地理研究 ›› 2013, Vol. 32 ›› Issue (4): 671-682.doi: 10.11821/yj2013040010

• 生态与土地利用 • 上一篇    下一篇

基于局部化转换规则的元胞自动机土地利用模型

宇林军1,2, 孙丹峰1, 彭仲仁3, 李红4   

  1. 1. 中国农业大学资源与环境学院, 北京100193;
    2. 中国科学院遥感与数字地球研究所, 北京100101;
    3. 美国佛罗里达大学区域与城市规划系, 美国盖恩斯威尔32611-5706;
    4. 北京市农林科学院农业综合发展研究所, 北京100097
  • 收稿日期:2012-03-09 修回日期:2012-11-15 出版日期:2013-04-10 发布日期:2013-04-10
  • 通讯作者: 孙丹峰(1971- ),男,山东人,教授,主要从事土地系统分析和资源环境信息技术。E-mail:sundf@cau.edu.cn E-mail:sundf@cau.edu.cn
  • 作者简介:宇林军(1981- ),男,博士,河北张家口人,主要研究领域为土地利用与信息技术。E-mail:yulinjun81@gmail.com
  • 基金资助:

    国家自然科学基金项目(41130526,41071146)

A cellular automata land use model based on localized transition rules

YU Linjun1,2, SUN Danfeng1, PENG Zhongren3, LI Hong4   

  1. 1. College of Resources and Environmental Sciences, China Agricultural University, Beijing 100193, China;
    2. Institute of Remote Sensing and Digital Earth, CAS, Beijing 100101, China;
    3. Department of Urban and Regional Planning, University of Florida, Gainesville, FL 32611-5706, USA;
    4. Institute of Agriculture Integration Development, Beijing Academy of Agriculture and Forestry Science, Beijing 100097, China
  • Received:2012-03-09 Revised:2012-11-15 Online:2013-04-10 Published:2013-04-10

摘要: 传统土地利用元胞自动机(Cellular automata,CA)模型基于空间同质性假设,使用全局性模型建立元胞转换规则,忽略了土地利用变化驱动因素的驱动作用在空间上的变化。以美国佛罗里达州的橙县(Orange County)2003-2009年土地利用变化为例,提出了基于局部化转化规则的CA土地利用模型,其中元胞的土地利用类型适宜性由地理加权多项logit模型(Geographically weighted multinomial logit,GWML)获得。结果表明:GWML模型较传统全局性多项logit(Multinomial logit,MNL)模型有更高的数据解释能力。基于GWML模型的土地利用CA模型能反映局部土地利用变化模式,因而较基于MNL模型的CA模型具有更高的模拟精度。所得结论对未来国内地区的研究有借鉴意义。

关键词: 土地利用变化, 元胞自动机模型, 空间异质性, 多项Logit模型, 地理加权多项Logit模型

Abstract: Spatial variations in the relationships between driving forces and land use change are ignored in traditional cellular automata(CA)based land use models which are based on assumption of spatial homogeneity.To address this issue,a geographically weighted multinomial logit(GWML)mode,developed by integrating a locally weighted regression procedure in the estimation of the multinomial logit(MNL)model,is applied in a CA land use model using a case study in Orange County,FL,USA.The results show that the GWML has higher data explanation power than conventional MNL model,and the GWML based CA land use model results in a higher simulation accuracy than the MNL based CA model because local land use change patterns are identified by the locally created transition rules. Although a region in the USA was used as a case study,the conclusion is still meaningful for future studies using regions of China as cases.

Key words: land use change, cellular automata model, spatial heterogeneity, multinomial logit model, geographically weighted multinomial logit model