全球变化知识资源中心

To mitigate global warming, the large volume of CO2 released into the atmosphere must be reduced. CO2 capture and storage (CCS) are promising means for this. In addition, CCS operators who inject and store CO2 into subseabed geological formations may be legally required to monitor CO2 leakage into seawater, even though such events are very unlikely; after detecting unusually high CO2 concentrations, their social responsibility implies specifying the position and flux of the associated CO2 seepage in the sea site. In this study, we adopted the adjoint marginal sensitivity method to predict such parameters based on the information recorded by a limited number of CO2 sensors located on the seafloor. In particular, we newly derived the adjoint marginal sensitivity from the relationship between a source term and the measured concentrations. We successfully conducted two-dimensional tests and three-dimensional simulations with a realistic topography; tidal and ocean currents with CO2 dispersion were time-forwardly simulated and the seepage information was efficiently evaluated via time-backward numerical tests. However, the limitation of this method also emerged. We accumulated knowledge on its practical utility and proposed some ways to reduce the estimation errors. It is also inferred that this method can be applied for the source detection of various marine pollutants and resources.