Abstract: Taking Lake Dianchi Watershed as study area, the dynamics change degree of land use types in the watershed was analyzed based on interpretation data from TM image in 1990 and 1999. The direction of land use conversion was quantitatively analyzed by land use transfer matrix based on the Markov model. Meanwhile, land use change tendency of 2008 was simulated and forecasted based on land use conversion rules using multi-criteria evaluation method from GIS and combining with CA-Markov model. Compared with the interpretation data of 2008, Kappa coefficient of the simulated results was 0.7338, indicating that the simulated results were credible. The spatial patterns of land use change in 2017 and 2026 were simulated by using CA-Markov model. The results show that land use change is also active in this watershed and ecological restoration and reconstruction is still confronted with pressure. Farmland, forest, water and unused land will reduce from 2008 to 2026, while the grassland and construction land will continue to increase. Compared with the trend of land use from 1990 to 2008, the simulated result indicate that farmland, forest and construction land from 2008 to 2026 will keep the same change trend while grassland, water and unused land show the opposite change trend. Construction land will expand on the basis of the present and the change of water is little. Although forest resources will decrease in future, it is still the main land use types in watershed. Construction land expansion significantly increases the load of non-point source pollution in Dianchi. Unreasonable land use is one of the major causes of the water quality degradation in Dianchi Lake currently, and it is going to increase the pressure on the water environment and ecology of water basin in the future. Therefore, to maintain enough arable land per capita, lower fertilization strength, improve land utilization levels, keep a sufficient amount of ecological land, use construction land frugally and intensively, are some important measures to reduce the potential of non-point source pollution in the area. The research can provide scientific support for rational planning and management of land use, the policies formulating of ecological restoration and economic development.

Key words: land use change, CA-Markov, MCE, dynamic simulation, Lake Dianchi Watershed

摘要： 以滇池流域为例, 基于1990和1999年TM遥感影像解译数据, 分析流域土地利用类型的动态变化程度, 并基于马尔可夫模型定量分析土地利用转化的方向。采用GIS中多标准评价(MCE)方法, 结合元胞自动机?马尔可夫(CA-Markov)模型, 对2008年土地利用变化进行模拟和预测, 预测结果与实际解译结果的Kappa系数达 0.7338, 表明预测结果可信。在此基础上, 运用CA-Markov 模型对2017和2026年的土地利用空间格局变化进行模拟, 结果表明, 该流域土地利用变化快, 生态恢复和重建的压力很大。2008?2026年, 耕地、林地、水域和未利用地面积将有所减少, 草地和建设用地面积将迅速增加; 与1990?2008年相比, 草地、水域和未利用地的变化趋势相反, 而耕地、林地和建设用地变化趋势相同。建设用地在原有基础上往外扩展, 水域变化则较小。林地资源虽然有所减少, 但仍然是流域主要的用地类型。建设用地扩张明显增加了滇池非点源污染的负荷。土地利用不合理是造成目前滇池水质退化的主要原因之一, 并将加剧未来流域的水环境和水生态压力。因此, 保有足够的人均耕地, 降低施肥强度, 提高土地利用水平, 保留足够量的生态用地, 节约集约利用建设用地, 是降低该区域非点源污染潜力的重要措施。研究结果可为流域土地利用合理规划管理、生态恢复和经济发展政策的制定提供科学支持。

关键词: 土地利用变化, 元胞自动机?马尔可夫模型, 多标准评价方法(MCE), 动态模拟, 滇池流域