HR: 1340h
AN: NG43B-0573 [Abstracts]
TI: Numerical Born kernels to account for finite-frequency effects in surface wave tomography
AU: * Peter, D B
EM: dpeter@erdw.ethz.ch
AF: Institute of Geophysics ETH, Schafmattstr. 30, Zurich, 8093
Switzerland
AU: Boschi, L
EM: lapo@erdw.ethz.ch
AF: Institute of Geophysics ETH, Schafmattstr. 30, Zurich, 8093
Switzerland
AB:
In order to conduct seismic tomography, one should
use the most adequate description of wave propagation; thus
far, almost all tomographic models rely on ray theory, a high-frequency
approximation. Particularly for surface waves, Born
theory, a single scattering theory, improves it by taking first
order scattering effects into account. The resulting sensitivity
kernels are typically derived from analytical far-field Green's
tensors, which lead to singularities at source and receiver
locations; numerical derivation of kernels in contrast can
avoid this, but full numerical integration of the equations of
motion in 3D is computationally too expensive when implemented on a realistic 3-D Earth.
If we restrict ourselves to the case of a smooth, laterally heterogeneous Earth,
we can use a 0-thickness membrane as an analogue for
Love and Rayleigh wave propagation, speeding up dramatically the numerical
integration. Here, following earlier work by Tanimoto (1990) and Tape (2003),
we implement ``membrane waves'' with
a new finite-difference algorithm that accounts for
finite-frequency effects in the smooth Earth approximation; thus leading to
the numerical derivation of Born sensitivity kernels for surface
wave phase velocity tomography.
DE: 7255 Surface waves and free oscillations
DE: 7270 Tomography (6982, 8180)
DE: 7290 Computational seismology
SC: Nonlinear Geophysics [NG]
MN: Fall Meeting 2005