153 1. INTRODUCTION The growing global demand for critical and base metals poses significant challenges for mineral exploration and the timely delivery of new resources. While many easily accessible deposits have already been discovered, remaining resources are often deeper, subtler, or concealed beneath complex geological settings. Traditional exploration workflows typically analyze geochemical, geophysical, remote sensing, and structural datasets independently before integrating them through expert-driven interpretation. Although conceptually robust, this approach can be time-consuming, subjective, and difficult to scale efficiently across new regions and exploration programs. Mineral Prospectivity Mapping (MPM) has emerged as a systematic framework (Lou & Liu, 2023; Shaw et al., 2022; Yan et al., 2025) to integrate diverse geological datasets and identify areas with higher mineralization potential. Early MPM approaches relied heavily on heuristic weighting and linear combinations derived from expert knowledge; these methods may struggle to capture complex, non-linear interactions among geological variables and are often sensitive to subjective parameter selection. Recent advances in Machine Learning (ML) provide powerful alternatives for modeling highdimensional geoscientific data. Algorithms such as Random Forest, Support Vector Machines, and Gradient Boosting (Lauzon & Gloaguen, 2024; Lou & Liu, 2023) have demonstrated strong capability in identifying predictive patterns within complex spatial datasets. By learning relationships directly from data, ML-based approaches can reduce interpretative bias and improve consistency in mineral prospectivity. However, despite growing adoption of ML in exploration, challenges remain regarding reproducibility, transparency, systematic integration of heterogeneous datasets, and spatial validation robustness. There is a need for automated and scalable workflows that objectively integrate multidisciplinary geological information while preserving geological interpretability. In this context, the objective of this study is to develop and evaluate an automated, reproducible Machine Learning workflow for copper porphyry prospectivity mapping in the Arequipa region of southern Peru. The proposed framework integrates multidisciplinary datasets within a unified and scalable spatial modeling environment to evaluate predictive performance and geological coherence. By reducing subjectivity, standardizing data integration, and accelerating target generation, it supports more efficient exploration investment decisions and contributes to faster, smarter, and more responsible mineral discovery. 2. MATERIALS AND METHODS 2.1. Study area and Multidimensional Datasets The study was conducted in the Arequipa region, southern Peru, a zone of significant Cu-Mo porphyry mineralization. The regional geology comprises Mesozoic sedimentary rocks (Yura Group), intrusive bodies of the Coastal Batholith, Cretaceous carbonate units, and Cenozoic volcanic sequences (Tacaza and Barroso Groups), as documented in regional geological quadrangle studies (Avendaño et al., 2000; Bellido & Guevara, 1963; García Márquez, 1968; Guevara Rosillo, 1969; Guizado Jol, 1968; Pecho Gutiérrez & Morales Serrano, 1969) , which host deposits such as Cerro Verde, Tía María, and Zafranal. All data were obtained from public repositories to ensure reproducibility and scalability and datasets were standardized and aggregated into a unified 500 × 500 m spatial grid to ensure consistent feature representation across variables. Eight ASTER satellite images (HDF format 15 x 15 m resolution) were downloaded from NASA's Earthdata platform. Public geochemical data (stream sediment and rock samples), structural data (lineaments, faults), and geological cartography were acquired from the GEOCATMIN portal of INGEMMET (Figure 1). The geophysical data consisted of a magnetic anomaly grid with Reduction-tothe-Pole (RTP) correction (TIF format).
RkJQdWJsaXNoZXIy MTM0Mzk2