日本落叶松引种区林地棕壤肥力的多变量分析

MULTIPLE VARIABLE ANALYSIS ON THE FERTILITY OF BROWN SOIL IN THE INTRODUCTION DISTRICT OF LARIX KAEMPFERI

本研究从日本落叶松引种区林地棕壤9个主要理化性状及与之配套的17株优势木材料入手，采用多变量分析方法筛选出影响该地区日本落叶松生长的主要肥力因子为氮及磷，並据此将该林地棕壤划分为9个肥力类型。在进行聚类分析时，本研究的创新在于以任意二样本点在各主成份上的坐标值为变量，以各主成份的贡献率为权重，重新定义欧氏距离公式，对主成份坐标值聚类，使其更为合理、直观。同时采用回归分析拟合出理化性状间的多个回归方程。此研究为该林地科学施肥、速生丰产、永续利用提供了可靠的数量依据。

Based on 9 main physiochemical properties of forest BROWN SOIL and the data of heights of 17 dominant trees from the introduction district of fLARIX KAEMPFERI,nitrogen and phosphorus among the main fertility actors whict affect LARIX KAEMPFERI in the district are picked out by using MULTIPLE VARIABLE ANALXSIS,and 9 fertility types of BROWN SOIL are classified,In the cluster analysis,the creativity is to redefine EUCLIDEAN DISTANCE formula using coordinate value of any sample points at the PRINCIPAL COMPONENT as variable,using CONTRIBUTION RATE OF THE PRINCIPAL COMPONENTS as weights,and through PRINCIPAL COMPONENT CLUSTER,to make the formular more reasonable and visualized,In the meantime,several regressional equations among physiochemical properties are modeled by usingregressional analysis.This provides a reliable quantitaive basis for scientific fertilization,fast-growing and high yield,and sustained utilization.Some mathematics models which reveal physiochemical properties of forest soil are established the first time in China,which can provide references for soil scientists to study further.Minimum distance formula defined by distribution rate and many analysis methods used are also the research's character.