Seasonal rainfall prediction, while usefull for some planning, does not have the temporal resolution needed for many applications. Information regarding subseasonal variations in rainfall are often needed for 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) a simulated using a regional climate model forced by reanalyses and also forced by general circulation model (GCM) ensemble realizations. Daily rainfall from station observations is employed to explore the subseasonal variations in rainfall, (e.g., precipitation time series, days with no rain, days with heavy rain, wet spells and dry spells) and to verify the model's ability to capture these variations. The analysis will focus on Northeast Brazil, the Amazon, and the southeast Brazil regions.