Abstract:
Objective 3D mineral resource potential evaluation serves as a vital tool for deep mineral exploration. In recent years, the integration of machine learning algorithms has significantly improved the accuracy of such 3D mineral resource potential evaluation, contributing substantially to advances in mineral exploration. The current mining elevation of the Weishan rare earth element (REE) deposit in Shandong is approximately −200 m, and the continuation of deep resources is facing challenges. This study aims to delineate the ore−controlling factors, develop a quantitative predictive model, and identify potential exploration targets.
Methods 3D geological modeling was constructed to delineate the spatial distribution of ore bodies, lithological units, and structural features. Key geological factors, including lithology, structure, and geochemical signatures, were quantitatively derived from the models. Mineralization prospectivity was then evaluated and compared using the weight−of−evidence (WoE) and random forest (RF) methods.
Results The WoE was applied to assess the relationships between geological factors and mineralization, followed by the calculation of corresponding weight values. Subsequently, a random forest (RF) model was constructed based on these weighted factors for deep mineral potential prediction. Integrated with the regional metallogenic framework, two prospective exploration targets were delineated.
Conclusions The integrated “3D modeling–weights of evidence–random forest” approach developed in this study provides an effective framework for enhancing the accuracy of deep mineral exploration. The results demonstrate that the Weishan REE deposit exhibits promising exploration potential at depth, and the delineated target areas offer practical guidance for further exploration.
下载: