As an important source of pollutants in Beijing, biomass burning (BB) aerosols emitted by BB have a strong absorption and scattering effect on surface solar radiation (SSR). However, the existence of meteorological factors has a powerful impact on the estimation of the BB aerosol radiation effect. In this paper, a deweathered technique based on machine learning is presented to quantitatively isolate the influence of weather conditions from the radiative effect of BB aerosols. Furthermore, the seasonal trend of SSR is constrained by removing the general seasonal changes by the detrending method. The BB event on October 7–11, 2014, in Beijing was used as a case to validate the effectiveness of the proposed method. Experimental results show that meteorological factors aggravate the influence of BB events on SSR. Specifically, during the BB event, meteorological factors led to an additional increase of 10.9% in diffuse radiation and an additional decrease of 17.5%, 13.1%, 181.3%, 9.7% and 86.9% in global radiation, net radiation, horizontal direct radiation, reflected radiation and vertical direct radiation, respectively. After isolating the impact of meteorological factors on SSR in the BB event, we found that the average effects of BB aerosols on global radiation, net radiation, diffuse radiation, horizontal direct solar radiation, reflected radiation and vertical direct solar radiation during the BB event were −35.6 Wm−2 (−3.1 MJm−2day−1), −14.1 Wm−2 (−1.2 MJm−2day−1), 20.7 Wm−2 (1.8 MJm−2day−1), −67.3 Wm−2 (−5.8 MJm−2day−1), −8.9 Wm−2 (−0.8MJm−2day−1), and −132.9 Wm−2 (−11.5 MJm−2day−1), respectively. In conclusion, the weather normalization and detrended methods in this study provide a new idea for the accurate estimation of the impact of aerosols on SSR. This provides a quantitative reference for further understanding the impacts of atmospheric pollution on SSR.
Read full abstract