地球信息科学学报 ›› 2020, Vol. 22 ›› Issue (3): 580-591.doi: 10.12082/dqxxkx.2020.190404

• “土地利用模拟”专栏 • 上一篇    下一篇

城镇-农业-生态空间划定的多情景模拟

柯新利*(), 肖邦勇, 郑伟伟, 马艳春, 李红艳   

  1. 华中农业大学公共管理学院,武汉 430070
  • 收稿日期:2019-07-28 修回日期:2019-12-14 出版日期:2020-03-25 发布日期:2020-05-18
  • 通讯作者: 柯新利 E-mail:kexl@mail.hzau.edu.cn
  • 作者简介:柯新利(1977— ),男,湖北天门人,博士,教授,主要研究方向为土地资源优化配置。E-mail:kexl@mail.hzau.edu.cn
  • 基金资助:
    项目基金:国家自然科学基金项目(41371113);国家社会科学基金项目(13CGL092)

Urban-agricultural-ecological Space Zoning based on Scenario Simulation

KE Xinli*(), XIAO Bangyong, ZHENG Weiwei, MA Yanchun, LI Hongyan   

  1. College of Public Administration, Huazhong Agricultural University, Wuhan 430070, China
  • Received:2019-07-28 Revised:2019-12-14 Online:2020-03-25 Published:2020-05-18
  • Contact: KE Xinli E-mail:kexl@mail.hzau.edu.cn
  • Supported by:
    National Natural Science Foundation of China(41371113);National Social Science Foundation of China(13CGL092)

摘要:

城镇—农业—生态空间划定(简称“三区”划定)是国土空间规划的核心内容,对于科学合理地规划、利用有限的国土资源具有重要的意义。已有研究主要根据区域内土地利用与社会经济发展现状构建指标体系进行“三区”划定,较少将未来土地利用变化纳入“三区”划定过程中,使得划定结果在指导实践过程中缺乏前瞻性。针对这一问题,本文提出一种基于土地利用情景模拟,结合指标体系评价与决策树特征挖掘的“三区”划定方法,并以武汉市2015年土地利用现状为基础,在土地利用变化情景模拟的基础上进行“三区”划定。通过对比,验证了本文提出方法的合理性。研究发现:① 不同情景下的“三区”空间在规模、空间分布上具有明显差异,将未来土地利用变化纳入“三区”划定过程中确有必要;② 不同土地利用情景下“三区”空间的差异主要出现在三类空间的交界区域,这些区域是国土空间规划应该关注的重点区域。

关键词: 国土空间规划, 三区划定, LANDSCAPE模型, 土地利用变化, 情景模拟, C5.0决策树, 数据挖掘, 武汉市

Abstract:

The urban-agricultural-ecological space zoning (referred to as the "three zones") is the core content of land space planning. It is important for scientific and rational planning and using of limited land resources. The former researches set up various indicator systems for "three zones", mainly based on the current regional land use and socio-economic development status. But they rarely incorporate future land use changes into the "three-zone" delineation process, making the results less forward-looking in guiding practice. To make up the gap of current research, we propose a new "three zones" delineation method in this paper. It is based on the simulation of land use scenarios and combines the advantages of indicator system and decision tree data mining, which is different from the traditional method of "indicator system - weighted comprehensive value". This paper selected Wuhan as the research area, and used this method to explore the spatial differences of the "three zones" under different land use scenarios in the future. We simulated four land use scenarios of Wuhan in 2035 (Natural development scenario, Farmland protection scenario, Ecological protection scenario and Balance development scenario) based on the land use of Wuhan in 2015. After that, we combined the constructed indicator system and multiple textual references to select typical samples for decision tree training. Then using the classification rule set generated by the decision tree (86.4% average accuracy) to identify the "three zones" spatial categories of the research units. Finally, we obtained the "three zones" distribution under different land use scenarios. Compared with the similar former researches, the method proposed in this article is more reasonable and feasible and can be used in specific research and practice. Besides, we found that: (1) there are obvious differences in the area, spatial distribution and main change types of the "three zones" caused by different land use scenarios. So, it is indeed necessary to incorporate future land-use changes into the "three zones" delineation process. (2) The spatial distribution of "three zones" in different land use scenarios shows similar characteristics. The differences in spatial distribution of "three zones" in different land use scenarios are mainly the border area where the main land use function changes. Therefore, these areas are the key areas that land space planning should focus on.

Key words: land space planning, "urban-agricultural-ecological space" zoning, LANDSCAPE model, LUCC, scenario simulation, the C5.0 decision tree, data mining, Wuhan City