Genetic algorithm and fuzzy neural networks combined with the hydrological modeling system for forecasting watershed runoff discharge

• DOI: 10.1007/s00521-015-1832-0
• Publication Date: 2015-02-03T05:18:14Z
• Authors: Chih-Chieh Young; Wen-Cheng Liu; Chung-En Chung
• Keywords: 13. Climate action; 0207 environmental engineering; 02 engineering and technology; 14. Life underwater; 6. Clean water

The accurate prediction of hourly runoff discharge in a river basin during typhoon events is of critical importance in operational flood control and management. This study utilizes three model approaches to predict runoff discharge in the Laonong Creek basin in southern Taiwan: the hydrological engineering center hydrological modeling system (HEC-HMS) model and two hybrid models which combine the HEC-HMS model with a genetic algorithm neural network (GANN) and an adaptive neuro-fuzzy inference system approach (ANFIS). Hourly runoff discharge data during seven heavy rainfall/typhoon events were collected for model calibration (training) and validation. Six statistical indicators [i.e., mean absolute error, root-mean-square error, coefficient of correlation, error of time to peak discharge, error of peak discharge, and coefficient of efficiency (CE)] were used to evaluate the prediction accuracy. The simulation results indicate that the HEC-HMS model cannot satisfactorily predict hourly runoff discharge during the typhoon events. Both hybrid approaches that use the HEC-HMS model in conjunction with the GANN and ANFIS models can significantly improve the prediction accuracy for the n-h-ahead runoff discharge.