Parametrização da propagação superficial do escoamento em bacias brasileiras

Abstract

Due to the scarce or non-existent hydrometeorological monitoring in Brazilian urban basins, rainfall-runoff modeling is used in the estimation of flows and hydrographs for the design of hydraulic and urban drainage structures. Among the methods used to simulate this rainfallrunoff transformation, the Curve Number (CN) for determining the volume of runoff, together with the Triangular Unit Hydrograph (HUT-NRCS) for the propagation of runoff, stand out. For the application of HUT-NRCS, parameters related to the basin response time to precipitation are necessary, such as the concentration time (tc), and the peak factor (PRF), which determines the shape and distribution of the hydrograph and is usually taken as a standard constant of the method equal to 484. Usually, these parameters are estimated by empirical equations and parameterizations adjusted for specific regions of the USA and Europe. Thus, this work aims to improve previous studies on tc in the Alto da Colina watershed, located in Santa Maria, Rio Grande do Sul (RS), and develop new results about the PRF of HUT-NRCS in the Vacacaí River basin, located in the central region of RS, and in sub-basins of the Arroio Dilúvio in Porto Alegre (RS). Regarding tc, values were estimated from observed hydrographs, and these results were compared with thirty empirical equations. A variation in values was observed according to precipitation events, however, without correlation with rainfall volume or with the preceding 5-day accumulated precipitation. However, higher tc's were verified during winter and fall, when the rains in the state of RS tend to be frontal, and lower tc values in spring and summer, when precipitation tends to be more intense, originated by the Mesoscale Convective Complexes. For all empirical methods of tc calculation, errors of at least 65% were observed, reaching 190% in some cases. As for the PRF analyses, values for 70 observed rainfall and flow events were calibrated using the HEC-HMS model. It was observed that PRF's lower than the standard presented lower errors in peak flow for most of the basins, and the use of PRF 484 can generate errors of up to 62% in the studied regions. However, no correlations were observed between the PRF and physical characteristics of the basins or the analyzed events. Therefore, the existing parameterizations do not allow the adequate calculation of tc and PRF of HUT-NRCS for Brazilian basins, with the need for further studies related to these parameters in regions with similar physical and climatological characteristics.