HR: 1340h
AN: S43B-1314    [Abstracts]
TI: Time-Reversal Location of the 2004 M6.0 Parkfield Earthquake Using the Vertical Component of Seismic Data.
AU: * Larmat, C S
EM: carene@lanl.gov
AF: Geophysics Group EES-11 Los Alamos National Laboratory of the University of California, MS D443, Los Alamos, NM 87545, United States
AU: Johnson, P
EM: paj@lanl.gov
AF: Geophysics Group EES-11 Los Alamos National Laboratory of the University of California, MS D443, Los Alamos, NM 87545, United States
AU: Huang, L
EM: ljh@lanl.gov
AF: Geophysics Group EES-11 Los Alamos National Laboratory of the University of California, MS D443, Los Alamos, NM 87545, United States
AU: Randall, G
EM: grandall@lanl.gov
AF: Geophysics Group EES-11 Los Alamos National Laboratory of the University of California, MS D443, Los Alamos, NM 87545, United States
AU: Patton, H
EM: patton@lanl.gov
AF: Geophysics Group EES-11 Los Alamos National Laboratory of the University of California, MS D443, Los Alamos, NM 87545, United States
AU: Montagner, J
EM: jpm@ipgp.jussieu.fr
AF: Institut de Physique du Globe de Paris, Case 89 4, place Jussieu, Paris, 75252-05, France
AB: In this work we describe Time Reversal experiments applying seismic waves recorded from the 2004 M6.0 Parkfield Earthquake. The reverse seismic wavefield is created by time-reversing recorded seismograms and then injecting them from the seismograph locations into a whole entire Earth velocity model. The concept is identical to acoustic Time-Reversal Mirror laboratory experiments except the seismic data are numerically backpropagated through a velocity model (Fink, 1996; Ulrich et al, 2007). Data are backpropagated using the finite element code SPECFEM3D (Komatitsch et al, 2002), employing the velocity model s20rts (Ritsema et al, 2000). In this paper, we backpropagate only the vertical component of seismic data from about 100 broadband surface stations located worldwide (FDSN), using the period band of 23-120s. We use those only waveforms that are highly correlated with forward-propagated synthetics. The focusing quality depends upon the type of waves back- propagated; for the vertical displacement component the possible types include body waves, Rayleigh waves, or their combination. We show that Rayleigh waves, both real and artifact, dominate the reverse movie in all cases. They are created during rebroadcast of the time reverse signals, including body wave phases, because we use point-like-force sources for injection. The artifact waves, termed "ghosts" manifest as surface waves, do not correspond to real wave phases during the forward propagation. The surface ghost waves can significantly blur the focusing at the source. We find that the ghosts cannot be easily eliminated in the manner described by Tsogka&Papanicolaou (2002). It is necessary to understand how they are created in order to remove them during TRM studies, particularly when using only the body waves. For this moderate magnitude of earthquake we demonstrate the robustness of the TRM as an alternative location method despite the restriction to vertical component phases. One advantage of TRM location is that it does not rely on a prior picking of specific phases (Larmat et al, 2006). In future work will be conducted TRM backpropagation using the horizontal displacement components of seismic data as well as study the source complexity (double couples). Our ultimate goal is to determine whether or not Time Reversal offers information about the source that cannot be obtained from other methods, or that complements other methods.
DE: 7215 Earthquake source observations (1240)
DE: 7255 Surface waves and free oscillations
DE: 7290 Computational seismology
DE: 7299 General or miscellaneous
SC: Seismology [S]
MN: 2007 Fall Meeting