地理科学进展 ›› 2008, Vol. 27 ›› Issue (5): 121-127.doi: 10.11820/dlkxjz.2008.05.016

• 区域发展 • 上一篇    

基于SOFM 的区域界线划分方法

郝成元1, 吴绍洪2, 李双成3   

  1. 1. 河南理工大学测绘与国土信息工程学院, 焦作454000;
    2. 中国科学院地理科学与资源研究所, 北京100101;
    3 北京大学城市与环境学院, 北京100871
  • 收稿日期:2008-06-01 修回日期:2008-08-01 出版日期:2008-09-25 发布日期:2008-09-25
  • 作者简介:郝成元( 1969- )| 男, 山东曹县人, 博士, 副教授, 中国地理学会会员, 主要从事自然地理综合研究. E- mail: haocy@hpu.edu.cn
  • 基金资助:

    河南省软科学研究( 082400440750) ; 国家重点基础研究发展计划(2003CB415101).

Study on the Method of Ar eal Differ entiation Based on SOFM

HAO Chengyuan1, WU Shaohong2, LI Shuangcheng3   

  1. 1. College of Surveying &|Land Information Engineering, Henan Polytechnic University, Jiaozuo 454000, Henan Province of China;
    2 .Institute of Geographic Sciences and Natural Resource Research, CAS, Beijing 100101, China;
    3. College of Urban and Environmental Sciences, Peking University, Beijing 100871, China
  • Received:2008-06-01 Revised:2008-08-01 Online:2008-09-25 Published:2008-09-25

摘要:

区域分异研究是人们对地理环境认知深度和自然地理研究水平的重要标志之一, 划定分区界线就成为一 项迫切而意义重大的工作, 尤其是在气候复杂、地貌多样的我国西南高原、山地组合区。云南省南部地区由于多季 风系统和大地形作用的影响, 气候复杂多样。雨季, 温暖湿润的西南夏季风给研究区西部带来大量降水, 东部雨量 少; 干季, 整个研究区主要在西风南支急流控制之下, 天气晴朗、少雨, 同时也使得植被种类及盖度差别较大。基于 研究区30 个气象台站的海拔高度、多年平均气温和降水、风速、活动积温、潜在蒸散以及MODIS- EVI 等数据, 利用 神经网络技术构建了非线性分类器, 即自组织特征映射模型( SOFM) , 对所有气象台站进行了聚类研究。结果显示, 哀牢山成为阻挡北来冷空气进入西南山地的屏障, 是我国冬季东北风和夏季西南风的分界线, 因此也成为研究区 东、西两类气候的分界线。SOFM网络应用于地形复杂、地貌多样的生态地理区域分异研究, 基本能反映不同区域之 间界线两侧的相似性和差异性, 能够揭示一个由量变到质量过程的连续性, 不失为一种较好的综合自然地理区划 方法。

关键词: "阻隔"作用, SOFM, 哀牢山, 气候复杂性, 区域分异

Abstract:

In the south of Yunnan Province in China, there are Mt. Laobie, Mt. Bangma, Mt. Wuliang and Mt. Ailao, which influenced its climate and vegetation. And the study area is commonly controlled by the Indian monsoon system and East Asian monsoon system, i.e., it is one of the main areas where the two summer monsoon systems merge. Therefore, the obvious regional difference of spatio - temporal heterogeneity of vegetation is induced by the complicated topographical terrain and the monsoon climate system, which cause various river hydrologic characteristics, soil types, vegetation types, etc. That is, these unique microhabitats formed by such kinds of combination of complicated topography and subtropical warm- humid air current are beneficial to differentiation and variation of local species to a certain extent, thus leading to even more diverse vegetation type. Clustering analysis based on self - organizing feature mapping (SOFM) network is a new unsupervised clustering method that develops from neutral networks. In this paper, the neural network has been trained to perform complex functions in various fields of application including elevation, temperature, precipitation, wind speed, active accumulated temperature, evapotranspiration potential and enhanced vegetation index (EVI) at 30 meteorological stations. The result reveals that Ailao Mountain is such a firm barrier blocking the cold air coming from northern into southwest mountain region, and such a possible boundary between summer southwestern monsoon and winter northeastern monsoon of our country that it becomes a demarcation boundary of climate sort between west and east regions. This way, SOFM network is used in areal differentiation study of ecological geographical, is a rather good comprehensive physical regionalization method. For it can reflect the similarities and differences of different areas near the basic boundaries, and reveal a continuous process from quantitative changes to qualitative changes.

Key words: Ailao Mountain, areal differentiation, barrier function, climate complexity, self- organizing feature mapping