Determinining a suitable subset of geochemical elements for Separation of lithological types of Poldasht waterspreading station basin

Because of many problems associated with traditional procedures for identifying sediment sources, fingerprinting techniques, based on physical, chemical and organic properties of sediment and source materials, are increasingly being used as a valuable and effective alternative approach to assemble such information. The first step involved in the application of this approach is the selection of a composite of diagnostic properties which clearly discriminate sediment sources. In this study, it is tried to choose a suitable composite of geochemical elements with the ability to discriminate lithological types in subareas of main subbasin of Pouldasht water spreading station basin, located in Makoo County, West Azarbaijan Province. The lithological maps were prepared, and lithological types as sediment sources were classified in five groups. Totally, 106 samples representing lithological characteristics, including uses, slope steepness and surface and subsurface erosions, were selected. Samples below 63 µm were separated and concentration of 18 geochemical elements was measured. Then, after being assured of the absence of outliers and insubstantial contribution of non-normality and inequality of covariance matrices to the results, the discriminant analysis was used to select the suitable subset of elements from the 18 geochemical elements. In order to eliminate the effects of multicollinearity on the discriminant analysis results, a subset (composite fingerprint) of elements, which enjoys both characteristics of least multicollinearity and highest discriminating power, was selected. The identified subset included 12 elements (Al, Co, Cr, Cu, Na, Zn, Mn, Ni, Pb, Mg, Sb and K) and was able to discriminate and classify over 87.7 percent of soil samples correctly. Meanwhile, all the 18 elements were ranked based on their relative ability to distinguish lithological groups, among which Ni was the most important and As and Sr played the least important role.