This study mobilizes machine learning (ML) and statistical models to predict Co concentrations and classify anomalies in the Bou Azzer ultramafic Complex of the, which hosts significant Co-Ni-Fe mineralization. Based on a dataset comprising geochemical analyses, enriched with ASTER-derived spectral indices (e.g., Serpentinite index, Ferric Index), we applied three regression models: LR as statistical baseline, PLS regression as multivariate statistical method and RF regression as primary ML algorithm in order to predict Co concentrations, while their corresponding version enabled binary anomaly. A 5-fold nested cross-validation framework, with internal feature selection, was used to ensure solid performance and unbiaised performance evaluation, addressing challenges such as small sample size, multicollinearity (e.g., Sc vs. MgO: r = 0.85), and outliers (e.g., 69 ppm Co). PLS Regression outperformed the others models, achieving a MAE of 5–7 ppm, an Area AUC-PR of 0.94, an AUC-ROC of 0.98, and an R2 of 0.60–0.70. Predictive maps align closely with known Co anomalies along the Bou Azzer Fault, highlighting the critical role of serpentinization as a process creating favorable lithological and alteration conditions for Co mineralization. Feature importance analysis confirms Sc (r = 0.883), MgO (0.867), and Ni (0.786) as key drivers of Co distribution. Despite limitations from the small dataset and complex mineralization, this study demonstrates ML's potential to enhance cost-effective, sustainable exploration, offering a scalable framework for targeting critical metals in ultramafic terrains globally.
Understanding magmatic-hydrothermal conditions, redox-halogen controls, and fluid evolution is essential for deciphering porphyry Cu-Mo genesis and assessing mineralization potential. The Mujicun porphyry deposit, a major intra-continental Cu-Mo system in the North Taihang Mountain region of the North China Craton, has been studied for magma sources and post-ore thermal events, but its physicochemical conditions, magmatic-hydrothermal evolution process, and magmatic fertility remain poorly understood. This study integrates in situ major and trace element analyses of biotite, apatite, and zircon from the ore-hosting diorite porphyry to constrain these factors. Magmatic biotite records the highest crystallization temperatures (782.6–879.6 °C) together with emplacement pressures of 1.12–2.11 kbar, corresponding to depths of 4.08–7.67 km. Zircon and rutile crystallized at somewhat lower temperatures (mean ~700 °C), indicating that these minerals crystallized during the early, high-temperature magmatic stages, prior to significant fluid exsolution, with magmatic biotite representing the earliest near-liquidus phase. Hydrothermal biotite yields slightly distinctly lower temperatures (457.2–478.3 °C) that overlap with the early ore-forming stage recorded by fluid inclusion (115–515 °C), marking the cooling path into the hydrothermal environment. Oxygen fugacity estimates from all three minerals indicate highly oxidized conditions, with a decreasing trend from magmatic to hydrothermal biotite. High fO₂ suppressed early sulfide saturation, enriching Cu and Mo in the residual melt. Halogen geochemistry displays magmatic biotite is Cl-enriched, whereas hydrothermal biotite is relatively F-enriched, reflected by increasing IV(F/Cl) ratios from magmatic to hydrothermal stages. This evolution records preferential Cl partitioning into the exsolved fluid, enhancing Cu transport as chloride complexes, while subsequent F enrichment in later fluid, facilitating Mo mobility as fluoride or molybdate complexes. Apatite displays light REE enrichment, variable Eu anomalies, high Sr contents and Sr/Y ratios, and low Y concentrations, features consistent with published Sr-Nd-Pb isotopic data that indicate a hybrid crust–mantle source involving slab-derived melts. Critically, zircon trace-element signatures, specifically high Ce/Ce* and Eu/Eu* ratios, elevated 10000×(Eu/Eu*)/Y values, and consistently low Dy/Yb ratios, collectively attest to high magmatic water contents and strongly oxidized conditions. Such high fO₂ suppressed early sulfide saturation, thereby pre-enriching Cu and Mo in the residual melt, while the elevated magmatic water content drove the timely exsolution of metal-rich fluids at ~700 °C. Collectively, this multi-mineral approach attests that the Mujicun mineralized system evolved under water-rich, highly oxidized conditions. The synergy of continuous temperature-driven fluid exsolution and a dual Cl–F complexing mechanism enabled the efficient extraction and transport of Cu and Mo, highlighting the utility of coupled biotite-zircon-apatite geochemistry in evaluating porphyry magmatic fertility.
Tectonothermal events associated with the Pan-African Orogeny in the northern West Nigerian Subshield (WNS) are poorly dated. However, in the Bakoshi-Gadanya (BAG) area of the WNS, Pan-African orogenic events were recently constrained to 710–646 Ma (i.e., Cryogenian) based on Rb-Sr and U-Pb isotopic ages. Whether or not this age range is the complete record of the orogenic event is untested, in view of younger (i.e., Ediacaran) ages reported from the East Nigerian Subshield and some parts of the WNS. Given that high-temperature events (<600 °C) can reset the U-Pb isotopic system in apatite, we selected apatite to investigate potential post-Cryogenian orogenic events in the BAG area. Our study reports apatite chemical compositions and U-Pb and Sm-Nd isotopic ages, providing new constraints on early to late Ediacaran metamorphic-metasomatic events of the WNS. Apatite from the country rocks (schists and gneisses) are well-formed and homogeneous, occasionally deformed, and have bright cores with dark rims, suggesting a late-stage metamorphic or metasomatic origin. Apatite grains from the granitic phases are also homogeneous and display bright cores enveloped by dark rims cut by late-stage veins. They show patchy zonation and display dissolution-reprecipitation features, suggesting a metasomatic process. Apatite from the granites yielded variable 238U/206Pb weighted mean ages of 603 ± 40 Ma and 551 ± 44 Ma (Bakoshi Granite), 682 ± 25 Ma and 607 ± 26 Ma (Shanono Granite), 683 ± 19 Ma (Yettiti Granite), and 549 ± 13 Ma (Gadanya Granite). Apatite from the Gadanya Granite gives a Sm-Nd isochron age of 550 ± 29 Ma. Apatite grains from the gneisses yielded 238U/206Pb weighted mean ages of 700 ± 30 and 644 ± 20 Ma, whereas apatite from the schists yielded an age of 645 Ma. Apatite from tourmalinite yielded 238U/206Pb ages of 641 ± 52 and 550 ± 20 Ma. Clustering of these isotopic ages implies (1) an early-stage metamorphic event (∼644–608 Ma), and (2) a late-stage metamorphic-metasomatic event (∼550 Ma) in the BAG area. The early-stage event was of regional metamorphic character (i.e., apatite in gneisses and schists), whereas the late-stage event recorded contact metamorphic-metasomatic processes (i.e., apatite in granites and tourmalinites). Both the metamorphic and metasomatic events affected LREE distributions in the BAG apatite compared to the host lithologies. The Ediacaran-aged events recorded by BAG apatite represent the youngest known Pan-African tectonothermal events in the Malumfashi Schist Belt of the WNS. Our study serves to demonstrate the utility of apatite in unravelling discrete events that other minerals may not record due to apatite's unique ability to reset its isotopic system at relatively low (<600 °C) temperatures.
Rapid and reliable identification of multivariate geochemical anomalies is critical for delineating prospective mineralized zones and reducing uncertainty in mineral exploration targeting. Extended isolation forest (EIF) is a powerful unsupervised ensemble learning algorithm that efficiently isolates anomalies from high-dimensional geochemical datasets using randomly oriented hyperplane partitions. Previous studies have demonstrated the effectiveness of EIF in multivariate geochemical anomaly detection and mineral potential modeling. However, its performance can be significantly affected by stochastic variability arising from random partitioning and random subsampling during isolation tree construction, which may result in unstable anomaly patterns and inconsistent exploration targets in complex geological environments. To mitigate this limitation, we developed a robust unsupervised framework for the identification of multivariate geochemical anomalies associated with gold mineralization in the Southwestern Yilgarn Craton, Australia. The proposed framework integrates robust factor analysis (RFA), a Jaccard-based stability index and EIF to enhance the reliability and reproducibility of anomaly detection. RFA was first applied to compositional soil geochemical data to identify the most significant pathfinder elements associated with gold mineralization, which were subsequently used as input variables for the EIF model. The model was then optimized using a Jaccard-based stability criterion to ensure consistent anomaly detection across repeated independent runs. Model performance was assessed using area under the receiver operating characteristic curve (AUC). The obtained AUC value of 0.82 indicates strong predictive capacity, confirming that the generated anomaly map effectively delineates mineralization-related geochemical patterns and provides a reliable proxy for mineral prospectivity mapping. Overall, the proposed framework offers a robust and reproducible unsupervised approach for multivariate geochemical anomaly detection with strong applicability in both greenfield and brownfield mineral exploration settings.
The genesis of the Cryogenian Datangpo-type Mn deposits in South China remains contentious, particularly concerning the source of Mn and its precipitation mechanism. This study conducts an integrated investigation, combining petrography, mineralogy, fluid inclusion analysis, and geochemistry, on the Lannitian deposit, a representative deposit in the world-class Hunan–Guizhou–Chongqing Mn Triangle. Results show that rhodochrosite and Mn-rich dolomite are the principal Mn-bearing minerals and contain considerable amounts of methane (CH₄)-bearing fluid inclusions. Integrated sedimentological and geochemical evidence, including high Mn/Al ratios, a distinct negative Ce anomaly, and a hydrothermal affinity in Co/Zn space, indicates a dominant submarine hydrothermal source of Mn within a highly restricted, redox-stratified basin during the post-Sturtian interglacial. The occurrence of tuffaceous horizons and regional tectono-magmatic activity further suggests that hydrothermal circulation was partly influenced by synsedimentary volcanic activity and associated magmatic heat and/or fluid inputs in the rift setting. We propose that hydrothermal Mn(II) was subsequently regulated by microbial oxidation, redox cycling, and early diagenetic processes, including organic matter degradation and methane cycling, leading to Mn‑carbonate precipitation. The Lannitian deposit therefore represents a hybrid hydrothermal–sedimentary system in which hydrothermal and minor magmatic influences supplied reactive Mn and modified local redox conditions, whereas basin restriction, redox stratification, and diagenesis governed final ore formation. These results provide new insights into the metallogenesis of Datangpo-type Mn deposits and highlight the interaction between hydrothermal activity, volcanic processes, and redox evolution in Neoproterozoic rift basins.
