Advanced time-lapse processing of continuous DAS VSP data for plume evolution monitoring: Stage 3 of the CO2CRC Otway project case study

Abstract

Active time-lapse seismic monitoring technology is essential for carbon storage projects due to its ability to track the CO2 plume in space and time. A particularly attractive implementation of this technology is permanent seismic reservoir monitoring (PRM) using permanent sources and receivers, which can track subsurface changes in near-real-time over decades. As the number of source points in such a setup is likely to be limited, repeatability and signal to noise ratio need to be improved by processing all the multiple vintages together rather than separately.We explore the advantages of this approach using a PRM dataset acquired to monitor injection of 15 kt of CO2 into a saline aquifer at the Otway International Test Centre (Australia). The monitoring employs continuous acquisition of multi-well offset VSP using nine permanent sources (surface orbital vibrators) and fibre-optic distributed acoustic sensors installed in five monitoring wells producing a vintage every two days for eighteen months before, during and after the injection.Since data reveals seasonal repeatability variations, mainly created by seasonal variations in precipitation levels, for each monitor, an optimal baseline is the one acquired in the same season. The signal-to-noise ratio is further improved by wavefield decomposition of P and S waves. The consistency of the source signature is improved using Wiener filtering. These algorithms improve the data repeatability from about 15–20% to 10–15% normalised root mean square. The results show the CO2 plume signal on the second day of the injection and subsequent time-lapse changes of a stable CO2 plume created at the same reservoir 650 m up-dip five years earlier as a part of the previous field experiment (Stage 2C) at the site.