Seismic images of the source area of the 2004 Mid-Niigata prefecture earthquake in Northeast Japan

To better understand the generation mechanism of the 23 October 2004 Mid-Niigata prefecture earthquake (M 6.8), we used 120,352 P-wave and 95,391 S-wave high-quality arrival times from 5013 earthquakes to determine the three-dimensional seismic velocity (Vp, Vs) and Poisson's ratio (σ) structures in and around the source area. The hypocenter locations of the aftershocks are relocated accurately by using absolute and relative travel time (double-difference) data. Our results demonstrate that the seismic velocity and Poisson's ratio vary markedly around the source area. Most active faults are located along the margins of low-velocity and high-velocity zones. An obvious change in seismic velocity and Poisson's ratio occurs between the northwest and southeast parts of the study area, and this boundary zone reflects the spatial distribution of active faults, being well consistent with the surface geological features. Most of the large historic crustal earthquakes are located in or around the low seismic velocity zones. The mainshock hypocenter is located near the margin of low-velocity (high-Poisson's ratio) and high-velocity (low-Poisson's ratio) anomalies along the active fault zone. A zone with pronounced low-velocity and high-Poisson's ratio is clearly imaged in the lower crust under the source area, reflecting the existence of fluids that are released due to the dehydration of the subducting Pacific slab. Such fluids might have reduced the mechanical strength of the fault zone, and thus initiated the Mid-Niigata prefecture earthquake.