*2.3. Back-Trajectory Analysis*

Backward trajectory analysis essentially follows a parcel of air backward in hourly time steps for a specified length of time [29]. The HYbrid-Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model developed by the National Oceanic and Atmospheric Administration (NOAA) was used to identify the potential source area of air pollution in a specific city and capture the vertical movement of the air masses from the sources to the receptor inside planetary boundary layer [30].

The HYSPLIT model was used to investigate the movement of air masses during a heavy particulate pollution observed in winter 2017. In order to understand the impact of the regional transmission process, three-dimensional 48 h backward trajectories arriving at 500 m above ground level (AGL) of the receptor sites were also calculated using 1◦ × 1◦ Global Data Assimilation System (GDAS) data from National Centers for Environmental Prediction (NCEP). Based on the Euclidean distance between the motion trajectories, the Ward layering method was used to assign the motion trajectories to different clusters according to their moving speed and direction. The hour with the highest PM concentration of coarse particles in Chengdu was selected as the start time of the trajectory, and the backward trajectory at a height of 500 m from 3 January to 6 January 2017 was calculated. The main transportation routes that caused severe air pollution in the winter of 2017 were identified by combining the trajectory with the corresponding average concentration of pollutants [31].

#### **3. Results**

#### *3.1. Spatio-Temporal Characteristics of the Air Quality*

The annual average concentrations of the pollutants in Sichuan Basin were determined by averaging the effective data from all stations. Their values are shown in Figure 2. The annual mean concentrations of CO, NO2, SO2, O3, PM2.5 and PM10 in the entire basin area ranged from 0.67–0.90 mg/m3, 24.33–30.4 μg/m3, 8.41–17.76 μg/m3, 80.08–91.4 μg/m3, 31.2–46.56 μg/m<sup>3</sup> and 47.87–75.19 μg/m3, respectively. During the same period, in Chengdu and Chongqing, two megacities of Sichuan Basin, the annual average concentrations of the six pollutants ranged from 0.69–1.08 (0.79–1.1) mg/m3, 33.75–49.46 (37.18–45.5) μg/m3, 6.56–15.75 (7.49–16.17) μg/m3, 87.86–101.56 (68.98–81.54) μg/m3, 39.23–61.85 (32.27–54.42) μg/m3, 61.27–103.83 (51.85–84.12) μg/m3, respectively.

The annual average SO2 concentration was the lowest in Bazhong, with a value of 4.71 μg/m3, and the highest in Panzhihua, with a value of 33.69 μg/m3. The lowest NO2 concentration of 9.21 μg/m<sup>3</sup> was observed in Aba Prefecture, and the highest in Chengdu, reached 43.73 μg/m3. The highest annual average CO concentration observed at 1.49 mg/m<sup>3</sup> in Panzhihua, and the lowest of 0.48 mg/m<sup>3</sup> was observed in Ganzi. The lowest annual average concentration of O3 was observed in Ya'an during 2015, which was 53.2 μg/m3. The highest annual average concentration of O3 was observed in Zigong during 2018, with a value of 105.19 μg/m3. The annual average concentration of PM2.5 and PM10 in Aba Prefecture was the lowest, 15.12 μg/m3 and 26.38 μg/m3, respectively. The

highest values of 84.94 μg/m<sup>3</sup> and 109.6 μg/m<sup>3</sup> was observed in Zigong, both of which appeared in 2015.

**Figure 2.** The annual average concentrations of pollutants in Sichuan Basin. The filled circle represents the mean concentration whereas the error bar denotes the range of the annual concentration. The data used in each city comes from every station during 2015–2020.

As shown in Figure 3, the concentrations of five pollutants other than O3 all showed a downward trend from 2015 to 2020 in Sichuan Basin. The concentration of NO2 was the highest in 2017, at 30.4 μg/m3, and the lowest in 2020, at 24.33 μg/m3, with an average annual decline rate of 2.72%. The highest concentrations of CO, SO2, PM2.5 and PM10 all appeared in 2015 with the value of 0.9 mg/m3, 17.76 μg/m3, 46.56 μg/m3 and 75.19 μg/m3, respectively. The lowest concentrations all appeared in 2020 with the value of 0.67 mg/m3, 8.41 μg/m3, 31.2 μg/m3, 47.87 μg/m3, and the average annual decline rates were 5.14%, 10.52%, 6.59%, and 7.27%, respectively.

At present, the environmental concentration of most air pollutants in China is declining, but the concentration of secondary pollutants such as O3 is increasing at both provincial and capital city levels [32,33]. Previous studies showed that the rising rate of O3 in China's 2 + 26 urban areas was almost 14 times that of the global O3 [34]. The lowest ozone concentration of 80.08 μg/m<sup>3</sup> in Sichuan Basin was observed in 2015, and reached the highest in 2018, with the value of 91.4 μg/m3. It declined slightly in the following two years, but still showed an upward trend during the overall study period. The annual growth rate was about 0.76%.

**Figure 3.** Annual average concentration trend of the six pollutants.

Ozone is not directly emitted by pollution sources in the environment [35]. It is a secondary pollutant generated by chemical reactions of nitrogen oxides and volatile organic compounds under strong ultraviolet light irradiation [36]. Although China has adopted strict control measures in recent years, which has made PM, NO2, SO2 and other atmospheric pollutants show a clear downward trend, the ozone concentration is still slowly increasing [37]. The main reason for this phenomenon is that emissions of NOx and VOCs (main precursors of O3) remain high in China [38]. And the meteorological conditions of high temperature and low rainfall are conducive to the generation of O3 in recent years [39]. At the same time, the global O3 background value has been continuously increasing, which also makes a certain promoting effect on China's ozone concentration [40].
