HR: 09:15h
AN: S31E-06    [Abstracts]
TI: Global P-wave Tomography Using Finite-Frequency Modeling and Finite-Frequency Data
AU: * Sigloch, K
EM: sigloch@princeton.edu
AF: Princeton University, Geosciences Department, 312 Guyot Hall, Princeton, NJ 08544, United States
AU: Nolet, G
EM: nolet@princeton.edu
AF: Princeton University, Geosciences Department, 312 Guyot Hall, Princeton, NJ 08544, United States
AB: We use our matched-filtering measurements of finite-frequency, teleseismic body-wave traveltimes and amplitudes as input data for global-scale "banana-doughnut" tomography. We present a new mantle model of V_p anomalies obtained from joint inversion of P-wave traveltimes and amplitudes. A Q_s model is also obtained since amplitudes are inverted simultaneously for focusing and attenuation. In accordance with their definition in the Born approximation, traveltimes and amplitudes were measured by cross-correlating observed and predicted seismograms. Observables are consistent across frequency bands since for any source-receiver path, finite-frequency measurements are derived by filtering a single broadband seismogram into distinct sub-bands. Our bandpass filters have central frequencies between 0.03 and 1~Hz. In this range, predicted waveforms depend heavily on accurate estimates of source time functions, a challenge explicitly addressed by our method. Frechet kernels are computed using the exact frequency responses of the bandpass filters used. Our data set comprises teleseismic P-wave data from 1999-2007, available from IRIS DMC. We have processed all events for which m_b ≥ 6.0, and most events for which 5.7 ≤ m_b < 6.0, for a total of >1,500 earthquakes. We have thus far concentrated on the inversion of teleseismic P-wave data observed in North America. The obtained models predict traveltime dispersion due to diffraction effects on the order of 0.1-0.5 sec, consistent with what we observe in the data. Amplitude dispersion is on the order of 20%. Our models explain roughly half of the variance associated with the amplitude data. For North America, we observe a close correspondence of upper mantle V_p patterns with known surface tectonics. Thanks to the excellent USArray data, features on the order of 200~km can be resolved in large parts of the upper mantle under the Western U.S. Deeper down and further east, the presumed fragments of the subducting Farallon plate are clearly visible. We expect to include the full data set with global source-receiver coverage and to present preliminary results for a global inversion in addition to the North American tomography.
DE: 7203 Body waves
DE: 7270 Tomography (6982, 8180)
DE: 7290 Computational seismology
SC: Seismology [S]
MN: 2007 Fall Meeting