Accuracy Assessment of a Satellite Based Method for Solar Radiation Estimation

Abstract

In this study a simple statistical model has been used to estimate hourly solar global radiation over Bangladesh for two different time periods. Only the visible channel (0.63 μm) data of NOAA-AVHRR digital images, obtained from Bangladesh Space Research and Remote Sensing Organization, has been processed to find the cloud index, a parameter which includes the horizontal extension of clouds and its optical thickness. Ground measured (pyranometer) data, obtained from Bangladesh Meteorological Department, has been used to obtain the total atmospheric transmission factor. One set of cloud index-atmospheric transmission factor data pairs has been used to find the regression coefficients using a statistical model. Once these coefficients are obtained these are then applied to the other set to calculate the ground irradiation. Ground solar global irradiation for two places has been estimated for two months through application of this method. Accuracy of the method for calculating surface insolation has been checked in two ways. Reasonably good correlation between the measured and estimated irradiation has been observed for all cases. The root mean square error (RMSE) varies between 7.99% and 13.62% and mean bias error (MBE) from -7.22% to +9.54%. The method can be used for calculation of hourly irradiation over areas in a tropical environment.