Modelagem da qualidade da água no trecho de vazão reduzida (TVR) devido à instalação do aproveitamento hidrelétrico da Capim Branco I no Rio Araguari, MG

Abstract

The Capim Branco I hydroelectric is located in Araguari river, in the Brazilian state of Minas Gerais. In this hydroelectric, there is the need of keeping a minimum flow in the former river bed, because the dam position and the power plant impose a different flow. The Capim Branco Ilow flow reach (LFR) has a length of 9 km and it was specified that a minimum flow of 7 m3 s-1 should be kept in the reach. In order to raise the water level in the low flow reach, the Environmental Agency FEAM determined the construction of some intermediate dikes along the river bed. The LFR´s tributary Terra Branca receives the discharge of treated wastewater coming from the city of Uberlândia. As part of this study, water quality of the low-flow reach was modeled, in order to gain insight into its possible behavior under different scenarios (without and with intermediate dikes). The constituents modeled were: dissolved oxygen (DO), biochemical oxygen demand (BOD), nitrogen fractions, phosphorus fractions, algae and thermotolerant coliforms. QUAL2E equations were used and implemented into a FORTRAN code. The model takes into account point-source pollution and diffuse pollution (incremental load along the river reach). Uncertainty analysis was performed, presenting probabilistic results and allowing identification of the most important coefficients in the LFR water-quality model. In this study, water quality data in the low-flow reach (still without the intermediate dykes) wascollected on a monthly basis, and operational records from the wastewater treatment plant were used. The simulated results indicate, in general, very good conditions for most of the water quality parameters, but a concern on the dissolved oxygen concentrations after the introduction of the dykes, with a possibility of non-compliance with the Brazilian environmental standards. From the sensitivity analysis, the variables most influenced in the model results were the conversion coefficients (without and with dikes), the initial conditions in the reach (without dikes) and the hydraulic characteristics of the reach (with dikes).