2001 Geiyo earthquake Last updated 03/Apr/2001

Rupture Process of the 2001 Geiyo Earthquake obtained from Strong Motion Data [Ver.1]
Sekiguchi, H.* and T. Iwata**

* Geological Survey of Japan
** DPRI, Kyoto Univ.

We conducted kinematic source inversions to obtain rupture process of the 2001 Geiyo earthquake (MJ6.4), occurred at 15:27JST, 24th March, using strong motion network data.

We assumed two-segment fault plane* from the aftershock distribution. The north plane was assumed to be N172E Strike direction of western dipping plane with angle of 57deg. For the south plane, Strike of N180E and dip angle of 60deg, respectively. Each fault length and width are 15km and 21km. Subfault-size is 3x3km in square. Totally, we use 10 (in length) x 7 (in width) numbers of subfaults. We used fourteen K-NET, KiK-net, and Freesia (strong motion) station data with uniform coverage. In
Fig. 1, map of station distribution, the assumed fault plane, and aftershock distribution determined by JMA for two days are shown. Rupture starting point is fixed to 34.120N, 132.709E, and 51.3km from JMA.

Target velocity waveforms were integrated from acceleration with BPF of 0.1 to 1.0Hz. (cf. Originally velocity waveforms were used for TMG, Freesia network data) for getting not only high-resolution source model but also source characterization for strong motion generation. We used surface station data for KiK-net. For calculating Green's function, we refered velocity structure that Shiraki Seismoc Observatory of ERI, Univ. of Tokoy, use for hypocener determination. Target waveforms are 16second window length of the S-wave portion from 1sec before S-wave onset. Multi-window linear waveform ingersion method (Hartzell and Heaton, 1983) with spatio-temporal constraint (Sekiguchi et al., 2001) was used for the analysis. We also took into account the moving dislocation effects in subfaults with a convolution technique (Sekiguchi, 1999; Sekiguchi et al., 2001). Slip history on each subfault is represented by six element source time functions with 1s duration succesively triggered at 0.5s interval. Rake angle constraint was between -25 and -110deg (-65+-45deg).

(*: We conducted one fault plane model with 172/60. Two-segment fault model gives slightly better fitting than the one fault plane model.)

In Fig. 2, final slip distribution is shown. Rupture velocity of the first time window was selected to be 2.8km/s with radial propagation. Slips occurred mainly in the deeper part. two asperities are observed at each segment. The asperity in the southern segment is larger than that in the northern segment. Maximum slip is about 3m. In Fig. 3, time progression of rupture is shown. Comparison between observed (Black) and synthetic (Red) waveforms are shown in Fig. 4. Numbers indicated each wawveforms are maximum velocity in m/s. Synthetics agree fairly well the observations. Pulse waves in NS component, which appeared at about 10s later from the S-wave onset, at the stations in the north direction, such as HRSH08, HRSH07, HRS012, HRSH01, and HRSH02, can not reproduce well by this source model. more complex rupture history could be suggested. (This is the reason of Ver.1.) Total seismic moment release is 3.3X10**19N.m or Mw =6.9. Still we need to tune the velocity model for Green function, too.


We use K-NET, KiK-net, and Freesia strong motion data.

If you have questions, please contact to Tomotaka Iwata (iwata@egmdpri01.dpri.kyoto-u.ac.jp).

goto English Laboratory page