横断山区森林土壤的数值分类

在土壤分类中综合运用多种数值分类方法可以取得较好的分类效果.本文应用了主组元分析、聚类分析,多组判别分析和最小支撑树等方法,将横断山区有代表性的30个森林土壤剖面分为六组,各组土壤间不仅在理化性质上有鲜明特征,而且与最重要的成土条件和森林植被类型基本吻合.在西南高山林区,树种与下木对土壤形成有重要作用,主要成土过程为酸性淋溶与腐殖化过程.通过选取的六个理化性质指标,主组元分析结果反映出了上述主导成土过程,而多组判别分析则进一步给予每组土壤以明确的分类域.这两种方法的结合可作为土壤数值分类的基本方法.

NUMERICAL CLASSIFICATION OF FOREST SOILS IN THE REGION OF HONGDUAN MOUNTAINS

Principal component analysis,cluster analysis,discriminant analysis and a graphtheoretical method were used for the classification of forest soils in region of the Hongduan Mountains.30 forest}oil profiles were divided into six groups based on their chemical and physical properties.The results obtained from the numerical methods well conformed to the most importanr genetic conditions of the soils,i.e.the forest vegetation types.In the high-mountain forest region of Southwest China,the major soil forming processes are eluviation (acid leaching) and humification.Results of principal component analysis of 6 physical and chemical indexes of the soils indicated that the soils grouping could reflect the major processes mentioned above,while discriminant analysis of the soils gave the distinct delimition of each soil group.The two methods in combination may be the basic methed for soil numerical classification.