Impacts of satellite-based precipitation datasets on rainfall-runoff modeling of the Western Amazon basin of Peru and Ecuador

• Autores: Zubieta, R., Getirana, A., Espinoza, J.C., Lavado-Casimiro, W.
• Formato: artículo
• Fecha de Publicación: 2015
• Institución: Servicio Nacional de Meteorología e Hidrología del Perú
• Repositorio: SENAMHI-Institucional
• Lenguaje: inglés
• OAI Identifier: oai:repositorio.senamhi.gob.pe:20.500.12542/54
• Enlace del recurso: https://hdl.handle.net/20.500.12542/54; https://doi.org/10.1016/j.jhydrol.2015.06.064
• Nivel de acceso: acceso cerrado
• Materia: Hydrological modeling; Precipitation dataset; Satellite; Amazon Basin

Satellites are an alternative source of rainfall data used as input to hydrological models in poorly gauged or ungauged regions. They are also useful in regions with highly heterogeneous precipitation, such as the tropical Andes. This paper evaluates three satellite precipitation datasets (TMPA, CMORPH, PERSIANN), as well as a dataset based only on rain gauge data (HYBAM), and their impacts on the water balance of the Western Amazon basin, a region where hydrological modeling and hydrological forecasting are poorly developed. These datasets were used as inputs in the MGB-IPH hydrological model to simulate streamflows for the 2003-2009 period. The impacts of precipitation on model parameterization and outputs were evaluated in two calibration experiments. In Experiment 1, parameter sets were separately defined for each catchment; in Experiment 2, a single parameter set was defined for the entire basin. TMPA shows overestimated precipitation over the northern region, while CMORPH and PERSIANN significantly underestimate rainfall in the same that region and along the Andes. TMPA and CMORPH lead to similar estimates of mean evapotranspiration (~2. mm/day) for different regions along the entire basin, while PERSIANN is the least accurate (~0.5. mm/day). Overall, better scores for streamflow simulations are obtained with Experiment 1 forced by HYBAM and TMPA. Nevertheless, results using the three satellite datasets indicate inter-basin differences, low performance in the northern and high in the southern regions. Low model performances are mainly related to scale issues and forcing errors in small basins over regions that present very low rainfall seasonality.