Forecasting flow discharge through time series analysis using SARIMA model for drought conditions, a case study of Jamishan River

Nowadays, water supply is more limited and providing water is more difficult due to increasing population and demand for water. Thus, due to rainfall shortage and impacts of drought, the need for forecasting monthly and annual rainfall and flow discharge through time series analysis is acutely felt. One of the key assumption in time series is their static condition. However, hydrological time series are sometimes dynamic due to independent occurrences and variations. The objective of this study was to evaluate a suitable forecast model for water flow discharge  through evaluating whether time series data of rainfall in Jamishan river, Kermanshah, Iran were static. The 25-year data of monthly rainfall (1988-2013) were used and assumed to be a dynamic series. Non-seasonal parameters of SARIMAS model were analyzed through MINTAB software. The accuracy, reliability, normality and independence of data were also evaluated. Finally, water flow discharge was modeled and predicted by SARIMA (3,0,1) x (3,0,1)12 which showed minimum mean error.