有序样方聚类在植被垂直划分中的应用

 本文介绍一种非等级分类方法——有序样方聚类法，并将其应用于山西绵山植被垂直带的划分，结果为：1，落叶阔叶林带(包含3个亚带：Ⅰ．落叶阔叶灌丛亚带，Ⅱ．松栎林亚带，Ⅲ．落叶阔叶林亚带)，2．寒温性针叶林带和3，亚高山草甸带。基于黄金分割法(或Fisher’s法)的有序样方聚类法，是按照样方在空间(或时间)先后出现的序号和组内相似性最大，组间相似性最小的标准，对样方进行分类，因此，所得结果是最优的。与TWINSPAN的结果比较，在绵山植被遭到严重扰动的情况下，有序样方聚类要优于TWINSPAN。

The Application of the Ordered Plot Clustering to the Vertical Belt Partitioning of Vegetation of Main Mountains Shanxi

The ordered plot clustering, belonging to the non-hierarchical agglomerative algorithms, was introduced in the vertical belt partitioning of Mian Mountains, Shanxi in this paper. The result of the vertical belt partitioning is as follows: 1. the deciduous broad leaved forest belt (800～2000m above sea level),including 3 subbelts: Ⅰ. the deciduous broad leaved shrub subbelt (800～1200m above sea level); Ⅱ. the pine and oak forest subbelt (1200～1700m above sea level) ;and Ⅲ. the deciduous broad leaved forest subbelt (1700～2000m above sea level)];2, the cold-temperate coniferous forest belt (2000～2200m above sea level) ;and 3, the subalpine meadow belt (2200～2440m above sea level).Based on the algorithm of the gold partitioning method(Fisher’s method),ordered plot clustering was carried out according to , firstly, the criterion of plot order from the first to last one, and secondly ,the criterion of the maximal similarity within each plot group (or the minimal similarity among plot groups). Thus, the classification is optimal. In order to compare with ordered plot clustering, TWINSPAN was applied to the vertical belt partitioning of the mountains. The results showed that the classification of the ordered plot clustering was superior to that of TWlNSPAN in the case of the serious disturbances of primary vegetation.