3D frequency-domain acoustic full waveform inversion

Jinji Li, Kristopher A. Innanen

Full waveform inversion (FWI) has established itself as a crucial technique in earth science, providing high accuracy and resolution in geophysical imaging. The increasing demand for 3D FWI stems from its ability to capture complex subsurface structures and azimuthal features by utilizing the complete seismic wavefield information. The frequency-domain approach to FWI offers advantages by modeling and matching data at specific frequencies, which helps to address the nonlinearity of FWI and manage frequency-dependent parameters more effectively. Despite these benefits, challenges remain in the 3D frequency-domain FWI, including the complexities of 3D finite-difference modeling, inversion term calculations, and high computational demands. In this research, we focus on forward modeling, inversion strategies, and specific techniques to address inherent challenges in the 3D frequency-domain FWI. Through synthetic examples, we demonstrate the efficacy of this method in reconstructing detailed 3D acoustic velocity models and analyze the associated computational requirements in both spatial and temporal domains. Our findings highlight the significant potential of 3D frequency-domain FWI in advancing geophysical research and its applications in subsurface exploration.