Evaluating Time Series Models to Estimate Monthly Temperature of Iran's Old Synoptic Stations During 1977-2005

Noting the temperature effect on climate of any region and its importance in environmental planning, using statistical methods to study and predict the changes of temperature has a wide application. Statistical methods are considered as useful and efficient tools to evaluate and understand the climate’s behaviors. The ARIMA family models can be mentioned as a group of the widely used statistical models, which are capable of modeling and forecasting the climatic data, based on their past behaviors. In the present study, initially different time series’ models were fitted to the monthly temperature data of Iran considering the existence and lack of trend and evaluating the autocorrelation and partial autocorrelation functions for the period 1977-2005. Then the optimum model was selected using the AIC and SBC criteria for each station. The results showed that annual time series’ models can be used for simulating and predicting the monthly temperature parameter.