Precise hypocenter determination of a natural earthquake by the reflection survey array in Japan

One large earthquake (Mj6.4) occurred in northern Japan on July 23, 2003, called the 2003 northern Miyagi earthquake. Locations of its aftershocks were determined using dense temporary seismic network by Tohoku University, Japan. Its spatial distribution showed that the fault plain tilts westward and that its extrapolation reaches to the Ishinomaki-wan fault at the surface. In 2003 and 2005, we conducted two seismic reflection surveys above and near the aftershock zone. The 2003 survey line was extended to 15 km in length from east to west across above the zone. The 2005 survey, by contrast, consisted of two crossed lines above the bottom of that; these lines run north-south and east-west. Each line had 192 stations and was 2 km long. The 2003 survey revealed detailed seismic structure across the source area. Two west-tilted faults were found to 3-km depth beneath the Ishinomaki-wan fault and the Asahiyama flexure in the depth section of the 2003 survey. During the 2005 survey period, a small aftershock of the 2003 earthquake was observed at the stations. We re-determined the hypocenter location of the aftershock by using observed first-arrival times. The velocity model used is the one determined by the 2003 survey. The hypocenter is relocated beneath the western stations of the 2005 survey at 5-km depth; this position is close to the lower extension of the flexure. It is 1.6 km west of the one determined by the Japan Meteorological Agency (JMA) and MEXT Japan, and 8 km shallower than that. We also installed one 3-component seismograph nearby during the 2005 survey period. The observed S-P time of the aftershock at the seismograph can be explained by the relocated hypocenter with a reasonable Vp/Vs ratio. On the contrary, the one determined by the JMA and MEXT is too far from the seismograph to explain this S-P time with realistic ratios. Previous reflection surveys showed a 2-3 km thick sedimentary layer in the west of the Ishinomaki-wan fault, which area contains the 2005 survey lines. A previous gravity survey also suggested this heterogeneous structure extends to a wider area. This heterogeneous structure will cause eastward and deeper bias of the hypocenters estimated by using laterally homogeneous velocity models with higher velocities than those estimated from the reflection surveys at a shallow depth. This mislocation is consistent with our result.