This study presents the first national-scale clay-hosted rare earth element (clay-hosted REE) mineral potential assessment of Australia, developed using a mineral systems approach. The study was undertaken as part of “Accelerating Development of Australia's Rare Earth Resources” (ADARER) project under the Australian Critical Minerals Research and Development Hub. The aim of the assessment is to promote exploration, discovery and inform land-use decision making in Australia by delineating areas of higher geological potential for the occurrence of clay-hosted REE mineral systems.Clay-hosted REE deposits are globally important economic and strategic resources, and represent a crucial source of REE, which are essential for modern technologies supporting economies, national security, and energy transition goals. Overseas deposits, such as those in south China, Myanmar and Madagascar, are commonly reported as being characterised by a high proportion of ion-adsorbed REE mineralisation bound to clays, with lesser proportions of primary or secondary mineral phases such as monazite and xenotime, and a range of carbonate, phosphate and silicate minerals (e.g., allanite, bastnäsite, cerianite, and rhabdophane). Ion-adsorption-dominated mineralisation has gained particular attention in recent times owing to the relatively low cost and technical simplicity of REE extraction. Clay-hosted REE mineralisation in Australia typically contains a higher proportion of REE in primary and secondary minerals, with a much lower ion-adsorption component.The clay-hosted REE mineral potential assessment presented here applies two models: (i) a knowledge-driven weighted sum (KDWS), and (ii) a random forest machine learning (RFML). The KDWS model used five mappable criteria representing four mineral system components and predicted 85 % of known deposits/occurrences in 57 % of the area; a moderately performing model that reduced the effective exploration search space by 43 %. The RFML model used eight mappable criteria representing four mineral system components and predicted 95 % of clay-hosted REE deposits and occurrences in 25 % of the area. It has a stronger predictive power relative to the KDWS model and reduced the exploration search space by 75 %. Both the KDWS and RFML models highlight known and previously unrecognised regions prospective for clay-hosted REE deposits. Areas of higher mineral potential occur mostly in highly weathered regions dominated by igneous intrusive lithologies, particularly those of higher alkali content, which have greater potential to host primary REE-bearing minerals. Known areas of high mineral potential are Precambrian cratons or orogens in the central and western parts of the continent. In eastern Australia higher mineral potential is primarily associated with Paleozoic to Cenozoic granitic and volcanic lithologies. Newly identified regions of elevated prospectivity include northern and central Queensland, the Eastern Highlands (Queensland, New South Wales and Victoria), the Pine Creek and Tennant Creek regions (Northern Territory), and within multiple geological regions straddling the Western Australia and Northern Territory border.This study highlights how precompetitive geoscience data can be integrated using a geologically robust framework to generate national-scale mineral potential models with strong predictive performance and identifies caveats appropriate to the interpretation of the model predictions. It also demonstrates that such models can be applied to less conventional mineral systems, such as regolith-hosted settings, where mineralisation has developed primarily through secondary chemical enrichment via weathering processes.
Australia) et al. (Thu,) studied this question.