Air Quality Response to COVID-19 Control Measures in the Arid Inland Region of China: A Case Study of Eastern Xinjiang

Abstract

This study examined the temporal changes and dispersion of potential sources of the six criteria air pollutants, namely, particulate matter with an aerodynamic diameter of less than 2.5 and 10 μm (PM_2.5 and PM₁₀), nitrogen dioxide (NO₂), sulfur dioxide (SO₂), carbon monoxide (CO), and ozone (O₃), in eastern Xinjiang, China, during the COVID-19 period in summer 2020 (16 July to 29 August ). Compared to the same periods in 2019 and 2021, the mean concentrations of all pollutants, except for SO₂ and O₃, and the air quality index (AQI) were lower in 2020 (relative changes: NO₂ 48.3–54.4%, PM₁₀ 35.8–49.6%, PM_2.5 19.3–43.5%, CO 16.5–34.8%, AQI 17.2–29.4%), which can be attributed to the reduced anthropogenic activities. Compared to the period before the lockdown in 2020 (16 June to 15 July), the mean NO₂ concentration showed the largest decrease during the lockdown (47.9%), followed by PM_2.5 (32.7%), PM₁₀ (37.6%), and CO (15.4%). In contrast, there were only minimal changes in O₃, with the mean concentrations falling slightly by 7.56%, and the mean concentration of SO₂ increased by 10.4%. The decrease in NO_x and the dry climate could have hindered O₃ formation, while vital industrial activities in eastern Xinjiang probably maintained SO₂ emissions. In the subsequent recovery period (30 August to 28 September), the mean NO₂ concentration increased the most at 59.3%, which was due to the rapid resumption of traffic-related emissions. During the lockdown in 2020, the diurnal profiles of PM_2.5, PM₁₀, NO₂, and CO concentrations showed lower peak concentrations in the morning (09:00–11:00) and evening (20:00–22:00), demonstrating a significant reduction in traffic-related emissions. The lower O₃ and higher SO₂ peak concentrations may have resulted from lower NO_x levels and higher electricity consumption due to the “stay-at-home” policy. The analysis of the distribution of potential sources showed that O₃ generally originated from widespread source areas, while the other pollutants mainly originated from local emissions. During the lockdown period, the source areas of PM_2.5 and PM₁₀ were more dispersed, with an enhanced contribution from long-range transport.

1. Introduction

Air pollution in China remains one of the most pressing environmental and public health problems. Numerous epidemiological studies have shown that exposure to elevated concentrations of criteria air pollutants (particulate matter with an aerodynamic diameter of less than 2.5 and 10 μm (PM_2.5 and PM₁₀), nitrogen dioxide (NO₂), sulfur dioxide (SO₂), carbon monoxide (CO), and ozone (O₃)) is associated with a variety of adverse health outcomes, including respiratory and cardiovascular diseases, increased mortality rates, and impaired cognitive development [1,2,3]. In 2021, an estimated 1.86 million deaths in China were attributable to PM_2.5 pollution, and 126,000 deaths were associated with O₃ pollution (https://www.stateofglobalair.org/health/pm, accessed on 26 May 2025). In response, China has launched a series of nationwide air pollution control initiatives, such as the “Air Pollution Prevention and Control Action Plan” (APPCAP, 2013) and the “Three-year Action Plan to Win the Blue Sky Defense War” (“Blue Sky Defense War”, 2018), which have led to significant improvements in urban air quality, especially in eastern and central cities [4,5,6,7]. However, comparatively little attention has been paid to western and remote regions, such as Xinjiang, where air quality is still a problem [8,9,10].

The outbreak of the COVID-19 pandemic in 2020 provided an unprecedented opportunity to study the response of air quality to a sudden and widespread reduction in anthropogenic activities. During the early phase of the pandemic (from late January to March 2020), air pollutants (e.g., NO₂ and PM_2.5) decreased significantly in the industrially densely populated regions of eastern China, such as the Yangtze River Delta (YRD) and the Beijing–Tianjin–Hebei (BTH) region. However, secondary pollutants such as O₃ increased by 102–125% and 76% in YRD and BHT, respectively [11,12,13,14,15]. This phenomenon was attributed to a sharp decrease in traffic emissions, which led to lower NO_x concentrations and weakened its titration effect on O₃. The cold and humid winter environment favored the accumulation of free radicals, which increased the efficiency of O₃ formation. During the pandemic, there was a significant decrease in nitrate concentration in urban areas such as Beijing and Suzhou, but the formation of sulfate and secondary organic aerosols (SOA) was enhanced [16,17]. In addition, the health hazard index (HAQI) of BTH and YRD decreased by 24.5% and 18.1%, respectively, due to the decrease in PM_2.5 concentrations [18]. The increase in O₃ in winter partially offset these health benefits and exacerbated respiratory risks, particularly at low temperatures [15,19].

The eastern region of Xinjiang in China, which is characterized by its dry climate and high level of industrial activity, serves as an important corridor connecting Central Asia with inland China. The special climatic conditions prevailing here (e.g., frequent dust storms) further exacerbate the problem of air quality. In contrast to central and eastern China, the COVID-19 outbreak and lockdown in eastern Xinjiang occurred later in the summer (16 July to 29 August 2020) with strong sunlight and high temperatures, which may lead to a different impact of the pandemic on air quality. In this work, variations in the six criteria air pollutants (PM_2.5, PM₁₀, NO₂, SO₂, CO and O₃) in eastern Xinjiang before, during, and after the COVID-19 lockdown in summer 2020 were analyzed and compared with those in the same period in 2019 and 2021. The MeteoInfo trajectory model was used in combination with the Potential Source Contribution Function (PSCF) and the Concentration Weighted Trajectory (CWT) method to determine local and regional contributions to air pollutants. The results are expected to improve our understanding of the impact of reduced emissions during the pandemic on air pollution in the remote western areas of China.

2. Materials and Methods

2.1. Study Area and Data Source

The study area of this work includes Turpan and Hami in eastern Xinjiang, which is geographically located between the Tianshan Mountains in the west and the Gobi Desert in the east (Figure 1) and forms a distinct transition zone in northwestern China. This area is characterized by a typical continental climate with hot summers and cold winters, and serves as an important industrial location and as a corridor connecting Central Asia with inland China.

Hourly concentration data for the six criteria air pollutants, including PM_2.5, PM₁₀, CO, NO₂, SO₂, and the maximum daily average 8 h (MDA8) O₃, were collected from national monitoring stations in Hami (Hami Normal School, HMNS) and Turpan (Turpan Environmental Protection Bureau, TREPB; Figure 1). Statistics of the six criteria air pollutants are listed in Table S1 in the Supplement. The air quality measurements were conducted using Thermo Fisher Scientific (Waltham, MA, USA) instruments: the 5014i Beta Monitor for PM_2.5 and PM₁₀ (beta-attenuation principle, flow rate 16.67 L/min, detection limit: <1 μg/m³ over 24 h), 48i for CO (NDIR photometry, range 0–10,000 ppm, detection limit: 0.04 ppm), 42i for NO₂ (chemiluminescence, range 50–1000 ppb, detection limit: 0.4 ppb), 49i for O₃ (UV photometry, range 50–1000 ppb, detection limit: 0.5 ppb), and 43i for SO₂ (pulsed fluorescence, range 50–1000 ppb, detection limit: <0.5 ppb) (https://www.thermofisher.cn, accessed on 3 September 2025). The devices were housed in a temperature-controlled enclosure (20–30 °C) with Teflon tubing and filters to prevent contamination. Synchronous meteorological parameters, including temperature, relative humidity, barometric pressure, and wind speed, were obtained from national weather stations in Hami and Turpan. According to the “Technical Specifications for the Operation and Quality Control of the Continuous Automated Monitoring System for SO₂, NO₂, O₃, and CO” (Ministry of Ecology and Environment of the People’s Republic of China, 2018) and the “Technical Guide for Automatic Monitoring of Particulate Matter in Ambient Air (PM₁₀ and PM_2.5) Using the Beta Ray Method” (Ministry of Ecology and Environment of the People’s Republic of China, 2020), the reference gas and particulate matter analyzers are subjected to periodic inspections and calibrations at specified intervals.

The data given here are the average values of the two stations in eastern Xinjiang. For the backward trajectory analysis, the meteorological input fields were derived from the NOAA Global Data Assimilation System (GDAS1), which provides four daily time datasets (00:00, 06:00, 12:00, and 18:00 UTC). Access was obtained via the official NOAA FTP site (https://www.ready.noaa.gov/data/archives/gdas1/; last accessed on 17 May 2025).

To assess the impact of COVID-19 on air quality in eastern Xinjiang, the study focused on the period from 16 June to 28 September 2020, which was divided into three phases according to the local timing of the outbreak. Phase 1 (P1, 16 June to 15 July) preceded the implementation of the main control measures (e.g., “stay-at-home” policy). Phase 2 (P2, 16 July to 29 August) marked the period of lockdown, during which strict restrictions were imposed on industrial and social activities. Phase 3 (P3, 30 August to 28 September) was the post-lockdown period, when restrictions were gradually eased and public and economic activities partially recovered, although some control measures remained in place. Air quality data for the same periods in 2019 and 2021 were also analyzed for comparison.

2.2. Source Analysis Models

Forty-eight hourly backward trajectories at an altitude of 500 m above the ground were calculated using the MeteoInfo trajectory model (http://meteothink.org/docs/introduction/index_chinese.html, accessed on 13 May 2025) for the measurement site Hami (42.84° N, 93.51° E), a city with flat terrain in eastern Xinjiang that is rather representative for the region [10,20], during the COVID-19 pandemic.

The PSCF method was applied to identify likely source regions of elevated pollutant concentrations by estimating the conditional probability of air masses passing through each grid cell [21,22]. The calculation proceeds as follows:

$$PSCF = {\displaystyle \frac{m_{ij}}{n_{ij}}}$$

(1)

where m_ij is the number of trajectory endpoints associated with high pollutant concentrations that fall within the grid cell (i, j), and n_ij represents the total number of trajectory endpoints in that grid cell. A weighting factor W_ij was used to minimize uncertainties in cells with low trajectory counts [23].

$${WPSCF}_{ij}={PSCF}_{ij}\times W_{ij}$$

(2)

$$W_{ij}\left\{\begin{array}{c}1.00, n_{ij}>80\hfill \\ 0.70, 80{ \ge n}_{ij}>20\hfill \\ 0.42, 20\ge n_{ij}>10\hfill \\ 0.05, 10\ge n_{ij} \hfill \end{array}\right.$$

(3)

Since the PSCF method identifies potential source regions based on the likelihood of high pollutant concentrations, it does not take into account the actual extent of pollution. To address this limitation, the concentration-weighted trajectory (CWT) method is used, which assigns a weighted average concentration to each grid cell by integrating the pollutant concentrations along the backward trajectories, allowing a more quantitative assessment of source contributions. The CWT value for grid cell (i, j) is calculated as follows:

$${C\mathrm{W}\mathrm{T}}_{ij}={\displaystyle \frac{\sum _{l=1}^M{C}_1\times {\tau }_{ijl}}{\sum _{l=1}^M{\tau }_{ijl}}}$$

(4)

where CWT_ij is the average contamination weight concentration of the grid (i, j); l is the trajectory of the air mass; M is the total number of traces; C_l is the corresponding pollutant mass concentration when trace l crosses the grid (i, j), τ_ijl is the time that track l stays in the grid (i, j). The same weighting scheme as for WPSCF is used to calculate the weighted CWT values (WCWT) [24].

$${WCWT}_{ij}={CWT}_{ij}\times W_{ij}$$

(5)

3. Results

3.1. Variations in Air Pollutants

3.1.1. Temporal Variations in Air Quality Index (AQI) and Air Pollutants

As shown in Figure 2a, the average AQI values during the P1 period (59–60) showed no significant differences between the three years (p > 0.05), indicating relatively stable atmospheric conditions and minimal interannual variation prior to the pandemic. The calculation method for the AQI is provided in Text S1 in the Supporting Information. During the P2 period, the mean AQI value in eastern Xinjiang decreased by 29.4% in 2020 (48) compared to 2019 (68), and the value in 2021 (58) remained below the 2019 level. The proportion of days with “excellent” air quality (AQI: 0–50) based on the “Technical Regulation on Ambient Air Quality Index (HJ 633–2012)” was 81.4% in P2 2020, higher than the values recorded in eastern China (about 50%) during the pandemic [25] and in eastern Xinjiang in P2 2019 (4.65%) and 2021 (30.2%). Due to its unique geographical location, eastern Xinjiang is influenced by both dust emissions and anthropogenic activities, with PM₁₀ being the main pollutant [10]. Compared to 2019 and 2021, the main pollutant shifted from PM₁₀ to O₃ during P2 2020 (Table S2). During the pandemic in the winter months, PM_2.5 dominated the air pollutants in eastern China [26]. The improvement of air quality in eastern Xinjiang in P2 2020 can be attributed to the reduction in anthropogenic activities, including local industrial and traffic emissions, as well as meteorological conditions in the summer months. The mean AQI value in P3 2020 (60) was also significantly (p < 0.05) lower than in 2019 (72) and 2021 (73), indicating a short-term lag in air pollution recovery and sustained benefits of the strict control measures.

Compared to P1 2020, the daily mean concentrations of PM_2.5, PM₁₀, NO₂, and CO decreased significantly (p < 0.05) in P2 2020 (Figure 2b–e). Among these pollutants, NO₂ showed the largest decrease (47.9%), which was also observed in eastern Chinese cities, suggesting that the reduction in traffic volume had a significant impact on NO₂ concentrations [27,28]. The mean concentrations of PM_2.5 and PM₁₀ decreased by 32.7% and 37.6%, respectively, which were lower than those observed in eastern Chinese cities, such as Hangzhou (PM_2.5, 59.0%; PM₁₀, 54.0%) [13]. The lower decrease in particulate matter (PM) concentrations may be attributed to increased emissions from natural sources and regional transport in summer. As shown in Figure S1c, the mean wind speed in P2 is lower than in P1 2020, which indicates better diffusion conditions in P1 2020. The decrease in mean CO concentration (15.4%) can be related to the decrease in all combustion emissions, including traffic and industry.

Unlike the air pollutants mentioned above, the mean concentration of SO₂ increased by 10.4% in P2 2020 (Figure 2g). This is probably mainly due to the fact that eastern Xinjiang is a major resource processing region with large-scale coal mining, coal-fired power plants, and petrochemical production. Industrial activities such as coal and energy production are vital to the local economy and cannot be allowed to cease completely. There is evidence that SO₂ concentrations have only decreased in 8 out of 16 cities in Xinjiang during the pandemic [29]. In the subsequent P3 period, concentrations of PM_2.5, PM₁₀ and CO increased by 27.8% to 39.0% compared to P2, with NO₂ showing the strongest rebound with an increase of 59.3%, reflecting the rapid resumption of traffic-related emissions.

O₃ showed no significant fluctuations, and its mean concentration decreased slightly by 7.56% in P2 (Figure 2f). This is due to the fact that O₃ is a secondary product whose concentration is not linearly related to the direct emissions of its precursors. Studies show that O₃ formation in eastern and northwestern China is limited to volatile organic compounds (VOCs) and NO_x, respectively [13,14,30]. The significant decrease in NO_x concentrations in P2 2020 could be an unfavorable condition for O₃ formation. In addition, this trend is also likely to be exacerbated by the dry climate and dust storms in Xinjiang, which reduce solar radiation and hinder O₃ formation [31].

To isolate the specific effects of the lockdown from interannual variability, the daily mean pollutant concentrations during the lockdown (P2) in 2020 were compared with those of the same calendar intervals in 2019 and 2021. During the lockdown in 2020, the mean NO₂ concentration fell to 9.48 μg/m³, which corresponds to a decrease of 54.4% and 48.3% compared to 2019 (20.8 μg/m³) and 2021 (18.3 μg/m³), respectively (Figure 3c). This decrease is significantly higher than the typical inter-annual fluctuations without lockdown (~12%). Considering that NO₂ is primarily associated with motor vehicle emissions, this remarkable decrease underlines the significant impact of the lockdown measures on local traffic [32]. Similarly, mean concentrations of PM₁₀ (43.0 μg/m³) and PM_2.5 (14.1 μg/m³) decreased by 49.6%/35.8% and 43.5%/19.3%, respectively, compared to P2 2019/2021 levels (Figure 3a,b). This decline is primarily attributed to lower primary emissions and the associated secondary formations. The mean winds peed was higher in P2 2020 (4.54 m/s) than in P2 2019 (4.18 m/s) and 2021 (4.32 m/s; Figure S1c), which indicates a stronger dispersion of air pollutants during the lockdown [12]. The comparisons in Figure 3d show that the mean CO concentration in P2 decreased less than the NO₂ concentration in 2020, indicating that the CO concentration in eastern Xinjiang is less dependent on traffic emissions. Besides the emissions from industrial production, the increase in mean SO₂ concentration in P2 2020 (Figure 3f) is likely due to higher electricity consumption for cooling and ventilation by people following the “stay-at-home” policy [33,34]. Unlike other pollutants, O₃ showed little fluctuation due to its complex formation mechanism and influencing factors.

3.1.2. Diurnal Variation of Air Pollutants

Figure 4 illustrates the diurnal variation of air pollutants during the P2 period in 2020 and the same period in 2019 and 2021. Due to the geographical location and longer daylight duration, peak concentrations of air pollutants generally occur 2–3 h later in eastern Xinjiang than in eastern and central China. In 2019, the diurnal profiles of PM_2.5, PM₁₀, NO₂ and CO exhibited a bimodal pattern with peak concentrations in the morning (09:00–11:00) and evening (20:00–22:00), which is due to traffic-related emissions during rush hour [35]. The decrease in concentrations at midday and the increase in the evening are due to changes in the height of the planetary boundary layer (PBL) [36]. The morning peak in NO₂ concentrations is narrower than for the other pollutants, which may be due to the fact that NO₂ is specifically associated with vehicle exhaust at this time of day. In contrast, PM_2.5 and PM₁₀ are influenced not only by direct vehicle emissions, but also by contributions from dust and secondary formation processes [37]. The diurnal variation of CO was similar to that of SO₂, and both showed an increase in early morning starting hours from 1:00 a.m., supporting that CO is more influenced by industrial emissions than NO₂.

In 2020, the peak concentrations of these four pollutants were reduced by strict lockdown measures, and their daily curves were flattened (Figure 4). For example, the PM_2.5 peak in the morning (09:00–11:00) almost completely disappeared and the evening increase (18:00–20:00) weakened, which can be attributed to the drastic reduction in traffic volumes as a result of the implementation of the “stay-at-home” policy [38]. However, a weak bimodal pattern with peak concentrations at the same time was observed for PM₁₀, NO₂ and CO, suggesting that diurnal variations in PM_2.5 concentrations are more sensitive to traffic emissions [12]. Also, in eastern Chinese cities (e.g., Beijing, Wuhan, and YRD), the lockdown measures only reduced peak concentrations of PM_2.5, PM₁₀, NO₂, and CO without changing the timing of peak occurrence during the day.

However, there was no significant decrease in peak concentrations of O₃ and SO₂ in 2020. O₃ showed a unimodal pattern from 2019 to 2021 with a peak in the afternoon (18:00–20:00), which is ascribed to enhanced photochemical reactions during the long hours of sunlight in summer [39]. In 2020, peak O₃ concentrations were lower than in 2019 and 2021, which contrasts with eastern Chinese cities where O₃ concentrations increased due to lower NO_x titration [12]. This could be partly due to the fact that in northwestern China, particularly in Xinjiang, O₃ formation is limited to NO_x levels [30]. In 2020, SO₂ showed an anomalous morning peak (09:00–10:00) compared to 2019 and 2021. As mentioned above, this is likely due to higher electricity consumption and increased emissions from coal-fired power plants following the “stay-at-home” policy. In 2021, the diurnal profiles of all pollutants were similar to those in 2019, but the mean concentrations were lower, with the exception of SO₂, possibly due to the continued effect of the implementation of the “Blue Sky Defense War” policy.

3.2. Analysis of Air Pollutant Sources During the COVID-19 Lockdown

3.2.1. Diagnostic Ratio Analysis

The PM_2.5/PM₁₀ ratio in eastern Xinjiang remained below 0.5 in summer 2019–2021 (Figure 5a), indicating that coarse particles, mainly from natural sources such as sand dust, predominate over fine PM from anthropogenic activities [40,41,42]. During the P2 period in 2020, the PM_2.5/PM₁₀ ratio increased slightly but remained below 0.5 overall. This can be attributed to the close relationship between anthropogenic activities and the sparse vegetation, dry climate, and numerous dust sources in eastern Xinjiang, where vehicle traffic not only contributes to fine PM emissions, but also causes the resuspension of dust. During the COVID-19 pandemic, the reduction in traffic volume may have led to a significant decrease in coarse PM concentrations in eastern Xinjiang. In contrast, the ratio generally decreased during the lockdown period in eastern Chinese cities due to reduced anthropogenic activities [11,43]. The PM_2.5/CO ratio is commonly used as an indicator of primary combustion emissions [44]. Its mean values in P1 and P3 2020 were 21.7% and 6.65% higher than in P2 2020, and the NO₂/CO ratio showed a similar temporal variation from P1 to P3 in summer 2020. A decrease in the PM_2.5/SO₂ ratio was observed along with a notable increase in the PM_2.5/NO₂ ratio in P2 2020, suggesting that despite the overall reduced anthropogenic activities, some industrial activities, particularly those with fixed emission sources such as coal combustion, may have persisted, resulting in continued SO₂ and primary PM emissions. This inference is also supported by the increased correlations between PM_2.5, PM₁₀, and CO in P2 2020 (Figure S2).

3.2.2. Regional Transport and Potential Contributing Sources

For the period P1 in 2020, the air masses mainly came from the west and southwest (Figure 6a). Clusters 1, 2, and 3 together accounted for 67.8% of the total, with the transport routes passing through Kazakhstan, the Mongolian Autonomous Prefecture of Bortala, Tacheng, and Changji. Cluster 5 (11.1%) was mainly associated with air masses from the Mongolian Autonomous Prefecture of Bayingolin. Cluster 4 (21.1%) had a relatively short transport distance, suggesting a significant contribution from local emissions. In period P2, the prevailing air mass paths generally matched those of P1 (Figure 6b). However, the influence of long-range transport became more evident. Clusters 1, 3, and 4 (70.0%) originated mainly from the northwest and crossed Kazakhstan, Tacheng, and Changji. Cluster 4 showed notably higher levels of PM_2.5 (18.6 μg/m³), PM₁₀ (69.3 μg/m³), and O₃ (93.8 μg/m³; Table S3) in P2 than in P1, indicating an increased contribution of regional transport to local air pollution. Meanwhile, the influence of local emissions seemed to decrease, as shown by the lower share of Cluster 2 (19.9%). Cluster 5 originated from the east, more precisely from the Alxa League in Inner Mongolia, and was transported to the receptor site via Jiuquan in Gansu province. In the P3 period, Clusters 1 and 2, which were characterized by relatively short transport distances, accounted for 58.6% of the total, suggesting a stronger local emission influence in this phase (Figure 6c). Clusters 5 and 6 (19.7%) came from the northwest and passed through Tacheng, Alxa, and Changji. Cluster 3 came from the east and followed a path through Alxa League and Jiuquan before reaching the receptor site.

The PSCF results showed that the high potential source region for PM_2.5 and PM₁₀ (WPSCF > 0.5) shrank in P2 compared to P1 2020 (Figure 7a,b,d,e). The high potential source regions were mainly located in Turpan and Hami in P1 and shifted towards Altay, Tacheng, and Changji, which is supported by the WCWT analysis in Figure S3. These results indicate lower local emissions and a greater influence of regional transport during the lockdown. Unlike PM, the high potential source region for NO₂ was mainly confined to the local area in all three periods (Figure 7g–i), which can be attributed to its relatively high reactivity [45,46].

Compared to other pollutants, the potential source region of O₃ exhibited a wider distribution over the three periods. The WPSCF analysis showed that the high potential source regions during the P1 period were mainly concentrated in Turpan and Hami, with some influence of long-range transport from areas such as northeastern Kazakhstan (Figure 7i). Compared to P1, the high potential source regions of O₃ in P2 were more dispersed and less concentrated (Figure 7k), indicating a greater influence of regional transport. During the P3 period, the potential source region became concentrated again in Altay, Tacheng, and Changji (Figure 7l).

4. Conclusions

This study investigates the impact of the COVID-19 pandemic on air pollution in eastern Xinjiang, the northwestern inland region of China, by analyzing the temporal changes of six major air pollutants (PM_2.5, PM₁₀, NO₂, SO₂, CO, and O₃) in the summer months of 2019–2021 and their potential source areas during the pandemic in 2020. A significant decrease in daily mean concentrations of PM_2.5, PM₁₀, NO₂, and CO was observed in P2 compared to the other phases in summer 2020. Compared to P2 in 2019 and 2021, the mean concentrations of the above four pollutants and AQI also decreased significantly in 2020. These results reflect the drastic decrease in anthropogenic activities during the pandemic. The significantly lower peak concentrations of PM_2.5, PM₁₀, NO₂, and CO in the morning and evening rush hours in P2 2020 compared to 2019 and 2021 illustrate the impact of reduced traffic-related emissions. In contrast to the cities in eastern China, O₃ concentrations in eastern Xinjiang fluctuated little during the P2 periods from 2019 to 2021, with a slight decrease in daily peak concentrations during the P2 period in 2020, likely due to lower NO_x levels, dry climate, and dust storms. The increase in SO₂ concentrations could be explained by ongoing industrial activities and higher electricity consumption during the lockdown period. The analysis of potential source areas showed that the contribution of regional sources for PM increased during the lockdown period, while the region with high source potential for NO₂ was always limited to the local area. O₃ exhibited a widespread regional source contribution during the summer months of 2020. Considering that O₃ was the main pollutant in eastern Xinjiang during the lockdown period and that its mean concentration and diurnal peak showed no significant reduction, further investigation into its formation mechanisms is needed, and strengthening regional cooperative control measures for precursor emissions should be a priority in the future.