多元统计分析在三明烟区土壤肥力评价中的运用

Multivariate Statistical Analysis Application in Evaluation Soil Fertility in Sanming Areas

以福建三明烟区11个县(市)的17个土壤肥力因子为例,采用多元统计的主成分分析法对土壤肥力等级进行综合地评价,为烟区科学合理的施肥提供依据。借助SPSS统计软件,可以得出土壤肥力因子前4个主成分的特征值λ>1,且方差贡献率分别为46.528%,20.422%,10.631%,8.164%,累计贡献率为85.745%,已足够反映原指标信息。这4个主成分分别概括为(1)土壤P、土壤交换性Ca以及土壤微量元素因子;(2)土壤粘粒和土壤Mg素因子;(3)土壤有机质因子;(4)土壤酸碱度因子。再借助SPSS统计软件得出方差极大旋转后主成分因子得分,最后再计算主成分因子的综合得分。结果显示:大田县土壤肥力等级排名第一,其土壤综合肥力水平最高。

The text adopts multivariate statistical principal component analysis to evaluate on soil fertility＇s rank taking example for 17 soil fertility＇s factors of 11 counties （cities） in Sanming Fujian tobacco areas. Soil fertility factor＇s former 4 principal components which 〉1 and contribution rate of variance are 46.528%,20.422%,10.631%,8.164% and cumulative contribution is 85.745%,which is enough to reflect original factor＇s information. 4 principal components are respectively generalized （1） soil P soil exchangeable Ca and soil microelement;（2）soil texture and soil Mg;（3）soil organic matter;（4）soil acidity and alkalinity.It comes out varimax rotation＇s principal component score in virtue of SPSS statistical software and integrated score is calculated.The results indicate that Datian county＇s soil fertility ranks first and its soil integrate fertility is the highest of all.