数理统计分析在土壤养分评价中的应用  

Application of mathematical statistical method in evaluation of soil nutrients

作　　者：朱青[1] 陈正刚 陈欣[1] 王兆骞[1] 王文华 

机构地区：[1]浙江大学生命科学学院,浙江杭州310029 [2]贵州省土壤肥料研究所,贵州贵阳550006

出　　处：《西南农业学报》2006年第5期847-852,共6页Southwest China Journal of Agricultural Sciences

基　　金：国家自然科学基金重点项目(30030030);钾磷研究所及加拿大钾磷研究所资助项目(Guizhou200101)

摘　　要：以贵州贫困地区为例,依据土壤系统研究法土壤养分分级指标,采用主成分分析-系统聚类分析决策优化模型,对该贫困地区的72个旱作土壤的14项指标进行合理的数据转换,构造土壤养分状况指标筛选模型,聚类分析评价该区的土壤养分状况。其结果显示,新筛选出的AA,OM,pH,N,K,Ca,Mg和P 8个土壤养分状况指标只为原指标的60%,但足以代表该区土壤养分状况信息,从而可减少40%的分析工作量,节省财力物力和时间。新划分的4大类别土壤为今后该地区的土地合理规划及土壤保育和利用提供了科学的依据。The 14 indexes of 72 upland soil samples from the poverty area are rationally changed the date by the principal component analysis and hierarchical clustering analysis according to the grading standard of soil nutrients,which are used to build the model for soil nutrient criterion selection and to investigate the soil nutrient status.The results show that the new soil nutrient indexes of AA, OM, pH, N, K, Ca, Mg and P are only about 60 % of the original indexes, which can represent the information of the soil nutrients. Therefore, the new mathematical .statistical method can save 40 % of work amount,cost and time.The four new classified kinds of soils can provide the useful information for land planning and managemeat in the tested area.

关 键 词：土壤养分状况 主成分分析 系统聚类分析 贫困地区 贵州 

分 类 号：S151.9[农业科学—土壤学]