Integrated 3D modelling and associated machine learning targeting: the Jaguar Greenstone Belt example.

Mira Geoscience completed an integrated interpretation in the Jaguar Greenstone Belt (JGB), in Western Australia, on behalf of Round Oak Minerals (now Aeris Resources). The 3D structural and stratigraphic regional model, consistent with geophysical data sets was the foundation for the exploration model. The targeting and prospectivity analysis were based on quantifying exploration criteria and explicitly representing these criteria in the exploration model for sub-seafloor replacement-style Volcanic Hosted Massive Sulfide (VHMS) deposits. First, the regional geological model was built from geological constraints (mapping, drill holes) but also developed in close integration with potential fields data, producing a viable starting model for geologically constrained inversion to solve for rock property variations within geological domains. When the model thus constructed was submitted to geologically constrained inversion to reconcile unexplained response as property variations within those domains, sensible/stable property variations were recovered in the inverted model, which it was possible to interpret in terms of alteration and potential targets. The exploration criteria were translated using the integrated 3D model to create exploration vectors that were representative of the mineral system. In other words, these vectors were numerical realisations of the various targeting criteria. The prospectivity analysis at Jaguar used a Machine Leaning approach, namely Random Forests, to generate a 3D Mineral Potential Index based on different combinations of input exploration vectors. This resulted in identification of 41 separate targets within the JGB.