A dynamic rupture simulation for the generation process of the 2011 Tohoku earthquake

A great earthquake of the 2011 Off the Pacific Coast of Tohoku Earthquake occurred at the subduction plate boundary east off northern Honshu on March 11, 2011. Its magnutide, Mw9.0, was far larger than the expected M7-class earthquakes on the asperities off the Miyagi Prefecture (e.g., Yamanaka and Kikuchi [2004]). The rupture process has been investigated by many researchers. Now we have a qualitatively common image of the rupture process, although the details are not clarified. The rupture started from the hypocenter approximately at 38N, 143E, like an ordinary small earthquake (Chu et al. [2011]), but it propagated through a huge region, the size of which was about 450 km in length and 200 km in width. The largest slip area was located in the east of the hypocenter, possibly near the Japan trench (e.g., Yoshida et al. [2011], Fujii et al. [2011]). On the other hand, mechanical aspects of the Tohoku earthquake have not been clarified yet. Clarification of the mechanical aspects is important for understanding the great earthquake. In particular, why the initial (apparently ) ordinary earthquake grew to the great event is the issue to be tackled. Here we show numerical simulations of dynamic rupture propagation for the 2011 Tohoku earthquake. A planar fault is set in a 3D elastic body. Following a scheme of the boundary integral equation method, the equation of motion is solved at numerical cells on the fault. Friction evolution on the fault is controlled by both changes of frictional coefficient and pore fluid pressure. We especially focus on the generation part of the great event. In most cases, an initial break does not grow to a large event, but in some cases, it can. We constrain the parameter spaces for great earthquakes, and imply that the location of the preceding M7-class events before the 2011 Tohoku earthquake played a key role for the occurrence of the great earthquake.