Integrating mineral-scale evidence with PMF-SHAP to interpret geogenic reservoirs and mining-enhanced trace-element dispersion in agricultural soils.

Distinguishing geogenic reservoirs from mining-enhanced trace-element dispersion remains challenging in mineralized agricultural landscapes. This study focuses on the Aihui area of the northwestern Lesser Khingan Mountains (Xiaoxing'anling), China. Surface soils, ore samples, and host rocks were characterized by bulk geochemical analysis, scanning electron microscopy with energy-dispersive spectroscopy (SEM-EDS), laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS), positive matrix factorization (PMF) and random forest combined with Shapley additive explanations (RF-SHAP) analysis. Arsenic (As), copper (Cu), lead (Pb), zinc (Zn), and selenium (Se) were enriched relative to regional background values, whereas cadmium (Cd) was generally depleted at the regional scale but locally enriched near mineralized and mining-affected zones. These patterns indicate that trace-element distributions are jointly controlled by parent-material lithology, weathering redistribution, mineralized geogenic reservoirs, and mining-enhanced disturbance. The retained five-factor PMF solution resolved signals associated with coal-bearing strata, lithogenic weathering, copper-molybdenum (Cu-Mo) mineralization, mafic-to-intermediate volcanic-rock weathering or sorption processes, and silver (Ag)-polymetallic mineralization. The convergence of PMF factor structure, mineral-scale evidence, spatial hotspot patterns, and lithological context supports the interpretation that mining amplifies trace-element release and dispersion superimposed on a geogenic background. The results provide a geology-constrained framework for source discrimination and risk-oriented soil management in geologically complex terrains.