Simulating spatial and temporal varying CO 2 signals from sources at the seafloor to help designing risk‐based monitoring programs

Risk‐based monitoring requires quantification of the probability of the design to detect the potentially adverse events. A component in designing the monitoring program will be to predict the varying signal caused by an event, here detection of a gas seep through the seafloor from an unknown location. The Bergen Ocean Model (BOM) is used to simulate dispersion of CO 2 leaking from different locations in the North Sea, focusing on temporal and spatial variability of the CO 2 concentration. It is shown that the statistical footprint depends on seep location and that this will have to be accounted for in designing a network of sensors with highest probability of detecting a seep. As a consequence, heterogeneous probabilistic predictions of CO 2 footprints should be available to subsea geological CO 2 storage projects in order to meet regulations.