A Geology-Informed Remote Sensing and Machine Learning Framework for System-Scale Porphyry Copper Exploration

Porphyry copper systems are characterized by kilometer-scale hydrothermal alteration footprints that extend well beyond mineralized cores and can be mapped regionally using multispectral remote sensing. Although the integration of remote sensing (RS) and machine learning (ML) has become increasingly common for porphyry copper targeting, many RS–ML studies remain weakly validated in geological terms and rely primarily on spatial coincidence with known deposits, providing limited assessment of whether predicted targets represent complete hydrothermal systems. This study presents an integrated ASTER-based remote sensing–machine learning (RS–ML) methodology for regional-scale porphyry copper exploration in the Copper Triangle, Arizona (2,486 km²), combining multi-algorithm alteration mineral mapping with system-scale geological validation. Six alteration minerals representing the proximal-to-distal zonation sequence were mapped using Spectral Angle Mapper (SAM), Adaptive Coherence Estimator (ACE), Matched Filtering (MF), and Constrained Energy Minimization (CEM), followed by algorithm fusion and spatial analysis using Kernel Density Estimation (KDE) and K-means clustering to delineate porphyry-favorable regions. Validation was conducted at two levels: deposit-based assessment and chlorite-based evaluation of system-scale geological consistency. Thirteen discrete chlorite halos were delineated (area: 4.46–233.14 km²; equivalent radii: 1.19–8.61 km). Deposit-based validation confirmed 100% recovery of major porphyry districts (5/5). Validation halos associated with known districts (n = 8; mean radius 3.6 ± 1.5 km) exhibit dimensions consistent with classical propylitic alteration models, while greenfield halos (n = 4) are interpreted as composite systems. Lithological analysis demonstrates that validation halos are hosted by volcanic, intrusive, and consolidated sedimentary units, whereas greenfield halos retain coherent footprints despite >70% surficial cover, indicating cover-masked alteration. The proposed framework advances RS–ML porphyry copper targeting toward a reproducible, geology-informed approach applicable to both exposed and covered terranes.