Coeficientes de desoxigenação e de reaeração superficial em trechos do rio Vacacaí Mirim

Abstract

This dissertation presents the experimental determination of the two most important coefficients in modeling of dissolved oxygen in water bodies: the reaeration-rate coefficient K2 and the deoxygenation coefficient K1. The deoxygenation coefficient was determined in three representative sections of the considered basin: the first with urban characteristics, the second mixed and the third rural characteristics. Three experiments were conducted, featuring different flows in the Vacacaí Mirim River. K1 was determined by constructing the oxygen depletion versus time curve, in a 20 days period. K1 ranged from 0.15 day-1 to 0.27 day-1. An increase in K1 value was observed due to flow increment, in mixed and rural characteristics areas, reflecting the major contribution of diffuse organic load in those areas. It was not possible to observe a significant variation of flow in the urban feature section, due to damping of the flow by the upstream reservoir. This section is considered to present a decrease in the K1 value as liquid discharge increases, since it is noticeable the sewage contribution. The reaeration-rate coefficient was determined using inert tracer technique, applying propane as gas tracer and Rhodamine WT as dye tracer. Two experiments were conducted in two reaches, both in the highest urbanization areas of the Vacacaí Mirim basin. In the first experiment, K2 value was 30.30 day-1 for the first reach, and 24.77 day-1 for the second. In the second test, K2 obtained was 14.91 day-1 and 12.73 day-1, respectively. An approach was made about alternative methods for experimental determination of the reaeration-rate coefficient and the predictive equations of K2. The results obtained in tests were compared with those predicted by the most important equations in literature. The predictive equations presented large discrepancy between them, and the latest formulas, derived from K2 measurements obtained by the gas tracer methods, presented the lowest variation, ranging from -32.95% to 14.51%. Based on results, an equation was proposed to estimate K2 restricted to the studied sites, adapting from the equations of Melching and Flores.