An efficient multiscale method for time-domain waveform tomography

摘要

This efficient multiscale method for time-domain waveformtomography incorporates filters that are more efficientthan Hamming-window filters.Astrategy for choosing optimalfrequency bands is proposed to achieve computationalefficiency in the time domain. A staggered-grid, explicit finite-difference method with fourth-order accuracy in spaceand second-order accuracy in time is used for forward modelingand the adjoint calculation. The adjoint method is utilizedin inverting for an efficient computation of the gradient directions.In the multiscale approach, multifrequency data andmultiple grid sizes are used to overcome somewhat the severelocal minima problem of waveform tomography. The methodis applied successfully to 1D and 2D heterogeneous models;it can accurately recover low- and high-wavenumber componentsof the velocity models. The inversion result for the 2Dmodel demonstrates that the multiscale method is computationallyefficient and converges faster than a conventional,single-scale method.