Answers are grounded in this paper and cite their sources.
SUMMARY Tomography based on full waveforms is an important tool for characterizing the subsurface. However, systemic artefacts in the recorded data must be removed prior to the imaging process to fully utilize the information contained in the data. Especially for near-surface surveys, the coupling between sources/receivers and the medium can introduce significant distortions in the recorded data. We present two novel time-domain full waveform inversion approaches that account for the interaction between sources, receivers and the subsurface. The first approach does not impose any restrictions on the shape of the source wavelet, except that it must be compact in time (TC inversion) such that the computation of synthetic seismograms is feasible. In the second approach, we assume that the source wavelet can be approximated with a Ricker wavelet, and we invert only for the three parameters describing a Ricker wavelet (source parametrized, SP inversion). Both algorithms have been tested with synthetic crosshole data. The SP approach is slightly superior to the TC inversion when the true wavelet is indeed a Ricker wavelet. However, the TC inversion outperforms the SP approach when the true source wavelet is not well approximated with a Ricker wavelet. This is demonstrated with a field data set acquired in boreholes at a CO₂ injection test site in Svelvik (Norway).
No takes yet. Share an insight, caveat, or question.
Maurer et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: