摘要
本文收集四十个盐碱土分析样品,经过一系列前处理工作,旨在应用聚类分析进行盐碱土基层单元分类的偿试,为此,计算了样品间的距离系数,变量间的相关系数,建立了相应的聚类谱系图。在一定的相似水平下,根据样品和变量间的亲疏关系,把四十个样品的原七个土属聚为五个,其中有十三个样品的土属发生了变化,占样品总数的32.5%,亚类级的变化仅有两个样品,占样品总数的5%。在选取的九个变量中,在0.525的相似水平下分为四组,即:CO_3^-、HCO_3^-、PH;Na^+、SO_4^-、Cl^-、全盐量;Ca^(++);Mg^(++)。以此反映盐碱士变量间的亲疏关系。
The present paper deals with the analytical data of forty profiles from saline-alkalisoil.Through a series of tretments.We have pre- liminarily stydied the application of the the cluster analysis method to the basic units of classification in saline-alkali soil.Therefore,the correlatione coefficients among the variables and the distance coefficients among the Samples are calculated.At same time,the cluster spectral figures are estab- lished correspondingly.The abstracts are as following: 1.According to the coefficients of distance matrix among the samples, the sevengenera of the original classified method can be changed into the five under 0.08 distance level. 2.By means of the cluster analysis,it is shown that the thirteen samples are alternated in genera about 32.5% of total samples and the two samples are alternated in the sub-group about 5% of total Samples. 3.Under 0.525 opproximated level the nine variables can be divided into the four groups.(1)CO_3^-、HCO_3^-、pH;(2)Na^+、SO_4^-、Cl^-, TSC;(3)Ca^(++); (4)Mg^(++)。These reflect the relationships among the variables. 4.The correlation analysis showed that the absolute values of corre lation coefficients between ions and pH values can be arranged from large- to small as following: CO_3^->HCO_3^->TSC>Na^+>SO_4^-8>Cl^->Ca^(++)>Mg^(++)
