Stochastic Modeling of Monthly River Flow Forecasting (Case study: Atrak River Basin, Iran)

Abstract

In order to analyze the hydrologic phenomena, a mathematic model of the stochastic hydrologic system to simulate the phenomena must be formulated. The three approaches of stochastic modeling which account for the effect of seasonality in different ways, i.e. autoregressive integrated moving average (ARIMA), Thomas-Fiering and spectral analysis models were used to model monthly flow time series of five different hydrometric stations in the Atrak river basin, northeastern Iran. The error estimates of forecasts from the three approaches were compared to identify the most suitable approach for the reliable forecast. The results show that the ARIMA model provides the most accurate forecasts  recommended for forecasting of monthly river flow in particular stations.