Many models based on the physics of radiation transfer through the clear atmosphere have been developed (Lacis and Hansen 1974; Atwater and Ball 1978; Hoyt 1978; Bird and Hulstrom 1981a, 1981b; Davies, McKay 1982; Gueymard 1993, 2008). These models can be compared with clear-sky measured data to determine if measured data deviate significantly from “expected” data.
Satellite-based remote sensors and the development of algorithms for estimating the solar flux at the surface has also lead to the possibility of using satellite-based estimates to evaluate the performance of ground-based sensors (Perez et al. 1997, 2002; Myers 2005). An example of a web-based means of evaluating ground-based measurements with respect to estimates derived from satellite data has been described (Geiger et al. 2002).
Of course, in all cases, model inputs or flux estimation algorithms have to be relatively accurate, and there must still be tolerances on measured data to account for possible instrumentation bias errors, as well as an idea of the additional sources of uncertainty in the models and algorithms themselves (Gueymard and Myers 2007).
We urge the reader to remember, as he or she studies the rest of this book, that there are no perfect measurements. However, the field of solar radiometry has critically progressed in the last few years, resulting in significantly improved measurement quality at research-class sites. Much more work yet needs to be done to obtain better instruments, reference scales, calibrations, characterizations, and corrections for measuring solar radiation accurately in the field, and improve the data quality at the vast majority of sites that still rely on suboptimal experimental techniques.
The basic uncertainties in the best practical solar radiation data available today are still on the order of 3% in direct beam, 5% in total global horizontal, 3% ±2Wm~2 in diffuse horizontal irradiance (measured with a black-and-white or corrected all-black pyranometer), 15% to 20% in diffuse radiation measured with uncorrected all-black pyranometers behind a shadow band, and perhaps 5 to 20% in sunshine duration, for digital (including pyrheliometer) and analog (burning) sunshine recorders, respectively. For the future, we can only hope for better models through better instrumentation and improved measurement techniques.