HR: 1330h
AN: S42E-0218    [PDF]
TI: Rupture Velocity of the 2001 Kunlun, China, Event Estimated From SEM Waveform Modeling
AU: * Hj\"{o}rleifsd\'{o}ttir, V
EM: vala@gps.caltech.edu
AF: California Institute of Technology, 1200 E. California Blvd, Pasadena, CA 91125 United States
AU: Kanamori, H
EM: hiroo@gps.caltech.edu
AF: California Institute of Technology, 1200 E. California Blvd, Pasadena, CA 91125 United States
AU: Tromp, J
EM: jtromp@gps.caltech.edu
AF: California Institute of Technology, 1200 E. California Blvd, Pasadena, CA 91125 United States
AB: The rupture speed during an earthquake is controlled by the energy dissipated mechanically during faulting. Rupture velocities close to the limiting speed imply that very little energy is going into the fracture and thus has implications for whether the rupture will continue to grow or stop. The M$_w$~7.9, November~14, 2001, Kunlun, China, strike-slip event ruptured unilaterally over 400 km along the Kunlun fault. The long rupture combined with worldwide broad-band instrumentation provides us with a unique opportunity to estimate the rupture speed. Unfortunately, the strike-slip nature of the faulting causes the body waves recorded at teleseismic distances to be nearly nodal, making body-wave modeling very difficult. We use the spectral-element method (SEM) and a 3D Earth model to accurately compute waveforms at periods of 18 seconds and longer for this event. We compute synthetics for two source models. We use (a) a body-wave model with an average rupture speed of 3.5 km/s, similar to the model reported by Kikuchi and Yamanaka (2001) and (b) a surface-wave model proposed by Bouchon and Vall\'{e}e (2003) with an average rupture speed of 4.3 km/s and a slip distribution constrained by measured surface offsets. Both models show similar fits to the surface-wave radiation pattern at periods above 100 seconds and thus we turn to long-period body waves to discriminate between the two. The event comprised several smaller subevents, with the largest subevent occurring about 60 seconds after the initiation of rupture. The azimuthal variation in arrival time of this phase depends on the rupture velocity. Using the Bouchon and Vall\'{e}e model we find that shear waves from the large subevent arrive about 10 seconds earlier than observed in the direction of rupture, whereas they arrive on time in the anti-rupture direction. A similar phase shift is observed in the short period surface waves. As this azimuthal variation is not evident for the Kikuchi and Yamanaka model, we conclude that the rupture velocity is too high in the Bouchon and Vall\'{e}e model. Note that by using the 3D SEM synthetics we account for the effect of 3D structure on the arrival time of the shear-waves. A lower bound of the rupture speed is being investigated in a similar manner.
DE: 7209 Earthquake dynamics and mechanics
DE: 7215 Earthquake parameters
DE: 7255 Surface waves and free oscillations
SC: Seismology [S]
MN: 2003 Fall Meeting