Leveraging Machine Learning and Geophysical Data for Automated Detection of Interior Structures of Cratons

The internal structures and discontinuities of cratons hold considerable economic value due to their tendency for reactivation and different horizontal stress, serving as conduits for fluid flow and mineral deposition over time. Detecting these structures at various depths is critical for accurately mapping prospective zones of metallic mineralisation. This study demonstrates the effectiveness of integrating signal processing, feature extraction, and clustering on magnetic and gravity data for mapping the internal structures of the Gawler Craton, which has undergone rifting, sedimentation, extension, and orogenic processes. This combined approach results in precise internal structural mapping. Validated by three distinct metrics and geological maps, the resulting clustered maps can serve as foundational tools for further exploration and support decision-making in mineral exploration. Our findings indicate that most known metallic mineral occurrences, including all significant ones, are formed near the boundaries of these clusters. Therefore, mapping and targeting these boundaries can significantly reduce the search area for structurally controlled, extension-related mineral systems. Our proposed framework addresses the challenges of mapping hidden shallow and deep crustal structures, enhancing the capabilities of exploration geophysicists and geologists to investigate geological settings and the interiors of cratons. It provides a rapid, reliable, and cost-efficient method for generating geophysical features, which can be used as input to supervised prospectivity mapping workflows to identify favourable sites for mineralisation at any stage of an exploration program.