In order to further improve the accuracy of fine particulate matter (PM2.5) source apportionment results, a hybrid source apportionment approach (CTM-RM) combining the capabilities of a receptor model (RM) and chemical transport model (CTM) was developed. The CTM-RM method was evaluated and applied according to a typical PM2.5 pollution process from January 21 to 27, 2019 in Chongqing. The average value of square prediction error based on CTM-RM was 84.58% lower than that of CAMx/PSAT during the campaign. Compared with that of CAMx/PSAT, the fractional error of PM2.5 and its chemical component concentrations decreased by 15.69%-92.86%. Furthermore, the temporal and spatial variations in PM2.5 source impacts could be obtained using the CTM-RM method in Chongqing. The average adjustment factor (R) values were 1.39±0.38 (agriculture sources), 1.54±0.48 (industrial sources), 1.01±0.13 (power sources), 1.02±0.58 (residential sources), 0.86±0.59 (transportation sources), and 0.58±0.67 (other sources) in the main urban areas of Chongqing. Additionally, the cumulative distribution functions of R were found to be distinct among the six sources. The residential and industrial sources were the main sources of PM2.5, with contributions of 46.23% and 28.23%, respectively. In contrast to that of the other sources, the transportation source impacts of PM2.5 (8.62%) increased significantly from the clear period to pollution period (P<0.001), indicating that the increase in PM2.5 concentrations was mainly driven by vehicular emissions during the pollution period in the main urban areas of Chongqing. The fitting functions between the initial simulated concentrations and R values of each source in the main urban areas of Chongqing could be used to evaluate PM2.5 concentrations at 47 air quality monitoring stations in Chongqing, and the correlation between the refined simulated concentrations and measured concentration of PM2.5 was significant (r=0.82, P<0.001). Compared with that during the clear period, the increases in the percentages of industrial source impacts of PM2.5 in Northeast Chongqing and residential source impacts of PM2.5 in Southeast Chongqing were 17.20% and 9.15% higher, respectively, than that in other areas during the pollution period. By contrast, the increasing percentage of transportation source impacts of PM2.5 in the main urban areas of Chongqing (66.39%) and Western Chongqing (84.16%) from the clear period to the pollution period were higher than that in other areas. The results of CTM-RM on January 26 indicated that the residential source impacts in Northeast Chongqing (64.56%) were higher than those in other areas, and the industry source impacts of PM2.5 were primarily observed in the main urban areas of Chongqing and Western Chongqing, with contributions of 25.26% and 21.20%, respectively.
Keywords: Chongqing; PM2.5; hybrid source apportionment; model evaluation and application; temporal and spatial variation.