Simulation of Regional Land Use Competition for Jiangxi Province

As an important research field of global change and sustainable development, land use change has drawn much attention. Spatial dynamics of regional land use modeling can help to understand the dynamics of land use and project future land use trajectories in order to target management decisions. In this study, a spatially explicit simulation model of regional land use competition is developed to simulate regional land use change. The model is sub-divided into two distinct modules, namely a non-spatial module and a spatially explicit module. The non-spatial module is developed under the modeling framework of Computable General Equilibrium (CGE) to calculate the demand of land use change at the regional level. Economic growth, social development and ecological environment conservation are the three fundamental factors driving variation in regional land use change. Land use change demand is determined by the entire system consisting of economic subsystem, social subsystem, and ecological and environmental subsystem, since links in the comprehensive framework provided by CGE are tightly connected with each other. The spatially explicit module is a raster-based system that identifies the pixel tending to change in land use and allocate land use change demand calculated by the non-spatial module. A Multinomial Logistic Regression (MNL) model is introduced in the spatially explicit module to calculate the probability of a certain grid cell to be devoted to a land use change given a set of driving factors including location factors, topographic factors, environmental factors and soil properties. The MNL model is a regression model that generalizes a Logistic Regression (LR) model by allowing more than two discrete and unordered response variables. This method empirically describes the statistical relationships between driving factors and land use change rather than land use. Coupling the two modules, simulating regional land use competition can explicitly simulate regional land use change based on an empirical analysis of location suitability combined with the dynamic simulation of regional land use change demand. The results of comparative analysis using remote sensing data for Jiangxi Province in 2005 shows that there are 86.87% of pixels of forest and 66.26% of pixels of cultivated land correctly simulated by simulation model of regional land use competition. The total simulation accuracy of the model reached 77.68%.