Aplicação de técnicas estatísticas para avaliação de dados de monitoramento de qualidade das águas superficiais da porção mineira do rio São Francisco

Abstract

Regular monitoring programs of water quality are essential for the proper management of water resources. However, they generate a complex set of data, often difficult to interpret and analyze. Multivariate statistical techniques and trend tests tends to emerge as efficient alternatives to extract information from these data. The overall objective of this research was to analyze monitoring water quality data of São Francisco river basin, in Minas Gerais, obtained by the monitoring program Projeto Águas de Minas (IGAM), to identify and evaluate spatio-temporal changes in the more impacted sub-basin. The results of exploratory statistical tools used in this search allowed for comparisons to determine the more impacted sub-basin between ten Units of the São Francisco basin. The techniques used (CA, PCA, FA, Kruskal-Wallis) indicated four groups of Units, allowed to characterize water quality in the São Francisco basin, to infer some characteristics and possible sources of pollution in each of the groupings found and to identified some parameters that are most significant to separate the groups. Furthermore, they indicated that the Units SF5 and SF10 are the sub-basins with characteristics that stand out among the ten Units. Other approaches were used to compare in detail all sub-basins such as the violation of legal limits and analysis of the IQA (Water Quality Index). It was confirmed, from the numerous analyzes performed, the greater degradation of Unit SF5 (sub-basin of Rio das Velhas), using asdecision criterion the anthropogenic impacts. Twenty-nine monitoring points from this more impacted sub-basin (SF5) were studied for spatial and temporal trends. Using the median values of IQA, it was observed that the groupings of these 29 stations were consistent with pollution levels (high, moderate and low) presented by watercourses from that sub-basin. Mann-Kendall and Seasonal Mann-Kendall trend test results showed that although there have been changes in water quality over ten years in the sub-basin of Rio das Velhas, there was a concentration of stations exhibiting variations in relation to the 11 parameters studied near large urban centers, the most populous region of sub-basin with a significant economic activity.