Shear partitioning near the central Japan triple junction: the 1923 great Kanto earthquake revisited—II

The faulting mechanism of the 1923 great Kanto earthquake is re-examined. The new fault model is constrained by: (1) the focal mechanism of the main shock; (2) triangulation data; (3) levelling data; (4) the distribution of aftershocks; (5) the geology of the ocean floor east of the Boso peninsula; and (6) the delineation of lateral variations in the dip of the subducting Philippine Sea plate. While factors (1)–(4) have been used to constrain earlier fault models, factor (5) leads us to propose the rupture of a major right-lateral fault (here named the Boso transform fault) which accommodated about 1.6 m of slip in the 1923 earthquake. Kinematic constraints derived from the 1923 main-shock fault geometry suggest that the Boso transform fault accommodates about 1.6 cm yr⁻¹ of the relative plate motion, in agreement with onland and marine geology and the uplift history of the Boso peninsula. All plausible fault models of the 1923 earthquake involve right-lateral slip of the Boso transform fault. Triangulation data are better explained by a model which specifies shallower dip of the subducting Philippine Sea slab southeast of the Miura peninsula, consistent with seismological observations.