Multivariate statistical analysis encompasses a range of methods that can fingerprint mineralization, alteration, and host-rock signatures within an ore system, thus assisting in interpretation of ore deposit models and supporting exploration programs. We utilize numeric interpolation of metals (Fe, Cu, and U), principal component analysis (PCA), and a Random Forest (RF) classification, applied to whole-rock geochemical data, to define metal distribution patterns and geochemical signatures of alteration/Fe oxide mineralization in the Wirrda Well iron oxide copper-gold (IOCG) prospect. The prospect is located in the Olympic Dam district, Gawler Craton (South Australia), bounded by NW- and NE-trending faults and characterized by two distinct residual gravity anomalies: Wirrda Well North (WW-North) and Wirrda Well South (WW-South). The mineralization is attributed to hydrothermal activity associated with magmatism at ~1.6 Ga that generated the Gawler Silicic Large Igneous Province (SLIP), corroborated by U-Pb dates of hematite. The mineralization is hosted by the ~1.85 Ga Donington Granitoid Suite and abundant mafic dikes, the majority of which are considered to predate mineralization.
Five lithogeochemical clusters are defined from PCA, hierarchical, and k-means clustering, efficiently discriminating least-altered felsic lithologies from altered, mixed lithologies (felsic and mafic) and, importantly, two distinct mineralization clusters, representing magnetite and hematite. The RF method is successfully applied to a larger data set with a smaller number of analyzed elements to extrapolate the results over the whole prospect. WW-North is characterized by an Fe-V-Ni-Co signature defined as “magnetite-type,” whereas WW-South contains higher-grade Cu-Au(± Bi-Ag) mineralization and has a marked Mo-W-U(± Sn) signature defined as “hematite-type.” The latter is considered a characteristic hypogene hydrothermal signature in the Olympic Dam District. Sodic and hydrolytic alteration are associated with Fe metasomatism and are depicted by the magnetite and hematite clusters, respectively. In combination, the results indicate that premineralization mafic dikes provided permeable structures for fluid focusing and metal deposition, since metals are concentrated by their contacts with host granite, specifically in WW-North. The metal interpolations and geochemical data analysis suggest strong lithological-structural control on IOCG mineralization at Wirrda Well.