Variability of Trends in Precipitation across the Amazon River Basin Determined from the CHIRPS Precipitation Product and from Station Records

• Autor(a) principal: PACA, Victor Hugo da Motta
• Data de Publicação: 2020
• Outros Autores: ESPINOZA-DÁVALOS, Gonzalo E., MOREIRA, Daniel Medeiros, COMAIR, Georges
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional de Geociências - RIGEO
• Texto Completo: https://rigeo.cprm.gov.br/handle/doc/21671; https://doi.org/10.3390/w12051244
• topic: RIO AMAZONAS; PRODUTOS DE PRECIPITAÇÃO; TENDÊNCIAS DE PRECIPITAÇÃO; AMAZON RIVER BASIN; PRECIPITATION PRODUCTS; PRECIPITATION TRENDS

The Amazon River Basin is the largest rainforest in the world. Long-term changes in precipitation trends in the basin can affect the continental water balance and the world’s climate. The precipitation trends in the basin are not spatially uniform; estimating these trends only at locations where station data are available has an inherent bias. In the present research, the spatially distributed annual precipitation trends were studied in the Amazon River Basin from the year 1981 to 2017 using the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS) product. The precipitation trends were also cross-validated at locations where station data were available. The research also identifies clusters within the basin where trends showed a larger increase (nine clusters) or decrease in precipitation (10 clusters). The overall precipitation trend in the Amazon River Basin over 37 years showed a 2.8 mm/year increase, with a maximum of 45.1 mm/year and minimum of −37.9 mm/year. The highest positive cluster was in Cuzco in the Ucayali River basin, and the lowest negative was in Santa Cruz de la Sierra, in the upstream Madeira River basin. The total volume of the incoming precipitation was 340,885.1 km3, with a withdrawal of −244,337.1 km3. Cross-validation was performed using 98 in situ stations with more than 20 years of recorded data, obtaining an R2 of 0.981, a slope of 1.027, and a root mean square error (RMSE) of 363.6 mm/year. The homogeneous, standardized, and continuous long-term time series provided by CHIRPS is a valuable product for basins with a low-density network of stations such as the Amazon Basin.