The Application of Linear Models for Estimation of Global Solar Radiation using Available Meteorological Parameters for Sokoto, Nigeria

Abstract

Solar radiation data base is an essential pre-requisite for designing, sizing and performance evaluation of any solar energy systems in any part of the globe, even though these data base are not readily available in many localities of both developed
and developing countries such as Nigeria. However, modeling techniques can be used to significantly evaluate daily global
solar radiation, using other more accomplished available meteorological data. In this study, sunshine hours, relative
humidity, minimum and maximum temperatures data for Sokoto location (latitude 13.030N, Longitude 5.270E and Altitude
296m) have been obtained from Nigerian meteorological Agency (NIMET) for a period of six years (2007-2012). The data
was subjected to eleven different linear models in which the coefficients were obtained using least square error method. The
values of global solar radiation estimated by the models were compared with the measured values and tested using some
statistical parameters (MBE, NDMBE, MAD, NDMAD, MPE, RMSE, R2 and t-statistic). The values were generally found to be in close agreement with measured values for the site. However, when only sunshine hours is available among the meteorological parameters, then Angstrom-Prescott model can be used as it shows reasonable agreement with the measured data. It performed better than the models that combine sunshine hours with temperatures; but excellent performance can be achieved by using the models that include sunshine hours, temperatures and natural logarithms of relative humidity. Thus, the resultant correlations an  linear regression relations developed may be used for other locations of similar meteorological/geographical characteristics and can also be used to estimate the missing data of solar radiation for the respective sites with existing weather stations.