Importance of the Source Estimation in FWI - Sensitivity and Examples

It has been recognised that the source estimation can play a crucial role in full waveform inversion (FWI). Often the source signature is estimated inside the iteration loop in FWI and is updated with iterations. Due to the trade-off between source signature and subsurface parameters this approach can result in wrong source signature and subsurface model. Exploiting the global source signature might introduce systematic errors in FWI results. We explore the potential of separating the source estimation from the FWI loop and suggest the source estimation for each shot in the dataset. To reduce the subsurface component in the source signature, we incorporate high frequency data in our approach. On various real data examples, we show the advantages of the proposed approach. Using the synthetic data examples, we access sensitivity of the FWI results to the errors in the source signature.