Land Use Transformation in Global Agri-Food Systems: China’s Role, Trade-Related Exposure, and Sustainable Land Use Pathways

Global agri-food systems increasingly spatially separate places of production and consumption, which is why national stabilization of agricultural land does not necessarily mean a reduction of the overall pressure on land resources. The aim of this paper is to link long-term cropland changes with crop production dynamics and international trade links. Digital paths are considered qualitatively as a supplementary interpretive layer, with a special focus on China. The primary land-use layer is based on FAOSTAT land use data for 2001–2023 and includes 218 countries and territories with complete endpoints. Production-land decoupling was analyzed in a narrower, fully reproductive sample of 140 countries. In the baseline classification, 44.3% of countries showed relative and 27.1% absolute decoupling; agreement with the baseline regimen was 95.0–97.9% with alternative thresholds and 91.4% when individual endpoints were replaced by 3-year averages. The validation sample is not compositionally neutral in relation to the 78 units without full production reconstruction; the median cropland change was +5.92% vs. −1.44% (Mann–Whitney U = 7350; p < 0.001), which is why the mode shares were not generalized to all 218 units. Between 2001 and 2023, China reduced cropland by 2.34%, while the Crop Production Index increased by 78.83%, which corresponds to absolute decoupling; the same regime is maintained in the three-year robustness specification. Among the top five sources of Chinese food imports in the WITS/UN Comtrade validation, the United States, Thailand, and Canada show absolute decoupling, while Brazil and Indonesia combine strong output growth with further cropland expansion and relative decoupling. The present results support a clear distinction between territorial production-land decoupling, trade-linked exposure, and full consumption-based land footprint. Digitization is not tested in this paper as an empirical determinant of land changes; it is considered as a potential, conditional mechanism based on the existing literature, while the empirical findings of this paper refer to patterns of land change, production, and trade-related exposure.