Erros na estimativa de eventos extremos de vazão usando precipitação estimada por radar

Abstract

The real-time availability and high spatial-temporal resolution of precipitation estimates promoted by meteorological radar tend to improve the estimation of extreme flood events by distributed hydrological models, especially in basins that present rapid discharges and lack of rainfall data in the appropriate temporal and spatial scale. However, such estimates continue to provide an important source of uncertainty for hydrological forecasts. In this context, this work evaluated the errors in the estimates of extreme short-term flood events through a distributed hydrological model, considering different forms of radar precipitation estimates in a tropical / semi-arid climate basin, which presents frequent flood events CAPPI and PPI products were used for radar rain estimates and the use of different Z-R and double polarization relationships. The local Z-R relations developed for Alagoas presented the best performance in rainfall-flow modeling. The CAPPI product was promising in the short-term flow estimates, while the PPI product was physically inconsistent. In addition, for the analyzed events the double polarization equation did not bring significant gains for rainfall estimation and consequently flow.