Evaluation of a novel fuzzy method and a conceptual model for a long-term daily streamflow simulation
Shahraiyni Taheri, Hamid; Ghafouri, Mohammad Reza; Saghafian, Bahram; Bagheri Shouraki, Saeed
HEC-HMS (Hydrologic Modeling System) has been designed by HEC (Hydrologic Engineering Center) for simulation of precipitation-runoff processes in a drainage basin. In this study, HEC-HMS as a conceptual model is compared with an Active Learning Method (ALM) as a novel fuzzy model for long term simulation of daily streamflow in Karoon III basin, Iran. The daily discharge data from 1991 to 1996 and from 1996 to 1999 were utilized for 'calibration and training' and 'validation and testing' the model approach, respectively. In addition, HEC-HMS was implemented using its new loss module (SMA, Soil Moisture Account). The following statistical error measures were determined for the validated conceptual model: the Nash-Sutcliffe (model efficiency) 0.8, Bias 36 m3s-1, R2 0.84, MPAE (Mean Percent of Absolute Error) 36.5% and PTVE (Percent of Total Volume Error) 11.3%. These error measures for the ALM model with 32 fuzzy rules were Nash-Sutcliffe = 0.75, Bias = 1.3 m3s-1, R2 = 0.75, MPAE = 13%, and PTVE = 0.46%. Results of this study demonstrated acceptable streamflow simulation by HEC-HMS (with SMA module) and ALM for the continuous streamflow modeling. In addition, training of the ALM is easier and more straight forward than the training of other artificial intelligence models, thus ALM appears suitable to be introduced as a new modeling method for streamflow simulation.