MODELING THE STOCHASTIC BEHAVIOR OF THE FARS RIVERS

Historical records for rivers in Fars Province are inadequate in comparison with the design period of hydraulic structures. In this study, time series techniques are applied to the records of three Iranian rivers in the Fars Province in order to generate forecast values of the mean monthly river flows. The autoregressive models (AR), moving average models (MA) and autoregressive moving average models (ARMA) are fitted to the stationary series and the optimum model for each river is formulated. Data generation is done and the synthetic sequences are tested individually against the corresponding historical data, and the optimum length of synthetic data is specified. Statistical tests including means, standard deviations, spectral diagrams and Hurst's coefficient are also provided