TIME SERIES ARIMA MODEL FOR PREDICTING MONTHLY NET RADIATION

Net radiation is not a climatic variable hence observed. Tedious numerical computations have been shown to characterize the methods used in its determination using data on some variables. This study aims at generating monthly synthetic net Ibadan, Benue and Kano, Nigeria Autoregressive Integrated Moving Average (ARIMA) model. performed Autocorrelation Function (ACF) Partial (PACF) analysis determining parameters of model while, residual plots Functions graphical backward predictions or estimates their respective actual values were validation. The reveals that, first difference can be represented by ARIMA (2, 1, 2) for Ibadan (1, 1) Benue. Further result showed that there significant fairly strong positive correlation between predicted across stations (p < 0.05). Lastly, Benue, Kano examined it was observed residuals within confidence intervals. affirms fact good fit.