Seasonal rainfall prediction, while useful for some planning, does not have the temporal resolution needed for many applications. Information regarding sub-seasonal variations in rainfall are often needed for sectors such as water management and agriculture. In this light, daily rainfall data from seasonal integrations with a nested modeling system for South America are examined and evaluated against observations. The two extreme seasons of 1983 (dry) and 1985 (wet) are simulated using a regional climate model model (RegCM2) driven by reanalyses and also by general circulation model (GCM) ensemble realizations. Daily rainfall from station observations is used to define spatial (% days having rainfall) and temporal characteristics (daily time series, frequency distributions) of rainfall and to verify the model’s ability to capture these characteristics.