Accelerating CCS site selection: Automated fault seal and stability analysis using seismic and AI

Effective site screening for geologic carbon capture and storage (CCS) demands rapid identification of locations with minimal risk of CO 2 migration along faults. Traditional seismic interpretation workflows, often manual and time-intensive, can significantly delay these crucial early screening decisions. In this study, we introduce a rapid, data-driven methodology that harnesses the power of deep learning to directly extract quantitative fault throw information from seismic data. This capability enables an unprecedented level of automated fault zone characterization. By seamlessly integrating these deep-learning-derived throw magnitudes with readily available shale volume data, we efficiently compute the shale gouge ratio (SGR), a widely recognized proxy for fault seal capacity. Furthermore, we leverage the seismic-derived fault geometry (dip and azimuth) to perform a robust mechanical stability assessment using established slip tendency and critical pore pressure criteria. Applied to a 3D seismic data set from the Exmouth subbasin, offshore Western Australia, our automated workflow generates continuous 3D volumes and insightful maps of SGR and fault stability. These outputs directly highlight the most secure injection zones and potential leakage or reactivation risks. This novel, automated workflow represents a significant leap forward, capable of dramatically accelerating the screening and derisking processes in large-scale CCS site evaluations.