Joint 4D single-iteration least-squares migration

Time-lapse (4D) seismic monitoring generally benefits from a dedicated co-processing where important inconsistencies between vintages are attenuated. We propose a joint formulation for least-squares migration (LSM) that further mitigates effects such as low repeatability and diverse data quality, by constraining the least-squares inversion with information shared among baseline and monitor data sets. It can be applied in both post-stack and pre-stack domains. We show results for two pairs of baseline/monitor data sets with different levels of repeatability in the Brazilian pre-salt. In both cases, joint 4D LSM gave an overall reduction of 4D noise.