It is widely known that hydrology models are rarely capable of perfectly simulating the watershed response. The difference between the simulated response and the actual observed response is termed uncertainty and may range from very small to excessively large. The uncertainty is usually due to a combination of factors including inadequate meteorologic data, incongruities in model representation of physical processes, and errors in identifying model parameter values. To a lesser degree, errors in observing the watershed response may also contribute to model uncertainty. A quantitative assessment of the model uncertainty is a key component of determining the reliability of the model predictions and critical to using the predictions with risk-based engineering methodologies.