Key points are not available for this paper at this time.
Abstract Precipitation plays a vital role in various fields, including hydroclimatic modeling, climate change studies, agricultural optimization, and water resources management. Precipitation data can be obtained through observational measurements using the rain gauge approach or as Gridded precipitation products (GPP) derived from satellites or atmospheric models. GPPs provide optimized global estimates of climate data without spatial or temporal gaps, making them a valuable solution for areas with sparse or nonexistent rain gauges. However, it is essential to assess their reliability and limitations across different time scales and regions before usage. This study aims to evaluate the accuracy of two specific GPP datasets, ERA5 and MERRA-2, in comparison with two observational datasets, focusing on the Tocantins-Araguaia watershed and Pará river estuary in Brazil. The results show that both GPPs, ERA5 and MERRA-2, captured the overall precipitation regime for the analyzed period. However, discrepancies emerged, particularly at the daily and annual scales, with better agreement observed at monthly and climatology scales when compared to observational datasets. ERA5 demonstrated a higher number of acceptable stations compared to MERRA-2. Although both reanalysis products showed good agreement in climatological analysis, a more detailed evaluation revealed shortcomings in simulating precipitation during the dry season. While GPPs offer consistent time series with higher temporal and spatial resolutions, the observational precipitation data is deemed the most suitable input for hydrological-hydrodynamic modeling in the Tocantins-Araguaia watershed. Its widespread coverage, numerous rain gauges, and accurate representation of reality make it an ideal choice for hydrological modeling in the region.
Pereira et al. (Sat,) studied this question.