stress changes due to megathrust earthquakes may modulate the shallow crustal seismicity in the overriding plate. Historical documents indicate the frequent occurrence of large shallow crustal earthquakes in the overriding continental plate 50 years before and 10 years after the megathrust earthquakes along the Nankai trough in southwest Japan (i.e., active period). We performed the earthquake simulation based on Coulomb failure stress (CFS) change assuming the strength of inland block boundary faults using the locking distribution of the megathrust fault and the relative motion rate of block boundary faults (hereafter "inland faults") estimated using geodetic data. The stress change is computed using numerical code by Fukahata and Matsu'ura (2006), assuming the medium composed of an elastic layer overlying a Maxwell viscoelastic half-space. The stress sources of inland faults are the slip (i.e., earthquake) and locking on the megathrust fault and inland faults. The stress change due to megathrust fault locking is calculated using geodetic backslip rates. On the other hand, the backslip rate of the inland faults is assumed by multiplying the geodetic rate of relative block motion by an arbitrary coupling ratio uniformly for all the inland faults. The megathrust earthquake releases backslip accumulated during the interseismic period every 120 years. The inland earthquake occurs when the CFS exceeds the assumed fault strength of an inland fault. As a result of the simulation, the "active period" of the past seismic activity in southwest Japan could be explained when the apparent friction coefficient and the coupling ratio of inland faults were about 0 and 0.1, respectively. The tendency of simulated long-term seismic activity is consistent with the past seismic activity at the low apparent friction coefficient. It suggests that the fault strength hardly changes over time. Furthermore, the small coupling ratio indicating low loading rates of the inland faults suggests that most part of geodetic observed deformations are inelastic deformations. Another interpretation is also possible that the low loading rates imply that the geodetic relative block motion is distributed among slips on not only major faults but also minor faults along the block boundary.