不同非空间模拟方法下CLUE-S模型土地利用预测——以秦淮河流域为例

基于快速城镇化背景下秦淮河流域土地利用历史状况,选择CLUE-S模型对其2020年土地利用情况进行模拟预测。分别使用线性回归、Markov模型、灰色GM（1,1）模型预测CLUE-S模型非空间模块的土地利用需求量,再嵌入CLUE-S中得到3种预测结果,对预测结果进行比较。另外设定“自然发展”情景与考虑规划政策影响的“优化格局”情景,模拟2020年不同情景下秦淮河流域土地利用格局情况,并进行景观格局分析。结果表明：线性回归模型、Markov模型、灰色GM（1,1）模型的Kappa指数分别为0.866、0.849、0.867,3种方法均满足模型精度要求;自然发展情景中2020年水域、水田、林地、城镇用地、旱地面积相对于2010年分别变化21.5%、-15.3%、-9.0%、51.5%、-28.9%,而优化格局情景下水域、水田、林地、城镇用地、旱地面积分别变化3.1%、-1.6%、10.8%、6.3%、-10.6%,相比于自然发展情景,优化情景土地利用状况更符合保护基本农田、增加生态用地连通性、提高雨水下渗能力以及缓解城市热岛效应的要求,为后期土地利用规划提供了依据。

Land Use Prediction Based on CLUE-S Model Under Different Non-spatial Simulation Methods: A Case Study of the Qinhuai River Watershed

Under the background of rapid urbanization in the Qinhuai River Watershed, models of land use change are primary tools for analyzing the causes and consequences of land use changes. We choose CLUE-S model to simulate the land use situation of it in 2020. We use linear regression model, Markov model and the gray GM (1, 1) model respectively to predict the demand for land use which is needed by the non-spatial module of CLUE-S model, then we compared the three forecast results.In order to further verify the influence of policy on land use change, two prediction scenarios were established, one is "natural development" scenario where land use will change according to historical trend and the other is "optimization" scenario which considered the effects of planning policy. We simulated the Qinhuai River Watershed land use pattern in 2020 under different scenarios, and analyze the landscape pattern of it. The results shows that the Kappa index of Linear regression model, Markov model, the gray GM (1, 1) model are 0.866, 0.849, 0.867 respectively, so three methods all satisfy the requirements of model accuracy; In “natural development” scenario, the water area, paddy field, forest land, urban land and the dry farm change, compared to 2010, by 21.5%, 15.3%, 9.0%, 9.0%, 9.0%, respectively, while in “optimization” scenario water area, paddy field, forest land, urban land and the dry farm change by 3.1%, 1.6%, 10.8%, 6.3%, 10.8%, respectively; Under the “optimization” scenario, the land use condition can meet the requirement of protection of basic farmland and ecological land, increasing infiltration capacity of rainwater, and alleviating the urban heat island effect. This work could be the reference for the choice of the method of non-spatial module and provide scientific support for land use planning and managements of the watershed.