Multi-elements linear discriminant analysis of herbaceous and woody plants in southwest china karst region using orthogonal partial least squares model

Abstract

The karst region in southwest China is one of world’s largest continuous karst landforms in the world, renowned for its unique landscapes and abundant biodiversity. This study collected 49 leaf samples (21 herbaceous plants and 28 woody plants) from the typical karst zone in Puding County, China, and determined the content of elements in plant leaves using ICP-OES. The concentration characteristics and discrepancy of trace elements (Cr, Cu, Fe, Mn, Pb, Sr, and Zn) in herbaceous and woody plants were analyzed employing statistical analysis models. The results revealed that there were significant differences in the concentrations of trace elements between herbaceous and woody plants. The median concentrations of trace elements in herbaceous plants and woody plants, ranked from high to low, were: Fe > Sr > Mn > Zn > Cr > Cu > Pb and Fe > Sr > Mn > Cr > Zn > Pb > Cu. The outcomes of the correlation analysis revealed discernible differences in the interactions of trace elements within the leaves of herbaceous and woody plants. Principal component analysis (PCA) indicated that Cu, Mn and Zn were influenced by different mechanisms from Cr, Fe, Pb and Sr in plant system. Orthogonal partial least squares discriminant analysis (OPLS-DA) showed that Pb and Cr had stronger distinguishing capabilities between herbaceous and woody plants compared to other elements. The OPLS-DA model was likely considered an optimized model for tracing element sources from different plant species, which has a greatly applied potential in source identification of plant-derived trace elements in a complex environment.