Multi-scenario simulation and carbon stock prediction for production-living-ecological spaces in Ordos City based on the PLUS-InVEST model

Introduction To support dual carbon (Carbon peak and carbon neutrality) goals, this study examines land use impacts on carbon storage in the Yellow River Basin’s semi-arid region using Ordos City as a case study. It explores Production-Living-Ecological Spaces (production, living, ecological) evolution patterns through multi-scenario simulations for regional ecological sustainability. Methods CLCD30 m high-resolution multi-year land use data were used to identify four types of Production-Living-Ecological Spaces. A driver index system for Production-Living-Ecological Spaces changes was constructed, and multiple machine learning algorithms were applied to analyze spatial suitability. The PLUS (Patch-generating Land Use Simulation) model simulated future multi-scenario Production-Living-Ecological Spaces patterns. The InVEST (Integrated Valuation of Ecosystem Services and Trade-offs) model carbon storage analysis module, incorporating biomass and soil carbon emission correction coefficients with carbon density parameters for different spatial types, quantitatively assessed the impact of land use transitions on carbon storage. Results The overall accuracy of the Production-Living-Ecological Spaces simulation reached 0.91, with a Kappa coefficient of 0.83. Ecological spaces were identified as primary carbon sinks, while county-level carbon sinks exhibited differentiated characteristics between “area-based” and “efficiency-based” patterns. Discussion This framework effectively reveals the spatiotemporal variation in land-use succession patterns and carbon storage under different scenarios. It provides a reference for further research in the same direction.