基于CLUE-S模型的煤矿城市土地利用变化模拟

以淮南市为研究区,选择1985、1995、2005、2016年土地利用数据,在分析土地利用动态变化特征的基础上,利用CLUE-S模型模拟预测了未来土地利用格局。结果表明：1985—2016年,研究区耕地面积减少11.62%;建设用地和水体面积百分比分别增加7.98个百分点和4.29个百分点。2005—2016年是各地类变化最强烈的阶段,其综合土地利用动态度最大,为13.46%。建设用地变化速率最快,其土地利用动态度为5.19%。土地转移主要发生在耕地、水体和建设用地之间,以耕地向建设用地和水体的转换为主。耕地转为建设用地的面积达207.61 km²,新增水体集中分布在潘谢矿区。加入空间自相关性和土壤质量因子后,耕地和建设用地的Logistics回归效果显著改善,ROC分别增加0.201和0.133。年均降水量是影响耕地变化的主要驱动因子,与耕地分布概率呈负相关;而建设用地变化主要驱动因子为GDP。土地利用模拟的Kappa系数为0.74,CLUE-S模型在研究区域具有较好的模拟能力。运用CLUE-S模型预测了研究区2028、2034、2040年土地利用空间分布,未来...

Simulation of Land Use Change in Typical Coal Mining City Based on CLUE-S Model

Huainan in Anhui Province was selected as study area. The dynamic change characteristics of land use was explored by using land use data of 1985, 1995, 2005 and 2016. And then the future land use patterns were simulated and predicted based on CLUE-S model. The results showed that from 1985 to 2016, the cultivated land area in the study area was decreased by 11.62%;the area percentage of construction land and water body was increased by 7.98 percentage points and 4.29 percentage points, respectively. From 2005 to 2016, the comprehensive dynamic degree was the largest, and it was the stage where the change of each land use type was the strongest, which was 13.46%. The change rate of construction land was the fastest, with land use dynamic index of 5.19%. Land use types mainly changed between cultivated land, water area and construction land. Cultivated land converted to construction land and water area were the dominant land use change types. The area of cultivated land converted to construction land reached 207.61km2, and the newly added water body was mainly distributed in the Panxie mining area. After adding soil quality factor and spatial autocorrelation, the Logistics regression effect of cultivated land and construction land was significantly improved, and the ROC was increased by 0.201 and 0.133, respectively. The main driving factor of cultivated land change was mean annual precipitation, which was negatively correlated with the cultivated land distribution probability;and the main driving force of construction land was GDP. Kappa index of land use simulation was 0.74, indicating that CLUE-S model had good capabilities of land use simulation in study area. On this basis, the CLUE-S model was used to predict the spatial distribution of land use in study area in 2028, 2034 and 2040. There was no significant change in the spatial distribution of land use in the future, and the area change of each land use was relatively stable.