This work compares five precipitation datasets for South America, for the period 2000-2004, considering the meridional variation, the variation across the main continental watersheds (Amazonas, Tocantins, São Francisco, Orinoco, Paraná/Prata, in addition to the main Patagônia basins) and across the main vegetation types (tropical evergreen forest, tropical deciduous forest, woodland, savanna and grassland/steppe). We compare three products based on a combination of rain gauge measurements and satellite data (TRMM, CMAP and GPCP) and two based on reanalyzed meteorological datasets (NCEP/NCAR and CPTEC). The results indicate that the annual average precipitation fields of the different products present different behavior among them. For example, the CPTEC reanalysis does not represent the main regimes of precipitation in the continent, overestimating the precipitation in the interior of northeast Brazil and underestimating the precipitation in the rest of the continent. The combined rain gauge-remote sensing products obtained similar patterns, mainly CMAP and GPCP.
Precipitation; Reanalysis; datasets