Source parameter estimation is an important aspect of full waveform inversion (FWI), both as an input and as a QC tool for velocity inversion. In active-source FWI, we often estimate a single source parameter for many available sources, by assuming that all sources excite the same source signature. Alternatively, we can estimate an individual signature per source, and evaluate the validity of the velocity distribution based on the coherency of the obtained source signatures. To invert for source parameters during conventional acoustic FWI, we assume an isotropic source mechanism, and focus on estimating source wavelet. When extending FWI to an elastic medium, we need to estimate both source mechanism and wavelet. In this talk, I will review and discuss source estimation methods for elastic FWI, and subsequently the trade-off between source parameters and velocity structures.
Back to Workshop II: Full Waveform Inversion and Velocity Analysis