Two 3D time-lapse seismic surveys were acquired in 2012 and 2013 at the Aquistore storage site prior to the start of injection. Using these surveys, we determined the background time-lapse noise at the site and assessed the feasibility of using a sparse areal permanent receiver array as a monitoring tool. Applying a standard processing sequence to these data, we adequately imaged the reservoir at 3150–3350 m depth. Evaluation of the impact of each processing step on the repeatability revealed a general monotonic increase in similarity between the data sets as a function of processing. The prestack processing sequence reduced the normalized root mean squared difference (nrms) from 1.13 between the raw stacks to 0.13 after poststack time migration. The postmigration cross-equalization sequence further reduced the global nrms to 0.07. A simulation of the changes in seismic response due to a range of injection scenarios suggested that was detectable within the reservoir at the Aquistore site provided that zones of greater thickness than 6–13 m have reached saturations of greater than 5%.