The previous techniques all have one problem in common; they are not really designed to be forecasting tools. Ensemble streamflow prediction, on the other hand, is just such a tool. The scenario-based approach uses a continuous hydrologic model and saves the current model state conditions. Once this is done, ESP has the flexibility to use short-term meteorological forecasts, historic or forecast climatic data, and future reservoir regulation.
The result is a product that shows a range of streamflow values along with an associated probability. It is important to note that ESP is not a stochastic model and that only model forcings uncertainty is captured in the probabilities.