This research is linked to the project 'Geochemical Background of the Itacaiúnas River Basin' under development by the Vale Institute of Technology Sustainable Development (ITVDS). In Brazil, geochemical surveys have been carried out mainly by the Companhia de Pesquisa de Recursos Minerais (CPRM – Brazilian geological survey), which has executed some projects focused on the region of Carajás, the largest mineral province in Brazil and in which the Itacaiúnas River basin is located. One of these projects, titled 'Evaluation of the Potential of the Strategic Mineral Resources of Brazil', carried out in the year 2012, covered the region of Canaã dos Carajás and generated a large collection of soil geochemical data. As such data were made available, the main purpose of this dissertation is to explore as much information as possible to elaborate multi-elemental geochemical maps based on interpolation techniques and to establish geochemical background values of the elements in soils of the target region of the CPRM survey. 225 soil samples, along with 32 duplicates, were collected in that project in an area of approximately 4,500 km2. The 80 mesh (0.177 mm) fraction of these samples was submitted to aqua regia digestion and 53 elements (Ag, Al, As, Au, B, Ba, Be, Bi, Ca, Cd, Ce, Co, Cr, Cs, Cu, Fe, Ga, Ge, Hf, Hg, In, K, La, Li, Mg, Mn, Mo, Na, Nb, Ni, P, Pb, Pd, Pt, Rb, Re, S, Sb, Sc, Se, Sn, Sr, Ta, Te, Th, Ti, Tl, U, V, W, Y, Zn, Zr) were analyzed by Inductively Coupled Plasma Atomic Emission Spectrometry (ICP-AES) for the major and minor elements and Inductively Coupled Plasma Mass Spectrometry (ICP-MS) for the trace elements. The analytical data provided by the CPRM were initially submitted to a quality control by means of Residual Standard Deviation (RSD) calculation for each duplicate. Then, the simple substitution method was used for the values beyond the limit of detection, followed by uniand multivariate statistical analyzis, including descriptive statistical treatments, identification of outliers via Grubb test, normality tests (Lilliefors and Kolmogrov- Smirnov), correlation and cluster analysis. All statistical analyzis previously highlighted were conducted by STATISTICA® and SPSS® softwares. To determine the values of threshold and geochemical background and graphical representation were used the methods boxplot and cumulative frequency curve methods (Minitab® and Excel® software) and the iterative techniques 2 and calculated distribution function (freeware Visual Basic macro denominada VB Background®). For the construction of geochemical maps, the World Geodetic System 1984 (WGS84) coordinate system was adopted as standard. Spatial representations followed techniques of interpolation according to the kriging or inverse distance weighting (IDW) methods. 43 analyzed elements were submitted to the statistical tests, having been left aside Au, B, Ge, Na, Pd, Pt, Re, Se, Ta e Te, because they presented ≥90% of the analytical data below the limit of detection. Correlation analyzis revealed 14 moderate correlations (0.650> r ≥ 0.575) and 19 high correlations (r ≥ 0.650), generally consistent with usual geochemical associations in geological environments such as Cr-Ni, Cu-Ni and U-Th. The cluster analysis discriminated seven clusters, at a linkage distance of 0.8, with evident geogenic origin, as for example the group composed of Cr, Ni and Mg, typical of mafic-ultramafic rocks. The geochemical maps allowed to identify occurrences of several anomalies in soils, interpreted as being of geogenic origin. Thus, U, Th, La and Ce enrichment was found in soils derived from sub-alkaline neoarchean granites, for example Estrela, Serra do Rabo and Planalto granites, and Cr and Ni in soils from mafic-ultramafic rocks of Luanga and Vermelho complexes and the Cateté Suite. Elements typically associated with anthropic activities, such as P, Zn, Mn, Ba and Pb, as well as potentially toxic elements (EPT), such as Al, As, Bi, Cd, Co, Hg, Mo and Sn, did not generally exhibit, anomalous concentrations and, when this occurred, these appear to be due to geological conditions. Among the methods used to estimate the background values of the 43 evaluated elements, the iterative technique 2 was the one that presented the best result. In cases where this method could not be used, the values obtained by the calculated distribution function (As) and the boxplot representation (Hg and S) were adopted as background. The cumulative frequency curve showed to be a useful graphical technique in the identification of outliers and multiple thresholds. The 225 samples were grouped into five classes that revealed the close linkage between the background values and the geology of the region: 1) Sub-alkaline granites (39 samples); 2) Sedimentary and metasedimentary rocks (28 samples); 3) Gneisses and associated granitoids (70 samples); 4) Metamafic, intermediate and banded iron formation (BIF) (80 samples); 5) mafic-ultramafic rocks (8 samples). With the results obtained, it is concluded that, in the scale of the survey carried out, there is no conclusive evidence of contamination related to human activity, but rather strong evidence of a marked geogenic contribution in the soils of the study area.
|
