Evaluation of a global precipitation product in the hydrological modeling of a river in the Amazon basin.

In rainfall-runoff modeling, the main input variable is precipitation, and the understanding of its temporal and spatial variation is the key for good hydrological simulation results. Conventionally, the precipitated volumes are measured by rain gauges, which are representative of its surroundings and, consequently, it is necessary to apply extrapolation techniques to obtain data in ungauged regions. However, classical techniques are based on mathematical interpolation and do not consider the physical evidence for the occurrence of precipitation. Remote sensing represents a valuable alternative to hydrological modeling due to its wide coverage, and from observations by meteorological satellites and radars, quantitative precipitation estimation is possible. In this sense, the integrated use of data from rain gauges and remote sensing has the potential to improve the accuracy of hydrological simulations. This study aims to evaluate the performance of a hydrological model in the Colider River basin (Brazil), when calibrated with a global product that provides precipitation data based on rain gauges observations, satellite and weather radar. The model used was the MGB-IPH and the data source of precipitation was MSWEP (Multi-Source Weighted-Ensemble Precipitation). Two different calibrations were performed: the first, considering only the precipitation data from rain gauges; the second, considering the precipitation estimated by the product. The comparison between the rain datasets indicates that MSWEP tends to overestimate the precipitation in most cases, except during periods of considerable drought, when it underestimates. Nevertheless, the results in the hydrological simulation were satisfactory, with the model calibrated with MSWEP presenting equivalente or slightly better performance metrics than the one with conventional data. This is an indication that the continuous development of remote sensing products can be the key to increase the reliability of tools that comprise hydrological modeling, such as forecasting hydrological events, climatic hazards and also commercialization of electric energy.

Acknowledgments: This work presents part of the results obtained during the project granted by the Brazilian National Electricity Regulatory Agency (ANEEL) under its Research and Development Project PD 6491-0503/2018 – “Previsão Hidroclimática com Abrangência no Sistema Interligado Nacional de Energia Elétrica” developed by the Paraná State electric company (COPEL GeT), the Meteorological System of Paraná (SIMEPAR) and the RHAMA Consulting company.