Extensional Failure of “Pre-Stressed” Lithosphere Above a Subduction Zone May Have Contributed to the Size of the Tohoku-Oki Earthquake and Tsunami

Abstract:

The Tohoku-oki earthquake was not only the costliest natural disaster in history it was the best monitored. The unprecedented data set showed that anomalously large lateral motion of the seafloor near the trench contributed to the size of the tsunami. Also, for the first time it was shown that a large subduction earthquake was followed by extensional aftershocks in a broad region of the upper plate (up to 250 km from the Japan Trench). Several observations suggest that the near-trench seafloor motion and the extensional aftershocks are linked. For example, a seismically imaged fault, just landward of the region of large seafloor motion, slipped in a normal sense during the earthquake. Also, inspired by the Tohoku data, researchers have searched for and found upper plate extensional aftershocks associated with several other subduction earthquakes that produced large tsunami.

Extension of the upper plate can be driven by a reduction in the dip of a subducting slab. Such a dip change is suggested by the post-Miocene westward migration of the volcanic arc in Honshu. Numerical models show that a long-term reduction in slab dip can generate enough extensional stress to cause normal faulting over a broad region of the upper plate. The time step of the numerical model is then reduced to treat the inter-seismic time scale of 100-1000 years, when the subduction interface is locked. The interface dip continues to be reduced during the inter-seismic period, but extensional fault slip is suppressed by the relative compression of the upper plate caused by continued convergence. The relief of compressional stresses during dynamic weakening of the megathrust triggers a release of bending-related extensional strain energy. This extensional yielding can add significantly to the co-seismic radiated seismic energy and seafloor deformation. This mechanism is analogous to the breaking of a pre-stressed concrete beam supporting a bending moment when the compressional pre-stress is removed. It is plausible that similar bending is occurring at a number of subduction zones. A testable prediction of this bending model is that inter-seismic stresses can be compressional near the surface of the upper plate, but should become extensional at depths accessible to drilling.