Abstract
Passive seismic tomography has been proved to be a great alternative solution for areas which is highly challenging to be explored by active seismic methods. As a modern high resolution tomographic algorithm, we consider Full Waveform Inversion (FWI) as a potential passive seismic method for Dehdasht embayment, South West, Iran. However, before the implementation of FWI process on the real dataset of Dehdasht area, we test our proposed multi-scale workflow on a synthetic model resembling our area of interest. Using this workflow, different factors affecting the FWI results such as source and receiver configurations, presence of noise in data and erroneous source location were evaluated. Fortunately, with a receiver spacing similar to Dehdasht acquisition network, using enough number of sources and especially in Vs model, it is possible to reconstruct such high-resolution features of the medium as hydrocarbon trap, high-velocity bedrock and shallow syncline structure. Moreover, despite a high level of random noise added to the data, the results demonstrate that our proposed workflow is robust enough to diminish the adverse effect of noise and present a clear subsurface tomographic image of the medium. However, simulations with erroneous source locations warns about the divergence and instability of FWI process caused by the sources with high location errors. As a result, before the implementation of FWI on the real dataset of Dehdasht project, we should dedicate a specific attention on source relocation or source modeling process in general. At the end, we propose a hierarchical processing workflow for the real dataset of Dehdasht to guarantee the convergence to a high-resolution tomographic model of the Dehdasht embayment area.