Integrated spatiotemporal dynamics and spatial prediction of medicinal quality of Aconitum tanguticum based on species distribution models and machine learning approaches

Aconitum tanguticum is an important component of traditional Tibetan medicine. Owing to its high medicinal value, rising market demand, and the difficulty of artificial cultivation, wild populations have experienced long-term overharvesting and habitat fragmentation. Consequently, substantial variation in herb quality has emerged across production regions, underscoring the urgent need for a systematic and scientifically grounded quality-assessment framework. Here, we conducted an integrated evaluation of A. tanguticum by combining ecological niche modeling, quantitative profiling of active compounds, multivariate statistical analysis, and machine-learning algorithms across three dimensions: habitat suitability, chemical-composition variability, and environmental drivers. High-suitability areas were concentrated in the eastern Qinghai–Tibet Plateau, including Qinghai, Sichuan, and Gansu, with precipitation identified as the primary determinant of species distribution, followed by elevation. Five major active compounds—atisine, hordenine, quercetin, kaempferol, and total flavonoids—are significantly enriched in suitable habitats; however, their concentrations vary markedly among production regions. Soil properties—particularly total exchangeable bases, available phosphorus, total phosphorus, organic carbon, clay content, exchangeable sodium percentage, and moisture content—play dominant roles in regulating active-ingredient accumulation. By integrating neural-network-based quality-classification models with habitat-suitability results, we delineated five functional zones, with the core zone concentrated in eastern Qinghai. This integrative framework provides a robust scientific basis for the conservation, cultivation management, and sustainable use of A. tanguticum .