基于FLUS的长春市土地利用动态变化与预测分析

Dynamic Change and Predictive Analysis of Land Use Types in Changchun City Based on FLUS Model

研究城乡土地利用变化规律与驱动机制,有利于实现区域土地资源可持续发展。本文以长春市为例,以监督分类与人工解译相结合的方式对1997、2007和2017年Landsat卫星影像进行分类,总体精度分别为93.06%,90.70%和94.12%。1997-2017年,草地、耕地和其他土地面积分别减少354.74、922.11和55.35km^2,建设用地、水域和林地面积分别增加1 154.14、70.38和107.54km^2,整体表现为建设用地向周边扩张,侵占其他用地类型面积。利用未来土地利用模拟(future land use simulation,FLUS)模型,以2007年分类数据为基础,结合地形、交通区位和社会经济等土地利用变化驱动因子,仿真2017年土地利用格局,仿真结果与真实情况吻合较好,仿真精度达85.10%,Kappa系数为0.821 2,验证了模型和驱动因子精度可靠,符合土地利用变化趋势。以此模型因子预测2027年土地利用格局,结果表明:在城镇周围,建设用地将持续侵占耕地、林地、草地和其他土地的面积,但趋势减缓,同时林地面积和水域面积增加。

Studying the pattern and driving factors of urban and rural land use change is conducive to the sustainable development of regional land resources.Taking Changchun City as an example,combined with manual interpretation,the overall accuracy of the Landsat satellite supervised classification images in 1997,2007 and 2017,is 93.06%,90.70%,and 94.12%,respectively.From 1997 to 2017,the grassland,cultivated land and other land area decreased by 354.74 km^2,922.11 km^2,and 55.35 km^2 respectively,and the construction land, water area,and forest land increased by 1 154.14 km^2,70.38 km^2,and 107.54 km^2 respectively.The overall trend is that the construction land expanded to the periphery and encroached on the area of other land types.Based on the 2007 classification data,combined with the driving factors of land use change of terrain,traffic location and social economy,the land use pattern was simulated for 2017 with FLUS(future land use simulation)model.The simulation is in good agreement with the real results.The simulation accuracy is 85.10%,Kappa coefficient is 0.821 2,and the verification model and the driving factors are reliable and consistent with the trend of land use change.The model is used to predict the land use pattern of 2027,showing that the construction land will slowly invade the area of cultivated land,forest land,grassland and other land around the town,and the area of forest land and water will increase.