The differentiation of soil types and mineralization from multi-element geochemistry using multivariate methods and digital topography

Abstract

Multi-element geochemical data can be effectively interpreted through the application of multivariate statistical techniques, imaging methods and integration with digital topographic information. These techniques have been applied to a suite of 1665 soil samples collected in a sampling program from the central Sumatra area of Indonesia. The selected samples were analyzed for Au, Cu, Pb, Zn, As, Sb, Ba, Ca, Cd, Co, Cr, Fe, Ga, K, La, Li, Mg, Mn, Nb, Ni, Sc, Sr, Ti, V, Y, Zr and Hg using aqua-regia digestion followed by ICP–IES determination. Plan maps of individual elements proved difficult to interpret with several elements displaying bimodal populations. The spatial patterns of individual elements appeared to be discontinuous and raised suspicion that the bimodal population reflected differences in sample media rather than features related to lithology or mineralization. Statistical methods were employed to test the hypothesis of differences in sample media types. Principal components analysis identified several distinct element associations and populations. The scores of the samples for each principal component were interpolated, imaged and plotted as plan maps. The first principal component was difficult to interpret in plan view and did not appear to reflect any lithological or known mineralization trends. It was concluded that the patterns associated with the first component may reflect differences in sample media types (saprolite and volcanic ash). The second component revealed a pattern associated with saprolitic soils reflecting elevated Cu concentrations and is coincident with the regional northwest trending fault structures that are parallel to the Great Sumatra Fault Zone. The third component was attributed to Au associated with the saprolitic soils possibly reflecting local epithermal processes associated with zones of dilation and the regional faulting. A digital elevation model with a resolution of 10 m was integrated with the results of the principal components analysis. The two populations of samples that occur within the first principal component show distinct spatial characteristics. The population of samples, interpreted to be volcanic ash, occurs along hill tops slopes and drainage channels and overly the saprolitic soil profile. This results of this study indicate that given a suitable multi-element set of data, the application of multivariate techniques and the integration with digital topography provides significantly more information than can be determined by evaluating single elements and standard plan-view maps. The application of these techniques also provides a more effective way of visualizing and interpreting multi-element geochemical data.