| dc.description.abstract | Full Waveform Inversion (FWI) is a seismic imaging technique that reconstructs high-resolution subsurface velocity models by minimizing the difference between observed and modelled seismic waveforms. This study applies FWI to both synthetic and field 2D seismic datasets using the open-source Seismic Waveform Inversion Toolbox (SWIT), modified for compatibility with field data formats and acquisition geometries. Synthetic tests using the Marmousi 2 model evaluate different inversion schemes, misfit functions, and optimization methods, highlighting the effectiveness of a multiscale strategy that begins with traveltime misfit and transitions to waveform misfit. The L-BFGS optimization method shows faster convergence in noise-free conditions, while the NLCG method demonstrates greater robustness to noise. For field data, the inversion workflow was adapted to handle moving-spread acquisition and limited frequency content. Although the results were less optimal due to time constraints and parameter limitations, the study identifies several critical factors affecting inversion performance, including preprocessing strategy, grid configuration, inversion parameter tuning, and wavelet estimation. The findings confirm that FWI, when implemented with appropriate preprocessing, parameter control, and multiscale strategies, can improve subsurface imaging resolution and provide a flexible framework adaptable to various seismic datasets. Recommendations are made for enhancing field-data performance through more extensive parameter testing, alternative misfit functions, and synthetic robustness trials. | en_US |
