Refined Local and Regional Seismic Velocity and Attenuation Models from Finite-Frequency Waveforms; Conference paper

摘要

In seismic tomography, the reference or starting models are commonly one-dimensional (1D) and the structural sensitivity kernels of seismic data are calculated without considering the finiteness of seismic waves in both time and frequency domains. These simplifications result in a theoretical limit, in addition to that from the data coverage, on the structural resolution in tomography inversions. In small-scale (such as the crustal) structural imaging, however, a higher resolution is desired, and the ID reference model is no longer adequate; therefore, accurate modeling of wave propagation in three- dimensional (3D) reference models is necessary. To address this problem, elaborate numerical methods, such as the finite-difference and spectral-element methods, have been adopted to calculate the full-wave, finite-frequency, banana- doughnut, structural sensitivity kernels of seismic data. The objective of this work is to develop refitted local and regional velocity and attenuation models for selected areas of interest (AOIs) in Eurasia. We have carried out systematic tests to validate the finite-frequency sensitivity kernels computed by a staggered-grid, finite-difference method. These teats result in corrections in the calculation of the structural sensitivity kernels as well as an important finding that anomalies in S-wave speed have a significant contribution to P-wave travel time perturbations (Zhang and Shea, 2008). Thus, current seismic tomography practices, in which P-wave travel times ate assumed to be unrelated to S-wave velocity anomalies, may lead to systematic biases in inversion results. Furthermore, the different components of the same arrival at the same receiver have different travel time and amplitude sensitivities to variations in the velocity Structure.