SOFM network land types physical regionalization transect of

  本文关键词：基于SOFM神经网络模型的土地类型分区尝试——以青藏高原东部样带为例，，由笔耕文化传播整理发布。

基于SOFM神经网络模型的土地类型分区尝试——以青藏高原东部样带为例

Zoning by land types based on SOFM network： A case study on transect of eastern Tibetan Plateau

[1] [2] [3] [4]

ZHANG Xueru, ZHANG Yili, LIU Linshan, ZHANG Jiping （1. Institute of Geographic Sciences and Natural Resources Research, CAS, Beijing 100101, China; 2. University of Chinese

[1]中国科学院地理科学与资源研究所,北京100101; [2]中国科学院大学,北京100049

文章摘要：基于土地类型自下而上的自然区划能够确立更加清晰的自然区划界线,是自然区划研究取得突破的关键。以青藏高原东部山区为研究区,采用神经网络模型与GIS技术,开展基于土地类型自下而上的区划研究。通过计算得到研究区地形综合指数、温暖指数、湿润指数、地被指数和水文指数5个自然指数指标,并将这些指标作为变量输入层,输入到建立的Self-Organizing Feature Maps神经网络模型中,对土地类型单元自下而上合并,生成青藏高原东部山区自然区划图,实现以土地类型单元为控制本底的定量化分区。结果表明：①可以将土地类型单元聚合成高原高寒稀疏植被区、高原高寒草甸草原区、高原高寒灌丛草甸区、高山深谷灌丛草甸区和高山深谷针叶林区5个自然带区域。②分区结果与中国生态地理区域划分的自然界线比较接近,相似性较高,分区结果较理想。

Abstr：The bottom-up physical regionalization based on land type units can establish more clear boundaries, which is dramatic breakthroughs of research in physical regionalization. The study area is located in the mountainous areas of eastern Tibetan Plateau. The research on bottom-up physical regionalization based on land type units is implemented by means of a combined method of SOFM network models and GIS. Topographic index, warmth index, humidity index, LCI index and hydrological index can be seen as input layer variables, and land type units and natural zones are regarded as background and regionalizing target. The results show that： （1） the land types are aggregated into five natural regions, namely alpine sparse vegetation plateau, alpine meadow plateau, alpine shrub-meadow plateau, alpine shrub-meadow in mountain-valleys, and coniferous forestland in mountain-valleys. （2） The boundaries of this zoning are close to scheme of ecological regionalization in China. Therefore, the results of network classification show that there is high concentration at spatial scale, which represents natural geographical characteristics in the mountainous areas of eastern Tibetan Plateau. This study can provide new ideas and methods for bottom-up physical regionalization based on land type units.

文章关键词：

Keyword：:SOFM network land types physical regionalization transect of eastern TibetanPlateau GIS

课题项目：国家自然科学基金项目（40901057,40771206）; 国家重点基础研究发展计划（2010CB951704）致谢：感谢中国科学院地理科学与资源研究所郑度研究员、申元村研究员和李秀彬研究员对研究工作的指导.

  本文关键词：基于SOFM神经网络模型的土地类型分区尝试——以青藏高原东部样带为例，由笔耕文化传播整理发布。