HR: 1330h
AN: S12A-0367    [PDF]
TI: Ground Motion Characteristics and Source Process of the 2002 Denali Earthquake Inferred from the Strong Motion Records
AU: * Asano, K
EM: k-asano@egmdpri01.dpri.kyoto-u.ac.jp
AF: Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji, Kyoto, 611-0011 Japan
AU: Iwata, T
EM: iwata@egmdpri01.dpri.kyoto-u.ac.jp
AF: Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji, Kyoto, 611-0011 Japan
AU: Irikura, K
EM: irikura@egmdpri01.dpri.kyoto-u.ac.jp
AF: Disaster Prevention Research Institute, Kyoto University, Gokasho, Uji, Kyoto, 611-0011 Japan
AB: The 2002 Denali earthquake, which occurred at the Denali Fault System in Alaska on 3 November, 2003, was one of the largest inland earthquakes all over the world. The Denali Fault System extends for more than 2000 km across south-central Alaska, Yukon Territory, northern British Columbia and southeastern Alaska (Lanphere, 1978). In this time the rupture started on the Susitna Glacier Fault with thrust motion, propagated eastward along the Denali Fault with right-lateral movement, and terminated on the Totschunda Fault (Fuis and Wald, 2003). From the particle motion at a strong motion site PS10 (pump station \#10 of Trans Alaska Pipeline), which is located at about 3 km distance from the fault, we could recognize a significant phase with fault-parallel movement before fault-normal motions. This motion shows the fault displacement in front of PS10. We have also checked the orientation of seismometer at PS10 with visiting the place in this summer. Therefore, the sense of ground motion at PS10 is reliable. Comparing the observed and calculated travel time of this phase, the average rupture propagation velocity between the rupture starting point and a sub-fault in front of PS10 was estimated to be 2700 m/s. However, it does not deny the possibility of supershear rupture on a certain portion of source area, so that we need to examine carefully the rupture velocity during the rupture. Whole source process was investigated with the multi-time window kinematic waveform inversion (Hartzell and Heaton, 1983; Sekiguchi et al., 2000). A 1-D laterally homogeneous underground structure model was assumed based on the result of refraction and wide-angle reflection survey by Beaudoin et al. (1992). Green's functions were calculated using the discrete wavenumber method (Bouchon, 1981) together with the reflection transmission matrix method (Kennett and Kerry, 1979). Spatio-time smoothing and slip constraints as pure-dip to right-lateral slip for Susitna Glacier fault, right-lateral $\pm 45 \deg$ slip for other faults, were included (Sekiguchi et al., 2000). The appropriate strength of smoothing constraint was evaluated using Akaike's Bayesian Information Criterion (Akaike, 1980). Three components of velocity waveforms obtained at 10 strong motion stations including PS10 were inverted. We tried to include the detailed information of the surface displacement distributions compiled by Eberhart-Phillips et al. (2003) as constraint information in our analysis to make up the insufficiency of strong motion data. We could reproduce observed ground motion in low-frequency range (0.1-0.5 Hz for Alyeska stations, 0.05-0.5 Hz for others) based on the source model obtained by the inversion analysis. The inversion result showed that the ground motion at PS10 was mainly controlled by the nearest asperity. Large slips on the fault from the inversion result were observed at about 90 km east and about 170 km east from the epicenter. We are grateful to United States Geological Survey, University of Alaska, and Alyeska Pipeline Service Company for releasing digital strong motion records and to Alaska Earthquake Information Center for which we retrieved hypocentral information.
DE: 7209 Earthquake dynamics and mechanics
DE: 7212 Earthquake ground motions and engineering
DE: 9350 North America
SC: Seismology [S]
MN: 2003 Fall Meeting