Performance assessment of empirical models for prediction of daily and monthly average global solar radiation: the case study of Ibadan, Nigeria

In this study, the performance assessment of empirical models for modelling global solar radiation in Ibadan is presented. The empirical models are derived from the three basic models: Angstrom–Prescott model, Garcia model and Hargreaves–Sammani model. The data used in the analysis consist of daily average global solar radiation, daily average sunshine hours, daily average maximum temperature and daily average minimum temperature collected over a period of nine years (2000–2008). Regression constants are determined for each of the model for each months of the year. The study reveals that Garcia Quadratic model puts up the best overall performance. The model can predict the daily average global solar irradiation with Mean Absolute Error of 1.86 MJ m−2 day−1, Root Mean Square Error of 2.7 MJ m−2 day−1, Mean Absolute Percentage Error 9.34% and Coefficient of Determination (R2) of 0.68.