Application of Full Waveform Inversion in passive seismic explorations: Some practical considerations

نویسندگان

1 International Institute of Earthquake Engineering and Seismology (IIEES), Tehran, Iran

2 National Iranian Oil Company (NIOC), Exploration Directorate, Tehran, Iran

3 International Institute of Earthquake Engineering and Seismology (IIEES), Tehran, Iran

doi
10.22059/geope.2025.396562.648828
چکیده

Passive seismic tomography has proven to be a valuable alternative in areas where active seismic methods face significant challenges. In this study, we investigate the application of Full Waveform Inversion (FWI) – a modern, high-resolution tomographic technique – as a potential passive seismic approach for the Dehdasht embayment, Southwest, Iran. Prior to applying FWI to the real dataset from Dehdasht, we test a proposed multi-scale workflow on a synthetic model that closely resembles the study area. Through this workflow, we examine key factors affecting the FWI performance, including source and receiver configurations, presence of noise in data and inaccuracies in source location. Encouragingly, with a receiver spacing comparable to Dehdasht acquisition network and a sufficient number of sources – particularly in Vs model– we are able to reconstruct high-resolution subsurface features such as hydrocarbon trap, high-velocity bedrock and shallow syncline structure. Even with the addition of substantial level of random noise, the results demonstrate that our proposed workflow is robust and capable of producing clear subsurface tomographic image. However, simulationsincorporating erroneous source locations reveal that significant source misplacement can lead to divergence and instability in FWI process. As a result, before applying FWI to the real dataset of Dehdasht, special attention must be given to source relocation or accurate source modeling. At the end, we propose a hierarchical processing workflow to ensure the convergence toward a reliable high-resolution tomographic model of the Dehdasht embayment.