Regional Analysis of Flow Duration Curves Using Adaptive Neuro-Fuzzy Inference System
View/ Open
Date
2015Author
Saeid Khosrobeigi Bozchaloei
Mehdi Vafakhah
Metadata
Show full item recordAbstract
This paper uses adaptive neuro-fuzzy inference system (ANFIS), artificial neural network (ANN) and nonlinear regression (NLR) for flow duration curve (FDC) estimation at ungauged sites. For this reason, all stations existing in the Namak Lake basin of Iran were considered due to their long period of data availability and minimum human activities. In total, 33 hydrometric stations were selected and the annual FDC was determined for each station. In selecting the effective factors on FDCs, 18 parameters were extracted such as physiographical, meteorological, land use, and geological characteristics using Arc/GIS. Six factors including weighted average height (H), area (A), rangeland area (RA), drainage density (DD), permeable formation (PF), and average stream slope (SS) using principal-component analysis (PCA) were selected, which illustrate 83.54% of variation of the data. The results showed that the ANFIS has generally the lower root-mean squared error (RMSE) and higher Nash criterion than the ANN and the NLR for regional analysis of FDCs.