Adaptive and Differentiated Land Governance for Sustainability: Spatiotemporal Dynamics and Explainable Machine Learning Analysis of Land Use Intensity in the Guanzhong Plain Urban Agglomeration

Urban agglomerations underpin regional economic growth and sustainability transitions, yet the spatial heterogeneity and drivers of land use intensity (LUI) remain insufficiently understood. Here, we develop a high-resolution (1 km hexagonal grid) framework to map LUI dynamics and identify nonlinear drivers in the Guanzhong Plain Urban Agglomeration, China, over 2000-2020. Composite indices for human settlement (HS), cropland (CS), and forest (FS) subsystems were derived from multi‐indicator metrics, and an XGBoost-SHAP workflow was employed to quantify the relative importance, threshold effects, and interactions of eleven natural, socioeconomic, urban–rural, and locational variables. HS experienced marked intensification and spatial expansion driven predominantly by economic growth and urbanization. CS transitioned from high- to moderate-intensity cultivation in response to agricultural policy reforms and shifting from biophysical to anthropogenic controls. FS high-intensity zones contracted, reflecting the success of ecological restoration and increasing influence of precipitation and spatial isolation. Urban-rural gradient analyses revealed that HS LUI declines radially from the core, CS peaks at the fringe, and FS intensifies toward rural and mountainous areas. Explainable machine learning illuminated key nonlinearities-such as the inverted U-shaped relationship between GDP per capita and HS LUI-and critical thresholds in population density and proximity effects. These insights advocate for adaptive, subsystem-specific governance: optimizing urban growth boundaries, integrating urban-rural planning, enforcing cropland protection, and tailoring forest management. Our high‐resolution, data‐driven framework offers a transferable basis for policy innovation and sustainable land‐use planning in rapidly urbanizing contexts.