Simulation and Multi-Scenario Prediction of Cultivated Land Spatial Patterns Based on Multi-Source Data

Abstract

In recent years, rapid urbanization and industrialization have imposed increasing pressures on cultivated land resources, manifested primarily in reductions in cultivated land area, intensified spatial fragmentation, and degradation of ecological functions. Scientifically predicting land use change and evaluating its implications for cultivated land protection are therefore essential for supporting territorial spatial planning and sustainable land management. As an important grain-producing area in Sichuan Province and a core part of the Chengdu-Chongqing Twin-City economic circle, Neijiang requires evidence-based support for cultivated land conservation and spatial optimization. Based on multi-source datasets, this study develops an integrated Markov–PLUS–CoCA multi-model framework that links land-demand prediction, spatial allocation, and model comparison. The framework is applied under three scenarios—natural development, urban expansion, and ecological protection—to assess cultivated land spatial patterns, fragmentation, and ecosystem service value. The results show that: (1) from 2019 to 2023, cultivated land in Neijiang is generally characterized by small plots and numerous patches, and localized spatial concentration. The PLUS-based contribution analysis identified precipitation, distance to rivers, slope, and distance to roads as the dominant predictors associated with cultivated land expansion, while nighttime light intensity shows the highest contribution to construction land expansion. (2) Comparative model evaluation indicates that the CoCA model performs better than the PLUS model in terms of spatial prediction accuracy and degree of spatial agreement, suggesting that it can more effectively capture future spatial evolution patterns of cultivated land. (3) In future land use, under the Urban Expansion Scenario (UEs), cultivated land fragmentation tends to intensify, patch connectivity declines, and ecosystem service value decreases, while under the Ecological Protection Scenario (EPs), cultivated land fragmentation is significantly reduced, patch connectivity is enhanced, ecosystem service functions are improved, and the total cultivated land area remains relatively stable. The proposed framework provides methodological support for cultivated land protection and spatial optimization.