地理科学 ›› 2019, Vol. 39 ›› Issue (2): 242-251.doi: 10.13249/j.cnki.sgs.2019.02.008

• • 上一篇    下一篇

基于大数据方法和SOFM聚类的中国经济-环境综合分区研究

冯喆1,2(), 蒋洪强1(), 卢亚灵1   

  1. 1.环境保护部环境规划院国家环境保护环境规划与政策模拟重点实验室,北京 100012
    2.中国地质大学(北京)土地科学技术学院,北京 100083
  • 收稿日期:2018-01-29 修回日期:2018-04-21 出版日期:2019-02-10 发布日期:2019-02-10
  • 作者简介:

    作者简介:冯喆(1984-),男,北京人,讲师,博士,主要研究方向为地理大数据。E-mail: zhefeng@cugb.edu.cn

  • 基金资助:
    国家环境保护环境规划与政策模拟重点实验室开放基金(ZDSYS201701)、国家自然科学基金项目(71603097, 71433007, 41771204)资助

China's Economic-environment Comprehensive Zoning Based on Big Data Method and SOFM Clustering

Zhe Feng1,2(), Hongqiang Jiang1(), Yaling Lu1   

  1. 1. State Environmental Protection Key Laboratory of Environmental Planning and Policy Simulation, Chinese Academy for Environmental Planning, Beijing 100012, China
    2. School of Land Science and Technology, China University of Geosciences, Beijing 100083, China
  • Received:2018-01-29 Revised:2018-04-21 Online:2019-02-10 Published:2019-02-10
  • Supported by:
    Open Foundation of Key Laboratory of National Environmental Protection Environmental Planning and Policy Simulation (ZDSYS201701),National Natural Science Foundation of China (71603097, 71433007, 41771204).

摘要:

研究使用经济和环境多源大数据,建立包含人口、GDP等经济指标和空气质量等环境指标的中国经济-环境关联体系,识别各指标的热点、冷点时空变化特征,采用人工神经网络聚类方法对中国现阶段经济-环境进行综合分区。研究结果如下: 灯光平均强度较高的省份主要集中在沿海地区,经济以长三角、珠三角和环渤海区域为主要拉动引擎,呈东南高西北低的发展态势,东部沿海地区经济发展优于东北、中部和西南地区。PM2.5浓度呈现先增后减趋势,高污染区主要集中在华北、华中等区域;东北方向逐步扩散,污染热点地区从辽东半岛、山海关一线向东北扩张;南方地区基本保持稳定态势。 采用自组织特征映射模型对2015年全国各地级市OLS灯光指数、人口、城市自然边界和年均PM2.5浓度4类指标进行聚类,第I类为经济极发达-环境恶化地区,主要位于华北平原和长江三角洲;第II类为经济发达-环境趋恶化地区,主要分布在第I类区域周边,特别是京津冀周边地区;第III类为经济发达-环境良好地区,广东、海南、江西、福建以及重庆等省市多属此类型;第IV类为经济不发达-环境优质地区,主要分布于东北地区北部、内蒙古、甘肃、贵州、新疆、青海、西藏等地。

关键词: 经济-环境关联体系, 夜间灯光指数, PM2.5, 自组织特征映射模型(SOFM)

Abstract:

In the past 40 years, China's economy has developed rapidly but has paid a heavy environmental cost. China's environment and economy have undergone tremendous changes. The imbalance in regional economic development has intensified, while the various kinds of environmental problems have taken place in different area. Therefore, it is necessary to formulate environmental policies in a targeted manner. To this end, big data methods should be applied to reveal the relationship between the environment and economy, which forms the basis of comprehensive zoning. In this study, multi-source big data such as nighttime light remote sensing, spatialized population data, urban natural boundary and simulated PM2.5 concentration was used to establish an economic-environmental linkage system. The time and spatial variation rules of economic and environment hotspots were identified by Getis-Ord index. The provincial administrative areas of China were divided into four groups by using an artificial neural network (Self-Organizing Feature Map, SOFM) clustering of both economic and environmental indicators. The results show that: 1) The provinces with higher average light intensity in China are mainly concentrated in coastal areas. The economy is mainly driven by the Yangtze River Delta, the Pearl River Delta, and the Bohai Rim. It is a developing trend that is low in the southeast, high in the northwest and the eastern coastal areas. The economic development of coastal area is superior to the northeast, central and southwest regions. 2) The concentration of PM2.5 in China shows a trend of increasing first and then decreasing. Highly polluted areas are mainly concentrated in the North China and Central China regions. In northeast China, the pollution hotspots expand from the Liaodong Peninsula and Shanhaiguan to the northeast, while the south China maintains stability. 3) Based on OLS light index, population, urban natural boundary, and annual average PM2.5 concentration in 2015, the prefecture-level cities can be grouped into four types by using the self-organizing feature mapping model. Type I region is highly economic developed and environment deteriorating areas, mainly located in the North China Plain and the Yangtze River Delta. Type II region is the economic developed and environment deteriorating areas, which is mainly distributed in the periphery of the Type I region, especially around Jing-jin-ji area. Type III region is economically developed and environmentally friendly, such as Guangdong, Hainan, Jiangxi, Fujian, and Chongqing. Type IV is economically underdeveloped and high environment quality areas, which are mainly distributed in the north of the Northeast, Inner Mongolia, Gansu, Guizhou, Xinjiang, Qinghai, and Tibet. The research results are hoped to provide reference for balancing economic development and environment conservation in different regions of China.

Key words: economic-environment system, big data, nighttime lighting index, PM2.5, Self-Organizing Feature Map (SOFM)

中图分类号: 

  • S601.9