Detecting Sandstones Affected by Hydrothermal Activity Related to Uranium and REE Mineralization by Multivariate Analysis of Lithogeochemical Data, Athabasca Basin, Canada
Loading...
Date
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Université d'Ottawa / University of Ottawa
Abstract
The Athabasca Basin in northern Saskatchewan, Canada, hosts the world's largest high-grade U resources along the unconformity between sandstones and underlying crystalline basement rocks. Finding U deposits is difficult in the interior of the Athabasca Basin where the sandstone cover is thick. This study focuses on detecting footprints of U and REE mineralization in sandstones in the Wheeler River property which hosts Phoenix and Gryphon U deposits and Maw Zone REE mineralization. The principal component analysis (PCA) reveals that U is positively associated with REEs-Y-Cu-B-Na-Mg-Ni-Be. The evidence suggests that U was dispersed into sandstones together with these elements during the uraniferous hydrothermal activity. The elements positively associated with U are high in concentrations above the deposit, forming a “chimney-like” or “hump-like” distribution in a vertical section. Their enrichment patterns are explained by the ascent of basement fluids through faults to sandstones. The Pb isotope data and the PCA suggest that the secondary dispersion has minor effect on the overall lithogeochemistry of sandstones. The composition of sandstones in the Wheeler River property (excluding sandstones overlying Phoenix deposit) shows that U is positively associated REEs-Y-Cu-Na-Zn-W-Pb-Co-Cr-Sc-Mo-V-Ni-B-Mg. This element association is similar to that of sandstones overlying the Phoenix deposit, indicating that the uraniferous hydrothermal activity was extensive in the entire property. In contrast, sandstones in unmineralized areas in Athabasca Basin show a positive association of U with Th-Ti-Zr-Hf, indicating that heavy minerals, such as zircon and oxide minerals, are major hosts of U. The sandstones in REE-rich Maw Zone show that U is strongly associated with V-Cr-Fe-Ni-Cu-Na-Li-Ba, but very weakly with HREEs. Relative enrichment of HREEs-Y-P in the Maw Zone suggests xenotime as the predominant host of the HREEs. The occurrence of xenotime is confirmed by a mineralogical study.
Using the elements associated with U, linear discriminant analysis and random forest are able to differentiate sandstones into three; above the Phoenix deposit, in the Wheeler River property and backgrounds with high accuracy. The receiver operating characteristic, precision/recall curve and area under cover suggest that random forest is preferred method to linear discriminant analysis. This work demonstrates that the statistical analysis of lithogeochemical data of sandstones is useful in detecting REE mineralization and footprints of deeply buried U mineralization.
Description
Keywords
Unconformity type uranium deposit, compositional data, Mineral exploration, Footprint, Machine learning
