Temporal Variations in the Air Quality Index and the Impact of the COVID-19 Event on Air Quality in Western China

This study investigated the AQI (air quality index) and atmospheric pollutants including PM2.5, PM10, CO, SO2, NO2 and O3 in Chongqing, Luzhou and Chengdu from 2017 to 2019. In addition, the impacts of the COVID-19 event on the air quality in the three cities in 2020 were compared and discussed. For the combined AQIs for the three cities, in spring, the daily AQIs ranged between 25 and 182 and averaged 72.1. In summer, the daily AQIs ranged between 24 and 206 and averaged 77.5. In autumn, the daily AQIs ranged between 22 and 170 and averaged 61.1, and in winter, the daily AQIs ranged between 28 and 375 and averaged 99.6. The distributions of the six AQI classes in spring were 3%, 94%, 3%, 0%, 0%, and 0%; in summer, they were 11%, 74%, 15%, 0%, 0% and 0%; in autumn, they were 29%, 70%, 1%, 0%, 0%, and 0%, and in winter, they were 1%, 52%, 44%, 3%, 0%, and 0%, respectively. The average AQIs, in order, were Chengdu (85.4) > Chongqing (73.8) > Luzhou (73.2). Both the highest AQIs and PM2.5 (as the major indicatory air pollutant) occurred mainly in the low temperature season (January, December, and February), while O3 was the main air pollutant in June and August when the weather was hot. In February 2020, during the epidemic prevention and control actions taken in response to COVID-19 for the three cities, the combined AQIs for the top five days with the highest AQIs in February 2020 was 79.4, which was 23.6% lower than that from 2017–2019 (AQI = 100.7), and the average concentrations of PM2.5, PM10, SO2, CO, and NO2 were 89.4 μg m, 106 μg m, 2.31 ppb, 0.72 ppm, and 12.3 ppb, respectively, and were 17.9%, 30.8%, 83.8%, 19.8%, and 62.1%, lower than those in February 2017–2019. However, the average O3 concentration (31.8 ppb) in February 2020 rather than decreasing, increased by 6.2%. This is because a lower NO2 concentration hindered the NO + O3 reaction and led to increase O3 concentration in the ambient air.


INTRODUCTION
Over the past few decades, China has developed rapidly in terms of economic growth and urbanization. Automobile exhaust, industrial activities, and biomass combustion have released a significant amount of pollutants into the atmosphere. Air pollution was a serious issue many cities (Liu et al., 2012;Liu and Wang, 2014;Li et al., 2017a).
An estimated 2.5 million people are killed each year by indoor and outdoor air pollution in China (Kulmala, 2015).
Previous research investigated the relationship between air pollutants and human health (Pope and Dochery, 2006;Cao et al, 2012;Heal et al., 2012;Pope and Dochery, 2013;Jin et al., 2017). Studies have shown that inhaling PM 2.5 can cause pneumonia and that it can dissolve in the bloodstream and cause heart and reproductive system diseases (Yang et al., 2017). PM 10 also has a significant impact on the mortality related to cardiovascular diseases and respiratory diseases (Abe et al., 2018). PM 2.5 is more harmful than PM 10 (Deng et al., 2013a, b). SO 2 is a common air pollutant that can cause health damage such as bronchitis and bronchial asthma and thus damage health (Hansell et al., 2011;Cerón-Bretón et al., 2018). CO in the atmosphere not only destroys the nerve function of the heart, but also affects the central nervous system and even leads to death from asphyxia (Yang et al., 2012). NO 2 can hurt the function of the human respiratory The AQI (Air Quality Index), describes the level of air cleanliness or pollution as well as the impact of air quality on health. This study investigated, compared, and discussed the AQIs and air pollutants PM 2.5 , PM 10 , SO 2 , NO 2 , CO, O 3 , and in Chongqing, Luzhou and Chengdu, near western China from 2017 to 2020 focusing on the impacts of epidemic prevention and control actions on air quality.

METHODS
In this study, the air quality in three cities in western China, the Chongqing municipality (longitude 106.54, latitude 29.40), Luzhou in Sichuan province (longitude 105.44, latitude 28.88) and Chengdu (longitude 104.10, latitude 30.66) (Fig. 1) was studied and analyzed for the months of January and February from 2017 to 2020. Chongqing is located in the upper reaches of the Yangtze river, and is characterized by a mild climate and subtropical monsoon humid climate. Chongqing is foggy, and is known as "fog Chongqing." In this study, Spring is defined as the months of March, April, and May; summer includes June, July, and August, and autumn includes September, October, and November, and winter includes January, February and December. The average temperature in Chongqing from 2017 to 2019 ranged between 2.0 and 40℃ and averaged 19.6℃; the temperature in spring ranged between 6.0 and 37℃ and averaged 20℃; in summer, the temperature ranged between 18 and 40℃ and averaged 29℃; in autumn, the temperature ranged between 7.0 and 39℃ and averaged 19℃; and in winter, the temperature ranged between 2.0 and 20℃ and averaged 10℃.
Located southeast of Sichuan province, Luzhou has a mild subtropical humid climate and four distinct seasons. The average temperature in Luzhou from 2017 to 2019 ranged between -2.0 and 39℃ and averaged 18.7℃; the temperature in spring ranged between 4.0 and 36℃ and averaged 20℃; in summer, the temperature ranged between 17 and 39℃ and averaged 28℃; in autumn, the temperature ranged between 6.0 and 35℃ and averaged 18.5℃; and in winter, the temperature ranged between -2.0 and 22℃ and averaged 9.0℃.
Chengdu is located in the central part of Sichuan province and the western part of the Sichuan basin. It has a subtropical monsoon climate and a warm winter climate. The average temperature in Chengdu from 2017 to 2019 ranged between -5.0 and 36℃ and averaged 17.3℃; the temperature in spring ranged between 4.0 and 35℃ and averaged 18℃; in summer, the temperature ranged between 17 and 36℃ and averaged 26℃; in autumn, the temperature ranged between 1.0 and 34℃ and averaged 17℃, and in winter, the temperature ranged between -5.0 and 21℃ and averaged 8.0℃.

Air Quality Index (AQI)
The AQI is a dimensionless index that quantitatively describes the air quality status. First, the sub-AQI of six standard pollutants (PM 2.5 , PM 10 , SO 2 , CO, NO 2 and O 3 ) is calculated based on the observed concentration, as shown in Eq. (1) (Shen et al., 2017;She et al., 2017). The overall AQI represents the maximum value of the sub-AQI of all pollutants. When the AQI is higher than 50, the highest contributor of the sub-AQI of the day is defined as the primary pollutant, as shown in Eq. (2) (Shen et al., 2017;She et al., 2017): (1) AQI = max(I 1 , I 2 , …, I n ) (2) IAQI p : the air quality sub index for air pollutant p; C p : the concentration of pollutant p; C low : the concentration breakpoint that is ≤ C p ; C high : the concentration breakpoint that is ≥ C p ; I low : the index breakpoint corresponding to C low ; I high : the index breakpoint corresponding to C high . The six standards for air pollutants have serious implications for human health. The daily AQI value is calculated from the average concentration of PM 2.5 , PM 10 , SO 2 , NO 2 , and CO for 24 hours, and the maximum concentration of O 3 eight hours per day. According to the U.S. Environmental Protection Agency (U.S. EPA) AQI, the range of AQI values related to air quality can be divided into six classes (Hu et al., 2015;Lanzafame et al., 2015;She et al., 2017;Zhao et al, 2018)

Wind Streamline and Wind Speed
In order to understand the pathway of airflow in Chongqing Municipality and Sichuan Province from January to March, we chose GrADS (Grid Analysis and Display System, http://cola.gmu.edu/grads/) to compute and draw the distribution of the monthly average near-surface streamlines and wind speed with NCEP GDAS/FNL 0.25 Degree Global Tropospheric Analyses data (https://rda.ucar.edu/datasets/ds 083.3/).
According to the above results, in 2017-2019, in Chongqing, the daily AQIs ranged between 27 and 204 and averaged 73.8; in Luzhou, the daily AQIs ranged between 22.7 and 208 and averaged 73.2, and in Chengdu, the daily AQIs ranged between 28.2 and 270 and averaged 85.4. The average AQIs in order were Chengdu (85.4) > Chongqing (73.8) > Luzhou (73.2).
In 2017-2019, in Chongqing, in spring, the daily AQIs ranged between 29 and 145 and averaged 66.6; in summer, those ranged between 30 and 203 and averaged 79.9; in autumn, those ranged between 25 and 132 and averaged 59.9, and in winter, those ranged between 28 and 220 and averaged 88.6. It can be seen that the average AQI was the highest in winter and the lowest was in autumn. In Luzhou, in spring, the daily AQIs ranged between 25 and 165 and averaged 69.3; in summer, those ranged between 24 and 177 and averaged 70.2; in autumn, those ranged between 22 and 134 and averaged 55.6, and in winter, those ranged between 28 and 258 and averaged 97.5. It can be seen that the average AQI was the highest in winter and was the lowest in autumn. In Chengdu, in spring, the daily AQIs ranged between 29 and 182 and averaged 80.3, in summer, those ranged between 28 and 206 and averaged 82.3; in autumn, those ranged between 27 and 170 and averaged 67.9, and in winter, those ranged between 32 and 375 and averaged 113. The seasonal distribution of the highest and lowest average AQIs in Chengdu were consistent with those in Chongqing and Luzhou.
According to the data for the three cities, the average AQI in winter was much higher than in other seasons. This is because the temperature in winter is low, and the atmospheric temperature inversion phenomenon is very dominant, which is not conducive to the, dispersion, dilution and diffusion of pollutants in the air, so the average AQI in winter is higher (Xu et al., 2020a, b).

The Top Five Days with the Highest AQIs from 2017 to 2019
Tables 1, 2, and 3 show the top five days with the highest AQIs each year from 2017 to 2019 in Chongqing, Luzhou and Chengdu, and the concentrations of PM 2.5 , PM 10 , SO 2 , CO, NO 2 , and O 3 , respectively.
In Table 1, the top five days with the highest AQIs in Chongqing in 2017 were 220, 207, 202, 202 and 202, on January 4, January 3, January 5, December 27, and December 28, respectively. On January 4, 2017, the concentrations of PM 2.5 , PM 10 , SO 2 , CO, NO 2 , and O 3 were 171 µg m -3 , 231 µg m -3 , 11.2 ppb, 1.36 ppm, 36.5 ppb, respectively, and on December 28, they were 156 µg m -3 , 229 µg m -3 , 4.9 ppb, 1.52 ppm, 33.1 ppb, and 9.33 ppb, respectively. The indicatory air pollutant on all of these days was PM 2.5 . The highest AQIs in 2018 were 190, 178, 172, 168, and 163, on January 13, December 19, June 7, January 15, and January 14, respectively. On January 13, 2018, the concentrations of PM 2.5 , PM 10 , SO 2 , CO, NO 2 , and O 3 were 143 µg m -3 , 205 µg m -3 , 7.35 ppb, 1.12 ppm, 37.5 ppb, and 3.73 ppb, respectively, and on January 14, they were 124 µg m -3 , 174 µg m -3 , 5.95 ppb, 1.04 ppm, 35.1 ppb and 21.5 ppb respectively. The indicatory air pollutants on these five days were PM 2.5 , PM 2.5 , O 3 , PM 2.5 , and PM 2.5 , respectively. The highest AQIs in 2019 were 203, 179, 175, 172, and 166, on August 12, August 15, January 26, August 17, and August 16, respectively. On August 12, 2019, the concentrations of PM 2.5 , PM 10 , SO 2 , CO, NO 2 , and O 3 were 26 µg m -3 , 55 µg m -3 , 2.8 ppb, 0.64 ppm, 21.4 ppb, and 129 ppb, respectively, and on August 16, they were 28 µg m -3 , 57 µg m -3 , 3.50 ppb, 0.56 ppm, 21.9 ppb, and 108 ppb, respectively, and the indicatory air pollutants for these five days were O 3 , O 3 , PM 2.5 , O 3 and O 3 , respectively. The most common indicatory air pollutant in summer was O 3 , but, in winter, it was PM 2.5 . Previous studies have shown that atmospheric relative humidity is negatively correlated with O 3 concentration and that a lower relative humidity is conducive to the formation of O 3 (Kato et al., 2016;Li et al., 2017b;Gong et al., 2018). In summer, low relative humidity and high temperatures and  wind speeds are conducive to the dispersion, diffusion, and dilution of air pollutants. Precursors (VOCs) also promote the formation of O 3 . In addition, high temperatures and strong ultraviolet radiation will increase the production rate of O 3 in summer; therefore, the concentration of O 3 in summer is much higher. In winter, a large amount of coal was used, and the exhaust gas from combustion greatly contributed to an increase in the concentration of atmospheric particulate matter. By contrast, the temperature in winter is low, and the inverse temperature phenomenon is prominent, which hinders the dilution and diffusion of pollutants in the air. Therefore, the PM 2.5 concentration in winter is higher.
According to Table 3, in Chongqing, the top five highest AQIs in 2017 occurred mainly in January and December. In 2018, they occurred mainly in January, June, and December, while in 2019, they were mainly in January and August. In Luzhou, the top five highest AQIs in 2017 occurred mainly in January and December. In 2018, they occurred mainly in January, February, and December, while in 2019, they were mainly in January, March, and August. In Chengdu, the top five highest AQIs in 2017 occurred mainly in January, in 2018, they were mainly in January, February, and December, and in 2019, the top six highest AQIs were mainly in August and December.
Among the 46 days with the highest AQIs over the three years under observation (2017-2019), the distributions were 52.0%, 24.0%, 17.0%, 4.30%, and 2.20% in January, December, August, February, and June, respectively. It can be seen that the highest AQIs occurred mainly in the season with lower temperatures (January, December, and February), which was consistent with the conclusion made earlier that PM 2.5 was the main indicatory air pollutant in January, February, and December due to low temperatures, and O 3 was the main air pollutant in June and August due to hot temperatures. In winter, a lower ground temperature will hinder the dispersion of air pollutants and lead to an increase in PM 2.5 concentrations in the atmosphere Xing et al., 2017;Lee et al., 2018;Wang et al., 2018). Under sufficient solar radiation intensity, NO 2 acts as the precursor of a photochemical reaction, and first decomposes into NO and O (3P): It can be seen that NO x is one of the important precursors of O 3 , and NO is a direct reactant with O 3 . A lower NO 2 concentration in the atmosphere will lead to a decrease in the level of NO, which will reduce the possibility of NO reacting with O 3 , and in turn will lead to an accumulation of O 3 . Due to strong solar radiation and a dominant photochemical reaction, summer is more suitable for the accumulation of O 3 , Therefore, O 3 concentrations in summer are much higher. The average AQIs during the 15 days with highest AQIs were 187, 194, and 227 in Chongqing, Luzhou, and Chengdu, respectively. These results indicate that Chengdu had the highest average AQI during the observation period, and thus more attention should be paid to improving the air quality in this city.
It can be seen that the average AQIs in Chengdu (227) were far greater than the average AQI values in Chongqing and Luzhou, while the results in Chongqing (187) and Luzhou (194) were very close. It can be concluded that the air quality in Chengdu was the worst, and the air quality in Chongqing and Luzhou were, in order, the second and the third. This result is consistent with conclusions drawn above.
In Chongqing, the average AQI values for the highest five days were 207, 174, and 179 in 2017, 2018, and 2019, respectively; in Luzhou, they were 240, 174, and 167, and in Chengdu, they were 300, 212 and 178 in 2017, 2018 and 2019, respectively. It can be seen that in Chongqing, the air quality was the worst in 2017, followed by 2019, and was the best in 2018; in Luzhou, the air quality was the worst in 2017, followed by 2018, and was the best in 2019; in Chengdu, the air quality was the worst in 2017, followed by 2018, and was the best in 2019. From 2017 to 2019, in general, the air quality of the three cities were all improved.

Impact of the COVID-19 Event on Air Quality
Figs. 2, 3, and 4 show the average monthly AQI distribution for Chongqing, Luzhou and Chengdu in January and February 2017-2019 during the non-epidemic period, respectively, and Fig. 5 shows the monthly AQI distribution in Chongqing, Luzhou and Chengdu in January and February 2020 for the epidemic prevention and control period.
As shown in Figs. 2 and 5, in Chongqing, the proportion of classes I, II, III, IV, V, and VI with the average AQI from January 2017-2019 were 0%, 42%, 58%, 0%, 0%, and 0%, respectively, and those in January 2020, were 19%, 74%, 7%, 0%, 0%, and 0%, respectively. It can be seen that compared with the average AQI from January 2017-2019, the proportion of Class I in January 2020 increased to 19.0%; that of Class II increased to 74.0%, and that of Class III decreased to 7.00%. In addition, the average AQI in January 2020 was 69.8, which was 35.7% lower than that in 2017-2019 (AQI = 98.8). In February 2017-2019, in Chongqing, the proportion of classes I, II, III, IV, V, and VI with the average AQI was 7%, 82%, 11%, 0%, 0%, and 0% respectively, and in February 2020, the proportion of classes I, II, III, IV, V, and VI with the average AQI was 21%, 65%, 14%, 0%, 0%, and 0%, respectively. It can be seen that compared with the average AQI in February 2017-2019, Class I in February 2020 increased to 21.0%, and Class II decreased to 65.0%. The average AQI in February 2020 was 68.8, which was 11.5% lower than that from 2017-2019 (AQI=77.2). It can be seen that the air quality in January and February 2020 (during the epidemic prevention and control period) improved significantly compared with January and February 2017-2019 (the non-epidemic period).
As shown in Figs. 4 and 5, in Chengdu, the proportion of classes I, II, III, IV, V, and VI with the average AQI in January 2017-2019 was 0%, 16%, 61%, 23%, 0%, and 0%, respectively, and the proportion of classes I, II, III, IV, V and VI with the average AQI in January 2020 was 0%, 58%, 39%, 3%, 0%, and 0%, respectively. It can be seen that compared with the average AQI in January 2017-2019, the proportion of AQI Class II in January 2020 increased to 58.0%, and that of Class III decreased to 39.0%. The average AQI in January 2020 was 92.9, which was 32.0% lower than that from 2017-2019 (AQI=128). In Chengdu, the proportion of classes I, II, III, IV, V, and VI with the average AQI in February 2017-2019 was 0%, 64%, 32%, 4%, 0%, and 0%, respectively, and the proportion in February 2020 was 17%, 59%, 24%, 0%, 0%, and 0%, respectively. It can be seen that compared with the average AQI in February 2017-2019, AQI Class I in February 2020 increased to 17.0%; Class II decreased to 59.0%; Class III decreased to 24.0%, and Class IV decreased to zero. The average AQI in February 2020 was 74.7, which was 25.0% lower than that during the period from 2017-2019 (AQI=96.0). It can be seen that the AQI in January and February 2020 (during the epidemic prevention and control period) decreased significantly compared with January and February 2017-2019 (during the non-epidemic period).
According to the above analysis, for the three cities, the combined AQI in February 2020 was 79.4, which was 23.6% lower than that in the period from 2017-2019 (AQI = 101). These results indicated that the prevention and control measures for COVID-19 greatly restricted the movement of people, transportation, engineering construction, industrial production and commercial trading activities Therefore, the stationary emissions, automobile exhaust, and fugitive emissions were also been greatly reduced, so the air quality was significantly improved.
The above results indicated that in Chongqing, the average value on the 15 days with the highest AQI in February 2017-2019 (non-epidemic period) was 125, and the corresponding  Table 4, the average AQI in 2020 was 105, which was 16.9% lower than that in the period from 2017-2019 (AQI = 125). Compared with the non-epidemic control period, the air quality in the epidemic prevention and control period was improved significantly. As shown in Table 5, in Luzhou, the highest AQIs in February 2017 were 177, 170, 168, 162, and 161 on February 18, February 16, February 20, February 14, and February 15, respectively. On February 18, 2017, the concentrations of PM 2.5 , PM 10 , SO 2 , CO, NO 2 , and O 3 were 139 µg m -3 , 188 µg m -3 , 13.3 ppb, 0.72 ppm, 24.4 ppb, and 30.8 ppb, respectively, and on February 15, they were 122 µg m -3 , 165 µg m -3 11.2 ppb, 0.56 ppm, 23.9 ppb, and 35.5 ppb, respectively. The indicatory air pollutant on all five days was PM 2.5 .
According to the above data, it can be seen that in Luzhou, the average AQI on the 15 days with the highest AQI in February 2017-2019 (non-epidemic period) was 142, and the corresponding average concentrations of PM 2.5 , PM 10 , SO 2 , CO, NO 2 , and O 3 were 108 µg m -3 , 143 µg m -3 , 7.70 ppb, 0.69 ppm, 18.4 ppb, and 31.7 ppb respectively. The average AQI on the top five days with the highest AQI in February 2020 (epidemic prevention and control period) was 132, and the corresponding average concentrations of PM 2.5 , PM 10 , SO 2 , CO, NO 2 , and O 3 were 100 µg m -3 , 113 µg m -3 , 2.38 ppb, 0.77 ppm, 8.86 ppb, and 30.0 ppb, which were 7.60%, 23.5%, 105.6%, -11.0%, 70.0%, and 5.60% lower than those in 2017-2019, respectively. In Table 5, it can be seen that the average AQI in 2020 was 142, which was 7.60% lower than that in the period from 2017-2019 (AQI = 132). Compared with the non-epidemic period, the air quality in the epidemic prevention and control period improved significantly.
According to the above data, the average of the 15 days with the highest AQI in February from 2017 to 2019 (the non-epidemic period) in Chengdu was 156, and the corresponding average concentrations of PM 2.5 , PM 10 , SO 2 , CO, NO 2 , and O 3 were 118.1 µg m -3 , 162 µg m -3 , 4.22 ppb, 0.99 ppm, 27.9 ppb, and 32.3 ppb, respectively. The average of the top five days with the highest AQI in February 2020 (the epidemic prevention and control period) was 118, and the corresponding average concentrations of PM 2.5 , PM 10 , SO 2 , CO, NO 2 , and O 3 were 89.2 µg m -3 , 108 µg m -3 , 2.38 ppb, 0.74 ppm, 15.7 ppb, and 40.5 ppb, which were 27. 9%, 40.2%, 55.8%, 29.6%, 55.9%, and -22.5% lower than those in the period from 2017-2019, respectively. In Table 6, the average AQI in 2020 was 118, which was 27.6% lower than that in the period from 2017-2019 (AQI = 155.7). Compared with the non-epidemic period, during the epidemic prevention and control period, the air quality was improved significantly.
In the combined results of three cities, during the 5 days with the highest AQI during the epidemic prevention and control action period (February 2020), the average concentrations of PM 2.5 , PM 10 , SO 2 , CO, and NO 2 , were 89.4 µg m -3 , 106 µg m -3 , 2.31 ppb, 0.72 ppm, and 12.3 ppb, and which were 17.9%, 30.8%, 83.8%, 19.8%, and 62.1%, lower than those in February 2017-2019, respectively. However, the average O 3 concentration (31.8 ppb) in February 2020 did not show a significant decrease, but rather increased by 6.2%. This decrease in the average concentration of PM 2.5 , PM 10 , SO 2 , CO, and NO 2 was attributed to the "lockdown" of the cities during the epidemic prevention and control action period (2020). The O 3 concentration increased because a lower NO 2 concentration hindered the NO + O 3 reaction and resulted in an accumulation of O 3 in the air.

The Wind Streamline and Wind Speed
During this period, the airflow pathway is usually affected by Siberia High (Siberia Anticyclone). Fig. 6 shows the distribution of monthly average near-surface streamlines in Chongqing Municipality and Sichuan Province (including Chengdu and Luzhou City) from January to March in 2019 and 2020, respectively. Fig. 7 shows the distribution of the monthly average nearsurface wind speed in the same regions. The results indicated that the monthly average wind speed in this region from January to March of 2019 was generally higher than the same period in 2020, and the monthly average wind speed in February 2019 was significantly higher than that in January and March of 2019 (about 0.5 m s -1 higher). From January to March of 2020, the monthly average wind speed in March was just slightly higher than that in January and February of 2020, and the difference between these three months was not obvious. In terms of geographical distribution, Chongqing Municipality is located on the east side of Sichuan province; Luzhou City is located on the southeast corner of Sichuan Province close to the southwest side of Chongqing Municipality, and Chengdu City is located east of the center of Sichuan Province. Due to the influence of the terrain, western Sichuan Province (relatively higher terrain) experiences relatively stronger wind speed compared to Chongqing, Chengdu, and Luzhou (relative lower terrain) in which the monthly average wind speed was below 2.5 m s -1 in 2019 and even less than 2.0 m s -1 in 2020. According to the distribution of the monthly average streamlines, based on the places where the confluences of streamlines generally have lower wind speed, these three cities are located in lower wind speed regions. In Chongqing, from January to March, the prevailing winds in the southeastern corner are usually southeasterly (SE) to easterly (E) wind, while the prevailing winds in the northern corner (relatively upwind) and southwestern corners (relatively downwind) are usually northeasterly (NE) to east-northeasterly (ENE), and the wind speed is usually lower at the streamline confluence where air pollution is easily accumulated. The prevailing wind in Luzhou from January to March is northerly (N). Luzhou is located downwind of Chongqing, which it means that the air pollutants from Chongqing were easily transported and passes through Luzhou in a northerly to southerly direction with the air flow. Between Chengdu (west side) and Chongqing and Luzhou (east side), there is a banded wind zone with relatively stronger wind speed from January to March, and the wind direction is generally northeasterly to northerly. In Chengdu, the western corner is affected by a relatively stronger wind zone as mentioned above. The prevailing winds are northeasterly to easterly, so it is not as easy for this banded zone to accumulate air pollutants as it is in the surrounding regions. In other parts of Chengdu, the wind direction is generally more chaotic, resulting in relatively lower wind speed, but it tends toward a northerly wind. Because of this lower wind speed, it is relatively easy for Chengdu to accumulate air pollutants. Generally speaking, because it is affected by Siberia High (Siberia Anticyclone) during the winter, the airflow transport in Chongqing Municipality and the eastern half of Sichuan Province (including Chengdu and Luzhou City) from January to March is dominated by a weaker northeasterly wind (entering from the northeast and leaving from the southwest), and the monthly average wind speed is less than 2.5 m s -1 . The wind streamline and wind speed data provide insights into the air transport of pollutants and their effects on the surrounding environment.

CONCLUSIONS
1. In 2017-2019, in Chongqing, the daily AQIs ranged between 27 and 204 and averaged 73.8; in Luzhou, the daily AQIs ranged between 22.7 and 208 and averaged 73.2; while in Chengdu, the daily AQIs ranged between 28.2 and 270 and averaged 85.4. The average AQIs, in order, were Chengdu (85.4) > Chongqing (73.8) > Luzhou (73.2). 2. For the combined AQIs for the three cities, in 2017-2019 in spring, the daily AQIs ranged between 25 and 182 and averaged 72.1; in summer, the daily AQIs ranged between 24 and 206 and averaged 77.5; in autumn, the daily AQIs ranged between 22 and 170 and averaged 61.1, and in winter, the daily AQIs ranged between 28 and 375 and averaged 99.6. The distributions of the six AQI classes in spring were 3%, 94%, 3%, 0%, 0%, and 0%, respectively; those in summer were 11%, 74%, 15%, 0%, 0%, and 0%, respectively; those in autumn were 29%, 70%, 1%, 0%, 0% and 0%, respectively, and those in winter were 1%, 52%, 44%, 3%, 0%, and 0%, respectively. The average AQI in winter was much higher than in the other seasons. This is because the temperature in winter is low, and the atmospheric temperature inversion phenomenon is prominent, which is not conducive to the dispersion and dilution of pollutants in the air, so the average AQI in winter tends to be higher. 3. In Chongqing, the average AQI values on the highest 5 days were 207, 174, and 179 in 2017, and 2019in Luzhou, they were 240, 174, and 167;in Chengdu, those were 300, 212, and178, in 2017, 2018, and 2019, respectively. It can be seen that in Chongqing, the air quality was the worst in 2017, followed by 2019, and was the best in 2018; in Luzhou, the air quality was the worst in 2017, followed by 2018, and was the best in 2019; in Chengdu, the air quality was the worst in 2017, followed by 2018, and was the best in 2019. From 2017 to 2019, in general, the air quality improved on an annual basis. 4. For the three cities, the combined AQI in February 2020 was 79.4, which was 23.6% lower than that in the period from 2017-2019 (AQI=101). These results indicated that the prevention and control measures for COVID-19 greatly restricted the movements of people, transportation, engineering construction, industrial production, and commercial trading activity Therefore, stationary emissions, automobile exhaust, and fugitive emissions were also greatly reduced, so the air quality was significantly improved. 5. The combined results for the three cities on the 5 days with the highest AQIs during the epidemic prevention and control action period (February 2020) reveal, the average concentrations of PM 2.5 , PM 10 , SO 2 , CO, and NO 2 of 89.4 µg m -3 , 106 µg m -3 , 2.31 ppb, 0.72 ppm, and 12.3 ppb, which were 17.9%, 30.8%, 83.8%, 19.8%, and 62.1%, lower than those in February 2017-2019, respectively. However, the average O 3 concentration (31.8 ppb) in February 2020 did not show a significant decrease, but rather increased by 6.2%. The decrease in the average concentrations of PM 2.5 , PM 10 , SO 2 , CO, and NO 2 was attributed to the "lockdown" of the cities during the epidemic prevention and control action period (February 2020). An increase in the O 3 concentration was because a lower NO 2 concentration hindered the NO + O 3 reaction and resulted in an accumulation of O 3 in the air. 6. In general, due to effects of the Siberia High (Siberia Anticyclone) during winter, the airflow transport features in Chongqing Municipality and the eastern half of Sichuan Province (including Chengdu and Luzhou City) from January to March was dominated by a weaker northeasterly wind (entering from the northeast and leaving from the southwest), and the monthly average wind speed was below 2.5 m s -1 . The wind streamline and wind speed data provides insights into the air transport of pollutants and their effects on the air quality in the surrounding environment. 7. This study provides useful information for the development of air pollution control strategies and adds to the body of research on this topic in the literature.