地理学报 ›› 2015, Vol. 70 ›› Issue (3): 461-475.doi: 10.11821/dlxb201503009

• • 上一篇    下一篇

基于元胞自动机模型的土地利用变化模拟——以大连经济技术开发区为例

杨俊1,2(), 解鹏1, 席建超2(), 葛全胜2, 李雪铭1, 马占东1   

  1. 1. 辽宁师范大学 自然地理与空间信息科学辽宁省重点实验室,大连 116029
    2. 中国科学院地理科学与资源研究所,北京 100101
  • 收稿日期:2014-06-10 修回日期:2015-01-10 出版日期:2015-03-20 发布日期:2015-06-14
  • 作者简介:

    作者简介:杨俊(1978-), 男, 湖北孝昌, 博士后, 副教授, 主要从事区域地表过程、城市人居环境与地理信息系统应用研究。E-mail: yangjun@lnnu.edu.cn

  • 基金资助:
    国家自然科学基金项目(41471140)

LUCC simulation based on the cellular automata simulation: A case study of Dalian Economic and Technological Development Zone

Jun YANG1,2(), Peng XIE1, Jianchao XI2(), Quansheng GE2, Xueming LI1, Zhandong MA1   

  1. 1. Liaoning Key Laboratory of Physical Geography and Geomatics, Liaoning Normal University, Dalian 116029, Liaoning, China
    2. Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101, China
  • Received:2014-06-10 Revised:2015-01-10 Online:2015-03-20 Published:2015-06-14
  • Supported by:
    National Natural Science Foundation of China, No.41471140

摘要:

元胞自动机模型已经成为模拟土地利用变化的重要方法。传统土地模拟方法中侧重于通过分析影响土地利用变化的因素来构建预测模型,较少从土地利用类型变化及其相互作用的空间角度来关注模型构建。本文以1998年、2004年和2009年1:10000土地利用数据,利用Python语言结合GDAL与Numpy类库实现局部土地利用竞争的元胞自动机模型原型开发,并用于模拟大连市经济技术开发区1998-2009年土地利用变化模拟。研究结果:① 建立了发掘多地类之间相互作用关系的试验方法,研究适用于具有明确物理意义的多地类元胞自动机模拟模型;② 该模型具有好的模拟精度,对建设用地、农用地和林地等3种不同类型用地进行同时模拟,其对应Kappa系数分别为0.762,0.634和0.678;③ 该模型建立了研究不同种地类协调作用的基本方法,可以用于进一步研究土地利用变化地类之间驱动原理。

关键词: 局部土地利用竞争, 元胞自动机模型, 土地利用变化, 模拟, 大连市经济技术开发区

Abstract:

China has been experiencing the exceptionally drastic land-use/cover changes (LUCC) in the last decades. Many techniques focused on the simulation of the single type of land-use change. However, simulating changes between different types of land-use is much more difficult than simulating change for a single type of land-use. The key to simulations based on the Cellular Automata (CA) model is how to define the transition rules. This paper built a localized land-use competition CA model to examine land-use change rules for several land-use candidates under different localized land-use patterns. By this method, the potential transformation amount of each land-use type was calculated and an optimum transition rule was used to balance the amount of land-use changes of complicated types during the simulation period. The total area of each land-use type was predicted by Markov analysis as the global variable during the whole simulation period. This CA-Markov method was adopted to simulate LUCC in Dalian Economic and Technological Development Zone in 1998 and 2009, followed by a comparative experiment. The results show that: (1) The established test method is feasible for exploring the interaction between various land-use classes; (2) This model had high simulation precision, and the Kappa coefficients in urban area, agricultural area and undeveloped area were 0.762, 0.634 and 0.678, respectively; (3) The model proposed in this paper shows a symbolic way to studying the conversion among different land-use types, which could be used to improve the interactive relationship of LUCC.

Key words: localized land-use competition, cellular automata, land-use change, simulation, Dalian Economic and Technological Development Zone