Time-lapse analysis of CaMI.FRS CO2 VSP data

Brendan Kolkman-Quinn, Donald C. Lawton, Marie Macquet

The Containment and Monitoring Institute Field Research Station (CaMI.FRS) operates a carbon sequestration experiment near Brooks, Alberta. By March 2021, 33 tonnes of CO2 had been injected into a Basal Belly River Fm sandstone at 300m depth, as a simulation of a shallow leak from a CO2 storage reservoir. Vertical Seismic Profiles (VSP) were collected between 2017 and 2021 to determine the detection threshold and monitor the plume of injected CO2. These field data had high repeatability, with permanent borehole geophones and identical shot coordinates, with only surface conditions adding variability. A finite-difference VSP forward model was created to predict the 2D time-lapse response of the CO2 plume and allow for testing of processing parameters with noiseless data. The 10Hz-150Hz field data required careful processing. A time-lapse compliant processing workflow was developed by paring down a standard VSP workflow, with the goal of avoiding unnecessary alterations to amplitude and phase. This simplified and minimized the cross-equalization between baseline and monitor amplitude spectra: Applying a pre-stack shaping-filter preserved most of the frequency content, at the cost of diminishing the amplitude of the CO2 time-lapse anomaly. Alternatively, applying custom high-cut filters to the baseline and monitor shot gather pairs reduced the overall frequency content but preserved amplitude contrasts in the reservoir, yielding a higher-confidence result. The 2019-2017 time-lapse difference for a 15t of injection did not unequivocally detect the plume, showing a faint CO2 anomaly amid background residuals of similar amplitude. The 2021-2017 results showed a higher-amplitude time-lapse anomaly similar to model’s predictions. The normalized-root-mean-square (NRMS) difference between reflection amplitudes ranged from 10%-20% in the field data, above the noiseless model’s NRMS of 8%-13%. The interpreted CO2 plume had a 45m-51m lateral extent with an asymmetric distribution around the injection well, indicating greater permeation towards the south-west than to the north-east. A detection threshold between 15t and 33t of CO2 in a 10% porosity reservoir has been confidently established. These findings will help inform Monitoring, Measurement and Verification (MMV) procedures at future CO2 sequestration operations.