| Abstract Submission No. | ABS-4.1-0146 |
| Title of Abstract | Application of Full Waveform Inversion (FWI) for imaging synthetic seismic data |
| Authors | Machitti Pavani*, Mrinal Kanti Sen, Prakash Kumar, Biswajit Mandal, Bijayananda Dalai |
| Organisation | CSIR-National Geophysical Research Institute |
| Address | CSIR-NGRI, Uppal road, Hubsiguda, Hyderabad. Hyderabad, Telangana, India Pincode: 500007 E-mail: pavanigeo23@gmail.com |
| Country | India |
| Presentation | Oral |
| Abstract | Full Waveform Inversion (FWI) is a high-resolution seismic imaging technique that goes beyond conventional approaches by exploiting the entire seismic waveform rather than relying solely on kinematic information. Conventional methods, such as travel-time tomography and migration, are based on ray theory approximations and often provide insufficient resolution for imaging complex subsurface structures. In contrast, FWI incorporates the full wavefield, enabling more accurate characterization even in geologically challenging environments.In this study, we applied FWI to suite of synthetic seismic data to investigate both shallow and deep subsurface features. Forward modelling generated shot gathers that served as input for the inversion, which was carried out under an elastic approximation. The inversion workflow employed an iterative local optimization scheme, where synthetic data were updated progressively to minimize the misfit with observed data. A multiscale strategy was adopted, beginning with low-frequency components to mitigate cycle skipping, followed by higher frequencies to sharpen structural details.Our results highlight the computational demands of the process, requiring high-performance parallel computing resources. Nonetheless, the synthetic case studies demonstrate that FWI effectively bridges the resolution gap between conventional tomography and reflection imaging, offering a unified framework for detailed subsurface characterization. |
| Keywords | Full Waveform Inversion, Seismic Imaging, Forward modelling, Synthetic seismic data, Cycle Skipping, Parallel computation. |