*3.1. Variations in PM2.5 Spatial Disparity*

The CVs of annual and seasonal PM2.5 concentrations are shown in Figure 2. The annual CVs showed a general decreasing trend from 2016 to 2019, except for slightly rebounding in 2017. This indicated that the differences in annual PM2.5 concentrations between 18 cities in SCB were narrowing. Similar variation trends were observed in other regions in China, such as North China Plain [41] and Northeastern China [42]. From the seasonal perspective, the CVs were the largest in summer among the four seasons, followed by spring and winter. This was partly due to the low average concentrations in summer (Table 1). The CVs in spring and summer decreased gradually from 2016 to 2019. The CV was 0.16 in the winter of 2016, but increased to 0.21 in 2017 and then gradually decreased to 0.13 in 2019. The spatial disparity varied more obviously in autumn than in other seasons. In the autumn of 2016 and 2017, the CVs were 0.25 and 0.26, respectively, and decreased to around 0.16 in 2018 and 2019. The smallest CV in 2018 was 38% lower than the largest value in 2017. In general, the spatial disparity of PM2.5 concentration in SCB decreased from 2016 to 2019, whether in terms of annual or seasonal average PM2.5 concentrations. In addition, it is worth noting that the CVs of the wintertime average PM2.5 concentration were close to those of annual averages. This revealed that the variation in PM2.5 distribution was dominated by wintertime PM2.5 distribution in SCB.

**Table 1.** The annual and seasonal averages and standard deviations of PM2.5 concentrations (μg/m3) in SCB from 2016 to 2019.


To explore the in-depth details about the variations in PM2.5 spatial disparity in SCB, the PM2.5 concentrations of 18 cities in SCB at annual and seasonal scales from 2016 to 2019 are shown in Figure 3. From the perspective of annual average PM2.5 concentrations, a rapid PM2.5 concentration decrease of 23 μg/m3, 15 μg/m3, 18 μg/m<sup>3</sup> and 20 μg/m<sup>3</sup> from 2016 to 2019 was observed in ZG, CD, LZ and MS, the most polluted cities in SCB. The PM2.5 concentrations of moderately polluted cities decreased more slowly. For example, the concentrations of CQ fell by 12 μg/m3 during these four years. However, the PM2.5 concentrations in GY and BZ, representing relatively lightly polluted cities, showed an increasing trend of 4 μg/m<sup>3</sup> and 1 μg/m3, respectively. Hence, the decrease in PM2.5 concentrations in severely polluted cities and the maintaining (even increasing) of concentrations in moderately and lightly polluted cities together made the spatial disparity of PM2.5 concentration gradually decrease in SCB.

**Figure 2.** The CVs of annual and seasonal average concentrations of PM2.5 in SCB from 2016 to 2019.

The inter-annual variations in wintertime PM2.5 concentrations presented a similar trend to the annual concentrations from 2017 to 2019, except before 2017. The PM2.5 concentrations in most cities increased from 2016 to 2017 and severely polluted cities showed a larger increase. The rising amount was 18 μg/m3, 29 μg/m<sup>3</sup> and 19 μg/m3 in ZG, CD and MS, respectively. In the representative cities of moderately polluted cities, CQ, and lightly polluted cities, GY increased by 7 μg/m3 and decreased by 3 μg/m3, respectively. As a result, the range of concentrations in 18 cities became wider and wintertime CV increased significantly in 2017. From then to 2019, PM2.5 concentrations in ZG, CD and MS decreased significantly, by 40 μg/m3, 39 μg/m3 and 39 μg/m3, respectively. The concentration in CQ decreased only by 12 μg/m3 and the concentrations in GY and BZ rose by 5 μg/m<sup>3</sup> and 2 μg/m3, respectively.

The inter-annual variations in PM2.5 concentrations in spring and summer were consistent with the annual trend. In autumn, the concentrations decreased in all cities at a similar rate from 2016 to 2017. However, much more obvious increasing trends were observed in moderately and lightly polluted cities from 2017 to 2018. The autumntime PM2.5 concentrations in BZ, GA, and GY increased by more than 10 μg/m3. Even the concentrations in CQ increased slightly as well. The reductions in autumntime concentrations in severely polluted cities, such as ZG, LZ and MS, were smaller than those in other seasons. This led to a significant reduction in spatial disparity in autumn of 2018 in SCB. From 2018 to 2019, the autumntime concentrations in most cities decreased at similar rates, hence, the spatial disparity remained stable, as shown in Figure 2.

**Figure 3.** Inter-annual variations in (**a**) annual, (**b**) wintertime, (**c**) springtime, (**d**) summertime and (**e**) autumntime average PM2.5 concentrations of 18 cities in SCB from 2016 to 2019.
