Abstract

In this study, we implement an elastic 2D Full Waveform Inversion (FWI) workflow in the time domain to obtain a high-resolution subsurface velocity-depth model. The FWI workflow uses the entire waveform data and requires an initial velocity model that is refined iteratively in a least-squares sense, and the misfit between the modelled and the observed seismic data is minimized to get an inverted subsurface velocitydepth model. We implement the FWI methodology first on a synthetic dataset that mimics a subduction zone environment with a marine acquisition setup. Implementation of the FWI workflow is a computationally intensive process; therefore, we utilize the PARAM SANGANAK facility at IITK, accessed under the NSM, Government of India, to reduce the time involved in forward modeling and subsequent inversion, thereby resulting in a higher-resolution subsurface velocity model. The final inverted velocity model using a multiscale FWI methodology for the synthetic dataset matches well with the true subsurface velocity model, validating the potential of FWI in resolving finescale features beneath the surface. We further implement the FWI methodology to the multicomponent Ocean Bottom Seismometer (OBS) datasets across the Andaman Sumatra subduction zone, where the survey profile line CD (the WG2 profile) crosses the entire subduction setup, and present the initial results.