Abstract
The need for adequate solar radiation is ever increasing for various applications. However there is an inadequate data of solar radiation in many countries due to the cost of instrument set up. Hence this study investigates two models for estimating solar radiation from routinely measured meteorological parameters. The data were obtained from the International Institute of Tropical Agriculture, Ibadan. The regression coefficients of the quadratic models were determined and used to estimate the global solar radiation for both forward and backward predictions. Their predictive accuracies were compared with four other models and the measured values using standard statistical error indicators. The results showed for forward as compared to backward predictions in bracket root mean square errors 1.2 (1.1); mean bias errors 1.1 (0.8) and mean percentage errors -4.8% (-2.9%) while for backward prediction 1.9 (1.7), 1.7 (1.4) and 7.9% (2.2%) measured in KJm−2day−1 respectively. A positive error value shows an over estimation while a negative value shows an under estimation. The models are versatile for estimating global solar radiation at the horizontal surface, fixing missing data and correcting outliers.