DynaSIRT: A Robust Dynamic Imaging Method Applied to CO2 Injection Monitoring

Dynamic imaging provides an effective way to integrate previous surveys seismic data in order to estimate current velocity model. It is particularly useful for time-lapse imaging, which has been successfully applied to reservoir monitoring over the years by oil industry and developed for CO2 sequestration monitoring more recently. This work aims to introduce a new dynamic imaging method for permanent acquisition systems applied to monitor CO2 injection for detection of undesired leaks that may cause environmental impact. We propose a method called DynaSIRT that integrates previous surveys data in an efficient way, without reprocessing older data. The proposed method keeps state variables that store the temporally damped effective illumination of previous surveys and timestamps in order to track each parameter update. We successfully applied DynaSIRT to image a synthetic time-lapse diffraction tomography dataset, providing a clear detection of a CO2 leakage even for sparse survey geometry, thus showing its relevance for CO2 injection safety assessment.