HR: 0800h
AN: U11A-0817 [Abstracts]
TI: Rupture process of the December 26th 2004 great Sumatra-Andaman earthquake inferred from the far field
body waves
AU: * Wu, C
EM: wuchang@eri.u-tokyo.ac.jp
AF: Earthquake Research Institute, University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo, Japan, Japan,
1130032
Japan
AU: Koketsu, K
EM: koketsu@eri.u-tokyo.ac.jp
AF: Earthquake Research Institute, University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo, Japan, Japan,
1130032
Japan
AB:
The December 26th 2004 great Sumatra-Andaman earthquake occurred on the interface between the India plate and the Burma
plate. This is the fourth largest earthquake in the world since 1900. The heavy damage and casualties caused by the tsunami
of this earthquake shocked the whole world. Simulation of tsunami propagation requires a detailed rupture process. Several
studies on the source rupture process using the far field body waves are available now. On the basis of the aftershock
distribution, the rupture area probably extended more than 1000km. Assuming a physically reasonable rupture propagation
velocity of 2.5km/s, we expect the rupture duration as long as 400 seconds.
Since the P-wave train from this earthquake includes sveral phases such as P, PP and PcP, we calculated the Green's functions
for rupture process inversion considering the P, PP (including PPP, PPPP and so on) and PcP phases. The previous studies
(e.g. Fukao et al., 2003) showed that the observed travel time difference (Tpp-Tp) varies from region to region because of
the anomaly of P-wave velicity in the upper and middle mantle. In this study, we select four earthquakes before the main
shock in this area to correct the calculated PP travel time. Since the PP wave shows a 90-degree phase shift compared with
the P phase, the observed PP arrival times are hence determined by searching the maximum correlation of the Hilbert
transformation of P phase with the observed PP phase.
Following the trench trace in this area, we divide the fault plane into three segments. The southernmost 350 km segment has a
strike direction of 320 degree; the middle segment (630 km long) rotates 20.5 degree clockwise, whereas the northernmost
segment (420 km long) further rotates 30 degree clockwise. This fault plane is continuous and the dip angle varies from the
south to the north. We then distribute the 120 grids on the fault plane. The grid spacing in the strike direction is 70km and
about 40km in the dip direction. Rupture velocity 2.5km/s is assumed. Note that this value does not control the real
rupture front but the front of the first source time window.
Our inversion results show that introduction of the PP and PcP phases into the synthetic seismogram improves the waveform
fittings. In about 550 seconds, the rupture propagated about 1300km from the hypocenter in the southern part to the northern
end. The slip distribution shows heterogeneity of rupture pattern. The released total moment leads to a moment magnitude Mw
of 9.1. The maximum slip is about 8.5 meter. Two asperities can be identified in this main shock. The first one is confined
in the southern part around the hypocenter, and the other in the northern fault.
DE: 7215 Earthquake source observations (1240)
SC: Union [U]
MN: Fall Meeting 2005