A deterministic model will produce a single hydrograph with a single value for any particular day.
On the other hand, an ensemble of streamflow values can be used to create a probability graph for various flows. Here, the gray represents a very high exceedance probability, or a good chance that the flow will be this level or higher since all the traces were above this level. The color scale shows progressively lower chances of higher flows occurring.
Now this is where deterministic and probabilistic forecasts differ. Say you would like to know streamflow conditions for a particular date of interest. If you used a deterministic approach, you would have a single value. This value is either right or wrong. However, with a probabilistic forecast, you would have a full range of flows and the associated chances of exceedance. In this case, the probabilistic forecast is neither right nor wrong. This is a verification challenge and is discussed later in this presentation.