Multivariate and principal component statistical analysis of contamination in urban and agricultural soils from north Jordan

Abstract

 A Multivariate analysis and principal component analysis was performed on 37 samples gathered from different regions in north Jordan. Twenty-eight descriptors (variables) for each sample were used in these calculations, among them were metal ion concentration, such as Pb, Cr, Co, Zn, Ca, Mg, Fe, Na, K, Al, Cu, Ni, Ti, Si and Mn. Other descriptors were pH, electrical conductivity (EC), and grain sizes. It was found that the samples form three clusters, namely, agriculture, industrial and waste treatment plane regions. Statistical analysis showed that the samples are classified into seven classes, with the majority of the samples classified into three clusters. The first three principal components explained 99.2% of the variance. This method offers a method to classify geographical regions on a pollution bases and to determine the possible sources of pollution for such regions. This study showed that some of the agricultural farms were remote from the pollution sources clusters with polluted regions such as industrial or waste treatment complexes.