ABSTRACT

The monthly daily average solar energy parameter measured in Umudike (Latitude 05^o29’N, Longitude 07^o33’E and altitude 122m above sea level) for a period of forty two (42) years (1972-2016) were obtained from the AGROMET Unit of National Root Crops Research Institute, Umudike. These data which include solar radiation, sunshine hour duration, relative humidity, rainfall, maximum and minimum temperatures were analyzed using both single-variable and multi-variable regression techniques of the Angstrom-Prescott type to generate several regression equations (models) that were used to estimate the total solar radiation reaching a horizontal surface in Umudike. The statistical error methods of the Mean Bias Error (MBE), Root Mean Square Error (RMSE), Mean Percentage Error (MPE) and T-test were used to test the performance of these models. The correlation coefficient (R) and coefficient of correlation (R²) were also determined for each model. Based on the highest values of correlation coefficient (R) and coefficient of determination (R²) and the least RMSE value, the equation (model 10) that show the best performance is K= -3.329+0-745ϴ +0.098T_max –0.135T_min +4.488 +8.33x10^-5R-4.0x10^-4RH-0.309ϴ. This model is recommended for designers and engineers of solar energy and other renewable energy devices in this area.