Uncertainty quantification of CO₂ plume migration using static connectivity of geologic features

During the operation of a geological carbon storage project, a critical question is whether injected CO₂ remains within the permitted zone. However, because a large suite of subsurface models are possible given very sparse static data, simulating flow in the entire suite to quantify the uncertainty in CO₂ plume migration is impractical. We propose a fast alternative that scans the suite of geologic models and groups them on the basis of static connectivity. Grouping is achieved simply by measuring the shape dissimilarity of permeable zones in the prior models using path skeletons. By selecting a specific group of models that reflect the performance observed in the field, it is possible to quantify the uncertainty in CO₂ plume migration. Our approach is compared against results obtained by flow simulation and subsequent model classification using principal component analysis (PCA) in characteristics of connectivity.