LAUSR

Abstract A controlled CO2 release experiment was carried out in order to mimic unintended leakage of geologically stored CO2, and to study methods for detecting these leak events. The experiment was carried out at 60 m depth in the Oslo Fjord over the course of one month. During the simulated leak events, the water chemistry was monitored by sensors mounted on stationary templates located 10 and 22 m horizontally from the source, as well as sensors mounted on an Autonomous Underwater Vehicle (AUV). During baseline conditions (no CO2 release), we observe a strong biogenic correlation between O2 and CO2. This correlation is lacking during CO2 releases, indicating that the CO2-O2 correlation can be used as a marker for CO2 leakage. The same deviations were not initially detected by the CO2 sensor mounted on the AUV due to the longer response time of membrane-based CO2 sensors. However, by applying response time correction in the post-processing of the AUV CO2 data, the generated CO2 plume was detected. Moreover, the plume was clearly detected by the AUV using the faster responding pH sensors, where the correlation with O2 again could be used to confirm the anomaly.