Monthly Rainfall Prediction Using ARIMA and Gene Expression Programming: A Case Study in Urmia, Iran

Precipitation is an essential parameter of the hydrological cycle that is known as the most important climatic variable in water resources management. In this study, monthly rainfall data obtained from Iran Meteorological Organization from ۱۹۵۱ to ۲۰۱۷ were used to model monthly rainfall in Urmia. To this end, two different methods, namely ARIMA and gene expression programming (GEP) were used. For ARIMA and GEP modeling, ۲۵ and ۱۰ different patterns were considered, respectively. The patterns were made by changing the delay in the rainfall time series. In both methods, %۸۰ of the data was used for training and the remaining %۲۰ was used for testing the models. Comparison of the models performances was done through three statistical indicators namely: root mean square error (RMSE), the coefficient of determination (R۲) and mean absolute error (MAE). The results showed that ARIMA model, which provides the lowest MAE and RMSE and the maximum R۲ is slightly superior to the GEP.

Keywords:

ARIMA , Gene expression programming , rainfall forecasting , statistical analysis