To obtain subsurface information from onshore seismic exploration data using full waveform inversion (FWI) based on the acoustic wave equation, elastic waves, such as ground rolls and mode-converted waves, should be suppressed through heavy preprocess...
To obtain subsurface information from onshore seismic exploration data using full waveform inversion (FWI) based on the acoustic wave equation, elastic waves, such as ground rolls and mode-converted waves, should be suppressed through heavy preprocessing. However, the preprocessing deforms not only the elastic waves but also the acoustic waves. Moreover, it is not easy to separate body waves and surface waves in seismic traces. For these reasons, in the modeling step, we need to generate both types of waves to obtain more similar seismic waves to the real seismic waves. Therefore, elastic full waveform inversion using elastic wave equation is necessary for more accurate full waveform inversion. In addition, elastic full waveform inversion can give better geological information than acoustic full waveform inversion because it inverts P-wave velocity, S-wave velocity and density. Laplace-Fourier domain FWI using time-domain modeling combines time-domain wave propagation modeling and Laplace-Fourier-domain FWI. To obtain forward wavefield and adjoint wavefield in the time domain, we implemented staggered grid finite difference method. The residuals between the recorded and modeled data, virtual sources, hessian matrices and gradient directions were calculated in the Laplace-Fourier domain. We used time domain wave propagation modeling for the forward and adjoint wavefield because it is more intuitive to treat the wavefield in the time domain than in the Laplace-Fourier domain. Moreover, time domain wave propagation modeling using staggered grid finite difference method does not need matrix solver which is necessary for the conventional Laplace-Fourier domain FWI. The optimization procedure is conducted in the Laplace-Fourier domain because Laplace-Fourier-domain FWI can be applied to real seismic data, which lacks low-frequency components. To validate our proposed algorithm, we performed numerical tests with synthetic data and real exploration data. We applied the algorithm to Model 94 synthetic onshore data and Benjamin Creek real onshore data.