Time domain FWI imaging in acoustic variable-density media

Ziguang Su, Daniel O. Trad

Amplitude variation with angle (AVA) response is crucial in seismic processing and interpretation, representing angle-dependent reflectivity. They require true-amplitude migration algorithms and accurate model parameters, such as velocity. Advances in high-performance computing and Full-Waveform Inversion (FWI) algorithms have made FWI effective in generating precise velocity models and other parameters across various geological settings. FWI is now used not only for velocity updates but also as a direct imaging tool, achieving both refined velocity models and high-quality seismic reflectivity images. This process, known as 'FWI imaging', has gained attention for its potential to deliver high-quality imaging and simplify workflows. FWI imaging modifies the FWI workflow to output subsurface images or reflectivity directly, bypassing the time-consuming seismic imaging process. By providing reflectivity alongside the velocity model, FWI offers a more detailed and accurate subsurface representation, aiding better decision-making and interpretation. FWI imaging surpasses conventional imaging methods by delivering seismic images with improved illumination, signal-to-noise ratio, focusing, and resolution. However, the FWI sensitivity kernel in acoustic media with density differs by parameterization method, and the time-domain FWI gradient equation in velocity-density media has multiple arguments. In this article, we tested all the published velocity-density gradient equations in time-domain FWI in acoustic variable-density media and derived a new FWI sensitivity kernel from scattering theory.