地理科学进展 ›› 2006, Vol. 25 ›› Issue (3): 79-85.doi: 10.11820/dlkxjz.2006.03.010

• 遥感与GIS应用 • 上一篇    下一篇

基于GIS和人工神经网络的区域贫困化空间模拟分析——以贵州省猫跳河流域为例

许月卿1,2, 李双成2, 蔡运龙2   

  1. 1. 中国农业大学资源与环境学院, 北京100094|
    2. 北京大学环境学院, 北京100871
  • 收稿日期:2006-01-01 修回日期:2006-04-01 出版日期:2006-05-25 发布日期:2006-05-25
  • 作者简介:许月卿( 1972- )| 女, 河北定州市人, 博士后, 副教授.主要从事土地利用变化及其环境效应等方面 研究.Email:xmoonq@sina.com
  • 基金资助:

    国家自然科学基金资助重点项目( 40335046) 和高等学校博士学科点专项科研基金资助课题 ( 20040001038) 资助.

Spatial Simulation Using GIS and Ar tificial Neur al Network for Regional Pover ty —A Case Study of Maotiaohe Water shed, Guizhou Province

XU Yueqing1,2, LI Shuangcheng2, CAI Yunlong2   

  1. 1. College of Resources and Environment, China Agricultural University, Beijing 100094, China|
    2. College of Environmental Sciences, Peking University, Beijing 100871, China
  • Received:2006-01-01 Revised:2006-04-01 Online:2006-05-25 Published:2006-05-25

摘要:

我国西南喀斯特地区山高坡陡, 地形破碎, 生境脆弱, 水土流失严重, 是我国典型的极贫困 代表区域之一。本文选择贵州省猫跳河流域作为研究区, 以乡镇为基本单元, 应用GIS 和ANN 技 术, 模拟区域自然致贫因子和消贫因子的空间分布, 计算各乡镇的贫困度, 揭示区域贫困的空间 分布格局, 以期为指导研究区早日脱贫及生态重建提供科学依据。结果表明, 地形、土壤侵蚀等自 然要素是主要的致贫因子, 而社会经济要素是缓解贫困的因子。贫困度较小的乡镇主要分布在研 究区的中部和东部, 贫困度较大的乡镇主要分布在研究区的南部和北部边缘。可见, 应用人工神 经网络模拟区域贫困化简便、实用, 避免了传统的单纯依靠统计数据进行贫困化研究的做法, 是 一种可行的方法与技术途径。

关键词: 空间模拟, 区域贫困, 人工神经网络

Abstract:

Owing to the fragile eco- environment, terrain fragmentation, and serious soil erosion, the karst area in Southwest China is one of the distinct poor regions. Selecting Maotiaohe watershed as study area, taking villages and towns as studying unit, and using GIS and ANN model, this paper simulates the spatial distribution of natural impoverishing index and socio- economic alleviating impoverishing index, calculates the poverty degree of villages and towns, and reveals the spatial distribution of poverty in order to provide scientific basis for eliminating poverty and ecological reconstruction. The results show that the natural factors such as soil erosion and so on are the main impoverishing indexes, and socio- economic factors are the main alleviating impoverishing indexes. The villages and towns with smaller poverty degree are mainly distributed in the middle and east area of Maotiaohe watershed, and those with larger poverty degree are mainly distributed in the southern and northern area of Maotiaohe watershed. The results also indicate that application of BP neural network to simulating regional poverty is convenient, precise and feasible, which can be an alternative approach to simulating regional poverty.

Key words: artificial neural network, regional poverty, spatial simulation

中图分类号: 

  • N945.1