With the dividing the plots into 25 square samples(20 m × 20 m),Multivariate regression trees(MRT) method was used to class the small Qinling Mountains forest communities and using the Correspondence analysis(CA), Detrended correspondence analysis(DCA),Redundancy analysis(RA)and Canonical Correspondence analysis(CCA) were used to sort the small Qinling Mountains forest communities. The results show that:MRT classification of 25 plots,the cross validationbased on the principle of nomenclature and classification of plant communities ,plant communities in this region can be divided into 3 categories;DCA sort explicitly reveal the habitat distribution range of each community types,and it reflect the relation between the forest communities of small Qinling Mountains nature reserve and environmental factors;If combining the classification and ordination to analyze the community characteristics,the effect would be better.%设置25个20 m ×20 m样地，用多元回归树（Multivariate regression trees，MRT）方法对小秦岭森林群落进行分类，采用对应分析（Correspondence analysis，CA）、除趋势对应分析（Detrended correspondence analysis，DCA）、冗余分析（Redundancy analysis，RDA）和典范对应分析（Canonical Correspondence analysis，CCA）方法对小秦岭森林群落进行排序，并比较了4种排序方法的优劣。研究结果表明，依据植物群落分类和命名原则，对25个样方进行多元回归树分类，本区植物群落可分为3类；样方除趋势对应分析排序明确地揭示各群落类型生境分布范围，较好地反映小秦岭自然保护区森林群落与环境因子的关系；分类与排序结合起来分析群落特征效果更好。
展开▼
