Exploitation of data-information content in elastic-waveform inversions

Abstract

Elastic-waveform inversions have the potential to provide detailed subsurface images of the elastic parameters (P- and S-wave velocities and density), but acquisition of suitable data sets and their inversion are nontrivial tasks. We explore the information content offered by elastic-waveform data by means of a 2D synthetic study. Comprehensive noise-free data sets that include recordings based on multicomponent (directed) sources and multicomponent (vector) receivers that fully surround the area of interest allow all elastic parameters to be reliably recovered. Results that are almost as good can be achieved with the more commonly used crosshole configuration. If only single-source components (e.g., those oriented perpendicular to the borehole walls) are used, then there is no significant quality degradation of the tomographic images. Crosshole experiments that include pressure sources and multicomponent receivers still allow P- and S-wave velocities to be recovered, but such data sets contain virtually no information about the density. Finally, seismic data collected with omnidirectional pressure sources and pressure receivers contain information about P- and S-wave velocities, but there are pronounced trade-offs between these parameters. This is demonstrated through formal model-resolution analyses. This study concludes that seismic data recorded with pressure sources and 2C receivers offer the best compromise between acquisition efficiency and data-information content.