Methods in operational hydrology for real-time flash-flood forecasting need to be simple enough to match requirements of real-time system management. For this reason, hydrologic routing methods are widely used in river engineering. Among them, the popular Muskingum method is the most extended one, due to its simplicity and parsimonious formulation involving only two parameters. In the present application, two simple conceptual models with an error correction scheme were used. They were applied in practice to a mountain catchment located in the central Pyrenees (North of Spain), where occasional flash flooding events take place. Several relevant historical flood events have been selected for calibration and validation purposes. The models were designed to produce real-time predictions at the downstream gauge station, with variable lead times during a flood event. They generated accurate estimates of forecasted discharges at the downstream end of the river reach. For the validation data set and 2 h lead time, the estimated Nash-Sutcliffe coefficient was 0.970 for both models tested. The quality of the results, together with the simplicity of the formulations proposed, suggests an interesting potential for the practical use of these schemes for operational hydrology purposes.
Efficient error correcting scheme for chaos shift keying signals
