半干旱盐渍化地区果园土壤盐分离子相关性研究

Correlation of Soil Salt Ions in Orchard in Semiarid Saline Regions

以甘肃省秦安县兴国镇郑川村果园土壤为研究对象,利用统计分析的方法对500个土样盐分离子含量的相关性进行了研究。结果表明:该研究区土壤阴离子主要是HCO3-、SO42-,阳离子主要是Na+。CO32-与HCO3-、Mg2+之间存在着显著的负相关,与Na+之间存在着极显著的正相关;HCO3-与Cl-之间存在着极显著的负相关,与Mg2+之间存在着极显著的正相关;Cl-与SO42-、Ca2+、Mg2+、Na+之间存在着显著的正相关,其中与Ca2+、Na+之间存在着极显著的正相关;SO42-与Ca2+、Mg2+、Na+之间存在着极显著的正相关;Ca2+与K+之间存在着极显著的正相关,与Na+之间存在着显著的负相关;Mg2+与K+之间存在着显著的负相关;K+与Na+之间存在着极显著的负相关。通过对盐分离子的聚类分析发现,Cl-、Mg2+、CO32-、Ca2+、K+和HCO3-在该区土壤中有着相似或相近的运移特征。对相关分析和偏相关分析这两种分析方法所得结果进行了对比分析,它们之间存在着明显的差异,偏相关分析能进一步消除由于其他变量的影响,因而能够更加准确地刻化多元变量其中的两个变量之间的相关性。

The object of this study was orchard soil in Zhengchuan Village of XinGuo Town in Qin＇an County of Gansu Province, statistical analysis was applied to explore the correlation of soil salt ions of a test site in semiarid saline area, 500 samples collected from 0 - 120 cm layer of soil were analyzed in laboratory. The results showed that the predominant anions in soil salt were HCO3^- and SO4^2- while the predominant cation was Na^＋. CO3^2- showed a significantly negative correlation with HCO3^- and Mg^2＋ but an extremely significantly positive correlation with Na^＋; HCO3^- showed a significantly negative correlation with Cl^- but an extremely significantly positive correlation with Mg^2＋; CI showed a significantly positive correlation with SO4^2-, Ca^2＋, Mg^2＋ and Na^＋ and an extremely significantly positive correlation with Ca^2＋ and Na^＋; SO4^2- showed an extremely significantly positive correlation with Ca^2＋, Mg^2＋ and Na^＋; Ca^2＋ showed an extremely significantly positive correlation with K^＋ but a significant negative correlation with Na^＋; a significant negative correlation existed between Mg^2＋ and K^＋, an extremely significant negative correlation existed between K^＋ and Na^＋. Similar characters of movement of Cl^-, Mg^2＋, CO3^2-, Ca^2＋, K^＋ and HCO3^- were found by clustering analysis. The analytical results of correlation analysis and partial correlation analysis were compared and significant differences were found between the two methods. Partial correlation analysis could eliminate more the effect of others variable, thus could depict more exactly the correlation of multivariate analysis between two variables.