A comparison of three approaches to estimate daily totals of global solar radiation in Australia using GMS data

Three approaches which estimate daily totals of global radiation using data from the Japanese Geostationary Meteorological Satellite (GMS) are compared. They consist of a detailed physical model which calculates the hourly global radiation resulting from absorption and scattering processes; a simple physical model describing the bulk daily transmittance of global solar radiation; and a linear regression relating daily transmittance of global solar radiation to a satellite-determined earth-atmosphere reflectance. The models were tested on 381 simultaneous observations of daily satellite-determined reflectance and pyranometer measurements taken in seven Australian capital cities. Results show that the regression and simple physical model had comparable performance and were marginally better than the detailed model. However, the standard error of the estimates, typically at 2.6 MJ m-2 d-1 for a mean value, is considerably higher than similar studies using GOES-2 data. This error mainly is a result of the lower frequency of temporal sampling of CMS, and to a lesser extent, of location problems. Regional differences in the estimates are discussed.