A semi-empirical model for estimating surface solar radiation from satellite data

Abstract

This paper presents a semi-empirical model for estimating surface solar radiation from satellite data for a tropical environment. The model expresses solar irradiance as a semi-empirical function of cloud index, aerosol optical depth, precipitable water, total column ozone and air mass. The cloud index data were derived from MTSAT-1R satellite, whereas the aerosol optical depth data were obtained from MODIS/Terra satellite. The total column ozone data were derived from OMI/AURA satellite and the precipitable water data were obtained from NCEP/NCAR. A five year period (2006-2010) of these data and global solar irradiance measured at four sites in Thailand namely, Chiang Mai (18.78 °N, 98.98 °E), Nakhon Pathom (13.82 °N, 100.04 °E), Ubon Ratchathani (15.25 °N, 104.87 °E) and Songkhla (7.20 °N, 100.60 °E), were used to derive the coefficients of the model. To evaluate its performance, the model was used to calculate solar radiation at four sites in Thailand namely, Phisanulok (16.93 °N, 100.24 °E), Kanchanaburi (14.02 °N, 99.54 °E), Nongkhai (17.87 °N, 102.72 °E) and Surat Thani (9.13 °N, 99.15 °E) and the results were compared with solar radiation measured at these sites. It was found that the root mean square difference (RMSD) between measured and calculated values of hourly solar radiation was in the range of 25.5-29.4%. The RMSD is reduced to 10.9-17.0% for the case of monthly average hourly radiation. The proposed model has the advantage in terms of the simplicity for applications and reasonable accuracy of the results. © 2013 AIP Publishing LLC.