Abstract
Medium-range streamflow forecasts, which are generated using rainfall-runoff models forced by numerical precipitation predictions, are very useful for the anticipation of hydrological events. Traditionally these predictions are deterministic, but in the last decade the movement for generating ensemble streamflow forecasts has been gaining strength. In the middle ground between these two techniques (deterministic and ensembles) the use of persistence information from deterministic streamflow forecasts for decision making, using it as an uncertainty measurement, appears to be an interesting strategy. This research investigates precisely these possible benefits of using forecasts based on persistence as uncertainty information measurement. Results suggest that the use of forecasts based on persistence have advantages over single deterministic forecasts on the higher lead times.
Keywords:
hydrological forecasting; persistence; ensemble forecasting
