The fall webworm ( Hyphantria cunea ), a globally significant quarantine pest, has undergone rapid range expansion across China, posing substantial threats to forest ecosystems, urban landscapes, and agricultural production. Its high ecological adaptability, broad host range, and strong migratory capacity highlight the urgent need to accurately forecast future habitat suitability and spatial diffusion trends. In this study, multi-source geospatial big data were integrated to model its spatiotemporal dynamics using historical distribution records (2016–2024), climate projections under the SSP245 scenario (2021–2040), Normalized Difference Vegetation Index (NDVI), and topographic variables. A Random Forest model was employed to predict potential high-suitability habitats, while the Standard Deviation Ellipse method quantified shifts in the historical distribution centroid. Meanwhile, an XGBoost regression model was applied to the historical Standard Deviation Ellipse parameters to forecast the future geometric trajectory and diffusion direction (2025–2040), elucidating diffusion pathways and directional mechanisms. Results indicated that high-risk areas are primarily located in North China, the Huang-Huai-Hai Plain, and the middle–lower Yangtze River basin. Elevation, temperature seasonality, mean temperature of the wettest quarter, and mean temperature of the driest quarter emerged as dominant environmental drivers, while NDVI contributed minimally due to the pest’s polyphagous nature. Centroid trajectories revealed a persistent southward shift, implying heightened invasion risks for Central and Southern China under future warming scenarios. This research proposes an integrated framework that links static habitat suitability modeling with dynamic diffusion analysis, offering a novel and transferable approach for the early warning of invasive pests. The findings provide a scientific basis for risk zoning, targeted resource allocation, and the safeguarding of ecological security.
He et al. (Sun,) studied this question.