HR: 0800h
AN: U11A-0812    [Abstracts]
TI: Inversion of long period seismic data for the source process of the great Sumtara-Andaman earthquake
AU: * Rhie, J
EM: rhie@seismo.berkeley.edu
AF: Berkeley Seismological Laboratory, 215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760 United States
AU: Romanowicz, B
EM: barbara@seismo.berkeley.edu
AF: Berkeley Seismological Laboratory, 215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760 United States
AU: Dreger, D
EM: dreger@seismo.berkeley.edu
AF: Berkeley Seismological Laboratory, 215 McCone Hall, UC Berkeley, Berkeley, CA 94720-4760 United States
AB: The December 26, 2004 Great Sumatra-Andaman, earthquake opened a new era for the seismologists to understand the complex source process of a great earthquake. This is the first great event since the deployment of high dynamic range broadband seismic and GPS sensors on the globe. For the first time, seismologists have a chance to look at the very complex source process by using various seismic and geodetic data sets with different frequency contents. While several works on the source process have been already reported during the last few months, some important parameters for the fault geometry and slip distribution, such as dip angle of the fault plane and evolution of rupture velocities over the whole fault plane, may not yet be well constrained. The purpose of this study is to derive more complete fault plane and slip distribution models, constrained from long period seismic data sets by using independent inversion methods. To do that, we start from the two different fault geometries inferred from GPS measurements (Barnerjee et al., 2005) and multiple Harvard CMT solution (Tsai et al., 2005) and invert the global long period (100-500s) seismic measurements for the slip distribution over the fault planes. In addition, line source models with several segments are tested for the variation in rupture velocity by fitting very long period seismic waveforms and normal mode spectra. Finally, we combine models derived from two different methods to get an optimal model which can explain the data sets. We perform sensitivity tests for the perturbation to the given fault plane geometries including the number of fault segments and the variation in rupture velocity. We also compute error estimations for several key models to show the region where the slip distribution is not well constrained. Final models are compared to starting models and their agreement and potential disagreement will be discussed. More quantitatively constrained source parameters may help us to understand more about the complex source process of this great event, as well as the devastating tsunami it generated.
DE: 4564 Tsunamis and storm surges
DE: 7200 SEISMOLOGY
DE: 7215 Earthquake source observations (1240)
DE: 7240 Subduction zones (1207, 1219, 1240)
DE: 7255 Surface waves and free oscillations
SC: Union [U]
MN: Fall Meeting 2005