Time-lapse full-waveform inversion for cases with seawater or near-surface velocity changes

Xin Fu, Kristopher A. Innanen

Time-lapse (4D) seismic full-waveform inversion (FWI) can provide high-resolution imaging of reservoir changes caused by the production of hydrocarbon (e.g., enhanced oil recovery) and the unground storage of CO₂. However, successful seismic monitoring depends on good repeatability between baseline and monitor surveys. However, the capacity of time-lapse FWI technology on solving the non-repeatability issue of seawater or near-surface velocity changes during baseline and monitor seismic surveys has not been demonstrated. In this paper, we investigate the capability of the parallel strategy, the double-difference strategy, the sequential strategy, and the common-model strategy in the non-repeatability issue of seawater or near-surface velocity changes using synthetic time-lapse marine streamer data, time-lapse OBN (ocean-bottom node) data, and time-lapse surface land data. The investigation shows that when using marine streamer data, both the double-difference strategy and the common-model strategy can adapt to relatively small seawater changes, and only the common-model strategy can adapt to relatively big seawater changes; when using OBN data, the parallel strategy, the double-difference strategy, and the common-model strategy all can adapt to relatively small seawater changes, and the best result is given by the double-difference strategy, but none of them can adapt to relatively big seawater changes; when using surface land data, the common-model strategy can adapt to the random near-surface changes best.