Continental-scale probabilistic resistivity imaging of Australia using deep learning: Implications for geology, groundwater, and critical minerals

This paper presents a continental-scale electrical resistivity model of Australia's shallow crust extending to approximately 650 meters depth, derived from probabilistic inversion using deep learning techniques. The model is built upon more than 26.6 million airborne electromagnetic (AEM) soundings collected across 352,600 line-kilometers from the AusAEM program. The approach provides quantified uncertainty estimates, enabling applications in subsurface structure mapping, lithological interpretation, fluid distribution analysis, and critical mineral exploration.