Injection and migration of CO2 within geologic formations and its reactions with reservoir rocks will cause changes in seismic velocities, density and attenuation, resulting in changes in time-lapse seismic scattering/diffraction. In this study, we use reverse-time migration of time-lapse walkaway vertical seismic profiling (VSP) data to study reservoir changes due to CO2 injection/migration. The data were acquired at the SACROC enhanced oil recovery (EOR) field in Texas using geophones at depths from 5000 to 5700 feet in a monitoring well. The CO2 injection started in October, 2008, and the time-lapse walkaway VSP data were acquired in July 2008 before the CO2 injection and April, 2009, six months after the CO2 injection. The objective of the project is to study the combined EOR and geologic carbon sequestration. We apply statics corrections to the time-lapse walkaway VSP data, and then use a spectrum analysis method to balance the amplitudes of the data to make migration images comparable. We conduct reverse-time migration using the upgoing waves of the baseline and balanced walkaway VSP data to study detailed changes in the reservoir. The results demonstrate that reverse-time migration of time-lapse walkaway VSP data can reveal reservoir changes due to CO2 injection and migration.
The sources of time-lapse offset vertical seismic profiling (VSP) surveys should be located exactly at the same positions to reliably monitor reservoir changes due to CO2 injection. However, there is often some uncertainty in source locations during time-lapse data acquisitions. We use double-difference tomography and downgoing waves of time-lapse offset VSP data to invert for the source locations and the velocity structures simultaneously. We use synthetic data for validation and apply doubledifference tomography to time-lapse offset VSP data acquired at the Aneth oil field in Utah for monitoring CO2 injection. Our synthetic studies demonstrate that doubledifference tomography can accurately invert for VSP source locations. Our results of Aneth field data show that real source locations of time-lapse offset VSP surveys are separated up to a few tens of meters. Accounting for these source location differences can improve reliability of timelapse VSP monitoring.