عنوان مقاله [English]
The Flow Duration Curve (FDC) is a classical method used to graphically represent the relationship between the frequency and magnitude of stream flow and is required as a prerequisite for water resources management projects. In this study, by analyzing daily data of hydrodynamic stations, 47 stations with the appropriate statistics and the common period between 1976 and 2011 was selected in a semi-arid region of the country. Using a topographic map with a scale of 1: 50000 and determining the position of the stations, the study area was determined and 11 physiographic parameters influencing the flow duration curve including: average height, basin area, gravilillus coefficient, basin slope, main river length and hydrological parameters including annual rainfall, Base flow index, hydrograph recession constant, curve number, permeability and the number of rainy days were extracted for each basin. The flow duration curve indices were then extracted using daily flow data. Then factor analysis was performed and independent factors influencing the flow duration curve were determined. Finally, homogeneity was performed based on independent main factors and the regression relations of the curve indices were extracted in each homogeneous region.In order to investigate the validity and accuracy of the models in homogeneous regions, error-independent test methods, normal distribution of errors and control stations were used. The results showed that the selected factors for factor analysis in semi-arid climatic zone (75.875 percent) of the variance of data were explained. The six parameters of precipitation, curve number, slope, rainy days, permeability and area were known as the most effective parameters. The results of the accuracy assessment of the models using the control stations showed that the relative error of the relations presented in the homogeneous region was 0.17, 2.17, 2.73, 1.53 and 1.94, respectively. The normal distribution of errors, the coefficient of determination of more than 0.90 and the Durbin Watson coefficient between 1.5 to 2.5 also Nash-Sutcliff near "one" indicate the reliability of the regression relations presented for estimating the flow duration curve indices in the ungagged catchments in semi-arid areas of the country.