Carbon Capture and Storage (CCS) projects could grow to become a major new industrial activity over the next few decades; but securing the associated climate benefit is critically dependent on ensuring high integrity containment of the injected CO2. Our technique, called LightSource, is based on commercially available optical gas sensors that measure path-averaged CO2 gas concentrations along beams scanned over part of an onshore CCS site. Inter-beam correlations are used to infer the current local ambient background concentration. Statistically significant discrepancies between the multiple beams' path-averaged concentration measurements can be used to infer the existence of a source by applying the methods of statistical process control. This allows the estimation of the anomalous concentration on each beam that is associated with the inferred source(s). Using these anomalous concentration data in conjunction with a gas dispersion model, high frequency wind velocity and turbulence intensity data, we can solve the inverse gas dispersion problem to estimate the location and mass emission rates of the source(s) that best explain the data. The system does not require sources to be situated within the beam pattern unlike tomographic approaches which in addition require more intensive instrumentation. We have evaluated the LightSource system's performance under field conditions at the Quest CCS project site in Alberta Canada. All calibrated releases of tempered CO2 at emission rates of up to 300 kg/hr were successfully detected. © 2017 The Authors.