Imaging the asperities of the 1997 northwestern Kagoshima earthquake (M6.5), southwestern Japan, by seismic waveform inversion and double-difference tomography

A large inland earthquake (1997 northwestern Kagoshima earthquake) occurred in the northwestern part of Kagoshima Prefecture, southwestern Japan, on March 26, 1997. We estimated the slip distribution of the mainshock rupture by a seismic waveform inversion method of Yagi et al. [2004]. The data used for the inversion were 7 teleseismic P-wave seismograms from 7 stations of the IRIS-DMC and 18 strong motion seismograms from 6 stations of the K-NET, NIED, Japan. We adopted the hypocentral parameters of (Lat., Lon., Depth)=(31.973N, 130.388E, 6.8km), the fault parameters of (Strike, Dip)=(280°, 90°), and the fault area of 15km x 15km divided into 15 x 10 subfaults. Obtained coseismic slip area spreads mainly through the west of the mainshock hypocenter. Spatial extent of the large coseismic slip area is about 3km x 3km. We also obtained detailed 3D seismic velocity structure in and around the fault plane of the mainshock using travel time data from a temporary seismic network, which was deployed for about three months just after the mainshock [Miyamachi et al., 1999]. We used double-difference tomography method of Zhang and Thurber [2003]. The obtained velocity structure shows that the mainshock hypocenter is located near the boundary between the low and high velocity portions along the fault plane. Comparison with the slip distribution shows that the asperity (large coseismic slip area) of the mainshock seems to correspond with a high velocity portion along the fault plane. This observation is consistent with the correspondence between high-velocity bodies and asperities for other earthquakes: the 2003 northern Miyagi earthquake (M6.4) in NE Japan, the 2000 western Tottori earthquake (M7.3) in SW Japan, the 1995 southern Hyogo (Kobe) earthquake, the 2001 Geiyo intraslab earthquake in SW Japan, and the 2004 Mid Niigata earthquake (M6.8) in Central Japan.