Impact of SARS-CoV-2 on Ambient Air Quality in Northwest China (NWC)

: SARS-CoV-2 was discovered in Wuhan (Hubei) in late 2019 and covered the globe by March 2020. To prevent the spread of the SARS-CoV-2 outbreak, China imposed a countrywide lockdown that signiﬁcantly improved the air quality. To investigate the collective effect of SARS-CoV-2 on air quality, we analyzed the ambient air quality in ﬁve provinces of northwest China (NWC): Shaanxi (SN), Xinjiang (XJ), Gansu (GS), Ningxia (NX) and Qinghai (QH), from January 2019 to December 2020. For this purpose, ﬁne particulate matter (PM 2.5 ), coarse particulate matter (PM 10 ), sulfur dioxide (SO 2 ), nitrogen dioxide (NO 2 ), carbon monoxide (CO), and ozone (O 3 ) were obtained from the China National Environmental Monitoring Center (CNEMC). In 2020, PM 2.5 , PM 10 , SO 2 , NO 2 , CO, and O 3 improved by 2.72%, 5.31%, 7.93%, 8.40%, 8.47%, and 2.15%, respectively, as compared with 2019. The PM 2.5 failed to comply in SN and XJ; PM 10 failed to comply in SN, XJ, and NX with CAAQS Grade II standards (35 µ g/m 3 , 70 µ g/m 3 , annual mean). In a seasonal variation, all the pollutants experienced signiﬁcant spatial and temporal distribution, e.g., highest in winter and lowest in summer, except O 3 . Moreover, the average air quality index (AQI) improved by 4.70%, with the highest improvement in SN followed by QH, GS, XJ, and NX. AQI improved in all seasons; signiﬁcant improvement occurred in winter (December to February) and spring (March to May) when lockdowns, industrial closure etc. were at their peak. The proportion of air quality Class I improved by 32.14%, and the number of days with PM 2.5 , SO 2 , and NO 2 as primary pollutants decreased while they increased for PM 10 , CO, and O 3 in 2020. This study indicates a signiﬁcant association between air quality improvement and the prevalence of SARS-CoV-2 in 2020.

The SARS-CoV-2 also known as "COVID-19" was first discovered in December 2019 in Wuhan, a city in the Hubei province of China [23][24][25]. This deadly virus covered the globe within two months, and the World Health Organization (WHO) declared a pandemic on 11 March 2020 (WHO 2020). The common symptoms of COVID-19 include fever, dry cough, tiredness, and shortness of breath. In contrast, severe symptoms include a runny nose, sore throat, chills, muscle aches, headache, diarrhea, nausea, chest pain, breathing difficulties, and organ failure [26,27]. So far, COVID-19 has infected more than 133.10 million people and killed 2.90 million people worldwide (https://ourworldindata.org/ covid-cases/, accessed on 14 April 2021).
During the lockdown period (January 2020 to March 2020), most of the studies focused on air quality assessment in central China and nearby areas while ignoring rapidly developing areas of NWC [27,30,[34][35][36][37][38]. However, very few studies assessed the air quality in northwest China (the industrial and manufacturing hub of China) to evaluate the influence of the COVID-19 outbreak and associated lockdowns on air quality. A study conducted by He et al. [31] observed that the AQI (SO 2 , PM 2.5 , PM 10 , NO 2 , and CO) improved by 7.8% (6.76%, 5.93%, 13.66%, 24.67%, and 4.58%) in Lanzhou (Gansu province) during the lockdown. Similarly, [39] observed that PM 1 decreased by 50% during the Lanzhou lockdown period (January to March). Depending on the spread of the viral outbreak, lockdowns, travel restrictions, etc. were extended and re-imposed in some areas, e.g., Kashgar city (October 2020). No study has extensively analyzed the influence of the viral outbreak and lockdowns on air quality in northwest China.
In this study, we assessed the spatial and temporal variation of air pollution in 53 cities of NWC. We examined six criteria pollutants (PM 2.5 , PM 10 , SO 2 , NO 2 , CO, and O 3 ), the air quality index (AQI), the proportion of AQI classes, and major pollutants etc. for a period of two years (January 2019 to December 2020) to illustrate the impact of SARS-CoV-2 and associated lockdowns on the spatial and temporal distribution of air pollution in NWC. We believe this is the first study focusing on the long-term impact of the COVID-19 outbreak on air quality in 53 cities of NWC and is of considerable significance to environmental protection and human health.

Data Collection
The hourly concentration of six criteria pollutants (PM 2.5 , PM 10 , SO 2 , NO 2 , CO, and O 3 ) in NWC was obtained from the China National Environmental Monitoring Center (CNEMC) for a period of two years (2019-2020). The online data-sharing platform had installed hundreds of ambient air monitoring stations, covering both urban and rural areas in 337 cities of China, and published information according to the Technical Guideline on Environmental Monitoring Quality Management (HJ 630-2011).

Air Quality Index (AQI)
The air quality index (AQI) is a color-coded scale that simplifies different pollutants' concentrations into a single numerical value to reflect overall air quality, health effects, sensitive groups, and required precautionary measures. The AQI includes 24-h average measurements of NO 2 , SO 2 , CO, PM2.5, PM10, and 8-h average concentrations of O 3 [12,13,40]. When the AQI is higher than 50, the highest Individual air quality index (IAQI) is considered a primary pollutant for that given day [13,30,[41][42][43]. The individual air quality index (IAQI) for six criteria pollutants is determined by using Equation (1), and the overall AQI is calculated based on the highest IAQI by using Equation (2) according to the instruction given in the technical regulation on ambient air quality index (on trial) (HJ-633-2012).
IAQI P = I high -I low /C high -C low * (C P -C low ) + I low (1) IAQI P = Individual sub air quality index of the pollutant p C P = concentration of the pollutant p C high = concentration breakpoint that is ≥ C p C low = concentration breakpoint that is ≤ C p I high = index breakpoint corresponding to C high I low = index breakpoint corresponding to C low AQI = max (I 1 , I 2 , . . . , In) AQI has the following six categories: Class I; 0-50 (green), good; Class II: 51-100 (yellow), moderate; Class III: 101-150 (orange), unhealthy for sensitive groups; Class IV: 151-200 (red), unhealthy; Class V: 201-300 (purple), very unhealthy; Class VI: 300-500 (maroon), hazardous.

Quality Assurance and Quality Control (QA&AR)
Quality assurance and control procedures for ambient air quality data were strictly by Chinese Ambient Air Quality Standards (CAAQS) (GB 3095 2012). The daily average value was calculated when we had valid data for more than 16 h of that day (except for O 3 , minimum 6-h values for 8-h O 3 value). The monthly average was calculated when we had 27 daily mean values; an annual value was calculated when we had 324 daily mean values. Besides this, manual inspection was carried out to remove abnormal values e.g., PM 2.5 values higher than PM 10 values.

Kriging (Ordinary/Universal)
Kriging is a geospatial interpolation technique that defines the unknown values depending on the available known values and considers both the distance and the degree of variation between known data points when estimating values in unknown areas. We used kriging (ordinary) to evaluate the spatial distribution of criteria pollutants (PM 2.5 , PM 10 , SO 2 , NO 2 , CO, O 3 ), AQI, etc., in NWC, and then applied reclassification to obtain the desired map format.

Statistical Analysis
In this study, we used the Statistical Package for Social Sciences (SPSS) for Windows (IBM SPSS Statistics, Version 25) to evaluate Pearson's correlation coefficient for criteria pollutants on an annual and seasonal basis. Pearson's correlation is a correlation coefficient commonly used in linear regression and used to measure the strength of relationships between the six air pollutants. The effect of a certain variable was considered statistically significant for P (0.01 and 0.05) (two tailed). Annual mean values of data were used for the analysis of six criteria pollutants between 2019 to 2020: mean absolute deviation (MAD), mean square error (MSE), root mean square error (RMSE), mean absolute percentage error (MAPE), and mean percentage error (MPE), and were calculated by Excel 2016.

Annual and Seasonal Changes in Criteria Pollutants
During the study period (2019-2020), the annual average concentration of PM 2.5 , PM 10 , SO 2 , NO 2 , CO, and O 3 (Table S1)

PM 2.5 /PM 10 Ratio
During the study period (2019-2020), the average PM 2.5 /PM 10 ratio in NWC ranged from 0.325 ± 0.135 to 0.640 ± 0.190 with an average of 0.472 ± 0.100, highest PM 2.5 /PM 10 ratio occurring in SN followed by QH, NX, GS, and XJ, and experiencing an average change of 2.56% in NWC in 2020 as compared with 2019 ( Figure 2h). In a seasonal variation, the highest PM 2.5 /PM 10 ratio occurred in winter followed by autumn, summer, and spring, respectively, and increased by 10.68% in winter in NWC during 2020 (Figure 4h). In 2020, 62.26% of cities observed an increase in the PM 2.5 /PM 10 ratio. Similarly, spring, summer, autumn and winter experienced an improvement in 33.96%, 52.83%, 54.72%, and 90.57% cities of NWC, respectively, in 2020 against 2019 ( Figure 5).

Air Quality Index (AQI)
During the study period (2019-2020), the average AQI in NWC ranged from 43.34 ± 10.15 to 194 ± 210.26 with an average of 79.65 ± 19.90, and the highest AQI occurred in XJ followed by SN, GS, NX, and QH. The AQI improved by 4.67% (10.26%, 2.25%, 2.73%, 0.31%, 9.74%) in NWC (SN, XJ, GS, NX, QH) during 2020 as compared with 2019 ( Figure 2g). In a seasonal variation, the highest AQI occurred in winter followed by spring, summer, and autumn, respectively, and improved by 6.69%, 2.91%, 8.57%, and 1.59%, respectively, in NWC during 2020 ( Figure 4g). Seasonal variation was not consistent throughout NWC e.g., SN and XJ experienced the highest AQI in winter, GS and NX in spring. At the same time, QH observed the highest AQI in summer. Figure 6 illustrates the annual and seasonal spatial distribution of the AQI in NWC. In 2020, 77.36% cities experienced an improvement in the AQI. Similarly, spring, summer, autumn and winter experienced an improvement of 83.02%, 13.21%, 52.83%, and 62.26% in the cities of NWC, respectively, in 2020 against 2019. Significant improvement in the AQI occurred in winter (December to February) and spring (March to April) when the viral outbreak, lockdown, and movement restrictions were at their peak.      Figure 7 explains the annual and seasonal (spring, summer, autumn, and winter) proportion of different AQI classes in NWC during 2019 and 2020. In 2020, the proportion of AQI "Class I", and "Class II" improved by 32.14%, and 4%, respectively, while they decreased by 9.13%, 21.35%, and 18.41% for "Class III", "Class IV", and "Class V", respectively, in NWC. In the seasonal variation, the proportion of Class I increased by 22.42%, 50.13%, and 41.95% in spring, summer, and winter, respectively, in NWC during 2020. In the case of monthly variation, the combined proportion of Class I & II was higher in summer (June, July, August), indicating better air quality compared with other seasons ( Figure S1).
In the case of seasonal variation, PM 10 was a major pollutant in spring (43.57%) and autumn (39.32%), PM 2.5 was a major pollutant in winter (46.98%), and O 3 was a major pollutant (67.31%) in summer. The number of days with PM 2.5 as a major pollutant decreased by 37.5% in spring, while they increased by 33.33%, 47.15%, and 17.65% in summer, autumn, and winter 2020, respectively. Similarly, the number of days with PM 10 as major pollutant increased by 0.61%, 21.33%, and 8.75% in spring, summer, and autumn 2020, respectively, while they decreased by 32.31% in winter 2020 as compared with 2019. The number of days with SO 2 , NO 2 and O 3 as major pollutant decreased in all seasons while increasing slightly for O3 in spring 2020 compared with 2019. The number of days with O 3 as a major pollutant was higher in the hotter months (April-September), and PM 2.5 was higher in colder months (November-February). PM 10 experienced a "U" shaped curve with higher concentrations in winter, spring, and autumn, and lower concentrations in summer ( Figure S2).
Generally, PM 2.5 , PM 10 (except XJ in 2020, GS (2019, 2020), NX (2019), QH (2020)), SO 2 , NO 2 , and CO observed the same seasonal variation, e.g., highest in winter and lowest in summer. Higher pollution in winter is associated with increased coal combustion, civil heating, power generation, fossil fuel burning, industrial activity, vehicular exhausts, and stagnant meteorology [13,44,46,51,63]. The exception was that, the concentration of PM 10 was higher in XJ during spring 2020 than in 2019, which indicates the ongoing influence of natural sources, e.g., sand storms, deserts, mineral dust, etc. [52][53][54][55]. In contrast to other pollutants, the concentration of O 3 was higher in summer than winter [13,46,51,64] due to lower NOx levels in winter, as NOx levels decrease the O 3 depletion and enhance the accumulation of O 3 , as well as higher temperatures favor ozone production [23,27,30,36,37,[65][66][67][68][69][70]. Lockdown caused minor changes in seasonal variation instead of major changes because the lockdown pattern was not regular and largely varied spatially and temporally. PM 2.5 /PM 10 reflect air quality, pollution sources and origin e.g., a higher PM 2.5 /PM 10 ratio indicates an increased proportion of PM 2.5 mainly emitted from anthropogenic activities [52,53,55]. In 2020, the PM 2.5 /PM 10 ratio increased by 2.56%, and 60.38% of cities experienced an increase in the PM 2.5 /PM 10 ratio in NWC. This increase is associated with a minor reduction in PM 2.5 as compared with PM 10 in 2020. As the share of PM 2.5 increases, the PM 2.5 /PM 10 ratio also increases [13]. In general, the PM 2.5 /PM 10 ratio was higher in winter (low temperature), compared to summer (high temperature) due to increased PM 2.5 emissions from coal combustion, civil heating, biomass burning, industrial emissions, and stable atmospheric conditions helped with stagnation and accumulation of pollution [13,[71][72][73]. In 2020, the PM 2.5 /PM 10 ratio decreased in spring and summer, while it increased in autumn and winter due to an increase in PM 2.5 and minor reduction of PM 2.5 compared with PM 10 .
In 2020, the number of days with PM 10 and O 3 as major pollutants increased, while the number of days with PM 2.5 , SO 2 , and NO 2 as major pollutants decreased in NWC concerning 2019. Due to limited anthropogenic activity e.g., industrial closures, traffic stagnation, movement restrictions etc., the concentration of PM 2.5 , SO 2 , and NO 2 decreased significantly [23,27,30,31,36,38]. PM 2.5 was a major pollutant in winter, indicating anthropogenic emissions, e.g., civil heating, industrial emissions, etc. [13,47,48,71]. PM 10 was a major pollutant in spring and autumn, while O 3 was a major pollutant in summer as higher temperatures favor ozone accumulation [35,[64][65][66][67][68][69]. In the southern part of XJ (Kashgar), the number of days with PM 10 as a major pollutant was higher due to emissions from natural sources, e.g., Taklimakan desert, sand storms [52][53][54][55]75]. Any day with one or more pollutants exceeding CAAQS (Grade II) standards is considered as a non-attainment/pollution day [27,30]. During 2020, the number of pollution days decreased by 13.59% in NWC, which indicates improved air quality associated with a reduction in anthropogenic activities. Similarly, the number of pollution days decreased by 15.41%, 26.79%, 9.66%, and 11.55% in spring, summer, autumn, and winter, respectively, which indicates that the ambient air quality was improved significantly in 2020. A strong correlation between all the criteria pollutants indicates mutual emission sources. PM 2.5, mainly originates from anthropogenic activity e.g., fossil fuels, developmental activity, industrial activity, etc. [7,8,11,15]. Such activities also contribute to SO, NO 2 , CO and PM 10 . Multiple studies have concluded that coal burning is a major source of PM 2.5 [8,76] and a major source of SO 2 ; vehicular emissions are a major source of PM 2.5 [49,77,78] and release NO 2 and CO as well. Similarly, road dust mainly consists of PM 10 , which contributes a significant share to PM 2.5 [10,18,49,77,79].
In short, ongoing rapid economic development, industrialization, urbanization, motorization, natural events, and adverse meteorology have deteriorated ambient air quality in NWC [2,9,16,18,[80][81][82][83]. The SARS-CoV-2 proved to be a blessing in disguise as the associated lockdowns put in place to prevent the spread of the viral outbreak resulted in a significant reduction in PM 2.5 , PM 10 , SO 2 , NO 2 , CO, O 3 , AQI, and the number of pollution days in NWC decreased during 2020 compared to 2019.

Conclusions
This study collected hourly monitoring data of ambient air pollutants from 53 cities located in five provinces of northwest China (NWC) from January 2019 to December 2020 to show the collective effect SARS-CoV-2 had on ambient air quality in 2020 as compared with 2019. The results showed that the average concentrations of PM 2.5 , PM 10 , SO 2 , NO 2 , CO, and O 3 improved by 2.72%, 5.31%, 7.93%, 8.40%, 8.47%, and 2.15% in NWC, respectively, during 2020. The annual average concentration of PM 2.5 failed to comply in SN, XJ, and NW; PM 10 failed to comply in SN, XJ, NX, and NW, while SO 2 and NO 2 complied with CAAQS Grade II standards (35 µg/m 3 , 70 µg/m 3 , 60 µg/m 3 and 40 µg/m 3 , annual mean) in SN, XJ, GS, NX, QH, and NWC. All the pollutants experienced their highest pollution level in winter except ozone, with varying degrees of spatial distribution. The AQI improved by 4.67% (higher in cities with low population, with some exceptions) in NWC and experienced the highest improvement in SN followed by QH, GS, XJ, and NX. The AQI improved in all seasons with the highest increase in summer, followed by winter, spring and autumn. Significant improvements in the AQI occurred in winter (December to February) and spring (March to April) when lockdowns, industrial closures etc. were at their peak. In NWC, O 3 was a major pollutant followed by PM 2.5 , PM 10 , NO 2 , SO 2 , and, CO with different spatial and temporal variations, e.g., PM 2.5 in winter, PM 10 in autumn and spring, and O 3 in summer. A strong correlation occurred between all pollutants except O 3 . This paper comprehensively discussed the impact of SARS-CoV-2, and associated lockdowns on air pollution in NWC and calls for future detailed assessment focusing on health risk assessment and the impact of meteorology, etc.

Data Availability Statement:
The data presented in this study are available upon request from the corresponding author.