Targeted nullspace shuttles for full waveform time-lapse seismic monitoring and CO₂ detection thresholds

Kimberly Pike, Kristopher A. Innanen, Scott Keating

Time-lapse seismic monitoring is a proven technique in hydrocarbon reservoir monitoring and optimization. Time-lapse monitoring can also be extended to full waveform inversion and applied to the measurement, monitoring, and validation of CO₂ sequestration projects. Traditional time-lapse implementations detect changes in the subsurface through differencing baseline and monitor surveys and relies on the repeatability of baseline and monitor survey geometries. The models obtained through full waveform inversion are non-unique, and therefore absolute inferences about time-lapse changes are difficult to make. Through targeted nullspace shuttling, we investigate an approach to find unique baseline and monitor models which minimize the time-lapse difference and preserve the data-fit, though explicitly navigating the nullspace, providing a minimum bound on CO₂ plume uncertainty. Using synthetic examples, this approach demonstrates the ability of nullspace shuttling to detect minimum CO₂ time-lapse changes in sparse monitoring scenarios that are avoided in traditional time-lapse applications.