This project aims to reconstruct time-lapse pre-stack seismic data to analyze and replicate anomalies caused by CO₂ injection. The primary focus is to develop innovative methodologies for data regularization and reconstruction, enabling the processing of undersampled seismic datasets to preserve critical amplitude and travel-time information. These efforts are directed toward designing efficient and fast-turnaround acquisition strategies with minimal source usage while maintaining high-resolution subsurface imaging capabilities.

The project integrates theoretical advancements in compressive sampling and Fresnel zone analysis to optimize source configurations for illuminating CO₂ plume targets effectively. Numerical reconstruction techniques will be developed to create data grids suitable for imaging, ensuring the fidelity of time-lapse anomalies.

Machine learning will be explored as a complementary tool to enhance denoising, separate time-lapse anomalies from background noise, and mitigate processing and migration artifacts. These efforts are supported by established expertise in geophysical ML/AI applications, enabling the project to adopt cutting-edge computational techniques.

The work is further strengthened through mentorship and collaboration with experienced researchers specializing in seismic data preconditioning and imaging. This research aims to contribute significantly to the advancement of seismic monitoring technologies for CO₂ storage by providing reliable and efficient data reconstruction methods.