Search

The long-term variability of direct and diffuse solar radiation at the surface remains unknown for many areas of the world due to the lack of measurements. In this study, two methods, an empirical model and a neural network, are used to estimate annual direct and diffuse solar radiation in North China for the 1959–2016 period. Results reveal that the two methods have a good performance in estimating direct solar radiation (R2 = 0.49–0.95), and the estimation accuracy of diffuse solar radiation can be improved by incorporating the influence of aerosols (R2 = 0.29–0.79). Generally, total and direct solar radiations have significantly decreased with the declining trend in sunshine duration since 1959 (p < 0.01). However, diffuse solar radiation shows a significant (p < 0.01) increasing tendency over the period of 1959–2016. The proportions of direct and diffuse solar radiation are almost equal from 1982 onward. Influenced by the atmospheric aerosols, two periods with large diffuse solar radiation values have been observed from 1959 to 1989 and from 2004 to 2016. Over the period of 2000–2016, sunshine duration and aerosols can explain 89 and 85% of direct and diffuse solar radiation variations, where aerosol concentration accounts for about 63% of changes in diffuse solar radiation. This study characterizes the long-term variability of direct and diffuse solar radiation in North China since 1959 and highlights the significance of aerosols to diffuse solar radiation from 2000 to 2016. [ABSTRACT FROM AUTHOR]