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