Determinação de clusters de enriquecimento/empobrecimento em solos tropicais – um caso de estudo na República Dominicana

Abstract

When considering complex scenarios involving several covariates, such as in environmental characterization, a clearer reality, through the dimensional reduction of data, can be achieved. Indeed, maps work as a visualization tool of contaminant spatial patterns, allowing the identification of enriched areas. Eighteen soil samples were gathered in two Dominican drainage basins, characterized by acidic, intermediate, and to a lesser extent basic lithologies. Detrital rock outcrops are also identified, which allowed the identification of distinct geochemical signatures, in both watersheds. A set, of 19 elements (Al, Ca, Fe, K, Mg, Mn, Na, Si, Ti, P, Ba, Co, Cr, Sr, V, Cu, Ni, Pb, Zn and LOI) was measured in the soil. Principal Components Analysis (PCA) was used for dimensionality reduction and for the evaluation of the attributes’ association grade. The selected covariates underwent geostatistical modelling. Ordinary kriging and Local G clustering allowed the definition of High/Low enrichment clusters and the spatial delimitation of enrichment/impoverishment compartments.