USE OF PRINCIPAL COMPONENT ANALYSIS FOR GROUPING SOIL SAMPLES BASED ON PARTICLE SIZE AND CHEMICAL AND MINERALOGICAL CHARACTERISTICS

Abstract

There are several factors that determine the horizontal and vertical variation in the chemical, physical and mineralogical soil attributes. The objective was to apply the technique of principal component analysis (PCA) to verify the similarity of soil samples in an area, based on particle size and their chemical and mineralogical characteristics, and establish relations with the source material, soil`s depth and antropic interferences. Samples were collected at three depths (0-10 cm, 10-20 cm and 20-40 cm) from eight different locations within an area of extraction and processing of heavy metals in the city of Adrianópolis-PR. Texture and the attributes of soil fertility were determined. The clay fraction was studied by X-ray diffraction, thermal and chemical extractions with ammonium oxalate and dithionite-citrate-bicarbonate. Based on the results, two PCA were generated, one with particle size and chemical data and another with mineralogic data. The first showed that the parent materials associated with the incorporation of heavy metal metallurgy wastes in soil profiles were determinant in grouping the samples. The second PCA separated allochthonous soil samples (colluvium sediments), but given the similarity of clay mineralogy, couldn’t not separated soils with different degrees of development (Entisols and Inceptisol). The concentration of hematite, contrasting to the contents of goethite, in addition to gibbsite levels, were predominant in the grouping of the samples. Thus, the use of PCA was recommend to discriminate soils in the conditions of the present study.