Numerical simulation of crustal deformation near the trench axis

Monday, 15 December 2014
Tsuyoshi Watanabe, Keiichi Tadokoro, Kenji Yasuda, Cosmo Fujii and Kenjiro Matsuhiro, Nagoya University, Nagoya, Japan
The Philippine Sea plate subducts beneath southwest Japan along the Nankai Trough with a rate of about 4-6 cm/yr, where megathrust earthquakes have repeatedly occurred with recurrence intervals of about 100-150 years. It is known that these earthquakes often ruptured adjacent segments and brought the Japanese Islands serious damage like the 2011 Tohoku earthquake. Especially the damage caused by tsunami which arises from co-seismic slip at the shallow part of plate boundary is serious. Thus, it is important to investigate the status of strain accumulation near the trench axis in inter-seismic period. For this issue, we started seafloor geodetic observations in 2013 using a GPS/Acoustic technique at two or three sites across the trench axis, which are located in the shallow extension of the rupture area of the 1944 Tonankai earthquake. Therefore, on the basis of these observations, we investigated the crustal movement at the shallow part of plate boundary through numerical simulation using finite element method. Until now back slip model have been used to interpret the crustal deformation derived from onshore geodetic observations such as GPS, and then this model achieved a good deal of successful results. However, in late years it is doubted for crustal deformation at the shallow part of plate boundary whether it is appropriate to use back slip model there. In this study, we assumed frictionless plate boundary from the trench to a certain depth and assigned back slip to plate boundary deeper than a certain depth as boundary conditions. The frictionless up-dip segment is dragged passively toward the deep part by the back slip of the down-dip segment. This model reproduces the unlocked zone at the shallow depth and locked zones at the deep depth. We investigated the horizontal and vertical displacement profiles on the surface accompanied with the variation of the depth of the up-dip limit assigning back slip. This result showed that the peaks of horizontal displacement profiles correspond to the locations of starting point of back slip. In addition, horizontal displacement near the trench axis increases as the depth of the up-dip limit of back slip become shallow. For vertical displacements, we cannot detect the significant differences among the models.