基于多目标决策和CA模型的土地利用变化预测模型及其应用

结合宏观用地总体需求与微观土地利用适宜性, 集成灰色预测模型、多目标决策模型、 元胞自动机模型、地理信息系统技术方法, 建立了GCMG 土地利用变化预测模型。GCMG 模型包括非空间和空间2 个模块, 非空间模块侧重依据宏观社会经济发展趋势预测研究区未 来的总体用地需求变化, 而空间模块集成多目标决策模型、元胞自动机、地理信息系统等技 术方法实现了基于土地适宜性的土地利用空间配置。运用该模型对龙海市2000-2010 年土地 利用变化进行了情景模拟, 结果表明园地和建设用地是该区域内变化最为显著的用地类型, 基本农田保护政策严格实施与否将对龙海市未来土地利用变化产生深远的影响。GCMG 模型 在龙海市的应用实例表明, 该模型将土地利用系统作为一个整体, 兼顾到区域宏观水平上的 土地利用需求与局部尺度上的土地利用适宜性, 能够较好地同时模拟不同土地利用类型以及 不同人类决策情景下的土地利用转换概率, 因而可为理解土地利用多尺度复杂系统提供一定 的帮助。

Land Use Change Simulation Model Based on MCDM and CA and Its Application

A macro-micro integrated land use change model, Grey-Cellular automata (CA) -Multi-Criterion Decision-Making (MCDM)-Geographic Information System (GIS) based model (GCMG for short) which can simulate human decision making process was proposed. The GCMG model borrows the theoretical hypothesis of CLUE-S model which supposes that regional land use change is driven by its land use requirement and the land use distribution is in dynamic balance with land use demands and regional natural resources and socio-economic conditions. The GCMG model consists of both non-spatial and spatial part. The non-spatial part, so called macro model, calculates the changes of land use demand in the future based on experiential relationship of land use and its dominating drivers using the grey model. The spatial part, also called micro model, completes the land use allocation process whose total quantity is calculated by the non-spatial part with a combined method of MCDM, GIS and CA model. In the spatial part, firstly MCDM method was used to simulate the human decision making process for land use change considering socio-economic and bio-physical conditions; the results of which was brought into conversion rule of CA model; and the integration was finally implemented in GIS to model the land use allocation. To illustrate the functioning of GCMG model and its validation, it is applied in Longhai County to simulate land use change in 2010. As one of the typical counties at coastal area of Fujian Province, great changes in land use have taken place in Longhai County over the past decades, including the garden plots expansion, town land for urban and rural housing, and land for industrial and mining purpose. Firstly the GCMG simulation results are compared with map of the actual distribution of land use in 2000 for validation. The Kappa equals to 0.93 in the simulation at 10 m×10 m grid level and has gained satisfactory results. Then the validated model is applied to simulate the land use conversion probabilities under different decision-making scenarios. The results show that the basic farmland protection policy will determine the future land use change pattern. The application of GCMG model indicated that it can both simulate the land use demand at macro level and land suitability at micro level, thus possessing the ability of studying the multi-level land use system.