More than four months of continuously recorded micro-earthquake data acquired at Cold Lake, Canada, was analyzed using advanced algorithms for microearthquake location and subsurface tomography. Robust determination of the spatial, temporal, and magnitude distribution of seismicity is the first step toward understanding the relationship between the stress perturbations caused by the cyclic steam stimulation (CSS) process and seismicity. Acquisition geometry was constrained because the receivers were located in a single vertical borehole. Despite this constraint, we were successful in improving event locations by use of the double-difference method, which highlights several tight event clusters. The deep cluster at a depth of , just above the oil reservoir, shows very high seismicity during the CSS processes. A second cluster is observed at shallower depths in the successive steam cycle. This suggests that repeated steaming causes the deformation to spread to shallower depths. The number of events, however, decreases in the second steam cycle. Even though some of the largest events occur below the Clearwater reservoir, we observed few events in the reservoir itself, indicating that the reservoir may be an aseismic region. The size and distribution of seismicity during the first cycle agrees with a Mohr's circle analysis using simple geomechanical modeling.