*3.2. DF and SD Trends*

Generally, DF is an important indicator of levels of diffuse radiation, which are closely related to the aerosol content in the air, and can uncover the effects of pollutants on SD measurements. The correlation coefficients between DF and SD, listed in Table 2, ranged from −0.04 to −0.62, indicating that there was a negative relationship between the DF and SD. Therefore, sunshine duration decreased when diffuse radiation increased, and vice versa.

However, the bar graph of the DF trend coefficients shown in Figure 5 provides a different perspective. Long-term trend coefficients of DF during 1981 to 2020 were both positive and negative, with no distinct relationship between SD and the DF. According to the previous analysis on seasonal trends in SD, most SD trends took 2010 as the node—first decreasing and then increasing. We recalculated the trend coefficients of DF according to the 2010 node. From Figure 5, the DF trend coefficients of all the studied stations were positive before 2010 and negative from 2011 to 2020. This indicates that the diffuse radiation component increased continuously from 1981 to 2010, and began to decrease after 2010, just contrary to the seasonal trends in SD depicted in Section 3.1.

**Figure 5.** Trend coefficients of DF from 1981 to 2020 for the ten selected stations.

The opposite trend observed between SD and DF might be closely related to anthropogenic emissions. To understand the impact of the anthropogenic emissions on the relationship between SD and DF, AOD, as an important physical property of particle pollutants, is introduced to validate the negative correlation. The spatial–temporal variation and trends in MERRA-2 AOD over China from 1980 to 2017 have been analyzed by Sun et al. [22]. High AOD values mainly appeared in agglomerations such as the North China Plain (the city cluster of Beijing–Tianjin–Hebei), Pearl River Delta (the city cluster of Guangzhou–Shenzhen–Zhuhai), Yangtze River delta (the city cluster of Shanghai–Nanjing– Wuhan) and part of the Sichuan Basin (the city cluster of Chengdu), while low AOD values mainly appeared in western China. The annual Mean MERRA-2 AOD of the whole of China showed a slight increase from the 1980s to the 1990s, a sharp increase from 2001 to 2010 and a decrease from 2010 to 2017—which is almost identical to the trends and turning points of the DF and SD. Furthermore, He et al. [23] investigated the AOD spatial–temporal distribution during 2003 to 2016 throughout China. Their statistical results showed that high AOD values mainly occurred in developed agglomerations; the overall annual mean AOD showed a declining trend of −0.0018 per year from 2004 to 2016 over the whole of China. A remarkable upward trend of 0.012 year−<sup>1</sup> from 2003 to 2008 and a significant decrease from 2008 to 2016 could be found; the turnaround appeared in 2008, which is adjacent to the inflection point of the SD and DF at around 2010. Obviously, the variation

in AOD had a good positive correlation with the DF, and an evident negative correlation with SD. The possible reason for this might be that the surged anthropogenic emissions could lead to an increase in aerosols and AOD; through scattering and absorbing, a large number of aerosols will increase the diffuse radiation in the atmosphere and reduce the direct radiation reaching the ground—thus reducing the SD. Regarding the turning point shown in the SD, DF and AOD trends, this is most likely due to the Chinese government taking a series of stringent pollution controls from around 2010.

In this study, we took annual PM2.5 concentration as an example to investigate the influence of anthropogenic emissions on the relationship between SD and DF. As the particle concentration data from throughout China before 2012 are not available, we could only illustrate the spatial distributions in annual PM2.5 concentrations during 2012 to 2020 in Figure 6—which correlate well with the trends in SD and DF after 2010 and are consistent with previous studies [22,23]. From 2012, high concentrations of PM2.5 have been distributed over the North China Plain (the city cluster of Beijing–Tianjin–Hebei), Yangtze River delta (the city cluster of Shanghai–Nanjing–Wuhan), Northeast Plain (the city cluster of Shenyang–Jilin–Harbin), Northwest Plateau (the city cluster of Xi'an–Lanzhou– Urumqi) and Sichuan Basin (the city cluster of Chengdu). All these regions are the fastest developing areas in China. However, the interannual trend in PM2.5 concentration is getting better year by year. This is likely to be closely related to the Air Pollution Prevention and Control Action Plan (2013–2017) and the Blue Sky Protection Campaign (2018–2020), implemented since 2013. According to statistical studies, the Air Pollution Prevention and Control Action Plan reduced the annual population-weighted mean PM2.5 from 62.5 μg/m<sup>3</sup> in 2013 to 44.4 μg/m<sup>3</sup> in 2017, and the Blue Sky Protection Campaign further reduced PM2.5 concentrations to 33.1 μg/m3 in 2020 [24–26]. The decrease in anthropogenic aerosols resulted in a decrease in the DF, which further increased the solar radiation reaching the ground and raised SD. Correspondingly, except for Urumqi—which had a decreasing trend of SD before 2015—the other nine cities had an overall decrease before 2010; after that, the SD trends of the cities rose gradually, which is consistent with the analysis results in Figure 5, validating the negative relationship between SD and DF related to the concentrations of aerosol particles.
