In environmental prediction, uncertainty typically exists in several places such as in the model inputs, the model structure and parameters, and in observations of the system behavior. In hydrology, emphasis on prediction uncertainty has traditionally been placed on the model parameters. As new remote sensing technologies from radar and satellite become available, it becomes necessary to recognize, evaluate, and quantify the uncertainty in the model inputs. A multisensor precipitation algorithm called Quantitative Precipitation Estimation and Segregation Using Multiple Sensors (QPE SUMS) supplies up to 9 different model inputs with each having various levels of complexity. Hydrologic predictions are cast in a probabilistic framework by inputting each QPE SUMS precipitation member in the Vflo hydrologic model independently and exploring the entire parameter space. In essence, a model parameter ensemble is created and then used to evaluate the input that yields the hydrologic predictions that most likely give the observed system behavior. In addition, the diversity in the 9 model inputs allows us to approximate the uncertainty in the model inputs, which is then combined with the model parameter uncertainty to provide a first estimate of the total prediction uncertainty.