Tempo de concentração e parâmetros da metodologia da curva número: uma avaliação para bacias brasileiras

Abstract

Among the available models for design flows estimation, the Curve-Number (CN) method in conjunction with the Triangular Unit Hydrograph Method (HUT) is highlighted. These are widely used methods, but there are also criticisms and doubts. In the CN method, the biggest difficulty is the choice of its CN parameter and there is contestation of the initial abstraction ratio (λ) defined as 0,2. For HUT, the time of concentration (tc) is an indispensable parameter and with uncertainty in the estimate. Errors in these parameters can influence the design of hydraulic structures. Therefore, for the six basins of the Arroio Dilúvio and the Alto da Colina basin, this study aimed to estimate tc, λ and CN, from observed data and through empirical methods or preestablished values, and to verify the simulated hydrograph with these information. Based on these data, the tc was determined and the estimation errors were verified through thirty methods. Also, from the data, λ and CN were determined and compared with the value of λ defined by the SCS and the estimated CN based on tables. The results showed that, in general, the tc of the data presented great dispersion and correlation with the intensity of the precipitation. For the tc estimated with the methods, the best overall result was with the use of Kerby and had an average error of 57%, compared to the tc of the seven basin data. For rural basins the Simas-Hawkins method produced the best tc results, along with SCS lag and Desbordes for one of these basins. In urban basins, for each case, different tc methods were the best option, with errors ranging from 35% in Bela Vista to 65% in Alto da Colina. From the data we obtained λ for the surface runoff obtained by three separation methods, and in all cases its median was lower than the 0,20 defined by the SCS. For almost all basins it was evident that λ = 0,05 is the most realistic value, except for Beco Carvalho and Arroio do Meio which had higher results when using the concave flow separation method. From the data, the CN was calibrated to λ = 0.2 and λ observed. For λ = 0.2, in all basins the calibrated CN was superior to that determined by the use of tables, demonstrating that the tabulated CN should be used with caution, because it underestimated the runoff. Finally, the calibrated CN with observed λ has provided a more realistic value so that it may be useful for designers or future studies. With tc, λ and CN, the hydrograph were simulated by the CN-HUT method. For the estimated and pre-established tc, λ and CN, the generated hydrographs underestimated the volumes in relation to the observations. With tc, λ and CN observed, there were better results, especially for the tc calibrated as a function of the observed tc. However, for some events, the results were still unacceptable and future studies peak factor rate and recession factor rate of the HUT are needed.