Comparison of Algorithms for Incoming Atmospheric Long-wave Radiation.

While numerous algorithms exist for predicting incident atmospheric long-wave radiation under clear (L_clr) and cloudy skies, few comparisons have been published to assess the accuracy of the different algorithms.  Virtually no comparisons have been made for both clear and cloudy skies across multiple sites.  This study evaluates the accuracy of twelve algorithms for predicting incident long-wave radiation under clear skies, ten cloud correction algorithms, and four algorithms for all-sky conditions using data from fourteen sites across North America and China.  Data from five research sites were combined with publicly available data from nine sites in the AmeriFlux network.  Clear sky algorithms that excelled in predicting L_clr were the Dilley-O’Brien, Prata, and Ångström algorithms.  Root mean square difference (RMSD) between predicted and measured L_clr averaged 22 to 23 Wm⁻² for these three algorithms across all sites.  Cloud-correction algorithms of Kimball, Unsworth-Monteith and Crawford described the data best when combined with the Dilley clear-sky algorithm.  Average RMSD across all sites for these three cloud corrections was 24 to 25 Wm⁻².  The Kimball and Unsworth-Monteith cloud corrections require an estimate of cloud cover while the Crawford algorithm corrects for cloud cover directly from measured solar radiation.  Optimum limits in the clearness index, defined as the ratio of observed solar radiation to theoretical terrestrial solar radiation, for complete cloud cover and clear skies were suggested for the Kimball and Unsworth-Monteith algorithms.  Based on the results, the recommended algorithms can be applied with reasonable accuracy for a wide range of climates, elevations and latitudes.