基于GWR模型的土地利用变化模拟研究——以郧西县城关镇为例

Simulation of Land Use Change Based on Geographically Weighted Regression:A Case Study of Chengguan Town,Yunxi County

模型模拟预测是研究土地利用变化的有效方法之一。该文以郧西县城关镇为研究区域,选用土地利用模拟常用模型CLUE-S模型作为土地利用变化模拟的模型。考虑到驱动因子对土地利用格局影响的空间不稳定性,用GWR(geographically weighted regression)模型取代常用的Logistic回归模型来建立各地类与其驱动因子之间的回归模型,实现了对CLUE-S模型的改进,并进行实证研究与模型比较。结果表明:运用GWR对CLUE-S模型进行改进,可以获得更高的模拟准确率,能够为未来城镇发展提供科学的决策支持。

Model simulation and prediction is one of the effective methods to study the land use change. In this paper,took Chengguan Town of Ynnxi County as a study area,chosen CLUE-S model as the simulation model of land use change. Considering the instability factor of driving spatial effect on the land use pattern ,using GWR model instead of the global logistic regression model to establish a model between urban space and its driving factors ,realize the improvement of the CLUE-S model. The research results showed that the simulation accuracy obtain higher with the use of GWR to improve the CLUE-S model, and provided the scientific decision support for urban development.