Comparative analysis of gridded precipitation products and the development of a blended product in the Andes and surrounding regions

Abstract

Accurate precipitation estimates are essential for land surface models and the management of water resources. However, traditional gauge-based precipitation measurements are limited by spatial and temporal coverage and are prone to errors and uncertainties. Gridded precipitation products derived from reanalysis data and satellite observations offer a more continuous and comprehensive alternative, though their performance varies regionally and is often uncertain. This study evaluates the performance of five widely available gridded precipitation products (MERRA-2, ERA5, CHIRPS, MSWEP, and IMERG) across the Andes and surrounding regions. Using gauges and the Extended Triple Collocation (ETC) method, we assess the accuracy and reliability of these products at daily, seasonal, and annual scales over the period 2001–2023. Our results indicate that MERRA-2 has the lowest errors when compared with gauges, followed by MSWEP and ERA5. CHIRPS shows a high probability of detecting precipitation events but tends to overestimate, resulting in higher errors. IMERG outperforms others in the ETC analysis, with the lowest errors and highest correlation coefficient. Recognizing the lack of a single best-performing product, we developed MIMERG, a new gridded product that combines MERRA-2 and IMERG through advanced interpolation techniques based on the Braseth algorithm. MIMERG demonstrates improved performance over MERRA-2 in ETC analysis and enhances IMERG’s alignment with gauges. This study highlights the importance of tailored precipitation product selection and the potential for innovative combinations like MIMERG to improve precipitation estimates, ultimately supporting more accurate environmental modeling and water resource management.