The quality of time-lapse analysis depends on the quality of acquisition repeatability between base and monitor surveys. In practice, it is almost impossible to replicate a base survey due to environmental conditions, especially for marine streamer surveys. In most cases where acquisition differences are minimal, standard four-dimensional (4D) processing techniques are sufficient. However, cases lacking good repeatability often result in strong 4D noise that requires more sophisticated approaches and technology. In our case study in the Northern Carnarvon Basin, severe weather conditions were encountered during the 4D monitor survey. Strong currents led to source positioning errors and significant streamer feathering that degraded the 4D signal-to-noise ratio. Thus, a deterministic 4D de-striping correction processing flow, instead of a conventional survey matching processing flow, was used to suppress the acquisition-related 4D noise. This flow decoupled two noise-contributing factors: (1) water velocity error (DV); and (2) mean sea level error (DZ) based on physical characteristics. Furthermore, to better reveal the underlying 4D signal, least-squares migration (LSM) reduced migration artifacts and compensated for illumination variation. Four-dimensional noise was reduced in the LSM process by mitigating position errors through demigration and remigration techniques. With a comprehensive 4D processing flow, we were able to achieve a reduction of NRMS from 0.7 to 0.1. This resulted in a better 4D signal for the interpretation of reservoir performance and activity, and in our case, much-improved visibility of the water ingress.

Open-Access Online Publication: May 31, 2023