*3.2. Variations in PM2.5 Spatial Distribution*

The spatial distribution of RAs in annual average concentrations can be found in Figure 4. In 2016, the PM2.5 concentrations of most cities in the western (CD, MS and LS) and south parts (ZG, YB and LZ) of SCB were highest, inferred from the distribution of positive RAs in these areas. The RAs of cities in the northeast part of SCB were slightly positive (DZ and NC) or negative (GY, BZ, GA), revealing relatively lower concentrations in this area. In 2017, in total, 10 cities held positive RAs, among which 9 cities were located in the western and southern parts of SCB, except DZ in the northeast part, including CD, DY, MY, MS, YA, LS, ZG, YB and LZ. Moreover, the RAs of these cities were higher than those in 2016, resulting in larger spatial disparity in annual average concentrations in 2017, as shown in Figure 2. In 2018 and 2019, cities with positive RAs were distributed in the south (ZG, YB and LS), the northwest (CD, DY and MY) and the northeast (DZ and NC) parts of SCB. Another feature was that the RA in MS (LZ) decreased (increased) significantly from 2018 to 2019. The variations in RA distribution further suggested that spatial disparity of annual average PM2.5 concentrations was declining in SCB because the maximum RA decreased from 2016 to 2019 and the variations in RAs showed obvious regional characteristics, decreasing in the western and southern basin and increasing in the northern basin.

**Figure 4.** The distribution of RAs in annual and seasonal average PM2.5 concentrations from 2016 to 2019 in SCB.

Generally, the variation features of RAs in seasonal average PM2.5 concentrations were coincident with those of the annual averages. In winter, the RAs in the western and southern parts of SCB were much higher than those in other parts and kept decreasing from 2016 to 2019. The lowest RAs were found in the northern part of SCB, such as GY and BZ and the RAs in this area increased significantly from 2016 to 2019, especially in 2019. In spring and autumn, the RAs in GY and BZ increased more significantly than in winter, while the RAs in the western and southern parts of SCB were decreasing. In summer, the RA in CQ was higher than that in other seasons, and the RAs in the northern part of SCB maintained the values in other seasons.

In summary, the spatial disparity of PM2.5 concentrations of 18 cities in SCB narrowed and the variations in concentrations showed prominent regional characteristics, namely decreasing in the western and southern basin, maintaining in the central and eastern basin, and slightly increasing in the northern basin. These regional characteristics suggested that the meteorological conditions might be important causes, especially the increase in concentrations in the northern basin, despite strict emission control measures. Therefore, in the next section, the synoptic patterns were identified and their impacts on the distribution of atmospheric diffusion conditions were analyzed, aiming to explain the regional maintaining or increasing trends of PM2.5 concentration in the northeastern basin.
