The high-resolution estimation of sulfur dioxide (SO2) concentration, health effect and monetary costs in Beijing.

Severe air pollution episodes with high SO2 loading have been frequently observed during the last decades in Beijing and have caused a noticeable damage to human health. To advance the spatiotemporal prediction of SO2 exposure in Beijing, we developed the monthly land use regression (LUR) models using daily SO2 concentration data collected from 34 monitoring stations during 2016 and 7 categories of potential independent variables (socio-economic factors, traffic and transport, emission source, land use, meteorological data, building morphology and Geographic location) in Beijing. The average adjusted R2 of 12 final LUR models was 0.62, and the root-mean-squared error (RMSE) was 4.12 μg/m3. The LOOCV R2 and RMSE of LUR models reached 0.56 and 5.43 μg/m3, respectively, suggesting that the LUR models achieved the satisfactory performance. The prediction results suggested that the average SO2 level in Beijing was 11.06 μg/m3 with the highest one up to 22.49 μg/m3 but the lowest one down to 3.86 μg/m3. The SO2 exposure showed strong spatial heterogeneity, which was much higher in the southern area than that in the northern in Beijing. The mortality and morbidity due to the excessive SO2 concentration were estimated to be 73 (95% CI:(38-125)) and 27854 (95% CI:(13852-41659)) cases per year in Beijing, leading to economic cost of 35.76 (95% CI:(16.45-54.06)) and 441.47 (95% CI:(318.31-562.04)) million RMB Yuan in 2016, respectively. This study clarified the intra- and inter-regional transport modeling of the SO2 pollution in Beijing and supplied an important support for the future air-quality and public health management strategies.