Technical Abstract: Four sources of precipitation data representing low- to high-spatial resolution were used to calibrate the Soil and Water Assessment Tool (SWAT) for the purposes of statistically evaluating the performance of SWAT on the basis of reproducing measured streamflow, and to evaluate changes in model parameters as a function of precipitation data set used to calibrate the model. The experiment was carried in the Ft. Cobb Reservoir experimental watershed (FCREW), located in southwestern Oklahoma. Precipitation data for the watershed included the National Weather Service Cooperative weather network (COOP), the National Weather Service’s Next generation radar precipitation estimates (NEXRAD), the University of Oklahoma and Oklahoma State University’s joint state-wide weather station network (Mesonet), and the USDA Agricultural Research Service’s weather station network (Micronet) deployed in the FCREW. The FCREW was divided into three main sub-watersheds (Cobb, Lake, and Willow Creeks), with SWAT calibrated for each sub-watershed and each precipitation data set. Results indicated that correlation between the COOP data and the other data sets decreased as the distance between the sub-watershed centroid and the COOP station increased. Model simulations were generally “good” to “very good” at both the daily and monthly time steps for all precipitation data sets, except in the Willow Creek sub-watershed, which scored as “satisfactory” at the monthly time step and “unsatisfactory” at the daily time step when the COOP data were used. Model calibration parameter values within the Cobb Creek sub-watershed changed little across precipitation data sets. In the Lake Creek and Willow Creek sub-watershed, the deep recharge calibration parameter values varied greatly, implying that precipitation data source could impact model assessments of conservation practices designed to ameliorate the movement of agro-chemicals from the surface to lower positions in the soil profile and eventually into the groundwater.